revert export of CIFPP_SHARE_DIR variable

exclude from all for gxrio

.gitignore

@@ -1,14 +1,9 @@
 build/
 .vscode/
 .vs/
-.pc/
-tools/symop-map-generator
-test/unit-test
-test/pdb2cif-test
-test/rename-compound-test
 tools/update-libcifpp-data
 data/components.cif*
 CMakeSettings.json
 msvc/
-Testing/
-
+src/revision.hpp
+test/test-create_sugar_?.cif

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "gxrio"]
+	path = gxrio
+	url = https://github.com/mhekkel/gxrio.git

.travis.yml

@@ -1,33 +0,0 @@
-language: cpp
-
-os:
-  - linux
-  - osx
-
-dist: focal
-
-osx_image:
-  - xcode12
-
-compiler:
-  - gcc
-  - clang
-
-addons:
-  apt:
-    packages:
-      - libboost-all-dev
-
-before_install:
-  - if [ "$TRAVIS_OS_NAME" = "osx" ]; then brew install make; fi
-
-script:
-  - if [ "$TRAVIS_OS_NAME" = "osx" ]; then ./configure --disable-shared --disable-revision --disable-download-ccd ; else ./configure --disable-revision --disable-download-ccd ; fi
-  - if [ "$TRAVIS_OS_NAME" = "osx" ]; then gmake                        ; else make             ; fi
-  - if [ "$TRAVIS_OS_NAME" = "osx" ]; then gmake test                   ; else make test        ; fi
-  - if [ "$TRAVIS_OS_NAME" = "osx" ]; then sudo gmake install           ; else sudo make install; fi
-
-# jobs:
-#   allow_failures:
-#     - os: osx
-

CMakeLists.txt

@@ -6,10 +6,10 @@
 # modification, are permitted provided that the following conditions are met:
 
 # 1. Redistributions of source code must retain the above copyright notice, this
-#    list of conditions and the following disclaimer
+# list of conditions and the following disclaimer
 # 2. Redistributions in binary form must reproduce the above copyright notice,
-#    this list of conditions and the following disclaimer in the documentation
-#    and/or other materials provided with the distribution.
+# this list of conditions and the following disclaimer in the documentation
+# and/or other materials provided with the distribution.
 
 # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
@@ -25,32 +25,25 @@
 cmake_minimum_required(VERSION 3.16)
 
 # set the project name
-project(cifpp VERSION 2.0.3 LANGUAGES CXX)
+project(cifpp VERSION 5.0.3.1 LANGUAGES CXX)
 
 list(PREPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
 
-enable_testing()
-
 include(GNUInstallDirs)
 include(CheckFunctionExists)
 include(CheckIncludeFiles)
 include(CheckLibraryExists)
 include(CMakePackageConfigHelpers)
-include(Dart)
-include(FindFilesystem)
-include(GenerateExportHeader)
+include(CheckCXXSourceCompiles)
 
 set(CXX_EXTENSIONS OFF)
-set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_STANDARD 20)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)
 
-find_package(Filesystem REQUIRED)
-
 if("${CMAKE_CXX_COMPILER_ID}" STREQUAL "GNU")
-	# https://stackoverflow.com/questions/63902528/program-crashes-when-filesystempath-is-destroyed
 	set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Wextra -Wno-unused-parameter -Wno-missing-field-initializers")
 elseif(MSVC)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /W4")
+	set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /W4")
 endif()
 
 # Building shared libraries?
@@ -61,6 +54,31 @@ set(CMAKE_WINDOWS_EXPORT_ALL_SYMBOLS ON)
 
 # Optionally build a version to be installed inside CCP4
 option(BUILD_FOR_CCP4 "Build a version to be installed in CCP4" OFF)
+
+# Lots of code depend on the availability of the components.cif file
+option(CIFPP_DOWNLOAD_CCD "Download the CCD file components.cif during installation" ON)
+
+# An optional cron script can be installed to keep the data files up-to-date
+if(UNIX)
+	option(CIFPP_INSTALL_UPDATE_SCRIPT "Install the script to update CCD and dictionary files" ON)
+endif()
+
+# When CCP4 is sourced in the environment, we can recreate the symmetry operations table
+if(EXISTS "$ENV{CCP4}")
+	if(EXISTS "$ENV{CLIBD}/syminfo.lib")
+		option(CIFPP_RECREATE_SYMOP_DATA "Recreate SymOp data table in case it is out of date" ON)
+	else()
+		set(CIFPP_RECREATE_SYMOP_DATA OFF)
+		message(WARNING "Symop data table recreation requested, but file syminfo.lib was not found in $ENV{CLIBD}")
+	endif()
+else()
+	set(CIFPP_RECREATE_SYMOP_DATA OFF)
+	message("Not trying to recreate symop_table_data.hpp since CCP4 is not defined")
+endif()
+
+# Unit tests
+option(ENABLE_TESTING "Build test exectuables" OFF)
+
 if(BUILD_FOR_CCP4)
 	if("$ENV{CCP4}" STREQUAL "" OR NOT EXISTS $ENV{CCP4})
 		message(FATAL_ERROR "A CCP4 built was requested but CCP4 was not sourced")
@@ -69,43 +87,18 @@ if(BUILD_FOR_CCP4)
 		list(APPEND CMAKE_PREFIX_PATH "$ENV{CCP4}")
 		set(CMAKE_INSTALL_PREFIX "$ENV{CCP4}")
 
-		# This is the only option:
 		if(WIN32)
 			set(BUILD_SHARED_LIBS ON)
 		endif()
-	endif("$ENV{CCP4}" STREQUAL "" OR NOT EXISTS $ENV{CCP4})
-endif()
-
-# Check if CCP4 is available
-if(EXISTS "$ENV{CCP4}")
-	set(CCP4 $ENV{CCP4})
-	set(CLIBD ${CCP4}/lib/data)
-endif()
-if(CCP4 AND NOT CLIBD)
-	set(CLIBD ${CCP4}/lib/data)
-endif()
-
-# When CCP4 is sourced in the environment, we can recreate the symmetry operations table
-if(EXISTS "${CCP4}")
-	if(CIFPP_RECREATE_SYMOP_DATA AND NOT EXISTS "${CLIBD}/syminfo.lib")
-		message(WARNING "Symop data table recreation requested, but file syminfo.lib was not found in ${CLIBD}")
-		set(CIFPP_RECREATE_SYMOP_DATA OFF)
-	else()
-		option(CIFPP_RECREATE_SYMOP_DATA "Recreate SymOp data table in case it is out of date" ON)
 	endif()
-else()
-	set(CIFPP_RECREATE_SYMOP_DATA OFF)
-	message("Not trying to recreate SymOpTable_data.hpp since CCP4 is not defined")
 endif()
 
-# set(CMAKE_DEBUG_POSTFIX d)
-
 if(MSVC)
-    # make msvc standards compliant...
-    add_compile_options(/permissive-)
+	# make msvc standards compliant...
+	add_compile_options(/permissive-)
 
 	macro(get_WIN32_WINNT version)
-		if (WIN32 AND CMAKE_SYSTEM_VERSION)
+		if(WIN32 AND CMAKE_SYSTEM_VERSION)
 			set(ver ${CMAKE_SYSTEM_VERSION})
 			string(REPLACE "." "" ver ${ver})
 			string(REGEX REPLACE "([0-9])" "0\\1" ver ${ver})
@@ -116,185 +109,176 @@ if(MSVC)
 
 	get_WIN32_WINNT(ver)
 	add_definitions(-D_WIN32_WINNT=${ver})
-
-	# On Windows, do not install in the system location
-	if(CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT AND NOT BUILD_FOR_CCP4)
-		message(STATUS "The library and auxiliary files will be installed in $ENV{LOCALAPPDATA}/${PROJECT_NAME}")
-		set(CMAKE_INSTALL_PREFIX "$ENV{LOCALAPPDATA}/${PROJECT_NAME}" CACHE PATH "..." FORCE)
-	endif()
-
-	# Find out the processor type for the target
-	if(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "AMD64")
-		set(COFF_TYPE "x64")
-	elseif(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "i386")
-		set(COFF_TYPE "x86")
-	elseif(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "ARM64")
-		set(COFF_TYPE "arm64")
-	else()
-		message(FATAL_ERROR "Unsupported or unknown processor type ${CMAKE_SYSTEM_PROCESSOR}")
-	endif()	
-
-	set(COFF_SPEC "--coff=${COFF_TYPE}")
-
-	# for mrc, just in case
-	list(APPEND CMAKE_PREFIX_PATH "$ENV{LOCALAPPDATA}/mrc")
-endif()
-
-if(UNIX AND NOT APPLE AND NOT BUILD_FOR_CCP4 AND CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT)
-	# On Linux, install in the $HOME/.local folder by default
-	message(STATUS "The library and auxiliary files will be installed in $ENV{HOME}/.local")
-	set(CMAKE_INSTALL_PREFIX "$ENV{HOME}/.local" CACHE PATH "..." FORCE)
-endif()
-
-# Optionally use mrc to create resources
-
-if(WIN32 AND BUILD_SHARED_LIBS)
-	message("Not using resources when building shared libraries for Windows")
-else()
-	find_program(MRC mrc)
-
-	if(MRC)
-		option(CIFPP_USE_RSRC "Use mrc to create resources" ON)
-	else()
-		message("Using resources not possible since mrc was not found")
-	endif()
-
-	if(CIFPP_USE_RSRC STREQUAL "ON")
-		set(CIFPP_USE_RSRC 1)
-
-		message("Using resources compiled with ${MRC}")
-		add_compile_definitions(CIFPP_USE_RSRC)
-	endif()
 endif()
 
 # Libraries
 
+# Start by finding out if std:regex is usable. Note that the current
+# implementation in GCC is not acceptable, it crashes on long lines.
+# The implementation in libc++ (clang) and MSVC seem to be OK.
+check_cxx_source_compiles("
+#include <iostream>
+#ifndef __GLIBCXX__
+#error
+#endif
+int main(int argc, char *argv[]) { return 0; }" GXX_LIBSTDCPP)
+
+if(GXX_LIBSTDCPP)
+	message(STATUS "Testing for known regex bug, since you're using GNU libstdc++")
+
+	try_run(STD_REGEX_RUNNING STD_REGEX_COMPILING
+		${CMAKE_CURRENT_BINARY_DIR}/test ${PROJECT_SOURCE_DIR}/cmake/test-rx.cpp)
+
+	if(STD_REGEX_RUNNING STREQUAL FAILED_TO_RUN)
+		message(STATUS "You are probably trying to compile using the g++ standard library which contains a crashing std::regex implementation. Will use boost::regex instead")
+
+		add_subdirectory(regex EXCLUDE_FROM_ALL)
+		set(BOOST_REGEX ON)
+	endif()
+endif()
+
 set(CMAKE_THREAD_PREFER_PTHREAD)
 set(THREADS_PREFER_PTHREAD_FLAG)
 find_package(Threads)
 
-set(Boost_DETAILED_FAILURE_MSG ON)
-if(NOT BUILD_SHARED_LIBS)
-	set(Boost_USE_STATIC_LIBS ON)
-endif()
-find_package(Boost 1.70.0 REQUIRED COMPONENTS system iostreams regex program_options)
+if(MSVC)
+	# Avoid linking the shared library of zlib
+	# Search ZLIB_ROOT first if it is set.
+	if(ZLIB_ROOT)
+		set(_ZLIB_SEARCH_ROOT PATHS ${ZLIB_ROOT} NO_DEFAULT_PATH)
+		list(APPEND _ZLIB_SEARCHES _ZLIB_SEARCH_ROOT)
+	endif()
 
-if(NOT MSVC AND Boost_USE_STATIC_LIBS)
-	find_package(ZLIB REQUIRED)
-	find_package(BZip2 REQUIRED)
+	# Normal search.
+	set(_ZLIB_x86 "(x86)")
+	set(_ZLIB_SEARCH_NORMAL
+		PATHS "[HKEY_LOCAL_MACHINE\\SOFTWARE\\GnuWin32\\Zlib;InstallPath]"
+			"$ENV{ProgramFiles}/zlib"
+			"$ENV{ProgramFiles${_ZLIB_x86}}/zlib")
+	unset(_ZLIB_x86)
+	list(APPEND _ZLIB_SEARCHES _ZLIB_SEARCH_NORMAL)
+
+	foreach(search ${_ZLIB_SEARCHES})
+    	find_library(ZLIB_LIBRARY NAMES zlibstatic NAMES_PER_DIR ${${search}} PATH_SUFFIXES lib)
+	endforeach()
 endif()
+find_package(ZLIB REQUIRED)
+
+include(FindFilesystem)
+list(APPEND CIFPP_REQUIRED_LIBRARIES ${STDCPPFS_LIBRARY})
+
+include(FindAtomic)
+list(APPEND CIFPP_REQUIRED_LIBRARIES ${STDCPPATOMIC_LIBRARY})
 
 # Create a revision file, containing the current git version info
-
-find_package(Git)
-if(GIT_FOUND AND EXISTS "${CMAKE_SOURCE_DIR}/.git")
-	include(GetGitRevisionDescription)
-    get_git_head_revision(REFSPEC COMMITHASH)
-
-	# Generate our own version string
-	git_describe_working_tree(BUILD_VERSION_STRING --match=build --dirty)
-else()
-    message(WARNING "Git not found, cannot set version info")
-
-    SET(BUILD_VERSION_STRING ${PROJECT_VERSION})
-endif()
-
-# generate version.h
-string(TIMESTAMP BUILD_DATE_TIME "%Y-%m-%dT%H:%M:%SZ" UTC)
-configure_file("${CMAKE_SOURCE_DIR}/src/revision.hpp.in" "${CMAKE_BINARY_DIR}/revision.hpp" @ONLY)
+include(VersionString)
+write_version_header(${PROJECT_SOURCE_DIR}/src/ "LibCIFPP")
 
 # SymOp data table
 if(CIFPP_RECREATE_SYMOP_DATA)
 	# The tool to create the table
+	add_executable(symop-map-generator "${PROJECT_SOURCE_DIR}/src/symop-map-generator.cpp")
 
-	add_executable(symop-map-generator "${CMAKE_SOURCE_DIR}/tools/symop-map-generator.cpp")
-
-	target_link_libraries(symop-map-generator Threads::Threads ${Boost_LIBRARIES} std::filesystem ${ZLIB_LIBRARIES} ${BZip2_LIBRARIES})
-	if(Boost_INCLUDE_DIR)
-		target_include_directories(symop-map-generator PUBLIC ${Boost_INCLUDE_DIR})
-	endif()
-
-	set($ENV{CLIBD} ${CLIBD})
-	
 	add_custom_command(
-		OUTPUT ${CMAKE_SOURCE_DIR}/src/SymOpTable_data.hpp
-		COMMAND $<TARGET_FILE:symop-map-generator> ${CLIBD}/syminfo.lib ${CMAKE_SOURCE_DIR}/src/SymOpTable_data.hpp
-		)
-	
+		OUTPUT ${PROJECT_SOURCE_DIR}/src/symop_table_data.hpp
+		COMMAND $<TARGET_FILE:symop-map-generator> $ENV{CLIBD}/syminfo.lib ${PROJECT_SOURCE_DIR}/src/symop_table_data.hpp
+	)
+
 	add_custom_target(
-		OUTPUT ${CMAKE_SOURCE_DIR}/src/SymOpTable_data.hpp
+		OUTPUT ${PROJECT_SOURCE_DIR}/src/symop_table_data.hpp
 		DEPENDS symop-map-generator "$ENV{CLIBD}/syminfo.lib"
 	)
 endif()
 
 # Sources
+set(project_sources
+	${PROJECT_SOURCE_DIR}/src/category.cpp
+	${PROJECT_SOURCE_DIR}/src/condition.cpp
+	${PROJECT_SOURCE_DIR}/src/datablock.cpp
+	${PROJECT_SOURCE_DIR}/src/dictionary_parser.cpp
+	${PROJECT_SOURCE_DIR}/src/file.cpp
+	${PROJECT_SOURCE_DIR}/src/item.cpp
+	${PROJECT_SOURCE_DIR}/src/parser.cpp
+	${PROJECT_SOURCE_DIR}/src/row.cpp
+	${PROJECT_SOURCE_DIR}/src/validate.cpp
+	${PROJECT_SOURCE_DIR}/src/text.cpp
+	${PROJECT_SOURCE_DIR}/src/utilities.cpp
 
-set(project_sources 
-	${PROJECT_SOURCE_DIR}/src/AtomType.cpp
-	${PROJECT_SOURCE_DIR}/src/BondMap.cpp
-	${PROJECT_SOURCE_DIR}/src/Cif++.cpp
-	${PROJECT_SOURCE_DIR}/src/Cif2PDB.cpp
-	${PROJECT_SOURCE_DIR}/src/CifParser.cpp
-	${PROJECT_SOURCE_DIR}/src/CifUtils.cpp
-	${PROJECT_SOURCE_DIR}/src/CifValidator.cpp
-	${PROJECT_SOURCE_DIR}/src/Compound.cpp
-	${PROJECT_SOURCE_DIR}/src/PDB2Cif.cpp
-	${PROJECT_SOURCE_DIR}/src/PDB2CifRemark3.cpp
-	${PROJECT_SOURCE_DIR}/src/Point.cpp
-	${PROJECT_SOURCE_DIR}/src/Secondary.cpp
-	${PROJECT_SOURCE_DIR}/src/Structure.cpp
-	${PROJECT_SOURCE_DIR}/src/Symmetry.cpp
-	${PROJECT_SOURCE_DIR}/src/TlsParser.cpp
+	${PROJECT_SOURCE_DIR}/src/atom_type.cpp
+	${PROJECT_SOURCE_DIR}/src/compound.cpp
+	${PROJECT_SOURCE_DIR}/src/point.cpp
+	${PROJECT_SOURCE_DIR}/src/symmetry.cpp
+
+	${PROJECT_SOURCE_DIR}/src/model.cpp
+
+	${PROJECT_SOURCE_DIR}/src/pdb/cif2pdb.cpp
+	${PROJECT_SOURCE_DIR}/src/pdb/pdb2cif.cpp
+	${PROJECT_SOURCE_DIR}/src/pdb/pdb2cif_remark_3.cpp
+	${PROJECT_SOURCE_DIR}/src/pdb/tls.cpp
 )
 
-set(project_headers 
-	${PROJECT_SOURCE_DIR}/include/cif++/AtomType.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/BondMap.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Cif++.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Cif2PDB.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/CifParser.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/CifUtils.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/CifValidator.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Compound.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Matrix.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/PDB2Cif.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/PDB2CifRemark3.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Point.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Secondary.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Structure.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/Symmetry.hpp
-	${PROJECT_SOURCE_DIR}/include/cif++/TlsParser.hpp
+set(project_headers
+	${PROJECT_SOURCE_DIR}/include/cif++.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/utilities.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/item.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/datablock.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/file.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/validate.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/list.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/iterator.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/parser.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/forward_decl.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/dictionary_parser.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/condition.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/category.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/row.hpp
+
+	${PROJECT_SOURCE_DIR}/include/cif++/atom_type.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/compound.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/point.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/symmetry.hpp
+
+	${PROJECT_SOURCE_DIR}/include/cif++/model.hpp
+
+	${PROJECT_SOURCE_DIR}/include/cif++/pdb/cif2pdb.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/pdb/io.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/pdb/pdb2cif.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/pdb/pdb2cif_remark_3.hpp
+	${PROJECT_SOURCE_DIR}/include/cif++/pdb/tls.hpp
 )
 
-add_library(cifpp ${project_sources} ${project_headers} ${CMAKE_SOURCE_DIR}/src/SymOpTable_data.hpp)
+add_library(cifpp ${project_sources} ${project_headers} ${PROJECT_SOURCE_DIR}/src/symop_table_data.hpp)
+add_library(cifpp::cifpp ALIAS cifpp)
+
+if(BOOST_REGEX)
+	target_compile_definitions(cifpp PRIVATE USE_BOOST_REGEX=1 BOOST_REGEX_STANDALONE=1)
+	target_include_directories(cifpp PRIVATE regex/include)
+endif()
+
+if(MSVC)
+	target_compile_definitions(cifpp PUBLIC NOMINMAX=1)
+endif()
+
 set_target_properties(cifpp PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 target_include_directories(cifpp
 	PUBLIC
 	"$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/include>"
 	"$<INSTALL_INTERFACE:${CMAKE_INSTALL_INCLUDEDIR}>"
-	${Boost_INCLUDE_DIR}
 )
 
-target_include_directories(cifpp
-	PRIVATE
-	${CMAKE_BINARY_DIR}
-)
+target_link_libraries(cifpp PUBLIC Threads::Threads ZLIB::ZLIB ${CIFPP_REQUIRED_LIBRARIES})
 
-target_link_libraries(cifpp Threads::Threads ${Boost_LIBRARIES} std::filesystem ${ZLIB_LIBRARIES} ${BZip2_LIBRARIES})
+if(CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
+	target_link_options(cifpp PRIVATE -undefined dynamic_lookup)
+endif(CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
 
-if (CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
-    target_link_options(cifpp PRIVATE -undefined dynamic_lookup)
-endif (CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
-
-option(CIFPP_DOWNLOAD_CCD "Download the CCD file components.cif during installation" OFF)
 if(CIFPP_DOWNLOAD_CCD)
 	# download the components.cif file from CCD
 	set(COMPONENTS_CIF ${PROJECT_SOURCE_DIR}/data/components.cif)
 
-	if (NOT EXISTS ${COMPONENTS_CIF})
-
-		if (NOT EXISTS ${PROJECT_SOURCE_DIR}/data)
+	if(NOT EXISTS ${COMPONENTS_CIF})
+		if(NOT EXISTS ${PROJECT_SOURCE_DIR}/data)
 			file(MAKE_DIRECTORY ${PROJECT_SOURCE_DIR}/data/)
 		endif()
 
@@ -305,7 +289,7 @@ if(CIFPP_DOWNLOAD_CCD)
 				SHOW_PROGRESS)
 			add_custom_command(OUTPUT ${COMPONENTS_CIF}
 				COMMAND ${GUNZIP} ${COMPONENTS_CIF}.gz
-				WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}/data/)
+				WORKING_DIRECTORY ${PROJECT_SOURCE_DIR}/data/)
 		else()
 			file(DOWNLOAD ftp://ftp.wwpdb.org/pub/pdb/data/monomers/components.cif ${COMPONENTS_CIF}
 				SHOW_PROGRESS)
@@ -315,25 +299,18 @@ if(CIFPP_DOWNLOAD_CCD)
 	add_custom_target(COMPONENTS ALL DEPENDS ${COMPONENTS_CIF})
 endif()
 
-if(UNIX)
-	option(CIFPP_INSTALL_UPDATE_SCRIPT "Install the script to update CCD and dictionary files" OFF)
-	set(CIFPP_CACHE_DIR "/var/cache/libcifpp" CACHE STRING "The cache directory to use")
-	target_compile_definitions(cifpp PUBLIC CACHE_DIR="${CIFPP_CACHE_DIR}")
-endif()
-
-generate_export_header(cifpp
-	EXPORT_FILE_NAME cif++/Cif++Export.hpp)
-
-set(INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR} )
-set(LIBRARY_INSTALL_DIR ${CMAKE_INSTALL_LIBDIR} )
-set(SHARE_INSTALL_DIR ${CMAKE_INSTALL_DATADIR}/libcifpp)
-
-set(CIFPP_DATA_DIR "${CMAKE_INSTALL_PREFIX}/${SHARE_INSTALL_DIR}" CACHE STRING "The directory containing the provided data files")
-
+# Installation directories
+set(CIFPP_DATA_DIR "${CMAKE_INSTALL_FULL_DATADIR}/libcifpp")
 target_compile_definitions(cifpp PUBLIC DATA_DIR="${CIFPP_DATA_DIR}")
 
-# Install rules
+if(UNIX)
+	set(CIFPP_CACHE_DIR "${CMAKE_INSTALL_FULL_LOCALSTATEDIR}/cache/libcifpp")
+	target_compile_definitions(cifpp PUBLIC CACHE_DIR="${CIFPP_CACHE_DIR}")
 
+	set(CIFPP_ETC_DIR "${CMAKE_INSTALL_FULL_SYSCONFDIR}")
+endif()
+
+# Install rules
 install(TARGETS cifpp
 	EXPORT cifppTargets
 	ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR}
@@ -341,6 +318,13 @@ install(TARGETS cifpp
 	RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR}
 	INCLUDES DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
 
+if(MSVC AND BUILD_SHARED_LIBS)
+	install(
+		FILES $<TARGET_PDB_FILE:${PROJECT_NAME}>
+		DESTINATION ${CMAKE_INSTALL_LIBDIR}
+		OPTIONAL)
+endif()
+
 install(EXPORT cifppTargets
 	FILE "cifppTargets.cmake"
 	NAMESPACE cifpp::
@@ -354,27 +338,31 @@ install(
 )
 
 install(
-	FILES "${CMAKE_CURRENT_BINARY_DIR}/cif++/Cif++Export.hpp"
-	DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/cif++
+	FILES include/cif++.hpp
+	DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
 	COMPONENT Devel
 )
 
 install(FILES
 	${PROJECT_SOURCE_DIR}/rsrc/mmcif_ddl.dic
-	${PROJECT_SOURCE_DIR}/rsrc/mmcif_pdbx_v50.dic
+	${PROJECT_SOURCE_DIR}/rsrc/mmcif_pdbx.dic
+	${PROJECT_SOURCE_DIR}/rsrc/mmcif_ma.dic
 	${COMPONENTS_CIF}
-	DESTINATION ${SHARE_INSTALL_DIR}
+	DESTINATION ${CIFPP_DATA_DIR}
 )
 
-configure_package_config_file(Config.cmake.in
+set(CONFIG_TEMPLATE_FILE ${PROJECT_SOURCE_DIR}/cmake/cifppConfig.cmake.in)
+
+configure_package_config_file(
+	${CONFIG_TEMPLATE_FILE}
 	${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfig.cmake
 	INSTALL_DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/cifpp
-	PATH_VARS INCLUDE_INSTALL_DIR LIBRARY_INSTALL_DIR SHARE_INSTALL_DIR
+	PATH_VARS CIFPP_DATA_DIR
 )
 
 install(FILES
-		"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfig.cmake"
-		"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfigVersion.cmake"
+	"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfig.cmake"
+	"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfigVersion.cmake"
 	DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/cifpp
 	COMPONENT Devel
 )
@@ -386,21 +374,20 @@ set_target_properties(cifpp PROPERTIES
 	INTERFACE_cifpp_MAJOR_VERSION ${cifpp_MAJOR_VERSION})
 
 set_property(TARGET cifpp APPEND PROPERTY
-  COMPATIBLE_INTERFACE_STRING cifpp_MAJOR_VERSION
+	COMPATIBLE_INTERFACE_STRING cifpp_MAJOR_VERSION
 )
 
 write_basic_package_version_file(
-  "${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfigVersion.cmake"
-  VERSION ${PROJECT_VERSION}
-  COMPATIBILITY AnyNewerVersion
+	"${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifppConfigVersion.cmake"
+	VERSION ${PROJECT_VERSION}
+	COMPATIBILITY AnyNewerVersion
 )
 
 # pkgconfig support
-
-set(prefix      ${CMAKE_INSTALL_PREFIX})
+set(prefix ${CMAKE_INSTALL_PREFIX})
 set(exec_prefix ${CMAKE_INSTALL_PREFIX})
-set(libdir      ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR})
-set(includedir  ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_INCLUDEDIR})
+set(libdir ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR})
+set(includedir ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_INCLUDEDIR})
 
 configure_file(${CMAKE_CURRENT_SOURCE_DIR}/libcifpp.pc.in
 	${CMAKE_CURRENT_BINARY_DIR}/libcifpp.pc.in @ONLY)
@@ -408,38 +395,22 @@ file(GENERATE OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.pc
 	INPUT ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.pc.in)
 install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.pc DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
 
-# Unit tests
+if(ENABLE_TESTING)
+	enable_testing()
 
-option(CIFPP_BUILD_TESTS "Build test exectuables" OFF)
+	find_package(Boost REQUIRED)
 
-if(CIFPP_BUILD_TESTS)
-
-	if(CIFPP_USE_RSRC)
-		add_custom_command(OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/cifpp_test_rsrc.obj
-			COMMAND ${MRC} -o ${CMAKE_CURRENT_BINARY_DIR}/cifpp_test_rsrc.obj ${CMAKE_SOURCE_DIR}/rsrc/mmcif_pdbx_v50.dic ${COFF_SPEC}
-		)
-		set(CIFPP_TEST_RESOURCE ${CMAKE_CURRENT_BINARY_DIR}/cifpp_test_rsrc.obj)
-	endif()
-
-	list(APPEND CIFPP_tests
-		# pdb2cif
-		rename-compound
-		structure
-		unit)
+	list(APPEND CIFPP_tests unit-v2 unit-3d format model rename-compound sugar
+	)
 
 	foreach(CIFPP_TEST IN LISTS CIFPP_tests)
 		set(CIFPP_TEST "${CIFPP_TEST}-test")
 		set(CIFPP_TEST_SOURCE "${CMAKE_CURRENT_SOURCE_DIR}/test/${CIFPP_TEST}.cpp")
 
-		add_executable(${CIFPP_TEST} ${CIFPP_TEST_SOURCE} ${CIFPP_TEST_RESOURCE})
+		add_executable(${CIFPP_TEST} ${CIFPP_TEST_SOURCE})
 
-		target_include_directories(${CIFPP_TEST} PRIVATE
-			${CMAKE_CURRENT_SOURCE_DIR}/include
-			${CMAKE_CURRENT_BINARY_DIR}  # for config.h
-		)
+		target_link_libraries(${CIFPP_TEST} PRIVATE Threads::Threads cifpp::cifpp Boost::boost)
 
-		target_link_libraries(${CIFPP_TEST} PRIVATE Threads::Threads cifpp ${Boost_LIBRARIES} std::filesystem ${ZLIB_LIBRARIES} ${BZip2_LIBRARIES})
-	
 		if(MSVC)
 			# Specify unwind semantics so that MSVC knowns how to handle exceptions
 			target_compile_options(${CIFPP_TEST} PRIVATE /EHsc)
@@ -453,18 +424,20 @@ if(CIFPP_BUILD_TESTS)
 
 		add_test(NAME ${CIFPP_TEST}
 			COMMAND $<TARGET_FILE:${CIFPP_TEST}> -- ${PROJECT_SOURCE_DIR}/test)
-
 	endforeach()
 endif()
 
-message("Will install in ${CMAKE_INSTALL_PREFIX}")
-
 # Optionally install the update scripts for CCD and dictionary files
-
 if(CIFPP_INSTALL_UPDATE_SCRIPT)
-	set(CIFPP_CRON_DIR "$ENV{DESTDIR}/etc/cron.weekly")
+	if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+		set(CIFPP_CRON_DIR "${CIFPP_ETC_DIR}/cron.weekly")
+	elseif(UNIX)	# assume all others are like FreeBSD... 
+		set(CIFPP_CRON_DIR "${CIFPP_ETC_DIR}/periodic/weekly")
+	else()
+		message(FATAL_ERROR "Don't know where to install the update script")
+	endif()
 
-	configure_file(${CMAKE_SOURCE_DIR}/tools/update-libcifpp-data.in update-libcifpp-data @ONLY)
+	configure_file(${PROJECT_SOURCE_DIR}/tools/update-libcifpp-data.in update-libcifpp-data @ONLY)
 	install(
 		FILES ${CMAKE_CURRENT_BINARY_DIR}/update-libcifpp-data
 		DESTINATION ${CIFPP_CRON_DIR}
@@ -472,17 +445,26 @@ if(CIFPP_INSTALL_UPDATE_SCRIPT)
 	)
 
 	install(DIRECTORY DESTINATION ${CIFPP_CACHE_DIR})
-	install(DIRECTORY DESTINATION "$ENV{DESTDIR}/etc/libcifpp/cache-update.d")
+	install(DIRECTORY DESTINATION "${CIFPP_ETC_DIR}/libcifpp/cache-update.d")
 
 	# a config to, to make it complete
-	if(NOT EXISTS "$ENV{DESTDIR}/etc/libcifpp.conf")
+	if(NOT EXISTS "${CIFPP_ETC_DIR}/libcifpp.conf")
 		file(WRITE ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf [[# Uncomment the next line to enable automatic updates
 # update=true
 ]])
-		install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf DESTINATION "$ENV{DESTDIR}/etc")
-		install(CODE "message(\"A configuration file has been written to $ENV{DESTDIR}/etc/libcifpp.conf, please edit this file to enable automatic updates\")")
+		install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libcifpp.conf DESTINATION "${CIFPP_ETC_DIR}")
+		install(CODE "message(\"A configuration file has been written to ${CIFPP_ETC_DIR}/libcifpp.conf, please edit this file to enable automatic updates\")")
 	endif()
 
 	target_compile_definitions(cifpp PUBLIC CACHE_DIR="${CIFPP_CACHE_DIR}")
 endif()
 
+set(CPACK_RESOURCE_FILE_LICENSE "${CMAKE_CURRENT_SOURCE_DIR}/LICENSE")
+set(CPACK_SOURCE_TGZ ON)
+set(CPACK_SOURCE_TBZ2 OFF)
+set(CPACK_SOURCE_TXZ OFF)
+set(CPACK_SOURCE_TZ OFF)
+set(CPACK_SOURCE_IGNORE_FILES "/data/components.cif;/build;/.vscode;/.git")
+set(CPACK_PACKAGE_FILE_NAME "${PROJECT_NAME}-${PROJECT_VERSION}")
+set(CPACK_SOURCE_PACKAGE_FILE_NAME ${CPACK_PACKAGE_FILE_NAME})
+include(CPack)

Config.cmake.in

@@ -1,16 +0,0 @@
-@PACKAGE_INIT@
-
-include(CMakeFindDependencyMacro)
-find_dependency(Boost 1.70.0 REQUIRED COMPONENTS system iostreams regex program_options)
-if(NOT WIN32)
-find_dependency(ZLIB)
-find_dependency(BZip2)
-endif()
-
-INCLUDE("${CMAKE_CURRENT_LIST_DIR}/cifppTargets.cmake")
-
-set_and_check(CIFPP_INCLUDE_DIR "@PACKAGE_INCLUDE_INSTALL_DIR@")
-set_and_check(CIFPP_LIBRARY_DIR "@PACKAGE_LIBRARY_INSTALL_DIR@")
-set_and_check(CIFPP_SHARE_DIR "@PACKAGE_SHARE_INSTALL_DIR@")
-
-check_required_components(cifpp)

LICENSE

@@ -1,6 +1,7 @@
-SPDX-License-Identifier: BSD-2-Clause
+BSD-2-Clause License
 
 Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+All rights reserved.
 
 Redistribution and use in source and binary forms, with or without
 modification, are permitted provided that the following conditions are met:
@@ -20,4 +21,4 @@ ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
 ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
-SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

README.md

@@ -12,7 +12,6 @@ have been used as well as MSVC version 2019.
 
 Other requirements are:
 
-- Boost libraries, at least version 1.70
 - [mrc](https://github.com/mhekkel/mrc), a resource compiler that
   allows including data files into the executable making them easier to
   install. Strictly this is optional, but at the expense of functionality.

changelog

@@ -1,3 +1,73 @@
+Version 5.0.3.1
+- Revert removal of CIFPP_SHARE_DIR export
+
+Version 5.0.3
+- Fix installation of libcifpp into the correct locations
+
+Version 5.0.2
+- Fix export of CISPEP records in PDB format
+- Better support for exporting package_source
+
+Version 5.0.1
+- Fix loading dictionaries
+- Support for cifv1.0 files
+
+Version 5.0.0
+- Total rewrite of cif part
+- Removed DSSP code, moved into dssp project itself
+
+Version 4.2.1
+- Improved REMARK 3 parser (for TLS in large molecules)
+
+Version 4.2.0
+- Yet another rewrite of resource loading
+
+Version 4.1.1
+- Fall back to zero charge for scattering factors if the atom
+  was not found in the table.
+- Improve code to locate resources, failing less.
+
+Version 4.1.0
+- Some interface changes for mmcif::Atom
+
+Version 4.0.1
+- Added a bunch of const methods to Datablock and Category.
+- Changed PDB writing interface to accept Datablock instead of File.
+
+Version 4.0.0
+- getResidue in mmcif::Structure now requires both a
+  sequence ID and an auth sequence ID. As a result the code was cleaned
+  up considerably.
+
+Version 3.0.5
+- mmcif::Structure redesign. It is now a wrapper around a cif::Datablock.
+
+Version 3.0.4
+- Fix in mmCIF parser, now correctly handles the unquoted
+  string ??
+
+Version 3.0.3
+- Better configuration checks, for atomic e.g.
+- Fixed a problem introduced in refactoring mmcif::Atom
+- Version string creation
+
+Version 3.0.2
+- refactored mmcif::Atom for performance reasons
+
+Version 3.0.1
+- Fixed processing of proline restraints file from CCP4, proline
+  is a peptide, really.
+- Added code to facilitate DSSP
+
+Version 3.0.0
+- Replaced many strings in the API with string_view for
+  performance reasons.
+- Upgraded mmcif::Structure
+- various other small fixes
+
+Version 2.0.5
+- Backporting updated CMakeLists.txt file
+
 Version 2.0.4
 - Reverted a too strict test when reading cif files.
 

cmake/FindAtomic.cmake

@@ -0,0 +1,62 @@
+# Simple check to see if we need a library for std::atomic
+
+if(TARGET std::atomic)
+	return()
+endif()
+
+cmake_minimum_required(VERSION 3.10)
+
+include(CMakePushCheckState)
+include(CheckIncludeFileCXX)
+include(CheckCXXSourceRuns)
+
+cmake_push_check_state()
+
+check_include_file_cxx("atomic" _CXX_ATOMIC_HAVE_HEADER)
+mark_as_advanced(_CXX_ATOMIC_HAVE_HEADER)
+
+set(code [[
+#include <atomic>
+int main(int argc, char** argv) {
+  std::atomic<long long> s;
+  ++s;
+  return 0;
+}
+]])
+
+check_cxx_source_runs("${code}" _CXX_ATOMIC_BUILTIN)
+
+if(_CXX_ATOMIC_BUILTIN)
+	set(_found 1)
+else()
+  list(APPEND CMAKE_REQUIRED_LIBRARIES atomic)
+  list(APPEND FOLLY_LINK_LIBRARIES atomic)
+
+  check_cxx_source_runs("${code}" _CXX_ATOMIC_LIB_NEEDED)
+  if (NOT _CXX_ATOMIC_LIB_NEEDED)
+    message(FATAL_ERROR "unable to link C++ std::atomic code: you may need \
+      to install GNU libatomic")
+  else()
+	set(_found 1)
+  endif()
+endif()
+
+if(_found)
+	add_library(std::atomic INTERFACE IMPORTED)
+	set_property(TARGET std::atomic APPEND PROPERTY INTERFACE_COMPILE_FEATURES cxx_std_14)
+
+	if(_CXX_ATOMIC_BUILTIN)
+		# Nothing to add...
+	elseif(_CXX_ATOMIC_LIB_NEEDED)
+		set_target_properties(std::atomic PROPERTIES IMPORTED_LIBNAME atomic)
+		set(STDCPPATOMIC_LIBRARY atomic)
+	endif()
+endif()
+
+cmake_pop_check_state()
+
+set(Atomic_FOUND ${_found} CACHE BOOL "TRUE if we can run a program using std::atomic" FORCE)
+
+if(Atomic_FIND_REQUIRED AND NOT Atomic_FOUND)
+    message(FATAL_ERROR "Cannot run simple program using std::atomic")
+endif()

cmake/FindFilesystem.cmake

@@ -12,8 +12,6 @@ include(CheckCXXSourceCompiles)
 
 cmake_push_check_state()
 
-set(CMAKE_CXX_STANDARD 17)
-
 check_include_file_cxx("filesystem" _CXX_FILESYSTEM_HAVE_HEADER)
 mark_as_advanced(_CXX_FILESYSTEM_HAVE_HEADER)
 
@@ -59,8 +57,10 @@ if(_found)
 		# Nothing to add...
 	elseif(CXX_FILESYSTEM_STDCPPFS_NEEDED)
 		set_target_properties(std::filesystem PROPERTIES IMPORTED_LIBNAME stdc++fs)
+		set(STDCPPFS_LIBRARY stdc++fs)
 	elseif(CXX_FILESYSTEM_CPPFS_NEEDED)
 		set_target_properties(std::filesystem PROPERTIES IMPORTED_LIBNAME c++fs)
+		set(STDCPPFS_LIBRARY c++fs)
 	endif()
 endif()
 

cmake/VersionString.cmake

@@ -0,0 +1,81 @@
+# SPDX-License-Identifier: BSD-2-Clause
+
+# Copyright (c) 2021 NKI/AVL, Netherlands Cancer Institute
+
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+
+# 1. Redistributions of source code must retain the above copyright notice, this
+#    list of conditions and the following disclaimer
+# 2. Redistributions in binary form must reproduce the above copyright notice,
+#    this list of conditions and the following disclaimer in the documentation
+#    and/or other materials provided with the distribution.
+
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+# ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+cmake_minimum_required(VERSION 3.15)
+
+# Create a revision file, containing the current git version info, if any
+function(write_version_header dir)
+	# parameter check
+	if(NOT IS_DIRECTORY ${dir})
+		message(FATAL_ERROR "First parameter to write_version_header should be a directory where the final revision.hpp file will be placed")
+	endif()
+
+	include(GetGitRevisionDescription)
+	if(NOT(GIT-NOTFOUND OR HEAD-HASH-NOTFOUND))
+		git_describe_working_tree(BUILD_VERSION_STRING --match=build --dirty)
+
+		if(BUILD_VERSION_STRING MATCHES "build-([0-9]+)-g([0-9a-f]+)(-dirty)?")
+			set(BUILD_GIT_TAGREF "${CMAKE_MATCH_2}")
+			if(CMAKE_MATCH_3)
+				set(BUILD_VERSION_STRING "${CMAKE_MATCH_1}*")
+			else()
+				set(BUILD_VERSION_STRING "${CMAKE_MATCH_1}")
+			endif()
+		endif()
+	else()
+		message(WARNING "no git info available, cannot update version string")
+	endif()
+
+	string(TIMESTAMP BUILD_DATE_TIME "%Y-%m-%dT%H:%M:%SZ" UTC)
+
+	if(ARGC GREATER 1)
+		set(VAR_PREFIX "${ARGV1}")
+	endif()
+
+	file(WRITE "${PROJECT_BINARY_DIR}/revision.hpp.in" [[// Generated revision file
+
+#pragma once
+
+#include <ostream>
+
+const char k@VAR_PREFIX@ProjectName[] = "@PROJECT_NAME@";
+const char k@VAR_PREFIX@VersionNumber[] = "@PROJECT_VERSION@";
+const char k@VAR_PREFIX@VersionGitTag[] = "@BUILD_GIT_TAGREF@";
+const char k@VAR_PREFIX@BuildInfo[] = "@BUILD_VERSION_STRING@";
+const char k@VAR_PREFIX@BuildDate[] = "@BUILD_DATE_TIME@";
+
+inline void write_version_string(std::ostream &os, bool verbose)
+{
+	os << k@VAR_PREFIX@ProjectName << " version " << k@VAR_PREFIX@VersionNumber << std::endl;
+	if (verbose)
+	{
+		os << "build: " << k@VAR_PREFIX@BuildInfo << ' ' << k@VAR_PREFIX@BuildDate << std::endl;
+		if (k@VAR_PREFIX@VersionGitTag[0] != 0)
+			os << "git tag: " << k@VAR_PREFIX@VersionGitTag << std::endl;
+	}
+}
+]])
+	configure_file("${PROJECT_BINARY_DIR}/revision.hpp.in" "${dir}/revision.hpp" @ONLY)
+endfunction()
+

cmake/cifppConfig.cmake.in

@@ -0,0 +1,13 @@
+@PACKAGE_INIT@
+
+include(CMakeFindDependencyMacro)
+find_dependency(Threads)
+
+find_dependency(ZLIB REQUIRED)
+find_dependency(LibLZMA REQUIRED)
+
+INCLUDE("${CMAKE_CURRENT_LIST_DIR}/cifppTargets.cmake")
+
+set_and_check(CIFPP_SHARE_DIR "@PACKAGE_CIFPP_DATA_DIR@")
+
+check_required_components(cifpp)

cmake/test-rx.cpp

@@ -0,0 +1,18 @@
+// See: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86164
+
+#include <iostream>
+#include <regex>
+
+int main()
+{
+	std::string s(100'000, '*');
+	std::smatch m;
+	std::regex r("^(.*?)$");
+
+	std::regex_search(s, m, r);
+
+	std::cout << s.substr(0, 10) << std::endl;
+	std::cout << m.str(1).substr(0, 10) << std::endl;
+
+	return 0;
+}

data/ccd-subset.cif

@@ -2394,3 +2394,322 @@ VAL "Create component"  1999-07-08 RCSB
 VAL "Modify descriptor" 2011-06-04 RCSB 
 # 
 
+data_NAG
+#
+
+_chem_comp.id                                   NAG
+_chem_comp.name                                 2-acetamido-2-deoxy-beta-D-glucopyranose
+_chem_comp.type                                 "D-saccharide, beta linking"
+_chem_comp.pdbx_type                            ATOMS
+_chem_comp.formula                              "C8 H15 N O6"
+_chem_comp.mon_nstd_parent_comp_id              ?
+_chem_comp.pdbx_synonyms
+;N-acetyl-beta-D-glucosamine; 2-acetamido-2-deoxy-beta-D-glucose; 2-acetamido-2-deoxy-D-glucose;
+2-acetamido-2-deoxy-glucose; N-ACETYL-D-GLUCOSAMINE
+;
+
+_chem_comp.pdbx_formal_charge                   0
+_chem_comp.pdbx_initial_date                    1999-07-08
+_chem_comp.pdbx_modified_date                   2020-07-17
+_chem_comp.pdbx_ambiguous_flag                  N
+_chem_comp.pdbx_release_status                  REL
+_chem_comp.pdbx_replaced_by                     ?
+_chem_comp.pdbx_replaces                        ?
+_chem_comp.formula_weight                       221.208
+_chem_comp.one_letter_code                      ?
+_chem_comp.three_letter_code                    NAG
+_chem_comp.pdbx_model_coordinates_details       ?
+_chem_comp.pdbx_model_coordinates_missing_flag  N
+_chem_comp.pdbx_ideal_coordinates_details       Corina
+_chem_comp.pdbx_ideal_coordinates_missing_flag  N
+_chem_comp.pdbx_model_coordinates_db_code       8PCH
+_chem_comp.pdbx_subcomponent_list               ?
+_chem_comp.pdbx_processing_site                 RCSB
+#   #
+loop_
+_pdbx_chem_comp_synonyms.ordinal
+_pdbx_chem_comp_synonyms.comp_id
+_pdbx_chem_comp_synonyms.name
+_pdbx_chem_comp_synonyms.provenance
+_pdbx_chem_comp_synonyms.type
+1  NAG  N-acetyl-beta-D-glucosamine         PDB  ?
+2  NAG  2-acetamido-2-deoxy-beta-D-glucose  PDB  ?
+3  NAG  2-acetamido-2-deoxy-D-glucose       PDB  ?
+4  NAG  2-acetamido-2-deoxy-glucose         PDB  ?
+5  NAG  N-ACETYL-D-GLUCOSAMINE              PDB  ?
+#   #
+loop_
+_chem_comp_atom.comp_id
+_chem_comp_atom.atom_id
+_chem_comp_atom.alt_atom_id
+_chem_comp_atom.type_symbol
+_chem_comp_atom.charge
+_chem_comp_atom.pdbx_align
+_chem_comp_atom.pdbx_aromatic_flag
+_chem_comp_atom.pdbx_leaving_atom_flag
+_chem_comp_atom.pdbx_stereo_config
+_chem_comp_atom.model_Cartn_x
+_chem_comp_atom.model_Cartn_y
+_chem_comp_atom.model_Cartn_z
+_chem_comp_atom.pdbx_model_Cartn_x_ideal
+_chem_comp_atom.pdbx_model_Cartn_y_ideal
+_chem_comp_atom.pdbx_model_Cartn_z_ideal
+_chem_comp_atom.pdbx_component_atom_id
+_chem_comp_atom.pdbx_component_comp_id
+_chem_comp_atom.pdbx_ordinal
+NAG  C1   C1   C  0  1  N  N  R  7.396  28.163  26.662   0.185   1.082  -0.421  C1   NAG   1
+NAG  C2   C2   C  0  1  N  N  R  6.973  29.233  27.644   0.790  -0.220   0.112  C2   NAG   2
+NAG  C3   C3   C  0  1  N  N  R  7.667  29.055  29.000  -0.124  -1.390  -0.265  C3   NAG   3
+NAG  C4   C4   C  0  1  N  N  S  7.573  27.588  29.490  -1.526  -1.129   0.294  C4   NAG   4
+NAG  C5   C5   C  0  1  N  N  R  7.902  26.592  28.373  -2.042   0.207  -0.246  C5   NAG   5
+NAG  C6   C6   C  0  1  N  N  N  7.599  25.173  28.797  -3.417   0.504   0.355  C6   NAG   6
+NAG  C7   C7   C  0  1  N  N  N  6.291  31.299  26.595   3.197   0.157   0.076  C7   NAG   7
+NAG  C8   C8   C  0  1  N  N  N  6.684  32.649  26.036   4.559  -0.052  -0.533  C8   NAG   8
+NAG  N2   N2   N  0  1  N  N  N  7.268  30.545  27.089   2.114  -0.422  -0.480  N2   NAG   9
+NAG  O1   O1   O  0  1  N  Y  N  6.676  28.363  25.419   1.003   2.185  -0.024  O1   NAG  10
+NAG  O3   O3   O  0  1  N  N  N  7.038  29.909  29.947   0.395  -2.600   0.291  O3   NAG  11
+NAG  O4   O4   O  0  1  N  N  N  8.494  27.358  30.574  -2.405  -2.180  -0.114  O4   NAG  12
+NAG  O5   O5   O  0  1  N  N  N  7.104  26.875  27.206  -1.130   1.248   0.113  O5   NAG  13
+NAG  O6   O6   O  0  1  N  N  N  6.232  25.040  29.165  -3.949   1.691  -0.236  O6   NAG  14
+NAG  O7   O7   O  0  1  N  N  N  5.114  30.936  26.562   3.074   0.845   1.067  O7   NAG  15
+NAG  H1   H1   H  0  1  N  N  N  8.477  28.257  26.481   0.133   1.040  -1.509  H1   NAG  16
+NAG  H2   H2   H  0  1  N  N  N  5.888  29.146  27.803   0.879  -0.163   1.197  H2   NAG  17
+NAG  H3   H3   H  0  1  N  N  N  8.729  29.321  28.892  -0.174  -1.478  -1.350  H3   NAG  18
+NAG  H4   H4   H  0  1  N  N  N  6.544  27.403  29.831  -1.483  -1.091   1.382  H4   NAG  19
+NAG  H5   H5   H  0  1  N  N  N  8.971  26.674  28.128  -2.123   0.154  -1.332  H5   NAG  20
+NAG  H61  H61  H  0  1  N  N  N  7.816  24.492  27.961  -4.088  -0.333   0.157  H61  NAG  21
+NAG  H62  H62  H  0  1  N  N  N  8.232  24.910  29.657  -3.320   0.645   1.431  H62  NAG  22
+NAG  H81  H81  H  0  1  N  N  N  5.791  33.159  25.646   4.560   0.320  -1.558  H81  NAG  23
+NAG  H82  H82  H  0  1  N  N  N  7.136  33.258  26.833   5.305   0.490   0.050  H82  NAG  24
+NAG  H83  H83  H  0  1  N  N  N  7.411  32.511  25.222   4.799  -1.115  -0.532  H83  NAG  25
+NAG  HN2  HN2  H  0  1  N  N  N  8.210  30.881  27.079   2.212  -0.973  -1.273  HN2  NAG  26
+NAG  HO1  HO1  H  0  1  N  Y  N  6.933  27.696  24.793   0.679   3.044  -0.328  HO1  NAG  27
+NAG  HO3  HO3  H  0  1  N  Y  N  7.459  29.809  30.793  -0.135  -3.384   0.091  HO3  NAG  28
+NAG  HO4  HO4  H  0  1  N  Y  N  8.425  26.456  30.863  -3.312  -2.079   0.206  HO4  NAG  29
+NAG  HO6  HO6  H  0  1  N  Y  N  6.060  24.143  29.428  -4.822   1.940   0.099  HO6  NAG  30
+#   #
+loop_
+_chem_comp_bond.comp_id
+_chem_comp_bond.atom_id_1
+_chem_comp_bond.atom_id_2
+_chem_comp_bond.value_order
+_chem_comp_bond.pdbx_aromatic_flag
+_chem_comp_bond.pdbx_stereo_config
+_chem_comp_bond.pdbx_ordinal
+NAG  C1  C2   SING  N  N   1
+NAG  C1  O1   SING  N  N   2
+NAG  C1  O5   SING  N  N   3
+NAG  C1  H1   SING  N  N   4
+NAG  C2  C3   SING  N  N   5
+NAG  C2  N2   SING  N  N   6
+NAG  C2  H2   SING  N  N   7
+NAG  C3  C4   SING  N  N   8
+NAG  C3  O3   SING  N  N   9
+NAG  C3  H3   SING  N  N  10
+NAG  C4  C5   SING  N  N  11
+NAG  C4  O4   SING  N  N  12
+NAG  C4  H4   SING  N  N  13
+NAG  C5  C6   SING  N  N  14
+NAG  C5  O5   SING  N  N  15
+NAG  C5  H5   SING  N  N  16
+NAG  C6  O6   SING  N  N  17
+NAG  C6  H61  SING  N  N  18
+NAG  C6  H62  SING  N  N  19
+NAG  C7  C8   SING  N  N  20
+NAG  C7  N2   SING  N  N  21
+NAG  C7  O7   DOUB  N  N  22
+NAG  C8  H81  SING  N  N  23
+NAG  C8  H82  SING  N  N  24
+NAG  C8  H83  SING  N  N  25
+NAG  N2  HN2  SING  N  N  26
+NAG  O1  HO1  SING  N  N  27
+NAG  O3  HO3  SING  N  N  28
+NAG  O4  HO4  SING  N  N  29
+NAG  O6  HO6  SING  N  N  30
+#   #
+loop_
+_pdbx_chem_comp_descriptor.comp_id
+_pdbx_chem_comp_descriptor.type
+_pdbx_chem_comp_descriptor.program
+_pdbx_chem_comp_descriptor.program_version
+_pdbx_chem_comp_descriptor.descriptor
+NAG  SMILES            ACDLabs               12.01  "O=C(NC1C(O)C(O)C(OC1O)CO)C"
+NAG  InChI             InChI                 1.03   "InChI=1S/C8H15NO6/c1-3(11)9-5-7(13)6(12)4(2-10)15-8(5)14/h4-8,10,12-14H,2H2,1H3,(H,9,11)/t4-,5-,6-,7-,8-/m1/s1"
+NAG  InChIKey          InChI                 1.03   OVRNDRQMDRJTHS-FMDGEEDCSA-N
+NAG  SMILES_CANONICAL  CACTVS                3.370  "CC(=O)N[C@H]1[C@H](O)O[C@H](CO)[C@@H](O)[C@@H]1O"
+NAG  SMILES            CACTVS                3.370  "CC(=O)N[CH]1[CH](O)O[CH](CO)[CH](O)[CH]1O"
+NAG  SMILES_CANONICAL  "OpenEye OEToolkits"  1.7.6  "CC(=O)N[C@@H]1[C@H]([C@@H]([C@H](O[C@H]1O)CO)O)O"
+NAG  SMILES            "OpenEye OEToolkits"  1.7.6  "CC(=O)NC1C(C(C(OC1O)CO)O)O"
+#   #
+loop_
+_pdbx_chem_comp_identifier.comp_id
+_pdbx_chem_comp_identifier.type
+_pdbx_chem_comp_identifier.program
+_pdbx_chem_comp_identifier.program_version
+_pdbx_chem_comp_identifier.identifier
+NAG  "SYSTEMATIC NAME"                      ACDLabs               12.01  "2-(acetylamino)-2-deoxy-beta-D-glucopyranose"
+NAG  "SYSTEMATIC NAME"                      "OpenEye OEToolkits"  1.7.6  "N-[(2R,3R,4R,5S,6R)-6-(hydroxymethyl)-2,4,5-tris(oxidanyl)oxan-3-yl]ethanamide"
+NAG  "CONDENSED IUPAC CARBOHYDRATE SYMBOL"  GMML                  1.0    DGlcpNAcb
+NAG  "COMMON NAME"                          GMML                  1.0    N-acetyl-b-D-glucopyranosamine
+NAG  "IUPAC CARBOHYDRATE SYMBOL"            PDB-CARE              1.0    b-D-GlcpNAc
+NAG  "SNFG CARBOHYDRATE SYMBOL"             GMML                  1.0    GlcNAc
+#   #
+loop_
+_pdbx_chem_comp_feature.comp_id
+_pdbx_chem_comp_feature.type
+_pdbx_chem_comp_feature.value
+_pdbx_chem_comp_feature.source
+_pdbx_chem_comp_feature.support
+NAG  "CARBOHYDRATE ISOMER"                  D         PDB  ?
+NAG  "CARBOHYDRATE RING"                    pyranose  PDB  ?
+NAG  "CARBOHYDRATE ANOMER"                  beta      PDB  ?
+NAG  "CARBOHYDRATE PRIMARY CARBONYL GROUP"  aldose    PDB  ?
+#   #
+loop_
+_pdbx_chem_comp_audit.comp_id
+_pdbx_chem_comp_audit.action_type
+_pdbx_chem_comp_audit.date
+_pdbx_chem_comp_audit.processing_site
+NAG  "Create component"          1999-07-08  RCSB
+NAG  "Modify descriptor"         2011-06-04  RCSB
+NAG  "Modify leaving atom flag"  2011-07-01  RCSB
+NAG  "Modify leaving atom flag"  2012-11-26  RCSB
+NAG  "Other modification"        2019-08-12  RCSB
+NAG  "Other modification"        2019-12-19  RCSB
+NAG  "Other modification"        2020-07-03  RCSB
+NAG  "Modify name"               2020-07-17  RCSB
+NAG  "Modify synonyms"           2020-07-17  RCSB
+##
+
+
+
+data_HIS
+# 
+_chem_comp.id                                    HIS 
+_chem_comp.name                                  HISTIDINE 
+_chem_comp.type                                  "L-PEPTIDE LINKING" 
+_chem_comp.pdbx_type                             ATOMP 
+_chem_comp.formula                               "C6 H10 N3 O2" 
+_chem_comp.mon_nstd_parent_comp_id               ? 
+_chem_comp.pdbx_synonyms                         ? 
+_chem_comp.pdbx_formal_charge                    1 
+_chem_comp.pdbx_initial_date                     1999-07-08 
+_chem_comp.pdbx_modified_date                    2011-06-04 
+_chem_comp.pdbx_ambiguous_flag                   N 
+_chem_comp.pdbx_release_status                   REL 
+_chem_comp.pdbx_replaced_by                      ? 
+_chem_comp.pdbx_replaces                         ? 
+_chem_comp.formula_weight                        156.162 
+_chem_comp.one_letter_code                       H 
+_chem_comp.three_letter_code                     HIS 
+_chem_comp.pdbx_model_coordinates_details        ? 
+_chem_comp.pdbx_model_coordinates_missing_flag   N 
+_chem_comp.pdbx_ideal_coordinates_details        "OpenEye/OEToolkits V1.4.2" 
+_chem_comp.pdbx_ideal_coordinates_missing_flag   N 
+_chem_comp.pdbx_model_coordinates_db_code        ? 
+_chem_comp.pdbx_subcomponent_list                ? 
+_chem_comp.pdbx_processing_site                  EBI 
+# 
+loop_
+_chem_comp_atom.comp_id 
+_chem_comp_atom.atom_id 
+_chem_comp_atom.alt_atom_id 
+_chem_comp_atom.type_symbol 
+_chem_comp_atom.charge 
+_chem_comp_atom.pdbx_align 
+_chem_comp_atom.pdbx_aromatic_flag 
+_chem_comp_atom.pdbx_leaving_atom_flag 
+_chem_comp_atom.pdbx_stereo_config 
+_chem_comp_atom.model_Cartn_x 
+_chem_comp_atom.model_Cartn_y 
+_chem_comp_atom.model_Cartn_z 
+_chem_comp_atom.pdbx_model_Cartn_x_ideal 
+_chem_comp_atom.pdbx_model_Cartn_y_ideal 
+_chem_comp_atom.pdbx_model_Cartn_z_ideal 
+_chem_comp_atom.pdbx_component_atom_id 
+_chem_comp_atom.pdbx_component_comp_id 
+_chem_comp_atom.pdbx_ordinal 
+HIS N   N   N 0 1 N N N 33.472 42.685 -4.610 -0.040 -1.210 0.053  N   HIS 1  
+HIS CA  CA  C 0 1 N N S 33.414 41.686 -5.673 1.172  -1.709 0.652  CA  HIS 2  
+HIS C   C   C 0 1 N N N 33.773 42.279 -7.040 1.083  -3.207 0.905  C   HIS 3  
+HIS O   O   O 0 1 N N N 33.497 43.444 -7.337 0.040  -3.770 1.222  O   HIS 4  
+HIS CB  CB  C 0 1 N N N 32.005 41.080 -5.734 1.484  -0.975 1.962  CB  HIS 5  
+HIS CG  CG  C 0 1 Y N N 31.888 39.902 -6.651 2.940  -1.060 2.353  CG  HIS 6  
+HIS ND1 ND1 N 1 1 Y N N 32.539 38.710 -6.414 3.380  -2.075 3.129  ND1 HIS 7  
+HIS CD2 CD2 C 0 1 Y N N 31.199 39.734 -7.804 3.960  -0.251 2.046  CD2 HIS 8  
+HIS CE1 CE1 C 0 1 Y N N 32.251 37.857 -7.382 4.693  -1.908 3.317  CE1 HIS 9  
+HIS NE2 NE2 N 0 1 Y N N 31.439 38.453 -8.237 5.058  -0.801 2.662  NE2 HIS 10 
+HIS OXT OXT O 0 1 N Y N 34.382 41.455 -7.879 2.247  -3.882 0.744  OXT HIS 11 
+HIS H   H   H 0 1 N N N 33.485 42.227 -3.721 -0.102 -1.155 -0.950 H   HIS 12 
+HIS H2  HN2 H 0 1 N Y N 34.301 43.234 -4.714 -0.715 -0.741 0.634  H2  HIS 13 
+HIS HA  HA  H 0 1 N N N 34.155 40.908 -5.439 1.965  -1.558 -0.089 HA  HIS 14 
+HIS HB2 1HB H 0 1 N N N 31.733 40.750 -4.721 1.215  0.087  1.879  HB2 HIS 15 
+HIS HB3 2HB H 0 1 N N N 31.337 41.860 -6.127 0.859  -1.368 2.775  HB3 HIS 16 
+HIS HD1 HD1 H 0 1 N N N 33.135 38.521 -5.633 2.828  -2.838 3.511  HD1 HIS 17 
+HIS HD2 HD2 H 0 1 N N N 30.577 40.470 -8.292 4.108  0.647  1.479  HD2 HIS 18 
+HIS HE1 HE1 H 0 1 N N N 32.618 36.844 -7.461 5.340  -2.550 3.892  HE1 HIS 19 
+HIS HE2 HE2 H 0 1 N N N 31.061 38.039 -9.065 6.002  -0.428 2.627  HE2 HIS 20 
+HIS HXT HXT H 0 1 N Y N 34.553 41.905 -8.698 2.188  -4.848 0.901  HXT HIS 21 
+# 
+loop_
+_chem_comp_bond.comp_id 
+_chem_comp_bond.atom_id_1 
+_chem_comp_bond.atom_id_2 
+_chem_comp_bond.value_order 
+_chem_comp_bond.pdbx_aromatic_flag 
+_chem_comp_bond.pdbx_stereo_config 
+_chem_comp_bond.pdbx_ordinal 
+HIS N   CA  SING N N 1  
+HIS N   H   SING N N 2  
+HIS N   H2  SING N N 3  
+HIS CA  C   SING N N 4  
+HIS CA  CB  SING N N 5  
+HIS CA  HA  SING N N 6  
+HIS C   O   DOUB N N 7  
+HIS C   OXT SING N N 8  
+HIS CB  CG  SING N N 9  
+HIS CB  HB2 SING N N 10 
+HIS CB  HB3 SING N N 11 
+HIS CG  ND1 SING Y N 12 
+HIS CG  CD2 DOUB Y N 13 
+HIS ND1 CE1 DOUB Y N 14 
+HIS ND1 HD1 SING N N 15 
+HIS CD2 NE2 SING Y N 16 
+HIS CD2 HD2 SING N N 17 
+HIS CE1 NE2 SING Y N 18 
+HIS CE1 HE1 SING N N 19 
+HIS NE2 HE2 SING N N 20 
+HIS OXT HXT SING N N 21 
+# 
+loop_
+_pdbx_chem_comp_descriptor.comp_id 
+_pdbx_chem_comp_descriptor.type 
+_pdbx_chem_comp_descriptor.program 
+_pdbx_chem_comp_descriptor.program_version 
+_pdbx_chem_comp_descriptor.descriptor 
+HIS SMILES           ACDLabs              10.04 "O=C(O)C(N)Cc1cnc[nH+]1"                                                                    
+HIS SMILES_CANONICAL CACTVS               3.341 "N[C@@H](Cc1c[nH]c[nH+]1)C(O)=O"                                                            
+HIS SMILES           CACTVS               3.341 "N[CH](Cc1c[nH]c[nH+]1)C(O)=O"                                                              
+HIS SMILES_CANONICAL "OpenEye OEToolkits" 1.5.0 "c1c([nH+]c[nH]1)C[C@@H](C(=O)O)N"                                                          
+HIS SMILES           "OpenEye OEToolkits" 1.5.0 "c1c([nH+]c[nH]1)CC(C(=O)O)N"                                                               
+HIS InChI            InChI                1.03  "InChI=1S/C6H9N3O2/c7-5(6(10)11)1-4-2-8-3-9-4/h2-3,5H,1,7H2,(H,8,9)(H,10,11)/p+1/t5-/m0/s1" 
+HIS InChIKey         InChI                1.03  HNDVDQJCIGZPNO-YFKPBYRVSA-O                                                                 
+# 
+loop_
+_pdbx_chem_comp_identifier.comp_id 
+_pdbx_chem_comp_identifier.type 
+_pdbx_chem_comp_identifier.program 
+_pdbx_chem_comp_identifier.program_version 
+_pdbx_chem_comp_identifier.identifier 
+HIS "SYSTEMATIC NAME" ACDLabs              10.04 "3-(1H-imidazol-3-ium-4-yl)-L-alanine"                  
+HIS "SYSTEMATIC NAME" "OpenEye OEToolkits" 1.5.0 "(2S)-2-amino-3-(1H-imidazol-3-ium-4-yl)propanoic acid" 
+# 
+loop_
+_pdbx_chem_comp_audit.comp_id 
+_pdbx_chem_comp_audit.action_type 
+_pdbx_chem_comp_audit.date 
+_pdbx_chem_comp_audit.processing_site 
+HIS "Create component"  1999-07-08 EBI  
+HIS "Modify descriptor" 2011-06-04 RCSB 
+# 
+

examples/example.cpp

@@ -1,28 +1,24 @@
 #include <iostream>
 #include <filesystem>
 
-#include <cif++/Cif++.hpp>
+#include <cif++.hpp>
 
 namespace fs = std::filesystem;
 
 int main()
 {
-	fs::path in("1cbs.cif.gz");
-
-	cif::File file;
-
-	file.loadDictionary("mmcif_pdbx_v50");
-
+	cif::file file;
 	file.load("1cbs.cif.gz");
 
-	auto& db = file.firstDatablock()["atom_site"];
-	auto n = db.find(cif::Key("label_atom_id") == "OXT").size();
+	auto& db = file.front();
+	auto &atom_site = db["atom_site"];
+	auto n = atom_site.find(cif::key("label_atom_id") == "OXT").size();
 
-	std::cout << "File contains " << db.size() << " atoms of which " << n << (n == 1 ? " is" : " are") << " OXT" << std::endl
+	std::cout << "File contains " << atom_site.size() << " atoms of which " << n << (n == 1 ? " is" : " are") << " OXT" << std::endl
 		<< "residues with an OXT are:" << std::endl;
 	
-	for (const auto& [asym, comp, seqnr]: db.find<std::string,std::string,int>(
-			cif::Key("label_atom_id") == "OXT", "label_asym_id", "label_comp_id", "label_seq_id"))
+	for (const auto& [asym, comp, seqnr]: atom_site.find<std::string,std::string,int>(
+			cif::key("label_atom_id") == "OXT", "label_asym_id", "label_comp_id", "label_seq_id"))
 	{
 		std::cout << asym << ' ' << comp << ' ' << seqnr << std::endl;
 	}

include/cif++.hpp

@@ -0,0 +1,41 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cif++/utilities.hpp>
+#include <cif++/file.hpp>
+#include <cif++/parser.hpp>
+#include <cif++/format.hpp>
+
+#include <cif++/compound.hpp>
+#include <cif++/point.hpp>
+#include <cif++/symmetry.hpp>
+
+#include <cif++/model.hpp>
+
+#include <cif++/pdb/io.hpp>
+#include <cif++/gzio.hpp>

include/cif++/AtomType.hpp

@@ -1,245 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-// Lib for working with structures as contained in mmCIF and PDB files
-
-#pragma once
-
-#include <cstdint>
-#include <string>
-#include <stdexcept>
-
-namespace mmcif
-{
-
-enum AtomType : uint8_t
-{
-	Nn = 0,		// Unknown
-	
-	H = 1,		// Hydro­gen
-	He = 2,		// He­lium
-
-	Li = 3,		// Lith­ium
-	Be = 4,		// Beryl­lium
-	B = 5,		// Boron
-	C = 6,		// Carbon
-	N = 7,		// Nitro­gen
-	O = 8,		// Oxy­gen
-	F = 9,		// Fluor­ine
-	Ne = 10,	// Neon
-
-	Na = 11,	// So­dium
-	Mg = 12,	// Magne­sium
-	Al = 13,	// Alumin­ium
-	Si = 14,	// Sili­con
-	P = 15,		// Phos­phorus
-	S = 16,		// Sulfur
-	Cl = 17,	// Chlor­ine
-	Ar = 18,	// Argon
-
-	K = 19,		// Potas­sium
-	Ca = 20,	// Cal­cium
-	Sc = 21,	// Scan­dium
-	Ti = 22,	// Tita­nium
-	V = 23,		// Vana­dium
-	Cr = 24,	// Chrom­ium
-	Mn = 25,	// Manga­nese
-	Fe = 26,	// Iron
-	Co = 27,	// Cobalt
-	Ni = 28,	// Nickel
-	Cu = 29,	// Copper
-	Zn = 30,	// Zinc
-	Ga = 31,	// Gallium
-	Ge = 32,	// Germa­nium
-	As = 33,	// Arsenic
-	Se = 34,	// Sele­nium
-	Br = 35,	// Bromine
-	Kr = 36,	// Kryp­ton
-
-	Rb = 37,	// Rubid­ium
-	Sr = 38,	// Stront­ium
-	Y = 39,		// Yttrium
-	Zr = 40,	// Zirco­nium
-	Nb = 41,	// Nio­bium
-	Mo = 42,	// Molyb­denum
-	Tc = 43,	// Tech­netium
-	Ru = 44,	// Ruthe­nium
-	Rh = 45,	// Rho­dium
-	Pd = 46,	// Pallad­ium
-	Ag = 47,	// Silver
-	Cd = 48,	// Cad­mium
-	In = 49,	// Indium
-	Sn = 50,	// Tin
-	Sb = 51,	// Anti­mony
-	Te = 52,	// Tellurium
-	I = 53,		// Iodine
-	Xe = 54,	// Xenon
-	Cs = 55,	// Cae­sium
-	Ba = 56,	// Ba­rium
-	La = 57,	// Lan­thanum
-
-	Hf = 72,	// Haf­nium
-	Ta = 73,	// Tanta­lum
-	W = 74,		// Tung­sten
-	Re = 75,	// Rhe­nium
-	Os = 76,	// Os­mium
-	Ir = 77,	// Iridium
-	Pt = 78,	// Plat­inum
-	Au = 79,	// Gold
-	Hg = 80,	// Mer­cury
-	Tl = 81,	// Thallium
-	Pb = 82,	// Lead
-	Bi = 83,	// Bis­muth
-	Po = 84,	// Polo­nium
-	At = 85,	// Asta­tine
-	Rn = 86,	// Radon
-	Fr = 87,	// Fran­cium
-	Ra = 88,	// Ra­dium
-	Ac = 89,	// Actin­ium
-
-	Rf = 104,	// Ruther­fordium
-	Db = 105,	// Dub­nium
-	Sg = 106,	// Sea­borgium
-	Bh = 107,	// Bohr­ium
-	Hs = 108,	// Has­sium
-	Mt = 109,	// Meit­nerium
-	Ds = 110,	// Darm­stadtium
-	Rg = 111,	// Roent­genium
-	Cn = 112,	// Coper­nicium
-	Nh = 113,	// Nihon­ium
-	Fl = 114,	// Flerov­ium
-	Mc = 115,	// Moscov­ium
-	Lv = 116,	// Liver­morium
-	Ts = 117,	// Tenness­ine
-	Og = 118,	// Oga­nesson
-
-	Ce = 58,	// Cerium
-	Pr = 59,	// Praseo­dymium
-	Nd = 60,	// Neo­dymium
-	Pm = 61,	// Prome­thium
-	Sm = 62,	// Sama­rium
-	Eu = 63,	// Europ­ium
-	Gd = 64,	// Gadolin­ium
-	Tb = 65,	// Ter­bium
-	Dy = 66,	// Dyspro­sium
-	Ho = 67,	// Hol­mium
-	Er = 68,	// Erbium
-	Tm = 69,	// Thulium
-	Yb = 70,	// Ytter­bium
-	Lu = 71,	// Lute­tium
-
-	Th = 90,	// Thor­ium
-	Pa = 91,	// Protac­tinium
-	U = 92,		// Ura­nium
-	Np = 93,	// Neptu­nium
-	Pu = 94,	// Pluto­nium
-	Am = 95,	// Ameri­cium
-	Cm = 96,	// Curium
-	Bk = 97,	// Berkel­ium
-	Cf = 98,	// Califor­nium
-	Es = 99,	// Einstei­nium
-	Fm = 100,	// Fer­mium
-	Md = 101,	// Mende­levium
-	No = 102,	// Nobel­ium
-	Lr = 103,	// Lawren­cium
-
-	D = 129,	// Deuterium
-};
-
-// --------------------------------------------------------------------
-// AtomTypeInfo
-
-enum RadiusType {
-	eRadiusCalculated,
-	eRadiusEmpirical,
-	eRadiusCovalentEmpirical,
-
-	eRadiusSingleBond,
-	eRadiusDoubleBond,
-	eRadiusTripleBond,
-
-	eRadiusVanderWaals,
-
-	eRadiusTypeCount
-};
-
-struct AtomTypeInfo
-{
-	AtomType		type;
-	std::string		name;
-	std::string		symbol;
-	float			weight;
-	bool			metal;
-	float			radii[eRadiusTypeCount];
-};
-
-extern const AtomTypeInfo kKnownAtoms[];
-
-// --------------------------------------------------------------------
-// AtomTypeTraits
-
-class AtomTypeTraits
-{
-  public:
-	AtomTypeTraits(AtomType a);
-	AtomTypeTraits(const std::string& symbol);
-	
-	AtomType type() const			{ return mInfo->type; }
-	std::string	name() const		{ return mInfo->name; }
-	std::string	symbol() const		{ return mInfo->symbol; }
-	float weight() const			{ return mInfo->weight; }
-	
-	bool isMetal() const			{ return mInfo->metal; }
-	
-	static bool isElement(const std::string& symbol);
-	static bool isMetal(const std::string& symbol);
-	
-	float radius(RadiusType type = eRadiusSingleBond) const
-	{
-		if (type >= eRadiusTypeCount)
-			throw std::invalid_argument("invalid radius requested");
-		return mInfo->radii[type] / 100.f;
-	}
-	
-	// data type encapsulating the Waasmaier & Kirfel scattering factors
-	// in a simplified form (only a and b).
-	// Added the electrion scattering factors as well
-	struct SFData
-	{
-		double a[6], b[6];
-	};
-	
-	// to get the Cval and Siva values, use this constant as charge:
-	enum { kWKSFVal = -99 };
-	
-	const SFData& wksf(int charge = 0) const;
-	const SFData& elsf() const;
-
-  private:
-	const struct AtomTypeInfo*	mInfo;
-};
-
-}

include/cif++/BondMap.hpp

@@ -1,101 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#pragma once
-
-#include <unordered_map>
-#include <filesystem>
-#include <stdexcept>
-
-#include "cif++/Structure.hpp"
-
-namespace mmcif
-{
-
-class BondMapException : public std::runtime_error
-{
-  public:
-	BondMapException(const std::string& msg)
-		: runtime_error(msg) {}
-};
-
-class BondMap
-{
-  public:
-	BondMap(const Structure& p);
-	
-	BondMap(const BondMap&) = delete;
-	BondMap& operator=(const BondMap&) = delete;
-
-	bool operator()(const Atom& a, const Atom& b) const
-	{
-		return isBonded(index.at(a.id()), index.at(b.id()));
-	}
-
-	bool is1_4(const Atom& a, const Atom& b) const
-	{
-		uint32_t ixa = index.at(a.id());
-		uint32_t ixb = index.at(b.id());
-	
-		return bond_1_4.count(key(ixa, ixb));
-	}
-	
-	// links coming from the struct_conn records:
-	std::vector<std::string> linked(const Atom& a) const;
-
-	// This list of atomID's is comming from either CCD or the CCP4 dictionaries loaded
-	static std::vector<std::string> atomIDsForCompound(const std::string& compoundID);
-	
-  private:
-
-	bool isBonded(uint32_t ai, uint32_t bi) const
-	{
-		return bond.count(key(ai, bi)) != 0;
-	}
-
-	uint64_t key(uint32_t a, uint32_t b) const
-	{
-		if (a > b)
-			std::swap(a, b);
-		return static_cast<uint64_t>(a) | (static_cast<uint64_t>(b) << 32);
-	}
-	
-	std::tuple<uint32_t,uint32_t> dekey(uint64_t k) const
-	{
-		return std::make_tuple(
-			static_cast<uint32_t>(k >> 32),
-			static_cast<uint32_t>(k)
-		);
-	}
-	
-	uint32_t dim;
-	std::unordered_map<std::string,uint32_t> index;
-	std::set<uint64_t> bond, bond_1_4;
-
-	std::map<std::string,std::set<std::string>> link;
-};
-
-}

include/cif++/CifParser.hpp

@@ -1,248 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#pragma once
-
-#include "cif++/Cif++.hpp"
-
-#include <stack>
-#include <map>
-
-namespace cif
-{
-
-// --------------------------------------------------------------------
-
-class CifParserError : public std::runtime_error
-{
-  public:
-	CifParserError(uint32_t lineNr, const std::string& message);
-};
-
-// --------------------------------------------------------------------
-
-extern const uint32_t kMaxLineLength;
-
-extern const uint8_t kCharTraitsTable[128];
-
-enum CharTraitsMask: uint8_t {
-	kOrdinaryMask = 1 << 0,
-	kNonBlankMask = 1 << 1,
-	kTextLeadMask = 1 << 2,
-	kAnyPrintMask = 1 << 3
-};
-
-inline bool isWhite(int ch)
-{
-	return std::isspace(ch) or ch == '#';
-}
-
-inline bool isOrdinary(int ch)
-{
-	return ch >= 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kOrdinaryMask) != 0;
-}
-
-inline bool isNonBlank(int ch)
-{
-	return ch > 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kNonBlankMask) != 0;
-}
-
-inline bool isTextLead(int ch)
-{
-	return ch >= 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kTextLeadMask) != 0;
-}
-
-inline bool isAnyPrint(int ch)	
-{
-	return ch == '\t' or 
-		(ch >= 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kAnyPrintMask) != 0);
-}
-
-inline bool isUnquotedString(const char* s)
-{
-	bool result = isOrdinary(*s++);
-	while (result and *s != 0)
-	{
-		result = isNonBlank(*s);
-		++s;
-	}
-	return result;
-}
-
-// --------------------------------------------------------------------
-
-std::tuple<std::string,std::string> splitTagName(const std::string& tag);
-
-// --------------------------------------------------------------------
-
-using DatablockIndex = std::map<std::string,std::size_t>;
-
-// --------------------------------------------------------------------
-// sac Parser, analogous to SAX Parser (simple api for xml)
-
-class SacParser
-{
-  public:
-	SacParser(std::istream& is, bool init = true);
-	virtual ~SacParser() {}
-
-	enum CIFToken
-	{
-		eCIFTokenUnknown,
-		
-		eCIFTokenEOF,
-	
-		eCIFTokenDATA,
-		eCIFTokenLOOP,
-		eCIFTokenGLOBAL,
-		eCIFTokenSAVE,
-		eCIFTokenSTOP,
-		eCIFTokenTag,
-		eCIFTokenValue,
-	};
-
-	static const char* kTokenName[];
-
-	enum CIFValueType
-	{
-		eCIFValueInt,
-		eCIFValueFloat,
-		eCIFValueNumeric,
-		eCIFValueString,
-		eCIFValueTextField,
-		eCIFValueInapplicable,
-		eCIFValueUnknown
-	};
-
-	static const char* kValueName[];
-	
-	int getNextChar();
-
-	void retract();
-	void restart();
-	
-	CIFToken getNextToken();
-	void match(CIFToken token);
-
-	bool parseSingleDatablock(const std::string& datablock);
-
-	DatablockIndex indexDatablocks();
-	bool parseSingleDatablock(const std::string& datablock, const DatablockIndex &index);
-
-	void parseFile();
-	void parseGlobal();
-	void parseDataBlock();
-
-	virtual void parseSaveFrame();
-	
-	void parseDictionary();
-	
-	void error(const std::string& msg);
-	
-	// production methods, these are pure virtual here
-	
-	virtual void produceDatablock(const std::string& name) = 0;
-	virtual void produceCategory(const std::string& name) = 0;
-	virtual void produceRow() = 0;
-	virtual void produceItem(const std::string& category, const std::string& item, const std::string& value) = 0;
-
-  protected:
-
-	enum State
-	{
-		eStateStart,
-		eStateWhite,
-		eStateComment,
-		eStateQuestionMark,
-		eStateDot,
-		eStateQuotedString,
-		eStateQuotedStringQuote,
-		eStateUnquotedString,
-		eStateTag,
-		eStateTextField,
-		eStateFloat = 100,
-		eStateInt = 110,
-//		eStateNumericSuffix = 200,
-		eStateValue = 300
-	};
-
-	std::istream&			mData;
-
-	// Parser state
-	bool					mValidate;
-	uint32_t					mLineNr;
-	bool					mBol;
-	int						mState, mStart;
-	CIFToken				mLookahead;
-	std::string				mTokenValue;
-	CIFValueType			mTokenType;
-	std::stack<int>			mBuffer;
-};
-
-// --------------------------------------------------------------------
-
-class Parser : public SacParser
-{
-  public:
-	Parser(std::istream& is, File& f, bool init = true);
-
-	virtual void produceDatablock(const std::string& name);
-	virtual void produceCategory(const std::string& name);
-	virtual void produceRow();
-	virtual void produceItem(const std::string& category, const std::string& item, const std::string& value);
-
-  protected:
-	File&					mFile;
-	Datablock*				mDataBlock;
-	Datablock::iterator		mCat;
-	Row						mRow;
-};
-
-// --------------------------------------------------------------------
-
-class DictParser : public Parser
-{
-  public:
-
-	DictParser(Validator& validator, std::istream& is);
-	~DictParser();
-	
-	void loadDictionary();
-	
-  private:
-
-	virtual void parseSaveFrame();
-	
-	bool collectItemTypes();
-	void linkItems();
-
-	Validator&						mValidator;
-	File							mFile;
-	struct DictParserDataImpl*		mImpl;
-	bool							mCollectedItemTypes = false;
-};
-
-}

include/cif++/CifValidator.hpp

@@ -1,198 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#pragma once
-
-#include "cif++/Cif++.hpp"
-
-// duh.. https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86164
-// #include <regex>
-#include <boost/regex.hpp>
-
-#include <set>
-
-namespace cif
-{
-	
-struct ValidateCategory;
-
-// --------------------------------------------------------------------
-
-class ValidationError : public std::exception
-{
-  public:
-	ValidationError(const std::string& msg);
-	ValidationError(const std::string& cat, const std::string& item,
-		const std::string& msg);
-	const char* what() const noexcept		{ return mMsg.c_str(); }
-	std::string mMsg;
-};
-
-// --------------------------------------------------------------------
-
-enum class DDL_PrimitiveType
-{
-	Char, UChar, Numb
-};
-
-DDL_PrimitiveType mapToPrimitiveType(const std::string& s);
-
-struct ValidateType
-{
-	std::string				mName;
-	DDL_PrimitiveType		mPrimitiveType;
-	// std::regex				mRx;
-	boost::regex			mRx;
-
-	bool operator<(const ValidateType& rhs) const
-	{
-		return icompare(mName, rhs.mName) < 0;
-	}
-
-	// compare values based on type	
-//	int compare(const std::string& a, const std::string& b) const
-//	{
-//		return compare(a.c_str(), b.c_str());
-//	}
-	
-	int compare(const char* a, const char* b) const;
-};
-
-struct ValidateItem
-{
-	std::string				mTag;
-	bool					mMandatory;
-	const ValidateType*		mType;
-	cif::iset				mEnums;
-	std::string				mDefault;
-	bool					mDefaultIsNull;
-	ValidateCategory*		mCategory = nullptr;
-
-	// ItemLinked is used for non-key links
-	struct ItemLinked
-	{
-		ValidateItem*		mParent;
-		std::string			mParentItem;
-		std::string			mChildItem;
-	};
-
-	std::vector<ItemLinked>	mLinked;
-	
-	bool operator<(const ValidateItem& rhs) const
-	{
-		return icompare(mTag, rhs.mTag) < 0;
-	}
-
-	bool operator==(const ValidateItem& rhs) const
-	{
-		return iequals(mTag, rhs.mTag);
-	}
-
-	void operator()(std::string value) const;
-};
-
-struct ValidateCategory
-{
-	std::string					mName;
-	std::vector<std::string>	mKeys;
-	cif::iset					mGroups;
-	cif::iset					mMandatoryFields;
-	std::set<ValidateItem>		mItemValidators;
-
-	bool operator<(const ValidateCategory& rhs) const
-	{
-		return icompare(mName, rhs.mName) < 0;
-	}
-
-	void addItemValidator(ValidateItem&& v);
-	
-	const ValidateItem* getValidatorForItem(std::string tag) const;
-	
-	const std::set<ValidateItem>& itemValidators() const
-	{
-		return mItemValidators;
-	}
-};
-
-struct ValidateLink
-{
-	int							mLinkGroupID;
-	std::string					mParentCategory;
-	std::vector<std::string>	mParentKeys;
-	std::string					mChildCategory;
-	std::vector<std::string>	mChildKeys;
-	std::string					mLinkGroupLabel;
-};
-
-// --------------------------------------------------------------------
-
-class Validator
-{
-  public:
-	friend class DictParser;
-
-	Validator();
-	~Validator();
-
-	Validator(const Validator& rhs) = delete;
-	Validator& operator=(const Validator& rhs) = delete;
-	
-	Validator(Validator&& rhs);
-	Validator& operator=(Validator&& rhs);
-	
-	void addTypeValidator(ValidateType&& v);
-	const ValidateType* getValidatorForType(std::string typeCode) const;
-
-	void addCategoryValidator(ValidateCategory&& v);
-	const ValidateCategory* getValidatorForCategory(std::string category) const;
-
-	void addLinkValidator(ValidateLink&& v);
-	std::vector<const ValidateLink*> getLinksForParent(const std::string& category) const;
-	std::vector<const ValidateLink*> getLinksForChild(const std::string& category) const;
-
-	void reportError(const std::string& msg, bool fatal);
-	
-	std::string dictName() const					{ return mName; }
-	void dictName(const std::string& name)			{ mName = name; }
-
-	std::string dictVersion() const				{ return mVersion; }
-	void dictVersion(const std::string& version)	{ mVersion = version; }
-
-  private:
-
-	// name is fully qualified here:
-	ValidateItem* getValidatorForItem(std::string name) const;
-
-	std::string					mName;
-	std::string					mVersion;
-	bool						mStrict = false;
-//	std::set<uint32_t>			mSubCategories;
-	std::set<ValidateType>		mTypeValidators;
-	std::set<ValidateCategory>	mCategoryValidators;
-	std::vector<ValidateLink>	mLinkValidators;
-};
-
-}

include/cif++/Matrix.hpp

@@ -1,391 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright Maarten L. Hekkelman, Radboud University 2008-2011.
- * Copyright (c) 2021 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-// --------------------------------------------------------------------
-// uBlas compatible matrix types
-
-#pragma once
-
-#include <iostream>
-#include <vector>
-
-// matrix is m x n, addressing i,j is 0 <= i < m and 0 <= j < n
-// element m i,j is mapped to [i * n + j] and thus storage is row major
-
-template <typename T>
-class MatrixBase
-{
-  public:
-	using value_type = T;
-
-	virtual ~MatrixBase() {}
-
-	virtual uint32_t dim_m() const = 0;
-	virtual uint32_t dim_n() const = 0;
-
-	virtual value_type &operator()(uint32_t i, uint32_t j) { throw std::runtime_error("unimplemented method"); }
-	virtual value_type operator()(uint32_t i, uint32_t j) const = 0;
-
-	MatrixBase &operator*=(const value_type &rhs);
-
-	MatrixBase &operator-=(const value_type &rhs);
-};
-
-template <typename T>
-MatrixBase<T> &MatrixBase<T>::operator*=(const T &rhs)
-{
-	for (uint32_t i = 0; i < dim_m(); ++i)
-	{
-		for (uint32_t j = 0; j < dim_n(); ++j)
-		{
-			operator()(i, j) *= rhs;
-		}
-	}
-
-	return *this;
-}
-
-template <typename T>
-MatrixBase<T> &MatrixBase<T>::operator-=(const T &rhs)
-{
-	for (uint32_t i = 0; i < dim_m(); ++i)
-	{
-		for (uint32_t j = 0; j < dim_n(); ++j)
-		{
-			operator()(i, j) -= rhs;
-		}
-	}
-
-	return *this;
-}
-
-template <typename T>
-std::ostream &operator<<(std::ostream &lhs, const MatrixBase<T> &rhs)
-{
-	lhs << '[' << rhs.dim_m() << ',' << rhs.dim_n() << ']' << '(';
-	for (uint32_t i = 0; i < rhs.dim_m(); ++i)
-	{
-		lhs << '(';
-		for (uint32_t j = 0; j < rhs.dim_n(); ++j)
-		{
-			if (j > 0)
-				lhs << ',';
-			lhs << rhs(i, j);
-		}
-		lhs << ')';
-	}
-	lhs << ')';
-
-	return lhs;
-}
-
-template <typename T>
-class Matrix : public MatrixBase<T>
-{
-  public:
-	using value_type = T;
-
-	template <typename T2>
-	Matrix(const MatrixBase<T2> &m)
-		: m_m(m.dim_m())
-		, m_n(m.dim_n())
-	{
-		m_data = new value_type[m_m * m_n];
-		for (uint32_t i = 0; i < m_m; ++i)
-		{
-			for (uint32_t j = 0; j < m_n; ++j)
-				operator()(i, j) = m(i, j);
-		}
-	}
-
-	Matrix()
-		: m_data(nullptr)
-		, m_m(0)
-		, m_n(0)
-	{
-	}
-
-	Matrix(const Matrix &m)
-		: m_m(m.m_m)
-		, m_n(m.m_n)
-	{
-		m_data = new value_type[m_m * m_n];
-		std::copy(m.m_data, m.m_data + (m_m * m_n), m_data);
-	}
-
-	Matrix &operator=(const Matrix &m)
-	{
-		value_type *t = new value_type[m.m_m * m.m_n];
-		std::copy(m.m_data, m.m_data + (m.m_m * m.m_n), t);
-
-		delete[] m_data;
-		m_data = t;
-		m_m = m.m_m;
-		m_n = m.m_n;
-
-		return *this;
-	}
-
-	Matrix(uint32_t m, uint32_t n, T v = T())
-		: m_m(m)
-		, m_n(n)
-	{
-		m_data = new value_type[m_m * m_n];
-		std::fill(m_data, m_data + (m_m * m_n), v);
-	}
-
-	virtual ~Matrix()
-	{
-		delete[] m_data;
-	}
-
-	virtual uint32_t dim_m() const { return m_m; }
-	virtual uint32_t dim_n() const { return m_n; }
-
-	virtual value_type operator()(uint32_t i, uint32_t j) const
-	{
-		assert(i < m_m);
-		assert(j < m_n);
-		return m_data[i * m_n + j];
-	}
-
-	virtual value_type &operator()(uint32_t i, uint32_t j)
-	{
-		assert(i < m_m);
-		assert(j < m_n);
-		return m_data[i * m_n + j];
-	}
-
-	template <typename Func>
-	void each(Func f)
-	{
-		for (uint32_t i = 0; i < m_m * m_n; ++i)
-			f(m_data[i]);
-	}
-
-	template <typename U>
-	Matrix &operator/=(U v)
-	{
-		for (uint32_t i = 0; i < m_m * m_n; ++i)
-			m_data[i] /= v;
-
-		return *this;
-	}
-
-  private:
-	value_type *m_data;
-	uint32_t m_m, m_n;
-};
-
-// --------------------------------------------------------------------
-
-template <typename T>
-class SymmetricMatrix : public MatrixBase<T>
-{
-  public:
-	typedef typename MatrixBase<T>::value_type value_type;
-
-	SymmetricMatrix(uint32_t n, T v = T())
-		: m_owner(true)
-		, m_n(n)
-	{
-		uint32_t N = (m_n * (m_n + 1)) / 2;
-		m_data = new value_type[N];
-		std::fill(m_data, m_data + N, v);
-	}
-
-	SymmetricMatrix(const T *data, uint32_t n)
-		: m_owner(false)
-		, m_data(const_cast<T *>(data))
-		, m_n(n)
-	{
-	}
-
-	virtual ~SymmetricMatrix()
-	{
-		if (m_owner)
-			delete[] m_data;
-	}
-
-	virtual uint32_t dim_m() const { return m_n; }
-	virtual uint32_t dim_n() const { return m_n; }
-
-	T operator()(uint32_t i, uint32_t j) const;
-	virtual T &operator()(uint32_t i, uint32_t j);
-
-	// erase two rows, add one at the end (for neighbour joining)
-	void erase_2(uint32_t i, uint32_t j);
-
-	template <typename Func>
-	void each(Func f)
-	{
-		uint32_t N = (m_n * (m_n + 1)) / 2;
-
-		for (uint32_t i = 0; i < N; ++i)
-			f(m_data[i]);
-	}
-
-	template <typename U>
-	SymmetricMatrix &operator/=(U v)
-	{
-		uint32_t N = (m_n * (m_n + 1)) / 2;
-
-		for (uint32_t i = 0; i < N; ++i)
-			m_data[i] /= v;
-
-		return *this;
-	}
-
-  private:
-	bool m_owner;
-	value_type *m_data;
-	uint32_t m_n;
-};
-
-template <typename T>
-inline T SymmetricMatrix<T>::operator()(uint32_t i, uint32_t j) const
-{
-	return i < j
-	           ? m_data[(j * (j + 1)) / 2 + i]
-	           : m_data[(i * (i + 1)) / 2 + j];
-}
-
-template <typename T>
-inline T &SymmetricMatrix<T>::operator()(uint32_t i, uint32_t j)
-{
-	if (i > j)
-		std::swap(i, j);
-	assert(j < m_n);
-	return m_data[(j * (j + 1)) / 2 + i];
-}
-
-template <typename T>
-void SymmetricMatrix<T>::erase_2(uint32_t di, uint32_t dj)
-{
-	uint32_t s = 0, d = 0;
-	for (uint32_t i = 0; i < m_n; ++i)
-	{
-		for (uint32_t j = 0; j < i; ++j)
-		{
-			if (i != di and j != dj and i != dj and j != di)
-			{
-				if (s != d)
-					m_data[d] = m_data[s];
-				++d;
-			}
-
-			++s;
-		}
-	}
-
-	--m_n;
-}
-
-template <typename T>
-class IdentityMatrix : public MatrixBase<T>
-{
-  public:
-	typedef typename MatrixBase<T>::value_type value_type;
-
-	IdentityMatrix(uint32_t n)
-		: m_n(n)
-	{
-	}
-
-	virtual uint32_t dim_m() const { return m_n; }
-	virtual uint32_t dim_n() const { return m_n; }
-
-	virtual value_type operator()(uint32_t i, uint32_t j) const
-	{
-		value_type result = 0;
-		if (i == j)
-			result = 1;
-		return result;
-	}
-
-  private:
-	uint32_t m_n;
-};
-
-// --------------------------------------------------------------------
-// matrix functions
-
-template <typename T>
-Matrix<T> operator*(const MatrixBase<T> &lhs, const MatrixBase<T> &rhs)
-{
-	Matrix<T> result(std::min(lhs.dim_m(), rhs.dim_m()), std::min(lhs.dim_n(), rhs.dim_n()));
-
-	for (uint32_t i = 0; i < result.dim_m(); ++i)
-	{
-		for (uint32_t j = 0; j < result.dim_n(); ++j)
-		{
-			for (uint32_t li = 0, rj = 0; li < lhs.dim_m() and rj < rhs.dim_n(); ++li, ++rj)
-				result(i, j) += lhs(li, j) * rhs(i, rj);
-		}
-	}
-
-	return result;
-}
-
-template <typename T>
-Matrix<T> operator*(const MatrixBase<T> &lhs, T rhs)
-{
-	Matrix<T> result(lhs);
-	result *= rhs;
-
-	return result;
-}
-
-template <typename T>
-Matrix<T> operator-(const MatrixBase<T> &lhs, const MatrixBase<T> &rhs)
-{
-	Matrix<T> result(std::min(lhs.dim_m(), rhs.dim_m()), std::min(lhs.dim_n(), rhs.dim_n()));
-
-	for (uint32_t i = 0; i < result.dim_m(); ++i)
-	{
-		for (uint32_t j = 0; j < result.dim_n(); ++j)
-		{
-			result(i, j) = lhs(i, j) - rhs(i, j);
-		}
-	}
-
-	return result;
-}
-
-template <typename T>
-Matrix<T> operator-(const MatrixBase<T> &lhs, T rhs)
-{
-	Matrix<T> result(lhs.dim_m(), lhs.dim_n());
-	result -= rhs;
-	return result;
-}
-
-// template <typename T>
-// symmetric_matrix<T> hammingDistance(const MatrixBase<T> &lhs, T rhs);
-
-// template <typename T>
-// std::vector<T> sum(const MatrixBase<T> &m);

include/cif++/Point.hpp

@@ -1,428 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#pragma once
-
-#include <functional>
-
-#if HAVE_LIBCLIPPER
-#include <clipper/core/coords.h>
-#endif
-
-#include <boost/math/quaternion.hpp>
-
-namespace mmcif
-{
-
-typedef boost::math::quaternion<float>	Quaternion;
-
-const double
-	kPI = 3.141592653589793238462643383279502884;
-
-// --------------------------------------------------------------------
-
-//	Point, a location with x, y and z coordinates as floating point.
-//	This one is derived from a tuple<float,float,float> so
-//	you can do things like:
-//
-//	float x, y, z;
-//	tie(x, y, z) = atom.loc();
-
-template<typename F>
-struct PointF
-{
-	typedef F FType;
-
-	FType mX, mY, mZ;
-	
-	PointF()							: mX(0), mY(0), mZ(0) {}
-	PointF(FType x, FType y, FType z)	: mX(x), mY(y), mZ(z) {}
-
-	template<typename PF>
-	PointF(const PointF<PF>& pt)
-		: mX(static_cast<F>(pt.mX))
-		, mY(static_cast<F>(pt.mY))
-		, mZ(static_cast<F>(pt.mZ)) {}
-
-#if HAVE_LIBCLIPPER	
-	PointF(const clipper::Coord_orth& pt): mX(pt[0]), mY(pt[1]), mZ(pt[2]) {}
-
-	PointF& operator=(const clipper::Coord_orth& rhs)
-	{
-		mX = rhs[0];
-		mY = rhs[1];
-		mZ = rhs[2];
-		return *this;
-	}
-#endif
-
-	template<typename PF>
-	PointF& operator=(const PointF<PF>& rhs)
-	{
-		mX = static_cast<F>(rhs.mX);
-		mY = static_cast<F>(rhs.mY);
-		mZ = static_cast<F>(rhs.mZ);
-		return *this;
-	}
-	
-	FType& getX()			{ return mX; }
-	FType getX() const		{ return mX; }
-	void setX(FType x)		{ mX = x; }
-
-	FType& getY()			{ return mY; }
-	FType getY() const		{ return mY; }
-	void setY(FType y)		{ mY = y; }
-
-	FType& getZ()			{ return mZ; }
-	FType getZ() const		{ return mZ; }
-	void setZ(FType z)		{ mZ = z; }
-	
-	PointF& operator+=(const PointF& rhs)
-	{
-		mX += rhs.mX;
-		mY += rhs.mY;
-		mZ += rhs.mZ;
-		
-		return *this;
-	}
-	
-	PointF& operator+=(FType d)
-	{
-		mX += d;
-		mY += d;
-		mZ += d;
-		
-		return *this;
-	}
-
-	PointF& operator-=(const PointF& rhs)
-	{
-		mX -= rhs.mX;
-		mY -= rhs.mY;
-		mZ -= rhs.mZ;
-		
-		return *this;
-	}
-
-	PointF& operator-=(FType d)
-	{
-		mX -= d;
-		mY -= d;
-		mZ -= d;
-		
-		return *this;
-	}
-
-	PointF& operator*=(FType rhs)
-	{
-		mX *= rhs;
-		mY *= rhs;
-		mZ *= rhs;
-		return *this;
-	}
-	
-	PointF& operator/=(FType rhs)
-	{
-		mX /= rhs;
-		mY /= rhs;
-		mZ /= rhs;
-		return *this;
-	}
-
-	FType normalize()
-	{
-		auto length = mX * mX + mY * mY + mZ * mZ;
-		if (length > 0)
-		{
-			length = std::sqrt(length);
-			operator/=(length);
-		}
-		return length;
-	}
-	
-	void rotate(const boost::math::quaternion<FType>& q)
-	{
-		boost::math::quaternion<FType> p(0, mX, mY, mZ);
-		
-		p = q * p * boost::math::conj(q);
-	
-		mX = p.R_component_2();
-		mY = p.R_component_3();
-		mZ = p.R_component_4();
-	}
-	
-#if HAVE_LIBCLIPPER
-	operator clipper::Coord_orth() const
-	{
-		return clipper::Coord_orth(mX, mY, mZ);
-	}
-#endif
-
-	operator std::tuple<const FType&, const FType&, const FType&>() const
-	{
-		return std::make_tuple(std::ref(mX), std::ref(mY), std::ref(mZ));
-	}
-
-	operator std::tuple<FType&,FType&,FType&>()
-	{
-		return std::make_tuple(std::ref(mX), std::ref(mY), std::ref(mZ));
-	}
-	
-	bool operator==(const PointF& rhs) const
-	{
-		return mX == rhs.mX and mY == rhs.mY and mZ == rhs.mZ;
-	}
-	
-	// consider point as a vector... perhaps I should rename Point?
-	FType lengthsq() const
-	{
-		return mX * mX + mY * mY + mZ * mZ;
-	}
-
-	FType length() const
-	{
-		return sqrt(mX * mX + mY * mY + mZ * mZ);
-	}
-};
-
-typedef PointF<float> Point;
-typedef PointF<double> DPoint;
-
-template<typename F>
-inline std::ostream& operator<<(std::ostream& os, const PointF<F>& pt)
-{
-	os << '(' << pt.mX << ',' << pt.mY << ',' << pt.mZ << ')';
-	return os; 
-}
-
-template<typename F>
-inline PointF<F> operator+(const PointF<F>& lhs, const PointF<F>& rhs)
-{
-	return PointF<F>(lhs.mX + rhs.mX, lhs.mY + rhs.mY, lhs.mZ + rhs.mZ);
-}
-
-template<typename F>
-inline PointF<F> operator-(const PointF<F>& lhs, const PointF<F>& rhs)
-{
-	return PointF<F>(lhs.mX - rhs.mX, lhs.mY - rhs.mY, lhs.mZ - rhs.mZ);
-}
-
-template<typename F>
-inline PointF<F> operator-(const PointF<F>& pt)
-{
-	return PointF<F>(-pt.mX, -pt.mY, -pt.mZ);
-}
-
-template<typename F>
-inline PointF<F> operator*(const PointF<F>& pt, F f)
-{
-	return PointF<F>(pt.mX * f, pt.mY * f, pt.mZ * f);
-}
-
-template<typename F>
-inline PointF<F> operator*(F f, const PointF<F>& pt)
-{
-	return PointF<F>(pt.mX * f, pt.mY * f, pt.mZ * f);
-}
-
-template<typename F>
-inline PointF<F> operator/(const PointF<F>& pt, F f)
-{
-	return PointF<F>(pt.mX / f, pt.mY / f, pt.mZ / f);
-}
-
-// --------------------------------------------------------------------
-// several standard 3d operations
-
-template<typename F>
-inline double DistanceSquared(const PointF<F>& a, const PointF<F>& b)
-{
-	return
-		(a.mX - b.mX) * (a.mX - b.mX) +
-		(a.mY - b.mY) * (a.mY - b.mY) +
-		(a.mZ - b.mZ) * (a.mZ - b.mZ);
-}
-
-template<typename F>
-inline double Distance(const PointF<F>& a, const PointF<F>& b)
-{
-	return sqrt(
-		(a.mX - b.mX) * (a.mX - b.mX) +
-		(a.mY - b.mY) * (a.mY - b.mY) +
-		(a.mZ - b.mZ) * (a.mZ - b.mZ));
-}
-
-template<typename F>
-inline F DotProduct(const PointF<F>& a, const PointF<F>& b)
-{
-	return a.mX * b.mX + a.mY * b.mY + a.mZ * b.mZ;
-}
-
-template<typename F>
-inline PointF<F> CrossProduct(const PointF<F>& a, const PointF<F>& b)
-{
-	return PointF<F>(a.mY * b.mZ - b.mY * a.mZ,
-				  a.mZ * b.mX - b.mZ * a.mX,
-				  a.mX * b.mY - b.mX * a.mY);
-}
-
-template<typename F>
-double Angle(const PointF<F>& p1, const PointF<F>& p2, const PointF<F>& p3)
-{
-	PointF<F> v1 = p1 - p2;
-	PointF<F> v2 = p3 - p2;
-	
-	return std::acos(DotProduct(v1, v2) / (v1.length() * v2.length())) * 180 / kPI;
-}
-
-template<typename F>
-double DihedralAngle(const PointF<F>& p1, const PointF<F>& p2, const PointF<F>& p3, const PointF<F>& p4)
-{
-	PointF<F> v12 = p1 - p2;	// vector from p2 to p1
-	PointF<F> v43 = p4 - p3;	// vector from p3 to p4
-	
-	PointF<F> z = p2 - p3;		// vector from p3 to p2
-	
-	PointF<F> p = CrossProduct(z, v12);
-	PointF<F> x = CrossProduct(z, v43);
-	PointF<F> y = CrossProduct(z, x);
-	
-	double u = DotProduct(x, x);
-	double v = DotProduct(y, y);
-	
-	double result = 360;
-	if (u > 0 and v > 0)
-	{
-		u = DotProduct(p, x) / sqrt(u);
-		v = DotProduct(p, y) / sqrt(v);
-		if (u != 0 or v != 0)
-			result = atan2(v, u) * 180 / kPI;
-	}
-	
-	return result;
-}
-
-template<typename F>
-double CosinusAngle(const PointF<F>& p1, const PointF<F>& p2, const PointF<F>& p3, const PointF<F>& p4)
-{
-	PointF<F> v12 = p1 - p2;
-	PointF<F> v34 = p3 - p4;
-	
-	double result = 0;
-	
-	double x = DotProduct(v12, v12) * DotProduct(v34, v34);
-	if (x > 0)
-		result = DotProduct(v12, v34) / sqrt(x);
-	
-	return result;
-}
-
-template<typename F>
-auto DistancePointToLine(const PointF<F> &l1, const PointF<F> &l2, const PointF<F> &p)
-{
-	auto line       = l2 - l1;
-    auto p_to_l1    = p - l1;
-    auto p_to_l2    = p - l2;
-    auto cross      = CrossProduct(p_to_l1, p_to_l2);
-    return cross.length() / line.length();
-}
-
-// --------------------------------------------------------------------
-// For e.g. simulated annealing, returns a new point that is moved in
-// a random direction with a distance randomly chosen from a normal
-// distribution with a stddev of offset.
-
-template<typename F>
-PointF<F> Nudge(PointF<F> p, F offset);
-
-// --------------------------------------------------------------------
-// We use quaternions to do rotations in 3d space
-
-Quaternion Normalize(Quaternion q);
-
-std::tuple<double,Point> QuaternionToAngleAxis(Quaternion q);
-Point Centroid(std::vector<Point>& Points);
-Point CenterPoints(std::vector<Point>& Points);
-Quaternion AlignPoints(const std::vector<Point>& a, const std::vector<Point>& b);
-double RMSd(const std::vector<Point>& a, const std::vector<Point>& b);
-
-// --------------------------------------------------------------------
-// Helper class to generate evenly divided Points on a sphere
-// we use a fibonacci sphere to calculate even distribution of the dots
-
-template<int N>
-class SphericalDots
-{
-  public:
-	enum { P = 2 * N + 1 };
-	typedef typename std::array<Point,P>	array_type;
-	typedef typename array_type::const_iterator	iterator;
-
-	static SphericalDots& instance()
-	{
-		static SphericalDots sInstance;
-		return sInstance;
-	}
-	
-	size_t size() const							{ return mPoints.size(); }
-	const Point operator[](uint32_t inIx) const	{ return mPoints[inIx]; }
-	iterator begin() const						{ return mPoints.begin(); }
-	iterator end() const						{ return mPoints.end(); }
-
-	double weight() const						{ return mWeight; }
-
-	SphericalDots()
-	{
-				
-		const double
-			kGoldenRatio = (1 + std::sqrt(5.0)) / 2;
-		
-		mWeight = (4 * kPI) / P;
-		
-		auto p = mPoints.begin();
-		
-		for (int32_t i = -N; i <= N; ++i)
-		{
-			double lat = std::asin((2.0 * i) / P);
-			double lon = std::fmod(i, kGoldenRatio) * 2 * kPI / kGoldenRatio;
-			
-			p->mX = sin(lon) * cos(lat);
-			p->mY = cos(lon) * cos(lat);
-			p->mZ =            sin(lat);
-
-			++p;
-		}
-	}
-
-  private:
-
-	array_type				mPoints;
-	double					mWeight;
-};
-
-typedef SphericalDots<50> SphericalDots_50;
-
-}

include/cif++/Secondary.hpp

@@ -1,218 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-// Calculate DSSP-like secondary structure information
-
-#pragma once
-
-namespace mmcif
-{
-	
-class Structure;
-class Monomer;
-
-struct Res;
-
-extern const float
-	kCouplingConstant, kMinHBondEnergy, kMaxHBondEnergy;
-
-enum SecondaryStructureType : char
-{
-	ssLoop			= ' ',
-	ssAlphahelix	= 'H',
-	ssBetabridge	= 'B',
-	ssStrand		= 'E',
-	ssHelix_3		= 'G',
-	ssHelix_5		= 'I',
-	ssHelix_PPII	= 'P',
-	ssTurn			= 'T',
-	ssBend			= 'S'
-};
-
-enum class HelixType
-{
-	rh_3_10, rh_alpha, rh_pi, rh_pp
-};
-
-enum class Helix
-{
-	None, Start, End, StartAndEnd, Middle
-};
-
-//struct HBond
-//{
-//	std::string 				labelAsymID;
-//	int							labelSeqID;
-//	double						energy;
-//};
-//
-//struct BridgePartner
-//{
-//	std::string					labelAsymID;
-//	int							labelSeqID;
-//	int							ladder;
-//	bool						parallel;
-//};
-
-struct SecondaryStructure
-{
-	SecondaryStructureType		type;
-//	HBond						donor[2], acceptor[2];
-//	BridgePartner				beta[2];
-//	int							sheet;
-//	bool						bend;
-};
-
-//void CalculateSecondaryStructure(Structure& s);
-
-const size_t
-	kHistogramSize = 30;
-
-struct DSSP_Statistics
-{
-	uint32_t nrOfResidues, nrOfChains, nrOfSSBridges, nrOfIntraChainSSBridges, nrOfHBonds;
-	uint32_t nrOfHBondsInAntiparallelBridges, nrOfHBondsInParallelBridges;
-	uint32_t nrOfHBondsPerDistance[11] = {};
-	double accessibleSurface = 0;
-
-	uint32_t residuesPerAlphaHelixHistogram[kHistogramSize] = {};
-	uint32_t parallelBridgesPerLadderHistogram[kHistogramSize] = {};
-	uint32_t antiparallelBridgesPerLadderHistogram[kHistogramSize] = {};
-	uint32_t laddersPerSheetHistogram[kHistogramSize] = {};
-};
-
-enum class ChainBreak
-{
-	None, NewChain, Gap
-};
-
-class DSSP
-{
-  public:
-	DSSP(const Structure& s, int min_poly_proline_stretch_length, bool calculateSurfaceAccessibility);
-	~DSSP();
-	
-	DSSP(const DSSP&) = delete;
-	DSSP& operator=(const DSSP&) = delete;
-	
-	SecondaryStructureType operator()(const std::string& inAsymID, int inSeqID) const;
-	SecondaryStructureType operator()(const Monomer& m) const;
-	
-	double accessibility(const std::string& inAsymID, int inSeqID) const;
-	double accessibility(const Monomer& m) const;
-
-	bool isAlphaHelixEndBeforeStart(const Monomer& m) const;
-	bool isAlphaHelixEndBeforeStart(const std::string& inAsymID, int inSeqID) const;
-
-	DSSP_Statistics GetStatistics() const;
-
-	class iterator;
-	using res_iterator = typename std::vector<Res>::iterator;
-
-	class ResidueInfo
-	{
-	  public:
-		friend class iterator;
-
-		explicit operator bool() const		{ return not empty(); }
-		bool empty() const					{ return mImpl == nullptr; }
-
-		const Monomer& residue() const;
-		std::string alt_id() const;
-
-		/// \brief return 0 if not a break, ' ' in case of a new chain and '*' in case of a broken chain
-		ChainBreak chainBreak() const;
-
-		/// \brief the internal number in DSSP
-		int nr() const;
-
-		SecondaryStructureType ss() const;
-		
-		int ssBridgeNr() const;
-
-		Helix helix(HelixType helixType) const;
-
-		bool bend() const;
-
-		double accessibility() const;
-
-		/// \brief returns resinfo, ladder and parallel
-		std::tuple<ResidueInfo,int,bool> bridgePartner(int i) const;
-
-		int sheet() const;
-
-		/// \brief return resinfo and the energy of the bond
-		std::tuple<ResidueInfo,double> acceptor(int i) const;
-		std::tuple<ResidueInfo,double> donor(int i) const;
-
-	  private:
-		ResidueInfo(Res* res) : mImpl(res) {}
-
-		Res* mImpl;
-	};
-
-	class iterator
-	{
-	  public:
-		using iterator_category = std::input_iterator_tag;
-		using value_type = ResidueInfo;
-		using difference_type = std::ptrdiff_t;
-		using pointer = value_type*;
-		using reference = value_type&;
-
-		iterator(const iterator& i);
-		iterator(Res* res);
-		iterator& operator=(const iterator& i);
-
-		reference operator*()		{ return mCurrent; }
-		pointer operator->()		{ return &mCurrent; }
-
-		iterator& operator++();
-		iterator operator++(int)
-		{
-			auto tmp(*this);
-			this->operator++();
-			return tmp;
-		}
-
-		bool operator==(const iterator& rhs) const		{ return mCurrent.mImpl == rhs.mCurrent.mImpl; }
-		bool operator!=(const iterator& rhs) const		{ return mCurrent.mImpl != rhs.mCurrent.mImpl; }
-
-	  private:
-		ResidueInfo	mCurrent;
-	};
-
-	iterator begin() const;
-	iterator end() const;
-
-	bool empty() const		{ return begin() == end(); }
-
-  private:
-	struct DSSPImpl* mImpl;
-};
-
-
-}

include/cif++/Structure.hpp

@@ -1,546 +0,0 @@
-/*-
- * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
- * 
- * 1. Redistributions of source code must retain the above copyright notice, this
- *    list of conditions and the following disclaimer
- * 2. Redistributions in binary form must reproduce the above copyright notice,
- *    this list of conditions and the following disclaimer in the documentation
- *    and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
- * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
- * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
- * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
- * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#pragma once
-
-#include <numeric>
-
-#include "cif++/AtomType.hpp"
-#include "cif++/Cif++.hpp"
-#include "cif++/Compound.hpp"
-#include "cif++/Point.hpp"
-
-/*
-	To modify a structure, you will have to use actions.
-	
-	The currently supported actions are:
-	
-//	- Move atom to new location
-	- Remove atom
-//	- Add new atom that was formerly missing
-//	- Add alternate Residue
-	- 
-	
-*/
-
-namespace mmcif
-{
-
-class Atom;
-class Residue;
-class Monomer;
-class Polymer;
-class Structure;
-class File;
-
-// --------------------------------------------------------------------
-
-class Atom
-{
-  public:
-	Atom();
-	Atom(struct AtomImpl *impl);
-	Atom(const Atom &rhs);
-
-	Atom(cif::Datablock &db, cif::Row &row);
-
-	// a special constructor to create symmetry copies
-	Atom(const Atom &rhs, const Point &symmmetry_location, const std::string &symmetry_operation);
-
-	~Atom();
-
-	explicit operator bool() const { return mImpl_ != nullptr; }
-
-	// return a copy of this atom, with data copied instead of referenced
-	Atom clone() const;
-
-	Atom &operator=(const Atom &rhs);
-
-	const std::string &id() const;
-	AtomType type() const;
-
-	Point location() const;
-	void location(Point p);
-
-	/// \brief Translate the position of this atom by \a t
-	void translate(Point t);
-
-	/// \brief Rotate the position of this atom by \a q
-	void rotate(Quaternion q);
-
-	// for direct access to underlying data, be careful!
-	const cif::Row getRow() const;
-	const cif::Row getRowAniso() const;
-
-	// Atom symmetryCopy(const Point& d, const clipper::RTop_orth& rt);
-	bool isSymmetryCopy() const;
-	std::string symmetry() const;
-	// const clipper::RTop_orth& symop() const;
-
-	const Compound &comp() const;
-	bool isWater() const;
-	int charge() const;
-
-	float uIso() const;
-	bool getAnisoU(float anisou[6]) const;
-	float occupancy() const;
-
-	template <typename T>
-	T property(const std::string &name) const;
-
-	void property(const std::string &name, const std::string &value);
-
-	template <typename T, std::enable_if_t<std::is_arithmetic_v<T>, int> = 0>
-	void property(const std::string &name, const T &value)
-	{
-		property(name, std::to_string(value));
-	}
-
-	// specifications
-	std::string labelAtomID() const;
-	std::string labelCompID() const;
-	std::string labelAsymID() const;
-	std::string labelEntityID() const;
-	int labelSeqID() const;
-	std::string labelAltID() const;
-	bool isAlternate() const;
-
-	std::string authAtomID() const;
-	std::string authCompID() const;
-	std::string authAsymID() const;
-	std::string authSeqID() const;
-	std::string pdbxAuthInsCode() const;
-	std::string pdbxAuthAltID() const;
-
-	std::string labelID() const; // label_comp_id + '_' + label_asym_id + '_' + label_seq_id
-	std::string pdbID() const;   // auth_comp_id + '_' + auth_asym_id + '_' + auth_seq_id + pdbx_PDB_ins_code
-
-	bool operator==(const Atom &rhs) const;
-
-	// // get clipper format Atom
-	// clipper::Atom toClipper() const;
-
-	// Radius calculation based on integrating the density until perc of electrons is found
-	void calculateRadius(float resHigh, float resLow, float perc);
-	float radius() const;
-
-	// access data in compound for this atom
-
-	// convenience routine
-	bool isBackBone() const
-	{
-		auto atomID = labelAtomID();
-		return atomID == "N" or atomID == "O" or atomID == "C" or atomID == "CA";
-	}
-
-	void swap(Atom &b)
-	{
-		std::swap(mImpl_, b.mImpl_);
-	}
-
-	int compare(const Atom &b) const;
-
-	bool operator<(const Atom &rhs) const
-	{
-		return compare(rhs) < 0;
-	}
-
-	friend std::ostream &operator<<(std::ostream &os, const Atom &atom);
-
-  private:
-	friend class Structure;
-	void setID(int id);
-
-	AtomImpl *impl();
-	const AtomImpl *impl() const;
-
-	struct AtomImpl *mImpl_;
-};
-
-inline void swap(mmcif::Atom &a, mmcif::Atom &b)
-{
-	a.swap(b);
-}
-
-inline double Distance(const Atom &a, const Atom &b)
-{
-	return Distance(a.location(), b.location());
-}
-
-inline double DistanceSquared(const Atom &a, const Atom &b)
-{
-	return DistanceSquared(a.location(), b.location());
-}
-
-typedef std::vector<Atom> AtomView;
-
-// --------------------------------------------------------------------
-
-class Residue
-{
-  public:
-	// constructors should be private, but that's not possible for now (needed in emplace)
-
-	// constructor for waters
-	Residue(const Structure &structure, const std::string &compoundID,
-		const std::string &asymID, const std::string &authSeqID);
-
-	// constructor for a residue without a sequence number
-	Residue(const Structure &structure, const std::string &compoundID,
-		const std::string &asymID);
-
-	// constructor for a residue with a sequence number
-	Residue(const Structure &structure, const std::string &compoundID,
-		const std::string &asymID, int seqID, const std::string &authSeqID);
-
-	Residue(const Residue &rhs) = delete;
-	Residue &operator=(const Residue &rhs) = delete;
-
-	Residue(Residue &&rhs);
-	Residue &operator=(Residue &&rhs);
-
-	virtual ~Residue();
-
-	const Compound &compound() const;
-	const AtomView &atoms() const;
-
-	/// \brief Unique atoms returns only the atoms without alternates and the first of each alternate atom id.
-	AtomView unique_atoms() const;
-
-	/// \brief The alt ID used for the unique atoms
-	std::string unique_alt_id() const;
-
-	Atom atomByID(const std::string &atomID) const;
-
-	const std::string &compoundID() const { return mCompoundID; }
-	const std::string &asymID() const { return mAsymID; }
-	int seqID() const { return mSeqID; }
-	std::string entityID() const;
-
-	std::string authAsymID() const;
-	std::string authSeqID() const;
-	std::string authInsCode() const;
-
-	// return a human readable PDB-like auth id (chain+seqnr+iCode)
-	std::string authID() const;
-
-	// similar for mmCIF space
-	std::string labelID() const;
-
-	// Is this residue a single entity?
-	bool isEntity() const;
-
-	bool isWater() const { return mCompoundID == "HOH"; }
-
-	const Structure &structure() const { return *mStructure; }
-
-	bool empty() const { return mStructure == nullptr; }
-
-	bool hasAlternateAtoms() const;
-
-	/// \brief Return the list of unique alt ID's present in this residue
-	std::set<std::string> getAlternateIDs() const;
-
-	/// \brief Return the list of unique atom ID's
-	std::set<std::string> getAtomIDs() const;
-
-	/// \brief Return the list of atoms having ID \a atomID
-	AtomView getAtomsByID(const std::string &atomID) const;
-
-	// some routines for 3d work
-	std::tuple<Point, float> centerAndRadius() const;
-
-	friend std::ostream &operator<<(std::ostream &os, const Residue &res);
-
-  protected:
-	Residue() {}
-
-	friend class Polymer;
-
-	const Structure *mStructure = nullptr;
-	std::string mCompoundID, mAsymID;
-	int mSeqID = 0;
-
-	// Watch out, this is used only to label waters... The rest of the code relies on
-	// MapLabelToAuth to get this info. Perhaps we should rename this member field.
-	std::string mAuthSeqID;
-	AtomView mAtoms;
-};
-
-// --------------------------------------------------------------------
-// a monomer models a single Residue in a protein chain
-
-class Monomer : public Residue
-{
-  public:
-	//	Monomer();
-	Monomer(const Monomer &rhs) = delete;
-	Monomer &operator=(const Monomer &rhs) = delete;
-
-	Monomer(Monomer &&rhs);
-	Monomer &operator=(Monomer &&rhs);
-
-	Monomer(const Polymer &polymer, size_t index, int seqID, const std::string &authSeqID,
-		const std::string &compoundID);
-
-	bool is_first_in_chain() const;
-	bool is_last_in_chain() const;
-
-	// convenience
-	bool has_alpha() const;
-	bool has_kappa() const;
-
-	// Assuming this is really an amino acid...
-
-	float phi() const;
-	float psi() const;
-	float alpha() const;
-	float kappa() const;
-	float tco() const;
-	float omega() const;
-
-	// torsion angles
-	size_t nrOfChis() const;
-	float chi(size_t i) const;
-
-	bool isCis() const;
-
-	/// \brief Returns true if the four atoms C, CA, N and O are present
-	bool isComplete() const;
-
-	/// \brief Returns true if any of the backbone atoms has an alternate
-	bool hasAlternateBackboneAtoms() const;
-
-	Atom CAlpha() const { return atomByID("CA"); }
-	Atom C() const { return atomByID("C"); }
-	Atom N() const { return atomByID("N"); }
-	Atom O() const { return atomByID("O"); }
-	Atom H() const { return atomByID("H"); }
-
-	bool isBondedTo(const Monomer &rhs) const
-	{
-		return this != &rhs and areBonded(*this, rhs);
-	}
-
-	static bool areBonded(const Monomer &a, const Monomer &b, float errorMargin = 0.5f);
-	static bool isCis(const Monomer &a, const Monomer &b);
-	static float omega(const Monomer &a, const Monomer &b);
-
-	// for LEU and VAL
-	float chiralVolume() const;
-
-  private:
-	const Polymer *mPolymer;
-	size_t mIndex;
-};
-
-// --------------------------------------------------------------------
-
-class Polymer : public std::vector<Monomer>
-{
-  public:
-	Polymer(const Structure &s, const std::string &entityID, const std::string &asymID);
-
-	Polymer(const Polymer &) = delete;
-	Polymer &operator=(const Polymer &) = delete;
-
-	//	Polymer(Polymer&& rhs) = delete;
-	//	Polymer& operator=(Polymer&& rhs) = de;
-
-	Monomer &getBySeqID(int seqID);
-	const Monomer &getBySeqID(int seqID) const;
-
-	Structure *structure() const { return mStructure; }
-
-	std::string asymID() const { return mAsymID; }
-	std::string entityID() const { return mEntityID; }
-
-	std::string chainID() const;
-
-	int Distance(const Monomer &a, const Monomer &b) const;
-
-  private:
-	Structure *mStructure;
-	std::string mEntityID;
-	std::string mAsymID;
-	cif::RowSet mPolySeq;
-};
-
-// --------------------------------------------------------------------
-// file is a reference to the data stored in e.g. the cif file.
-// This object is not copyable.
-
-class File : public std::enable_shared_from_this<File>
-{
-  public:
-	File();
-	File(const std::filesystem::path &path);
-	File(const char *data, size_t length); // good luck trying to find out what it is...
-	~File();
-
-	File(const File &) = delete;
-	File &operator=(const File &) = delete;
-
-	cif::Datablock& createDatablock(const std::string &name);
-
-	void load(const std::filesystem::path &path);
-	void save(const std::filesystem::path &path);
-
-	Structure *model(size_t nr = 1);
-
-	struct FileImpl &impl() const { return *mImpl; }
-
-	cif::Datablock &data();
-	cif::File &file();
-
-  private:
-	struct FileImpl *mImpl;
-};
-
-// --------------------------------------------------------------------
-
-enum class StructureOpenOptions
-{
-	SkipHydrogen = 1 << 0
-};
-
-inline bool operator&(StructureOpenOptions a, StructureOpenOptions b)
-{
-	return static_cast<int>(a) bitand static_cast<int>(b);
-}
-
-// --------------------------------------------------------------------
-
-class Structure
-{
-  public:
-	Structure(File &p, size_t modelNr = 1, StructureOpenOptions options = {});
-	Structure &operator=(const Structure &) = delete;
-	~Structure();
-
-	// Create a read-only clone of the current structure (for multithreaded calculations that move atoms)
-	Structure(const Structure &);
-
-	File &getFile() const;
-
-	const AtomView &atoms() const { return mAtoms; }
-	AtomView waters() const;
-
-	const std::list<Polymer> &polymers() const { return mPolymers; }
-	std::list<Polymer> &polymers() { return mPolymers; }
-
-	const std::vector<Residue> &nonPolymers() const { return mNonPolymers; }
-	const std::vector<Residue> &branchResidues() const { return mBranchResidues; }
-
-	Atom getAtomByID(std::string id) const;
-	// Atom getAtomByLocation(Point pt, float maxDistance) const;
-
-	Atom getAtomByLabel(const std::string &atomID, const std::string &asymID,
-		const std::string &compID, int seqID, const std::string &altID = "");
-
-	/// \brief Get a residue, if \a seqID is zero, the non-polymers are searched
-	const Residue &getResidue(const std::string &asymID, const std::string &compID, int seqID = 0) const;
-
-	// map between auth and label locations
-
-	std::tuple<std::string, int, std::string> MapAuthToLabel(const std::string &asymID,
-		const std::string &seqID, const std::string &compID, const std::string &insCode = "");
-
-	std::tuple<std::string, std::string, std::string, std::string> MapLabelToAuth(
-		const std::string &asymID, int seqID, const std::string &compID);
-
-	// returns chain, seqnr, icode
-	std::tuple<char, int, char> MapLabelToAuth(
-		const std::string &asymID, int seqID) const;
-
-	// returns chain,seqnr,comp,iCode
-	std::tuple<std::string, int, std::string, std::string> MapLabelToPDB(
-		const std::string &asymID, int seqID, const std::string &compID,
-		const std::string &authSeqID) const;
-
-	std::tuple<std::string, int, std::string> MapPDBToLabel(
-		const std::string &asymID, int seqID, const std::string &compID, const std::string &iCode) const;
-
-	// Actions
-	void removeAtom(Atom &a);
-	void swapAtoms(Atom &a1, Atom &a2); // swap the labels for these atoms
-	void moveAtom(Atom &a, Point p);    // move atom to a new location
-	void changeResidue(const Residue &res, const std::string &newCompound,
-		const std::vector<std::tuple<std::string, std::string>> &remappedAtoms);
-
-	/// \brief Create a new non-polymer entity, returns new ID
-	/// \param mon_id	The mon_id for the new nonpoly, must be an existing and known compound from CCD
-	/// \return			The ID of the created entity
-	std::string createNonPolyEntity(const std::string &mon_id);
-
-	/// \brief Create a new NonPolymer struct_asym with atoms constructed from \a atoms, returns asym_id.
-	/// This method assumes you are copying data from one cif file to another.
-	///
-	/// \param entity_id	The entity ID of the new nonpoly
-	/// \param atoms		The array of atom_site rows containing the data.
-	/// \return				The newly create asym ID
-	std::string createNonpoly(const std::string &entity_id, const std::vector<mmcif::Atom> &atoms);
-
-	/// \brief To sort the atoms in order of model > asym-id > res-id > atom-id
-	/// Will asssign new atom_id's to all atoms. Be carefull
-	void sortAtoms();
-
-	/// \brief Translate the coordinates of all atoms in the structure by \a t
-	void translate(Point t);
-
-	/// \brief Rotate the coordinates of all atoms in the structure by \a q
-	void rotate(Quaternion t);
-
-	const std::vector<Residue> &getNonPolymers() const { return mNonPolymers; }
-	const std::vector<Residue> &getBranchResidues() const { return mBranchResidues; }
-
-	void cleanupEmptyCategories();
-
-  private:
-	friend Polymer;
-	friend Residue;
-	// friend residue_view;
-	// friend residue_iterator;
-
-	cif::Category &category(const char *name) const;
-	cif::Datablock &datablock() const;
-
-	std::string insertCompound(const std::string &compoundID, bool isEntity);
-
-	void loadData();
-	void updateAtomIndex();
-
-	void loadAtomsForModel(StructureOpenOptions options);
-
-	File &mFile;
-	size_t mModelNr;
-	AtomView mAtoms;
-	std::vector<size_t> mAtomIndex;
-	std::list<Polymer> mPolymers;
-	std::vector<Residue> mNonPolymers, mBranchResidues;
-};
-
-} // namespace mmcif

include/cif++/atom_type.hpp

@@ -0,0 +1,277 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+// Lib for working with structures as contained in mmCIF and PDB files
+
+#pragma once
+
+#include <cstdint>
+#include <stdexcept>
+#include <string>
+
+namespace cif
+{
+
+enum atom_type : uint8_t
+{
+	Nn = 0, // Unknown
+
+	H = 1,  // Hydro­gen
+	He = 2, // He­lium
+
+	Li = 3,  // Lith­ium
+	Be = 4,  // Beryl­lium
+	B = 5,   // Boron
+	C = 6,   // Carbon
+	N = 7,   // Nitro­gen
+	O = 8,   // Oxy­gen
+	F = 9,   // Fluor­ine
+	Ne = 10, // Neon
+
+	Na = 11, // So­dium
+	Mg = 12, // Magne­sium
+	Al = 13, // Alumin­ium
+	Si = 14, // Sili­con
+	P = 15,  // Phos­phorus
+	S = 16,  // Sulfur
+	Cl = 17, // Chlor­ine
+	Ar = 18, // Argon
+
+	K = 19,  // Potas­sium
+	Ca = 20, // Cal­cium
+	Sc = 21, // Scan­dium
+	Ti = 22, // Tita­nium
+	V = 23,  // Vana­dium
+	Cr = 24, // Chrom­ium
+	Mn = 25, // Manga­nese
+	Fe = 26, // Iron
+	Co = 27, // Cobalt
+	Ni = 28, // Nickel
+	Cu = 29, // Copper
+	Zn = 30, // Zinc
+	Ga = 31, // Gallium
+	Ge = 32, // Germa­nium
+	As = 33, // Arsenic
+	Se = 34, // Sele­nium
+	Br = 35, // Bromine
+	Kr = 36, // Kryp­ton
+
+	Rb = 37, // Rubid­ium
+	Sr = 38, // Stront­ium
+	Y = 39,  // Yttrium
+	Zr = 40, // Zirco­nium
+	Nb = 41, // Nio­bium
+	Mo = 42, // Molyb­denum
+	Tc = 43, // Tech­netium
+	Ru = 44, // Ruthe­nium
+	Rh = 45, // Rho­dium
+	Pd = 46, // Pallad­ium
+	Ag = 47, // Silver
+	Cd = 48, // Cad­mium
+	In = 49, // Indium
+	Sn = 50, // Tin
+	Sb = 51, // Anti­mony
+	Te = 52, // Tellurium
+	I = 53,  // Iodine
+	Xe = 54, // Xenon
+	Cs = 55, // Cae­sium
+	Ba = 56, // Ba­rium
+	La = 57, // Lan­thanum
+
+	Hf = 72, // Haf­nium
+	Ta = 73, // Tanta­lum
+	W = 74,  // Tung­sten
+	Re = 75, // Rhe­nium
+	Os = 76, // Os­mium
+	Ir = 77, // Iridium
+	Pt = 78, // Plat­inum
+	Au = 79, // Gold
+	Hg = 80, // Mer­cury
+	Tl = 81, // Thallium
+	Pb = 82, // Lead
+	Bi = 83, // Bis­muth
+	Po = 84, // Polo­nium
+	At = 85, // Asta­tine
+	Rn = 86, // Radon
+	Fr = 87, // Fran­cium
+	Ra = 88, // Ra­dium
+	Ac = 89, // Actin­ium
+
+	Rf = 104, // Ruther­fordium
+	Db = 105, // Dub­nium
+	Sg = 106, // Sea­borgium
+	Bh = 107, // Bohr­ium
+	Hs = 108, // Has­sium
+	Mt = 109, // Meit­nerium
+	Ds = 110, // Darm­stadtium
+	Rg = 111, // Roent­genium
+	Cn = 112, // Coper­nicium
+	Nh = 113, // Nihon­ium
+	Fl = 114, // Flerov­ium
+	Mc = 115, // Moscov­ium
+	Lv = 116, // Liver­morium
+	Ts = 117, // Tenness­ine
+	Og = 118, // Oga­nesson
+
+	Ce = 58, // Cerium
+	Pr = 59, // Praseo­dymium
+	Nd = 60, // Neo­dymium
+	Pm = 61, // Prome­thium
+	Sm = 62, // Sama­rium
+	Eu = 63, // Europ­ium
+	Gd = 64, // Gadolin­ium
+	Tb = 65, // Ter­bium
+	Dy = 66, // Dyspro­sium
+	Ho = 67, // Hol­mium
+	Er = 68, // Erbium
+	Tm = 69, // Thulium
+	Yb = 70, // Ytter­bium
+	Lu = 71, // Lute­tium
+
+	Th = 90,  // Thor­ium
+	Pa = 91,  // Protac­tinium
+	U = 92,   // Ura­nium
+	Np = 93,  // Neptu­nium
+	Pu = 94,  // Pluto­nium
+	Am = 95,  // Ameri­cium
+	Cm = 96,  // Curium
+	Bk = 97,  // Berkel­ium
+	Cf = 98,  // Califor­nium
+	Es = 99,  // Einstei­nium
+	Fm = 100, // Fer­mium
+	Md = 101, // Mende­levium
+	No = 102, // Nobel­ium
+	Lr = 103, // Lawren­cium
+
+	D = 119, // Deuterium
+};
+
+// --------------------------------------------------------------------
+// atom_type_info
+
+enum class radius_type
+{
+	calculated,
+	empirical,
+	covalent_empirical,
+
+	single_bond,
+	double_bond,
+	triple_bond,
+
+	van_der_waals,
+
+	type_count
+};
+
+constexpr size_t kRadiusTypeCount = static_cast<size_t>(radius_type::type_count);
+
+enum class ionic_radius_type
+{
+	effective, crystal
+};
+
+struct atom_type_info
+{
+	atom_type type;
+	std::string name;
+	std::string symbol;
+	float weight;
+	bool metal;
+	float radii[kRadiusTypeCount];
+};
+
+extern const atom_type_info kKnownAtoms[];
+
+// --------------------------------------------------------------------
+// AtomTypeTraits
+
+class atom_type_traits
+{
+  public:
+	atom_type_traits(atom_type a);
+	atom_type_traits(const std::string &symbol);
+
+	atom_type type() const { return m_info->type; }
+	std::string name() const { return m_info->name; }
+	std::string symbol() const { return m_info->symbol; }
+	float weight() const { return m_info->weight; }
+
+	bool is_metal() const { return m_info->metal; }
+
+	static bool is_element(const std::string &symbol);
+	static bool is_metal(const std::string &symbol);
+
+	float radius(radius_type type = radius_type::single_bond) const
+	{
+		if (type >= radius_type::type_count)
+			throw std::invalid_argument("invalid radius requested");
+		return m_info->radii[static_cast<size_t>(type)] / 100.f;
+	}
+
+	/// \brief Return the radius for a charged version of this atom in a solid crystal
+	///
+	/// \param charge  The charge of the ion
+	/// \return        The radius of the ion
+	float crystal_ionic_radius(int charge) const;
+
+	/// \brief Return the radius for a charged version of this atom in a non-solid environment
+	///
+	/// \param charge  The charge of the ion
+	/// \return        The radius of the ion
+	float effective_ionic_radius(int charge) const;
+
+	/// \brief Return the radius for a charged version of this atom, returns the effective radius by default
+	///
+	/// \param charge  The charge of the ion
+	/// \return        The radius of the ion
+	float ionic_radius(int charge, ionic_radius_type type = ionic_radius_type::effective) const
+	{
+		return type == ionic_radius_type::effective ? effective_ionic_radius(charge) : crystal_ionic_radius(charge);
+	}
+
+	// data type encapsulating the Waasmaier & Kirfel scattering factors
+	// in a simplified form (only a and b).
+	// Added the electrion scattering factors as well
+	struct SFData
+	{
+		double a[6], b[6];
+	};
+
+	// to get the Cval and Siva values, use this constant as charge:
+	enum
+	{
+		kWKSFVal = -99
+	};
+
+	const SFData &wksf(int charge = 0) const;
+	const SFData &elsf() const;
+
+  private:
+	const struct atom_type_info *m_info;
+};
+
+} // namespace pdbx

include/cif++/category.hpp

@@ -0,0 +1,598 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <array>
+
+#include <cif++/forward_decl.hpp>
+
+#include <cif++/condition.hpp>
+#include <cif++/iterator.hpp>
+#include <cif++/row.hpp>
+#include <cif++/validate.hpp>
+
+// TODO: implement all of:
+// https://en.cppreference.com/w/cpp/named_req/Container
+// https://en.cppreference.com/w/cpp/named_req/SequenceContainer
+// and more?
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+// special exception
+class duplicate_key_error : public std::runtime_error
+{
+  public:
+	duplicate_key_error(const std::string &msg)
+		: std::runtime_error(msg) {}
+};
+
+// --------------------------------------------------------------------
+
+class category
+{
+  public:
+	friend class row_handle;
+
+	template <typename, typename...>
+	friend class iterator_impl;
+
+	using value_type = row_handle;
+	using reference = value_type;
+	using const_reference = const value_type;
+	using iterator = iterator_impl<category>;
+	using const_iterator = iterator_impl<const category>;
+
+	category() = default;
+
+	category(std::string_view name);
+
+	category(const category &rhs);
+
+	category(category &&rhs);
+
+	category &operator=(const category &rhs);
+
+	category &operator=(category &&rhs);
+
+	~category();
+
+	// --------------------------------------------------------------------
+
+	const std::string &name() const { return m_name; }
+
+	iset key_fields() const;
+
+	std::set<uint16_t> key_field_indices() const;
+
+	void set_validator(const validator *v, datablock &db);
+	void update_links(datablock &db);
+
+	const validator *get_validator() const { return m_validator; }
+	const category_validator *get_cat_validator() const { return m_cat_validator; }
+
+	bool is_valid() const;
+	bool validate_links() const;
+
+	bool operator==(const category &rhs) const;
+	bool operator!=(const category &rhs) const
+	{
+		return not operator==(rhs);
+	}
+
+	// --------------------------------------------------------------------
+
+	reference front()
+	{
+		return {*this, *m_head};
+	}
+
+	const_reference front() const
+	{
+		return {const_cast<category &>(*this), const_cast<row &>(*m_head)};
+	}
+
+	reference back()
+	{
+		return {*this, *m_tail};
+	}
+
+	const_reference back() const
+	{
+		return {const_cast<category &>(*this), const_cast<row &>(*m_tail)};
+	}
+
+	iterator begin()
+	{
+		return {*this, m_head};
+	}
+
+	iterator end()
+	{
+		return {*this, nullptr};
+	}
+
+	const_iterator begin() const
+	{
+		return {*this, m_head};
+	}
+
+	const_iterator end() const
+	{
+		return {*this, nullptr};
+	}
+
+	const_iterator cbegin() const
+	{
+		return {*this, m_head};
+	}
+
+	const_iterator cend() const
+	{
+		return {*this, nullptr};
+	}
+
+	size_t size() const
+	{
+		return std::distance(cbegin(), cend());
+	}
+
+	bool empty() const
+	{
+		return m_head == nullptr;
+	}
+
+	// --------------------------------------------------------------------
+	// A category can have a key, as defined by the validator/dictionary
+
+	/// @brief The key type
+	using key_type = row_initializer;
+
+	/// @brief Return a row_handle for the row specified by \a key
+	/// @param key The value for the key, fields specified in the dictionary should have a value
+	/// @return The row found in the index, or an undefined row_handle
+	row_handle operator[](const key_type &key);
+
+	const row_handle operator[](const key_type &key) const
+	{
+		return const_cast<category *>(this)->operator[](key);
+	}
+
+	// --------------------------------------------------------------------
+
+	template <typename... Ts, typename... Ns>
+	iterator_proxy<const category, Ts...> rows(Ns... names) const
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Ns), "The number of column titles should be equal to the number of types to return");
+		return iterator_proxy<const category, Ts...>(*this, begin(), {names...});
+	}
+
+	template <typename... Ts, typename... Ns>
+	iterator_proxy<category, Ts...> rows(Ns... names)
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Ns), "The number of column titles should be equal to the number of types to return");
+		return iterator_proxy<category, Ts...>(*this, begin(), {names...});
+	}
+
+	// --------------------------------------------------------------------
+
+	conditional_iterator_proxy<category> find(condition &&cond)
+	{
+		return find(begin(), std::forward<condition>(cond));
+	}
+
+	conditional_iterator_proxy<category> find(iterator pos, condition &&cond)
+	{
+		return {*this, pos, std::forward<condition>(cond)};
+	}
+
+	conditional_iterator_proxy<const category> find(condition &&cond) const
+	{
+		return find(cbegin(), std::forward<condition>(cond));
+	}
+
+	conditional_iterator_proxy<const category> find(const_iterator pos, condition &&cond) const
+	{
+		return conditional_iterator_proxy<const category>{*this, pos, std::forward<condition>(cond)};
+	}
+
+	template <typename... Ts, typename... Ns>
+	conditional_iterator_proxy<category, Ts...> find(condition &&cond, Ns... names)
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Ns), "The number of column titles should be equal to the number of types to return");
+		return find<Ts...>(cbegin(), std::forward<condition>(cond), std::forward<Ns>(names)...);
+	}
+
+	template <typename... Ts, typename... Ns>
+	conditional_iterator_proxy<const category, Ts...> find(condition &&cond, Ns... names) const
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Ns), "The number of column titles should be equal to the number of types to return");
+		return find<Ts...>(cbegin(), std::forward<condition>(cond), std::forward<Ns>(names)...);
+	}
+
+	template <typename... Ts, typename... Ns>
+	conditional_iterator_proxy<category, Ts...> find(const_iterator pos, condition &&cond, Ns... names)
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Ns), "The number of column titles should be equal to the number of types to return");
+		return {*this, pos, std::forward<condition>(cond), std::forward<Ns>(names)...};
+	}
+
+	template <typename... Ts, typename... Ns>
+	conditional_iterator_proxy<const category, Ts...> find(const_iterator pos, condition &&cond, Ns... names) const
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Ns), "The number of column titles should be equal to the number of types to return");
+		return {*this, pos, std::forward<condition>(cond), std::forward<Ns>(names)...};
+	}
+
+	// --------------------------------------------------------------------
+	// if you only expect a single row
+
+	row_handle find1(condition &&cond)
+	{
+		return find1(begin(), std::forward<condition>(cond));
+	}
+
+	row_handle find1(iterator pos, condition &&cond)
+	{
+		auto h = find(pos, std::forward<condition>(cond));
+
+		return h.size() != 1 ? row_handle{} : *h.begin();
+	}
+
+	const row_handle find1(condition &&cond) const
+	{
+		return find1(cbegin(), std::forward<condition>(cond));
+	}
+
+	const row_handle find1(const_iterator pos, condition &&cond) const
+	{
+		auto h = find(pos, std::forward<condition>(cond));
+
+		return h.size() != 1 ? row_handle{} : *h.begin();
+	}
+
+	template <typename T>
+	T find1(condition &&cond, const char *column) const
+	{
+		return find1<T>(cbegin(), std::forward<condition>(cond), column);
+	}
+
+	template <typename T>
+	T find1(const_iterator pos, condition &&cond, const char *column) const
+	{
+		auto h = find<T>(pos, std::forward<condition>(cond), column);
+
+		return h.size() == 1 ? *h.begin() : T{};
+	}
+
+	template <typename... Ts, typename... Cs, typename U = std::enable_if_t<sizeof...(Ts) != 1>>
+	std::tuple<Ts...> find1(condition &&cond, Cs... columns) const
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Cs), "The number of column titles should be equal to the number of types to return");
+		// static_assert(std::is_same_v<Cs, const char*>..., "The column names should be const char");
+		return find1<Ts...>(cbegin(), std::forward<condition>(cond), std::forward<Cs>(columns)...);
+	}
+
+	template <typename... Ts, typename... Cs, typename U = std::enable_if_t<sizeof...(Ts) != 1>>
+	std::tuple<Ts...> find1(const_iterator pos, condition &&cond, Cs... columns) const
+	{
+		static_assert(sizeof...(Ts) == sizeof...(Cs), "The number of column titles should be equal to the number of types to return");
+		auto h = find<Ts...>(pos, std::forward<condition>(cond), std::forward<Cs>(columns)...);
+
+		return h.size() == 1 ? *h.begin() : std::tuple<Ts...>{};
+	}
+
+	bool exists(condition &&cond) const
+	{
+		bool result = false;
+
+		if (cond)
+		{
+			cond.prepare(*this);
+
+			auto sh = cond.single();
+
+			if (sh.has_value() and *sh)
+				result = true;
+			else
+			{
+				for (auto r : *this)
+				{
+					if (cond(r))
+					{
+						result = true;
+						break;
+					}
+				}
+			}
+		}
+
+		return result;
+	}
+
+	// --------------------------------------------------------------------
+
+	bool has_children(row_handle r) const;
+	bool has_parents(row_handle r) const;
+
+	std::vector<row_handle> get_children(row_handle r, const category &childCat) const;
+	std::vector<row_handle> get_parents(row_handle r, const category &parentCat) const;
+	std::vector<row_handle> get_linked(row_handle r, const category &cat) const;
+
+	// --------------------------------------------------------------------
+
+	// void insert(const_iterator pos, const row_initializer &row)
+	// {
+	// 	insert_impl(pos, row);
+	// }
+
+	// void insert(const_iterator pos, row_initializer &&row)
+	// {
+	// 	insert_impl(pos, std::move(row));
+	// }
+
+	iterator erase(iterator pos);
+	void erase(row_handle rh)
+	{
+		erase(iterator(*this, rh.m_row));
+	}
+
+	size_t erase(condition &&cond);
+	size_t erase(condition &&cond, std::function<void(row_handle)> &&visit);
+
+	iterator emplace(row_initializer &&ri)
+	{
+		return this->emplace(ri.begin(), ri.end());
+	}
+
+	template <typename ItemIter>
+	iterator emplace(ItemIter b, ItemIter e)
+	{
+		row *r = this->create_row();
+
+		try
+		{
+			for (auto i = b; i != e; ++i)
+			{
+				// item_value *new_item = this->create_item(*i);
+				r->append(add_column(i->name()), { i->value() });
+			}
+		}
+		catch (...)
+		{
+			if (r != nullptr)
+				this->delete_row(r);
+			throw;
+		}
+
+		return insert_impl(cend(), r);
+	}
+
+	void clear();
+
+	// --------------------------------------------------------------------
+	/// \brief generate a new, unique ID. Pass it an ID generating function
+	/// based on a sequence number. This function will be called until the
+	/// result is unique in the context of this category
+	std::string get_unique_id(std::function<std::string(int)> generator = cif::cif_id_for_number);
+	std::string get_unique_id(const std::string &prefix)
+	{
+		return get_unique_id([prefix](int nr)
+			{ return prefix + std::to_string(nr + 1); });
+	}
+
+	// --------------------------------------------------------------------
+
+	/// \brief Rename a single column in the rows that match \a cond to value \a value
+	/// making sure the linked categories are updated according to the link.
+	/// That means, child categories are updated if the links are absolute
+	/// and unique. If they are not, the child category rows are split.
+
+	void update_value(condition &&cond, std::string_view tag, std::string_view value)
+	{
+		auto rs = find(std::move(cond));
+		std::vector<row_handle> rows;
+		std::copy(rs.begin(), rs.end(), std::back_inserter(rows));
+		update_value(rows, tag, value);
+	}
+
+	void update_value(const std::vector<row_handle> &rows, std::string_view tag, std::string_view value);
+
+	// --------------------------------------------------------------------
+	/// \brief Return the index number for \a column_name
+
+	uint16_t get_column_ix(std::string_view column_name) const
+	{
+		uint16_t result;
+
+		for (result = 0; result < m_columns.size(); ++result)
+		{
+			if (iequals(column_name, m_columns[result].m_name))
+				break;
+		}
+
+		if (VERBOSE > 0 and result == m_columns.size() and m_cat_validator != nullptr) // validate the name, if it is known at all (since it was not found)
+		{
+			auto iv = m_cat_validator->get_validator_for_item(column_name);
+			if (iv == nullptr)
+				std::cerr << "Invalid name used '" << column_name << "' is not a known column in " + m_name << std::endl;
+		}
+
+		return result;
+	}
+
+	std::string_view get_column_name(uint16_t ix) const
+	{
+		if (ix >= m_columns.size())
+			throw std::out_of_range("column index is out of range");
+
+		return m_columns[ix].m_name;
+	}
+
+	uint16_t add_column(std::string_view column_name)
+	{
+		using namespace std::literals;
+
+		uint16_t result = get_column_ix(column_name);
+
+		if (result == m_columns.size())
+		{
+			const item_validator *item_validator = nullptr;
+
+			if (m_cat_validator != nullptr)
+			{
+				item_validator = m_cat_validator->get_validator_for_item(column_name);
+				if (item_validator == nullptr)
+					m_validator->report_error("tag " + std::string(column_name) + " not allowed in category " + m_name, false);
+			}
+
+			m_columns.emplace_back(column_name, item_validator);
+		}
+
+		return result;
+	}
+
+	bool has_column(std::string_view name) const
+	{
+		return get_column_ix(name) < m_columns.size();
+	}
+
+	iset get_columns() const;
+
+	// --------------------------------------------------------------------
+
+	void sort(std::function<int(row_handle,row_handle)> f);
+	void reorder_by_index();
+
+	// --------------------------------------------------------------------
+
+	std::vector<std::string> get_tag_order() const;
+
+	void write(std::ostream &os) const;
+	void write(std::ostream &os, const std::vector<std::string> &order, bool addMissingColumns = true);
+
+  private:
+	void write(std::ostream &os, const std::vector<uint16_t> &order, bool includeEmptyColumns) const;
+
+  public:
+	friend std::ostream &operator<<(std::ostream &os, const category &cat)
+	{
+		cat.write(os);
+		return os;
+	}
+
+  private:
+	void update_value(row *row, uint16_t column, std::string_view value, bool updateLinked, bool validate = true);
+
+  private:
+	void erase_orphans(condition &&cond, category &parent);
+
+	using allocator_type = std::allocator<void>;
+
+	constexpr allocator_type get_allocator() const
+	{
+		return {};
+	}
+
+	using char_allocator_type = typename std::allocator_traits<allocator_type>::template rebind_alloc<char>;
+	using char_allocator_traits = std::allocator_traits<char_allocator_type>;
+
+	using row_allocator_type = typename std::allocator_traits<allocator_type>::template rebind_alloc<row>;
+	using row_allocator_traits = std::allocator_traits<row_allocator_type>;
+
+	row_allocator_traits::pointer get_row()
+	{
+		row_allocator_type ra(get_allocator());
+		return row_allocator_traits::allocate(ra, 1);
+	}
+
+	row *create_row()
+	{
+		auto p = this->get_row();
+		row_allocator_type ra(get_allocator());
+		row_allocator_traits::construct(ra, p);
+		return p;
+	}
+
+	row *clone_row(const row &r);
+
+	void delete_row(row *r);
+
+	row_handle create_copy(row_handle r);
+
+	struct item_column
+	{
+		std::string m_name;
+		const item_validator *m_validator;
+
+		item_column(std::string_view name, const item_validator *validator)
+			: m_name(name)
+			, m_validator(validator)
+		{
+		}
+	};
+
+	struct link
+	{
+		link(category *linked, const link_validator *v)
+			: linked(linked)
+			, v(v)
+		{
+		}
+
+		category *linked;
+		const link_validator *v;
+	};
+
+	// proxy methods for every insertion
+	iterator insert_impl(const_iterator pos, row *n);
+	iterator erase_impl(const_iterator pos);
+
+	// --------------------------------------------------------------------
+
+	condition get_parents_condition(row_handle rh, const category &parentCat) const;
+	condition get_children_condition(row_handle rh, const category &childCat) const;
+
+	// --------------------------------------------------------------------
+
+	void swap_item(uint16_t column_ix, row_handle &a, row_handle &b);
+
+	// --------------------------------------------------------------------
+	
+	std::string m_name;
+	std::vector<item_column> m_columns;
+	const validator *m_validator = nullptr;
+	const category_validator *m_cat_validator = nullptr;
+	std::vector<link> m_parent_links, m_child_links;
+	bool m_cascade = true;
+	uint32_t m_last_unique_num = 0;
+	class category_index *m_index = nullptr;
+	row *m_head = nullptr, *m_tail = nullptr;
+};
+
+} // namespace cif

include/cif++/compound.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
+ *
+ * Copyright (c) 2020-2022 NKI/AVL, Netherlands Cancer Institute
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
- * 
+ *
  * 1. Redistributions of source code must retain the above copyright notice, this
  *    list of conditions and the following disclaimer
  * 2. Redistributions in binary form must reproduce the above copyright notice,
  *    this list of conditions and the following disclaimer in the documentation
  *    and/or other materials provided with the distribution.
- * 
+ *
  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -26,7 +26,7 @@
 
 #pragma once
 
-/// \file This file contains the definition for the class Compound, encapsulating
+/// \file This file contains the definition for the class compound, encapsulating
 /// the information found for compounds in the CCD.
 
 #include <map>
@@ -34,20 +34,20 @@
 #include <tuple>
 #include <vector>
 
-#include "cif++/AtomType.hpp"
-#include "cif++/Cif++.hpp"
+#include <cif++.hpp>
+#include <cif++/atom_type.hpp>
 
-namespace mmcif
+namespace cif
 {
 
 // --------------------------------------------------------------------
 
-class Compound;
-struct CompoundAtom;
-class CompoundFactoryImpl;
+class compound;
+struct compound_atom;
+class compound_factory_impl;
 
-/// \brief The bond type as defined in the CCD, possible values taken from the mmcif_pdbx_v50 file
-enum class BondType
+/// \brief The bond type as defined in the CCD, possible values taken from the mmcif_pdbx file
+enum class bond_type
 {
 	sing, // 'single bond'
 	doub, // 'double bond'
@@ -59,32 +59,32 @@ enum class BondType
 	pi,   // 'pi bond'
 };
 
-std::string to_string(BondType bondType);
-BondType from_string(const std::string& bondType);
+std::string to_string(bond_type bondType);
+bond_type from_string(const std::string &bondType);
 
 /// --------------------------------------------------------------------
 /// \brief struct containing information about an atom in a chemical compound.
 /// This is a subset of the available information. Contact the author if you need more fields.
 
-struct CompoundAtom
+struct compound_atom
 {
 	std::string id;
-	AtomType typeSymbol;
+	atom_type type_symbol;
 	int charge = 0;
 	bool aromatic = false;
-	bool leavingAtom = false;
-	bool stereoConfig = false;
+	bool leaving_atom = false;
+	bool stereo_config = false;
 	float x, y, z;
 };
 
 /// --------------------------------------------------------------------
 /// \brief struct containing information about the bonds
 
-struct CompoundBond
+struct compound_bond
 {
-	std::string atomID[2];
-	BondType type;
-	bool aromatic = false, stereoConfig = false;
+	std::string atom_id[2];
+	bond_type type;
+	bool aromatic = false, stereo_config = false;
 };
 
 /// --------------------------------------------------------------------
@@ -95,62 +95,56 @@ struct CompoundBond
 /// compound definitions by calling the addExtraComponents function and
 /// pass it a valid CCD formatted file.
 
-class Compound
+class compound
 {
   public:
-
 	// accessors
 
-	std::string id() const { return mID; }
-	std::string name() const { return mName; }
-	std::string type() const { return mType; }
-	std::string formula() const { return mFormula; }
-	float formulaWeight() const { return mFormulaWeight; }
-	int formalCharge() const { return mFormalCharge; }
+	std::string id() const { return m_id; }
+	std::string name() const { return m_name; }
+	std::string type() const { return m_type; }
+	std::string group() const { return m_group; }
+	std::string formula() const { return m_formula; }
+	float formula_weight() const { return m_formula_weight; }
+	int formal_charge() const { return m_formal_charge; }
 
-	const std::vector<CompoundAtom> &atoms() const { return mAtoms; }
-	const std::vector<CompoundBond> &bonds() const { return mBonds; }
+	const std::vector<compound_atom> &atoms() const { return m_atoms; }
+	const std::vector<compound_bond> &bonds() const { return m_bonds; }
 
-	CompoundAtom getAtomByID(const std::string &atomID) const;
+	compound_atom get_atom_by_atom_id(const std::string &atom_id) const;
 
-	bool atomsBonded(const std::string &atomId_1, const std::string &atomId_2) const;
-	// float atomBondValue(const std::string &atomId_1, const std::string &atomId_2) const;
-	// float bondAngle(const std::string &atomId_1, const std::string &atomId_2, const std::string &atomId_3) const;
-	// float chiralVolume(const std::string &centreID) const;
+	bool atoms_bonded(const std::string &atomId_1, const std::string &atomId_2) const;
 
-	bool isWater() const
+	bool is_water() const
 	{
-		return mID == "HOH" or mID == "H2O" or mID == "WAT";
+		return m_id == "HOH" or m_id == "H2O" or m_id == "WAT";
 	}
 
   private:
+	friend class compound_factory_impl;
+	friend class CCD_compound_factory_impl;
+	friend class CCP4_compound_factory_impl;
 
-	friend class CompoundFactoryImpl;
-	friend class CCDCompoundFactoryImpl;
-	friend class CCP4CompoundFactoryImpl;
+	compound(cif::datablock &db);
+	compound(cif::datablock &db, const std::string &id, const std::string &name, const std::string &type, const std::string &group);
 
-	Compound(cif::Datablock &db);
-	Compound(cif::Datablock &db, const std::string &id, const std::string &name, const std::string &type);
-
-	std::string mID;
-	std::string mName;
-	std::string mType;
-	std::string mFormula;
-	float mFormulaWeight = 0;
-	int mFormalCharge = 0;
-	std::vector<CompoundAtom> mAtoms;
-	std::vector<CompoundBond> mBonds;
+	std::string m_id;
+	std::string m_name;
+	std::string m_type;
+	std::string m_group;
+	std::string m_formula;
+	float m_formula_weight = 0;
+	int m_formal_charge = 0;
+	std::vector<compound_atom> m_atoms;
+	std::vector<compound_bond> m_bonds;
 };
 
 // --------------------------------------------------------------------
-// Factory class for Compound and Link objects
+// Factory class for compound and Link objects
 
-CIFPP_EXPORT extern const std::map<std::string, char> kAAMap, kBaseMap;
-
-class CompoundFactory
+class compound_factory
 {
   public:
-
 	/// \brief Initialise a singleton instance.
 	///
 	/// If you have a multithreaded application and want to have different
@@ -159,37 +153,39 @@ class CompoundFactory
 	/// flag to true.
 
 	static void init(bool useThreadLocalInstanceOnly);
-	static CompoundFactory &instance();
+	static compound_factory &instance();
 	static void clear();
 
-	void setDefaultDictionary(const std::filesystem::path &inDictFile);
-	void pushDictionary(const std::filesystem::path &inDictFile);
-	void popDictionary();
+	void set_default_dictionary(const std::filesystem::path &inDictFile);
+	void push_dictionary(const std::filesystem::path &inDictFile);
+	void pop_dictionary();
 
-	bool isKnownPeptide(const std::string &res_name) const;
-	bool isKnownBase(const std::string &res_name) const;
+	bool is_known_peptide(const std::string &res_name) const;
+	bool is_known_base(const std::string &res_name) const;
 
-	/// \brief Create the Compound object for \a id
+	/// \brief Create the compound object for \a id
 	///
-	/// This will create the Compound instance for \a id if it doesn't exist already.
+	/// This will create the compound instance for \a id if it doesn't exist already.
 	/// The result is owned by this factory and should not be deleted by the user.
-	/// \param id	The Compound ID, a three letter code usually
+	/// \param id	The compound ID, a three letter code usually
 	/// \result		The compound, or nullptr if it could not be created (missing info)
-	const Compound *create(std::string id);
+	const compound *create(std::string id);
 
-	~CompoundFactory();
+	~compound_factory();
+
+	static const std::map<std::string, char> kAAMap, kBaseMap;
 
   private:
-	CompoundFactory();
+	compound_factory();
 
-	CompoundFactory(const CompoundFactory &) = delete;
-	CompoundFactory &operator=(const CompoundFactory &) = delete;
+	compound_factory(const compound_factory &) = delete;
+	compound_factory &operator=(const compound_factory &) = delete;
 
-	static std::unique_ptr<CompoundFactory> sInstance;
-	static thread_local std::unique_ptr<CompoundFactory> tlInstance;
-	static bool sUseThreadLocalInstance;
+	static std::unique_ptr<compound_factory> s_instance;
+	static thread_local std::unique_ptr<compound_factory> tl_instance;
+	static bool s_use_thread_local_instance;
 
-	std::shared_ptr<CompoundFactoryImpl> mImpl;
+	std::shared_ptr<compound_factory_impl> m_impl;
 };
 
-} // namespace mmcif
+} // namespace pdbx

include/cif++/condition.hpp

@@ -0,0 +1,744 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cassert>
+#include <functional>
+#include <iostream>
+#include <regex>
+#include <utility>
+
+#include <cif++/row.hpp>
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+// let's make life easier
+
+iset get_category_fields(const category &cat);
+uint16_t get_column_ix(const category &cat, std::string_view col);
+bool is_column_type_uchar(const category &cat, std::string_view col);
+
+// --------------------------------------------------------------------
+// some more templates to be able to do querying
+
+namespace detail
+{
+	struct condition_impl
+	{
+		virtual ~condition_impl() {}
+
+		virtual condition_impl *prepare(const category &) { return this; }
+		virtual bool test(row_handle) const = 0;
+		virtual void str(std::ostream &) const = 0;
+		virtual std::optional<row_handle> single() const { return {}; };
+	};
+
+	struct all_condition_impl : public condition_impl
+	{
+		bool test(row_handle) const override { return true; }
+		void str(std::ostream &os) const override { os << "*"; }
+	};
+
+	struct or_condition_impl;
+	struct and_condition_impl;
+	struct not_condition_impl;
+} // namespace detail
+
+class condition
+{
+  public:
+	using condition_impl = detail::condition_impl;
+
+	condition()
+		: m_impl(nullptr)
+	{
+	}
+
+	explicit condition(condition_impl *impl)
+		: m_impl(impl)
+	{
+	}
+
+	condition(const condition &) = delete;
+
+	condition(condition &&rhs) noexcept
+		: m_impl(nullptr)
+	{
+		std::swap(m_impl, rhs.m_impl);
+	}
+
+	condition &operator=(const condition &) = delete;
+
+	condition &operator=(condition &&rhs) noexcept
+	{
+		std::swap(m_impl, rhs.m_impl);
+		return *this;
+	}
+
+	~condition()
+	{
+		delete m_impl;
+		m_impl = nullptr;
+	}
+
+	void prepare(const category &c);
+
+	bool operator()(row_handle r) const
+	{
+		assert(this->m_impl != nullptr);
+		assert(this->m_prepared);
+		return m_impl ? m_impl->test(r) : false;
+	}
+
+	explicit operator bool() { return not empty(); }
+	bool empty() const { return m_impl == nullptr; }
+
+	std::optional<row_handle> single() const
+	{
+		return m_impl ? m_impl->single() : std::optional<row_handle>();
+	}
+
+	friend condition operator||(condition &&a, condition &&b);
+	friend condition operator&&(condition &&a, condition &&b);
+
+	friend struct detail::or_condition_impl;
+	friend struct detail::and_condition_impl;
+	friend struct detail::not_condition_impl;
+
+	void swap(condition &rhs)
+	{
+		std::swap(m_impl, rhs.m_impl);
+		std::swap(m_prepared, rhs.m_prepared);
+	}
+
+	friend std::ostream &operator<<(std::ostream &os, const condition &cond)
+	{
+		if (cond.m_impl)
+			cond.m_impl->str(os);
+		return os;
+	}
+
+  private:
+
+	void optimise(condition_impl *&impl);
+
+	condition_impl *m_impl;
+	bool m_prepared = false;
+};
+
+namespace detail
+{
+	struct key_is_empty_condition_impl : public condition_impl
+	{
+		key_is_empty_condition_impl(const std::string &item_tag)
+			: m_item_tag(item_tag)
+		{
+		}
+
+		condition_impl *prepare(const category &c) override
+		{
+			m_item_ix = get_column_ix(c, m_item_tag);
+			return this;
+		}
+
+		bool test(row_handle r) const override
+		{
+			return r[m_item_ix].empty();
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << m_item_tag << " IS NULL";
+		}
+
+		std::string m_item_tag;
+		uint16_t m_item_ix = 0;
+	};
+
+	struct key_equals_condition_impl : public condition_impl
+	{
+		key_equals_condition_impl(item &&i)
+			: m_item_tag(i.name())
+			, m_value(i.value())
+		{
+		}
+
+		condition_impl *prepare(const category &c) override;
+
+		bool test(row_handle r) const override
+		{
+			return m_single_hit.has_value() ?
+				*m_single_hit == r :
+				r[m_item_ix].compare(m_value, m_icase) == 0;
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << m_item_tag << (m_icase ? "^ " : " ") << " == " << m_value;
+		}
+
+		virtual std::optional<row_handle> single() const override
+		{
+			return m_single_hit;
+		}
+
+		std::string m_item_tag;
+		uint16_t m_item_ix = 0;
+		bool m_icase = false;
+		std::string m_value;
+		std::optional<row_handle> m_single_hit;
+	};
+
+	struct key_equals_or_empty_condition_impl : public condition_impl
+	{
+		key_equals_or_empty_condition_impl(key_equals_condition_impl *equals)
+			: m_item_tag(equals->m_item_tag)
+			, m_value(equals->m_value)
+			, m_icase(equals->m_icase)
+			, m_single_hit(equals->m_single_hit)
+		{
+		}
+
+		condition_impl *prepare(const category &c) override
+		{
+			m_item_ix = get_column_ix(c, m_item_tag);
+			m_icase = is_column_type_uchar(c, m_item_tag);
+			return this;
+		}
+
+		bool test(row_handle r) const override
+		{
+			bool result = false;
+			if (m_single_hit.has_value())
+				result = *m_single_hit == r;
+			else
+				result = r[m_item_ix].empty() or r[m_item_ix].compare(m_value, m_icase) == 0;
+			return result;
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << m_item_tag << (m_icase ? "^ " : " ") << " == " << m_value << " OR " << m_item_tag << " IS NULL";
+		}
+
+		virtual std::optional<row_handle> single() const override
+		{
+			return m_single_hit;
+		}
+
+		std::string m_item_tag;
+		uint16_t m_item_ix = 0;
+		std::string m_value;
+		bool m_icase = false;
+		std::optional<row_handle> m_single_hit;
+	};	
+
+	struct key_compare_condition_impl : public condition_impl
+	{
+		template <typename COMP>
+		key_compare_condition_impl(const std::string &item_tag, COMP &&comp, const std::string &s)
+			: m_item_tag(item_tag)
+			, m_compare(std::move(comp))
+			, m_str(s)
+		{
+		}
+
+		condition_impl *prepare(const category &c) override
+		{
+			m_item_ix = get_column_ix(c, m_item_tag);
+			m_icase = is_column_type_uchar(c, m_item_tag);
+			return this;
+		}
+
+		bool test(row_handle r) const override
+		{
+			return m_compare(r, m_icase);
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << m_item_tag << (m_icase ? "^ " : " ") << m_str;
+		}
+
+		std::string m_item_tag;
+		uint16_t m_item_ix = 0;
+		bool m_icase = false;
+		std::function<bool(row_handle, bool)> m_compare;
+		std::string m_str;
+	};
+
+	struct key_matches_condition_impl : public condition_impl
+	{
+		key_matches_condition_impl(const std::string &item_tag, const std::regex &rx)
+			: m_item_tag(item_tag)
+			, m_item_ix(0)
+			, mRx(rx)
+		{
+		}
+
+		condition_impl *prepare(const category &c) override
+		{
+			m_item_ix = get_column_ix(c, m_item_tag);
+			return this;
+		}
+
+		bool test(row_handle r) const override
+		{
+			std::string_view txt = r[m_item_ix].text();
+			return std::regex_match(txt.begin(), txt.end(), mRx);
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << m_item_tag << " =~ expression";
+		}
+
+		std::string m_item_tag;
+		uint16_t m_item_ix;
+		std::regex mRx;
+	};
+
+	template <typename T>
+	struct any_is_condition_impl : public condition_impl
+	{
+		typedef T valueType;
+
+		any_is_condition_impl(const valueType &value)
+			: mValue(value)
+		{
+		}
+
+		bool test(row_handle r) const override
+		{
+			auto &c = r.get_category();
+
+			bool result = false;
+			for (auto &f : get_category_fields(c))
+			{
+				try
+				{
+					if (r[f].compare(mValue) == 0)
+					{
+						result = true;
+						break;
+					}
+				}
+				catch (...)
+				{
+				}
+			}
+
+			return result;
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << "<any> == " << mValue;
+		}
+
+		valueType mValue;
+	};
+
+	struct any_matches_condition_impl : public condition_impl
+	{
+		any_matches_condition_impl(const std::regex &rx)
+			: mRx(rx)
+		{
+		}
+
+		bool test(row_handle r) const override
+		{
+			auto &c = r.get_category();
+
+			bool result = false;
+			for (auto &f : get_category_fields(c))
+			{
+				try
+				{
+					std::string_view txt = r[f].text();
+					if (std::regex_match(txt.begin(), txt.end(), mRx))
+					{
+						result = true;
+						break;
+					}
+				}
+				catch (...)
+				{
+				}
+			}
+
+			return result;
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << "<any> =~ expression";
+		}
+
+		std::regex mRx;
+	};
+
+	// TODO: Optimize and_condition by having a list of sub items.
+	// That way you can also collapse multiple _is_ conditions in
+	// case they make up an indexed tuple.
+	struct and_condition_impl : public condition_impl
+	{
+		and_condition_impl(condition &&a, condition &&b)
+		{
+			mSub.emplace_back(std::exchange(a.m_impl, nullptr));
+			mSub.emplace_back(std::exchange(b.m_impl, nullptr));
+		}
+
+		~and_condition_impl()
+		{
+			for (auto sub : mSub)
+				delete sub;
+		}
+
+		condition_impl *prepare(const category &c) override;
+
+		bool test(row_handle r) const override
+		{
+			bool result = true;
+
+			for (auto sub : mSub)
+			{
+				if (sub->test(r))
+					continue;
+		
+				result = false;
+				break;
+			}
+
+			return result;
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << '(';
+
+			bool first = true;
+			for (auto sub : mSub)
+			{
+				if (first)
+					first = false;
+				else
+					os << " AND ";
+
+				sub->str(os);
+			}
+
+			os << ')';
+		}
+
+		virtual std::optional<row_handle> single() const override
+		{
+			std::optional<row_handle> result;
+
+			for (auto sub : mSub)
+			{
+				auto s = sub->single();
+
+				if (not result.has_value())
+				{
+					result = s;
+					continue;
+				}
+				
+				if (s == result)
+					continue;
+
+				result.reset();
+				break;
+			}
+
+			return result;
+		}
+
+		std::vector<condition_impl *> mSub;
+	};
+
+	struct or_condition_impl : public condition_impl
+	{
+		or_condition_impl(condition &&a, condition &&b)
+			: mA(nullptr)
+			, mB(nullptr)
+		{
+			std::swap(mA, a.m_impl);
+			std::swap(mB, b.m_impl);
+		}
+
+		~or_condition_impl()
+		{
+			delete mA;
+			delete mB;
+		}
+
+		condition_impl *prepare(const category &c) override;
+
+		bool test(row_handle r) const override
+		{
+			return mA->test(r) or mB->test(r);
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << '(';
+			mA->str(os);
+			os << ") OR (";
+			mB->str(os);
+			os << ')';
+		}
+
+		virtual std::optional<row_handle> single() const override
+		{
+			auto sa = mA->single();
+			auto sb = mB->single();
+			
+			if (sa.has_value() and sb.has_value() and sa != sb)
+				sa.reset();
+			else if (not sa.has_value())
+				sa = sb;
+
+			return sa;
+		}
+
+		condition_impl *mA;
+		condition_impl *mB;
+	};
+
+	struct not_condition_impl : public condition_impl
+	{
+		not_condition_impl(condition &&a)
+			: mA(nullptr)
+		{
+			std::swap(mA, a.m_impl);
+		}
+
+		~not_condition_impl()
+		{
+			delete mA;
+		}
+
+		condition_impl *prepare(const category &c) override
+		{
+			mA = mA->prepare(c);
+			return this;
+		}
+
+		bool test(row_handle r) const override
+		{
+			return not mA->test(r);
+		}
+
+		void str(std::ostream &os) const override
+		{
+			os << "NOT (";
+			mA->str(os);
+			os << ')';
+		}
+
+		condition_impl *mA;
+	};
+
+} // namespace detail
+
+inline condition operator&&(condition &&a, condition &&b)
+{
+	if (a.m_impl and b.m_impl)
+		return condition(new detail::and_condition_impl(std::move(a), std::move(b)));
+	if (a.m_impl)
+		return condition(std::move(a));
+	return condition(std::move(b));
+}
+
+inline condition operator||(condition &&a, condition &&b)
+{
+	if (a.m_impl and b.m_impl)
+		return condition(new detail::or_condition_impl(std::move(a), std::move(b)));
+	if (a.m_impl)
+		return condition(std::move(a));
+	return condition(std::move(b));
+}
+
+struct empty_type
+{
+};
+
+/// \brief A helper to make it possible to have conditions like ("id"_key == cif::null)
+
+inline constexpr empty_type null = empty_type();
+
+struct key
+{
+	explicit key(const std::string &itemTag)
+		: m_item_tag(itemTag)
+	{
+	}
+
+	explicit key(const char *itemTag)
+		: m_item_tag(itemTag)
+	{
+	}
+
+	key(const key &) = delete;
+	key &operator=(const key &) = delete;
+
+	std::string m_item_tag;
+};
+
+template <typename T>
+condition operator==(const key &key, const T &v)
+{
+	return condition(new detail::key_equals_condition_impl({ key.m_item_tag, v }));
+}
+
+inline condition operator==(const key &key, const char *value)
+{
+	if (value != nullptr and *value != 0)
+		return condition(new detail::key_equals_condition_impl({ key.m_item_tag, value }));
+	else
+		return condition(new detail::key_is_empty_condition_impl(key.m_item_tag));
+}
+
+// inline condition_t operator==(const key& key, const detail::ItemReference& v)
+// {
+// 	if (v.empty())
+// 		return condition_t(new detail::key_is_empty_condition_impl(key.m_item_tag));
+// 	else
+// 		return condition_t(new detail::key_compare_condition_impl(key.m_item_tag, [tag = key.m_item_tag, v](const category& c, const row& r, bool icase)
+// 			{ return r[tag].template compare<(v, icase) == 0; }));
+// }
+
+template <typename T>
+condition operator!=(const key &key, const T &v)
+{
+	return condition(new detail::not_condition_impl(operator==(key, v)));
+}
+
+inline condition operator!=(const key &key, const char *v)
+{
+	std::string value(v ? v : "");
+	return condition(new detail::not_condition_impl(operator==(key, value)));
+}
+
+template <typename T>
+condition operator>(const key &key, const T &v)
+{
+	std::ostringstream s;
+	s << " > " << v;
+
+	return condition(new detail::key_compare_condition_impl(
+		key.m_item_tag, [tag = key.m_item_tag, v](row_handle r, bool icase)
+		{ return r[tag].template compare<T>(v, icase) > 0; },
+		s.str()));
+}
+
+template <typename T>
+condition operator>=(const key &key, const T &v)
+{
+	std::ostringstream s;
+	s << " >= " << v;
+
+	return condition(new detail::key_compare_condition_impl(
+		key.m_item_tag, [tag = key.m_item_tag, v](row_handle r, bool icase)
+		{ return r[tag].template compare<T>(v, icase) >= 0; },
+		s.str()));
+}
+
+template <typename T>
+condition operator<(const key &key, const T &v)
+{
+	std::ostringstream s;
+	s << " < " << v;
+
+	return condition(new detail::key_compare_condition_impl(
+		key.m_item_tag, [tag = key.m_item_tag, v](row_handle r, bool icase)
+		{ return r[tag].template compare<T>(v, icase) < 0; },
+		s.str()));
+}
+
+template <typename T>
+condition operator<=(const key &key, const T &v)
+{
+	std::ostringstream s;
+	s << " <= " << v;
+
+	return condition(new detail::key_compare_condition_impl(
+		key.m_item_tag, [tag = key.m_item_tag, v](row_handle r, bool icase)
+		{ return r[tag].template compare<T>(v, icase) <= 0; },
+		s.str()));
+}
+
+inline condition operator==(const key &key, const std::regex &rx)
+{
+	return condition(new detail::key_matches_condition_impl(key.m_item_tag, rx));
+}
+
+inline condition operator==(const key &key, const empty_type &)
+{
+	return condition(new detail::key_is_empty_condition_impl(key.m_item_tag));
+}
+
+inline condition operator !(condition &&rhs)
+{
+	return condition(new detail::not_condition_impl(std::move(rhs)));
+}
+
+struct any_type
+{
+};
+
+inline constexpr any_type any = any_type{};
+
+template <typename T>
+condition operator==(const any_type &, const T &v)
+{
+	return condition(new detail::any_is_condition_impl<T>(v));
+}
+
+inline condition operator==(const any_type &, const std::regex &rx)
+{
+	return condition(new detail::any_matches_condition_impl(rx));
+}
+
+inline condition all()
+{
+	return condition(new detail::all_condition_impl());
+}
+
+namespace literals
+{
+	inline key operator""_key(const char *text, size_t length)
+	{
+		return key(std::string(text, length));
+	}
+} // namespace literals
+
+} // namespace cif

include/cif++/datablock.hpp

@@ -0,0 +1,100 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cif++/forward_decl.hpp>
+
+#include <cif++/category.hpp>
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+class datablock : public std::list<category>
+{
+  public:
+	datablock() = default;
+
+	datablock(std::string_view name)
+		: m_name(name)
+	{
+	}
+
+	datablock(const datablock &) = default;
+
+	datablock(datablock &&) = default;
+
+	datablock &operator=(const datablock &) = default;
+	datablock &operator=(datablock &&) = default;
+
+	// --------------------------------------------------------------------
+
+	const std::string &name() const { return m_name; }
+
+	void set_name(std::string_view name)
+	{
+		m_name = name;
+	}
+
+	void set_validator(const validator *v);
+
+	const validator *get_validator() const;
+
+	bool is_valid() const;
+	bool validate_links() const;
+
+	// --------------------------------------------------------------------
+
+	category &operator[](std::string_view name);
+	const category &operator[](std::string_view name) const;
+
+	category *get(std::string_view name);
+	const category *get(std::string_view name) const;
+
+	std::tuple<iterator, bool> emplace(std::string_view name);
+
+	std::vector<std::string> get_tag_order() const;
+	void write(std::ostream &os) const;
+	void write(std::ostream &os, const std::vector<std::string> &tag_order);
+
+	friend std::ostream &operator<<(std::ostream &os, const datablock &db)
+	{
+		db.write(os);
+		return os;
+	}
+
+	// --------------------------------------------------------------------
+	
+	bool operator==(const datablock &rhs) const;
+
+  private:
+	std::string m_name;
+	const validator *m_validator = nullptr;
+};
+
+} // namespace cif

include/cif++/dictionary_parser.hpp

@@ -0,0 +1,36 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cif++/validate.hpp>
+
+namespace cif
+{
+
+validator parse_dictionary(std::string_view name, std::istream &is);
+
+} // namespace cif

include/cif++/file.hpp

@@ -0,0 +1,122 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <list>
+
+#include <cif++/datablock.hpp>
+#include <cif++/parser.hpp>
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+class file : public std::list<datablock>
+{
+  public:
+	file() = default;
+
+	explicit file(const std::filesystem::path &p)
+	{
+		load(p);
+	}
+
+	explicit file(std::istream &is)
+	{
+		load(is);
+	}
+
+	explicit file(const char *data, size_t length)
+	{
+		struct membuf : public std::streambuf
+		{
+			membuf(char *text, size_t length)
+			{
+				this->setg(text, text, text + length);
+			}
+		} buffer(const_cast<char *>(data), length);
+
+		std::istream is(&buffer);
+		load(is);
+	}
+
+	file(const file &) = default;
+	file(file &&) = default;
+	file &operator=(const file &) = default;
+	file &operator=(file &&) = default;
+
+	void set_validator(const validator *v);
+
+	const validator *get_validator() const
+	{
+		return m_validator;
+	}
+
+	bool is_valid() const;
+	bool is_valid();
+	bool validate_links() const;
+
+	void load_dictionary();
+	void load_dictionary(std::string_view name);
+
+	bool contains(std::string_view name) const;
+
+	datablock &front()
+	{
+		assert(not empty());
+		return std::list<datablock>::front();
+	}
+
+	const datablock &front() const
+	{
+		assert(not empty());
+		return std::list<datablock>::front();
+	}
+
+	datablock &operator[](std::string_view name);
+	const datablock &operator[](std::string_view name) const;
+
+	std::tuple<iterator, bool> emplace(std::string_view name);
+
+	void load(const std::filesystem::path &p);
+	void load(std::istream &is);
+
+	void save(const std::filesystem::path &p) const;
+	void save(std::ostream &os) const;
+
+	friend std::ostream &operator<<(std::ostream &os, const file &f)
+	{
+		f.save(os);
+		return os;
+	}
+
+  private:
+	const validator *m_validator = nullptr;
+};
+
+} // namespace cif

include/cif++/format.hpp

@@ -0,0 +1,206 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <string>
+
+/// \file format.hpp
+/// File containing a basic reimplementation of boost::format
+/// but then a bit more simplistic. Still this allowed me to move my code
+/// from using boost::format to something without external dependency easily.
+
+namespace cif
+{
+
+namespace detail
+{
+	template <typename T>
+	struct to_varg
+	{
+		using type = T;
+
+		to_varg(const T &v)
+			: m_value(v)
+		{
+		}
+
+		type operator*() { return m_value; }
+
+		T m_value;
+	};
+
+	template <>
+	struct to_varg<const char *>
+	{
+		using type = const char *;
+
+		to_varg(const char *v)
+			: m_value(v)
+		{
+		}
+
+		type operator*() { return m_value.c_str(); }
+
+		std::string m_value;
+	};
+
+	template <>
+	struct to_varg<std::string>
+	{
+		using type = const char *;
+
+		to_varg(const std::string &v)
+			: m_value(v)
+		{
+		}
+
+		type operator*() { return m_value.c_str(); }
+
+		std::string m_value;
+	};
+
+} // namespace
+
+template <typename... Args>
+class format_plus_arg
+{
+  public:
+	using args_vector_type = std::tuple<detail::to_varg<Args>...>;
+	using vargs_vector_type = std::tuple<typename detail::to_varg<Args>::type...>;
+
+	format_plus_arg(const format_plus_arg &) = delete;
+	format_plus_arg &operator=(const format_plus_arg &) = delete;
+
+
+	format_plus_arg(std::string_view fmt, Args... args)
+		: m_fmt(fmt)
+		, m_args(std::forward<Args>(args)...)
+	{
+		auto ix = std::make_index_sequence<sizeof...(Args)>();
+		copy_vargs(ix);
+	}
+
+	std::string str()
+	{
+		char buffer[1024];
+		std::string::size_type r = std::apply(snprintf, std::tuple_cat(std::make_tuple(buffer, sizeof(buffer), m_fmt.c_str()), m_vargs));
+		return { buffer, r };
+	}
+
+	friend std::ostream &operator<<(std::ostream &os, const format_plus_arg &f)
+	{
+		char buffer[1024];
+		std::string::size_type r = std::apply(snprintf, std::tuple_cat(std::make_tuple(buffer, sizeof(buffer), f.m_fmt.c_str()), f.m_vargs));
+		os.write(buffer, r);
+		return os;
+	}
+
+  private:
+
+	template <size_t... I>
+	void copy_vargs(std::index_sequence<I...>)
+	{
+		((std::get<I>(m_vargs) = *std::get<I>(m_args)), ...);
+	}
+
+	std::string m_fmt;
+	args_vector_type m_args;
+	vargs_vector_type m_vargs;
+};
+
+template <typename... Args>
+constexpr auto format(std::string_view fmt, Args... args)
+{
+	return format_plus_arg(fmt, std::forward<Args>(args)...);
+}
+
+// --------------------------------------------------------------------
+/// A streambuf that fills out lines with spaces up until a specified width
+
+class fill_out_streambuf : public std::streambuf
+{
+  public:
+	using base_type = std::streambuf;
+	using int_type = base_type::int_type;
+	using char_type = base_type::char_type;
+	using traits_type = base_type::traits_type;
+
+	fill_out_streambuf(std::ostream &os, int width = 80)
+		: m_os(os)
+		, m_upstream(os.rdbuf())
+		, m_width(width)
+	{
+	}
+
+	~fill_out_streambuf()
+	{
+		m_os.rdbuf(m_upstream);
+	}
+
+	virtual int_type
+	overflow(int_type ic = traits_type::eof())
+	{
+		char ch = traits_type::to_char_type(ic);
+
+		int_type result = ic;
+
+		if (ch == '\n')
+		{
+			for (int i = m_column_count; result != traits_type::eof() and i < m_width; ++i)
+				result = m_upstream->sputc(' ');
+		}
+
+		if (result != traits_type::eof())
+			result = m_upstream->sputc(ch);
+
+		if (result != traits_type::eof())
+		{
+			if (ch == '\n')
+			{
+				m_column_count = 0;
+				++m_line_count;
+			}
+			else
+				++m_column_count;
+		}
+
+		return result;
+	}
+
+	std::streambuf *get_upstream() const { return m_upstream; }
+
+	int get_line_count() const { return m_line_count; }
+
+  private:
+	std::ostream &m_os;
+	std::streambuf *m_upstream;
+	int m_width;
+	int m_line_count = 0;
+	int m_column_count = 0;
+};
+
+} // namespace pdbx

include/cif++/forward_decl.hpp

@@ -0,0 +1,46 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <string>
+#include <vector>
+
+namespace cif
+{
+
+class category;
+class datablock;
+class file;
+class parser;
+
+class row;
+class row_handle;
+
+class item;
+struct item_handle;
+
+} // namespace cif

include/cif++/item.hpp

@@ -0,0 +1,567 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <charconv>
+#include <cstring>
+#include <iomanip>
+#include <iostream>
+#include <limits>
+#include <memory>
+#include <optional>
+#include <utility>
+
+#include <cif++/forward_decl.hpp>
+#include <cif++/text.hpp>
+
+/// \file item.hpp
+/// This file contains the declaration of item but also the item_value and item_handle
+/// These handle the storage of and access to the data for a single data field. 
+
+namespace cif
+{
+
+extern int VERBOSE;
+
+// --------------------------------------------------------------------
+/// \brief item is a transient class that is used to pass data into rows
+///        but it also takes care of formatting data. 
+class item
+{
+  public:
+	/// \brief Default constructor, empty item
+	item() = default;
+
+	/// \brief constructor for an item with name \a name and as
+	/// content a single character string with content \a value
+	item(std::string_view name, char value)
+		: m_name(name)
+		, m_value({ value })
+	{
+	}
+
+	/// \brief constructor for an item with name \a name and as
+	/// content a the formatted floating point value \a value with
+	/// precision \a precision
+	template <typename T, std::enable_if_t<std::is_floating_point_v<T>, int> = 0>
+	item(std::string_view name, const T &value, int precision)
+		: m_name(name)
+	{
+		using namespace std;
+		using namespace cif;
+
+		char buffer[32];
+
+		auto r = to_chars(buffer, buffer + sizeof(buffer) - 1, value, chars_format::fixed, precision);
+		if (r.ec != std::errc())
+			throw std::runtime_error("Could not format number");
+
+		assert(r.ptr >= buffer and r.ptr < buffer + sizeof(buffer));
+		*r.ptr = 0;
+		m_value.assign(buffer, r.ptr - buffer);
+	}
+
+	/// \brief constructor for an item with name \a name and as
+	/// content a formatted floating point value \a value with
+	/// so-called general formatting
+	template <typename T, std::enable_if_t<std::is_floating_point_v<T>, int> = 0>
+	item(const std::string_view name, const T &value)
+		: m_name(name)
+	{
+		using namespace std;
+		using namespace cif;
+
+		char buffer[32];
+
+		auto r = to_chars(buffer, buffer + sizeof(buffer) - 1, value, chars_format::general);
+		if (r.ec != std::errc())
+			throw std::runtime_error("Could not format number");
+
+		assert(r.ptr >= buffer and r.ptr < buffer + sizeof(buffer));
+		*r.ptr = 0;
+		m_value.assign(buffer, r.ptr - buffer);
+	}
+
+	/// \brief constructor for an item with name \a name and as
+	/// content a the formatted integral value \a value
+	template <typename T, std::enable_if_t<std::is_integral_v<T> and not std::is_same_v<T,bool>, int> = 0>
+	item(const std::string_view name, const T &value)
+		: m_name(name)
+	{
+		char buffer[32];
+
+		auto r = std::to_chars(buffer, buffer + sizeof(buffer) - 1, value);
+		if (r.ec != std::errc())
+			throw std::runtime_error("Could not format number");
+
+		assert(r.ptr >= buffer and r.ptr < buffer + sizeof(buffer));
+		*r.ptr = 0;
+		m_value.assign(buffer, r.ptr - buffer);
+	}
+
+	/// \brief constructor for an item with name \a name and as
+	/// content a the formatted boolean value \a value
+	template <typename T, std::enable_if_t<std::is_same_v<T,bool>, int> = 0>
+	item(const std::string_view name, const T &value)
+		: m_name(name)
+	{
+		m_value.assign(value ? "y" : "n");
+	}
+
+	/// \brief constructor for an item with name \a name and as
+	/// content value \a value
+	item(const std::string_view name, const std::string_view value)
+		: m_name(name)
+		, m_value(value)
+	{
+	}
+
+	item(const item &rhs) = default;
+
+	item(item &&rhs) noexcept = default;
+
+	item &operator=(const item &rhs) = default;
+
+	item &operator=(item &&rhs) noexcept = default;
+
+	std::string_view name() const { return m_name; }
+	std::string_view value() const { return m_value; }
+
+	/// \brief replace the content of the stored value with \a v
+	void value(std::string_view v) { m_value = v; }
+
+	/// \brief empty means either null or unknown
+	bool empty() const { return m_value.empty(); }
+
+	/// \brief returns true if the field contains '.'
+	bool is_null() const { return m_value == "."; }
+
+	/// \brief returns true if the field contains '?'
+	bool is_unknown() const { return m_value == "?"; }
+
+	/// \brief the length of the value string
+	size_t length() const { return m_value.length(); }
+
+	/// \brief support for structured binding
+	template<size_t N>
+	decltype(auto) get() const
+	{
+		     if constexpr (N == 0) return name();
+		else if constexpr (N == 1) return value();
+	}
+
+  private:
+	std::string_view m_name;
+	std::string m_value;
+};
+
+// --------------------------------------------------------------------
+/// \brief the internal storage for items in a category
+///
+/// Internal storage, strictly forward linked list with minimal space
+/// requirements. Strings of size 7 or shorter are stored internally.
+/// Typically, more than 99% of the strings in an mmCIF file are less
+/// than 8 bytes in length.
+
+struct item_value
+{
+	item_value() = default;
+
+	/// \brief constructor
+	item_value(std::string_view text)
+		: m_length(text.length())
+	{
+		if (m_length >= kBufferSize)
+		{
+			m_data = new char[m_length + 1];
+			std::copy(text.begin(), text.end(), m_data);
+			m_data[m_length] = 0;
+		}
+		else
+		{
+			std::copy(text.begin(), text.end(), m_local_data);
+			m_local_data[m_length] = 0;
+		}
+	}
+
+	item_value(item_value &&rhs)
+		: m_length(std::exchange(rhs.m_length, 0))
+		, m_data(std::exchange(rhs.m_data, nullptr))
+	{
+	}
+
+	item_value &operator=(item_value &&rhs)
+	{
+		if (this != &rhs)
+		{
+			m_length = std::exchange(rhs.m_length, m_length);
+			m_data = std::exchange(rhs.m_data, m_data);
+		}
+		return *this;
+	}
+
+	~item_value()
+	{
+		if (m_length >= kBufferSize)
+			delete[] m_data;
+		m_data = nullptr;
+		m_length = 0;
+	}
+
+	item_value(const item_value &) = delete;
+	item_value &operator=(const item_value &) = delete;
+
+	explicit operator bool() const
+	{
+		return m_length != 0;
+	}
+
+	size_t m_length = 0;
+	union
+	{
+		char m_local_data[8];
+		char *m_data;
+	};
+
+	static constexpr size_t kBufferSize = sizeof(m_local_data);
+
+	// By using std::string_view instead of c_str we obain a
+	// nice performance gain since we avoid many calls to strlen.
+	constexpr inline std::string_view text() const
+	{
+		return { m_length >= kBufferSize ? m_data : m_local_data, m_length };
+	}
+};
+
+// static_assert(sizeof(item_value) == 24, "sizeof(item_value) should be 24 bytes");
+static_assert(sizeof(item_value) == 16, "sizeof(item_value) should be 16 bytes");
+
+// --------------------------------------------------------------------
+// Transient object to access stored data
+
+/// \brief This is item_handle, it is used to access the data stored in item_value.
+
+struct item_handle
+{
+  public:
+	// conversion helper class
+	template <typename T, typename = void>
+	struct item_value_as;
+
+	template <typename T>
+	item_handle &operator=(const T &value)
+	{
+		item v{ "", value };
+		assign_value(v);
+		return *this;
+	}
+
+	template <typename... Ts>
+	void os(const Ts &...v)
+	{
+		std::ostringstream ss;
+		((ss << v), ...);
+		this->operator=(ss.str());
+	}
+
+	void swap(item_handle &b);
+
+	template <typename T = std::string>
+	auto as() const -> T
+	{
+		using value_type = std::remove_cv_t<std::remove_reference_t<T>>;
+		return item_value_as<value_type>::convert(*this);
+	}
+
+	template <typename T>
+	auto value_or(const T &dv) const
+	{
+		return empty() ? dv : this->as<T>();
+	}
+
+	template <typename T>
+	int compare(const T &value, bool icase = true) const
+	{
+		return item_value_as<T>::compare(*this, value, icase);
+	}
+
+	template <typename T>
+	bool operator==(const T &value) const
+	{
+		// TODO: icase or not icase?
+		return item_value_as<T>::compare(*this, value, true) == 0;
+	}
+
+	// We may not have C++20 yet...
+	template <typename T>
+	bool operator!=(const T &value) const
+	{
+		return not operator==(value);
+	}
+
+	// empty means either null or unknown
+	bool empty() const
+	{
+		auto txt = text();
+		return txt.empty() or (txt.length() == 1 and (txt.front() == '.' or txt.front() == '?'));
+	}
+
+	explicit operator bool() const { return not empty(); }
+
+	// is_null means the field contains '.'
+	bool is_null() const
+	{
+		auto txt = text();
+		return txt.length() == 1 and txt.front() == '.';
+	}
+
+	// is_unknown means the field contains '?'
+	bool is_unknown() const
+	{
+		auto txt = text();
+		return txt.length() == 1 and txt.front() == '?';
+	}
+
+	std::string_view text() const;
+
+	item_handle(uint16_t column, row_handle &row)
+		: m_column(column)
+		, m_row_handle(row)
+	{
+	}
+
+	static const item_handle s_null_item;
+
+	friend void swap(item_handle a, item_handle b)
+	{
+		a.swap(b);
+	}
+
+  private:
+	item_handle();
+
+	uint16_t m_column;
+	row_handle &m_row_handle;
+
+	void assign_value(const item &value);
+};
+
+// So sad that older gcc implementations of from_chars did not support floats yet...
+
+template <typename T>
+struct item_handle::item_value_as<T, std::enable_if_t<std::is_arithmetic_v<T> and not std::is_same_v<T, bool>>>
+{
+	using value_type = std::remove_reference_t<std::remove_cv_t<T>>;
+
+	static value_type convert(const item_handle &ref)
+	{
+		value_type result = {};
+
+		if (not ref.empty())
+		{
+			auto txt = ref.text();
+
+			std::from_chars_result r = selected_charconv<value_type>::from_chars(txt.data(), txt.data() + txt.size(), result);
+
+			if (r.ec != std::errc())
+			{
+				result = {};
+				if (cif::VERBOSE)
+				{
+					if (r.ec == std::errc::invalid_argument)
+						std::cerr << "Attempt to convert " << std::quoted(txt) << " into a number" << std::endl;
+					else if (r.ec == std::errc::result_out_of_range)
+						std::cerr << "Conversion of " << std::quoted(txt) << " into a type that is too small" << std::endl;
+				}
+			}
+		}
+
+		return result;
+	}
+
+	static int compare(const item_handle &ref, const T &value, bool icase)
+	{
+		int result = 0;
+
+		auto txt = ref.text();
+
+		if (txt.empty())
+			result = 1;
+		else
+		{
+			value_type v = {};
+
+			std::from_chars_result r = selected_charconv<value_type>::from_chars(txt.data(), txt.data() + txt.size(), v);
+
+			if (r.ec != std::errc())
+			{
+				if (cif::VERBOSE)
+				{
+					if (r.ec == std::errc::invalid_argument)
+						std::cerr << "Attempt to convert " << std::quoted(txt) << " into a number" << std::endl;
+					else if (r.ec == std::errc::result_out_of_range)
+						std::cerr << "Conversion of " << std::quoted(txt) << " into a type that is too small" << std::endl;
+				}
+				result = 1;
+			}
+			else if (v < value)
+				result = -1;
+			else if (v > value)
+				result = 1;
+		}
+
+		return result;
+	}
+};
+
+template <typename T>
+struct item_handle::item_value_as<std::optional<T>>
+{
+	static std::optional<T> convert(const item_handle &ref)
+	{
+		std::optional<T> result;
+		if (ref)
+			result = ref.as<T>();
+		return result;
+	}
+
+	static int compare(const item_handle &ref, std::optional<T> value, bool icase)
+	{
+		if (ref.empty() and not value)
+			return 0;
+
+		if (ref.empty())
+			return -1;
+		else if (not value)
+			return 1;
+		else
+			return ref.compare(*value, icase);
+	}
+};
+
+template <typename T>
+struct item_handle::item_value_as<T, std::enable_if_t<std::is_same_v<T, bool>>>
+{
+	static bool convert(const item_handle &ref)
+	{
+		bool result = false;
+		if (not ref.empty())
+			result = iequals(ref.text(), "y");
+		return result;
+	}
+
+	static int compare(const item_handle &ref, bool value, bool icase)
+	{
+		bool rv = convert(ref);
+		return value && rv ? 0
+		                   : (rv < value ? -1 : 1);
+	}
+};
+
+template <size_t N>
+struct item_handle::item_value_as<char[N]>
+{
+	static std::string convert(const item_handle &ref)
+	{
+		if (ref.empty())
+			return {};
+		return { ref.text().data(), ref.text().size() };
+	}
+
+	static int compare(const item_handle &ref, const char (&value)[N], bool icase)
+	{
+		return icase ? cif::icompare(ref.text(), value) : ref.text().compare(value);
+	}
+};
+
+template <typename T>
+struct item_handle::item_value_as<T, std::enable_if_t<std::is_same_v<T, const char *>>>
+{
+	static std::string convert(const item_handle &ref)
+	{
+		if (ref.empty())
+			return {};
+		return { ref.text().data(), ref.text().size() };
+	}
+
+	static int compare(const item_handle &ref, const char *value, bool icase)
+	{
+		return icase ? cif::icompare(ref.text(), value) : ref.text().compare(value);
+	}
+};
+
+template <typename T>
+struct item_handle::item_value_as<T, std::enable_if_t<std::is_same_v<T, std::string_view>>>
+{
+	static std::string convert(const item_handle &ref)
+	{
+		if (ref.empty())
+			return {};
+		return { ref.text().data(), ref.text().size() };
+	}
+
+	static int compare(const item_handle &ref, const std::string_view &value, bool icase)
+	{
+		return icase ? cif::icompare(ref.text(), value) : ref.text().compare(value);
+	}
+};
+
+template <typename T>
+struct item_handle::item_value_as<T, std::enable_if_t<std::is_same_v<T, std::string>>>
+{
+	static std::string convert(const item_handle &ref)
+	{
+		if (ref.empty())
+			return {};
+		return { ref.text().data(), ref.text().size() };
+	}
+
+	static int compare(const item_handle &ref, const std::string &value, bool icase)
+	{
+		return icase ? cif::icompare(ref.text(), value) : ref.text().compare(value);
+	}
+};
+
+} // namespace cif
+
+namespace std
+{
+
+template<> struct tuple_size<::cif::item>
+            : public std::integral_constant<std::size_t, 2> {};
+
+template<> struct tuple_element<0, ::cif::item>
+{
+	using type = decltype(std::declval<::cif::item>().name());
+};
+
+template<> struct tuple_element<1, ::cif::item>
+{
+	using type = decltype(std::declval<::cif::item>().value());
+};
+
+}

include/cif++/iterator.hpp

@@ -0,0 +1,676 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cif++/row.hpp>
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+template <typename Category, typename... Ts>
+class iterator_impl
+{
+  public:
+	template <typename, typename...>
+	friend class iterator_impl;
+
+	friend class category;
+
+	static constexpr size_t N = sizeof...(Ts);
+
+	using category_type = std::remove_cv_t<Category>;
+	using row_type = std::conditional_t<std::is_const_v<Category>, const row, row>;
+
+	using tuple_type = std::tuple<Ts...>;
+
+	using iterator_category = std::forward_iterator_tag;
+	using value_type = tuple_type;
+	using difference_type = std::ptrdiff_t;
+	using pointer = value_type *;
+	using reference = value_type &;
+
+	iterator_impl() = default;
+
+	iterator_impl(const iterator_impl &rhs) = default;
+
+	template <typename C2, typename... T2s>
+	iterator_impl(const iterator_impl<C2, T2s...> &rhs)
+		: m_category(rhs.m_category)
+		, m_current(rhs.m_current)
+		, m_value(rhs.m_value)
+		, m_column_ix(rhs.m_column_ix)
+	{
+	}
+
+	template <typename IRowType>
+	iterator_impl(iterator_impl<IRowType, Ts...> &rhs)
+		: m_category(rhs.m_category)
+		, m_current(const_cast<row_type *>(rhs.m_current))
+		, m_value(rhs.m_value)
+		, m_column_ix(rhs.m_column_ix)
+	{
+		m_value = get(std::make_index_sequence<N>());
+	}
+
+	template <typename IRowType>
+	iterator_impl(const iterator_impl<IRowType> &rhs, const std::array<uint16_t, N> &cix)
+		: m_category(rhs.m_category)
+		, m_current(rhs.m_current)
+		, m_column_ix(cix)
+	{
+		m_value = get(std::make_index_sequence<N>());
+	}
+
+	iterator_impl &operator=(const iterator_impl &i)
+	{
+		m_category = i.m_category;
+		m_current = i.m_current;
+		m_column_ix = i.m_column_ix;
+		m_value = i.m_value;
+		return *this;
+	}
+
+	virtual ~iterator_impl() = default;
+
+	reference operator*()
+	{
+		return m_value;
+	}
+
+	pointer operator->()
+	{
+		return &m_value;
+	}
+
+	operator const row_handle() const
+	{
+		return { *m_category, *m_current };
+	}
+
+	operator row_handle()
+	{
+		return { *m_category, *m_current };
+	}
+
+	iterator_impl &operator++()
+	{
+		if (m_current != nullptr)
+			m_current = m_current->m_next;
+
+		m_value = get(std::make_index_sequence<N>());
+
+		return *this;
+	}
+
+	iterator_impl operator++(int)
+	{
+		iterator_impl result(*this);
+		this->operator++();
+		return result;
+	}
+
+	bool operator==(const iterator_impl &rhs) const { return m_current == rhs.m_current; }
+	bool operator!=(const iterator_impl &rhs) const { return m_current != rhs.m_current; }
+
+	template <typename IRowType, typename... ITs>
+	bool operator==(const iterator_impl<IRowType, ITs...> &rhs) const
+	{
+		return m_current == rhs.m_current;
+	}
+
+	template <typename IRowType, typename... ITs>
+	bool operator!=(const iterator_impl<IRowType, ITs...> &rhs) const
+	{
+		return m_current != rhs.m_current;
+	}
+
+  private:
+	template <size_t... Is>
+	tuple_type get(std::index_sequence<Is...>) const
+	{
+		if (m_current != nullptr)
+		{
+			row_handle rh{*m_category, *m_current};
+			return tuple_type{rh[m_column_ix[Is]].template as<Ts>()...};
+		}
+
+		return {};
+	}
+
+	category_type *m_category = nullptr;
+	row_type *m_current = nullptr;
+	value_type m_value;
+	std::array<uint16_t, N> m_column_ix;
+};
+
+template<typename Category>
+class iterator_impl<Category>
+{
+  public:
+	template <typename, typename...>
+	friend class iterator_impl;
+
+	friend class category;
+	using category_type = std::remove_cv_t<Category>;
+	using row_type = std::conditional_t<std::is_const_v<Category>, const row, row>;
+
+	using iterator_category = std::forward_iterator_tag;
+	using value_type = row_handle;
+	using difference_type = std::ptrdiff_t;
+	using pointer = row_handle;
+	using reference = row_handle;
+
+	iterator_impl() = default;
+
+	iterator_impl(const iterator_impl &rhs) = default;
+
+	template <typename C2>
+	iterator_impl(const iterator_impl<C2> &rhs)
+		: m_category(rhs.m_category)
+		, m_current(const_cast<row_type*>(rhs.m_current))
+	{
+	}
+
+	iterator_impl(Category &cat, row *current)
+		: m_category(const_cast<category_type *>(&cat))
+		, m_current(current)
+	{
+	}
+
+	template <typename IRowType>
+	iterator_impl(const iterator_impl<IRowType> &rhs, const std::array<uint16_t, 0> &)
+		: m_category(rhs.m_category)
+		, m_current(rhs.m_current)
+	{
+	}
+
+	iterator_impl &operator=(const iterator_impl &i)
+	{
+		m_category = i.m_category;
+		m_current = i.m_current;
+		return *this;
+	}
+
+	virtual ~iterator_impl() = default;
+
+	reference operator*()
+	{
+		return {*m_category, *m_current};
+	}
+
+	pointer operator->()
+	{
+		return &m_current;
+	}
+
+	operator const row_handle() const
+	{
+		return { *m_category, *m_current };
+	}
+
+	operator row_handle()
+	{
+		return { *m_category, *m_current };
+	}
+
+	iterator_impl &operator++()
+	{
+		if (m_current != nullptr)
+			m_current = m_current->m_next;
+
+		return *this;
+	}
+
+	iterator_impl operator++(int)
+	{
+		iterator_impl result(*this);
+		this->operator++();
+		return result;
+	}
+
+	bool operator==(const iterator_impl &rhs) const { return m_current == rhs.m_current; }
+	bool operator!=(const iterator_impl &rhs) const { return m_current != rhs.m_current; }
+
+	template <typename IRowType, typename... ITs>
+	bool operator==(const iterator_impl<IRowType, ITs...> &rhs) const
+	{
+		return m_current == rhs.m_current;
+	}
+
+	template <typename IRowType, typename... ITs>
+	bool operator!=(const iterator_impl<IRowType, ITs...> &rhs) const
+	{
+		return m_current != rhs.m_current;
+	}
+
+  private:
+	category_type *m_category = nullptr;
+	row_type *m_current = nullptr;
+};
+
+
+template<typename Category, typename T>
+class iterator_impl<Category, T>
+{
+  public:
+	template <typename, typename...>
+	friend class iterator_impl;
+
+	friend class category;
+
+	using category_type = std::remove_cv_t<Category>;
+	using row_type = std::conditional_t<std::is_const_v<Category>, const row, row>;
+
+	using iterator_category = std::forward_iterator_tag;
+	using value_type = T;
+	using difference_type = std::ptrdiff_t;
+	using pointer = value_type *;
+	using reference = value_type &;
+
+	iterator_impl() = default;
+
+	iterator_impl(const iterator_impl &rhs) = default;
+
+	template <typename C2, typename T2>
+	iterator_impl(const iterator_impl<C2, T2> &rhs)
+		: m_category(rhs.m_category)
+		, m_current(rhs.m_current)
+		, m_value(rhs.m_value)
+		, m_column_ix(rhs.m_column_ix)
+	{
+	}
+
+	template <typename IRowType>
+	iterator_impl(iterator_impl<IRowType, T> &rhs)
+		: m_category(rhs.m_category)
+		, m_current(const_cast<row_type *>(rhs.m_current))
+		, m_value(rhs.m_value)
+		, m_column_ix(rhs.m_column_ix)
+	{
+		m_value = get(m_current);
+	}
+
+	template <typename IRowType>
+	iterator_impl(const iterator_impl<IRowType> &rhs, const std::array<uint16_t, 1> &cix)
+		: m_category(rhs.m_category)
+		, m_current(rhs.m_current)
+		, m_column_ix(cix[0])
+	{
+		m_value = get();
+	}
+
+	iterator_impl &operator=(const iterator_impl &i)
+	{
+		m_category = i.m_category;
+		m_current = i.m_current;
+		m_column_ix = i.m_column_ix;
+		m_value = i.m_value;
+		return *this;
+	}
+
+	virtual ~iterator_impl() = default;
+
+	reference operator*()
+	{
+		return m_value;
+	}
+
+	pointer operator->()
+	{
+		return &m_value;
+	}
+
+	operator const row_handle() const
+	{
+		return { *m_category, *m_current };
+	}
+
+	operator row_handle()
+	{
+		return { *m_category, *m_current };
+	}
+
+	iterator_impl &operator++()
+	{
+		if (m_current != nullptr)
+			m_current = m_current->m_next;
+
+		m_value = get();
+
+		return *this;
+	}
+
+	iterator_impl operator++(int)
+	{
+		iterator_impl result(*this);
+		this->operator++();
+		return result;
+	}
+
+	bool operator==(const iterator_impl &rhs) const { return m_current == rhs.m_current; }
+	bool operator!=(const iterator_impl &rhs) const { return m_current != rhs.m_current; }
+
+	template <typename IRowType, typename... ITs>
+	bool operator==(const iterator_impl<IRowType, ITs...> &rhs) const
+	{
+		return m_current == rhs.m_current;
+	}
+
+	template <typename IRowType, typename... ITs>
+	bool operator!=(const iterator_impl<IRowType, ITs...> &rhs) const
+	{
+		return m_current != rhs.m_current;
+	}
+
+  private:
+	value_type get() const
+	{
+		if (m_current != nullptr)
+		{
+			row_handle rh{*m_category, *m_current};
+			return rh[m_column_ix].template as<T>();
+		}
+
+		return {};
+	}
+
+	category_type *m_category = nullptr;
+	row_type *m_current = nullptr;
+	value_type m_value;
+	uint16_t m_column_ix;
+};
+
+// --------------------------------------------------------------------
+// iterator proxy
+
+template <typename Category, typename... Ts>
+class iterator_proxy
+{
+  public:
+	static constexpr const size_t N = sizeof...(Ts);
+
+	using category_type = Category;
+	using row_type = std::conditional_t<std::is_const_v<category_type>, const row, row>;
+
+	using iterator = iterator_impl<category_type, Ts...>;
+	using row_iterator = iterator_impl<category_type>;
+
+	iterator_proxy(category_type &cat, row_iterator pos, char const *const columns[N]);
+	iterator_proxy(category_type &cat, row_iterator pos, std::initializer_list<char const *> columns);
+
+	iterator_proxy(iterator_proxy &&p);
+	iterator_proxy &operator=(iterator_proxy &&p);
+
+	iterator_proxy(const iterator_proxy &) = delete;
+	iterator_proxy &operator=(const iterator_proxy &) = delete;
+
+	iterator begin() const { return iterator(m_begin, m_column_ix); }
+	iterator end() const { return iterator(m_end, m_column_ix); }
+
+	bool empty() const { return m_begin == m_end; }
+
+	explicit operator bool() const { return not empty(); }
+
+	size_t size() const { return std::distance(begin(), end()); }
+
+	// row front() { return *begin(); }
+	// row back() { return *(std::prev(end())); }
+
+	category_type &category() const { return *m_category; }
+
+	void swap(iterator_proxy &rhs)
+	{
+		std::swap(m_category, rhs.m_category);
+		std::swap(m_begin, rhs.m_begin);
+		std::swap(m_end, rhs.m_end);
+		std::swap(m_column_ix, rhs.m_column_ix);
+	}
+
+  private:
+	category_type *m_category;
+	row_iterator m_begin, m_end;
+	std::array<uint16_t, N> m_column_ix;
+};
+
+// --------------------------------------------------------------------
+// conditional iterator proxy
+
+template <typename CategoryType, typename... Ts>
+class conditional_iterator_proxy
+{
+  public:
+	static constexpr const size_t N = sizeof...(Ts);
+
+	using category_type = std::remove_cv_t<CategoryType>;
+
+	using base_iterator = iterator_impl<CategoryType, Ts...>;
+	using value_type = typename base_iterator::value_type;
+	using row_type = typename base_iterator::row_type;
+	using row_iterator = iterator_impl<CategoryType>;
+
+	class conditional_iterator_impl
+	{
+	  public:
+		using iterator_category = std::forward_iterator_tag;
+		using value_type = conditional_iterator_proxy::value_type;
+		using difference_type = std::ptrdiff_t;
+		using pointer = value_type *;
+		using reference = value_type;
+
+		conditional_iterator_impl(CategoryType &cat, row_iterator pos, const condition &cond, const std::array<uint16_t, N> &cix);
+		conditional_iterator_impl(const conditional_iterator_impl &i) = default;
+		conditional_iterator_impl &operator=(const conditional_iterator_impl &i) = default;
+
+		virtual ~conditional_iterator_impl() = default;
+
+		reference operator*()
+		{
+			return *mBegin;
+		}
+
+		pointer operator->()
+		{
+			return &*mBegin;
+		}
+
+		conditional_iterator_impl &operator++()
+		{
+			while (mBegin != mEnd)
+			{
+				if (++mBegin == mEnd)
+					break;
+
+				if (m_condition->operator()(mBegin))
+					break;
+			}
+
+			return *this;
+		}
+
+		conditional_iterator_impl operator++(int)
+		{
+			conditional_iterator_impl result(*this);
+			this->operator++();
+			return result;
+		}
+
+		bool operator==(const conditional_iterator_impl &rhs) const { return mBegin == rhs.mBegin; }
+		bool operator!=(const conditional_iterator_impl &rhs) const { return mBegin != rhs.mBegin; }
+
+		template <typename IRowType, typename... ITs>
+		bool operator==(const iterator_impl<IRowType, ITs...> &rhs) const { return mBegin == rhs; }
+
+		template <typename IRowType, typename... ITs>
+		bool operator!=(const iterator_impl<IRowType, ITs...> &rhs) const { return mBegin != rhs; }
+
+	  private:
+		CategoryType *mCat;
+		base_iterator mBegin, mEnd;
+		const condition *m_condition;
+	};
+
+	using iterator = conditional_iterator_impl;
+	using reference = typename iterator::reference;
+
+	template <typename... Ns>
+	conditional_iterator_proxy(CategoryType &cat, row_iterator pos, condition &&cond, Ns... names);
+
+	conditional_iterator_proxy(conditional_iterator_proxy &&p);
+	conditional_iterator_proxy &operator=(conditional_iterator_proxy &&p);
+
+	conditional_iterator_proxy(const conditional_iterator_proxy &) = delete;
+	conditional_iterator_proxy &operator=(const conditional_iterator_proxy &) = delete;
+
+	iterator begin() const;
+	iterator end() const;
+
+	bool empty() const;
+
+	explicit operator bool() const { return not empty(); }
+
+	size_t size() const { return std::distance(begin(), end()); }
+
+	row_handle front() { return *begin(); }
+	// row_handle back() { return *begin(); }
+
+	CategoryType &category() const { return *m_cat; }
+
+	void swap(conditional_iterator_proxy &rhs);
+
+  private:
+	CategoryType *m_cat;
+	condition m_condition;
+	row_iterator mCBegin, mCEnd;
+	std::array<uint16_t, N> mCix;
+};
+
+// --------------------------------------------------------------------
+
+template <typename Category, typename... Ts>
+iterator_proxy<Category, Ts...>::iterator_proxy(Category &cat, row_iterator pos, char const *const columns[N])
+	: m_category(&cat)
+	, m_begin(pos)
+	, m_end(cat.end())
+{
+	for (uint16_t i = 0; i < N; ++i)
+		m_column_ix[i] = m_category->get_column_ix(columns[i]);
+}
+
+template <typename Category, typename... Ts>
+iterator_proxy<Category, Ts...>::iterator_proxy(Category &cat, row_iterator pos, std::initializer_list<char const *> columns)
+	: m_category(&cat)
+	, m_begin(pos)
+	, m_end(cat.end())
+{
+	// static_assert(columns.size() == N, "The list of column names should be exactly the same as the list of requested columns");
+
+	std::uint16_t i = 0;
+	for (auto column : columns)
+		m_column_ix[i++] = m_category->get_column_ix(column);
+}
+
+// --------------------------------------------------------------------
+
+template <typename Category, typename... Ts>
+conditional_iterator_proxy<Category, Ts...>::conditional_iterator_impl::conditional_iterator_impl(
+	Category &cat, row_iterator pos, const condition &cond, const std::array<uint16_t, N> &cix)
+	: mCat(&cat)
+	, mBegin(pos, cix)
+	, mEnd(cat.end(), cix)
+	, m_condition(&cond)
+{
+}
+
+template <typename Category, typename... Ts>
+conditional_iterator_proxy<Category, Ts...>::conditional_iterator_proxy(conditional_iterator_proxy &&p)
+	: m_cat(nullptr)
+	, mCBegin(p.mCBegin)
+	, mCEnd(p.mCEnd)
+	, mCix(p.mCix)
+{
+	std::swap(m_cat, p.m_cat);
+	std::swap(mCix, p.mCix);
+	m_condition.swap(p.m_condition);
+}
+
+template <typename Category, typename... Ts>
+template <typename... Ns>
+conditional_iterator_proxy<Category, Ts...>::conditional_iterator_proxy(Category &cat, row_iterator pos, condition &&cond, Ns... names)
+	: m_cat(&cat)
+	, m_condition(std::move(cond))
+	, mCBegin(pos)
+	, mCEnd(cat.end())
+{
+	static_assert(sizeof...(Ts) == sizeof...(Ns), "Number of column names should be equal to number of requested value types");
+
+	m_condition.prepare(cat);
+
+	while (mCBegin != mCEnd and not m_condition(*mCBegin))
+		++mCBegin;
+
+	uint16_t i = 0;
+	((mCix[i++] = m_cat->get_column_ix(names)), ...);
+}
+
+template <typename Category, typename... Ts>
+conditional_iterator_proxy<Category, Ts...> &conditional_iterator_proxy<Category, Ts...>::operator=(conditional_iterator_proxy &&p)
+{
+	swap(p);
+	return *this;
+}
+
+template <typename Category, typename... Ts>
+typename conditional_iterator_proxy<Category, Ts...>::iterator conditional_iterator_proxy<Category, Ts...>::begin() const
+{
+	return iterator(*m_cat, mCBegin, m_condition, mCix);
+}
+
+template <typename Category, typename... Ts>
+typename conditional_iterator_proxy<Category, Ts...>::iterator conditional_iterator_proxy<Category, Ts...>::end() const
+{
+	return iterator(*m_cat, mCEnd, m_condition, mCix);
+}
+
+template <typename Category, typename... Ts>
+bool conditional_iterator_proxy<Category, Ts...>::empty() const
+{
+	return mCBegin == mCEnd;
+}
+
+template <typename Category, typename... Ts>
+void conditional_iterator_proxy<Category, Ts...>::swap(conditional_iterator_proxy &rhs)
+{
+	std::swap(m_cat, rhs.m_cat);
+	m_condition.swap(rhs.m_condition);
+	std::swap(mCBegin, rhs.mCBegin);
+	std::swap(mCEnd, rhs.mCEnd);
+	std::swap(mCix, rhs.mCix);
+}
+
+} // namespace cif

include/cif++/list.hpp

@@ -0,0 +1,79 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <memory>
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+template<typename Allocator = std::allocator<void>>
+class list
+{
+  public:
+
+
+
+  protected:
+
+	struct list_item
+	{
+		list_item *m_next = nullptr;
+	};
+
+	using list_item_allocator_type = typename std::allocator_traits<Alloc>::template rebind_alloc<list_item>;
+	using list_item_allocator_traits = std::allocator_traits<item_allocator_type>;
+
+	list_item_allocator_traits::pointer get_item()
+	{
+		list_item_allocator_type ia(get_allocator());
+		return list_item_allocator_traits::allocate(ia, 1);
+	}
+
+	template<typename ...Arguments>
+	list_item *create_list_item(uint16_t column_ix, Arguments... args)
+	{
+		auto p = this->get_item();
+		list_item_allocator_type ia(get_allocator());
+		list_item_allocator_traits::construct(ia, p, std::forward<Arguments>(args)...);
+		return p;
+	}
+
+	void delete_list_item(list_item *iv)
+	{
+		list_item_allocator_type ia(get_allocator());
+		list_item_allocator_traits::destroy(ia, iv);
+		list_item_allocator_traits::deallocate(ia, iv, 1);
+	}
+
+	list_item *m_head = nullptr, *m_tail = nullptr;
+};
+
+
+} // namespace cif

include/cif++/model.hpp

@@ -0,0 +1,928 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <numeric>
+
+#include <cif++/atom_type.hpp>
+
+#if __cpp_lib_format
+#include <format>
+#endif
+
+namespace cif::mm
+{
+
+class atom;
+class residue;
+class monomer;
+class polymer;
+class structure;
+
+// --------------------------------------------------------------------
+
+class atom
+{
+  private:
+	struct atom_impl : public std::enable_shared_from_this<atom_impl>
+	{
+		atom_impl(datablock &db, std::string_view id)
+			: m_db(db)
+			, m_cat(db["atom_site"])
+			, m_id(id)
+		{
+			auto r = row();
+			if (r)
+				tie(m_location.m_x, m_location.m_y, m_location.m_z) = r.get("Cartn_x", "Cartn_y", "Cartn_z");
+		}
+
+		// constructor for a symmetry copy of an atom
+		atom_impl(const atom_impl &impl, const point &loc, const std::string &sym_op)
+			: atom_impl(impl)
+		{
+			m_location = loc;
+			m_symop = sym_op;
+		}
+
+		atom_impl(const atom_impl &i) = default;
+
+		void prefetch();
+
+		int compare(const atom_impl &b) const;
+
+		// bool getAnisoU(float anisou[6]) const;
+
+		int get_charge() const;
+
+		void moveTo(const point &p);
+
+		// const compound *compound() const;
+
+		std::string get_property(std::string_view name) const;
+		int get_property_int(std::string_view name) const;
+		float get_property_float(std::string_view name) const;
+
+		void set_property(const std::string_view name, const std::string &value);
+
+		row_handle row()
+		{
+			return m_cat[{{"id", m_id}}];
+		}
+
+		const row_handle row() const
+		{
+			return m_cat[{{"id", m_id}}];
+		}
+
+		row_handle row_aniso()
+		{
+			auto cat = m_db.get("atom_site_anisotrop");
+			return cat ? cat->find1(key("id") == m_id) : row_handle{};
+		}
+
+		const row_handle row_aniso() const
+		{
+			auto cat = m_db.get("atom_site_anisotrop");
+			return cat ? cat->find1(key("id") == m_id) : row_handle{};
+		}
+
+		const datablock &m_db;
+		category &m_cat;
+		std::string m_id;
+		point m_location;
+		std::string m_symop = "1_555";
+	};
+
+  public:
+	atom() {}
+
+	atom(std::shared_ptr<atom_impl> impl)
+		: m_impl(impl)
+	{
+	}
+
+	atom(const atom &rhs)
+		: m_impl(rhs.m_impl)
+	{
+	}
+
+	atom(datablock &db, row_handle &row)
+		: atom(std::make_shared<atom_impl>(db, row["id"].as<std::string>()))
+	{
+	}
+
+	// a special constructor to create symmetry copies
+	atom(const atom &rhs, const point &symmmetry_location, const std::string &symmetry_operation)
+		: atom(std::make_shared<atom_impl>(*rhs.m_impl, symmmetry_location, symmetry_operation))
+	{
+	}
+
+	explicit operator bool() const { return (bool)m_impl; }
+
+	// // return a copy of this atom, with data copied instead of referenced
+	// atom clone() const
+	// {
+	// 	auto copy = std::make_shared<atom_impl>(*m_impl);
+	// 	copy->mClone = true;
+	// 	return atom(copy);
+	// }
+
+	atom &operator=(const atom &rhs) = default;
+
+	// template <typename T>
+	// T get_property(const std::string_view name) const;
+
+	std::string get_property(std::string_view name) const
+	{
+		if (not m_impl)
+			throw std::logic_error("Error trying to fetch a property from an uninitialized atom");
+		return m_impl->get_property(name);
+	}
+
+	int get_property_int(std::string_view name) const
+	{
+		if (not m_impl)
+			throw std::logic_error("Error trying to fetch a property from an uninitialized atom");
+		return m_impl->get_property_int(name);
+	}
+
+	float get_property_float(std::string_view name) const
+	{
+		if (not m_impl)
+			throw std::logic_error("Error trying to fetch a property from an uninitialized atom");
+		return m_impl->get_property_float(name);
+	}
+
+	void set_property(const std::string_view name, const std::string &value)
+	{
+		if (not m_impl)
+			throw std::logic_error("Error trying to modify an uninitialized atom");
+		m_impl->set_property(name, value);
+	}
+
+	template <typename T, std::enable_if_t<std::is_arithmetic_v<T>, int> = 0>
+	void set_property(const std::string_view name, const T &value)
+	{
+		set_property(name, std::to_string(value));
+	}
+
+	const std::string &id() const { return impl().m_id; }
+
+	cif::atom_type get_type() const { return atom_type_traits(get_property("type_symbol")).type(); }
+
+	point get_location() const { return impl().m_location; }
+	void set_location(point p)
+	{
+		if (not m_impl)
+			throw std::logic_error("Error trying to modify an uninitialized atom");
+		m_impl->moveTo(p);
+	}
+
+	/// \brief Translate the position of this atom by \a t
+	void translate(point t)
+	{
+		set_location(get_location() + t);
+	}
+
+	/// \brief Rotate the position of this atom by \a q
+	void rotate(quaternion q)
+	{
+		auto loc = get_location();
+		loc.rotate(q);
+		set_location(loc);
+	}
+
+	/// \brief Translate and rotate the position of this atom by \a t and \a q
+	void translate_and_rotate(point t, quaternion q)
+	{
+		auto loc = get_location();
+		loc += t;
+		loc.rotate(q);
+		set_location(loc);
+	}
+
+	/// \brief Translate, rotate and translate again the coordinates this atom by \a t1 , \a q and \a t2
+	void translate_rotate_and_translate(point t1, quaternion q, point t2)
+	{
+		auto loc = get_location();
+		loc += t1;
+		loc.rotate(q);
+		loc += t2;
+		set_location(loc);
+	}
+
+	// for direct access to underlying data, be careful!
+	const row_handle get_row() const { return impl().row(); }
+	const row_handle get_row_aniso() const { return impl().row_aniso(); }
+
+	bool is_symmetry_copy() const { return impl().m_symop != "1_555"; }
+	std::string symmetry() const { return impl().m_symop; }
+
+	// const compound &compound() const;
+
+	bool is_water() const
+	{
+		auto comp_id = get_label_comp_id();
+		return comp_id == "HOH" or comp_id == "H2O" or comp_id == "WAT";
+	}
+
+	int get_charge() const { return impl().get_charge(); }
+
+	// float uIso() const;
+	// bool getAnisoU(float anisou[6]) const { return impl().getAnisoU(anisou); }
+	
+	float get_occupancy() const { return get_property_float("occupancy"); }
+
+	// specifications
+
+	std::string get_label_asym_id() const { return get_property("label_asym_id"); }
+	int get_label_seq_id() const { return get_property_int("label_seq_id"); }
+	std::string get_label_atom_id() const { return get_property("label_atom_id"); }
+	std::string get_label_alt_id() const { return get_property("label_alt_id"); }
+	std::string get_label_comp_id() const { return get_property("label_comp_id"); }
+	std::string get_label_entity_id() const { return get_property("label_entity_id"); }
+
+	std::string get_auth_asym_id() const { return get_property("auth_asym_id"); }
+	std::string get_auth_seq_id() const { return get_property("auth_seq_id"); }
+	std::string get_auth_atom_id() const { return get_property("auth_atom_id"); }
+	std::string get_auth_alt_id() const { return get_property("auth_alt_id"); }
+	std::string get_auth_comp_id() const { return get_property("auth_comp_id"); }
+	std::string get_pdb_ins_code() const { return get_property("pdbx_PDB_ins_code"); }
+
+	bool is_alternate() const { return not get_label_alt_id().empty(); }
+
+	// std::string labelID() const; // label_comp_id + '_' + label_asym_id + '_' + label_seq_id
+	
+	std::string pdb_id() const
+	{
+		return get_label_comp_id() + '_' + get_auth_asym_id() + '_' + get_auth_seq_id() + get_pdb_ins_code();
+	}
+
+	bool operator==(const atom &rhs) const
+	{
+		if (m_impl == rhs.m_impl)
+			return true;
+
+		if (not(m_impl and rhs.m_impl))
+			return false;
+
+		return &m_impl->m_db == &rhs.m_impl->m_db and m_impl->m_id == rhs.m_impl->m_id;
+	}
+
+	bool operator!=(const atom &rhs) const
+	{
+		return not operator==(rhs);
+	}
+
+	// // access data in compound for this atom
+
+	// convenience routine
+	bool is_back_bone() const
+	{
+		auto atomID = get_label_atom_id();
+		return atomID == "N" or atomID == "O" or atomID == "C" or atomID == "CA";
+	}
+
+	void swap(atom &b)
+	{
+		std::swap(m_impl, b.m_impl);
+	}
+
+	int compare(const atom &b) const { return impl().compare(*b.m_impl); }
+
+	bool operator<(const atom &rhs) const
+	{
+		return compare(rhs) < 0;
+	}
+
+	friend std::ostream &operator<<(std::ostream &os, const atom &atom);
+
+	// /// \brief Synchronize data with underlying cif data
+	// void sync()
+	// {
+	// 	if (m_impl)
+	// 		m_impl->prefetch();
+	// }
+
+  private:
+	friend class structure;
+
+	const atom_impl &impl() const
+	{
+		if (not m_impl)
+			throw std::runtime_error("Uninitialized atom, not found?");
+		return *m_impl;
+	}
+
+	std::shared_ptr<atom_impl> m_impl;
+};
+
+// template <>
+// inline std::string atom::get_property<std::string>(const std::string_view name) const
+// {
+// 	return get_property(name);
+// }
+
+// template <>
+// inline int atom::get_property<int>(const std::string_view name) const
+// {
+// 	auto v = impl().get_property(name);
+// 	return v.empty() ? 0 : stoi(v);
+// }
+
+// template <>
+// inline float atom::get_property<float>(const std::string_view name) const
+// {
+// 	return stof(impl().get_property(name));
+// }
+
+inline void swap(atom &a, atom &b)
+{
+	a.swap(b);
+}
+
+inline float distance(const atom &a, const atom &b)
+{
+	return distance(a.get_location(), b.get_location());
+}
+
+inline float distance_squared(const atom &a, const atom &b)
+{
+	return distance_squared(a.get_location(), b.get_location());
+}
+
+// --------------------------------------------------------------------
+
+enum class EntityType
+{
+	polymer,
+	NonPolymer,
+	Macrolide,
+	Water,
+	Branched
+};
+
+// --------------------------------------------------------------------
+
+class residue
+{
+  public:
+	friend class structure;
+
+	// constructor
+	residue(const structure &structure, const std::string &compoundID,
+		const std::string &asymID, int seqID,
+		const std::string &authAsymID, const std::string &authSeqID,
+		const std::string &pdbInsCode)
+		: m_structure(&structure)
+		, m_compound_id(compoundID)
+		, m_asym_id(asymID)
+		, m_seq_id(seqID)
+		, m_auth_asym_id(authAsymID)
+		, m_auth_seq_id(authSeqID)
+		, m_pdb_ins_code(pdbInsCode)
+	{
+	}
+
+	residue(const residue &rhs) = delete;
+	residue &operator=(const residue &rhs) = delete;
+
+	residue(residue &&rhs) = default;
+	residue &operator=(residue &&rhs) = default;
+
+	virtual ~residue() = default;
+
+	std::string get_entity_id() const;
+
+	EntityType entity_type() const;
+
+	const std::string &get_asym_id() const { return m_asym_id; }
+	int get_seq_id() const { return m_seq_id; }
+
+	const std::string get_auth_asym_id() const { return m_auth_asym_id; }
+	const std::string get_auth_seq_id() const { return m_auth_seq_id; }
+	std::string get_pdb_ins_code() const { return m_pdb_ins_code; }
+
+	const std::string &get_compound_id() const { return m_compound_id; }
+	void set_compound_id(const std::string &id) { m_compound_id = id; }
+
+	const structure *get_structure() const { return m_structure; }
+
+	// const compound &compound() const;
+
+	std::vector<atom> &atoms()
+	{
+		return m_atoms;
+	}
+
+	const std::vector<atom> &atoms() const
+	{
+		return m_atoms;
+	}
+
+	void add_atom(atom &atom);
+
+	/// \brief Unique atoms returns only the atoms without alternates and the first of each alternate atom id.
+	std::vector<atom> unique_atoms() const;
+
+	/// \brief The alt ID used for the unique atoms
+	std::string unique_alt_id() const;
+
+	atom get_atom_by_atom_id(const std::string &atomID) const;
+
+	// Is this residue a single entity?
+	bool is_entity() const;
+	bool is_water() const { return m_compound_id == "HOH"; }
+	// bool empty() const { return m_structure == nullptr; }
+
+	bool has_alternate_atoms() const;
+
+	/// \brief Return the list of unique alt ID's present in this residue
+	std::set<std::string> get_alternate_ids() const;
+
+	/// \brief Return the list of unique atom ID's
+	std::set<std::string> get_atom_ids() const;
+
+	/// \brief Return the list of atoms having ID \a atomID
+	std::vector<atom> get_atoms_by_id(const std::string &atomID) const;
+
+	// some routines for 3d work
+	std::tuple<point, float> center_and_radius() const;
+
+	friend std::ostream &operator<<(std::ostream &os, const residue &res);
+
+	bool operator==(const residue &rhs) const
+	{
+		return this == &rhs or (m_structure == rhs.m_structure and
+								   m_seq_id == rhs.m_seq_id and
+								   m_asym_id == rhs.m_asym_id and
+								   m_compound_id == rhs.m_compound_id and
+								   m_auth_seq_id == rhs.m_auth_seq_id);
+	}
+
+  protected:
+	residue() {}
+
+	const structure *m_structure = nullptr;
+	std::string m_compound_id, m_asym_id;
+	int m_seq_id = 0;
+	std::string m_auth_asym_id, m_auth_seq_id, m_pdb_ins_code;
+	std::vector<atom> m_atoms;
+};
+
+// --------------------------------------------------------------------
+// a monomer models a single residue in a protein chain
+
+class monomer : public residue
+{
+  public:
+	//	monomer();
+	monomer(const monomer &rhs) = delete;
+	monomer &operator=(const monomer &rhs) = delete;
+
+	monomer(monomer &&rhs);
+	monomer &operator=(monomer &&rhs);
+
+	monomer(const polymer &polymer, size_t index, int seqID, const std::string &authSeqID,
+		const std::string &pdbInsCode, const std::string &compoundID);
+
+	bool is_first_in_chain() const;
+	bool is_last_in_chain() const;
+
+	// convenience
+	bool has_alpha() const;
+	bool has_kappa() const;
+
+	// Assuming this is really an amino acid...
+
+	float phi() const;
+	float psi() const;
+	float alpha() const;
+	float kappa() const;
+	float tco() const;
+	float omega() const;
+
+	// torsion angles
+	size_t nr_of_chis() const;
+	float chi(size_t i) const;
+
+	bool is_cis() const;
+
+	/// \brief Returns true if the four atoms C, CA, N and O are present
+	bool is_complete() const;
+
+	/// \brief Returns true if any of the backbone atoms has an alternate
+	bool has_alternate_backbone_atoms() const;
+
+	atom CAlpha() const { return get_atom_by_atom_id("CA"); }
+	atom C() const { return get_atom_by_atom_id("C"); }
+	atom N() const { return get_atom_by_atom_id("N"); }
+	atom O() const { return get_atom_by_atom_id("O"); }
+	atom H() const { return get_atom_by_atom_id("H"); }
+
+	bool is_bonded_to(const monomer &rhs) const
+	{
+		return this != &rhs and are_bonded(*this, rhs);
+	}
+
+	static bool are_bonded(const monomer &a, const monomer &b, float errorMargin = 0.5f);
+	static bool is_cis(const monomer &a, const monomer &b);
+	static float omega(const monomer &a, const monomer &b);
+
+	// for LEU and VAL
+	float chiral_volume() const;
+
+	bool operator==(const monomer &rhs) const
+	{
+		return m_polymer == rhs.m_polymer and m_index == rhs.m_index;
+	}
+
+  private:
+	const polymer *m_polymer;
+	size_t m_index;
+};
+
+// --------------------------------------------------------------------
+
+class polymer : public std::vector<monomer>
+{
+  public:
+	polymer(const structure &s, const std::string &entityID, const std::string &asymID, const std::string &auth_asym_id);
+
+	polymer(const polymer &) = delete;
+	polymer &operator=(const polymer &) = delete;
+
+	// monomer &getBySeqID(int seqID);
+	// const monomer &getBySeqID(int seqID) const;
+
+	const structure *get_structure() const { return m_structure; }
+
+	std::string get_asym_id() const { return m_asym_id; }
+	std::string get_auth_asym_id() const { return m_auth_asym_id; }	// The PDB chain ID, actually
+	std::string get_entity_id() const { return m_entity_id; }
+
+	// int Distance(const monomer &a, const monomer &b) const;
+
+  private:
+	const structure *m_structure;
+	std::string m_entity_id;
+	std::string m_asym_id;
+	std::string m_auth_asym_id;
+};
+
+// --------------------------------------------------------------------
+// sugar and branch, to describe glycosylation sites
+
+class branch;
+
+class sugar : public residue
+{
+  public:
+	sugar(const branch &branch, const std::string &compoundID,
+		const std::string &asymID, int authSeqID);
+
+	sugar(sugar &&rhs);
+	sugar &operator=(sugar &&rhs);
+
+	int num() const {
+		int result;
+		auto r = std::from_chars(m_auth_seq_id.data(), m_auth_seq_id.data() + m_auth_seq_id.length(), result);
+		if (r.ec != std::errc())
+			throw std::runtime_error("The auth_seq_id should be a number for a sugar");
+		return result;
+	}
+	std::string name() const;
+
+	/// \brief Return the atom the C1 is linked to
+	atom get_link() const { return m_link; }
+	void set_link(atom link) { m_link = link; }
+
+	size_t get_link_nr() const
+	{
+		size_t result = 0;
+		if (m_link)
+			result = m_link.get_property_int("auth_seq_id");
+		return result;
+	}
+
+  private:
+	const branch *m_branch;
+	atom m_link;
+};
+
+class branch : public std::vector<sugar>
+{
+  public:
+	branch(structure &structure, const std::string &asymID);
+
+	void link_atoms();
+
+	std::string name() const;
+	float weight() const;
+	std::string get_asym_id() const { return m_asym_id; }
+
+	structure &get_structure() { return *m_structure; }
+	const structure &get_structure() const { return *m_structure; }
+
+	sugar &getSugarByNum(int nr);
+	const sugar &getSugarByNum(int nr) const;
+
+  private:
+	friend sugar;
+
+	std::string name(const sugar &s) const;
+
+	structure *m_structure;
+	std::string m_asym_id;
+};
+
+// // --------------------------------------------------------------------
+// // file is a reference to the data stored in e.g. the cif file.
+// // This object is not copyable.
+
+// class File : public file
+// {
+//   public:
+// 	File() {}
+
+// 	// File(const std::filesystem::path &path)
+// 	// {
+// 	// 	load(path);
+// 	// }
+
+// 	// File(const char *data, size_t length)
+// 	// {
+// 	// 	load(data, length);
+// 	// }
+
+// 	File(const File &) = delete;
+// 	File &operator=(const File &) = delete;
+
+// 	// void load(const std::filesystem::path &p) override;
+// 	// void save(const std::filesystem::path &p) override;
+
+// 	// using file::load;
+// 	// using file::save;
+
+// 	datablock &data() { return front(); }
+// };
+
+// --------------------------------------------------------------------
+
+enum class StructureOpenOptions
+{
+	SkipHydrogen = 1 << 0
+};
+
+inline bool operator&(StructureOpenOptions a, StructureOpenOptions b)
+{
+	return static_cast<int>(a) bitand static_cast<int>(b);
+}
+
+// --------------------------------------------------------------------
+
+class structure
+{
+  public:
+	structure(file &p, size_t modelNr = 1, StructureOpenOptions options = {});
+
+	structure(datablock &db, size_t modelNr = 1, StructureOpenOptions options = {});
+
+	structure(structure &&s) = default;
+
+	// Create a read-only clone of the current structure (for multithreaded calculations that move atoms)
+	// NOTE: removed, simply create a new structure for each thread
+	structure(const structure &) = delete;
+
+	structure &operator=(const structure &) = delete;
+	// Structure &operator=(Structure &&s) = default;
+
+	~structure() = default;
+
+	size_t get_model_nr() const { return m_model_nr; }
+
+	const std::vector<atom> &atoms() const { return m_atoms; }
+	// std::vector<atom> &atoms() { return m_atoms; }
+
+	EntityType get_entity_type_for_entity_id(const std::string entityID) const;
+	EntityType get_entity_type_for_asym_id(const std::string asymID) const;
+
+	// std::vector<atom> waters() const;
+
+	const std::list<polymer> &polymers() const { return m_polymers; }
+	std::list<polymer> &polymers() { return m_polymers; }
+
+	polymer &get_polymer_by_asym_id(const std::string &asymID);
+
+	const polymer &get_polymer_by_asym_id(const std::string &asymID) const
+	{
+		return const_cast<structure *>(this)->get_polymer_by_asym_id(asymID);
+	}
+
+	const std::list<branch> &branches() const { return m_branches; }
+	std::list<branch> &branches() { return m_branches; }
+
+	branch &get_branch_by_asym_id(const std::string &asymID);
+	const branch &get_branch_by_asym_id(const std::string &asymID) const;
+
+	const std::vector<residue> &non_polymers() const { return m_non_polymers; }
+
+	atom get_atom_by_id(const std::string &id) const;
+	// atom getAtomByLocation(point pt, float maxDistance) const;
+
+	atom get_atom_by_label(const std::string &atomID, const std::string &asymID,
+		const std::string &compID, int seqID, const std::string &altID = "");
+
+	// /// \brief Return the atom closest to point \a p
+	atom get_atom_by_position(point p) const;
+
+	/// \brief Return the atom closest to point \a p with atom type \a type in a residue of type \a res_type
+	atom get_atom_by_position_and_type(point p, std::string_view type, std::string_view res_type) const;
+
+	/// \brief Get a non-poly residue for an asym with id \a asymID
+	residue &get_residue(const std::string &asymID)
+	{
+		return get_residue(asymID, 0, "");
+	}
+
+	/// \brief Get a non-poly residue for an asym with id \a asymID
+	const residue &get_residue(const std::string &asymID) const
+	{
+		return get_residue(asymID, 0, "");
+	}
+
+	/// \brief Get a residue for an asym with id \a asymID seq id \a seqID and authSeqID \a authSeqID
+	residue &get_residue(const std::string &asymID, int seqID, const std::string &authSeqID);
+
+	/// \brief Get a the single residue for an asym with id \a asymID seq id \a seqID and authSeqID \a authSeqID
+	const residue &get_residue(const std::string &asymID, int seqID, const std::string &authSeqID) const
+	{
+		return const_cast<structure *>(this)->get_residue(asymID, seqID, authSeqID);
+	}
+
+	/// \brief Get a residue for an asym with id \a asymID, compound id \a compID, seq id \a seqID and authSeqID \a authSeqID
+	residue &get_residue(const std::string &asymID, const std::string &compID, int seqID, const std::string &authSeqID);
+
+	/// \brief Get a residue for an asym with id \a asymID, compound id \a compID, seq id \a seqID and authSeqID \a authSeqID
+	const residue &get_residue(const std::string &asymID, const std::string &compID, int seqID, const std::string &authSeqID) const
+	{
+		return const_cast<structure *>(this)->get_residue(asymID, compID, seqID, authSeqID);
+	}
+
+	/// \brief Get a the residue for atom \a atom
+	residue &get_residue(const atom &atom)
+	{
+		return get_residue(atom.get_label_asym_id(), atom.get_label_comp_id(), atom.get_label_seq_id(), atom.get_auth_seq_id());
+	}
+
+	/// \brief Get a the residue for atom \a atom
+	const residue &get_residue(const atom &atom) const
+	{
+		return get_residue(atom.get_label_asym_id(), atom.get_label_comp_id(), atom.get_label_seq_id(), atom.get_auth_seq_id());
+	}
+
+	// Actions
+	void remove_atom(atom &a)
+	{
+		remove_atom(a, true);
+	}
+
+	void swap_atoms(atom a1, atom a2); // swap the labels for these atoms
+	void move_atom(atom a, point p);   // move atom to a new location
+	void change_residue(residue &res, const std::string &newcompound,
+		const std::vector<std::tuple<std::string, std::string>> &remappedAtoms);
+
+	/// \brief Remove a residue, can be monomer or nonpoly
+	///
+	/// \param asym_id     The asym ID
+	/// \param seq_id      The sequence ID
+	void remove_residue(const std::string &asym_id, int seq_id, const std::string &auth_seq_id)
+	{
+		remove_residue(get_residue(asym_id, seq_id, auth_seq_id));
+	}
+
+	/// \brief Create a new non-polymer entity, returns new ID
+	/// \param mon_id	The mon_id for the new nonpoly, must be an existing and known compound from CCD
+	/// \return			The ID of the created entity
+	std::string create_non_poly_entity(const std::string &mon_id);
+
+	/// \brief Create a new NonPolymer struct_asym with atoms constructed from \a atoms, returns asym_id.
+	/// This method assumes you are copying data from one cif file to another.
+	///
+	/// \param entity_id	The entity ID of the new nonpoly
+	/// \param atoms		The array of atom_site rows containing the data.
+	/// \return				The newly create asym ID
+	std::string create_non_poly(const std::string &entity_id, const std::vector<atom> &atoms);
+
+	/// \brief Create a new NonPolymer struct_asym with atoms constructed from info in \a atom_info, returns asym_id.
+	/// This method creates new atom records filled with info from the info.
+	///
+	/// \param entity_id	The entity ID of the new nonpoly
+	/// \param atoms		The array of sets of item data containing the data for the atoms.
+	/// \return				The newly create asym ID
+	std::string create_non_poly(const std::string &entity_id, std::vector<row_initializer> atoms);
+
+	/// \brief Create a new (sugar) branch with one first NAG containing atoms constructed from \a atoms
+	branch &create_branch(std::vector<row_initializer> atoms);
+
+	/// \brief Extend an existing (sugar) branch identified by \a asymID with one sugar containing atoms constructed from \a atom_info
+	///
+	/// \param asym_id      The asym id of the branch to extend
+	/// \param atom_info    Array containing the info for the atoms to construct for the new sugar
+	/// \param link_sugar   The sugar to link to, note: this is the sugar number (1 based)
+	/// \param link_atom    The atom id of the atom linked in the sugar
+	branch &extend_branch(const std::string &asym_id, std::vector<row_initializer> atom_info,
+		int link_sugar, const std::string &link_atom);
+
+	/// \brief Remove \a branch
+	void remove_branch(branch &branch);
+
+	/// \brief Remove residue \a res
+	///
+	/// \param res         The residue to remove
+	void remove_residue(residue &res);
+
+	/// \brief Translate the coordinates of all atoms in the structure by \a t
+	void translate(point t);
+
+	/// \brief Rotate the coordinates of all atoms in the structure by \a q
+	void rotate(quaternion t);
+
+	/// \brief Translate and rotate the coordinates of all atoms in the structure by \a t and \a q
+	void translate_and_rotate(point t, quaternion q);
+
+	/// \brief Translate, rotate and translate again the coordinates of all atoms in the structure by \a t1 , \a q and \a t2
+	void translate_rotate_and_translate(point t1, quaternion q, point t2);
+
+	void cleanup_empty_categories();
+
+	/// \brief Direct access to underlying data
+	category &get_category(std::string_view name) const
+	{
+		return m_db[name];
+	}
+
+	datablock &get_datablock() const
+	{
+		return m_db;
+	}
+
+	void validate_atoms() const;
+
+  private:
+	friend polymer;
+	friend residue;
+
+	std::string insert_compound(const std::string &compoundID, bool is_entity);
+
+	std::string create_entity_for_branch(branch &branch);
+
+	void load_data();
+
+	void load_atoms_for_model(StructureOpenOptions options);
+
+	template <typename... Args>
+	atom &emplace_atom(Args... args)
+	{
+		return emplace_atom(atom{ std::forward<Args>(args)... });
+	}
+
+	atom &emplace_atom(atom &&atom);
+
+	void remove_atom(atom &a, bool removeFromResidue);
+	void remove_sugar(sugar &sugar);
+
+	datablock &m_db;
+	size_t m_model_nr;
+	std::vector<atom> m_atoms;
+	std::vector<size_t> m_atom_index;
+	std::list<polymer> m_polymers;
+	std::list<branch> m_branches;
+	std::vector<residue> m_non_polymers;
+};
+
+} // namespace cif::mm

include/cif++/parser.hpp

@@ -0,0 +1,293 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <map>
+
+#include <cif++/row.hpp>
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+class parse_error : public std::runtime_error
+{
+  public:
+	parse_error(uint32_t line_nr, const std::string &message)
+		: std::runtime_error("parse error at line " + std::to_string(line_nr) + ": " + message)
+	{
+	}
+};
+
+// --------------------------------------------------------------------
+
+// TODO: Need to implement support for transformed long lines
+
+class sac_parser
+{
+  public:
+	using datablock_index = std::map<std::string, std::size_t>;
+
+	sac_parser(std::istream &is, bool init = true);
+
+	virtual ~sac_parser() = default;
+
+	enum CharTraitsMask : uint8_t
+	{
+		kOrdinaryMask = 1 << 0,
+		kNonBlankMask = 1 << 1,
+		kTextLeadMask = 1 << 2,
+		kAnyPrintMask = 1 << 3
+	};
+
+	static bool is_white(int ch)
+	{
+		return std::isspace(ch) or ch == '#';
+	}
+
+	static constexpr bool is_ordinary(int ch)
+	{
+		return ch >= 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kOrdinaryMask) != 0;
+	}
+
+	static constexpr bool is_non_blank(int ch)
+	{
+		return ch > 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kNonBlankMask) != 0;
+	}
+
+	static constexpr bool is_text_lead(int ch)
+	{
+		return ch >= 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kTextLeadMask) != 0;
+	}
+
+	static constexpr bool is_any_print(int ch)
+	{
+		return ch == '\t' or
+		       (ch >= 0x20 and ch <= 0x7f and (kCharTraitsTable[ch - 0x20] & kAnyPrintMask) != 0);
+	}
+
+	static bool is_unquoted_string(std::string_view text)
+	{
+		auto s = text.begin();
+
+		bool result = is_ordinary(*s++);
+		while (result and s != text.end())
+		{
+			result = is_non_blank(*s);
+			++s;
+		}
+
+		// but be careful it does not contain e.g. stop_
+		if (result)
+		{
+			static const std::regex reservedRx(R"((^(?:data|save)|.*(?:loop|stop|global))_.+)", std::regex_constants::icase);
+			result = not std::regex_match(text.begin(), text.end(), reservedRx);
+		}
+
+		return result;
+	}
+
+  protected:
+	static constexpr uint8_t kCharTraitsTable[128] = {
+		//	0	1	2	3	4	5	6	7	8	9	a	b	c	d	e	f
+		14, 15, 14, 14, 14, 15, 15, 14, 15, 15, 15, 15, 15, 15, 15, 15, //	2
+		15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 10, 15, 15, 15, 15, //	3
+		15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, //	4
+		15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 14, 15, 14, 15, 14, //	5
+		15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, //	6
+		15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 0,  //	7
+	};
+
+	enum class CIFToken
+	{
+		Unknown,
+
+		Eof,
+
+		DATA,
+		LOOP,
+		GLOBAL,
+		SAVE,
+		STOP,
+		Tag,
+		Value
+	};
+
+	static constexpr const char *get_token_name(CIFToken token)
+	{
+		switch (token)
+		{
+			case CIFToken::Unknown: return "Unknown";
+			case CIFToken::Eof: return "Eof";
+			case CIFToken::DATA: return "DATA";
+			case CIFToken::LOOP: return "LOOP";
+			case CIFToken::GLOBAL: return "GLOBAL";
+			case CIFToken::SAVE: return "SAVE";
+			case CIFToken::STOP: return "STOP";
+			case CIFToken::Tag: return "Tag";
+			case CIFToken::Value: return "Value";
+			default: return "Invalid token parameter";
+		}
+	}
+
+	enum class CIFValue
+	{
+		Int,
+		Float,
+		Numeric,
+		String,
+		TextField,
+		Inapplicable,
+		Unknown
+	};
+
+	static constexpr const char *get_value_name(CIFValue type)
+	{
+		switch (type)
+		{
+			case CIFValue::Int: return "Int";
+			case CIFValue::Float: return "Float";
+			case CIFValue::Numeric: return "Numeric";
+			case CIFValue::String: return "String";
+			case CIFValue::TextField: return "TextField";
+			case CIFValue::Inapplicable: return "Inapplicable";
+			case CIFValue::Unknown: return "Unknown";
+			default: return "Invalid type parameter";
+		}
+	}
+
+	// get_next_char takes a char from the buffer, or if it is empty
+	// from the istream. This function also does carriage/linefeed
+	// translation.
+	int get_next_char();
+
+	void retract();
+
+	int restart(int start);
+
+	CIFToken get_next_token();
+
+	void match(CIFToken token);
+
+  public:
+	bool parse_single_datablock(const std::string &datablock);
+
+	datablock_index index_datablocks();
+
+	bool parse_single_datablock(const std::string &datablock, const datablock_index &index);
+
+	void parse_file();
+
+  protected:
+	void parse_global();
+
+	void parse_datablock();
+
+	virtual void parse_save_frame();
+
+	void error(const std::string &msg)
+	{
+		if (cif::VERBOSE > 0)
+			std::cerr << "Error parsing mmCIF: " << msg << std::endl;
+
+		throw parse_error(m_line_nr, msg);
+	}
+
+	void warning(const std::string &msg)
+	{
+		if (cif::VERBOSE > 0)
+			std::cerr << "parser warning at line " << m_line_nr << ": " << msg << std::endl;
+	}
+
+	// production methods, these are pure virtual here
+
+	virtual void produce_datablock(const std::string &name) = 0;
+	virtual void produce_category(const std::string &name) = 0;
+	virtual void produce_row() = 0;
+	virtual void produce_item(const std::string &category, const std::string &item, const std::string &value) = 0;
+
+  protected:
+	enum State
+	{
+		Start,
+		White,
+		Esc,
+		Comment,
+		QuestionMark,
+		Dot,
+		QuotedString,
+		QuotedStringQuote,
+		UnquotedString,
+		Tag,
+		TextField,
+		Float = 100,
+		Int = 110,
+		Value = 300,
+		DATA,
+		SAVE
+	};
+
+	std::streambuf &m_source;
+
+	// Parser state
+	bool m_validate;
+	uint32_t m_line_nr;
+	bool m_bol;
+	CIFToken m_lookahead;
+	std::string m_token_value;
+	CIFValue mTokenType;
+	std::vector<int> m_buffer;	// retract buffer, used to be a stack<char>
+};
+
+// --------------------------------------------------------------------
+
+class parser : public sac_parser
+{
+  public:
+	parser(std::istream &is, file &file)
+		: sac_parser(is)
+		, m_file(file)
+	{
+	}
+
+	void produce_datablock(const std::string &name) override;
+
+	void produce_category(const std::string &name) override;
+
+	void produce_row() override;
+
+	void produce_item(const std::string &category, const std::string &item, const std::string &value) override;
+
+  protected:
+	file &m_file;
+	datablock *m_datablock = nullptr;
+	category *m_category = nullptr;
+	row_handle m_row;
+};
+
+} // namespace cif

include/cif++/pdb/cif2pdb.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
+ *
  * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
- * 
+ *
  * 1. Redistributions of source code must retain the above copyright notice, this
  *    list of conditions and the following disclaimer
  * 2. Redistributions in binary form must reproduce the above copyright notice,
  *    this list of conditions and the following disclaimer in the documentation
  *    and/or other materials provided with the distribution.
- * 
+ *
  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -26,14 +26,17 @@
 
 #pragma once
 
-#include "cif++/Cif++.hpp"
+#include <cif++.hpp>
 
-void WritePDBFile(std::ostream& pdbFile, cif::File& cifFile);
+namespace cif::pdb
+{
 
 /// \brief Just the HEADER, COMPND, SOURCE and AUTHOR lines
-void WritePDBHeaderLines(std::ostream& os, cif::File& cifFile);
+void write_header_lines(std::ostream &os, const datablock &data);
 
-std::string GetPDBHEADERLine(cif::File& cifFile, std::string::size_type truncate_at = 127);
-std::string GetPDBCOMPNDLine(cif::File& cifFile, std::string::size_type truncate_at = 127);
-std::string GetPDBSOURCELine(cif::File& cifFile, std::string::size_type truncate_at = 127);
-std::string GetPDBAUTHORLine(cif::File& cifFile, std::string::size_type truncate_at = 127);
+std::string get_HEADER_line(const datablock &data, std::string::size_type truncate_at = 127);
+std::string get_COMPND_line(const datablock &data, std::string::size_type truncate_at = 127);
+std::string get_SOURCE_line(const datablock &data, std::string::size_type truncate_at = 127);
+std::string get_AUTHOR_line(const datablock &data, std::string::size_type truncate_at = 127);
+
+} // namespace pdbx

include/cif++/pdb/io.hpp

@@ -0,0 +1,60 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ * 
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * 
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cif++.hpp>
+
+namespace cif::pdb
+{
+
+/// \brief Read a file in either mmCIF or PDB format, compressed or not,
+/// depending on the content.
+file read(const std::filesystem::path &file);
+
+/// \brief Read a file in either mmCIF or PDB format, compressed or not,
+/// depending on the content.
+file read(std::istream &is);
+
+/// \brief Write out a file in PDB format
+void write(std::ostream &os, const datablock &db);
+
+/// \brief Write out a file in PDB format
+inline void write(std::ostream &os, const file &f)
+{
+	write(os, f.front());
+}
+
+/// \brief Write out a file in PDB format or mmCIF format, depending on the filename extension
+void write(const std::filesystem::path &file, const datablock &db);
+
+/// \brief Write out a file in PDB format or mmCIF format, depending on the filename extension
+inline void write(const std::filesystem::path &p, const file &f)
+{
+	write(p, f.front());
+}
+
+}

include/cif++/pdb/pdb2cif.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
+ *
  * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
- * 
+ *
  * 1. Redistributions of source code must retain the above copyright notice, this
  *    list of conditions and the following disclaimer
  * 2. Redistributions in binary form must reproduce the above copyright notice,
  *    this list of conditions and the following disclaimer in the documentation
  *    and/or other materials provided with the distribution.
- * 
+ *
  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -26,7 +26,10 @@
 
 #pragma once
 
-#include "cif++/Cif++.hpp"
+#include <cif++.hpp>
+
+namespace cif::pdb
+{
 
 // --------------------------------------------------------------------
 
@@ -57,4 +60,6 @@ struct PDBRecord
 
 // --------------------------------------------------------------------
 
-void ReadPDBFile(std::istream &pdbFile, cif::File &cifFile);
+void ReadPDBFile(std::istream &pdbFile, file &cifFile);
+
+} // namespace pdbx

include/cif++/pdb/pdb2cif_remark_3.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
+ *
  * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
- * 
+ *
  * 1. Redistributions of source code must retain the above copyright notice, this
  *    list of conditions and the following disclaimer
  * 2. Redistributions in binary form must reproduce the above copyright notice,
  *    this list of conditions and the following disclaimer in the documentation
  *    and/or other materials provided with the distribution.
- * 
+ *
  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -26,10 +26,13 @@
 
 #pragma once
 
-#include "cif++/PDB2Cif.hpp"
+#include <cif++/pdb/pdb2cif.hpp>
 
 // --------------------------------------------------------------------
 
+namespace cif::pdb
+{
+
 struct TemplateLine;
 
 class Remark3Parser
@@ -37,37 +40,36 @@ class Remark3Parser
   public:
 	virtual ~Remark3Parser() {}
 
-	static bool parse(const std::string& expMethod, PDBRecord* r, cif::Datablock& db);
+	static bool parse(const std::string &expMethod, PDBRecord *r, cif::datablock &db);
 
 	virtual std::string program();
 	virtual std::string version();
 
   protected:
-
-	Remark3Parser(const std::string& name, const std::string& expMethod, PDBRecord* r, cif::Datablock& db,
-			const TemplateLine templatelines[], uint32_t templateLineCount, std::regex programVersion);
+	Remark3Parser(const std::string &name, const std::string &expMethod, PDBRecord *r, cif::datablock &db,
+		const TemplateLine templatelines[], uint32_t templateLineCount, std::regex programVersion);
 
 	virtual float parse();
 	std::string nextLine();
 
-	bool match(const char* expr, int nextState);
-	void storeCapture(const char* category, std::initializer_list<const char*> items, bool createNew = false);
-	void storeRefineLsRestr(const char* type, std::initializer_list<const char*> values);
-	void updateRefineLsRestr(const char* type, std::initializer_list<const char*> values);
+	bool match(const char *expr, int nextState);
+	void storeCapture(const char *category, std::initializer_list<const char *> items, bool createNew = false);
+	void storeRefineLsRestr(const char *type, std::initializer_list<const char *> values);
+	void updateRefineLsRestr(const char *type, std::initializer_list<const char *> values);
 
 	virtual void fixup() {}
 
-	std::string		mName;
-	std::string		mExpMethod;
-	PDBRecord*		mRec;
-	cif::Datablock	mDb;
-	std::string		mLine;
-	std::smatch		mM;
-	uint32_t			mState;
+	std::string mName;
+	std::string mExpMethod;
+	PDBRecord *mRec;
+	cif::datablock mDb;
+	std::string mLine;
+	std::smatch mM;
+	uint32_t mState;
 
-	const TemplateLine*	mTemplate;
-	uint32_t				mTemplateCount;
-	std::regex			mProgramVersion;
+	const TemplateLine *mTemplate;
+	uint32_t mTemplateCount;
+	std::regex mProgramVersion;
 };
 
-
+} // namespace pdbx

include/cif++/pdb/tls.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
+ *
  * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
- * 
+ *
  * 1. Redistributions of source code must retain the above copyright notice, this
  *    list of conditions and the following disclaimer
  * 2. Redistributions in binary form must reproduce the above copyright notice,
  *    this list of conditions and the following disclaimer in the documentation
  *    and/or other materials provided with the distribution.
- * 
+ *
  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -26,32 +26,30 @@
 
 #pragma once
 
-#include <vector>
 #include <string>
 #include <tuple>
+#include <vector>
 
-#include "cif++/Cif++.hpp"
+#include <cif++.hpp>
 
 namespace cif
 {
-	
+
 extern const int
 	kResidueNrWildcard,
 	kNoSeqNum;
 
-struct TLSSelection;
-typedef std::unique_ptr<TLSSelection> TLSSelectionPtr;
+struct tls_selection;
+struct tls_residue;
 
-struct TLSResidue;
-
-struct TLSSelection
+struct tls_selection
 {
-	virtual ~TLSSelection() {}
-	virtual void CollectResidues(cif::Datablock& db, std::vector<TLSResidue>& residues, std::size_t indentLevel = 0) const = 0;
-	std::vector<std::tuple<std::string,int,int>> GetRanges(cif::Datablock& db, bool pdbNamespace) const;
+	virtual ~tls_selection() {}
+	virtual void collect_residues(cif::datablock &db, std::vector<tls_residue> &residues, std::size_t indentLevel = 0) const = 0;
+	std::vector<std::tuple<std::string, int, int>> get_ranges(cif::datablock &db, bool pdbNamespace) const;
 };
 
 // Low level: get the selections
-TLSSelectionPtr ParseSelectionDetails(const std::string& program, const std::string& selection);
+std::unique_ptr<tls_selection> parse_tls_selection_details(const std::string &program, const std::string &selection);
 
-}
+} // namespace cif

include/cif++/point.hpp

@@ -0,0 +1,736 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cmath>
+#include <complex>
+#include <functional>
+#include <valarray>
+
+#if __has_include(<clipper/core/coords.h>)
+#define HAVE_LIBCLIPPER 1
+#include <clipper/core/coords.h>
+#endif
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+const double
+	kPI = 3.141592653589793238462643383279502884;
+
+// --------------------------------------------------------------------
+// A stripped down quaternion implementation, based on boost::math::quaternion
+// We use quaternions to do rotations in 3d space
+
+template <typename T>
+class quaternion_type
+{
+  public:
+	using value_type = T;
+
+	constexpr explicit quaternion_type(value_type const &value_a = value_type(), value_type const &value_b = value_type(), value_type const &value_c = value_type(), value_type const &value_d = value_type())
+		: a(value_a)
+		, b(value_b)
+		, c(value_c)
+		, d(value_d)
+	{
+	}
+
+	constexpr explicit quaternion_type(std::complex<value_type> const &z0, std::complex<value_type> const &z1 = std::complex<value_type>())
+		: a(z0.real())
+		, b(z0.imag())
+		, c(z1.real())
+		, d(z1.imag())
+	{
+	}
+
+	constexpr quaternion_type(quaternion_type const &) = default;
+	constexpr quaternion_type(quaternion_type &&) = default;
+
+	template <typename X>
+	constexpr explicit quaternion_type(quaternion_type<X> const &rhs)
+		: a(static_cast<value_type>(rhs.a))
+		, b(static_cast<value_type>(rhs.b))
+		, c(static_cast<value_type>(rhs.c))
+		, d(static_cast<value_type>(rhs.d))
+	{
+	}
+
+	// accessors
+	//
+	// Note:    Like complex number, quaternions do have a meaningful notion of "real part",
+	//            but unlike them there is no meaningful notion of "imaginary part".
+	//            Instead there is an "unreal part" which itself is a quaternion, and usually
+	//            nothing simpler (as opposed to the complex number case).
+	//            However, for practicality, there are accessors for the other components
+	//            (these are necessary for the templated copy constructor, for instance).
+
+	constexpr value_type real() const
+	{
+		return a;
+	}
+
+	constexpr quaternion_type unreal() const
+	{
+		return { 0, b, c, d };
+	}
+
+	constexpr void swap(quaternion_type &o)
+	{
+		std::swap(a, o.a);
+		std::swap(b, o.b);
+		std::swap(c, o.c);
+		std::swap(d, o.d);
+	}
+
+	// assignment operators
+
+	template <typename X>
+	constexpr quaternion_type &operator=(quaternion_type<X> const &rhs)
+	{
+		a = static_cast<value_type>(rhs.a);
+		b = static_cast<value_type>(rhs.b);
+		c = static_cast<value_type>(rhs.c);
+		d = static_cast<value_type>(rhs.d);
+
+		return *this;
+	}
+
+	constexpr quaternion_type &operator=(quaternion_type const &rhs)
+	{
+		a = rhs.a;
+		b = rhs.b;
+		c = rhs.c;
+		d = rhs.d;
+
+		return *this;
+	}
+
+	constexpr quaternion_type &operator=(value_type const &rhs)
+	{
+		a = rhs;
+
+		b = c = d = static_cast<value_type>(0);
+
+		return *this;
+	}
+
+	constexpr quaternion_type &operator=(std::complex<value_type> const &rhs)
+	{
+		a = rhs.real();
+		b = rhs.imag();
+
+		c = d = static_cast<value_type>(0);
+
+		return *this;
+	}
+
+	// other assignment-related operators
+	//
+	// NOTE:    Quaternion multiplication is *NOT* commutative;
+	//            symbolically, "q *= rhs;" means "q = q * rhs;"
+	//            and "q /= rhs;" means "q = q * inverse_of(rhs);"
+
+	constexpr quaternion_type &operator+=(value_type const &rhs)
+	{
+		a += rhs;
+		return *this;
+	}
+
+	constexpr quaternion_type &operator+=(std::complex<value_type> const &rhs)
+	{
+		a += std::real(rhs);
+		b += std::imag(rhs);
+		return *this;
+	}
+
+	template <class X>
+	constexpr quaternion_type &operator+=(quaternion_type<X> const &rhs)
+	{
+		a += rhs.a;
+		b += rhs.b;
+		c += rhs.c;
+		d += rhs.d;
+		return *this;
+	}
+
+	constexpr quaternion_type &operator-=(value_type const &rhs)
+	{
+		a -= rhs;
+		return *this;
+	}
+
+	constexpr quaternion_type &operator-=(std::complex<value_type> const &rhs)
+	{
+		a -= std::real(rhs);
+		b -= std::imag(rhs);
+		return *this;
+	}
+
+	template <class X>
+	constexpr quaternion_type &operator-=(quaternion_type<X> const &rhs)
+	{
+		a -= rhs.a;
+		b -= rhs.b;
+		c -= rhs.c;
+		d -= rhs.d;
+		return *this;
+	}
+
+	constexpr quaternion_type &operator*=(value_type const &rhs)
+	{
+		a *= rhs;
+		b *= rhs;
+		c *= rhs;
+		d *= rhs;
+		return *this;
+	}
+
+	constexpr quaternion_type &operator*=(std::complex<value_type> const &rhs)
+	{
+		value_type ar = rhs.real();
+		value_type br = rhs.imag();
+		quaternion_type result(a * ar - b * br, a * br + b * ar, c * ar + d * br, -c * br + d * ar);
+		swap(result);
+		return *this;
+	}
+
+	constexpr friend quaternion_type operator*(const quaternion_type &a, const quaternion_type &b)
+	{
+		auto result = a;
+		result *= b;
+		return result;
+	}
+
+	template <typename X>
+	constexpr quaternion_type &operator*=(quaternion_type<X> const &rhs)
+	{
+		value_type ar = static_cast<value_type>(rhs.a);
+		value_type br = static_cast<value_type>(rhs.b);
+		value_type cr = static_cast<value_type>(rhs.c);
+		value_type dr = static_cast<value_type>(rhs.d);
+
+		quaternion_type result(a * ar - b * br - c * cr - d * dr, a * br + b * ar + c * dr - d * cr, a * cr - b * dr + c * ar + d * br, a * dr + b * cr - c * br + d * ar);
+		swap(result);
+		return *this;
+	}
+
+	constexpr quaternion_type &operator/=(value_type const &rhs)
+	{
+		a /= rhs;
+		b /= rhs;
+		c /= rhs;
+		d /= rhs;
+		return *this;
+	}
+
+	constexpr quaternion_type &operator/=(std::complex<value_type> const &rhs)
+	{
+		value_type ar = rhs.real();
+		value_type br = rhs.imag();
+		value_type denominator = ar * ar + br * br;
+		quaternion_type result((+a * ar + b * br) / denominator, (-a * br + b * ar) / denominator, (+c * ar - d * br) / denominator, (+c * br + d * ar) / denominator);
+		swap(result);
+		return *this;
+	}
+
+	template <typename X>
+	constexpr quaternion_type &operator/=(quaternion_type<X> const &rhs)
+	{
+		value_type ar = static_cast<value_type>(rhs.a);
+		value_type br = static_cast<value_type>(rhs.b);
+		value_type cr = static_cast<value_type>(rhs.c);
+		value_type dr = static_cast<value_type>(rhs.d);
+
+		value_type denominator = ar * ar + br * br + cr * cr + dr * dr;
+		quaternion_type result((+a * ar + b * br + c * cr + d * dr) / denominator, (-a * br + b * ar - c * dr + d * cr) / denominator, (-a * cr + b * dr + c * ar - d * br) / denominator, (-a * dr - b * cr + c * br + d * ar) / denominator);
+		swap(result);
+		return *this;
+	}
+
+	constexpr friend quaternion_type normalize(quaternion_type q)
+	{
+		std::valarray<value_type> t(4);
+
+		t[0] = q.a;
+		t[1] = q.b;
+		t[2] = q.c;
+		t[3] = q.d;
+
+		t *= t;
+
+		value_type length = std::sqrt(t.sum());
+
+		if (length > 0.001)
+			q /= static_cast<value_type>(length);
+		else
+			q = quaternion_type(1, 0, 0, 0);
+
+		return q;
+	}
+
+	constexpr friend quaternion_type conj(quaternion_type q)
+	{
+		return quaternion_type{ +q.a, -q.b, -q.c, -q.d };
+	}
+
+	constexpr value_type get_a() const { return a; }
+	constexpr value_type get_b() const { return b; }
+	constexpr value_type get_c() const { return c; }
+	constexpr value_type get_d() const { return d; }
+
+  private:
+	value_type a, b, c, d;
+};
+
+template <typename T>
+inline quaternion_type<T> spherical(T const &rho, T const &theta, T const &phi1, T const &phi2)
+{
+	T cos_phi1 = std::cos(phi1);
+	T cos_phi2 = std::cos(phi2);
+
+	T a = std::cos(theta) * cos_phi1 * cos_phi2;
+	T b = std::sin(theta) * cos_phi1 * cos_phi2;
+	T c = std::sin(phi1) * cos_phi2;
+	T d = std::sin(phi2);
+
+	quaternion_type result(a, b, c, d);
+	result *= rho;
+
+	return result;
+}
+
+using quaternion = quaternion_type<float>;
+
+// --------------------------------------------------------------------
+
+//	point, a location with x, y and z coordinates as floating point.
+//	This one is derived from a tuple<float,float,float> so
+//	you can do things like:
+//
+//	float x, y, z;
+//	tie(x, y, z) = atom.loc();
+
+template <typename F>
+struct point_type
+{
+	using value_type = F;
+
+	value_type m_x, m_y, m_z;
+
+	constexpr point_type()
+		: m_x(0)
+		, m_y(0)
+		, m_z(0)
+	{
+	}
+
+	constexpr point_type(value_type x, value_type y, value_type z)
+		: m_x(x)
+		, m_y(y)
+		, m_z(z)
+	{
+	}
+
+	template <typename PF>
+	constexpr point_type(const point_type<PF> &pt)
+		: m_x(static_cast<F>(pt.m_x))
+		, m_y(static_cast<F>(pt.m_y))
+		, m_z(static_cast<F>(pt.m_z))
+	{
+	}
+
+	constexpr point_type(const std::tuple<value_type, value_type, value_type> &pt)
+		: point_type(std::get<0>(pt), std::get<1>(pt), std::get<2>(pt))
+	{
+	}
+
+#if HAVE_LIBCLIPPER
+	constexpr point_type(const clipper::Coord_orth &pt)
+		: m_x(pt[0])
+		, m_y(pt[1])
+		, m_z(pt[2])
+	{
+	}
+
+	constexpr point_type &operator=(const clipper::Coord_orth &rhs)
+	{
+		m_x = rhs[0];
+		m_y = rhs[1];
+		m_z = rhs[2];
+		return *this;
+	}
+#endif
+
+	template <typename PF>
+	constexpr point_type &operator=(const point_type<PF> &rhs)
+	{
+		m_x = static_cast<F>(rhs.m_x);
+		m_y = static_cast<F>(rhs.m_y);
+		m_z = static_cast<F>(rhs.m_z);
+		return *this;
+	}
+
+	constexpr value_type &get_x() { return m_x; }
+	constexpr value_type get_x() const { return m_x; }
+	constexpr void set_x(value_type x) { m_x = x; }
+
+	constexpr value_type &get_y() { return m_y; }
+	constexpr value_type get_y() const { return m_y; }
+	constexpr void set_y(value_type y) { m_y = y; }
+
+	constexpr value_type &get_z() { return m_z; }
+	constexpr value_type get_z() const { return m_z; }
+	constexpr void set_z(value_type z) { m_z = z; }
+
+	constexpr point_type &operator+=(const point_type &rhs)
+	{
+		m_x += rhs.m_x;
+		m_y += rhs.m_y;
+		m_z += rhs.m_z;
+
+		return *this;
+	}
+
+	constexpr point_type &operator+=(value_type d)
+	{
+		m_x += d;
+		m_y += d;
+		m_z += d;
+
+		return *this;
+	}
+
+	constexpr point_type &operator-=(const point_type &rhs)
+	{
+		m_x -= rhs.m_x;
+		m_y -= rhs.m_y;
+		m_z -= rhs.m_z;
+
+		return *this;
+	}
+
+	constexpr point_type &operator-=(value_type d)
+	{
+		m_x -= d;
+		m_y -= d;
+		m_z -= d;
+
+		return *this;
+	}
+
+	constexpr point_type &operator*=(value_type rhs)
+	{
+		m_x *= rhs;
+		m_y *= rhs;
+		m_z *= rhs;
+		return *this;
+	}
+
+	constexpr point_type &operator/=(value_type rhs)
+	{
+		m_x /= rhs;
+		m_y /= rhs;
+		m_z /= rhs;
+		return *this;
+	}
+
+	constexpr value_type normalize()
+	{
+		auto length = m_x * m_x + m_y * m_y + m_z * m_z;
+		if (length > 0)
+		{
+			length = std::sqrt(length);
+			operator/=(length);
+		}
+		return length;
+	}
+
+	constexpr void rotate(const quaternion &q)
+	{
+		quaternion_type<value_type> p(0, m_x, m_y, m_z);
+
+		p = q * p * conj(q);
+
+		m_x = p.get_b();
+		m_y = p.get_c();
+		m_z = p.get_d();
+	}
+
+#if HAVE_LIBCLIPPER
+	operator clipper::Coord_orth() const
+	{
+		return clipper::Coord_orth(m_x, m_y, m_z);
+	}
+#endif
+
+	constexpr operator std::tuple<const value_type &, const value_type &, const value_type &>() const
+	{
+		return std::make_tuple(std::ref(m_x), std::ref(m_y), std::ref(m_z));
+	}
+
+	constexpr operator std::tuple<value_type &, value_type &, value_type &>()
+	{
+		return std::make_tuple(std::ref(m_x), std::ref(m_y), std::ref(m_z));
+	}
+
+	constexpr bool operator==(const point_type &rhs) const
+	{
+		return m_x == rhs.m_x and m_y == rhs.m_y and m_z == rhs.m_z;
+	}
+
+	// consider point as a vector... perhaps I should rename point?
+	constexpr value_type length_sq() const
+	{
+		return m_x * m_x + m_y * m_y + m_z * m_z;
+	}
+
+	constexpr value_type length() const
+	{
+		return std::sqrt(m_x * m_x + m_y * m_y + m_z * m_z);
+	}
+};
+
+using point = point_type<float>;
+
+template <typename F>
+inline constexpr std::ostream &operator<<(std::ostream &os, const point_type<F> &pt)
+{
+	os << '(' << pt.m_x << ',' << pt.m_y << ',' << pt.m_z << ')';
+	return os;
+}
+
+template <typename F>
+inline constexpr point_type<F> operator+(const point_type<F> &lhs, const point_type<F> &rhs)
+{
+	return point_type<F>(lhs.m_x + rhs.m_x, lhs.m_y + rhs.m_y, lhs.m_z + rhs.m_z);
+}
+
+template <typename F>
+inline constexpr point_type<F> operator-(const point_type<F> &lhs, const point_type<F> &rhs)
+{
+	return point_type<F>(lhs.m_x - rhs.m_x, lhs.m_y - rhs.m_y, lhs.m_z - rhs.m_z);
+}
+
+template <typename F>
+inline constexpr point_type<F> operator-(const point_type<F> &pt)
+{
+	return point_type<F>(-pt.m_x, -pt.m_y, -pt.m_z);
+}
+
+template <typename F>
+inline constexpr point_type<F> operator*(const point_type<F> &pt, F f)
+{
+	return point_type<F>(pt.m_x * f, pt.m_y * f, pt.m_z * f);
+}
+
+template <typename F>
+inline constexpr point_type<F> operator*(F f, const point_type<F> &pt)
+{
+	return point_type<F>(pt.m_x * f, pt.m_y * f, pt.m_z * f);
+}
+
+template <typename F>
+inline constexpr point_type<F> operator/(const point_type<F> &pt, F f)
+{
+	return point_type<F>(pt.m_x / f, pt.m_y / f, pt.m_z / f);
+}
+
+// --------------------------------------------------------------------
+// several standard 3d operations
+
+template <typename F>
+inline constexpr auto distance_squared(const point_type<F> &a, const point_type<F> &b)
+{
+	return (a.m_x - b.m_x) * (a.m_x - b.m_x) +
+	       (a.m_y - b.m_y) * (a.m_y - b.m_y) +
+	       (a.m_z - b.m_z) * (a.m_z - b.m_z);
+}
+
+template <typename F>
+inline constexpr auto distance(const point_type<F> &a, const point_type<F> &b)
+{
+	return std::sqrt(
+		(a.m_x - b.m_x) * (a.m_x - b.m_x) +
+		(a.m_y - b.m_y) * (a.m_y - b.m_y) +
+		(a.m_z - b.m_z) * (a.m_z - b.m_z));
+}
+
+template <typename F>
+inline constexpr auto dot_product(const point_type<F> &a, const point_type<F> &b)
+{
+	return a.m_x * b.m_x + a.m_y * b.m_y + a.m_z * b.m_z;
+}
+
+template <typename F>
+inline constexpr point_type<F> cross_product(const point_type<F> &a, const point_type<F> &b)
+{
+	return point_type<F>(a.m_y * b.m_z - b.m_y * a.m_z,
+		a.m_z * b.m_x - b.m_z * a.m_x,
+		a.m_x * b.m_y - b.m_x * a.m_y);
+}
+
+template <typename F>
+constexpr auto angle(const point_type<F> &p1, const point_type<F> &p2, const point_type<F> &p3)
+{
+	point_type<F> v1 = p1 - p2;
+	point_type<F> v2 = p3 - p2;
+
+	return std::acos(dot_product(v1, v2) / (v1.length() * v2.length())) * 180 / kPI;
+}
+
+template <typename F>
+constexpr auto dihedral_angle(const point_type<F> &p1, const point_type<F> &p2, const point_type<F> &p3, const point_type<F> &p4)
+{
+	point_type<F> v12 = p1 - p2; // vector from p2 to p1
+	point_type<F> v43 = p4 - p3; // vector from p3 to p4
+
+	point_type<F> z = p2 - p3; // vector from p3 to p2
+
+	point_type<F> p = cross_product(z, v12);
+	point_type<F> x = cross_product(z, v43);
+	point_type<F> y = cross_product(z, x);
+
+	auto u = dot_product(x, x);
+	auto v = dot_product(y, y);
+
+	F result = 360;
+	if (u > 0 and v > 0)
+	{
+		u = dot_product(p, x) / std::sqrt(u);
+		v = dot_product(p, y) / std::sqrt(v);
+		if (u != 0 or v != 0)
+			result = std::atan2(v, u) * static_cast<F>(180 / kPI);
+	}
+
+	return result;
+}
+
+template <typename F>
+constexpr auto cosinus_angle(const point_type<F> &p1, const point_type<F> &p2, const point_type<F> &p3, const point_type<F> &p4)
+{
+	point_type<F> v12 = p1 - p2;
+	point_type<F> v34 = p3 - p4;
+
+	auto x = dot_product(v12, v12) * dot_product(v34, v34);
+
+	return x > 0 ? dot_product(v12, v34) / std::sqrt(x) : 0;
+}
+
+template <typename F>
+constexpr auto distance_point_to_line(const point_type<F> &l1, const point_type<F> &l2, const point_type<F> &p)
+{
+	auto line = l2 - l1;
+	auto p_to_l1 = p - l1;
+	auto p_to_l2 = p - l2;
+	auto cross = cross_product(p_to_l1, p_to_l2);
+	return cross.length() / line.length();
+}
+
+// --------------------------------------------------------------------
+// For e.g. simulated annealing, returns a new point that is moved in
+// a random direction with a distance randomly chosen from a normal
+// distribution with a stddev of offset.
+
+point nudge(point p, float offset);
+
+// --------------------------------------------------------------------
+
+quaternion construct_from_angle_axis(float angle, point axis);
+std::tuple<double, point> quaternion_to_angle_axis(quaternion q);
+
+point centroid(const std::vector<point> &Points);
+point center_points(std::vector<point> &Points);
+
+/// \brief Returns how the two sets of points \a a and \b b can be aligned
+///
+/// \param a	The first set of points
+/// \param b    The second set of points
+/// \result     The quaternion which should be applied to the points in \a a to
+///             obtain the best superposition.
+quaternion align_points(const std::vector<point> &a, const std::vector<point> &b);
+
+/// \brief The RMSd for the points in \a a and \a b
+double RMSd(const std::vector<point> &a, const std::vector<point> &b);
+
+// --------------------------------------------------------------------
+// Helper class to generate evenly divided points on a sphere
+// we use a fibonacci sphere to calculate even distribution of the dots
+
+template <int N>
+class spherical_dots
+{
+  public:
+
+	constexpr static int P = 2 * N * 1;
+
+	using array_type = typename std::array<point, P>;
+	using iterator = typename array_type::const_iterator;
+
+	static spherical_dots &instance()
+	{
+		static spherical_dots sInstance;
+		return sInstance;
+	}
+
+	size_t size() const { return m_points.size(); }
+	const point operator[](uint32_t inIx) const { return m_points[inIx]; }
+	iterator begin() const { return m_points.begin(); }
+	iterator end() const { return m_points.end(); }
+
+	double weight() const { return m_weight; }
+
+	spherical_dots()
+	{
+		const double
+			kGoldenRatio = (1 + std::sqrt(5.0)) / 2;
+
+		m_weight = (4 * kPI) / P;
+
+		auto p = m_points.begin();
+
+		for (int32_t i = -N; i <= N; ++i)
+		{
+			double lat = std::asin((2.0 * i) / P);
+			double lon = std::fmod(i, kGoldenRatio) * 2 * kPI / kGoldenRatio;
+
+			p->m_x = std::sin(lon) * std::cos(lat);
+			p->m_y = std::cos(lon) * std::cos(lat);
+			p->m_z = std::sin(lat);
+
+			++p;
+		}
+	}
+
+  private:
+	array_type m_points;
+	double m_weight;
+};
+
+} // namespace cif

include/cif++/row.hpp

@@ -0,0 +1,305 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <cif++/item.hpp>
+
+namespace cif
+{
+
+namespace detail
+{
+
+	// some helper classes to help create tuple result types
+	template <typename... C>
+	struct get_row_result
+	{
+		static constexpr size_t N = sizeof...(C);
+
+		get_row_result(const row_handle &r, std::array<uint16_t, N> &&columns)
+			: m_row(r)
+			, m_columns(std::move(columns))
+		{
+		}
+
+		const item_handle operator[](uint16_t ix) const
+		{
+			return m_row[m_columns[ix]];
+		}
+
+		template <typename... Ts, std::enable_if_t<N == sizeof...(Ts), int> = 0>
+		operator std::tuple<Ts...>() const
+		{
+			return get<Ts...>(std::index_sequence_for<Ts...>{});
+		}
+
+		template <typename... Ts, size_t... Is>
+		std::tuple<Ts...> get(std::index_sequence<Is...>) const
+		{
+			return std::tuple<Ts...>{ m_row[m_columns[Is]].template as<Ts>()... };
+		}
+
+		const row_handle &m_row;
+		std::array<uint16_t, N> m_columns;
+	};
+
+	// we want to be able to tie some variables to a get_row_result, for this we use tiewraps
+	template <typename... Ts>
+	struct tie_wrap
+	{
+		tie_wrap(Ts... args)
+			: m_value(args...)
+		{
+		}
+
+		template <typename RR>
+		void operator=(const RR &&rr)
+		{
+			// get_row_result will do the conversion, but only if the types
+			// are compatible. That means the number of parameters to the get()
+			// of the row should be equal to the number of items in the tuple
+			// you are trying to tie.
+
+			using RType = std::tuple<typename std::remove_reference<Ts>::type...>;
+
+			m_value = static_cast<RType>(rr);
+		}
+
+		std::tuple<Ts...> m_value;
+	};
+
+} // namespace detail
+
+template <typename... Ts>
+auto tie(Ts &...v)
+{
+	return detail::tie_wrap<Ts &...>(std::forward<Ts &>(v)...);
+}
+
+// --------------------------------------------------------------------
+/// \brief the row class, this one is not directly accessible from the outside
+
+class row : public std::vector<item_value>
+{
+  public:
+	row() = default;
+
+	item_value* get(uint16_t ix)
+	{
+		return ix < size() ? &at(ix) : nullptr;
+	}
+
+	const item_value* get(uint16_t ix) const
+	{
+		return ix < size() ? &at(ix) : nullptr;
+	}
+
+  private:
+	friend class category;
+	friend class category_index;
+
+	template <typename, typename...>
+	friend class iterator_impl;
+
+	void append(uint16_t ix, item_value &&iv)
+	{
+		if (ix >= size())
+			resize(ix + 1);
+		
+		at(ix) = std::move(iv);
+	}
+
+	void remove(uint16_t ix)
+	{
+		if (ix < size())
+			at(ix) = item_value{};
+	}
+
+	row *m_next = nullptr;
+};
+
+// --------------------------------------------------------------------
+/// \brief row_handle is the way to access data stored in rows
+
+class row_handle
+{
+  public:
+	friend struct item_handle;
+	friend class category;
+	friend class category_index;
+	friend class row_initializer;
+
+	row_handle() = default;
+
+	row_handle(const row_handle &) = default;
+	row_handle(row_handle &&) = default;
+
+	row_handle &operator=(const row_handle &) = default;
+	row_handle &operator=(row_handle &&) = default;
+
+	row_handle(const category &cat, const row &r)
+		: m_category(const_cast<category *>(&cat))
+		, m_row(const_cast<row *>(&r))
+	{
+	}
+
+	const category &get_category() const
+	{
+		return *m_category;
+	}
+
+	bool empty() const
+	{
+		return m_category == nullptr or m_row == nullptr;
+	}
+
+	explicit operator bool() const
+	{
+		return not empty();
+	}
+
+	item_handle operator[](uint16_t column_ix)
+	{
+		return empty() ? item_handle::s_null_item : item_handle(column_ix, *this);
+	}
+
+	const item_handle operator[](uint16_t column_ix) const
+	{
+		return empty() ? item_handle::s_null_item : item_handle(column_ix, const_cast<row_handle &>(*this));
+	}
+
+	item_handle operator[](std::string_view column_name)
+	{
+		return empty() ? item_handle::s_null_item : item_handle(add_column(column_name), *this);
+	}
+
+	const item_handle operator[](std::string_view column_name) const
+	{
+		return empty() ? item_handle::s_null_item : item_handle(get_column_ix(column_name), const_cast<row_handle &>(*this));
+	}
+
+	template <typename... C>
+	auto get(C... columns) const
+	{
+		return detail::get_row_result<C...>(*this, { get_column_ix(columns)... });
+	}
+
+	template <typename... Ts, typename... C, std::enable_if_t<sizeof...(Ts) == sizeof...(C), int> = 0>
+	std::tuple<Ts...> get(C... columns) const
+	{
+		return detail::get_row_result<Ts...>(*this, { get_column_ix(columns)... });
+	}
+
+	template <typename T>
+	T get(const char *column)
+	{
+		return operator[](get_column_ix(column)).template as<T>();
+	}
+
+	void assign(const std::vector<item> &values)
+	{
+		for (auto &value : values)
+			assign(value, true);
+	}
+
+	void assign(std::string_view name, std::string_view value, bool updateLinked, bool validate = true)
+	{
+		assign(add_column(name), value, updateLinked, validate);
+	}
+
+	void assign(uint16_t column, std::string_view value, bool updateLinked, bool validate = true);
+
+	bool operator==(const row_handle &rhs) const { return m_category == rhs.m_category and m_row == rhs.m_row; }
+	bool operator!=(const row_handle &rhs) const { return m_category != rhs.m_category or m_row != rhs.m_row; }
+
+  private:
+	uint16_t get_column_ix(std::string_view name) const;
+	std::string_view get_column_name(uint16_t ix) const;
+
+	uint16_t add_column(std::string_view name);
+
+	row *get_row()
+	{
+		return m_row;
+	}
+
+	const row *get_row() const
+	{
+		return m_row;
+	}
+
+	void assign(const item &i, bool updateLinked)
+	{
+		assign(i.name(), i.value(), updateLinked);
+	}
+
+	void swap(uint16_t column, row_handle &r);
+
+	category *m_category = nullptr;
+	row *m_row = nullptr;
+};
+
+// --------------------------------------------------------------------
+
+class row_initializer : public std::vector<item>
+{
+  public:
+	friend class category;
+
+	row_initializer() = default;
+	row_initializer(const row_initializer &) = default;
+	row_initializer(row_initializer &&) = default;
+	row_initializer &operator=(const row_initializer &) = default;
+	row_initializer &operator=(row_initializer &&) = default;
+
+	row_initializer(std::initializer_list<item> items)
+		: std::vector<item>(items)
+	{
+	}
+
+	template <typename ItemIter, std::enable_if_t<std::is_same_v<typename ItemIter::value_type, item>, int> = 0>
+	row_initializer(ItemIter b, ItemIter e)
+		: std::vector<item>(b, e)
+	{
+	}
+
+	row_initializer(row_handle rh);
+
+	void set_value(std::string_view name, std::string_view value);
+	void set_value(const item &i)
+	{
+		set_value(i.name(), i.value());
+	}
+
+	void set_value_if_empty(std::string_view name, std::string_view value);
+	void set_value_if_empty(const item &i)
+	{
+		set_value_if_empty(i.name(), i.value());
+	}
+};
+
+} // namespace cif

include/cif++/symmetry.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
+ *
  * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- * 
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
- * 
+ *
  * 1. Redistributions of source code must retain the above copyright notice, this
  *    list of conditions and the following disclaimer
  * 2. Redistributions in binary form must reproduce the above copyright notice,
  *    this list of conditions and the following disclaimer in the documentation
  *    and/or other materials provided with the distribution.
- * 
+ *
  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
@@ -26,57 +26,67 @@
 
 #pragma once
 
-#include <string>
-#include <cstdint>
 #include <array>
+#include <cstdint>
+#include <string>
 
-#include "CifUtils.hpp"
-
-namespace mmcif
+namespace cif
 {
 
 // --------------------------------------------------------------------
 
-struct Spacegroup
+enum class space_group_name
 {
-	const char* name;
-	const char* xHM;
-	const char* Hall;
+	full,
+	xHM,
+	Hall
+};
+
+struct space_group
+{
+	const char *name;
+	const char *xHM;
+	const char *Hall;
 	int nr;
 };
 
-CIFPP_EXPORT extern const Spacegroup kSpaceGroups[];
-CIFPP_EXPORT extern const std::size_t kNrOfSpaceGroups;
+extern const space_group kSpaceGroups[];
+extern const std::size_t kNrOfSpaceGroups;
 
 // --------------------------------------------------------------------
 
-struct SymopData
+struct symop_data
 {
-	constexpr SymopData(const std::array<int,15>& data)
-		: m_packed((data[ 0] & 0x03ULL) << 34 bitor
-				   (data[ 1] & 0x03ULL) << 32 bitor
-				   (data[ 2] & 0x03ULL) << 30 bitor
-				   (data[ 3] & 0x03ULL) << 28 bitor
-				   (data[ 4] & 0x03ULL) << 26 bitor
-				   (data[ 5] & 0x03ULL) << 24 bitor
-				   (data[ 6] & 0x03ULL) << 22 bitor
-				   (data[ 7] & 0x03ULL) << 20 bitor
-				   (data[ 8] & 0x03ULL) << 18 bitor
-				   (data[ 9] & 0x07ULL) << 15 bitor
+	constexpr symop_data(const std::array<int, 15> &data)
+		: m_packed((data[0] & 0x03ULL) << 34 bitor
+				   (data[1] & 0x03ULL) << 32 bitor
+				   (data[2] & 0x03ULL) << 30 bitor
+				   (data[3] & 0x03ULL) << 28 bitor
+				   (data[4] & 0x03ULL) << 26 bitor
+				   (data[5] & 0x03ULL) << 24 bitor
+				   (data[6] & 0x03ULL) << 22 bitor
+				   (data[7] & 0x03ULL) << 20 bitor
+				   (data[8] & 0x03ULL) << 18 bitor
+				   (data[9] & 0x07ULL) << 15 bitor
 				   (data[10] & 0x07ULL) << 12 bitor
-				   (data[11] & 0x07ULL) <<  9 bitor
-				   (data[12] & 0x07ULL) <<  6 bitor
-				   (data[13] & 0x07ULL) <<  3 bitor
-				   (data[14] & 0x07ULL) <<  0)
+				   (data[11] & 0x07ULL) << 9 bitor
+				   (data[12] & 0x07ULL) << 6 bitor
+				   (data[13] & 0x07ULL) << 3 bitor
+				   (data[14] & 0x07ULL) << 0)
 	{
 	}
 
-	bool operator==(const SymopData& rhs) const
+	bool operator==(const symop_data &rhs) const
 	{
 		return m_packed == rhs.m_packed;
 	}
 
-	std::array<int,15> data() const
+	bool operator<(const symop_data &rhs) const
+	{
+		return m_packed < rhs.m_packed;
+	}
+
+	std::array<int, 15> data() const
 	{
 		return {
 			static_cast<int>(m_packed >> 34) bitand 0x03,
@@ -90,49 +100,51 @@ struct SymopData
 			static_cast<int>(m_packed >> 18) bitand 0x03,
 			static_cast<int>(m_packed >> 15) bitand 0x07,
 			static_cast<int>(m_packed >> 12) bitand 0x07,
-			static_cast<int>(m_packed >>  9) bitand 0x07,
-			static_cast<int>(m_packed >>  6) bitand 0x07,
-			static_cast<int>(m_packed >>  3) bitand 0x07,
-			static_cast<int>(m_packed >>  0) bitand 0x07,
+			static_cast<int>(m_packed >> 9) bitand 0x07,
+			static_cast<int>(m_packed >> 6) bitand 0x07,
+			static_cast<int>(m_packed >> 3) bitand 0x07,
+			static_cast<int>(m_packed >> 0) bitand 0x07,
 		};
 	}
 
   private:
+	friend struct symop_datablock;
 
-	friend struct SymopDataBlock;
+	const uint64_t kPackMask = (~0ULL >> (64 - 36));
 
-	const uint64_t kPackMask = (~0ULL >> (64-36));
-
-	SymopData(uint64_t v)
-		: m_packed(v bitand kPackMask) {}
+	symop_data(uint64_t v)
+		: m_packed(v bitand kPackMask)
+	{
+	}
 
 	uint64_t m_packed;
 };
 
-struct SymopDataBlock
+struct symop_datablock
 {
-	constexpr SymopDataBlock(int spacegroup, int rotational_number, const std::array<int,15>& rt_data)
+	constexpr symop_datablock(int spacegroup, int rotational_number, const std::array<int, 15> &rt_data)
 		: m_v((spacegroup & 0xffffULL) << 48 bitor
 			  (rotational_number & 0xffULL) << 40 bitor
-			  SymopData(rt_data).m_packed)
+			  symop_data(rt_data).m_packed)
 	{
 	}
 
-	uint16_t spacegroup() const			{ return m_v >> 48; }
-	SymopData symop() const				{ return SymopData(m_v); }
-	uint8_t rotational_number() const	{ return (m_v >> 40) bitand 0xff; }
+	uint16_t spacegroup() const { return m_v >> 48; }
+	symop_data symop() const { return symop_data(m_v); }
+	uint8_t rotational_number() const { return (m_v >> 40) bitand 0xff; }
 
   private:
 	uint64_t m_v;
 };
 
-static_assert(sizeof(SymopDataBlock) == sizeof(uint64_t), "Size of SymopData is wrong");
+static_assert(sizeof(symop_datablock) == sizeof(uint64_t), "Size of symop_data is wrong");
 
-CIFPP_EXPORT extern const SymopDataBlock kSymopNrTable[];
-CIFPP_EXPORT extern const std::size_t kSymopNrTableSize;
+extern const symop_datablock kSymopNrTable[];
+extern const std::size_t kSymopNrTableSize;
 
 // --------------------------------------------------------------------
 
-int GetSpacegroupNumber(std::string spacegroup);	// alternative for clipper's parsing code
+int get_space_group_number(std::string spacegroup);                        // alternative for clipper's parsing code, using space_group_name::full
+int get_space_group_number(std::string spacegroup, space_group_name type); // alternative for clipper's parsing code
 
-}
+} // namespace cif

include/cif++/text.hpp

@@ -0,0 +1,521 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <charconv>
+#include <cmath>
+#include <limits>
+#include <set>
+#include <sstream>
+#include <tuple>
+#include <vector>
+
+
+#if __has_include(<experimental/type_traits>)
+#include <experimental/type_traits>
+#else
+#include <type_traits>
+#endif
+
+#if (not defined(__cpp_lib_experimental_detect) or (__cpp_lib_experimental_detect < 201505)) and (not defined(_LIBCPP_VERSION) or _LIBCPP_VERSION < 5000)
+// This code is copied from:
+// https://ld2015.scusa.lsu.edu/cppreference/en/cpp/experimental/is_detected.html
+
+namespace std
+{
+	template< class... >
+	using void_t = void;
+
+	namespace experimental
+	{
+		namespace detail
+		{
+			template <class Default, class AlwaysVoid,
+					template<class...> class Op, class... Args>
+			struct detector
+			{
+				using value_t = false_type;
+				using type = Default;
+			};
+			
+			template <class Default, template<class...> class Op, class... Args>
+			struct detector<Default, void_t<Op<Args...>>, Op, Args...> {
+				// Note that std::void_t is a c++17 feature
+				using value_t = true_type;
+				using type = Op<Args...>;
+			};
+		} // namespace detail
+
+		struct nonesuch
+		{
+			nonesuch() = delete;
+			~nonesuch() = delete;
+			nonesuch(nonesuch const&) = delete;
+			void operator=(nonesuch const&) = delete;
+		};
+
+		template <template<class...> class Op, class... Args>
+		using is_detected = typename detail::detector<nonesuch, void, Op, Args...>::value_t;
+
+		template <template<class...> class Op, class... Args>
+		constexpr inline bool is_detected_v = is_detected<Op,Args...>::value;
+
+		template <template<class...> class Op, class... Args>
+		using detected_t = typename detail::detector<nonesuch, void, Op, Args...>::type;
+
+		template <class Default, template<class...> class Op, class... Args>
+		using detected_or = detail::detector<Default, void, Op, Args...>;
+
+		template <class Expected, template <class...> class Op, class... Args>
+		using is_detected_exact = std::is_same<Expected, detected_t<Op, Args...>>;
+
+		template <class Expected, template<class...> class Op, class... Args>
+		constexpr inline bool is_detected_exact_v = is_detected_exact<Expected, Op, Args...>::value;
+	}
+}
+
+#endif
+
+namespace cif
+{
+
+// --------------------------------------------------------------------
+
+// some basic utilities: Since we're using ASCII input only, we define for optimisation
+// our own case conversion routines.
+
+bool iequals(std::string_view a, std::string_view b);
+int icompare(std::string_view a, std::string_view b);
+
+bool iequals(const char *a, const char *b);
+int icompare(const char *a, const char *b);
+
+void to_lower(std::string &s);
+std::string to_lower_copy(std::string_view s);
+
+void to_upper(std::string &s);
+// std::string toUpperCopy(const std::string &s);
+
+template <typename IterType>
+std::string join(IterType b, IterType e, std::string_view sep)
+{
+	std::ostringstream s;
+
+	if (b != e)
+	{
+		auto ai = b;
+		auto ni = std::next(ai);
+
+		for (;;)
+		{
+			s << *ai;
+
+			if (ni == e)
+				break;
+
+			ai = ni;
+			ni = std::next(ai);
+
+			s << sep;
+		}
+	}
+
+	return s.str();
+}
+
+template <typename V>
+std::string join(const V &arr, std::string_view sep)
+{
+	return join(arr.begin(), arr.end(), sep);
+}
+
+template <typename StringType = std::string_view>
+std::vector<StringType> split(std::string_view s, std::string_view separators, bool suppress_empty = false)
+{
+	std::vector<StringType> result;
+
+	auto b = s.data();
+	auto e = b;
+
+	while (e != s.data() + s.length())
+	{
+		if (separators.find(*e) != std::string_view::npos)
+		{
+			if (e > b or not suppress_empty)
+				result.emplace_back(b, e - b);
+			b = e = e + 1;
+			continue;
+		}
+
+		++e;
+	}
+
+	if (e > b or not suppress_empty)
+		result.emplace_back(b, e - b);
+
+	return result;
+}
+
+void replace_all(std::string &s, std::string_view what, std::string_view with = {});
+
+#if defined(__cpp_lib_starts_ends_with)
+
+inline bool starts_with(std::string s, std::string_view with)
+{
+	return s.starts_with(with);
+}
+
+inline bool ends_with(std::string_view s, std::string_view with)
+{
+	return s.ends_with(with);
+}
+
+#else
+
+inline bool starts_with(std::string s, std::string_view with)
+{
+	return s.compare(0, with.length(), with) == 0;
+}
+
+inline bool ends_with(std::string_view s, std::string_view with)
+{
+	return s.length() >= with.length() and s.compare(s.length() - with.length(), with.length(), with) == 0;
+}
+
+#endif
+
+#if defined(__cpp_lib_string_contains)
+
+inline bool contains(std::string_view s, std::string_view q)
+{
+	return s.contains(q);
+}
+
+#else
+
+inline bool contains(std::string_view s, std::string_view q)
+{
+	return s.find(q) != std::string_view::npos;
+}
+
+#endif
+
+bool icontains(std::string_view s, std::string_view q);
+
+void trim_left(std::string &s);
+void trim_right(std::string &s);
+void trim(std::string &s);
+
+std::string trim_left_copy(std::string_view s);
+std::string trim_right_copy(std::string_view s);
+std::string trim_copy(std::string_view s);
+
+// To make life easier, we also define iless and iset using iequals
+
+struct iless
+{
+	bool operator()(const std::string &a, const std::string &b) const
+	{
+		return icompare(a, b) < 0;
+	}
+};
+
+typedef std::set<std::string, iless> iset;
+
+// --------------------------------------------------------------------
+// This really makes a difference, having our own tolower routines
+
+extern const uint8_t kCharToLowerMap[256];
+
+inline char tolower(int ch)
+{
+	return static_cast<char>(kCharToLowerMap[static_cast<uint8_t>(ch)]);
+}
+
+// --------------------------------------------------------------------
+
+std::tuple<std::string, std::string> split_tag_name(std::string_view tag);
+
+// --------------------------------------------------------------------
+// generate a cif name, mainly used to generate asym_id's
+
+std::string cif_id_for_number(int number);
+
+// --------------------------------------------------------------------
+//	custom wordwrapping routine
+
+std::vector<std::string> word_wrap(const std::string &text, size_t width);
+
+// --------------------------------------------------------------------
+/// std::from_chars for floating point types.
+/// These are optional, there's a selected_charconv class below that selects
+/// the best option to used based on support by the stl library
+/// I.e. that in case of GNU < 12 (or something) the cif implementation will
+/// be used, all other cases will use the stl version.
+
+template <typename FloatType, std::enable_if_t<std::is_floating_point_v<FloatType>, int> = 0>
+std::from_chars_result from_chars(const char *first, const char *last, FloatType &value)
+{
+	std::from_chars_result result{ first, {} };
+
+	enum State
+	{
+		IntegerSign,
+		Integer,
+		Fraction,
+		ExponentSign,
+		Exponent
+	} state = IntegerSign;
+	int sign = 1;
+	unsigned long long vi = 0;
+	long double f = 1;
+	int exponent_sign = 1;
+	int exponent = 0;
+	bool done = false;
+
+	while (not done and result.ec == std::errc())
+	{
+		char ch = result.ptr != last ? *result.ptr : 0;
+		++result.ptr;
+
+		switch (state)
+		{
+			case IntegerSign:
+				if (ch == '-')
+				{
+					sign = -1;
+					state = Integer;
+				}
+				else if (ch == '+')
+					state = Integer;
+				else if (ch >= '0' and ch <= '9')
+				{
+					vi = ch - '0';
+					state = Integer;
+				}
+				else if (ch == '.')
+					state = Fraction;
+				else
+					result.ec = std::errc::invalid_argument;
+				break;
+
+			case Integer:
+				if (ch >= '0' and ch <= '9')
+					vi = 10 * vi + (ch - '0');
+				else if (ch == 'e' or ch == 'E')
+					state = ExponentSign;
+				else if (ch == '.')
+					state = Fraction;
+				else
+				{
+					done = true;
+					--result.ptr;
+				}
+				break;
+
+			case Fraction:
+				if (ch >= '0' and ch <= '9')
+				{
+					vi = 10 * vi + (ch - '0');
+					f /= 10;
+				}
+				else if (ch == 'e' or ch == 'E')
+					state = ExponentSign;
+				else
+				{
+					done = true;
+					--result.ptr;
+				}
+				break;
+
+			case ExponentSign:
+				if (ch == '-')
+				{
+					exponent_sign = -1;
+					state = Exponent;
+				}
+				else if (ch == '+')
+					state = Exponent;
+				else if (ch >= '0' and ch <= '9')
+				{
+					exponent = ch - '0';
+					state = Exponent;
+				}
+				else
+					result.ec = std::errc::invalid_argument;
+				break;
+
+			case Exponent:
+				if (ch >= '0' and ch <= '9')
+					exponent = 10 * exponent + (ch - '0');
+				else
+				{
+					done = true;
+					--result.ptr;
+				}
+				break;
+		}
+	}
+
+	if (result.ec == std::errc())
+	{
+		long double v = f * vi * sign;
+		if (exponent != 0)
+			v *= std::pow(10, exponent * exponent_sign);
+
+		if (std::isnan(v))
+			result.ec = std::errc::invalid_argument;
+		else if (std::abs(v) > std::numeric_limits<FloatType>::max())
+			result.ec = std::errc::result_out_of_range;
+
+		value = static_cast<FloatType>(v);
+	}
+
+	return result;
+}
+
+enum class chars_format
+{
+	scientific = 1,
+	fixed = 2,
+	// hex,
+	general = fixed | scientific
+};
+
+template <typename FloatType, std::enable_if_t<std::is_floating_point_v<FloatType>, int> = 0>
+std::to_chars_result to_chars(char *first, char *last, FloatType &value, chars_format fmt)
+{
+	int size = static_cast<int>(last - first);
+	int r = 0;
+
+	switch (fmt)
+	{
+		case chars_format::scientific:
+			if constexpr (std::is_same_v<FloatType, long double>)
+				r = snprintf(first, last - first, "%le", value);
+			else
+				r = snprintf(first, last - first, "%e", value);
+			break;
+
+		case chars_format::fixed:
+			if constexpr (std::is_same_v<FloatType, long double>)
+				r = snprintf(first, last - first, "%lf", value);
+			else
+				r = snprintf(first, last - first, "%f", value);
+			break;
+
+		case chars_format::general:
+			if constexpr (std::is_same_v<FloatType, long double>)
+				r = snprintf(first, last - first, "%lg", value);
+			else
+				r = snprintf(first, last - first, "%g", value);
+			break;
+	}
+
+	std::to_chars_result result;
+	if (r < 0 or r >= size)
+		result = { first, std::errc::value_too_large };
+	else
+		result = { first + r, std::errc() };
+
+	return result;
+}
+
+template <typename FloatType, std::enable_if_t<std::is_floating_point_v<FloatType>, int> = 0>
+std::to_chars_result to_chars(char *first, char *last, FloatType &value, chars_format fmt, int precision)
+{
+	int size = last - first;
+	int r;
+
+	switch (fmt)
+	{
+		case chars_format::scientific:
+			if constexpr (std::is_same_v<FloatType, long double>)
+				r = snprintf(first, last - first, "%.*le", precision, value);
+			else
+				r = snprintf(first, last - first, "%.*e", precision, value);
+			break;
+
+		case chars_format::fixed:
+			if constexpr (std::is_same_v<FloatType, long double>)
+				r = snprintf(first, last - first, "%.*lf", precision, value);
+			else
+				r = snprintf(first, last - first, "%.*f", precision, value);
+			break;
+
+		case chars_format::general:
+			if constexpr (std::is_same_v<FloatType, long double>)
+				r = snprintf(first, last - first, "%.*lg", precision, value);
+			else
+				r = snprintf(first, last - first, "%.*g", precision, value);
+			break;
+	}
+
+	std::to_chars_result result;
+	if (r < 0 or r >= size)
+		result = { first, std::errc::value_too_large };
+	else
+		result = { first + r, std::errc() };
+
+	return result;
+}
+
+template <typename T>
+struct my_charconv
+{
+	static std::from_chars_result from_chars(const char *a, const char *b, T &d)
+	{
+		return cif::from_chars(a, b, d);
+	}
+
+	static std::to_chars_result to_chars(char *first, char *last, T &value, chars_format fmt)
+	{
+		return cif::to_chars(first, last, value, fmt);
+	}
+};
+
+template <typename T>
+struct std_charconv
+{
+	static std::from_chars_result from_chars(const char *a, const char *b, T &d)
+	{
+		return std::from_chars(a, b, d);
+	}
+
+	static std::to_chars_result to_chars(char *first, char *last, T &value, chars_format fmt)
+	{
+		return std::to_chars(first, last, value, fmt);
+	}
+};
+
+template <typename T>
+using from_chars_function = decltype(std::from_chars(std::declval<const char *>(), std::declval<const char *>(), std::declval<T &>()));
+
+template <typename T>
+using selected_charconv = typename std::conditional_t<std::experimental::is_detected_v<from_chars_function, T>, std_charconv<T>, my_charconv<T>>;
+
+} // namespace cif

include/cif++/utilities.hpp

@@ -26,13 +26,7 @@
 
 #pragma once
 
-#include <cassert>
 #include <filesystem>
-#include <iostream>
-#include <list>
-#include <memory>
-#include <set>
-#include <vector>
 
 #ifndef STDOUT_FILENO
 #define STDOUT_FILENO 1
@@ -45,8 +39,6 @@
 #include <unistd.h>
 #endif
 
-#include "cif++/Cif++Export.hpp"
-
 #if _MSC_VER
 #pragma warning(disable : 4996) // unsafe function or variable	(strcpy e.g.)
 #pragma warning(disable : 4068) // unknown pragma
@@ -58,60 +50,12 @@
 namespace cif
 {
 
+extern int VERBOSE;
+
 // the git 'build' number
 std::string get_version_nr();
 // std::string get_version_date();
 
-// --------------------------------------------------------------------
-
-// some basic utilities: Since we're using ASCII input only, we define for optimisation
-// our own case conversion routines.
-
-bool iequals(const std::string &a, const std::string &b);
-int icompare(const std::string &a, const std::string &b);
-
-bool iequals(const char *a, const char *b);
-int icompare(const char *a, const char *b);
-
-void toLower(std::string &s);
-std::string toLowerCopy(const std::string &s);
-
-// To make life easier, we also define iless and iset using iequals
-
-struct iless
-{
-	bool operator()(const std::string &a, const std::string &b) const
-	{
-		return icompare(a, b) < 0;
-	}
-};
-
-typedef std::set<std::string, iless> iset;
-
-// --------------------------------------------------------------------
-// This really makes a difference, having our own tolower routines
-
-extern const uint8_t kCharToLowerMap[256];
-
-inline char tolower(int ch)
-{
-	return static_cast<char>(kCharToLowerMap[static_cast<uint8_t>(ch)]);
-}
-
-// --------------------------------------------------------------------
-
-std::tuple<std::string, std::string> splitTagName(const std::string &tag);
-
-// --------------------------------------------------------------------
-// generate a cif name, mainly used to generate asym_id's
-
-std::string cifIdForNumber(int number);
-
-// --------------------------------------------------------------------
-//	custom wordwrapping routine
-
-std::vector<std::string> wordWrap(const std::string &text, size_t width);
-
 // --------------------------------------------------------------------
 //	Code helping with terminal i/o
 
@@ -226,14 +170,14 @@ class Progress
 	Progress(const Progress &) = delete;
 	Progress &operator=(const Progress &) = delete;
 
-	struct ProgressImpl *mImpl;
+	struct ProgressImpl *m_impl;
 };
 
 // --------------------------------------------------------------------
 // Resources
 
-std::unique_ptr<std::istream> loadResource(std::filesystem::path name);
-void addFileResource(const std::string &name, std::filesystem::path dataFile);
-void addDataDirectory(std::filesystem::path dataDir);
+std::unique_ptr<std::istream> load_resource(std::filesystem::path name);
+void add_file_resource(const std::string &name, std::filesystem::path dataFile);
+void add_data_directory(std::filesystem::path dataDir);
 
 } // namespace cif

include/cif++/validate.hpp

@@ -0,0 +1,242 @@
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ *
+ * Copyright (c) 2022 NKI/AVL, Netherlands Cancer Institute
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ *    list of conditions and the following disclaimer
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#pragma once
+
+#include <filesystem>
+#include <list>
+#include <mutex>
+#include <utility>
+
+#include <cif++/text.hpp>
+
+namespace cif
+{
+
+struct category_validator;
+
+// --------------------------------------------------------------------
+
+class validation_error : public std::exception
+{
+  public:
+	validation_error(const std::string &msg);
+	validation_error(const std::string &cat, const std::string &item,
+		const std::string &msg);
+	const char *what() const noexcept { return m_msg.c_str(); }
+	std::string m_msg;
+};
+
+// --------------------------------------------------------------------
+
+enum class DDL_PrimitiveType
+{
+	Char,
+	UChar,
+	Numb
+};
+
+DDL_PrimitiveType map_to_primitive_type(std::string_view s);
+
+struct regex_impl;
+
+struct type_validator
+{
+	std::string m_name;
+	DDL_PrimitiveType m_primitive_type;
+	regex_impl *m_rx;
+
+	type_validator() = delete;
+	type_validator(std::string_view name, DDL_PrimitiveType type, std::string_view rx);
+
+	type_validator(const type_validator &) = delete;
+	type_validator(type_validator &&rhs)
+		: m_name(std::move(rhs.m_name))
+		, m_primitive_type(rhs.m_primitive_type)
+	{
+		m_rx = std::exchange(rhs.m_rx, nullptr);
+	}
+
+	type_validator &operator=(const type_validator &) = delete;
+	type_validator &operator=(type_validator &&rhs)
+	{
+		m_name = std::move(rhs.m_name);
+		m_primitive_type = rhs.m_primitive_type;
+		m_rx = std::exchange(rhs.m_rx, nullptr);
+
+		return *this;
+	}
+
+	~type_validator();
+
+	bool operator<(const type_validator &rhs) const
+	{
+		return icompare(m_name, rhs.m_name) < 0;
+	}
+
+	int compare(std::string_view a, std::string_view b) const;
+};
+
+struct item_validator
+{
+	std::string m_tag;
+	bool m_mandatory;
+	const type_validator *m_type;
+	cif::iset m_enums;
+	std::string m_default;
+	bool m_default_is_null;
+	category_validator *m_category = nullptr;
+
+	// ItemLinked is used for non-key links
+	struct item_link
+	{
+		item_validator *m_parent;
+		std::string m_parent_item;
+		std::string m_child_item;
+	};
+
+	std::vector<item_link> mLinked;
+
+	bool operator<(const item_validator &rhs) const
+	{
+		return icompare(m_tag, rhs.m_tag) < 0;
+	}
+
+	bool operator==(const item_validator &rhs) const
+	{
+		return iequals(m_tag, rhs.m_tag);
+	}
+
+	void operator()(std::string_view value) const;
+};
+
+struct category_validator
+{
+	std::string m_name;
+	std::vector<std::string> m_keys;
+	cif::iset m_groups;
+	cif::iset m_mandatory_fields;
+	std::set<item_validator> m_item_validators;
+
+	bool operator<(const category_validator &rhs) const
+	{
+		return icompare(m_name, rhs.m_name) < 0;
+	}
+
+	void addItemValidator(item_validator &&v);
+
+	const item_validator *get_validator_for_item(std::string_view tag) const;
+
+	const std::set<item_validator> &item_validators() const
+	{
+		return m_item_validators;
+	}
+};
+
+struct link_validator
+{
+	int m_link_group_id;
+	std::string m_parent_category;
+	std::vector<std::string> m_parent_keys;
+	std::string m_child_category;
+	std::vector<std::string> m_child_keys;
+	std::string m_link_group_label;
+};
+
+// --------------------------------------------------------------------
+
+class validator
+{
+  public:
+	validator(std::string_view name)
+		: m_name(name)
+	{
+	}
+
+	~validator() = default;
+
+	validator(const validator &rhs) = delete;
+	validator &operator=(const validator &rhs) = delete;
+
+	validator(validator &&rhs) = default;
+	validator &operator=(validator &&rhs) = default;
+
+	friend class dictionary_parser;
+
+	void add_type_validator(type_validator &&v);
+	const type_validator *get_validator_for_type(std::string_view type_code) const;
+
+	void add_category_validator(category_validator &&v);
+	const category_validator *get_validator_for_category(std::string_view category) const;
+
+	void add_link_validator(link_validator &&v);
+	std::vector<const link_validator *> get_links_for_parent(std::string_view category) const;
+	std::vector<const link_validator *> get_links_for_child(std::string_view category) const;
+
+	void report_error(const std::string &msg, bool fatal) const;
+
+	const std::string &name() const { return m_name; }
+	void set_name(const std::string &name) { m_name = name; }
+
+	const std::string &version() const { return m_version; }
+	void version(const std::string &version) { m_version = version; }
+
+  private:
+	// name is fully qualified here:
+	item_validator *get_validator_for_item(std::string_view name) const;
+
+	std::string m_name;
+	std::string m_version;
+	bool m_strict = false;
+	std::set<type_validator> m_type_validators;
+	std::set<category_validator> m_category_validators;
+	std::vector<link_validator> m_link_validators;
+};
+
+// --------------------------------------------------------------------
+class validator_factory
+{
+  public:
+	static validator_factory &instance()
+	{
+		static validator_factory s_instance;
+		return s_instance;
+	}
+
+	const validator &operator[](std::string_view dictionary_name);
+
+  private:
+	void construct_validator(std::string_view name, std::istream &is);
+
+	// --------------------------------------------------------------------
+
+	validator_factory() = default;
+
+	std::mutex m_mutex;
+	std::list<validator> m_validators;
+};
+
+} // namespace cif

libcifpp.pc.in

@@ -8,5 +8,6 @@ Name: libcifpp
 Description: C++ library for the manipulation of mmCIF files.
 Version: @PACKAGE_VERSION@
 
-Libs: -L${libdir} -lcifpp -lboost_regex -lboost_iostreams
+Requires.private: zlib, liblzma
+Libs: -L${libdir} -lcifpp
 Cflags: -I${includedir} -pthread

regex/CMakeLists.txt

@@ -0,0 +1,69 @@
+# Copyright 2018 Mike Dev
+# Copyright 2019 Peter Dimov
+# Distributed under the Boost Software License, Version 1.0.
+# See accompanying file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt
+
+# We support CMake 3.5, but prefer 3.16 policies and behavior
+cmake_minimum_required(VERSION 3.5...3.16)
+
+project(boost_regex VERSION "${BOOST_SUPERPROJECT_VERSION}" LANGUAGES CXX)
+
+add_library(boost_regex INTERFACE)
+add_library(Boost::regex ALIAS boost_regex)
+
+target_include_directories(boost_regex INTERFACE include)
+
+option(BOOST_REGEX_STANDALONE "Boost.Regex: Enable Standalone Mode (i.e. no Boost dependencies)")
+
+if(NOT BOOST_REGEX_STANDALONE)
+
+   target_link_libraries(boost_regex
+     INTERFACE
+       Boost::config
+       Boost::throw_exception
+       Boost::predef
+       Boost::assert
+   )
+
+else()
+
+   target_compile_definitions(boost_regex
+     INTERFACE BOOST_REGEX_STANDALONE
+   )
+
+endif()
+
+find_package(ICU COMPONENTS data i18n uc QUIET)
+#option(BOOST_REGEX_ENABLE_ICU "Boost.Regex: enable ICU support" ${ICU_FOUND})
+
+if(ICU_FOUND)
+
+   add_library(boost_regex_icu INTERFACE)
+   add_library(Boost::regex_icu ALIAS boost_regex_icu)
+
+   target_include_directories(boost_regex_icu INTERFACE include)
+
+   if(NOT BOOST_REGEX_STANDALONE)
+
+      target_link_libraries(boost_regex_icu
+        INTERFACE
+          Boost::config
+          Boost::throw_exception
+          Boost::predef
+          Boost::assert
+      )
+
+   else()
+
+      target_compile_definitions(boost_regex_icu
+        INTERFACE BOOST_REGEX_STANDALONE
+      )
+
+   endif()
+
+   find_package(ICU COMPONENTS data i18n uc REQUIRED)
+
+   target_link_libraries(boost_regex_icu INTERFACE ICU::data ICU::i18n ICU::uc)
+
+endif()
+

regex/LICENSE_1_0.txt

@@ -0,0 +1,23 @@
+Boost Software License - Version 1.0 - August 17th, 2003
+
+Permission is hereby granted, free of charge, to any person or organization
+obtaining a copy of the software and accompanying documentation covered by
+this license (the "Software") to use, reproduce, display, distribute,
+execute, and transmit the Software, and to prepare derivative works of the
+Software, and to permit third-parties to whom the Software is furnished to
+do so, all subject to the following:
+
+The copyright notices in the Software and this entire statement, including
+the above license grant, this restriction and the following disclaimer,
+must be included in all copies of the Software, in whole or in part, and
+all derivative works of the Software, unless such copies or derivative
+works are solely in the form of machine-executable object code generated by
+a source language processor.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT
+SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE
+FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+DEALINGS IN THE SOFTWARE.

regex/README.md

@@ -0,0 +1,47 @@
+Boost Regex Library
+============================
+
+The Boost Regex library provides regular expression support for C++, this library is the ancestor to std::regex and still goes beyond
+and offers some advantages to, the standard version.
+
+The full documentation is available on [boost.org](http://www.boost.org/doc/libs/release/libs/regex/index.html).
+
+## Standalone Mode ##
+
+ This library may now be used in "standalone" mode without the rest of the Boost C++ libraries, in order to do this you must either:
+
+* Have a C++17 compiler that supports __has_include, in this case if <boost/config.hpp> is not present then the library will automoatically enter standalone mode. Or:
+* Define BOOST_REGEX_STANDALONE when building.
+
+The main difference between the 2 modes, is that when Boost.Config is present the library will automatically configure itself around various compiler defects. In particular in order to use the library with exception support turned off, you will either need a copy of Boost.Config in your include path, or else manually define BOOST_NO_EXCEPTIONS when building.
+
+In any event, to obtain a standalone version of this library, simply download a .zip of the "master" branch of this repository.
+
+## Support, bugs and feature requests ##
+
+Bugs and feature requests can be reported through the [Gitub issue tracker](https://github.com/boostorg/regex/issues)
+(see [open issues](https://github.com/boostorg/regex/issues) and
+[closed issues](https://github.com/boostorg/regex/issues?utf8=%E2%9C%93&q=is%3Aissue+is%3Aclosed)).
+
+You can submit your changes through a [pull request](https://github.com/boostorg/regex/pulls).
+
+There is no mailing-list specific to Boost Regex, although you can use the general-purpose Boost [mailing-list](http://lists.boost.org/mailman/listinfo.cgi/boost-users) using the tag [regex].
+
+
+## Development ##
+
+Clone the whole boost project, which includes the individual Boost projects as submodules ([see boost+git doc](https://github.com/boostorg/boost/wiki/Getting-Started)): 
+
+    git clone https://github.com/boostorg/boost
+    cd boost
+    git submodule update --init
+
+The Boost Regex Library is located in `libs/regex/`. 
+
+### Running tests ###
+First, make sure you are in `libs/regex/test`. 
+You can either run all the tests listed in `Jamfile.v2` or run a single test:
+
+    ../../../b2                        <- run all tests
+    ../../../b2 regex_regress          <- single test
+

regex/include/boost/cregex.hpp

@@ -0,0 +1,43 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org/libs/regex for most recent version.
+  *   FILE         cregex.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares POSIX API functions
+  *                + boost::RegEx high level wrapper.
+  */
+
+#ifndef BOOST_RE_CREGEX_HPP
+#define BOOST_RE_CREGEX_HPP
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+
+#ifdef BOOST_REGEX_CXX03
+#include <boost/regex/v4/cregex.hpp>
+#else
+#include <boost/regex/v5/cregex.hpp>
+#endif
+
+#endif /* include guard */
+
+
+
+
+
+
+
+
+
+

regex/include/boost/regex.h

@@ -0,0 +1,100 @@
+/*
+ *
+ * Copyright (c) 1998-2000
+ * Dr John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org/libs/regex for documentation.
+  *   FILE         regex.h
+  *   VERSION      3.12
+  *   DESCRIPTION: Declares POSIX API functions
+  */
+
+#ifndef BOOST_RE_REGEX_H
+#define BOOST_RE_REGEX_H
+
+#include <boost/cregex.hpp>
+
+/*
+*  add using declarations to bring POSIX API functions into
+* global scope, only if this is C++ (and not C).
+*/
+#ifdef __cplusplus
+
+using boost::regoff_t;
+using boost::regex_tA;
+using boost::regmatch_t;
+using boost::REG_BASIC;
+using boost::REG_EXTENDED;
+using boost::REG_ICASE;
+using boost::REG_NOSUB;
+using boost::REG_NEWLINE;
+using boost::REG_NOSPEC;
+using boost::REG_PEND;
+using boost::REG_DUMP;
+using boost::REG_NOCOLLATE;
+using boost::REG_ESCAPE_IN_LISTS;
+using boost::REG_NEWLINE_ALT;
+using boost::REG_PERL;
+using boost::REG_AWK;
+using boost::REG_GREP;
+using boost::REG_EGREP;
+using boost::REG_ASSERT;
+using boost::REG_INVARG;
+using boost::REG_ATOI;
+using boost::REG_ITOA;
+
+using boost::REG_NOTBOL;
+using boost::REG_NOTEOL;
+using boost::REG_STARTEND;
+
+using boost::reg_comp_flags;
+using boost::reg_exec_flags;
+using boost::regcompA;
+using boost::regerrorA;
+using boost::regexecA;
+using boost::regfreeA;
+
+#ifndef BOOST_NO_WREGEX
+using boost::regcompW;
+using boost::regerrorW;
+using boost::regexecW;
+using boost::regfreeW;
+using boost::regex_tW;
+#endif
+
+using boost::REG_NOERROR;
+using boost::REG_NOMATCH;
+using boost::REG_BADPAT;
+using boost::REG_ECOLLATE;
+using boost::REG_ECTYPE;
+using boost::REG_EESCAPE;
+using boost::REG_ESUBREG;
+using boost::REG_EBRACK;
+using boost::REG_EPAREN;
+using boost::REG_EBRACE;
+using boost::REG_BADBR;
+using boost::REG_ERANGE;
+using boost::REG_ESPACE;
+using boost::REG_BADRPT;
+using boost::REG_EEND;
+using boost::REG_ESIZE;
+using boost::REG_ERPAREN;
+using boost::REG_EMPTY;
+using boost::REG_E_MEMORY;
+using boost::REG_E_UNKNOWN;
+using boost::reg_errcode_t;
+
+#endif /* __cplusplus */
+
+#endif /* BOOST_RE_REGEX_H */
+
+
+
+

regex/include/boost/regex.hpp

@@ -0,0 +1,41 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org/libs/regex for documentation.
+  *   FILE         regex.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares boost::basic_regex<> and associated
+  *                functions and classes. This header is the main
+  *                entry point for the template regex code.
+  */
+
+
+/* start with C compatibility API */
+
+#ifndef BOOST_RE_REGEX_HPP
+#define BOOST_RE_REGEX_HPP
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+
+#ifdef BOOST_REGEX_CXX03
+#include <boost/regex/v4/regex.hpp>
+#else
+#include <boost/regex/v5/regex.hpp>
+#endif
+
+#endif  // include
+
+
+
+

regex/include/boost/regex/config.hpp

@@ -0,0 +1,480 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the
+ * Boost Software License, Version 1.0. (See accompanying file
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         config.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: regex extended config setup.
+  */
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#define BOOST_REGEX_CONFIG_HPP
+
+#if !((__cplusplus >= 201103L) || (defined(_MSC_VER) && (_MSC_VER >= 1600)) || defined(BOOST_REGEX_CXX03))
+#  define BOOST_REGEX_CXX03
+#endif
+
+#if defined(BOOST_REGEX_RECURSIVE) && !defined(BOOST_REGEX_CXX03)
+#  define BOOST_REGEX_CXX03
+#endif
+
+#if defined(__has_include)
+#if !defined(BOOST_REGEX_STANDALONE) && !__has_include(<boost/version.hpp>)
+#define BOOST_REGEX_STANDALONE
+#endif
+#endif
+
+/*
+ * Borland C++ Fix/error check
+ * this has to go *before* we include any std lib headers:
+ */
+#if defined(__BORLANDC__) && !defined(__clang__)
+#  include <boost/regex/config/borland.hpp>
+#endif
+#ifndef BOOST_REGEX_STANDALONE
+#include <boost/version.hpp>
+#endif
+
+/*************************************************************************
+*
+* Asserts:
+*
+*************************************************************************/
+
+#ifdef BOOST_REGEX_STANDALONE
+#include <cassert>
+#  define BOOST_REGEX_ASSERT(x) assert(x)
+#else
+#include <boost/assert.hpp>
+#  define BOOST_REGEX_ASSERT(x) BOOST_ASSERT(x)
+#endif
+
+/*****************************************************************************
+ *
+ *  Include all the headers we need here:
+ *
+ ****************************************************************************/
+
+#ifdef __cplusplus
+
+#  ifndef BOOST_REGEX_USER_CONFIG
+#     define BOOST_REGEX_USER_CONFIG <boost/regex/user.hpp>
+#  endif
+
+#  include BOOST_REGEX_USER_CONFIG
+
+#ifndef BOOST_REGEX_STANDALONE
+#  include <boost/config.hpp>
+#  include <boost/predef.h>
+#endif
+
+#else
+   /*
+    * C build,
+    * don't include <boost/config.hpp> because that may
+    * do C++ specific things in future...
+    */
+#  include <stdlib.h>
+#  include <stddef.h>
+#  ifdef _MSC_VER
+#     define BOOST_MSVC _MSC_VER
+#  endif
+#endif
+
+
+/****************************************************************************
+*
+* Legacy support:
+*
+*******************************************************************************/
+
+#if defined(BOOST_NO_STD_LOCALE) || defined(BOOST_NO_CXX11_HDR_MUTEX) || defined(BOOST_NO_CXX11_HDR_TYPE_TRAITS) \
+   || defined(BOOST_NO_CXX11_HDR_ATOMIC) || defined(BOOST_NO_CXX11_ALLOCATOR) || defined(BOOST_NO_CXX11_SMART_PTR) \
+   || defined(BOOST_NO_CXX11_STATIC_ASSERT) || defined(BOOST_NO_NOEXCEPT)
+#ifndef BOOST_REGEX_CXX03
+#  define BOOST_REGEX_CXX03
+#endif
+#endif
+
+/*****************************************************************************
+ *
+ *  Boilerplate regex config options:
+ *
+ ****************************************************************************/
+
+/* Obsolete macro, use BOOST_VERSION instead: */
+#define BOOST_RE_VERSION 500
+
+/* fix: */
+#if defined(_UNICODE) && !defined(UNICODE)
+#define UNICODE
+#endif
+
+#define BOOST_REGEX_JOIN(X, Y) BOOST_REGEX_DO_JOIN(X, Y)
+#define BOOST_REGEX_DO_JOIN(X, Y) BOOST_REGEX_DO_JOIN2(X,Y)
+#define BOOST_REGEX_DO_JOIN2(X, Y) X##Y
+
+#ifdef BOOST_FALLTHROUGH
+#  define BOOST_REGEX_FALLTHROUGH BOOST_FALLTHROUGH
+#else
+
+#if defined(__clang__) && (__cplusplus >= 201103L) && defined(__has_warning)
+#  if __has_feature(cxx_attributes) && __has_warning("-Wimplicit-fallthrough")
+#    define BOOST_REGEX_FALLTHROUGH [[clang::fallthrough]]
+#  endif
+#endif
+#if !defined(BOOST_REGEX_FALLTHROUGH) && defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 1800) && (__cplusplus >= 201703)
+#  define BOOST_REGEX_FALLTHROUGH [[fallthrough]]
+#endif
+#if !defined(BOOST_REGEX_FALLTHROUGH) && defined(__GNUC__) && (__GNUC__ >= 7)
+#  define BOOST_REGEX_FALLTHROUGH __attribute__((fallthrough))
+#endif
+
+#if !defined(BOOST_REGEX_FALLTHROUGH)
+#  define BOOST_REGEX_FALLTHROUGH
+#endif
+#endif
+
+#ifdef BOOST_NORETURN
+#  define BOOST_REGEX_NORETURN BOOST_NORETURN
+#else
+#  define BOOST_REGEX_NORETURN
+#endif
+
+
+/*
+* Define a macro for the namespace that details are placed in, this includes the Boost
+* version number to avoid mismatched header and library versions:
+*/
+#define BOOST_REGEX_DETAIL_NS BOOST_REGEX_JOIN(re_detail_, BOOST_RE_VERSION)
+
+/*
+ * Fix for gcc prior to 3.4: std::ctype<wchar_t> doesn't allow
+ * masks to be combined, for example:
+ * std::use_facet<std::ctype<wchar_t> >.is(std::ctype_base::lower|std::ctype_base::upper, L'a');
+ * returns *false*.
+ */
+#if defined(__GLIBCPP__) && defined(BOOST_REGEX_CXX03)
+#  define BOOST_REGEX_BUGGY_CTYPE_FACET
+#endif
+
+/*
+ * If there isn't good enough wide character support then there will
+ * be no wide character regular expressions:
+ */
+#if (defined(BOOST_NO_CWCHAR) || defined(BOOST_NO_CWCTYPE) || defined(BOOST_NO_STD_WSTRING))
+#  if !defined(BOOST_NO_WREGEX)
+#     define BOOST_NO_WREGEX
+#  endif
+#else
+#  if defined(__sgi) && (defined(__SGI_STL_PORT) || defined(_STLPORT_VERSION))
+      /* STLPort on IRIX is misconfigured: <cwctype> does not compile
+       * as a temporary fix include <wctype.h> instead and prevent inclusion
+       * of STLPort version of <cwctype> */
+#     include <wctype.h>
+#     define __STLPORT_CWCTYPE
+#     define _STLP_CWCTYPE
+#  endif
+
+#if defined(__cplusplus) && defined(BOOST_REGEX_CXX03)
+#  include <boost/regex/config/cwchar.hpp>
+#endif
+
+#endif
+
+/*
+ * If Win32 support has been disabled for boost in general, then
+ * it is for regex in particular:
+ */
+#if defined(BOOST_DISABLE_WIN32) && !defined(BOOST_REGEX_NO_W32)
+#  define BOOST_REGEX_NO_W32
+#endif
+
+/* disable our own file-iterators and mapfiles if we can't
+ * support them: */
+#if defined(_WIN32)
+#  if defined(BOOST_REGEX_NO_W32) || BOOST_PLAT_WINDOWS_RUNTIME
+#    define BOOST_REGEX_NO_FILEITER
+#  endif
+#else /* defined(_WIN32) */
+#  if !defined(BOOST_HAS_DIRENT_H)
+#    define BOOST_REGEX_NO_FILEITER
+#  endif
+#endif
+
+/* backwards compatibitity: */
+#if defined(BOOST_RE_NO_LIB)
+#  define BOOST_REGEX_NO_LIB
+#endif
+
+#if defined(__GNUC__) && !defined(_MSC_VER) && (defined(_WIN32) || defined(__CYGWIN__))
+/* gcc on win32 has problems if you include <windows.h>
+   (sporadically generates bad code). */
+#  define BOOST_REGEX_NO_W32
+#endif
+#if defined(__COMO__) && !defined(BOOST_REGEX_NO_W32) && !defined(_MSC_EXTENSIONS)
+#  define BOOST_REGEX_NO_W32
+#endif
+
+#ifdef BOOST_REGEX_STANDALONE
+#  if defined(_MSC_VER) && !defined(__clang__) && !defined(__GNUC__)
+#     define BOOST_REGEX_MSVC _MSC_VER
+#endif
+#elif defined(BOOST_MSVC)
+#  define BOOST_REGEX_MSVC BOOST_MSVC
+#endif
+
+
+/*****************************************************************************
+ *
+ *  Set up dll import/export options:
+ *
+ ****************************************************************************/
+
+#if (defined(BOOST_REGEX_DYN_LINK) || defined(BOOST_ALL_DYN_LINK)) && !defined(BOOST_REGEX_STATIC_LINK) && defined(BOOST_SYMBOL_IMPORT)
+#  if defined(BOOST_REGEX_SOURCE)
+#     define BOOST_REGEX_BUILD_DLL
+#     define BOOST_REGEX_DECL BOOST_SYMBOL_EXPORT
+#  else
+#     define BOOST_REGEX_DECL BOOST_SYMBOL_IMPORT
+#  endif
+#else
+#  define BOOST_REGEX_DECL
+#endif
+
+#ifdef BOOST_REGEX_CXX03
+#if !defined(BOOST_REGEX_NO_LIB) && !defined(BOOST_REGEX_SOURCE) && !defined(BOOST_ALL_NO_LIB) && defined(__cplusplus)
+#  define BOOST_LIB_NAME boost_regex
+#  if defined(BOOST_REGEX_DYN_LINK) || defined(BOOST_ALL_DYN_LINK)
+#     define BOOST_DYN_LINK
+#  endif
+#  ifdef BOOST_REGEX_DIAG
+#     define BOOST_LIB_DIAGNOSTIC
+#  endif
+#  include <boost/config/auto_link.hpp>
+#endif
+#endif
+
+/*****************************************************************************
+ *
+ *  Set up function call type:
+ *
+ ****************************************************************************/
+
+#if defined(_MSC_VER) && defined(_MSC_EXTENSIONS)
+#if defined(_DEBUG) || defined(__MSVC_RUNTIME_CHECKS) || defined(_MANAGED) || defined(BOOST_REGEX_NO_FASTCALL)
+#  define BOOST_REGEX_CALL __cdecl
+#else
+#  define BOOST_REGEX_CALL __fastcall
+#endif
+#  define BOOST_REGEX_CCALL __cdecl
+#endif
+
+#if defined(__BORLANDC__) && !defined(BOOST_DISABLE_WIN32)
+#if defined(__clang__)
+#  define BOOST_REGEX_CALL __cdecl
+#  define BOOST_REGEX_CCALL __cdecl
+#else
+#  define BOOST_REGEX_CALL __fastcall
+#  define BOOST_REGEX_CCALL __stdcall
+#endif
+#endif
+
+#ifndef BOOST_REGEX_CALL
+#  define BOOST_REGEX_CALL
+#endif
+#ifndef BOOST_REGEX_CCALL
+#define BOOST_REGEX_CCALL
+#endif
+
+/*****************************************************************************
+ *
+ *  Set up localisation model:
+ *
+ ****************************************************************************/
+
+/* backwards compatibility: */
+#ifdef BOOST_RE_LOCALE_C
+#  define BOOST_REGEX_USE_C_LOCALE
+#endif
+
+#ifdef BOOST_RE_LOCALE_CPP
+#  define BOOST_REGEX_USE_CPP_LOCALE
+#endif
+
+#if defined(__CYGWIN__)
+#  define BOOST_REGEX_USE_C_LOCALE
+#endif
+
+/* use C++ locale when targeting windows store */
+#if BOOST_PLAT_WINDOWS_RUNTIME
+#  define BOOST_REGEX_USE_CPP_LOCALE
+#  define BOOST_REGEX_NO_WIN32_LOCALE
+#endif
+
+/* Win32 defaults to native Win32 locale: */
+#if defined(_WIN32) && \
+    !defined(BOOST_REGEX_USE_WIN32_LOCALE) && \
+    !defined(BOOST_REGEX_USE_C_LOCALE) && \
+    !defined(BOOST_REGEX_USE_CPP_LOCALE) && \
+    !defined(BOOST_REGEX_NO_W32) && \
+    !defined(BOOST_REGEX_NO_WIN32_LOCALE)
+#  define BOOST_REGEX_USE_WIN32_LOCALE
+#endif
+/* otherwise use C++ locale if supported: */
+#if !defined(BOOST_REGEX_USE_WIN32_LOCALE) && !defined(BOOST_REGEX_USE_C_LOCALE) && !defined(BOOST_REGEX_USE_CPP_LOCALE) && !defined(BOOST_NO_STD_LOCALE)
+#  define BOOST_REGEX_USE_CPP_LOCALE
+#endif
+/* otherwise use C locale: */
+#if !defined(BOOST_REGEX_USE_WIN32_LOCALE) && !defined(BOOST_REGEX_USE_C_LOCALE) && !defined(BOOST_REGEX_USE_CPP_LOCALE)
+#  define BOOST_REGEX_USE_C_LOCALE
+#endif
+
+#ifndef BOOST_REGEX_MAX_STATE_COUNT
+#  define BOOST_REGEX_MAX_STATE_COUNT 100000000
+#endif
+
+
+/*****************************************************************************
+ *
+ *  Error Handling for exception free compilers:
+ *
+ ****************************************************************************/
+
+#ifdef BOOST_NO_EXCEPTIONS
+/*
+ * If there are no exceptions then we must report critical-errors
+ * the only way we know how; by terminating.
+ */
+#include <stdexcept>
+#include <string>
+#include <boost/throw_exception.hpp>
+
+#  define BOOST_REGEX_NOEH_ASSERT(x)\
+if(0 == (x))\
+{\
+   std::string s("Error: critical regex++ failure in: ");\
+   s.append(#x);\
+   std::runtime_error e(s);\
+   boost::throw_exception(e);\
+}
+#else
+/*
+ * With exceptions then error handling is taken care of and
+ * there is no need for these checks:
+ */
+#  define BOOST_REGEX_NOEH_ASSERT(x)
+#endif
+
+
+/*****************************************************************************
+ *
+ *  Stack protection under MS Windows:
+ *
+ ****************************************************************************/
+
+#if !defined(BOOST_REGEX_NO_W32) && !defined(BOOST_REGEX_V3)
+#  if(defined(_WIN32) || defined(_WIN64) || defined(_WINCE)) \
+        && !(defined(__GNUC__) || defined(__BORLANDC__) && defined(__clang__)) \
+        && !(defined(__BORLANDC__) && (__BORLANDC__ >= 0x600)) \
+        && !(defined(__MWERKS__) && (__MWERKS__ <= 0x3003))
+#     define BOOST_REGEX_HAS_MS_STACK_GUARD
+#  endif
+#elif defined(BOOST_REGEX_HAS_MS_STACK_GUARD)
+#  undef BOOST_REGEX_HAS_MS_STACK_GUARD
+#endif
+
+#if defined(__cplusplus) && defined(BOOST_REGEX_HAS_MS_STACK_GUARD)
+
+namespace boost{
+namespace BOOST_REGEX_DETAIL_NS{
+
+BOOST_REGEX_DECL void BOOST_REGEX_CALL reset_stack_guard_page();
+
+}
+}
+
+#endif
+
+
+/*****************************************************************************
+ *
+ *  Algorithm selection and configuration.
+ *  These options are now obsolete for C++11 and later (regex v5).
+ *
+ ****************************************************************************/
+
+#if !defined(BOOST_REGEX_RECURSIVE) && !defined(BOOST_REGEX_NON_RECURSIVE)
+#  if defined(BOOST_REGEX_HAS_MS_STACK_GUARD) && !defined(_STLP_DEBUG) && !defined(__STL_DEBUG) && !(defined(_MSC_VER) && (_MSC_VER >= 1400)) && defined(BOOST_REGEX_CXX03)
+#     define BOOST_REGEX_RECURSIVE
+#  else
+#     define BOOST_REGEX_NON_RECURSIVE
+#  endif
+#endif
+
+#ifdef BOOST_REGEX_NON_RECURSIVE
+#  ifdef BOOST_REGEX_RECURSIVE
+#     error "Can't set both BOOST_REGEX_RECURSIVE and BOOST_REGEX_NON_RECURSIVE"
+#  endif
+#  ifndef BOOST_REGEX_BLOCKSIZE
+#     define BOOST_REGEX_BLOCKSIZE 4096
+#  endif
+#  if BOOST_REGEX_BLOCKSIZE < 512
+#     error "BOOST_REGEX_BLOCKSIZE must be at least 512"
+#  endif
+#  ifndef BOOST_REGEX_MAX_BLOCKS
+#     define BOOST_REGEX_MAX_BLOCKS 1024
+#  endif
+#  ifdef BOOST_REGEX_HAS_MS_STACK_GUARD
+#     undef BOOST_REGEX_HAS_MS_STACK_GUARD
+#  endif
+#  ifndef BOOST_REGEX_MAX_CACHE_BLOCKS
+#     define BOOST_REGEX_MAX_CACHE_BLOCKS 16
+#  endif
+#endif
+
+
+/*****************************************************************************
+ *
+ *  Diagnostics:
+ *
+ ****************************************************************************/
+
+#ifdef BOOST_REGEX_CONFIG_INFO
+BOOST_REGEX_DECL void BOOST_REGEX_CALL print_regex_library_info();
+#endif
+
+#if defined(BOOST_REGEX_DIAG)
+#  pragma message ("BOOST_REGEX_DECL" BOOST_STRINGIZE(=BOOST_REGEX_DECL))
+#  pragma message ("BOOST_REGEX_CALL" BOOST_STRINGIZE(=BOOST_REGEX_CALL))
+#  pragma message ("BOOST_REGEX_CCALL" BOOST_STRINGIZE(=BOOST_REGEX_CCALL))
+#ifdef BOOST_REGEX_USE_C_LOCALE
+#  pragma message ("Using C locale in regex traits class")
+#elif BOOST_REGEX_USE_CPP_LOCALE
+#  pragma message ("Using C++ locale in regex traits class")
+#else
+#  pragma message ("Using Win32 locale in regex traits class")
+#endif
+#if defined(BOOST_REGEX_DYN_LINK) || defined(BOOST_ALL_DYN_LINK)
+#  pragma message ("Dynamic linking enabled")
+#endif
+#if defined(BOOST_REGEX_NO_LIB) || defined(BOOST_ALL_NO_LIB)
+#  pragma message ("Auto-linking disabled")
+#endif
+#ifdef BOOST_REGEX_NO_EXTERNAL_TEMPLATES
+#  pragma message ("Extern templates disabled")
+#endif
+
+#endif
+
+#endif
+

regex/include/boost/regex/config/borland.hpp

@@ -0,0 +1,72 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         boost/regex/config/borland.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: regex borland-specific config setup.
+  */
+
+
+#if defined(__BORLANDC__) && !defined(__clang__)
+#  if (__BORLANDC__ == 0x550) || (__BORLANDC__ == 0x551)
+      // problems with std::basic_string and dll RTL:
+#     if defined(_RTLDLL) && defined(_RWSTD_COMPILE_INSTANTIATE)
+#        ifdef BOOST_REGEX_BUILD_DLL
+#           error _RWSTD_COMPILE_INSTANTIATE must not be defined when building regex++ as a DLL
+#        else
+#           pragma message("Defining _RWSTD_COMPILE_INSTANTIATE when linking to the DLL version of the RTL may produce memory corruption problems in std::basic_string, as a result of separate versions of basic_string's static data in the RTL and you're exe/dll: be warned!!")
+#        endif
+#     endif
+#     ifndef _RTLDLL
+         // this is harmless for a staic link:
+#        define _RWSTD_COMPILE_INSTANTIATE
+#     endif
+      // external templates cause problems for some reason:
+#     define BOOST_REGEX_NO_EXTERNAL_TEMPLATES
+#  endif
+#  if (__BORLANDC__ <= 0x540) && !defined(BOOST_REGEX_NO_LIB) && !defined(_NO_VCL)
+      // C++ Builder 4 and earlier, we can't tell whether we should be using
+      // the VCL runtime or not, do a static link instead:
+#     define BOOST_REGEX_STATIC_LINK
+#  endif
+   //
+   // VCL support:
+   // if we're building a console app then there can't be any VCL (can there?)
+#  if !defined(__CONSOLE__) && !defined(_NO_VCL)
+#     define BOOST_REGEX_USE_VCL
+#  endif
+   //
+   // if this isn't Win32 then don't automatically select link
+   // libraries:
+   //
+#  ifndef _Windows
+#     ifndef BOOST_REGEX_NO_LIB
+#        define BOOST_REGEX_NO_LIB
+#     endif
+#     ifndef BOOST_REGEX_STATIC_LINK
+#        define BOOST_REGEX_STATIC_LINK
+#     endif
+#  endif
+
+#if __BORLANDC__ < 0x600
+//
+// string workarounds:
+//
+#include <cstring>
+#undef strcmp
+#undef strcpy
+#endif
+
+#endif
+
+

regex/include/boost/regex/config/cwchar.hpp

@@ -0,0 +1,207 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         boost/regex/config/cwchar.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: regex wide character string fixes.
+  */
+
+#ifndef BOOST_REGEX_CONFIG_CWCHAR_HPP
+#define BOOST_REGEX_CONFIG_CWCHAR_HPP
+
+#include <cwchar>
+#include <cwctype>
+#include <boost/config.hpp>
+
+#if defined(__STD_RWCOMPILER_H__) || defined(_RWSTD_VER)
+// apparently this is required for the RW STL on Linux:
+#undef iswalnum
+#undef iswalpha
+#undef iswblank
+#undef iswcntrl
+#undef iswdigit
+#undef iswgraph
+#undef iswlower
+#undef iswprint
+#undef iswprint
+#undef iswpunct
+#undef iswspace
+#undef iswupper
+#undef iswxdigit
+#undef iswctype
+#undef towlower
+#undef towupper
+#undef towctrans
+#undef wctrans
+#undef wctype
+#endif
+
+namespace std{
+
+#ifndef BOOST_NO_STDC_NAMESPACE
+extern "C"{
+#endif
+
+#ifdef iswalnum
+inline int (iswalnum)(wint_t i)
+{ return iswalnum(i); }
+#undef iswalnum
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswalnum;
+#endif
+
+#ifdef iswalpha
+inline int (iswalpha)(wint_t i)
+{ return iswalpha(i); }
+#undef iswalpha
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswalpha;
+#endif
+
+#ifdef iswcntrl
+inline int (iswcntrl)(wint_t i)
+{ return iswcntrl(i); }
+#undef iswcntrl
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswcntrl;
+#endif
+
+#ifdef iswdigit
+inline int (iswdigit)(wint_t i)
+{ return iswdigit(i); }
+#undef iswdigit
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswdigit;
+#endif
+
+#ifdef iswgraph
+inline int (iswgraph)(wint_t i)
+{ return iswgraph(i); }
+#undef iswgraph
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswgraph;
+#endif
+
+#ifdef iswlower
+inline int (iswlower)(wint_t i)
+{ return iswlower(i); }
+#undef iswlower
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswlower;
+#endif
+
+#ifdef iswprint
+inline int (iswprint)(wint_t i)
+{ return iswprint(i); }
+#undef iswprint
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswprint;
+#endif
+
+#ifdef iswpunct
+inline int (iswpunct)(wint_t i)
+{ return iswpunct(i); }
+#undef iswpunct
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswpunct;
+#endif
+
+#ifdef iswspace
+inline int (iswspace)(wint_t i)
+{ return iswspace(i); }
+#undef iswspace
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswspace;
+#endif
+
+#ifdef iswupper
+inline int (iswupper)(wint_t i)
+{ return iswupper(i); }
+#undef iswupper
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswupper;
+#endif
+
+#ifdef iswxdigit
+inline int (iswxdigit)(wint_t i)
+{ return iswxdigit(i); }
+#undef iswxdigit
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::iswxdigit;
+#endif
+
+#ifdef towlower
+inline wint_t (towlower)(wint_t i)
+{ return towlower(i); }
+#undef towlower
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::towlower;
+#endif
+
+#ifdef towupper
+inline wint_t (towupper)(wint_t i)
+{ return towupper(i); }
+#undef towupper
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using :: towupper;
+#endif
+
+#ifdef wcscmp
+inline int (wcscmp)(const wchar_t *p1, const wchar_t *p2)
+{ return wcscmp(p1,p2); }
+#undef wcscmp
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::wcscmp;
+#endif
+
+#ifdef wcscoll
+inline int (wcscoll)(const wchar_t *p1, const wchar_t *p2)
+{ return wcscoll(p1,p2); }
+#undef wcscoll
+#elif defined(BOOST_NO_STDC_NAMESPACE) && !defined(UNDER_CE)
+using ::wcscoll;
+#endif
+
+#ifdef wcscpy
+inline wchar_t *(wcscpy)(wchar_t *p1, const wchar_t *p2)
+{ return wcscpy(p1,p2); }
+#undef wcscpy
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::wcscpy;
+#endif
+
+#ifdef wcslen
+inline size_t (wcslen)(const wchar_t *p)
+{ return wcslen(p); }
+#undef wcslen
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::wcslen;
+#endif
+
+#ifdef wcsxfrm
+size_t wcsxfrm(wchar_t *p1, const wchar_t *p2, size_t s)
+{ return wcsxfrm(p1,p2,s); }
+#undef wcsxfrm
+#elif defined(BOOST_NO_STDC_NAMESPACE)
+using ::wcsxfrm;
+#endif
+
+
+#ifndef BOOST_NO_STDC_NAMESPACE
+} // extern "C"
+#endif
+
+} // namespace std
+
+#endif
+

regex/include/boost/regex/icu.hpp

@@ -0,0 +1,30 @@
+/*
+ *
+ * Copyright (c) 2020
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         icu.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Unicode regular expressions on top of the ICU Library.
+  */
+
+#ifndef BOOST_REGEX_ICU_HPP
+#define BOOST_REGEX_ICU_HPP
+
+#include <boost/regex/config.hpp>
+
+#ifdef BOOST_REGEX_CXX03
+#include <boost/regex/v4/icu.hpp>
+#else
+#include <boost/regex/v5/icu.hpp>
+#endif
+
+#endif

regex/include/boost/regex/mfc.hpp

@@ -0,0 +1,186 @@
+/*
+ *
+ * Copyright (c) 2004
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         mfc.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Overloads and helpers for using MFC/ATL string types with Boost.Regex.
+  */
+
+#ifndef BOOST_REGEX_MFC_HPP
+#define BOOST_REGEX_MFC_HPP
+
+#include <atlsimpstr.h>
+#include <boost/regex.hpp>
+
+namespace boost{
+
+//
+// define the types used for TCHAR's:
+typedef basic_regex<TCHAR> tregex;
+typedef match_results<TCHAR const*> tmatch;
+typedef regex_iterator<TCHAR const*> tregex_iterator;
+typedef regex_token_iterator<TCHAR const*> tregex_token_iterator;
+
+// Obsolete. Remove
+#define SIMPLE_STRING_PARAM class B, bool b
+#define SIMPLE_STRING_ARG_LIST B, b
+
+//
+// define regex creation functions:
+//
+template <class B, bool b>
+inline basic_regex<B> 
+make_regex(const ATL::CSimpleStringT<B, b>& s, ::boost::regex_constants::syntax_option_type f = boost::regex_constants::normal)
+{
+   basic_regex<B> result(s.GetString(), s.GetString() + s.GetLength(), f);
+   return result;
+}
+//
+// regex_match overloads:
+//
+template <class B, bool b, class A, class T>
+inline bool regex_match(const ATL::CSimpleStringT<B, b>& s,
+                 match_results<const B*, A>& what,
+                 const basic_regex<B, T>& e,
+                 boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   return ::boost::regex_match(s.GetString(),
+                               s.GetString() + s.GetLength(),
+                               what,
+                               e,
+                               f);
+}
+
+template <class B, bool b, class T>
+inline bool regex_match(const ATL::CSimpleStringT<B, b>& s,
+                 const basic_regex<B, T>& e,
+                 boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   return ::boost::regex_match(s.GetString(),
+                               s.GetString() + s.GetLength(),
+                               e,
+                               f);
+}
+//
+// regex_search overloads:
+//
+template <class B, bool b, class A, class T>
+inline bool regex_search(const ATL::CSimpleStringT<B, b>& s,
+                 match_results<const B*, A>& what,
+                 const basic_regex<B, T>& e,
+                 boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   return ::boost::regex_search(s.GetString(),
+                               s.GetString() + s.GetLength(),
+                               what,
+                               e,
+                               f);
+}
+
+template <class B, bool b, class T>
+inline bool regex_search(const ATL::CSimpleStringT<B, b>& s,
+                 const basic_regex<B, T>& e,
+                 boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   return ::boost::regex_search(s.GetString(),
+                               s.GetString() + s.GetLength(),
+                               e,
+                               f);
+}
+//
+// regex_iterator creation:
+//
+template <class B, bool b>
+inline regex_iterator<B const*> 
+make_regex_iterator(const ATL::CSimpleStringT<B, b>& s, const basic_regex<B>& e, ::boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   regex_iterator<B const*> result(s.GetString(), s.GetString() + s.GetLength(), e, f);
+   return result;
+}
+
+template <class B, bool b>
+inline regex_token_iterator<B const*> 
+   make_regex_token_iterator(const ATL::CSimpleStringT<B, b>& s, const basic_regex<B>& e, int sub = 0, ::boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   regex_token_iterator<B const*> result(s.GetString(), s.GetString() + s.GetLength(), e, sub, f);
+   return result;
+}
+
+template <class B, bool b>
+inline regex_token_iterator<B const*> 
+make_regex_token_iterator(const ATL::CSimpleStringT<B, b>& s, const basic_regex<B>& e, const std::vector<int>& subs, ::boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   regex_token_iterator<B const*> result(s.GetString(), s.GetString() + s.GetLength(), e, subs, f);
+   return result;
+}
+
+template <class B, bool b, std::size_t N>
+inline regex_token_iterator<B const*> 
+make_regex_token_iterator(const ATL::CSimpleStringT<B, b>& s, const basic_regex<B>& e, const int (& subs)[N], ::boost::regex_constants::match_flag_type f = boost::regex_constants::match_default)
+{
+   regex_token_iterator<B const*> result(s.GetString(), s.GetString() + s.GetLength(), e, subs, f);
+   return result;
+}
+
+template <class OutputIterator, class BidirectionalIterator, class traits,
+          class B, bool b>
+OutputIterator regex_replace(OutputIterator out,
+                           BidirectionalIterator first,
+                           BidirectionalIterator last,
+                           const basic_regex<B, traits>& e,
+                           const ATL::CSimpleStringT<B, b>& fmt,
+                           match_flag_type flags = match_default)
+{
+   return ::boost::regex_replace(out, first, last, e, fmt.GetString(), flags);
+}
+
+namespace BOOST_REGEX_DETAIL_NS{
+
+template <class B, bool b>
+class mfc_string_out_iterator
+{
+   ATL::CSimpleStringT<B, b>* out;
+public:
+   mfc_string_out_iterator(ATL::CSimpleStringT<B, b>& s) : out(&s) {}
+   mfc_string_out_iterator& operator++() { return *this; }
+   mfc_string_out_iterator& operator++(int) { return *this; }
+   mfc_string_out_iterator& operator*() { return *this; }
+   mfc_string_out_iterator& operator=(B v) 
+   { 
+      out->AppendChar(v); 
+      return *this; 
+   }
+   typedef std::ptrdiff_t difference_type;
+   typedef B value_type;
+   typedef value_type* pointer;
+   typedef value_type& reference;
+   typedef std::output_iterator_tag iterator_category;
+};
+
+}
+
+template <class traits, class B, bool b>
+ATL::CSimpleStringT<B, b> regex_replace(const ATL::CSimpleStringT<B, b>& s,
+                            const basic_regex<B, traits>& e,
+                            const ATL::CSimpleStringT<B, b>& fmt,
+                            match_flag_type flags = match_default)
+{
+   ATL::CSimpleStringT<B, b> result(s.GetManager());
+   BOOST_REGEX_DETAIL_NS::mfc_string_out_iterator<B, b> i(result);
+   regex_replace(i, s.GetString(), s.GetString() + s.GetLength(), e, fmt.GetString(), flags);
+   return result;
+}
+
+} // namespace boost.
+
+#endif

regex/include/boost/regex/pattern_except.hpp

@@ -0,0 +1,32 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         pattern_except.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares pattern-matching exception classes.
+  */
+
+#ifndef BOOST_RE_PAT_EXCEPT_HPP
+#define BOOST_RE_PAT_EXCEPT_HPP
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+
+#ifdef BOOST_REGEX_CXX03
+#include <boost/regex/v4/pattern_except.hpp>
+#else
+#include <boost/regex/v5/pattern_except.hpp>
+#endif
+
+#endif

regex/include/boost/regex/pending/object_cache.hpp

@@ -0,0 +1,29 @@
+/*
+ *
+ * Copyright (c) 2004
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         object_cache.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Implements a generic object cache.
+  */
+
+#ifndef BOOST_REGEX_OBJECT_CACHE_HPP
+#define BOOST_REGEX_OBJECT_CACHE_HPP
+
+#include <boost/regex/config.hpp>
+#ifdef BOOST_REGEX_CXX03
+#include <boost/regex/v4/object_cache.hpp>
+#else
+#include <boost/regex/v5/object_cache.hpp>
+#endif
+
+#endif

regex/include/boost/regex/pending/static_mutex.hpp

@@ -0,0 +1,182 @@
+/*
+ *
+ * Copyright (c) 2004
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         static_mutex.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares static_mutex lock type, there are three different
+  *                implementations: POSIX pthreads, WIN32 threads, and portable,
+  *                these are described in more detail below.
+  */
+
+#ifndef BOOST_REGEX_STATIC_MUTEX_HPP
+#define BOOST_REGEX_STATIC_MUTEX_HPP
+
+#include <boost/config.hpp>
+#include <boost/regex/config.hpp> // dll import/export options.
+
+#ifdef BOOST_HAS_PTHREADS
+#include <pthread.h>
+#endif
+
+#if defined(BOOST_HAS_PTHREADS) && defined(PTHREAD_MUTEX_INITIALIZER)
+//
+// pthreads version:
+// simple wrap around a pthread_mutex_t initialized with
+// PTHREAD_MUTEX_INITIALIZER.
+//
+namespace boost{
+
+class static_mutex;
+
+#define BOOST_STATIC_MUTEX_INIT { PTHREAD_MUTEX_INITIALIZER, }
+
+class BOOST_REGEX_DECL scoped_static_mutex_lock
+{
+public:
+   scoped_static_mutex_lock(static_mutex& mut, bool lk = true);
+   ~scoped_static_mutex_lock();
+   inline bool locked()const
+   {
+      return m_have_lock;
+   }
+   inline operator void const*()const
+   {
+      return locked() ? this : 0;
+   }
+   void lock();
+   void unlock();
+private:
+   static_mutex& m_mutex;
+   bool m_have_lock;
+};
+
+class static_mutex
+{
+public:
+   typedef scoped_static_mutex_lock scoped_lock;
+   pthread_mutex_t m_mutex;
+};
+
+} // namespace boost
+#elif defined(BOOST_HAS_WINTHREADS)
+//
+// Win32 version:
+// Use a 32-bit int as a lock, along with a test-and-set
+// implementation using InterlockedCompareExchange.
+//
+
+#include <boost/cstdint.hpp>
+
+namespace boost{
+
+class BOOST_REGEX_DECL scoped_static_mutex_lock;
+
+class static_mutex
+{
+public:
+   typedef scoped_static_mutex_lock scoped_lock;
+   boost::int32_t m_mutex;
+};
+
+#define BOOST_STATIC_MUTEX_INIT { 0, }
+
+class BOOST_REGEX_DECL scoped_static_mutex_lock
+{
+public:
+   scoped_static_mutex_lock(static_mutex& mut, bool lk = true);
+   ~scoped_static_mutex_lock();
+   operator void const*()const
+   {
+      return locked() ? this : 0;
+   }
+   bool locked()const
+   {
+      return m_have_lock;
+   }
+   void lock();
+   void unlock();
+private:
+   static_mutex& m_mutex;
+   bool m_have_lock;
+   scoped_static_mutex_lock(const scoped_static_mutex_lock&);
+   scoped_static_mutex_lock& operator=(const scoped_static_mutex_lock&);
+};
+
+} // namespace
+
+#else
+//
+// Portable version of a static mutex based on Boost.Thread library:
+// This has to use a single mutex shared by all instances of static_mutex
+// because boost::call_once doesn't alow us to pass instance information
+// down to the initialisation proceedure.  In fact the initialisation routine
+// may need to be called more than once - but only once per instance.
+//
+// Since this preprocessor path is almost never taken, we hide these header
+// dependencies so that build tools don't find them.
+//
+#define BOOST_REGEX_H1 <boost/thread/once.hpp>
+#define BOOST_REGEX_H2 <boost/thread/recursive_mutex.hpp>
+#define BOOST_REGEX_H3 <boost/thread/lock_types.hpp>
+#include BOOST_REGEX_H1
+#include BOOST_REGEX_H2
+#include BOOST_REGEX_H3
+#undef BOOST_REGEX_H1
+#undef BOOST_REGEX_H2
+#undef BOOST_REGEX_H3
+
+namespace boost{
+
+class BOOST_REGEX_DECL scoped_static_mutex_lock;
+extern "C" BOOST_REGEX_DECL void boost_regex_free_static_mutex();
+
+class BOOST_REGEX_DECL static_mutex
+{
+public:
+   typedef scoped_static_mutex_lock scoped_lock;
+   static void init();
+   static boost::recursive_mutex* m_pmutex;
+   static boost::once_flag m_once;
+};
+
+#define BOOST_STATIC_MUTEX_INIT {  }
+
+class BOOST_REGEX_DECL scoped_static_mutex_lock
+{
+public:
+   scoped_static_mutex_lock(static_mutex& mut, bool lk = true);
+   ~scoped_static_mutex_lock();
+   operator void const*()const;
+   bool locked()const;
+   void lock();
+   void unlock();
+private:
+   boost::unique_lock<boost::recursive_mutex>* m_plock;
+   bool m_have_lock;
+};
+
+inline scoped_static_mutex_lock::operator void const*()const
+{
+   return locked() ? this : 0;
+}
+
+inline bool scoped_static_mutex_lock::locked()const
+{
+   return m_have_lock;
+}
+
+} // namespace
+
+#endif
+
+#endif

regex/include/boost/regex/pending/unicode_iterator.hpp

@@ -0,0 +1,32 @@
+/*
+ *
+ * Copyright (c) 2020
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         unicode_iterator.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Iterator adapters for converting between different Unicode encodings.
+  */
+
+#ifndef BOOST_REGEX_PENDING_UNICODE_ITERATOR_HPP
+#define BOOST_REGEX_PENDING_UNICODE_ITERATOR_HPP
+
+#include <boost/regex/config.hpp>
+
+#if defined(BOOST_REGEX_CXX03)
+#include <boost/regex/v4/unicode_iterator.hpp>
+#else
+#include <boost/regex/v5/unicode_iterator.hpp>
+#endif
+
+
+#endif // BOOST_REGEX_PENDING_UNICODE_ITERATOR_HPP
+

regex/include/boost/regex/regex_traits.hpp

@@ -0,0 +1,39 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_traits.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares regular expression traits classes.
+  */
+
+#ifndef BOOST_REGEX_TRAITS_HPP
+#define BOOST_REGEX_TRAITS_HPP
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#  include <boost/regex/config.hpp>
+#endif
+
+#  ifndef BOOST_REGEX_TRAITS_HPP_INCLUDED
+#ifdef BOOST_REGEX_CXX03
+#     include <boost/regex/v4/regex_traits.hpp>
+#else
+#     include <boost/regex/v5/regex_traits.hpp>
+#endif
+#  endif
+
+#endif // include
+
+
+
+
+

regex/include/boost/regex/user.hpp

@@ -0,0 +1,95 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         user.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: User settable options.
+  */
+
+// define if you want the regex library to use the C locale
+// even on Win32:
+// #define BOOST_REGEX_USE_C_LOCALE
+
+// define this is you want the regex library to use the C++
+// locale:
+// #define BOOST_REGEX_USE_CPP_LOCALE
+
+// define this if the runtime library is a dll, and you
+// want BOOST_REGEX_DYN_LINK to set up dll exports/imports
+// with __declspec(dllexport)/__declspec(dllimport.)
+// #define BOOST_REGEX_HAS_DLL_RUNTIME
+
+// define this if you want to dynamically link to regex,
+// if the runtime library is also a dll (Probably Win32 specific,
+// and has no effect unless BOOST_REGEX_HAS_DLL_RUNTIME is set):
+// #define BOOST_REGEX_DYN_LINK
+
+// define this if you don't want the lib to automatically
+// select its link libraries:
+// #define BOOST_REGEX_NO_LIB
+
+// define this if templates with switch statements cause problems:
+// #define BOOST_REGEX_NO_TEMPLATE_SWITCH_MERGE
+ 
+// define this to disable Win32 support when available:
+// #define BOOST_REGEX_NO_W32
+
+// define this if bool is not a real type:
+// #define BOOST_REGEX_NO_BOOL
+
+// define this if no template instances are to be placed in
+// the library rather than users object files:
+// #define BOOST_REGEX_NO_EXTERNAL_TEMPLATES
+
+// define this if the forward declarations in regex_fwd.hpp
+// cause more problems than they are worth:
+// #define BOOST_REGEX_NO_FWD
+
+// define this if your compiler supports MS Windows structured
+// exception handling.
+// #define BOOST_REGEX_HAS_MS_STACK_GUARD
+
+// define this if you want to use the recursive algorithm
+// even if BOOST_REGEX_HAS_MS_STACK_GUARD is not defined.
+// NOTE: OBSOLETE!!
+// #define BOOST_REGEX_RECURSIVE
+
+// define this if you want to use the non-recursive
+// algorithm, even if the recursive version would be the default.
+// NOTE: OBSOLETE!!
+// #define BOOST_REGEX_NON_RECURSIVE
+
+// define this if you want to set the size of the memory blocks
+// used by the non-recursive algorithm.
+// #define BOOST_REGEX_BLOCKSIZE 4096
+
+// define this if you want to set the maximum number of memory blocks
+// used by the non-recursive algorithm.
+// #define BOOST_REGEX_MAX_BLOCKS 1024
+
+// define this if you want to set the maximum number of memory blocks
+// cached by the non-recursive algorithm: Normally this is 16, but can be 
+// higher if you have multiple threads all using boost.regex, or lower 
+// if you don't want boost.regex to cache memory.
+// #define BOOST_REGEX_MAX_CACHE_BLOCKS 16
+
+// define this if you want to be able to access extended capture
+// information in your sub_match's (caution this will slow things
+// down quite a bit).
+// #define BOOST_REGEX_MATCH_EXTRA
+
+// define this if you want to enable support for Unicode via ICU.
+// #define BOOST_HAS_ICU
+
+// define this if you want regex to use __cdecl calling convensions, even when __fastcall is available:
+// #define BOOST_REGEX_NO_FASTCALL

regex/include/boost/regex/v5/basic_regex.hpp

@@ -0,0 +1,734 @@
+/*
+ *
+ * Copyright (c) 1998-2004 John Maddock
+ * Copyright 2011 Garmin Ltd. or its subsidiaries
+ *
+ * Distributed under the Boost Software License, Version 1.0.
+ * (See accompanying file LICENSE_1_0.txt or copy at
+ * http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org/ for most recent version.
+  *   FILE         basic_regex.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares template class basic_regex.
+  */
+
+#ifndef BOOST_REGEX_V5_BASIC_REGEX_HPP
+#define BOOST_REGEX_V5_BASIC_REGEX_HPP
+
+#include <vector>
+
+namespace boost{
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable : 4251)
+#if BOOST_REGEX_MSVC < 1700
+#     pragma warning(disable : 4231)
+#endif
+#if BOOST_REGEX_MSVC < 1600
+#pragma warning(disable : 4660)
+#endif
+#if BOOST_REGEX_MSVC < 1910
+#pragma warning(disable:4800)
+#endif
+#endif
+
+namespace BOOST_REGEX_DETAIL_NS{
+
+//
+// forward declaration, we will need this one later:
+//
+template <class charT, class traits>
+class basic_regex_parser;
+
+template <class I>
+void bubble_down_one(I first, I last)
+{
+   if(first != last)
+   {
+      I next = last - 1;
+      while((next != first) && (*next < *(next-1)))
+      {
+         (next-1)->swap(*next);
+         --next;
+      }
+   }
+}
+
+static const int hash_value_mask = 1 << (std::numeric_limits<int>::digits - 1);
+
+template <class Iterator>
+inline int hash_value_from_capture_name(Iterator i, Iterator j)
+{
+   std::size_t r = 0;
+   while (i != j)
+   {
+      r ^= *i + 0x9e3779b9 + (r << 6) + (r >> 2);
+      ++i;
+   }
+   r %= ((std::numeric_limits<int>::max)());
+   return static_cast<int>(r) | hash_value_mask;
+}
+
+class named_subexpressions
+{
+public:
+   struct name
+   {
+      template <class charT>
+      name(const charT* i, const charT* j, int idx)
+         : index(idx) 
+      { 
+         hash = hash_value_from_capture_name(i, j); 
+      }
+      name(int h, int idx)
+         : index(idx), hash(h)
+      { 
+      }
+      int index;
+      int hash;
+      bool operator < (const name& other)const
+      {
+         return hash < other.hash;
+      }
+      bool operator == (const name& other)const
+      {
+         return hash == other.hash; 
+      }
+      void swap(name& other)
+      {
+         std::swap(index, other.index);
+         std::swap(hash, other.hash);
+      }
+   };
+
+   typedef std::vector<name>::const_iterator const_iterator;
+   typedef std::pair<const_iterator, const_iterator> range_type;
+
+   named_subexpressions(){}
+
+   template <class charT>
+   void set_name(const charT* i, const charT* j, int index)
+   {
+      m_sub_names.push_back(name(i, j, index));
+      bubble_down_one(m_sub_names.begin(), m_sub_names.end());
+   }
+   template <class charT>
+   int get_id(const charT* i, const charT* j)const
+   {
+      name t(i, j, 0);
+      typename std::vector<name>::const_iterator pos = std::lower_bound(m_sub_names.begin(), m_sub_names.end(), t);
+      if((pos != m_sub_names.end()) && (*pos == t))
+      {
+         return pos->index;
+      }
+      return -1;
+   }
+   template <class charT>
+   range_type equal_range(const charT* i, const charT* j)const
+   {
+      name t(i, j, 0);
+      return std::equal_range(m_sub_names.begin(), m_sub_names.end(), t);
+   }
+   int get_id(int h)const
+   {
+      name t(h, 0);
+      std::vector<name>::const_iterator pos = std::lower_bound(m_sub_names.begin(), m_sub_names.end(), t);
+      if((pos != m_sub_names.end()) && (*pos == t))
+      {
+         return pos->index;
+      }
+      return -1;
+   }
+   range_type equal_range(int h)const
+   {
+      name t(h, 0);
+      return std::equal_range(m_sub_names.begin(), m_sub_names.end(), t);
+   }
+private:
+   std::vector<name> m_sub_names;
+};
+
+//
+// class regex_data:
+// represents the data we wish to expose to the matching algorithms.
+//
+template <class charT, class traits>
+struct regex_data : public named_subexpressions
+{
+   typedef regex_constants::syntax_option_type   flag_type;
+   typedef std::size_t                           size_type;  
+
+   regex_data(const ::std::shared_ptr<
+      ::boost::regex_traits_wrapper<traits> >& t) 
+      : m_ptraits(t), m_flags(0), m_status(0), m_expression(0), m_expression_len(0),
+         m_mark_count(0), m_first_state(0), m_restart_type(0),
+         m_startmap{ 0 },
+         m_can_be_null(0), m_word_mask(0), m_has_recursions(false), m_disable_match_any(false) {}
+   regex_data() 
+      : m_ptraits(new ::boost::regex_traits_wrapper<traits>()), m_flags(0), m_status(0), m_expression(0), m_expression_len(0), 
+         m_mark_count(0), m_first_state(0), m_restart_type(0), 
+      m_startmap{ 0 },
+         m_can_be_null(0), m_word_mask(0), m_has_recursions(false), m_disable_match_any(false) {}
+
+   ::std::shared_ptr<
+      ::boost::regex_traits_wrapper<traits>
+      >                        m_ptraits;                 // traits class instance
+   flag_type                   m_flags;                   // flags with which we were compiled
+   int                         m_status;                  // error code (0 implies OK).
+   const charT*                m_expression;              // the original expression
+   std::ptrdiff_t              m_expression_len;          // the length of the original expression
+   size_type                   m_mark_count;              // the number of marked sub-expressions
+   BOOST_REGEX_DETAIL_NS::re_syntax_base*  m_first_state;             // the first state of the machine
+   unsigned                    m_restart_type;            // search optimisation type
+   unsigned char               m_startmap[1 << CHAR_BIT]; // which characters can start a match
+   unsigned int                m_can_be_null;             // whether we can match a null string
+   BOOST_REGEX_DETAIL_NS::raw_storage      m_data;                    // the buffer in which our states are constructed
+   typename traits::char_class_type    m_word_mask;       // mask used to determine if a character is a word character
+   std::vector<
+      std::pair<
+      std::size_t, std::size_t> > m_subs;                 // Position of sub-expressions within the *string*.
+   bool                        m_has_recursions;          // whether we have recursive expressions;
+   bool                        m_disable_match_any;       // when set we need to disable the match_any flag as it causes different/buggy behaviour.
+};
+//
+// class basic_regex_implementation
+// pimpl implementation class for basic_regex.
+//
+template <class charT, class traits>
+class basic_regex_implementation
+   : public regex_data<charT, traits>
+{
+public:
+   typedef regex_constants::syntax_option_type   flag_type;
+   typedef std::ptrdiff_t                        difference_type;
+   typedef std::size_t                           size_type; 
+   typedef typename traits::locale_type          locale_type;
+   typedef const charT*                          const_iterator;
+
+   basic_regex_implementation(){}
+   basic_regex_implementation(const ::std::shared_ptr<
+      ::boost::regex_traits_wrapper<traits> >& t)
+      : regex_data<charT, traits>(t) {}
+   void assign(const charT* arg_first,
+                          const charT* arg_last,
+                          flag_type f)
+   {
+      regex_data<charT, traits>* pdat = this;
+      basic_regex_parser<charT, traits> parser(pdat);
+      parser.parse(arg_first, arg_last, f);
+   }
+
+   locale_type  imbue(locale_type l)
+   { 
+      return this->m_ptraits->imbue(l); 
+   }
+   locale_type  getloc()const
+   { 
+      return this->m_ptraits->getloc(); 
+   }
+   std::basic_string<charT>  str()const
+   {
+      std::basic_string<charT> result;
+      if(this->m_status == 0)
+         result = std::basic_string<charT>(this->m_expression, this->m_expression_len);
+      return result;
+   }
+   const_iterator  expression()const
+   {
+      return this->m_expression;
+   }
+   std::pair<const_iterator, const_iterator>  subexpression(std::size_t n)const
+   {
+      const std::pair<std::size_t, std::size_t>& pi = this->m_subs.at(n);
+      std::pair<const_iterator, const_iterator> p(expression() + pi.first, expression() + pi.second);
+      return p;
+   }
+   //
+   // begin, end:
+   const_iterator  begin()const
+   { 
+      return (this->m_status ? 0 : this->m_expression); 
+   }
+   const_iterator  end()const
+   { 
+      return (this->m_status ? 0 : this->m_expression + this->m_expression_len); 
+   }
+   flag_type  flags()const
+   {
+      return this->m_flags;
+   }
+   size_type  size()const
+   {
+      return this->m_expression_len;
+   }
+   int  status()const
+   {
+      return this->m_status;
+   }
+   size_type  mark_count()const
+   {
+      return this->m_mark_count - 1;
+   }
+   const BOOST_REGEX_DETAIL_NS::re_syntax_base* get_first_state()const
+   {
+      return this->m_first_state;
+   }
+   unsigned get_restart_type()const
+   {
+      return this->m_restart_type;
+   }
+   const unsigned char* get_map()const
+   {
+      return this->m_startmap;
+   }
+   const ::boost::regex_traits_wrapper<traits>& get_traits()const
+   {
+      return *(this->m_ptraits);
+   }
+   bool can_be_null()const
+   {
+      return this->m_can_be_null;
+   }
+   const regex_data<charT, traits>& get_data()const
+   {
+      basic_regex_implementation<charT, traits> const* p = this;
+      return *static_cast<const regex_data<charT, traits>*>(p);
+   }
+};
+
+} // namespace BOOST_REGEX_DETAIL_NS
+//
+// class basic_regex:
+// represents the compiled
+// regular expression:
+//
+
+#ifdef BOOST_REGEX_NO_FWD
+template <class charT, class traits = regex_traits<charT> >
+#else
+template <class charT, class traits >
+#endif
+class basic_regex : public regbase
+{
+public:
+   // typedefs:
+   typedef std::size_t                           traits_size_type;
+   typedef typename traits::string_type          traits_string_type;
+   typedef charT                                 char_type;
+   typedef traits                                traits_type;
+
+   typedef charT                                 value_type;
+   typedef charT&                                reference;
+   typedef const charT&                          const_reference;
+   typedef const charT*                          const_iterator;
+   typedef const_iterator                        iterator;
+   typedef std::ptrdiff_t                        difference_type;
+   typedef std::size_t                           size_type;   
+   typedef regex_constants::syntax_option_type   flag_type;
+   // locale_type
+   // placeholder for actual locale type used by the
+   // traits class to localise *this.
+   typedef typename traits::locale_type          locale_type;
+   
+public:
+   explicit basic_regex(){}
+   explicit basic_regex(const charT* p, flag_type f = regex_constants::normal)
+   {
+      assign(p, f);
+   }
+   basic_regex(const charT* p1, const charT* p2, flag_type f = regex_constants::normal)
+   {
+      assign(p1, p2, f);
+   }
+   basic_regex(const charT* p, size_type len, flag_type f)
+   {
+      assign(p, len, f);
+   }
+   basic_regex(const basic_regex& that)
+      : m_pimpl(that.m_pimpl) {}
+   ~basic_regex(){}
+   basic_regex&  operator=(const basic_regex& that)
+   {
+      return assign(that);
+   }
+   basic_regex&  operator=(const charT* ptr)
+   {
+      return assign(ptr);
+   }
+
+   //
+   // assign:
+   basic_regex& assign(const basic_regex& that)
+   { 
+      m_pimpl = that.m_pimpl;
+      return *this; 
+   }
+   basic_regex& assign(const charT* p, flag_type f = regex_constants::normal)
+   {
+      return assign(p, p + traits::length(p), f);
+   }
+   basic_regex& assign(const charT* p, size_type len, flag_type f)
+   {
+      return assign(p, p + len, f);
+   }
+private:
+   basic_regex& do_assign(const charT* p1,
+                          const charT* p2,
+                          flag_type f);
+public:
+   basic_regex& assign(const charT* p1,
+                          const charT* p2,
+                          flag_type f = regex_constants::normal)
+   {
+      return do_assign(p1, p2, f);
+   }
+
+   template <class ST, class SA>
+   unsigned int  set_expression(const std::basic_string<charT, ST, SA>& p, flag_type f = regex_constants::normal)
+   { 
+      return set_expression(p.data(), p.data() + p.size(), f); 
+   }
+
+   template <class ST, class SA>
+   explicit basic_regex(const std::basic_string<charT, ST, SA>& p, flag_type f = regex_constants::normal)
+   { 
+      assign(p, f); 
+   }
+
+   template <class InputIterator>
+   basic_regex(InputIterator arg_first, InputIterator arg_last, flag_type f = regex_constants::normal)
+   {
+      typedef typename traits::string_type seq_type;
+      seq_type a(arg_first, arg_last);
+      if(!a.empty())
+         assign(static_cast<const charT*>(&*a.begin()), static_cast<const charT*>(&*a.begin() + a.size()), f);
+      else
+         assign(static_cast<const charT*>(0), static_cast<const charT*>(0), f);
+   }
+
+   template <class ST, class SA>
+   basic_regex&  operator=(const std::basic_string<charT, ST, SA>& p)
+   {
+      return assign(p.data(), p.data() + p.size(), regex_constants::normal);
+   }
+
+   template <class string_traits, class A>
+   basic_regex&  assign(
+       const std::basic_string<charT, string_traits, A>& s,
+       flag_type f = regex_constants::normal)
+   {
+      return assign(s.data(), s.data() + s.size(), f);
+   }
+
+   template <class InputIterator>
+   basic_regex&  assign(InputIterator arg_first,
+                          InputIterator arg_last,
+                          flag_type f = regex_constants::normal)
+   {
+      typedef typename traits::string_type seq_type;
+      seq_type a(arg_first, arg_last);
+      if(a.size())
+      {
+         const charT* p1 = &*a.begin();
+         const charT* p2 = &*a.begin() + a.size();
+         return assign(p1, p2, f);
+      }
+      return assign(static_cast<const charT*>(0), static_cast<const charT*>(0), f);
+   }
+
+   //
+   // locale:
+   locale_type  imbue(locale_type l);
+   locale_type  getloc()const
+   { 
+      return m_pimpl.get() ? m_pimpl->getloc() : locale_type(); 
+   }
+   //
+   // getflags:
+   // retained for backwards compatibility only, "flags"
+   // is now the preferred name:
+   flag_type  getflags()const
+   { 
+      return flags();
+   }
+   flag_type  flags()const
+   { 
+      return m_pimpl.get() ? m_pimpl->flags() : 0;
+   }
+   //
+   // str:
+   std::basic_string<charT>  str()const
+   {
+      return m_pimpl.get() ? m_pimpl->str() : std::basic_string<charT>();
+   }
+   //
+   // begin, end, subexpression:
+   std::pair<const_iterator, const_iterator>  subexpression(std::size_t n)const
+   {
+#ifdef BOOST_REGEX_STANDALONE
+      if (!m_pimpl.get())
+         throw std::logic_error("Can't access subexpressions in an invalid regex.");
+#else
+      if(!m_pimpl.get())
+         boost::throw_exception(std::logic_error("Can't access subexpressions in an invalid regex."));
+#endif
+      return m_pimpl->subexpression(n);
+   }
+   const_iterator  begin()const
+   { 
+      return (m_pimpl.get() ? m_pimpl->begin() : 0); 
+   }
+   const_iterator  end()const
+   { 
+      return (m_pimpl.get() ? m_pimpl->end() : 0); 
+   }
+   //
+   // swap:
+   void  swap(basic_regex& that)throw()
+   {
+      m_pimpl.swap(that.m_pimpl);
+   }
+   //
+   // size:
+   size_type  size()const
+   { 
+      return (m_pimpl.get() ? m_pimpl->size() : 0); 
+   }
+   //
+   // max_size:
+   size_type  max_size()const
+   { 
+      return UINT_MAX; 
+   }
+   //
+   // empty:
+   bool  empty()const
+   { 
+      return (m_pimpl.get() ? 0 != m_pimpl->status() : true); 
+   }
+
+   size_type  mark_count()const 
+   { 
+      return (m_pimpl.get() ? m_pimpl->mark_count() : 0); 
+   }
+
+   int status()const
+   {
+      return (m_pimpl.get() ? m_pimpl->status() : regex_constants::error_empty);
+   }
+
+   int  compare(const basic_regex& that) const
+   {
+      if(m_pimpl.get() == that.m_pimpl.get())
+         return 0;
+      if(!m_pimpl.get())
+         return -1;
+      if(!that.m_pimpl.get())
+         return 1;
+      if(status() != that.status())
+         return status() - that.status();
+      if(flags() != that.flags())
+         return flags() - that.flags();
+      return str().compare(that.str());
+   }
+   bool  operator==(const basic_regex& e)const
+   { 
+      return compare(e) == 0; 
+   }
+   bool  operator != (const basic_regex& e)const
+   { 
+      return compare(e) != 0; 
+   }
+   bool  operator<(const basic_regex& e)const
+   { 
+      return compare(e) < 0; 
+   }
+   bool  operator>(const basic_regex& e)const
+   { 
+      return compare(e) > 0; 
+   }
+   bool  operator<=(const basic_regex& e)const
+   { 
+      return compare(e) <= 0; 
+   }
+   bool  operator>=(const basic_regex& e)const
+   { 
+      return compare(e) >= 0; 
+   }
+
+   //
+   // The following are deprecated as public interfaces
+   // but are available for compatibility with earlier versions.
+   const charT*  expression()const 
+   { 
+      return (m_pimpl.get() && !m_pimpl->status() ? m_pimpl->expression() : 0); 
+   }
+   unsigned int  set_expression(const charT* p1, const charT* p2, flag_type f = regex_constants::normal)
+   {
+      assign(p1, p2, f | regex_constants::no_except);
+      return status();
+   }
+   unsigned int  set_expression(const charT* p, flag_type f = regex_constants::normal) 
+   { 
+      assign(p, f | regex_constants::no_except); 
+      return status();
+   }
+   unsigned int  error_code()const
+   {
+      return status();
+   }
+   //
+   // private access methods:
+   //
+   const BOOST_REGEX_DETAIL_NS::re_syntax_base* get_first_state()const
+   {
+      BOOST_REGEX_ASSERT(0 != m_pimpl.get());
+      return m_pimpl->get_first_state();
+   }
+   unsigned get_restart_type()const
+   {
+      BOOST_REGEX_ASSERT(0 != m_pimpl.get());
+      return m_pimpl->get_restart_type();
+   }
+   const unsigned char* get_map()const
+   {
+      BOOST_REGEX_ASSERT(0 != m_pimpl.get());
+      return m_pimpl->get_map();
+   }
+   const ::boost::regex_traits_wrapper<traits>& get_traits()const
+   {
+      BOOST_REGEX_ASSERT(0 != m_pimpl.get());
+      return m_pimpl->get_traits();
+   }
+   bool can_be_null()const
+   {
+      BOOST_REGEX_ASSERT(0 != m_pimpl.get());
+      return m_pimpl->can_be_null();
+   }
+   const BOOST_REGEX_DETAIL_NS::regex_data<charT, traits>& get_data()const
+   {
+      BOOST_REGEX_ASSERT(0 != m_pimpl.get());
+      return m_pimpl->get_data();
+   }
+   std::shared_ptr<BOOST_REGEX_DETAIL_NS::named_subexpressions > get_named_subs()const
+   {
+      return m_pimpl;
+   }
+
+private:
+   std::shared_ptr<BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits> > m_pimpl;
+};
+
+//
+// out of line members;
+// these are the only members that mutate the basic_regex object,
+// and are designed to provide the strong exception guarantee
+// (in the event of a throw, the state of the object remains unchanged).
+//
+template <class charT, class traits>
+basic_regex<charT, traits>& basic_regex<charT, traits>::do_assign(const charT* p1,
+                        const charT* p2,
+                        flag_type f)
+{
+   std::shared_ptr<BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits> > temp;
+   if(!m_pimpl.get())
+   {
+      temp = std::shared_ptr<BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits> >(new BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits>());
+   }
+   else
+   {
+      temp = std::shared_ptr<BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits> >(new BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits>(m_pimpl->m_ptraits));
+   }
+   temp->assign(p1, p2, f);
+   temp.swap(m_pimpl);
+   return *this;
+}
+
+template <class charT, class traits>
+typename basic_regex<charT, traits>::locale_type  basic_regex<charT, traits>::imbue(locale_type l)
+{ 
+   std::shared_ptr<BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits> > temp(new BOOST_REGEX_DETAIL_NS::basic_regex_implementation<charT, traits>());
+   locale_type result = temp->imbue(l);
+   temp.swap(m_pimpl);
+   return result;
+}
+
+//
+// non-members:
+//
+template <class charT, class traits>
+void swap(basic_regex<charT, traits>& e1, basic_regex<charT, traits>& e2)
+{
+   e1.swap(e2);
+}
+
+template <class charT, class traits, class traits2>
+std::basic_ostream<charT, traits>& 
+   operator << (std::basic_ostream<charT, traits>& os, 
+                const basic_regex<charT, traits2>& e)
+{
+   return (os << e.str());
+}
+
+//
+// class reg_expression:
+// this is provided for backwards compatibility only,
+// it is deprecated, no not use!
+//
+#ifdef BOOST_REGEX_NO_FWD
+template <class charT, class traits = regex_traits<charT> >
+#else
+template <class charT, class traits >
+#endif
+class reg_expression : public basic_regex<charT, traits>
+{
+public:
+   typedef typename basic_regex<charT, traits>::flag_type flag_type;
+   typedef typename basic_regex<charT, traits>::size_type size_type;
+   explicit reg_expression(){}
+   explicit reg_expression(const charT* p, flag_type f = regex_constants::normal)
+      : basic_regex<charT, traits>(p, f){}
+   reg_expression(const charT* p1, const charT* p2, flag_type f = regex_constants::normal)
+      : basic_regex<charT, traits>(p1, p2, f){}
+   reg_expression(const charT* p, size_type len, flag_type f)
+      : basic_regex<charT, traits>(p, len, f){}
+   reg_expression(const reg_expression& that)
+      : basic_regex<charT, traits>(that) {}
+   ~reg_expression(){}
+   reg_expression&  operator=(const reg_expression& that)
+   {
+      return this->assign(that);
+   }
+
+   template <class ST, class SA>
+   explicit reg_expression(const std::basic_string<charT, ST, SA>& p, flag_type f = regex_constants::normal)
+   : basic_regex<charT, traits>(p, f)
+   { 
+   }
+
+   template <class InputIterator>
+   reg_expression(InputIterator arg_first, InputIterator arg_last, flag_type f = regex_constants::normal)
+   : basic_regex<charT, traits>(arg_first, arg_last, f)
+   {
+   }
+
+   template <class ST, class SA>
+   reg_expression&  operator=(const std::basic_string<charT, ST, SA>& p)
+   {
+      this->assign(p);
+      return *this;
+   }
+
+};
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning (pop)
+#endif
+
+} // namespace boost
+
+#endif

regex/include/boost/regex/v5/c_regex_traits.hpp

@@ -0,0 +1,474 @@
+/*
+ *
+ * Copyright (c) 2004
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         c_regex_traits.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares regular expression traits class that wraps the global C locale.
+  */
+
+#ifndef BOOST_C_REGEX_TRAITS_HPP_INCLUDED
+#define BOOST_C_REGEX_TRAITS_HPP_INCLUDED
+
+#include <boost/regex/config.hpp>
+#include <boost/regex/v5/regex_workaround.hpp>
+#include <cctype>
+
+namespace boost{
+
+   namespace BOOST_REGEX_DETAIL_NS {
+
+      enum
+      {
+         char_class_space = 1 << 0,
+         char_class_print = 1 << 1,
+         char_class_cntrl = 1 << 2,
+         char_class_upper = 1 << 3,
+         char_class_lower = 1 << 4,
+         char_class_alpha = 1 << 5,
+         char_class_digit = 1 << 6,
+         char_class_punct = 1 << 7,
+         char_class_xdigit = 1 << 8,
+         char_class_alnum = char_class_alpha | char_class_digit,
+         char_class_graph = char_class_alnum | char_class_punct,
+         char_class_blank = 1 << 9,
+         char_class_word = 1 << 10,
+         char_class_unicode = 1 << 11,
+         char_class_horizontal = 1 << 12,
+         char_class_vertical = 1 << 13
+      };
+
+   }
+
+template <class charT>
+struct c_regex_traits;
+
+template<>
+struct c_regex_traits<char>
+{
+   c_regex_traits(){}
+   typedef char char_type;
+   typedef std::size_t size_type;
+   typedef std::string string_type;
+   struct locale_type{};
+   typedef std::uint32_t char_class_type;
+
+   static size_type length(const char_type* p) 
+   { 
+      return (std::strlen)(p); 
+   }
+
+   char translate(char c) const 
+   { 
+      return c; 
+   }
+   char translate_nocase(char c) const 
+   { 
+      return static_cast<char>((std::tolower)(static_cast<unsigned char>(c))); 
+   }
+
+   static string_type  transform(const char* p1, const char* p2);
+   static string_type  transform_primary(const char* p1, const char* p2);
+
+   static char_class_type  lookup_classname(const char* p1, const char* p2);
+   static string_type  lookup_collatename(const char* p1, const char* p2);
+
+   static bool  isctype(char, char_class_type);
+   static int  value(char, int);
+
+   locale_type imbue(locale_type l)
+   { return l; }
+   locale_type getloc()const
+   { return locale_type(); }
+
+private:
+   // this type is not copyable:
+   c_regex_traits(const c_regex_traits&);
+   c_regex_traits& operator=(const c_regex_traits&);
+};
+
+#ifndef BOOST_NO_WREGEX
+template<>
+struct c_regex_traits<wchar_t>
+{
+   c_regex_traits(){}
+   typedef wchar_t char_type;
+   typedef std::size_t size_type;
+   typedef std::wstring string_type;
+   struct locale_type{};
+   typedef std::uint32_t char_class_type;
+
+   static size_type length(const char_type* p) 
+   { 
+      return (std::wcslen)(p); 
+   }
+
+   wchar_t translate(wchar_t c) const 
+   { 
+      return c; 
+   }
+   wchar_t translate_nocase(wchar_t c) const 
+   { 
+      return (std::towlower)(c); 
+   }
+
+   static string_type  transform(const wchar_t* p1, const wchar_t* p2);
+   static string_type  transform_primary(const wchar_t* p1, const wchar_t* p2);
+
+   static char_class_type  lookup_classname(const wchar_t* p1, const wchar_t* p2);
+   static string_type  lookup_collatename(const wchar_t* p1, const wchar_t* p2);
+
+   static bool  isctype(wchar_t, char_class_type);
+   static int  value(wchar_t, int);
+
+   locale_type imbue(locale_type l)
+   { return l; }
+   locale_type getloc()const
+   { return locale_type(); }
+
+private:
+   // this type is not copyable:
+   c_regex_traits(const c_regex_traits&);
+   c_regex_traits& operator=(const c_regex_traits&);
+};
+
+#endif // BOOST_NO_WREGEX
+
+inline c_regex_traits<char>::string_type  c_regex_traits<char>::transform(const char* p1, const char* p2)
+{
+   std::string result(10, ' ');
+   std::size_t s = result.size();
+   std::size_t r;
+   std::string src(p1, p2);
+   while (s < (r = std::strxfrm(&*result.begin(), src.c_str(), s)))
+   {
+#if defined(_CPPLIB_VER)
+      //
+      // A bug in VC11 and 12 causes the program to hang if we pass a null-string
+      // to std::strxfrm, but only for certain locales :-(
+      // Probably effects Intel and Clang or any compiler using the VC std library (Dinkumware).
+      //
+      if (r == INT_MAX)
+      {
+         result.erase();
+         result.insert(result.begin(), static_cast<char>(0));
+         return result;
+      }
+#endif
+      result.append(r - s + 3, ' ');
+      s = result.size();
+   }
+   result.erase(r);
+   return result;
+}
+
+inline c_regex_traits<char>::string_type  c_regex_traits<char>::transform_primary(const char* p1, const char* p2)
+{
+   static char s_delim;
+   static const int s_collate_type = ::boost::BOOST_REGEX_DETAIL_NS::find_sort_syntax(static_cast<c_regex_traits<char>*>(0), &s_delim);
+   std::string result;
+   //
+   // What we do here depends upon the format of the sort key returned by
+   // sort key returned by this->transform:
+   //
+   switch (s_collate_type)
+   {
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_C:
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_unknown:
+      // the best we can do is translate to lower case, then get a regular sort key:
+   {
+      result.assign(p1, p2);
+      for (std::string::size_type i = 0; i < result.size(); ++i)
+         result[i] = static_cast<char>((std::tolower)(static_cast<unsigned char>(result[i])));
+      result = transform(&*result.begin(), &*result.begin() + result.size());
+      break;
+   }
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_fixed:
+   {
+      // get a regular sort key, and then truncate it:
+      result = transform(p1, p2);
+      result.erase(s_delim);
+      break;
+   }
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_delim:
+      // get a regular sort key, and then truncate everything after the delim:
+      result = transform(p1, p2);
+      if ((!result.empty()) && (result[0] == s_delim))
+         break;
+      std::size_t i;
+      for (i = 0; i < result.size(); ++i)
+      {
+         if (result[i] == s_delim)
+            break;
+      }
+      result.erase(i);
+      break;
+   }
+   if (result.empty())
+      result = std::string(1, char(0));
+   return result;
+}
+
+inline c_regex_traits<char>::char_class_type  c_regex_traits<char>::lookup_classname(const char* p1, const char* p2)
+{
+   using namespace BOOST_REGEX_DETAIL_NS;
+   static const char_class_type masks[] =
+   {
+      0,
+      char_class_alnum,
+      char_class_alpha,
+      char_class_blank,
+      char_class_cntrl,
+      char_class_digit,
+      char_class_digit,
+      char_class_graph,
+      char_class_horizontal,
+      char_class_lower,
+      char_class_lower,
+      char_class_print,
+      char_class_punct,
+      char_class_space,
+      char_class_space,
+      char_class_upper,
+      char_class_unicode,
+      char_class_upper,
+      char_class_vertical,
+      char_class_alnum | char_class_word,
+      char_class_alnum | char_class_word,
+      char_class_xdigit,
+   };
+
+   int idx = ::boost::BOOST_REGEX_DETAIL_NS::get_default_class_id(p1, p2);
+   if (idx < 0)
+   {
+      std::string s(p1, p2);
+      for (std::string::size_type i = 0; i < s.size(); ++i)
+         s[i] = static_cast<char>((std::tolower)(static_cast<unsigned char>(s[i])));
+      idx = ::boost::BOOST_REGEX_DETAIL_NS::get_default_class_id(&*s.begin(), &*s.begin() + s.size());
+   }
+   BOOST_REGEX_ASSERT(std::size_t(idx) + 1u < sizeof(masks) / sizeof(masks[0]));
+   return masks[idx + 1];
+}
+
+inline bool  c_regex_traits<char>::isctype(char c, char_class_type mask)
+{
+   using namespace BOOST_REGEX_DETAIL_NS;
+   return
+      ((mask & char_class_space) && (std::isspace)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_print) && (std::isprint)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_cntrl) && (std::iscntrl)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_upper) && (std::isupper)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_lower) && (std::islower)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_alpha) && (std::isalpha)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_digit) && (std::isdigit)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_punct) && (std::ispunct)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_xdigit) && (std::isxdigit)(static_cast<unsigned char>(c)))
+      || ((mask & char_class_blank) && (std::isspace)(static_cast<unsigned char>(c)) && !::boost::BOOST_REGEX_DETAIL_NS::is_separator(c))
+      || ((mask & char_class_word) && (c == '_'))
+      || ((mask & char_class_vertical) && (::boost::BOOST_REGEX_DETAIL_NS::is_separator(c) || (c == '\v')))
+      || ((mask & char_class_horizontal) && (std::isspace)(static_cast<unsigned char>(c)) && !::boost::BOOST_REGEX_DETAIL_NS::is_separator(c) && (c != '\v'));
+}
+
+inline c_regex_traits<char>::string_type  c_regex_traits<char>::lookup_collatename(const char* p1, const char* p2)
+{
+   std::string s(p1, p2);
+   s = ::boost::BOOST_REGEX_DETAIL_NS::lookup_default_collate_name(s);
+   if (s.empty() && (p2 - p1 == 1))
+      s.append(1, *p1);
+   return s;
+}
+
+inline int  c_regex_traits<char>::value(char c, int radix)
+{
+   char b[2] = { c, '\0', };
+   char* ep;
+   int result = std::strtol(b, &ep, radix);
+   if (ep == b)
+      return -1;
+   return result;
+}
+
+#ifndef BOOST_NO_WREGEX
+
+inline c_regex_traits<wchar_t>::string_type  c_regex_traits<wchar_t>::transform(const wchar_t* p1, const wchar_t* p2)
+{
+   std::size_t r;
+   std::size_t s = 10;
+   std::wstring src(p1, p2);
+   std::wstring result(s, L' ');
+   while (s < (r = std::wcsxfrm(&*result.begin(), src.c_str(), s)))
+   {
+#if defined(_CPPLIB_VER)
+      //
+      // A bug in VC11 and 12 causes the program to hang if we pass a null-string
+      // to std::strxfrm, but only for certain locales :-(
+      // Probably effects Intel and Clang or any compiler using the VC std library (Dinkumware).
+      //
+      if (r == INT_MAX)
+      {
+         result.erase();
+         result.insert(result.begin(), static_cast<wchar_t>(0));
+         return result;
+      }
+#endif
+      result.append(r - s + 3, L' ');
+      s = result.size();
+   }
+   result.erase(r);
+   return result;
+}
+
+inline c_regex_traits<wchar_t>::string_type  c_regex_traits<wchar_t>::transform_primary(const wchar_t* p1, const wchar_t* p2)
+{
+   static wchar_t s_delim;
+   static const int s_collate_type = ::boost::BOOST_REGEX_DETAIL_NS::find_sort_syntax(static_cast<const c_regex_traits<wchar_t>*>(0), &s_delim);
+   std::wstring result;
+   //
+   // What we do here depends upon the format of the sort key returned by
+   // sort key returned by this->transform:
+   //
+   switch (s_collate_type)
+   {
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_C:
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_unknown:
+      // the best we can do is translate to lower case, then get a regular sort key:
+   {
+      result.assign(p1, p2);
+      for (std::wstring::size_type i = 0; i < result.size(); ++i)
+         result[i] = (std::towlower)(result[i]);
+      result = c_regex_traits<wchar_t>::transform(&*result.begin(), &*result.begin() + result.size());
+      break;
+   }
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_fixed:
+   {
+      // get a regular sort key, and then truncate it:
+      result = c_regex_traits<wchar_t>::transform(&*result.begin(), &*result.begin() + result.size());
+      result.erase(s_delim);
+      break;
+   }
+   case ::boost::BOOST_REGEX_DETAIL_NS::sort_delim:
+      // get a regular sort key, and then truncate everything after the delim:
+      result = c_regex_traits<wchar_t>::transform(&*result.begin(), &*result.begin() + result.size());
+      if ((!result.empty()) && (result[0] == s_delim))
+         break;
+      std::size_t i;
+      for (i = 0; i < result.size(); ++i)
+      {
+         if (result[i] == s_delim)
+            break;
+      }
+      result.erase(i);
+      break;
+   }
+   if (result.empty())
+      result = std::wstring(1, char(0));
+   return result;
+}
+
+inline c_regex_traits<wchar_t>::char_class_type  c_regex_traits<wchar_t>::lookup_classname(const wchar_t* p1, const wchar_t* p2)
+{
+   using namespace BOOST_REGEX_DETAIL_NS;
+   static const char_class_type masks[] =
+   {
+      0,
+      char_class_alnum,
+      char_class_alpha,
+      char_class_blank,
+      char_class_cntrl,
+      char_class_digit,
+      char_class_digit,
+      char_class_graph,
+      char_class_horizontal,
+      char_class_lower,
+      char_class_lower,
+      char_class_print,
+      char_class_punct,
+      char_class_space,
+      char_class_space,
+      char_class_upper,
+      char_class_unicode,
+      char_class_upper,
+      char_class_vertical,
+      char_class_alnum | char_class_word,
+      char_class_alnum | char_class_word,
+      char_class_xdigit,
+   };
+
+   int idx = ::boost::BOOST_REGEX_DETAIL_NS::get_default_class_id(p1, p2);
+   if (idx < 0)
+   {
+      std::wstring s(p1, p2);
+      for (std::wstring::size_type i = 0; i < s.size(); ++i)
+         s[i] = (std::towlower)(s[i]);
+      idx = ::boost::BOOST_REGEX_DETAIL_NS::get_default_class_id(&*s.begin(), &*s.begin() + s.size());
+   }
+   BOOST_REGEX_ASSERT(idx + 1 < static_cast<int>(sizeof(masks) / sizeof(masks[0])));
+   return masks[idx + 1];
+}
+
+inline bool  c_regex_traits<wchar_t>::isctype(wchar_t c, char_class_type mask)
+{
+   using namespace BOOST_REGEX_DETAIL_NS;
+   return
+      ((mask & char_class_space) && (std::iswspace)(c))
+      || ((mask & char_class_print) && (std::iswprint)(c))
+      || ((mask & char_class_cntrl) && (std::iswcntrl)(c))
+      || ((mask & char_class_upper) && (std::iswupper)(c))
+      || ((mask & char_class_lower) && (std::iswlower)(c))
+      || ((mask & char_class_alpha) && (std::iswalpha)(c))
+      || ((mask & char_class_digit) && (std::iswdigit)(c))
+      || ((mask & char_class_punct) && (std::iswpunct)(c))
+      || ((mask & char_class_xdigit) && (std::iswxdigit)(c))
+      || ((mask & char_class_blank) && (std::iswspace)(c) && !::boost::BOOST_REGEX_DETAIL_NS::is_separator(c))
+      || ((mask & char_class_word) && (c == '_'))
+      || ((mask & char_class_unicode) && (c & ~static_cast<wchar_t>(0xff)))
+      || ((mask & char_class_vertical) && (::boost::BOOST_REGEX_DETAIL_NS::is_separator(c) || (c == L'\v')))
+      || ((mask & char_class_horizontal) && (std::iswspace)(c) && !::boost::BOOST_REGEX_DETAIL_NS::is_separator(c) && (c != L'\v'));
+}
+
+inline c_regex_traits<wchar_t>::string_type  c_regex_traits<wchar_t>::lookup_collatename(const wchar_t* p1, const wchar_t* p2)
+{
+   std::string name;
+   // Usual msvc warning suppression does not work here with std::string template constructor.... use a workaround instead:
+   for (const wchar_t* pos = p1; pos != p2; ++pos)
+      name.push_back((char)*pos);
+   name = ::boost::BOOST_REGEX_DETAIL_NS::lookup_default_collate_name(name);
+   if (!name.empty())
+      return string_type(name.begin(), name.end());
+   if (p2 - p1 == 1)
+      return string_type(1, *p1);
+   return string_type();
+}
+
+inline int  c_regex_traits<wchar_t>::value(wchar_t c, int radix)
+{
+#ifdef BOOST_BORLANDC
+   // workaround for broken wcstol:
+   if ((std::iswxdigit)(c) == 0)
+      return -1;
+#endif
+   wchar_t b[2] = { c, '\0', };
+   wchar_t* ep;
+   int result = std::wcstol(b, &ep, radix);
+   if (ep == b)
+      return -1;
+   return result;
+}
+
+#endif
+
+}
+
+#endif
+
+
+

regex/include/boost/regex/v5/char_regex_traits.hpp

@@ -0,0 +1,59 @@
+/*
+ *
+ * Copyright (c) 2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the
+ * Boost Software License, Version 1.0. (See accompanying file
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         char_regex_traits.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares deprecated traits classes char_regex_traits<>.
+  */
+
+
+#ifndef BOOST_REGEX_V5_CHAR_REGEX_TRAITS_HPP
+#define BOOST_REGEX_V5_CHAR_REGEX_TRAITS_HPP
+
+namespace boost{
+
+namespace deprecated{
+//
+// class char_regex_traits_i
+// provides case insensitive traits classes (deprecated):
+template <class charT>
+class char_regex_traits_i : public regex_traits<charT> {};
+
+template<>
+class char_regex_traits_i<char> : public regex_traits<char>
+{
+public:
+   typedef char char_type;
+   typedef unsigned char uchar_type;
+   typedef unsigned int size_type;
+   typedef regex_traits<char> base_type;
+
+};
+
+#ifndef BOOST_NO_WREGEX
+template<>
+class char_regex_traits_i<wchar_t> : public regex_traits<wchar_t>
+{
+public:
+   typedef wchar_t char_type;
+   typedef unsigned short uchar_type;
+   typedef unsigned int size_type;
+   typedef regex_traits<wchar_t> base_type;
+
+};
+#endif
+} // namespace deprecated
+} // namespace boost
+
+#endif // include
+

regex/include/boost/regex/v5/cregex.hpp

@@ -0,0 +1,195 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the
+ * Boost Software License, Version 1.0. (See accompanying file
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         cregex.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares POSIX API functions
+  *                + boost::RegEx high level wrapper.
+  */
+
+#ifndef BOOST_RE_CREGEX_HPP_INCLUDED
+#define BOOST_RE_CREGEX_HPP_INCLUDED
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/error_type.hpp>
+
+#ifndef BOOST_REGEX_STANDALONE
+#if !defined(BOOST_REGEX_NO_LIB) && !defined(BOOST_REGEX_SOURCE) && !defined(BOOST_ALL_NO_LIB) && defined(__cplusplus)
+#  define BOOST_LIB_NAME boost_regex
+#  if defined(BOOST_REGEX_DYN_LINK) || defined(BOOST_ALL_DYN_LINK)
+#     define BOOST_DYN_LINK
+#  endif
+#  ifdef BOOST_REGEX_DIAG
+#     define BOOST_LIB_DIAGNOSTIC
+#  endif
+#  include <boost/config/auto_link.hpp>
+#endif
+#endif
+
+#ifdef __cplusplus
+#include <cstddef>
+#else
+#include <stddef.h>
+#endif
+
+/* include these defs only for POSIX compatablity */
+#ifdef __cplusplus
+namespace boost{
+extern "C" {
+#endif
+
+#if defined(__cplusplus)
+typedef std::ptrdiff_t regoff_t;
+typedef std::size_t regsize_t;
+#else
+typedef ptrdiff_t regoff_t;
+typedef size_t regsize_t;
+#endif
+
+typedef struct
+{
+   unsigned int re_magic;
+#ifdef __cplusplus
+   std::size_t  re_nsub;      /* number of parenthesized subexpressions */
+#else
+   size_t re_nsub; 
+#endif
+   const char*  re_endp;       /* end pointer for REG_PEND */
+   void* guts;                /* none of your business :-) */
+   match_flag_type eflags;        /* none of your business :-) */
+} regex_tA;
+
+#ifndef BOOST_NO_WREGEX
+typedef struct
+{
+   unsigned int re_magic;
+#ifdef __cplusplus
+   std::size_t  re_nsub;         /* number of parenthesized subexpressions */
+#else
+   size_t re_nsub;
+#endif
+   const wchar_t* re_endp;       /* end pointer for REG_PEND */
+   void* guts;                   /* none of your business :-) */
+   match_flag_type eflags;           /* none of your business :-) */
+} regex_tW;
+#endif
+
+typedef struct
+{
+   regoff_t rm_so;      /* start of match */
+   regoff_t rm_eo;      /* end of match */
+} regmatch_t;
+
+/* regcomp() flags */
+typedef enum{
+   REG_BASIC = 0000,
+   REG_EXTENDED = 0001,
+   REG_ICASE = 0002,
+   REG_NOSUB = 0004,
+   REG_NEWLINE = 0010,
+   REG_NOSPEC = 0020,
+   REG_PEND = 0040,
+   REG_DUMP = 0200,
+   REG_NOCOLLATE = 0400,
+   REG_ESCAPE_IN_LISTS = 01000,
+   REG_NEWLINE_ALT = 02000,
+   REG_PERLEX = 04000,
+
+   REG_PERL = REG_EXTENDED | REG_NOCOLLATE | REG_ESCAPE_IN_LISTS | REG_PERLEX,
+   REG_AWK = REG_EXTENDED | REG_ESCAPE_IN_LISTS,
+   REG_GREP = REG_BASIC | REG_NEWLINE_ALT,
+   REG_EGREP = REG_EXTENDED | REG_NEWLINE_ALT,
+
+   REG_ASSERT = 15,
+   REG_INVARG = 16,
+   REG_ATOI = 255,   /* convert name to number (!) */
+   REG_ITOA = 0400   /* convert number to name (!) */
+} reg_comp_flags;
+
+/* regexec() flags */
+typedef enum{
+   REG_NOTBOL =    00001,
+   REG_NOTEOL =    00002,
+   REG_STARTEND =  00004
+} reg_exec_flags;
+
+/*
+ * POSIX error codes:
+ */
+typedef unsigned reg_error_t;
+typedef reg_error_t reg_errcode_t;  /* backwards compatibility */
+
+static const reg_error_t REG_NOERROR = 0;   /* Success.  */
+static const reg_error_t REG_NOMATCH = 1;   /* Didn't find a match (for regexec).  */
+
+  /* POSIX regcomp return error codes.  (In the order listed in the
+     standard.)  */
+static const reg_error_t REG_BADPAT = 2;    /* Invalid pattern.  */
+static const reg_error_t REG_ECOLLATE = 3;  /* Undefined collating element.  */
+static const reg_error_t REG_ECTYPE = 4;    /* Invalid character class name.  */
+static const reg_error_t REG_EESCAPE = 5;   /* Trailing backslash.  */
+static const reg_error_t REG_ESUBREG = 6;   /* Invalid back reference.  */
+static const reg_error_t REG_EBRACK = 7;    /* Unmatched left bracket.  */
+static const reg_error_t REG_EPAREN = 8;    /* Parenthesis imbalance.  */
+static const reg_error_t REG_EBRACE = 9;    /* Unmatched \{.  */
+static const reg_error_t REG_BADBR = 10;    /* Invalid contents of \{\}.  */
+static const reg_error_t REG_ERANGE = 11;   /* Invalid range end.  */
+static const reg_error_t REG_ESPACE = 12;   /* Ran out of memory.  */
+static const reg_error_t REG_BADRPT = 13;   /* No preceding re for repetition op.  */
+static const reg_error_t REG_EEND = 14;     /* unexpected end of expression */
+static const reg_error_t REG_ESIZE = 15;    /* expression too big */
+static const reg_error_t REG_ERPAREN = 8;   /* = REG_EPAREN : unmatched right parenthesis */
+static const reg_error_t REG_EMPTY = 17;    /* empty expression */
+static const reg_error_t REG_E_MEMORY = 15; /* = REG_ESIZE : out of memory */
+static const reg_error_t REG_ECOMPLEXITY = 18; /* complexity too high */
+static const reg_error_t REG_ESTACK = 19;   /* out of stack space */
+static const reg_error_t REG_E_PERL = 20;   /* Perl (?...) error */
+static const reg_error_t REG_E_UNKNOWN = 21; /* unknown error */
+static const reg_error_t REG_ENOSYS = 21;   /* = REG_E_UNKNOWN : Reserved. */
+
+BOOST_REGEX_DECL int BOOST_REGEX_CCALL regcompA(regex_tA*, const char*, int);
+BOOST_REGEX_DECL regsize_t BOOST_REGEX_CCALL regerrorA(int, const regex_tA*, char*, regsize_t);
+BOOST_REGEX_DECL int BOOST_REGEX_CCALL regexecA(const regex_tA*, const char*, regsize_t, regmatch_t*, int);
+BOOST_REGEX_DECL void BOOST_REGEX_CCALL regfreeA(regex_tA*);
+
+#ifndef BOOST_NO_WREGEX
+BOOST_REGEX_DECL int BOOST_REGEX_CCALL regcompW(regex_tW*, const wchar_t*, int);
+BOOST_REGEX_DECL regsize_t BOOST_REGEX_CCALL regerrorW(int, const regex_tW*, wchar_t*, regsize_t);
+BOOST_REGEX_DECL int BOOST_REGEX_CCALL regexecW(const regex_tW*, const wchar_t*, regsize_t, regmatch_t*, int);
+BOOST_REGEX_DECL void BOOST_REGEX_CCALL regfreeW(regex_tW*);
+#endif
+
+#ifdef UNICODE
+#define regcomp regcompW
+#define regerror regerrorW
+#define regexec regexecW
+#define regfree regfreeW
+#define regex_t regex_tW
+#else
+#define regcomp regcompA
+#define regerror regerrorA
+#define regexec regexecA
+#define regfree regfreeA
+#define regex_t regex_tA
+#endif
+
+#ifdef __cplusplus
+} /* extern "C" */
+} /* namespace */
+#endif
+
+#endif /* include guard */
+

regex/include/boost/regex/v5/error_type.hpp

@@ -0,0 +1,59 @@
+/*
+ *
+ * Copyright (c) 2003-2005
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the
+ * Boost Software License, Version 1.0. (See accompanying file
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         error_type.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares regular expression error type enumerator.
+  */
+
+#ifndef BOOST_REGEX_ERROR_TYPE_HPP
+#define BOOST_REGEX_ERROR_TYPE_HPP
+
+#ifdef __cplusplus
+namespace boost{
+#endif
+
+#ifdef __cplusplus
+namespace regex_constants{
+
+enum error_type{
+
+   error_ok = 0,         /* not used */
+   error_no_match = 1,   /* not used */
+   error_bad_pattern = 2,
+   error_collate = 3,
+   error_ctype = 4,
+   error_escape = 5,
+   error_backref = 6,
+   error_brack = 7,
+   error_paren = 8,
+   error_brace = 9,
+   error_badbrace = 10,
+   error_range = 11,
+   error_space = 12,
+   error_badrepeat = 13,
+   error_end = 14,    /* not used */
+   error_size = 15,
+   error_right_paren = 16,  /* not used */
+   error_empty = 17,
+   error_complexity = 18,
+   error_stack = 19,
+   error_perl_extension = 20,
+   error_unknown = 21
+};
+
+}
+}
+#endif /* __cplusplus */
+
+#endif

regex/include/boost/regex/v5/iterator_category.hpp

@@ -0,0 +1,84 @@
+/*
+ *
+ * Copyright (c) 2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_match.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Iterator traits for selecting an iterator type as
+  *                an integral constant expression.
+  */
+
+
+#ifndef BOOST_REGEX_ITERATOR_CATEGORY_HPP
+#define BOOST_REGEX_ITERATOR_CATEGORY_HPP
+
+#include <iterator>
+#include <type_traits>
+
+namespace boost{
+namespace detail{
+
+template <class I>
+struct is_random_imp
+{
+private:
+   typedef typename std::iterator_traits<I>::iterator_category cat;
+public:
+   static const bool value = (std::is_convertible<cat*, std::random_access_iterator_tag*>::value);
+};
+
+template <class I>
+struct is_random_pointer_imp
+{
+   static const bool value = true;
+};
+
+template <bool is_pointer_type>
+struct is_random_imp_selector
+{
+   template <class I>
+   struct rebind
+   {
+      typedef is_random_imp<I> type;
+   };
+};
+
+template <>
+struct is_random_imp_selector<true>
+{
+   template <class I>
+   struct rebind
+   {
+      typedef is_random_pointer_imp<I> type;
+   };
+};
+
+}
+
+template <class I>
+struct is_random_access_iterator
+{
+private:
+   typedef detail::is_random_imp_selector< std::is_pointer<I>::value> selector;
+   typedef typename selector::template rebind<I> bound_type;
+   typedef typename bound_type::type answer;
+public:
+   static const bool value = answer::value;
+};
+
+template <class I>
+const bool is_random_access_iterator<I>::value;
+
+}
+
+#endif
+

regex/include/boost/regex/v5/iterator_traits.hpp

@@ -0,0 +1,32 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         iterator_traits.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares iterator traits workarounds.
+  */
+
+#ifndef BOOST_REGEX_V5_ITERATOR_TRAITS_HPP
+#define BOOST_REGEX_V5_ITERATOR_TRAITS_HPP
+
+namespace boost{
+namespace BOOST_REGEX_DETAIL_NS{
+
+template <class T>
+struct regex_iterator_traits : public std::iterator_traits<T> {};
+
+} // namespace BOOST_REGEX_DETAIL_NS
+} // namespace boost
+
+#endif
+

regex/include/boost/regex/v5/match_flags.hpp

@@ -0,0 +1,156 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         match_flags.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares match_flags type.
+  */
+
+#ifndef BOOST_REGEX_V5_MATCH_FLAGS
+#define BOOST_REGEX_V5_MATCH_FLAGS
+
+#ifdef __cplusplus
+#  include <cstdint>
+#endif
+
+#ifdef __cplusplus
+namespace boost{
+   namespace regex_constants{
+#endif
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#if BOOST_REGEX_MSVC >= 1800
+#pragma warning(disable : 26812)
+#endif
+#endif
+
+typedef enum _match_flags
+{
+   match_default = 0,
+   match_not_bol = 1,                                /* first is not start of line */
+   match_not_eol = match_not_bol << 1,               /* last is not end of line */
+   match_not_bob = match_not_eol << 1,               /* first is not start of buffer */
+   match_not_eob = match_not_bob << 1,               /* last is not end of buffer */
+   match_not_bow = match_not_eob << 1,               /* first is not start of word */
+   match_not_eow = match_not_bow << 1,               /* last is not end of word */
+   match_not_dot_newline = match_not_eow << 1,       /* \n is not matched by '.' */
+   match_not_dot_null = match_not_dot_newline << 1,  /* '\0' is not matched by '.' */
+   match_prev_avail = match_not_dot_null << 1,       /* *--first is a valid expression */
+   match_init = match_prev_avail << 1,               /* internal use */
+   match_any = match_init << 1,                      /* don't care what we match */
+   match_not_null = match_any << 1,                  /* string can't be null */
+   match_continuous = match_not_null << 1,           /* each grep match must continue from */
+                                                     /* uninterrupted from the previous one */
+   match_partial = match_continuous << 1,            /* find partial matches */
+   
+   match_stop = match_partial << 1,                  /* stop after first match (grep) V3 only */
+   match_not_initial_null = match_stop,              /* don't match initial null, V4 only */
+   match_all = match_stop << 1,                      /* must find the whole of input even if match_any is set */
+   match_perl = match_all << 1,                      /* Use perl matching rules */
+   match_posix = match_perl << 1,                    /* Use POSIX matching rules */
+   match_nosubs = match_posix << 1,                  /* don't trap marked subs */
+   match_extra = match_nosubs << 1,                  /* include full capture information for repeated captures */
+   match_single_line = match_extra << 1,             /* treat text as single line and ignore any \n's when matching ^ and $. */
+   match_unused1 = match_single_line << 1,           /* unused */
+   match_unused2 = match_unused1 << 1,               /* unused */
+   match_unused3 = match_unused2 << 1,               /* unused */
+   match_max = match_unused3,
+
+   format_perl = 0,                                  /* perl style replacement */
+   format_default = 0,                               /* ditto. */
+   format_sed = match_max << 1,                      /* sed style replacement. */
+   format_all = format_sed << 1,                     /* enable all extensions to syntax. */
+   format_no_copy = format_all << 1,                 /* don't copy non-matching segments. */
+   format_first_only = format_no_copy << 1,          /* Only replace first occurrence. */
+   format_is_if = format_first_only << 1,            /* internal use only. */
+   format_literal = format_is_if << 1,               /* treat string as a literal */
+
+   match_not_any = match_not_bol | match_not_eol | match_not_bob 
+      | match_not_eob | match_not_bow | match_not_eow | match_not_dot_newline 
+      | match_not_dot_null | match_prev_avail | match_init | match_not_null
+      | match_continuous | match_partial | match_stop | match_not_initial_null 
+      | match_stop | match_all | match_perl | match_posix | match_nosubs
+      | match_extra | match_single_line | match_unused1 | match_unused2 
+      | match_unused3 | match_max | format_perl | format_default | format_sed
+      | format_all | format_no_copy | format_first_only | format_is_if
+      | format_literal
+
+
+} match_flags;
+
+typedef match_flags match_flag_type;
+
+#ifdef __cplusplus
+inline match_flags operator&(match_flags m1, match_flags m2)
+{ return static_cast<match_flags>(static_cast<std::int32_t>(m1) & static_cast<std::int32_t>(m2)); }
+inline match_flags operator|(match_flags m1, match_flags m2)
+{ return static_cast<match_flags>(static_cast<std::int32_t>(m1) | static_cast<std::int32_t>(m2)); }
+inline match_flags operator^(match_flags m1, match_flags m2)
+{ return static_cast<match_flags>(static_cast<std::int32_t>(m1) ^ static_cast<std::int32_t>(m2)); }
+inline match_flags operator~(match_flags m1)
+{ return static_cast<match_flags>(~static_cast<std::int32_t>(m1)); }
+inline match_flags& operator&=(match_flags& m1, match_flags m2)
+{ m1 = m1&m2; return m1; }
+inline match_flags& operator|=(match_flags& m1, match_flags m2)
+{ m1 = m1|m2; return m1; }
+inline match_flags& operator^=(match_flags& m1, match_flags m2)
+{ m1 = m1^m2; return m1; }
+#endif
+
+#ifdef __cplusplus
+} /* namespace regex_constants */
+/*
+ * import names into boost for backwards compatibility:
+ */
+using regex_constants::match_flag_type;
+using regex_constants::match_default;
+using regex_constants::match_not_bol;
+using regex_constants::match_not_eol;
+using regex_constants::match_not_bob;
+using regex_constants::match_not_eob;
+using regex_constants::match_not_bow;
+using regex_constants::match_not_eow;
+using regex_constants::match_not_dot_newline;
+using regex_constants::match_not_dot_null;
+using regex_constants::match_prev_avail;
+/* using regex_constants::match_init; */
+using regex_constants::match_any;
+using regex_constants::match_not_null;
+using regex_constants::match_continuous;
+using regex_constants::match_partial;
+/*using regex_constants::match_stop; */
+using regex_constants::match_all;
+using regex_constants::match_perl;
+using regex_constants::match_posix;
+using regex_constants::match_nosubs;
+using regex_constants::match_extra;
+using regex_constants::match_single_line;
+/*using regex_constants::match_max; */
+using regex_constants::format_all;
+using regex_constants::format_sed;
+using regex_constants::format_perl;
+using regex_constants::format_default;
+using regex_constants::format_no_copy;
+using regex_constants::format_first_only;
+/*using regex_constants::format_is_if;*/
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+
+
+} /* namespace boost */
+#endif /* __cplusplus */
+#endif /* include guard */
+

regex/include/boost/regex/v5/match_results.hpp

@@ -0,0 +1,667 @@
+/*
+ *
+ * Copyright (c) 1998-2009
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         match_results.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares template class match_results.
+  */
+
+#ifndef BOOST_REGEX_V5_MATCH_RESULTS_HPP
+#define BOOST_REGEX_V5_MATCH_RESULTS_HPP
+
+namespace boost{
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable : 4251 4459)
+#if BOOST_REGEX_MSVC < 1700
+#     pragma warning(disable : 4231)
+#endif
+#  if BOOST_REGEX_MSVC < 1600
+#     pragma warning(disable : 4660)
+#  endif
+#endif
+
+namespace BOOST_REGEX_DETAIL_NS{
+
+class named_subexpressions;
+
+}
+
+template <class BidiIterator, class Allocator>
+class match_results
+{ 
+private:
+   typedef          std::vector<sub_match<BidiIterator>, Allocator> vector_type;
+public: 
+   typedef          sub_match<BidiIterator>                         value_type;
+   typedef typename std::allocator_traits<Allocator>::value_type const &    const_reference;
+   typedef          const_reference                                         reference;
+   typedef typename vector_type::const_iterator                             const_iterator;
+   typedef          const_iterator                                          iterator;
+   typedef typename std::iterator_traits<
+                                    BidiIterator>::difference_type          difference_type;
+   typedef typename std::allocator_traits<Allocator>::size_type             size_type;
+   typedef          Allocator                                               allocator_type;
+   typedef typename std::iterator_traits<
+                                    BidiIterator>::value_type               char_type;
+   typedef          std::basic_string<char_type>                            string_type;
+   typedef          BOOST_REGEX_DETAIL_NS::named_subexpressions             named_sub_type;
+
+   // construct/copy/destroy:
+   explicit match_results(const Allocator& a = Allocator())
+      : m_subs(a), m_base(), m_null(), m_last_closed_paren(0), m_is_singular(true) {}
+   //
+   // IMPORTANT: in the code below, the crazy looking checks around m_is_singular are
+   // all required because it is illegal to copy a singular iterator.
+   // See https://svn.boost.org/trac/boost/ticket/3632.
+   //
+   match_results(const match_results& m)
+      : m_subs(m.m_subs), m_base(), m_null(), m_named_subs(m.m_named_subs), m_last_closed_paren(m.m_last_closed_paren), m_is_singular(m.m_is_singular)
+   {
+      if(!m_is_singular)
+      {
+         m_base = m.m_base;
+         m_null = m.m_null;
+      }
+   }
+   match_results& operator=(const match_results& m)
+   {
+      m_subs = m.m_subs;
+      m_named_subs = m.m_named_subs;
+      m_last_closed_paren = m.m_last_closed_paren;
+      m_is_singular = m.m_is_singular;
+      if(!m_is_singular)
+      {
+         m_base = m.m_base;
+         m_null = m.m_null;
+      }
+      return *this;
+   }
+   ~match_results(){}
+
+   // size:
+   size_type size() const
+   { return empty() ? 0 : m_subs.size() - 2; }
+   size_type max_size() const
+   { return m_subs.max_size(); }
+   bool empty() const
+   { return m_subs.size() < 2; }
+   // element access:
+   difference_type length(int sub = 0) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      sub += 2;
+      if((sub < (int)m_subs.size()) && (sub > 0))
+         return m_subs[sub].length();
+      return 0;
+   }
+   difference_type length(const char_type* sub) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      const char_type* sub_end = sub;
+      while(*sub_end) ++sub_end;
+      return length(named_subexpression_index(sub, sub_end));
+   }
+   template <class charT>
+   difference_type length(const charT* sub) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      const charT* sub_end = sub;
+      while(*sub_end) ++sub_end;
+      return length(named_subexpression_index(sub, sub_end));
+   }
+   template <class charT, class Traits, class A>
+   difference_type length(const std::basic_string<charT, Traits, A>& sub) const
+   {
+      return length(sub.c_str());
+   }
+   difference_type position(size_type sub = 0) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      sub += 2;
+      if(sub < m_subs.size())
+      {
+         const sub_match<BidiIterator>& s = m_subs[sub];
+         if(s.matched || (sub == 2))
+         {
+            return std::distance((BidiIterator)(m_base), (BidiIterator)(s.first));
+         }
+      }
+      return ~static_cast<difference_type>(0);
+   }
+   difference_type position(const char_type* sub) const
+   {
+      const char_type* sub_end = sub;
+      while(*sub_end) ++sub_end;
+      return position(named_subexpression_index(sub, sub_end));
+   }
+   template <class charT>
+   difference_type position(const charT* sub) const
+   {
+      const charT* sub_end = sub;
+      while(*sub_end) ++sub_end;
+      return position(named_subexpression_index(sub, sub_end));
+   }
+   template <class charT, class Traits, class A>
+   difference_type position(const std::basic_string<charT, Traits, A>& sub) const
+   {
+      return position(sub.c_str());
+   }
+   string_type str(int sub = 0) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      sub += 2;
+      string_type result;
+      if(sub < (int)m_subs.size() && (sub > 0))
+      {
+         const sub_match<BidiIterator>& s = m_subs[sub];
+         if(s.matched)
+         {
+            result = s.str();
+         }
+      }
+      return result;
+   }
+   string_type str(const char_type* sub) const
+   {
+      return (*this)[sub].str();
+   }
+   template <class Traits, class A>
+   string_type str(const std::basic_string<char_type, Traits, A>& sub) const
+   {
+      return (*this)[sub].str();
+   }
+   template <class charT>
+   string_type str(const charT* sub) const
+   {
+      return (*this)[sub].str();
+   }
+   template <class charT, class Traits, class A>
+   string_type str(const std::basic_string<charT, Traits, A>& sub) const
+   {
+      return (*this)[sub].str();
+   }
+   const_reference operator[](int sub) const
+   {
+      if(m_is_singular && m_subs.empty())
+         raise_logic_error();
+      sub += 2;
+      if(sub < (int)m_subs.size() && (sub >= 0))
+      {
+         return m_subs[sub];
+      }
+      return m_null;
+   }
+   //
+   // Named sub-expressions:
+   //
+   const_reference named_subexpression(const char_type* i, const char_type* j) const
+   {
+      //
+      // Scan for the leftmost *matched* subexpression with the specified named:
+      //
+      if(m_is_singular)
+         raise_logic_error();
+      BOOST_REGEX_DETAIL_NS::named_subexpressions::range_type r = m_named_subs->equal_range(i, j);
+      while((r.first != r.second) && ((*this)[r.first->index].matched == false))
+         ++r.first;
+      return r.first != r.second ? (*this)[r.first->index] : m_null;
+   }
+   template <class charT>
+   const_reference named_subexpression(const charT* i, const charT* j) const
+   {
+      static_assert(sizeof(charT) <= sizeof(char_type), "Failed internal logic");
+      if(i == j)
+         return m_null;
+      std::vector<char_type> s;
+      while(i != j)
+         s.insert(s.end(), *i++);
+      return named_subexpression(&*s.begin(), &*s.begin() + s.size());
+   }
+   int named_subexpression_index(const char_type* i, const char_type* j) const
+   {
+      //
+      // Scan for the leftmost *matched* subexpression with the specified named.
+      // If none found then return the leftmost expression with that name,
+      // otherwise an invalid index:
+      //
+      if(m_is_singular)
+         raise_logic_error();
+      BOOST_REGEX_DETAIL_NS::named_subexpressions::range_type s, r;
+      s = r = m_named_subs->equal_range(i, j);
+      while((r.first != r.second) && ((*this)[r.first->index].matched == false))
+         ++r.first;
+      if(r.first == r.second)
+         r = s;
+      return r.first != r.second ? r.first->index : -20;
+   }
+   template <class charT>
+   int named_subexpression_index(const charT* i, const charT* j) const
+   {
+      static_assert(sizeof(charT) <= sizeof(char_type), "Failed internal logic");
+      if(i == j)
+         return -20;
+      std::vector<char_type> s;
+      while(i != j)
+         s.insert(s.end(), *i++);
+      return named_subexpression_index(&*s.begin(), &*s.begin() + s.size());
+   }
+   template <class Traits, class A>
+   const_reference operator[](const std::basic_string<char_type, Traits, A>& s) const
+   {
+      return named_subexpression(s.c_str(), s.c_str() + s.size());
+   }
+   const_reference operator[](const char_type* p) const
+   {
+      const char_type* e = p;
+      while(*e) ++e;
+      return named_subexpression(p, e);
+   }
+
+   template <class charT>
+   const_reference operator[](const charT* p) const
+   {
+      static_assert(sizeof(charT) <= sizeof(char_type), "Failed internal logic");
+      if(*p == 0)
+         return m_null;
+      std::vector<char_type> s;
+      while(*p)
+         s.insert(s.end(), *p++);
+      return named_subexpression(&*s.begin(), &*s.begin() + s.size());
+   }
+   template <class charT, class Traits, class A>
+   const_reference operator[](const std::basic_string<charT, Traits, A>& ns) const
+   {
+      static_assert(sizeof(charT) <= sizeof(char_type), "Failed internal logic");
+      if(ns.empty())
+         return m_null;
+      std::vector<char_type> s;
+      for(unsigned i = 0; i < ns.size(); ++i)
+         s.insert(s.end(), ns[i]);
+      return named_subexpression(&*s.begin(), &*s.begin() + s.size());
+   }
+
+   const_reference prefix() const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      return (*this)[-1];
+   }
+
+   const_reference suffix() const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      return (*this)[-2];
+   }
+   const_iterator begin() const
+   {
+      return (m_subs.size() > 2) ? (m_subs.begin() + 2) : m_subs.end();
+   }
+   const_iterator end() const
+   {
+      return m_subs.end();
+   }
+   // format:
+   template <class OutputIterator, class Functor>
+   OutputIterator format(OutputIterator out,
+                         Functor fmt,
+                         match_flag_type flags = format_default) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      typedef typename BOOST_REGEX_DETAIL_NS::compute_functor_type<Functor, match_results<BidiIterator, Allocator>, OutputIterator>::type F;
+      F func(fmt);
+      return func(*this, out, flags);
+   }
+   template <class Functor>
+   string_type format(Functor fmt, match_flag_type flags = format_default) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      std::basic_string<char_type> result;
+      BOOST_REGEX_DETAIL_NS::string_out_iterator<std::basic_string<char_type> > i(result);
+
+      typedef typename BOOST_REGEX_DETAIL_NS::compute_functor_type<Functor, match_results<BidiIterator, Allocator>, BOOST_REGEX_DETAIL_NS::string_out_iterator<std::basic_string<char_type> > >::type F;
+      F func(fmt);
+
+      func(*this, i, flags);
+      return result;
+   }
+   // format with locale:
+   template <class OutputIterator, class Functor, class RegexT>
+   OutputIterator format(OutputIterator out,
+                         Functor fmt,
+                         match_flag_type flags,
+                         const RegexT& re) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      typedef ::boost::regex_traits_wrapper<typename RegexT::traits_type> traits_type;
+      typedef typename BOOST_REGEX_DETAIL_NS::compute_functor_type<Functor, match_results<BidiIterator, Allocator>, OutputIterator, traits_type>::type F;
+      F func(fmt);
+      return func(*this, out, flags, re.get_traits());
+   }
+   template <class RegexT, class Functor>
+   string_type format(Functor fmt,
+                      match_flag_type flags,
+                      const RegexT& re) const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      typedef ::boost::regex_traits_wrapper<typename RegexT::traits_type> traits_type;
+      std::basic_string<char_type> result;
+      BOOST_REGEX_DETAIL_NS::string_out_iterator<std::basic_string<char_type> > i(result);
+
+      typedef typename BOOST_REGEX_DETAIL_NS::compute_functor_type<Functor, match_results<BidiIterator, Allocator>, BOOST_REGEX_DETAIL_NS::string_out_iterator<std::basic_string<char_type> >, traits_type >::type F;
+      F func(fmt);
+
+      func(*this, i, flags, re.get_traits());
+      return result;
+   }
+
+   const_reference get_last_closed_paren()const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      return m_last_closed_paren == 0 ? m_null : (*this)[m_last_closed_paren];
+   }
+
+   allocator_type get_allocator() const
+   {
+      return m_subs.get_allocator();
+   }
+   void swap(match_results& that)
+   {
+      std::swap(m_subs, that.m_subs);
+      std::swap(m_named_subs, that.m_named_subs);
+      std::swap(m_last_closed_paren, that.m_last_closed_paren);
+      if(m_is_singular)
+      {
+         if(!that.m_is_singular)
+         {
+            m_base = that.m_base;
+            m_null = that.m_null;
+         }
+      }
+      else if(that.m_is_singular)
+      {
+         that.m_base = m_base;
+         that.m_null = m_null;
+      }
+      else
+      {
+         std::swap(m_base, that.m_base);
+         std::swap(m_null, that.m_null);
+      }
+      std::swap(m_is_singular, that.m_is_singular);
+   }
+   bool operator==(const match_results& that)const
+   {
+      if(m_is_singular)
+      {
+         return that.m_is_singular;
+      }
+      else if(that.m_is_singular)
+      {
+         return false;
+      }
+      return (m_subs == that.m_subs) && (m_base == that.m_base) && (m_last_closed_paren == that.m_last_closed_paren);
+   }
+   bool operator!=(const match_results& that)const
+   { return !(*this == that); }
+
+#ifdef BOOST_REGEX_MATCH_EXTRA
+   typedef typename sub_match<BidiIterator>::capture_sequence_type capture_sequence_type;
+
+   const capture_sequence_type& captures(int i)const
+   {
+      if(m_is_singular)
+         raise_logic_error();
+      return (*this)[i].captures();
+   }
+#endif
+
+   //
+   // private access functions:
+   void  set_second(BidiIterator i)
+   {
+      BOOST_REGEX_ASSERT(m_subs.size() > 2);
+      m_subs[2].second = i;
+      m_subs[2].matched = true;
+      m_subs[0].first = i;
+      m_subs[0].matched = (m_subs[0].first != m_subs[0].second);
+      m_null.first = i;
+      m_null.second = i;
+      m_null.matched = false;
+      m_is_singular = false;
+   }
+
+   void  set_second(BidiIterator i, size_type pos, bool m = true, bool escape_k = false)
+   {
+      if(pos)
+         m_last_closed_paren = static_cast<int>(pos);
+      pos += 2;
+      BOOST_REGEX_ASSERT(m_subs.size() > pos);
+      m_subs[pos].second = i;
+      m_subs[pos].matched = m;
+      if((pos == 2) && !escape_k)
+      {
+         m_subs[0].first = i;
+         m_subs[0].matched = (m_subs[0].first != m_subs[0].second);
+         m_null.first = i;
+         m_null.second = i;
+         m_null.matched = false;
+         m_is_singular = false;
+      }
+   }
+   void  set_size(size_type n, BidiIterator i, BidiIterator j)
+   {
+      value_type v(j);
+      size_type len = m_subs.size();
+      if(len > n + 2)
+      {
+         m_subs.erase(m_subs.begin()+n+2, m_subs.end());
+         std::fill(m_subs.begin(), m_subs.end(), v);
+      }
+      else
+      {
+         std::fill(m_subs.begin(), m_subs.end(), v);
+         if(n+2 != len)
+            m_subs.insert(m_subs.end(), n+2-len, v);
+      }
+      m_subs[1].first = i;
+      m_last_closed_paren = 0;
+   }
+   void  set_base(BidiIterator pos)
+   {
+      m_base = pos;
+   }
+   BidiIterator base()const
+   {
+      return m_base;
+   }
+   void  set_first(BidiIterator i)
+   {
+      BOOST_REGEX_ASSERT(m_subs.size() > 2);
+      // set up prefix:
+      m_subs[1].second = i;
+      m_subs[1].matched = (m_subs[1].first != i);
+      // set up $0:
+      m_subs[2].first = i;
+      // zero out everything else:
+      for(size_type n = 3; n < m_subs.size(); ++n)
+      {
+         m_subs[n].first = m_subs[n].second = m_subs[0].second;
+         m_subs[n].matched = false;
+      }
+   }
+   void  set_first(BidiIterator i, size_type pos, bool escape_k = false)
+   {
+      BOOST_REGEX_ASSERT(pos+2 < m_subs.size());
+      if(pos || escape_k)
+      {
+         m_subs[pos+2].first = i;
+         if(escape_k)
+         {
+            m_subs[1].second = i;
+            m_subs[1].matched = (m_subs[1].first != m_subs[1].second);
+         }
+      }
+      else
+         set_first(i);
+   }
+   void  maybe_assign(const match_results<BidiIterator, Allocator>& m);
+
+   void  set_named_subs(std::shared_ptr<named_sub_type> subs)
+   {
+      m_named_subs = subs;
+   }
+
+private:
+   //
+   // Error handler called when an uninitialized match_results is accessed:
+   //
+   static void raise_logic_error()
+   {
+      std::logic_error e("Attempt to access an uninitialized boost::match_results<> class.");
+#ifndef BOOST_REGEX_STANDALONE
+      boost::throw_exception(e);
+#else
+      throw e;
+#endif
+   }
+
+
+   vector_type            m_subs;                      // subexpressions
+   BidiIterator   m_base;                              // where the search started from
+   sub_match<BidiIterator> m_null;                     // a null match
+   std::shared_ptr<named_sub_type> m_named_subs;     // Shared copy of named subs in the regex object
+   int m_last_closed_paren;                            // Last ) to be seen - used for formatting
+   bool m_is_singular;                                 // True if our stored iterators are singular
+};
+
+template <class BidiIterator, class Allocator>
+void  match_results<BidiIterator, Allocator>::maybe_assign(const match_results<BidiIterator, Allocator>& m)
+{
+   if(m_is_singular)
+   {
+      *this = m;
+      return;
+   }
+   const_iterator p1, p2;
+   p1 = begin();
+   p2 = m.begin();
+   //
+   // Distances are measured from the start of *this* match, unless this isn't
+   // a valid match in which case we use the start of the whole sequence.  Note that
+   // no subsequent match-candidate can ever be to the left of the first match found.
+   // This ensures that when we are using bidirectional iterators, that distances 
+   // measured are as short as possible, and therefore as efficient as possible
+   // to compute.  Finally note that we don't use the "matched" data member to test
+   // whether a sub-expression is a valid match, because partial matches set this
+   // to false for sub-expression 0.
+   //
+   BidiIterator l_end = this->suffix().second;
+   BidiIterator l_base = (p1->first == l_end) ? this->prefix().first : (*this)[0].first;
+   difference_type len1 = 0;
+   difference_type len2 = 0;
+   difference_type base1 = 0;
+   difference_type base2 = 0;
+   std::size_t i;
+   for(i = 0; i < size(); ++i, ++p1, ++p2)
+   {
+      //
+      // Leftmost takes priority over longest; handle special cases
+      // where distances need not be computed first (an optimisation
+      // for bidirectional iterators: ensure that we don't accidently
+      // compute the length of the whole sequence, as this can be really
+      // expensive).
+      //
+      if(p1->first == l_end)
+      {
+         if(p2->first != l_end)
+         {
+            // p2 must be better than p1, and no need to calculate
+            // actual distances:
+            base1 = 1;
+            base2 = 0;
+            break;
+         }
+         else
+         {
+            // *p1 and *p2 are either unmatched or match end-of sequence,
+            // either way no need to calculate distances:
+            if((p1->matched == false) && (p2->matched == true))
+               break;
+            if((p1->matched == true) && (p2->matched == false))
+               return;
+            continue;
+         }
+      }
+      else if(p2->first == l_end)
+      {
+         // p1 better than p2, and no need to calculate distances:
+         return;
+      }
+      base1 = std::distance(l_base, p1->first);
+      base2 = std::distance(l_base, p2->first);
+      BOOST_REGEX_ASSERT(base1 >= 0);
+      BOOST_REGEX_ASSERT(base2 >= 0);
+      if(base1 < base2) return;
+      if(base2 < base1) break;
+
+      len1 = std::distance((BidiIterator)p1->first, (BidiIterator)p1->second);
+      len2 = std::distance((BidiIterator)p2->first, (BidiIterator)p2->second);
+      BOOST_REGEX_ASSERT(len1 >= 0);
+      BOOST_REGEX_ASSERT(len2 >= 0);
+      if((len1 != len2) || ((p1->matched == false) && (p2->matched == true)))
+         break;
+      if((p1->matched == true) && (p2->matched == false))
+         return;
+   }
+   if(i == size())
+      return;
+   if(base2 < base1)
+      *this = m;
+   else if((len2 > len1) || ((p1->matched == false) && (p2->matched == true)) )
+      *this = m;
+}
+
+template <class BidiIterator, class Allocator>
+void swap(match_results<BidiIterator, Allocator>& a, match_results<BidiIterator, Allocator>& b)
+{
+   a.swap(b);
+}
+
+template <class charT, class traits, class BidiIterator, class Allocator>
+std::basic_ostream<charT, traits>&
+   operator << (std::basic_ostream<charT, traits>& os,
+                const match_results<BidiIterator, Allocator>& s)
+{
+   return (os << s.str());
+}
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+} // namespace boost
+
+#endif
+
+

regex/include/boost/regex/v5/mem_block_cache.hpp

@@ -0,0 +1,173 @@
+ /*
+ * Copyright (c) 2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         mem_block_cache.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: memory block cache used by the non-recursive matcher.
+  */
+
+#ifndef BOOST_REGEX_V5_MEM_BLOCK_CACHE_HPP
+#define BOOST_REGEX_V5_MEM_BLOCK_CACHE_HPP
+
+#include <new>
+#ifdef BOOST_HAS_THREADS
+#include <mutex>
+#endif
+
+#ifndef BOOST_NO_CXX11_HDR_ATOMIC
+  #include <atomic>
+  #if ATOMIC_POINTER_LOCK_FREE == 2
+    #define BOOST_REGEX_MEM_BLOCK_CACHE_LOCK_FREE
+    #define BOOST_REGEX_ATOMIC_POINTER std::atomic
+  #endif
+#endif
+
+namespace boost{
+namespace BOOST_REGEX_DETAIL_NS{
+
+#if BOOST_REGEX_MAX_CACHE_BLOCKS != 0
+#ifdef BOOST_REGEX_MEM_BLOCK_CACHE_LOCK_FREE /* lock free implementation */
+struct mem_block_cache
+{
+  std::atomic<void*> cache[BOOST_REGEX_MAX_CACHE_BLOCKS];
+
+   ~mem_block_cache()
+   {
+     for (size_t i = 0;i < BOOST_REGEX_MAX_CACHE_BLOCKS; ++i) {
+       if (cache[i].load()) ::operator delete(cache[i].load());
+     }
+   }
+   void* get()
+   {
+     for (size_t i = 0;i < BOOST_REGEX_MAX_CACHE_BLOCKS; ++i) {
+       void* p = cache[i].load();
+       if (p != NULL) {
+         if (cache[i].compare_exchange_strong(p, NULL)) return p;
+       }
+     }
+     return ::operator new(BOOST_REGEX_BLOCKSIZE);
+   }
+   void put(void* ptr)
+   {
+     for (size_t i = 0;i < BOOST_REGEX_MAX_CACHE_BLOCKS; ++i) {
+       void* p = cache[i].load();
+       if (p == NULL) {
+         if (cache[i].compare_exchange_strong(p, ptr)) return;
+       }
+     }
+     ::operator delete(ptr);
+   }
+
+   static mem_block_cache& instance()
+   {
+      static mem_block_cache block_cache = { { {nullptr} } };
+      return block_cache;
+   }
+};
+
+
+#else /* lock-based implementation */
+
+
+struct mem_block_node
+{
+   mem_block_node* next;
+};
+
+struct mem_block_cache
+{
+   // this member has to be statically initialsed:
+   mem_block_node* next { nullptr };
+   unsigned cached_blocks { 0 };
+#ifdef BOOST_HAS_THREADS
+   std::mutex mut;
+#endif
+
+   ~mem_block_cache()
+   {
+      while(next)
+      {
+         mem_block_node* old = next;
+         next = next->next;
+         ::operator delete(old);
+      }
+   }
+   void* get()
+   {
+#ifdef BOOST_HAS_THREADS
+      std::lock_guard<std::mutex> g(mut);
+#endif
+     if(next)
+      {
+         mem_block_node* result = next;
+         next = next->next;
+         --cached_blocks;
+         return result;
+      }
+      return ::operator new(BOOST_REGEX_BLOCKSIZE);
+   }
+   void put(void* p)
+   {
+#ifdef BOOST_HAS_THREADS
+      std::lock_guard<std::mutex> g(mut);
+#endif
+      if(cached_blocks >= BOOST_REGEX_MAX_CACHE_BLOCKS)
+      {
+         ::operator delete(p);
+      }
+      else
+      {
+         mem_block_node* old = static_cast<mem_block_node*>(p);
+         old->next = next;
+         next = old;
+         ++cached_blocks;
+      }
+   }
+   static mem_block_cache& instance()
+   {
+      static mem_block_cache block_cache;
+      return block_cache;
+   }
+};
+#endif
+#endif
+
+#if BOOST_REGEX_MAX_CACHE_BLOCKS == 0
+
+inline void*  get_mem_block()
+{
+   return ::operator new(BOOST_REGEX_BLOCKSIZE);
+}
+
+inline void  put_mem_block(void* p)
+{
+   ::operator delete(p);
+}
+
+#else
+
+inline void*  get_mem_block()
+{
+   return mem_block_cache::instance().get();
+}
+
+inline void  put_mem_block(void* p)
+{
+   mem_block_cache::instance().put(p);
+}
+
+#endif
+}
+} // namespace boost
+
+#endif
+

regex/include/boost/regex/v5/object_cache.hpp

@@ -0,0 +1,160 @@
+/*
+ *
+ * Copyright (c) 2004
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         object_cache.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Implements a generic object cache.
+  */
+
+#ifndef BOOST_REGEX_OBJECT_CACHE_HPP
+#define BOOST_REGEX_OBJECT_CACHE_HPP
+
+#include <boost/regex/config.hpp>
+#include <memory>
+#include <map>
+#include <list>
+#include <stdexcept>
+#include <string>
+#ifdef BOOST_HAS_THREADS
+#include <mutex>
+#endif
+
+namespace boost{
+
+template <class Key, class Object>
+class object_cache
+{
+public:
+   typedef std::pair< ::std::shared_ptr<Object const>, Key const*> value_type;
+   typedef std::list<value_type> list_type;
+   typedef typename list_type::iterator list_iterator;
+   typedef std::map<Key, list_iterator> map_type;
+   typedef typename map_type::iterator map_iterator;
+   typedef typename list_type::size_type size_type;
+   static std::shared_ptr<Object const> get(const Key& k, size_type l_max_cache_size);
+
+private:
+   static std::shared_ptr<Object const> do_get(const Key& k, size_type l_max_cache_size);
+
+   struct data
+   {
+      list_type   cont;
+      map_type    index;
+   };
+
+   // Needed by compilers not implementing the resolution to DR45. For reference,
+   // see http://www.open-std.org/JTC1/SC22/WG21/docs/cwg_defects.html#45.
+   friend struct data;
+};
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable: 4702)
+#endif
+template <class Key, class Object>
+std::shared_ptr<Object const> object_cache<Key, Object>::get(const Key& k, size_type l_max_cache_size)
+{
+#ifdef BOOST_HAS_THREADS
+   static std::mutex mut;
+   std::lock_guard<std::mutex> l(mut);
+   return do_get(k, l_max_cache_size);
+#else
+   return do_get(k, l_max_cache_size);
+#endif
+}
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+
+template <class Key, class Object>
+std::shared_ptr<Object const> object_cache<Key, Object>::do_get(const Key& k, size_type l_max_cache_size)
+{
+   typedef typename object_cache<Key, Object>::data object_data;
+   typedef typename map_type::size_type map_size_type;
+   static object_data s_data;
+
+   //
+   // see if the object is already in the cache:
+   //
+   map_iterator mpos = s_data.index.find(k);
+   if(mpos != s_data.index.end())
+   {
+      //
+      // Eureka! 
+      // We have a cached item, bump it up the list and return it:
+      //
+      if(--(s_data.cont.end()) != mpos->second)
+      {
+         // splice out the item we want to move:
+         list_type temp;
+         temp.splice(temp.end(), s_data.cont, mpos->second);
+         // and now place it at the end of the list:
+         s_data.cont.splice(s_data.cont.end(), temp, temp.begin());
+         BOOST_REGEX_ASSERT(*(s_data.cont.back().second) == k);
+         // update index with new position:
+         mpos->second = --(s_data.cont.end());
+         BOOST_REGEX_ASSERT(&(mpos->first) == mpos->second->second);
+         BOOST_REGEX_ASSERT(&(mpos->first) == s_data.cont.back().second);
+      }
+      return s_data.cont.back().first;
+   }
+   //
+   // if we get here then the item is not in the cache,
+   // so create it:
+   //
+   std::shared_ptr<Object const> result(new Object(k));
+   //
+   // Add it to the list, and index it:
+   //
+   s_data.cont.push_back(value_type(result, static_cast<Key const*>(0)));
+   s_data.index.insert(std::make_pair(k, --(s_data.cont.end())));
+   s_data.cont.back().second = &(s_data.index.find(k)->first);
+   map_size_type s = s_data.index.size();
+   BOOST_REGEX_ASSERT(s_data.index[k]->first.get() == result.get());
+   BOOST_REGEX_ASSERT(&(s_data.index.find(k)->first) == s_data.cont.back().second);
+   BOOST_REGEX_ASSERT(s_data.index.find(k)->first == k);
+   if(s > l_max_cache_size)
+   {
+      //
+      // We have too many items in the list, so we need to start
+      // popping them off the back of the list, but only if they're
+      // being held uniquely by us:
+      //
+      list_iterator pos = s_data.cont.begin();
+      list_iterator last = s_data.cont.end();
+      while((pos != last) && (s > l_max_cache_size))
+      {
+         if(pos->first.use_count() == 1)
+         {
+            list_iterator condemmed(pos);
+            ++pos;
+            // now remove the items from our containers, 
+            // then order has to be as follows:
+            BOOST_REGEX_ASSERT(s_data.index.find(*(condemmed->second)) != s_data.index.end());
+            s_data.index.erase(*(condemmed->second));
+            s_data.cont.erase(condemmed); 
+            --s;
+         }
+         else
+            ++pos;
+      }
+      BOOST_REGEX_ASSERT(s_data.index[k]->first.get() == result.get());
+      BOOST_REGEX_ASSERT(&(s_data.index.find(k)->first) == s_data.cont.back().second);
+      BOOST_REGEX_ASSERT(s_data.index.find(k)->first == k);
+   }
+   return result;
+}
+
+}
+
+#endif

regex/include/boost/regex/v5/pattern_except.hpp

@@ -0,0 +1,106 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         pattern_except.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares pattern-matching exception classes.
+  */
+
+#ifndef BOOST_RE_V5_PAT_EXCEPT_HPP
+#define BOOST_RE_V5_PAT_EXCEPT_HPP
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+
+#include <cstddef>
+#include <stdexcept>
+#include <boost/regex/v5/error_type.hpp>
+#include <boost/regex/v5/regex_traits_defaults.hpp>
+
+namespace boost{
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable : 4275)
+#if BOOST_REGEX_MSVC >= 1800
+#pragma warning(disable : 26812 4459)
+#endif
+#endif
+class regex_error : public std::runtime_error
+{
+public:
+   explicit regex_error(const std::string& s, regex_constants::error_type err = regex_constants::error_unknown, std::ptrdiff_t pos = 0)
+      : std::runtime_error(s)
+      , m_error_code(err)
+      , m_position(pos)
+   {
+   }
+   explicit regex_error(regex_constants::error_type err)
+      : std::runtime_error(::boost::BOOST_REGEX_DETAIL_NS::get_default_error_string(err))
+      , m_error_code(err)
+      , m_position(0)
+   {
+   }
+   ~regex_error() noexcept override {}
+   regex_constants::error_type code()const
+   { return m_error_code; }
+   std::ptrdiff_t position()const
+   { return m_position; }
+   void raise()const 
+   {
+#ifndef BOOST_NO_EXCEPTIONS
+#ifndef BOOST_REGEX_STANDALONE
+      ::boost::throw_exception(*this);
+#else
+      throw* this;
+#endif
+#endif
+   }
+private:
+   regex_constants::error_type m_error_code;
+   std::ptrdiff_t m_position;
+};
+
+typedef regex_error bad_pattern;
+typedef regex_error bad_expression;
+
+namespace BOOST_REGEX_DETAIL_NS{
+
+template <class E>
+inline void raise_runtime_error(const E& ex)
+{
+#ifndef BOOST_REGEX_STANDALONE
+   ::boost::throw_exception(ex);
+#else
+   throw ex;
+#endif
+}
+
+template <class traits>
+void raise_error(const traits& t, regex_constants::error_type code)
+{
+   (void)t;  // warning suppression
+   regex_error e(t.error_string(code), code, 0);
+   ::boost::BOOST_REGEX_DETAIL_NS::raise_runtime_error(e);
+}
+
+}
+
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+
+} // namespace boost
+
+#endif

regex/include/boost/regex/v5/perl_matcher.hpp

@@ -0,0 +1,576 @@
+/*
+ *
+ * Copyright (c) 2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+#ifndef BOOST_REGEX_MATCHER_HPP
+#define BOOST_REGEX_MATCHER_HPP
+
+#include <boost/regex/v5/iterator_category.hpp>
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(push)
+#pragma warning(disable : 4251 4459)
+#if BOOST_REGEX_MSVC < 1700
+#     pragma warning(disable : 4231)
+#endif
+#  if BOOST_REGEX_MSVC < 1600
+#     pragma warning(disable : 4660)
+#  endif
+#if BOOST_REGEX_MSVC < 1910
+#pragma warning(disable:4800)
+#endif
+#endif
+
+namespace boost{
+namespace BOOST_REGEX_DETAIL_NS{
+
+//
+// error checking API:
+//
+inline void  verify_options(boost::regex_constants::syntax_option_type, match_flag_type mf)
+{
+   //
+   // can't mix match_extra with POSIX matching rules:
+   //
+   if ((mf & match_extra) && (mf & match_posix))
+   {
+      std::logic_error msg("Usage Error: Can't mix regular expression captures with POSIX matching rules");
+#ifndef BOOST_REGEX_STANDALONE
+      throw_exception(msg);
+#else
+      throw msg;
+#endif
+   }
+}
+//
+// function can_start:
+//
+template <class charT>
+inline bool can_start(charT c, const unsigned char* map, unsigned char mask)
+{
+   return ((c < static_cast<charT>(0)) ? true : ((c >= static_cast<charT>(1 << CHAR_BIT)) ? true : map[c] & mask));
+}
+inline bool can_start(char c, const unsigned char* map, unsigned char mask)
+{
+   return map[(unsigned char)c] & mask;
+}
+inline bool can_start(signed char c, const unsigned char* map, unsigned char mask)
+{
+   return map[(unsigned char)c] & mask;
+}
+inline bool can_start(unsigned char c, const unsigned char* map, unsigned char mask)
+{
+   return map[c] & mask;
+}
+inline bool can_start(unsigned short c, const unsigned char* map, unsigned char mask)
+{
+   return ((c >= (1 << CHAR_BIT)) ? true : map[c] & mask);
+}
+#if defined(WCHAR_MIN) && (WCHAR_MIN == 0) && !defined(BOOST_NO_INTRINSIC_WCHAR_T)
+inline bool can_start(wchar_t c, const unsigned char* map, unsigned char mask)
+{
+   return ((c >= static_cast<wchar_t>(1u << CHAR_BIT)) ? true : map[c] & mask);
+}
+#endif
+#if !defined(BOOST_NO_INTRINSIC_WCHAR_T)
+inline bool can_start(unsigned int c, const unsigned char* map, unsigned char mask)
+{
+   return (((c >= static_cast<unsigned int>(1u << CHAR_BIT)) ? true : map[c] & mask));
+}
+#endif
+
+template <class C, class T, class A>
+inline int string_compare(const std::basic_string<C,T,A>& s, const C* p)
+{ 
+   if(0 == *p)
+   {
+      if(s.empty() || ((s.size() == 1) && (s[0] == 0)))
+         return 0;
+   }
+   return s.compare(p); 
+}
+template <class Seq, class C>
+inline int string_compare(const Seq& s, const C* p)
+{
+   std::size_t i = 0;
+   while((i < s.size()) && (p[i] == s[i]))
+   {
+      ++i;
+   }
+   return (i == s.size()) ? -(int)p[i] : (int)s[i] - (int)p[i];
+}
+# define STR_COMP(s,p) string_compare(s,p)
+
+template<class charT>
+inline const charT* re_skip_past_null(const charT* p)
+{
+  while (*p != static_cast<charT>(0)) ++p;
+  return ++p;
+}
+
+template <class iterator, class charT, class traits_type, class char_classT>
+iterator  re_is_set_member(iterator next, 
+                          iterator last, 
+                          const re_set_long<char_classT>* set_, 
+                          const regex_data<charT, traits_type>& e, bool icase)
+{   
+   const charT* p = reinterpret_cast<const charT*>(set_+1);
+   iterator ptr;
+   unsigned int i;
+   //bool icase = e.m_flags & regex_constants::icase;
+
+   if(next == last) return next;
+
+   typedef typename traits_type::string_type traits_string_type;
+   const ::boost::regex_traits_wrapper<traits_type>& traits_inst = *(e.m_ptraits);
+   
+   // dwa 9/13/00 suppress incorrect MSVC warning - it claims this is never
+   // referenced
+   (void)traits_inst;
+
+   // try and match a single character, could be a multi-character
+   // collating element...
+   for(i = 0; i < set_->csingles; ++i)
+   {
+      ptr = next;
+      if(*p == static_cast<charT>(0))
+      {
+         // treat null string as special case:
+         if(traits_inst.translate(*ptr, icase))
+         {
+            ++p;
+            continue;
+         }
+         return set_->isnot ? next : (ptr == next) ? ++next : ptr;
+      }
+      else
+      {
+         while(*p && (ptr != last))
+         {
+            if(traits_inst.translate(*ptr, icase) != *p)
+               break;
+            ++p;
+            ++ptr;
+         }
+
+         if(*p == static_cast<charT>(0)) // if null we've matched
+            return set_->isnot ? next : (ptr == next) ? ++next : ptr;
+
+         p = re_skip_past_null(p);     // skip null
+      }
+   }
+
+   charT col = traits_inst.translate(*next, icase);
+
+
+   if(set_->cranges || set_->cequivalents)
+   {
+      traits_string_type s1;
+      //
+      // try and match a range, NB only a single character can match
+      if(set_->cranges)
+      {
+         if((e.m_flags & regex_constants::collate) == 0)
+            s1.assign(1, col);
+         else
+         {
+            charT a[2] = { col, charT(0), };
+            s1 = traits_inst.transform(a, a + 1);
+         }
+         for(i = 0; i < set_->cranges; ++i)
+         {
+            if(STR_COMP(s1, p) >= 0)
+            {
+               do{ ++p; }while(*p);
+               ++p;
+               if(STR_COMP(s1, p) <= 0)
+                  return set_->isnot ? next : ++next;
+            }
+            else
+            {
+               // skip first string
+               do{ ++p; }while(*p);
+               ++p;
+            }
+            // skip second string
+            do{ ++p; }while(*p);
+            ++p;
+         }
+      }
+      //
+      // try and match an equivalence class, NB only a single character can match
+      if(set_->cequivalents)
+      {
+         charT a[2] = { col, charT(0), };
+         s1 = traits_inst.transform_primary(a, a +1);
+         for(i = 0; i < set_->cequivalents; ++i)
+         {
+            if(STR_COMP(s1, p) == 0)
+               return set_->isnot ? next : ++next;
+            // skip string
+            do{ ++p; }while(*p);
+            ++p;
+         }
+      }
+   }
+   if(traits_inst.isctype(col, set_->cclasses) == true)
+      return set_->isnot ? next : ++next;
+   if((set_->cnclasses != 0) && (traits_inst.isctype(col, set_->cnclasses) == false))
+      return set_->isnot ? next : ++next;
+   return set_->isnot ? ++next : next;
+}
+
+template <class BidiIterator>
+class repeater_count
+{
+   repeater_count** stack;
+   repeater_count* next;
+   int state_id;
+   std::size_t count;        // the number of iterations so far
+   BidiIterator start_pos;   // where the last repeat started
+
+   repeater_count* unwind_until(int n, repeater_count* p, int current_recursion_id)
+   { 
+      while(p && (p->state_id != n))
+      {
+         if(-2 - current_recursion_id == p->state_id)
+            return 0;
+         p = p->next;
+         if(p && (p->state_id < 0))
+         {
+            p = unwind_until(p->state_id, p, current_recursion_id);
+            if(!p)
+               return p;
+            p = p->next;
+         }
+      }
+      return p;
+   }
+public:
+   repeater_count(repeater_count** s) : stack(s), next(0), state_id(-1), count(0), start_pos() {}
+   
+   repeater_count(int i, repeater_count** s, BidiIterator start, int current_recursion_id)
+      : start_pos(start)
+   {
+      state_id = i;
+      stack = s;
+      next = *stack;
+      *stack = this;
+      if((state_id > next->state_id) && (next->state_id >= 0))
+         count = 0;
+      else
+      {
+         repeater_count* p = next;
+         p = unwind_until(state_id, p, current_recursion_id);
+         if(p)
+         {
+            count = p->count;
+            start_pos = p->start_pos;
+         }
+         else
+            count = 0;
+      }
+   }
+   ~repeater_count()
+   {
+      if(next)
+         *stack = next;
+   }
+   std::size_t get_count() { return count; }
+   int get_id() { return state_id; }
+   std::size_t operator++() { return ++count; }
+   bool check_null_repeat(const BidiIterator& pos, std::size_t max)
+   {
+      // this is called when we are about to start a new repeat,
+      // if the last one was NULL move our count to max,
+      // otherwise save the current position.
+      bool result = (count == 0) ? false : (pos == start_pos);
+      if(result)
+         count = max;
+      else
+         start_pos = pos;
+      return result;
+   }
+};
+
+struct saved_state;
+
+enum saved_state_type
+{
+   saved_type_end = 0,
+   saved_type_paren = 1,
+   saved_type_recurse = 2,
+   saved_type_assertion = 3,
+   saved_state_alt = 4,
+   saved_state_repeater_count = 5,
+   saved_state_extra_block = 6,
+   saved_state_greedy_single_repeat = 7,
+   saved_state_rep_slow_dot = 8,
+   saved_state_rep_fast_dot = 9,
+   saved_state_rep_char = 10,
+   saved_state_rep_short_set = 11,
+   saved_state_rep_long_set = 12,
+   saved_state_non_greedy_long_repeat = 13, 
+   saved_state_count = 14
+};
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(push)
+#if BOOST_REGEX_MSVC >= 1800
+#pragma warning(disable:26495)
+#endif
+#endif
+template <class Results>
+struct recursion_info
+{
+   typedef typename Results::value_type value_type;
+   typedef typename value_type::iterator iterator;
+   int idx;
+   const re_syntax_base* preturn_address;
+   Results results;
+   repeater_count<iterator>* repeater_stack;
+   iterator location_of_start;
+};
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(pop)
+#endif
+
+template <class BidiIterator, class Allocator, class traits>
+class perl_matcher
+{
+public:
+   typedef typename traits::char_type char_type;
+   typedef perl_matcher<BidiIterator, Allocator, traits> self_type;
+   typedef bool (self_type::*matcher_proc_type)();
+   typedef std::size_t traits_size_type;
+   typedef typename is_byte<char_type>::width_type width_type;
+   typedef typename std::iterator_traits<BidiIterator>::difference_type difference_type;
+   typedef match_results<BidiIterator, Allocator> results_type;
+
+   perl_matcher(BidiIterator first, BidiIterator end, 
+      match_results<BidiIterator, Allocator>& what, 
+      const basic_regex<char_type, traits>& e,
+      match_flag_type f,
+      BidiIterator l_base)
+      :  m_result(what), base(first), last(end), 
+         position(first), backstop(l_base), re(e), traits_inst(e.get_traits()), 
+         m_independent(false), next_count(&rep_obj), rep_obj(&next_count)
+      , m_recursions(0)
+   {
+      construct_init(e, f);
+   }
+
+   bool match();
+   bool find();
+
+   void setf(match_flag_type f)
+   { m_match_flags |= f; }
+   void unsetf(match_flag_type f)
+   { m_match_flags &= ~f; }
+
+private:
+   void construct_init(const basic_regex<char_type, traits>& e, match_flag_type f);
+
+   bool find_imp();
+   bool match_imp();
+   void estimate_max_state_count(std::random_access_iterator_tag*);
+   void estimate_max_state_count(void*);
+   bool match_prefix();
+   bool match_all_states();
+
+   // match procs, stored in s_match_vtable:
+   bool match_startmark();
+   bool match_endmark();
+   bool match_literal();
+   bool match_start_line();
+   bool match_end_line();
+   bool match_wild();
+   bool match_match();
+   bool match_word_boundary();
+   bool match_within_word();
+   bool match_word_start();
+   bool match_word_end();
+   bool match_buffer_start();
+   bool match_buffer_end();
+   bool match_backref();
+   bool match_long_set();
+   bool match_set();
+   bool match_jump();
+   bool match_alt();
+   bool match_rep();
+   bool match_combining();
+   bool match_soft_buffer_end();
+   bool match_restart_continue();
+   bool match_long_set_repeat();
+   bool match_set_repeat();
+   bool match_char_repeat();
+   bool match_dot_repeat_fast();
+   bool match_dot_repeat_slow();
+   bool match_dot_repeat_dispatch()
+   {
+      return ::boost::is_random_access_iterator<BidiIterator>::value ? match_dot_repeat_fast() : match_dot_repeat_slow();
+   }
+   bool match_backstep();
+   bool match_assert_backref();
+   bool match_toggle_case();
+   bool match_recursion();
+   bool match_fail();
+   bool match_accept();
+   bool match_commit();
+   bool match_then();
+   bool skip_until_paren(int index, bool match = true);
+
+   // find procs stored in s_find_vtable:
+   bool find_restart_any();
+   bool find_restart_word();
+   bool find_restart_line();
+   bool find_restart_buf();
+   bool find_restart_lit();
+
+private:
+   // final result structure to be filled in:
+   match_results<BidiIterator, Allocator>& m_result;
+   // temporary result for POSIX matches:
+   std::unique_ptr<match_results<BidiIterator, Allocator> > m_temp_match;
+   // pointer to actual result structure to fill in:
+   match_results<BidiIterator, Allocator>* m_presult;
+   // start of sequence being searched:
+   BidiIterator base;
+   // end of sequence being searched:
+   BidiIterator last; 
+   // current character being examined:
+   BidiIterator position;
+   // where to restart next search after failed match attempt:
+   BidiIterator restart;
+   // where the current search started from, acts as base for $` during grep:
+   BidiIterator search_base;
+   // how far we can go back when matching lookbehind:
+   BidiIterator backstop;
+   // the expression being examined:
+   const basic_regex<char_type, traits>& re;
+   // the expression's traits class:
+   const ::boost::regex_traits_wrapper<traits>& traits_inst;
+   // the next state in the machine being matched:
+   const re_syntax_base* pstate;
+   // matching flags in use:
+   match_flag_type m_match_flags;
+   // how many states we have examined so far:
+   std::ptrdiff_t state_count;
+   // max number of states to examine before giving up:
+   std::ptrdiff_t max_state_count;
+   // whether we should ignore case or not:
+   bool icase;
+   // set to true when (position == last), indicates that we may have a partial match:
+   bool m_has_partial_match;
+   // set to true whenever we get a match:
+   bool m_has_found_match;
+   // set to true whenever we're inside an independent sub-expression:
+   bool m_independent;
+   // the current repeat being examined:
+   repeater_count<BidiIterator>* next_count;
+   // the first repeat being examined (top of linked list):
+   repeater_count<BidiIterator> rep_obj;
+   // the mask to pass when matching word boundaries:
+   typename traits::char_class_type m_word_mask;
+   // the bitmask to use when determining whether a match_any matches a newline or not:
+   unsigned char match_any_mask;
+   // recursion information:
+   std::vector<recursion_info<results_type> > recursion_stack;
+   //
+   // additional members for non-recursive version:
+   //
+   typedef bool (self_type::*unwind_proc_type)(bool);
+
+   void extend_stack();
+   bool unwind(bool);
+   bool unwind_end(bool);
+   bool unwind_paren(bool);
+   bool unwind_recursion_stopper(bool);
+   bool unwind_assertion(bool);
+   bool unwind_alt(bool);
+   bool unwind_repeater_counter(bool);
+   bool unwind_extra_block(bool);
+   bool unwind_greedy_single_repeat(bool);
+   bool unwind_slow_dot_repeat(bool);
+   bool unwind_fast_dot_repeat(bool);
+   bool unwind_char_repeat(bool);
+   bool unwind_short_set_repeat(bool);
+   bool unwind_long_set_repeat(bool);
+   bool unwind_non_greedy_repeat(bool);
+   bool unwind_recursion(bool);
+   bool unwind_recursion_pop(bool);
+   bool unwind_commit(bool);
+   bool unwind_then(bool);
+   bool unwind_case(bool);
+   void destroy_single_repeat();
+   void push_matched_paren(int index, const sub_match<BidiIterator>& sub);
+   void push_recursion_stopper();
+   void push_assertion(const re_syntax_base* ps, bool positive);
+   void push_alt(const re_syntax_base* ps);
+   void push_repeater_count(int i, repeater_count<BidiIterator>** s);
+   void push_single_repeat(std::size_t c, const re_repeat* r, BidiIterator last_position, int state_id);
+   void push_non_greedy_repeat(const re_syntax_base* ps);
+   void push_recursion(int idx, const re_syntax_base* p, results_type* presults, results_type* presults2);
+   void push_recursion_pop();
+   void push_case_change(bool);
+
+   // pointer to base of stack:
+   saved_state* m_stack_base;
+   // pointer to current stack position:
+   saved_state* m_backup_state;
+   // how many memory blocks have we used up?:
+   unsigned used_block_count;
+   // determines what value to return when unwinding from recursion,
+   // allows for mixed recursive/non-recursive algorithm:
+   bool m_recursive_result;
+   // We have unwound to a lookahead/lookbehind, used by COMMIT/PRUNE/SKIP:
+   bool m_unwound_lookahead;
+   // We have unwound to an alternative, used by THEN:
+   bool m_unwound_alt;
+   // We are unwinding a commit - used by independent subs to determine whether to stop there or carry on unwinding:
+   //bool m_unwind_commit;
+   // Recursion limit:
+   unsigned m_recursions;
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(push)
+#if BOOST_REGEX_MSVC >= 1800
+#pragma warning(disable:26495)
+#endif
+#endif
+   // these operations aren't allowed, so are declared private,
+   // bodies are provided to keep explicit-instantiation requests happy:
+   perl_matcher& operator=(const perl_matcher&)
+   {
+      return *this;
+   }
+   perl_matcher(const perl_matcher& that)
+      : m_result(that.m_result), re(that.re), traits_inst(that.traits_inst), rep_obj(0) {}
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(pop)
+#endif
+};
+
+} // namespace BOOST_REGEX_DETAIL_NS
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(pop)
+#endif
+
+} // namespace boost
+
+//
+// include the implementation of perl_matcher:
+//
+#include <boost/regex/v5/perl_matcher_non_recursive.hpp>
+// this one has to be last:
+#include <boost/regex/v5/perl_matcher_common.hpp>
+
+#endif

regex/include/boost/regex/v5/perl_matcher_common.hpp

@@ -0,0 +1,921 @@
+/*
+ *
+ * Copyright (c) 2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         perl_matcher_common.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Definitions of perl_matcher member functions that are 
+  *                common to both the recursive and non-recursive versions.
+  */
+
+#ifndef BOOST_REGEX_V5_PERL_MATCHER_COMMON_HPP
+#define BOOST_REGEX_V5_PERL_MATCHER_COMMON_HPP
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(push)
+#pragma warning(disable:4459)
+#if BOOST_REGEX_MSVC < 1910
+#pragma warning(disable:4800)
+#endif
+#endif
+
+namespace boost{
+namespace BOOST_REGEX_DETAIL_NS{
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(push)
+#pragma warning(disable:26812)
+#endif
+   template <class BidiIterator, class Allocator, class traits>
+void perl_matcher<BidiIterator, Allocator, traits>::construct_init(const basic_regex<char_type, traits>& e, match_flag_type f)
+{ 
+   typedef typename std::iterator_traits<BidiIterator>::iterator_category category;
+   typedef typename basic_regex<char_type, traits>::flag_type expression_flag_type;
+   
+   if(e.empty())
+   {
+      // precondition failure: e is not a valid regex.
+      std::invalid_argument ex("Invalid regular expression object");
+#ifndef BOOST_REGEX_STANDALONE
+      boost::throw_exception(ex);
+#else
+      throw e;
+#endif
+   }
+   pstate = 0;
+   m_match_flags = f;
+   estimate_max_state_count(static_cast<category*>(0));
+   expression_flag_type re_f = re.flags();
+   icase = re_f & regex_constants::icase;
+   if(!(m_match_flags & (match_perl|match_posix)))
+   {
+      if((re_f & (regbase::main_option_type|regbase::no_perl_ex)) == 0)
+         m_match_flags |= match_perl;
+      else if((re_f & (regbase::main_option_type|regbase::emacs_ex)) == (regbase::basic_syntax_group|regbase::emacs_ex))
+         m_match_flags |= match_perl;
+      else if((re_f & (regbase::main_option_type|regbase::literal)) == (regbase::literal))
+         m_match_flags |= match_perl;
+      else
+         m_match_flags |= match_posix;
+   }
+   if(m_match_flags & match_posix)
+   {
+      m_temp_match.reset(new match_results<BidiIterator, Allocator>());
+      m_presult = m_temp_match.get();
+   }
+   else
+      m_presult = &m_result;
+   m_stack_base = 0;
+   m_backup_state = 0;
+   // find the value to use for matching word boundaries:
+   m_word_mask = re.get_data().m_word_mask; 
+   // find bitmask to use for matching '.':
+   match_any_mask = static_cast<unsigned char>((f & match_not_dot_newline) ? BOOST_REGEX_DETAIL_NS::test_not_newline : BOOST_REGEX_DETAIL_NS::test_newline);
+   // Disable match_any if requested in the state machine:
+   if(e.get_data().m_disable_match_any)
+      m_match_flags &= regex_constants::match_not_any;
+}
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(pop)
+#endif
+
+template <class BidiIterator, class Allocator, class traits>
+void perl_matcher<BidiIterator, Allocator, traits>::estimate_max_state_count(std::random_access_iterator_tag*)
+{
+   //
+   // How many states should we allow our machine to visit before giving up?
+   // This is a heuristic: it takes the greater of O(N^2) and O(NS^2)
+   // where N is the length of the string, and S is the number of states
+   // in the machine.  It's tempting to up this to O(N^2S) or even O(N^2S^2)
+   // but these take unreasonably amounts of time to bale out in pathological
+   // cases.
+   //
+   // Calculate NS^2 first:
+   //
+   static const std::ptrdiff_t k = 100000;
+   std::ptrdiff_t dist = std::distance(base, last);
+   if(dist == 0)
+      dist = 1;
+   std::ptrdiff_t states = re.size();
+   if(states == 0)
+      states = 1;
+   if ((std::numeric_limits<std::ptrdiff_t>::max)() / states < states)
+   {
+      max_state_count = (std::min)((std::ptrdiff_t)BOOST_REGEX_MAX_STATE_COUNT, (std::numeric_limits<std::ptrdiff_t>::max)() - 2);
+      return;
+   }
+   states *= states;
+   if((std::numeric_limits<std::ptrdiff_t>::max)() / dist < states)
+   {
+      max_state_count = (std::min)((std::ptrdiff_t)BOOST_REGEX_MAX_STATE_COUNT, (std::numeric_limits<std::ptrdiff_t>::max)() - 2);
+      return;
+   }
+   states *= dist;
+   if((std::numeric_limits<std::ptrdiff_t>::max)() - k < states)
+   {
+      max_state_count = (std::min)((std::ptrdiff_t)BOOST_REGEX_MAX_STATE_COUNT, (std::numeric_limits<std::ptrdiff_t>::max)() - 2);
+      return;
+   }
+   states += k;
+
+   max_state_count = states;
+
+   //
+   // Now calculate N^2:
+   //
+   states = dist;
+   if((std::numeric_limits<std::ptrdiff_t>::max)() / dist < states)
+   {
+      max_state_count = (std::min)((std::ptrdiff_t)BOOST_REGEX_MAX_STATE_COUNT, (std::numeric_limits<std::ptrdiff_t>::max)() - 2);
+      return;
+   }
+   states *= dist;
+   if((std::numeric_limits<std::ptrdiff_t>::max)() - k < states)
+   {
+      max_state_count = (std::min)((std::ptrdiff_t)BOOST_REGEX_MAX_STATE_COUNT, (std::numeric_limits<std::ptrdiff_t>::max)() - 2);
+      return;
+   }
+   states += k;
+   //
+   // N^2 can be a very large number indeed, to prevent things getting out
+   // of control, cap the max states:
+   //
+   if(states > BOOST_REGEX_MAX_STATE_COUNT)
+      states = BOOST_REGEX_MAX_STATE_COUNT;
+   //
+   // If (the possibly capped) N^2 is larger than our first estimate,
+   // use this instead:
+   //
+   if(states > max_state_count)
+      max_state_count = states;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+inline void perl_matcher<BidiIterator, Allocator, traits>::estimate_max_state_count(void*)
+{
+   // we don't know how long the sequence is:
+   max_state_count = BOOST_REGEX_MAX_STATE_COUNT;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+inline bool perl_matcher<BidiIterator, Allocator, traits>::match()
+{
+   return match_imp();
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_imp()
+{
+   // initialise our stack if we are non-recursive:
+   save_state_init init(&m_stack_base, &m_backup_state);
+   used_block_count = BOOST_REGEX_MAX_BLOCKS;
+#if !defined(BOOST_NO_EXCEPTIONS)
+   try{
+#endif
+
+   // reset our state machine:
+   position = base;
+   search_base = base;
+   state_count = 0;
+   m_match_flags |= regex_constants::match_all;
+   m_presult->set_size((m_match_flags & match_nosubs) ? 1u : static_cast<typename results_type::size_type>(1u + re.mark_count()), search_base, last);
+   m_presult->set_base(base);
+   m_presult->set_named_subs(this->re.get_named_subs());
+   if(m_match_flags & match_posix)
+      m_result = *m_presult;
+   verify_options(re.flags(), m_match_flags);
+   if(0 == match_prefix())
+      return false;
+   return (m_result[0].second == last) && (m_result[0].first == base);
+
+#if !defined(BOOST_NO_EXCEPTIONS)
+   }
+   catch(...)
+   {
+      // unwind all pushed states, apart from anything else this
+      // ensures that all the states are correctly destructed
+      // not just the memory freed.
+      while(unwind(true)){}
+      throw;
+   }
+#endif
+}
+
+template <class BidiIterator, class Allocator, class traits>
+inline bool perl_matcher<BidiIterator, Allocator, traits>::find()
+{
+   return find_imp();
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::find_imp()
+{
+   static matcher_proc_type const s_find_vtable[7] = 
+   {
+      &perl_matcher<BidiIterator, Allocator, traits>::find_restart_any,
+      &perl_matcher<BidiIterator, Allocator, traits>::find_restart_word,
+      &perl_matcher<BidiIterator, Allocator, traits>::find_restart_line,
+      &perl_matcher<BidiIterator, Allocator, traits>::find_restart_buf,
+      &perl_matcher<BidiIterator, Allocator, traits>::match_prefix,
+      &perl_matcher<BidiIterator, Allocator, traits>::find_restart_lit,
+      &perl_matcher<BidiIterator, Allocator, traits>::find_restart_lit,
+   };
+
+   // initialise our stack if we are non-recursive:
+   save_state_init init(&m_stack_base, &m_backup_state);
+   used_block_count = BOOST_REGEX_MAX_BLOCKS;
+#if !defined(BOOST_NO_EXCEPTIONS)
+   try{
+#endif
+
+   state_count = 0;
+   if((m_match_flags & regex_constants::match_init) == 0)
+   {
+      // reset our state machine:
+      search_base = position = base;
+      pstate = re.get_first_state();
+      m_presult->set_size((m_match_flags & match_nosubs) ? 1u : static_cast<typename results_type::size_type>(1u + re.mark_count()), base, last);
+      m_presult->set_base(base);
+      m_presult->set_named_subs(this->re.get_named_subs());
+      m_match_flags |= regex_constants::match_init;
+   }
+   else
+   {
+      // start again:
+      search_base = position = m_result[0].second;
+      // If last match was null and match_not_null was not set then increment
+      // our start position, otherwise we go into an infinite loop:
+      if(((m_match_flags & match_not_null) == 0) && (m_result.length() == 0))
+      {
+         if(position == last)
+            return false;
+         else 
+            ++position;
+      }
+      // reset $` start:
+      m_presult->set_size((m_match_flags & match_nosubs) ? 1u : static_cast<typename results_type::size_type>(1u + re.mark_count()), search_base, last);
+      //if((base != search_base) && (base == backstop))
+      //   m_match_flags |= match_prev_avail;
+   }
+   if(m_match_flags & match_posix)
+   {
+      m_result.set_size(static_cast<typename results_type::size_type>(1u + re.mark_count()), base, last);
+      m_result.set_base(base);
+   }
+
+   verify_options(re.flags(), m_match_flags);
+   // find out what kind of expression we have:
+   unsigned type = (m_match_flags & match_continuous) ? 
+      static_cast<unsigned int>(regbase::restart_continue) 
+         : static_cast<unsigned int>(re.get_restart_type());
+
+   // call the appropriate search routine:
+   matcher_proc_type proc = s_find_vtable[type];
+   return (this->*proc)();
+
+#if !defined(BOOST_NO_EXCEPTIONS)
+   }
+   catch(...)
+   {
+      // unwind all pushed states, apart from anything else this
+      // ensures that all the states are correctly destructed
+      // not just the memory freed.
+      while(unwind(true)){}
+      throw;
+   }
+#endif
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_prefix()
+{
+   m_has_partial_match = false;
+   m_has_found_match = false;
+   pstate = re.get_first_state();
+   m_presult->set_first(position);
+   restart = position;
+   match_all_states();
+   if(!m_has_found_match && m_has_partial_match && (m_match_flags & match_partial))
+   {
+      m_has_found_match = true;
+      m_presult->set_second(last, 0, false);
+      position = last;
+      if((m_match_flags & match_posix) == match_posix)
+      {
+         m_result.maybe_assign(*m_presult);
+      }
+   }
+#ifdef BOOST_REGEX_MATCH_EXTRA
+   if(m_has_found_match && (match_extra & m_match_flags))
+   {
+      //
+      // we have a match, reverse the capture information:
+      //
+      for(unsigned i = 0; i < m_presult->size(); ++i)
+      {
+         typename sub_match<BidiIterator>::capture_sequence_type & seq = ((*m_presult)[i]).get_captures();
+         std::reverse(seq.begin(), seq.end());
+      }
+   }
+#endif
+   if(!m_has_found_match)
+      position = restart; // reset search postion
+   return m_has_found_match;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_literal()
+{
+   unsigned int len = static_cast<const re_literal*>(pstate)->length;
+   const char_type* what = reinterpret_cast<const char_type*>(static_cast<const re_literal*>(pstate) + 1);
+   //
+   // compare string with what we stored in
+   // our records:
+   for(unsigned int i = 0; i < len; ++i, ++position)
+   {
+      if((position == last) || (traits_inst.translate(*position, icase) != what[i]))
+         return false;
+   }
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_start_line()
+{
+   if(position == backstop)
+   {
+      if((m_match_flags & match_prev_avail) == 0)
+      {
+         if((m_match_flags & match_not_bol) == 0)
+         {
+            pstate = pstate->next.p;
+            return true;
+         }
+         return false;
+      }
+   }
+   else if(m_match_flags & match_single_line)
+      return false;
+
+   // check the previous value character:
+   BidiIterator t(position);
+   --t;
+   if(position != last)
+   {
+      if(is_separator(*t) && !((*t == static_cast<char_type>('\r')) && (*position == static_cast<char_type>('\n'))) )
+      {
+         pstate = pstate->next.p;
+         return true;
+      }
+   }
+   else if(is_separator(*t))
+   {
+      pstate = pstate->next.p;
+      return true;
+   }
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_end_line()
+{
+   if(position != last)
+   {
+      if(m_match_flags & match_single_line)
+         return false;
+      // we're not yet at the end so *first is always valid:
+      if(is_separator(*position))
+      {
+         if((position != backstop) || (m_match_flags & match_prev_avail))
+         {
+            // check that we're not in the middle of \r\n sequence
+            BidiIterator t(position);
+            --t;
+            if((*t == static_cast<char_type>('\r')) && (*position == static_cast<char_type>('\n')))
+            {
+               return false;
+            }
+         }
+         pstate = pstate->next.p;
+         return true;
+      }
+   }
+   else if((m_match_flags & match_not_eol) == 0)
+   {
+      pstate = pstate->next.p;
+      return true;
+   }
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_wild()
+{
+   if(position == last) 
+      return false;
+   if(is_separator(*position) && ((match_any_mask & static_cast<const re_dot*>(pstate)->mask) == 0))
+      return false;
+   if((*position == char_type(0)) && (m_match_flags & match_not_dot_null))
+      return false;
+   pstate = pstate->next.p;
+   ++position;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_word_boundary()
+{
+   bool b; // indcates whether next character is a word character
+   if(position != last)
+   {
+      // prev and this character must be opposites:
+      b = traits_inst.isctype(*position, m_word_mask);
+   }
+   else
+   {
+      if (m_match_flags & match_not_eow)
+         return false;
+      b = false;
+   }
+   if((position == backstop) && ((m_match_flags & match_prev_avail) == 0))
+   {
+      if(m_match_flags & match_not_bow)
+         return false;
+      else
+         b ^= false;
+   }
+   else
+   {
+      --position;
+      b ^= traits_inst.isctype(*position, m_word_mask);
+      ++position;
+   }
+   if(b)
+   {
+      pstate = pstate->next.p;
+      return true;
+   }
+   return false; // no match if we get to here...
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_within_word()
+{
+   bool b = !match_word_boundary();
+   if(b)
+      pstate = pstate->next.p;
+   return b;
+   /*
+   if(position == last)
+      return false;
+   // both prev and this character must be m_word_mask:
+   bool prev = traits_inst.isctype(*position, m_word_mask);
+   {
+      bool b;
+      if((position == backstop) && ((m_match_flags & match_prev_avail) == 0)) 
+         return false;
+      else
+      {
+         --position;
+         b = traits_inst.isctype(*position, m_word_mask);
+         ++position;
+      }
+      if(b == prev)
+      {
+         pstate = pstate->next.p;
+         return true;
+      }
+   }
+   return false;
+   */
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_word_start()
+{
+   if(position == last)
+      return false; // can't be starting a word if we're already at the end of input
+   if(!traits_inst.isctype(*position, m_word_mask))
+      return false; // next character isn't a word character
+   if((position == backstop) && ((m_match_flags & match_prev_avail) == 0))
+   {
+      if(m_match_flags & match_not_bow)
+         return false; // no previous input
+   }
+   else
+   {
+      // otherwise inside buffer:
+      BidiIterator t(position);
+      --t;
+      if(traits_inst.isctype(*t, m_word_mask))
+         return false; // previous character not non-word
+   }
+   // OK we have a match:
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_word_end()
+{
+   if((position == backstop) && ((m_match_flags & match_prev_avail) == 0))
+      return false;  // start of buffer can't be end of word
+   BidiIterator t(position);
+   --t;
+   if(traits_inst.isctype(*t, m_word_mask) == false)
+      return false;  // previous character wasn't a word character
+
+   if(position == last)
+   {
+      if(m_match_flags & match_not_eow)
+         return false; // end of buffer but not end of word
+   }
+   else
+   {
+      // otherwise inside buffer:
+      if(traits_inst.isctype(*position, m_word_mask))
+         return false; // next character is a word character
+   }
+   pstate = pstate->next.p;
+   return true;      // if we fall through to here then we've succeeded
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_buffer_start()
+{
+   if((position != backstop) || (m_match_flags & match_not_bob))
+      return false;
+   // OK match:
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_buffer_end()
+{
+   if((position != last) || (m_match_flags & match_not_eob))
+      return false;
+   // OK match:
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_backref()
+{
+   //
+   // Compare with what we previously matched.
+   // Note that this succeeds if the backref did not partisipate
+   // in the match, this is in line with ECMAScript, but not Perl
+   // or PCRE.
+   //
+   int index = static_cast<const re_brace*>(pstate)->index;
+   if(index >= hash_value_mask)
+   {
+      named_subexpressions::range_type r = re.get_data().equal_range(index);
+      BOOST_REGEX_ASSERT(r.first != r.second);
+      do
+      {
+         index = r.first->index;
+         ++r.first;
+      }while((r.first != r.second) && ((*m_presult)[index].matched != true));
+   }
+
+   if((m_match_flags & match_perl) && !(*m_presult)[index].matched)
+      return false;
+
+   BidiIterator i = (*m_presult)[index].first;
+   BidiIterator j = (*m_presult)[index].second;
+   while(i != j)
+   {
+      if((position == last) || (traits_inst.translate(*position, icase) != traits_inst.translate(*i, icase)))
+         return false;
+      ++i;
+      ++position;
+   }
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_long_set()
+{
+   typedef typename traits::char_class_type char_class_type;
+   // let the traits class do the work:
+   if(position == last)
+      return false;
+   BidiIterator t = re_is_set_member(position, last, static_cast<const re_set_long<char_class_type>*>(pstate), re.get_data(), icase);
+   if(t != position)
+   {
+      pstate = pstate->next.p;
+      position = t;
+      return true;
+   }
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_set()
+{
+   if(position == last)
+      return false;
+   if(static_cast<const re_set*>(pstate)->_map[static_cast<unsigned char>(traits_inst.translate(*position, icase))])
+   {
+      pstate = pstate->next.p;
+      ++position;
+      return true;
+   }
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_jump()
+{
+   pstate = static_cast<const re_jump*>(pstate)->alt.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_combining()
+{
+   if(position == last)
+      return false;
+   if(is_combining(traits_inst.translate(*position, icase)))
+      return false;
+   ++position;
+   while((position != last) && is_combining(traits_inst.translate(*position, icase)))
+      ++position;
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_soft_buffer_end()
+{
+   if(m_match_flags & match_not_eob)
+      return false;
+   BidiIterator p(position);
+   while((p != last) && is_separator(traits_inst.translate(*p, icase)))++p;
+   if(p != last)
+      return false;
+   pstate = pstate->next.p;
+   return true;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_restart_continue()
+{
+   if(position == search_base)
+   {
+      pstate = pstate->next.p;
+      return true;
+   }
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_backstep()
+{
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable:4127)
+#endif
+   if( ::boost::is_random_access_iterator<BidiIterator>::value)
+   {
+      std::ptrdiff_t maxlen = std::distance(backstop, position);
+      if(maxlen < static_cast<const re_brace*>(pstate)->index)
+         return false;
+      std::advance(position, -static_cast<const re_brace*>(pstate)->index);
+   }
+   else
+   {
+      int c = static_cast<const re_brace*>(pstate)->index;
+      while(c--)
+      {
+         if(position == backstop)
+            return false;
+         --position;
+      }
+   }
+   pstate = pstate->next.p;
+   return true;
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+}
+
+template <class BidiIterator, class Allocator, class traits>
+inline bool perl_matcher<BidiIterator, Allocator, traits>::match_assert_backref()
+{
+   // return true if marked sub-expression N has been matched:
+   int index = static_cast<const re_brace*>(pstate)->index;
+   bool result = false;
+   if(index == 9999)
+   {
+      // Magic value for a (DEFINE) block:
+      return false;
+   }
+   else if(index > 0)
+   {
+      // Have we matched subexpression "index"?
+      // Check if index is a hash value:
+      if(index >= hash_value_mask)
+      {
+         named_subexpressions::range_type r = re.get_data().equal_range(index);
+         while(r.first != r.second)
+         {
+            if((*m_presult)[r.first->index].matched)
+            {
+               result = true;
+               break;
+            }
+            ++r.first;
+         }
+      }
+      else
+      {
+         result = (*m_presult)[index].matched;
+      }
+      pstate = pstate->next.p;
+   }
+   else
+   {
+      // Have we recursed into subexpression "index"?
+      // If index == 0 then check for any recursion at all, otherwise for recursion to -index-1.
+      int idx = -(index+1);
+      if(idx >= hash_value_mask)
+      {
+         named_subexpressions::range_type r = re.get_data().equal_range(idx);
+         int stack_index = recursion_stack.empty() ? -1 : recursion_stack.back().idx;
+         while(r.first != r.second)
+         {
+            result |= (stack_index == r.first->index);
+            if(result)break;
+            ++r.first;
+         }
+      }
+      else
+      {
+         result = !recursion_stack.empty() && ((recursion_stack.back().idx == idx) || (index == 0));
+      }
+      pstate = pstate->next.p;
+   }
+   return result;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_fail()
+{
+   // Just force a backtrack:
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::match_accept()
+{
+   if(!recursion_stack.empty())
+   {
+      return skip_until_paren(recursion_stack.back().idx);
+   }
+   else
+   {
+      return skip_until_paren(INT_MAX);
+   }
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_any()
+{
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable:4127)
+#endif
+   const unsigned char* _map = re.get_map();
+   while(true)
+   {
+      // skip everything we can't match:
+      while((position != last) && !can_start(*position, _map, (unsigned char)mask_any) )
+         ++position;
+      if(position == last)
+      {
+         // run out of characters, try a null match if possible:
+         if(re.can_be_null())
+            return match_prefix();
+         break;
+      }
+      // now try and obtain a match:
+      if(match_prefix())
+         return true;
+      if(position == last)
+         return false;
+      ++position;
+   }
+   return false;
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_word()
+{
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(push)
+#pragma warning(disable:4127)
+#endif
+   // do search optimised for word starts:
+   const unsigned char* _map = re.get_map();
+   if((m_match_flags & match_prev_avail) || (position != base))
+      --position;
+   else if(match_prefix())
+      return true;
+   do
+   {
+      while((position != last) && traits_inst.isctype(*position, m_word_mask))
+         ++position;
+      while((position != last) && !traits_inst.isctype(*position, m_word_mask))
+         ++position;
+      if(position == last)
+         break;
+
+      if(can_start(*position, _map, (unsigned char)mask_any) )
+      {
+         if(match_prefix())
+            return true;
+      }
+      if(position == last)
+         break;
+   } while(true);
+   return false;
+#ifdef BOOST_REGEX_MSVC
+#pragma warning(pop)
+#endif
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_line()
+{
+   // do search optimised for line starts:
+   const unsigned char* _map = re.get_map();
+   if(match_prefix())
+      return true;
+   while(position != last)
+   {
+      while((position != last) && !is_separator(*position))
+         ++position;
+      if(position == last)
+         return false;
+      ++position;
+      if(position == last)
+      {
+         if(re.can_be_null() && match_prefix())
+            return true;
+         return false;
+      }
+
+      if( can_start(*position, _map, (unsigned char)mask_any) )
+      {
+         if(match_prefix())
+            return true;
+      }
+      if(position == last)
+         return false;
+      //++position;
+   }
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_buf()
+{
+   if((position == base) && ((m_match_flags & match_not_bob) == 0))
+      return match_prefix();
+   return false;
+}
+
+template <class BidiIterator, class Allocator, class traits>
+bool perl_matcher<BidiIterator, Allocator, traits>::find_restart_lit()
+{
+   return false;
+}
+
+} // namespace BOOST_REGEX_DETAIL_NS
+
+} // namespace boost
+
+#ifdef BOOST_REGEX_MSVC
+#  pragma warning(pop)
+#endif
+
+#endif
+

regex/include/boost/regex/v5/primary_transform.hpp

@@ -0,0 +1,120 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+ 
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE:        primary_transform.hpp
+  *   VERSION:     see <boost/version.hpp>
+  *   DESCRIPTION: Heuristically determines the sort string format in use
+  *                by the current locale.
+  */
+
+#ifndef BOOST_REGEX_PRIMARY_TRANSFORM
+#define BOOST_REGEX_PRIMARY_TRANSFORM
+
+namespace boost{
+   namespace BOOST_REGEX_DETAIL_NS{
+
+
+enum{
+   sort_C,
+   sort_fixed,
+   sort_delim,
+   sort_unknown
+};
+
+template <class S, class charT>
+unsigned count_chars(const S& s, charT c)
+{
+   //
+   // Count how many occurrences of character c occur
+   // in string s: if c is a delimeter between collation
+   // fields, then this should be the same value for all
+   // sort keys:
+   //
+   unsigned int count = 0;
+   for(unsigned pos = 0; pos < s.size(); ++pos)
+   {
+      if(s[pos] == c) ++count;
+   }
+   return count;
+}
+
+
+template <class traits, class charT>
+unsigned find_sort_syntax(const traits* pt, charT* delim)
+{
+   //
+   // compare 'a' with 'A' to see how similar they are,
+   // should really use a-accute but we can't portably do that,
+   //
+   typedef typename traits::string_type string_type;
+   typedef typename traits::char_type char_type;
+
+   // Suppress incorrect warning for MSVC
+   (void)pt;
+
+   char_type a[2] = {'a', '\0', };
+   string_type sa(pt->transform(a, a+1));
+   if(sa == a)
+   {
+      *delim = 0;
+      return sort_C;
+   }
+   char_type A[2] = { 'A', '\0', };
+   string_type sA(pt->transform(A, A+1));
+   char_type c[2] = { ';', '\0', };
+   string_type sc(pt->transform(c, c+1));
+
+   int pos = 0;
+   while((pos <= static_cast<int>(sa.size())) && (pos <= static_cast<int>(sA.size())) && (sa[pos] == sA[pos])) ++pos;
+   --pos;
+   if(pos < 0)
+   {
+      *delim = 0;
+      return sort_unknown;
+   }
+   //
+   // at this point sa[pos] is either the end of a fixed width field
+   // or the character that acts as a delimiter:
+   //
+   charT maybe_delim = sa[pos];
+   if((pos != 0) && (count_chars(sa, maybe_delim) == count_chars(sA, maybe_delim)) && (count_chars(sa, maybe_delim) == count_chars(sc, maybe_delim)))
+   {
+      *delim = maybe_delim;
+      return sort_delim;
+   }
+   //
+   // OK doen't look like a delimiter, try for fixed width field:
+   //
+   if((sa.size() == sA.size()) && (sa.size() == sc.size()))
+   {
+      // note assumes that the fixed width field is less than
+      // (numeric_limits<charT>::max)(), should be true for all types
+      // I can't imagine 127 character fields...
+      *delim = static_cast<charT>(++pos);
+      return sort_fixed;
+   }
+   //
+   // don't know what it is:
+   //
+   *delim = 0;
+   return sort_unknown;
+}
+
+
+   } // namespace BOOST_REGEX_DETAIL_NS
+} // namespace boost
+
+#endif
+
+
+

regex/include/boost/regex/v5/regbase.hpp

@@ -0,0 +1,158 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regbase.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares class regbase.
+  */
+
+#ifndef BOOST_REGEX_V5_REGBASE_HPP
+#define BOOST_REGEX_V5_REGBASE_HPP
+
+namespace boost{
+//
+// class regbase
+// handles error codes and flags
+//
+class regbase
+{
+public:
+   enum flag_type_
+   {
+      //
+      // Divide the flags up into logical groups:
+      // bits 0-7 indicate main synatx type.
+      // bits 8-15 indicate syntax subtype.
+      // bits 16-31 indicate options that are common to all
+      // regex syntaxes.
+      // In all cases the default is 0.
+      //
+      // Main synatx group:
+      //
+      perl_syntax_group = 0,                      // default
+      basic_syntax_group = 1,                     // POSIX basic
+      literal = 2,                                // all characters are literals
+      main_option_type = literal | basic_syntax_group | perl_syntax_group, // everything!
+      //
+      // options specific to perl group:
+      //
+      no_bk_refs = 1 << 8,                        // \d not allowed
+      no_perl_ex = 1 << 9,                        // disable perl extensions
+      no_mod_m = 1 << 10,                         // disable Perl m modifier
+      mod_x = 1 << 11,                            // Perl x modifier
+      mod_s = 1 << 12,                            // force s modifier on (overrides match_not_dot_newline)
+      no_mod_s = 1 << 13,                         // force s modifier off (overrides match_not_dot_newline)
+
+      //
+      // options specific to basic group:
+      //
+      no_char_classes = 1 << 8,                   // [[:CLASS:]] not allowed
+      no_intervals = 1 << 9,                      // {x,y} not allowed
+      bk_plus_qm = 1 << 10,                       // uses \+ and \?
+      bk_vbar = 1 << 11,                          // use \| for alternatives
+      emacs_ex = 1 << 12,                         // enables emacs extensions
+
+      //
+      // options common to all groups:
+      //
+      no_escape_in_lists = 1 << 16,                     // '\' not special inside [...]
+      newline_alt = 1 << 17,                            // \n is the same as |
+      no_except = 1 << 18,                              // no exception on error
+      failbit = 1 << 19,                                // error flag
+      icase = 1 << 20,                                  // characters are matched regardless of case
+      nocollate = 0,                                    // don't use locale specific collation (deprecated)
+      collate = 1 << 21,                                // use locale specific collation
+      nosubs = 1 << 22,                                 // don't mark sub-expressions
+      save_subexpression_location = 1 << 23,            // save subexpression locations
+      no_empty_expressions = 1 << 24,                   // no empty expressions allowed
+      optimize = 0,                                     // not really supported
+      
+
+
+      basic = basic_syntax_group | collate | no_escape_in_lists,
+      extended = no_bk_refs | collate | no_perl_ex | no_escape_in_lists,
+      normal = 0,
+      emacs = basic_syntax_group | collate | emacs_ex | bk_vbar,
+      awk = no_bk_refs | collate | no_perl_ex,
+      grep = basic | newline_alt,
+      egrep = extended | newline_alt,
+      sed = basic,
+      perl = normal,
+      ECMAScript = normal,
+      JavaScript = normal,
+      JScript = normal
+   };
+   typedef unsigned int flag_type;
+
+   enum restart_info
+   {
+      restart_any = 0,
+      restart_word = 1,
+      restart_line = 2,
+      restart_buf = 3,
+      restart_continue = 4,
+      restart_lit = 5,
+      restart_fixed_lit = 6, 
+      restart_count = 7
+   };
+};
+
+//
+// provide std lib proposal compatible constants:
+//
+namespace regex_constants{
+
+   enum flag_type_
+   {
+
+      no_except = ::boost::regbase::no_except,
+      failbit = ::boost::regbase::failbit,
+      literal = ::boost::regbase::literal,
+      icase = ::boost::regbase::icase,
+      nocollate = ::boost::regbase::nocollate,
+      collate = ::boost::regbase::collate,
+      nosubs = ::boost::regbase::nosubs,
+      optimize = ::boost::regbase::optimize,
+      bk_plus_qm = ::boost::regbase::bk_plus_qm,
+      bk_vbar = ::boost::regbase::bk_vbar,
+      no_intervals = ::boost::regbase::no_intervals,
+      no_char_classes = ::boost::regbase::no_char_classes,
+      no_escape_in_lists = ::boost::regbase::no_escape_in_lists,
+      no_mod_m = ::boost::regbase::no_mod_m,
+      mod_x = ::boost::regbase::mod_x,
+      mod_s = ::boost::regbase::mod_s,
+      no_mod_s = ::boost::regbase::no_mod_s,
+      save_subexpression_location = ::boost::regbase::save_subexpression_location,
+      no_empty_expressions = ::boost::regbase::no_empty_expressions,
+
+      basic = ::boost::regbase::basic,
+      extended = ::boost::regbase::extended,
+      normal = ::boost::regbase::normal,
+      emacs = ::boost::regbase::emacs,
+      awk = ::boost::regbase::awk,
+      grep = ::boost::regbase::grep,
+      egrep = ::boost::regbase::egrep,
+      sed = basic,
+      perl = normal,
+      ECMAScript = normal,
+      JavaScript = normal,
+      JScript = normal
+   };
+   typedef ::boost::regbase::flag_type syntax_option_type;
+
+} // namespace regex_constants
+
+} // namespace boost
+
+#endif
+

regex/include/boost/regex/v5/regex.hpp

@@ -0,0 +1,106 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Declares boost::basic_regex<> and associated
+  *                functions and classes. This header is the main
+  *                entry point for the template regex code.
+  */
+
+#ifndef BOOST_RE_REGEX_HPP_INCLUDED
+#define BOOST_RE_REGEX_HPP_INCLUDED
+
+#ifdef __cplusplus
+
+// what follows is all C++ don't include in C builds!!
+
+#include <boost/regex/config.hpp>
+#include <boost/regex/v5/regex_workaround.hpp>
+#include <boost/regex_fwd.hpp>
+#include <boost/regex/regex_traits.hpp>
+#include <boost/regex/v5/error_type.hpp>
+#include <boost/regex/v5/match_flags.hpp>
+#include <boost/regex/v5/regex_raw_buffer.hpp>
+#include <boost/regex/pattern_except.hpp>
+#include <boost/regex/v5/char_regex_traits.hpp>
+#include <boost/regex/v5/states.hpp>
+#include <boost/regex/v5/regbase.hpp>
+#include <boost/regex/v5/basic_regex.hpp>
+#include <boost/regex/v5/basic_regex_creator.hpp>
+#include <boost/regex/v5/basic_regex_parser.hpp>
+#include <boost/regex/v5/sub_match.hpp>
+#include <boost/regex/v5/regex_format.hpp>
+#include <boost/regex/v5/match_results.hpp>
+#include <boost/regex/v5/perl_matcher.hpp>
+
+namespace boost{
+#ifdef BOOST_REGEX_NO_FWD
+typedef basic_regex<char, regex_traits<char> > regex;
+#ifndef BOOST_NO_WREGEX
+typedef basic_regex<wchar_t, regex_traits<wchar_t> > wregex;
+#endif
+#endif
+
+typedef match_results<const char*> cmatch;
+typedef match_results<std::string::const_iterator> smatch;
+#ifndef BOOST_NO_WREGEX
+typedef match_results<const wchar_t*> wcmatch;
+typedef match_results<std::wstring::const_iterator> wsmatch;
+#endif
+
+} // namespace boost
+
+#include <boost/regex/v5/regex_match.hpp>
+#include <boost/regex/v5/regex_search.hpp>
+#include <boost/regex/v5/regex_iterator.hpp>
+#include <boost/regex/v5/regex_token_iterator.hpp>
+#include <boost/regex/v5/regex_grep.hpp>
+#include <boost/regex/v5/regex_replace.hpp>
+#include <boost/regex/v5/regex_merge.hpp>
+#include <boost/regex/v5/regex_split.hpp>
+
+#endif  // __cplusplus
+
+#endif  // include
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

regex/include/boost/regex/v5/regex_fwd.hpp

@@ -0,0 +1,73 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_fwd.cpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Forward declares boost::basic_regex<> and
+  *                associated typedefs.
+  */
+
+#ifndef BOOST_REGEX_FWD_HPP_INCLUDED
+#define BOOST_REGEX_FWD_HPP_INCLUDED
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+
+//
+// define BOOST_REGEX_NO_FWD if this
+// header doesn't work!
+//
+#ifdef BOOST_REGEX_NO_FWD
+#  ifndef BOOST_RE_REGEX_HPP
+#     include <boost/regex.hpp>
+#  endif
+#else
+
+namespace boost{
+
+template <class charT>
+class cpp_regex_traits;
+template <class charT>
+struct c_regex_traits;
+template <class charT>
+class w32_regex_traits;
+
+#ifdef BOOST_REGEX_USE_WIN32_LOCALE
+template <class charT, class implementationT = w32_regex_traits<charT> >
+struct regex_traits;
+#elif defined(BOOST_REGEX_USE_CPP_LOCALE)
+template <class charT, class implementationT = cpp_regex_traits<charT> >
+struct regex_traits;
+#else
+template <class charT, class implementationT = c_regex_traits<charT> >
+struct regex_traits;
+#endif
+
+template <class charT, class traits = regex_traits<charT> >
+class basic_regex;
+
+typedef basic_regex<char, regex_traits<char> > regex;
+#ifndef BOOST_NO_WREGEX
+typedef basic_regex<wchar_t, regex_traits<wchar_t> > wregex;
+#endif
+
+} // namespace boost
+
+#endif  // BOOST_REGEX_NO_FWD
+
+#endif
+
+
+
+

regex/include/boost/regex/v5/regex_grep.hpp

@@ -0,0 +1,98 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_grep.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Provides regex_grep implementation.
+  */
+
+#ifndef BOOST_REGEX_V5_REGEX_GREP_HPP
+#define BOOST_REGEX_V5_REGEX_GREP_HPP
+
+
+namespace boost{
+
+//
+// regex_grep:
+// find all non-overlapping matches within the sequence first last:
+//
+template <class Predicate, class BidiIterator, class charT, class traits>
+inline unsigned int regex_grep(Predicate foo, 
+                               BidiIterator first, 
+                               BidiIterator last, 
+                               const basic_regex<charT, traits>& e, 
+                               match_flag_type flags = match_default)
+{
+   if(e.flags() & regex_constants::failbit)
+      return false;
+
+   typedef typename match_results<BidiIterator>::allocator_type match_allocator_type;
+
+   match_results<BidiIterator> m;
+   BOOST_REGEX_DETAIL_NS::perl_matcher<BidiIterator, match_allocator_type, traits> matcher(first, last, m, e, flags, first);
+   unsigned int count = 0;
+   while(matcher.find())
+   {
+      ++count;
+      if(0 == foo(m))
+         return count; // caller doesn't want to go on
+      if(m[0].second == last)
+         return count; // we've reached the end, don't try and find an extra null match.
+      if(m.length() == 0)
+      {
+         if(m[0].second == last)
+            return count;
+         // we found a NULL-match, now try to find
+         // a non-NULL one at the same position:
+         match_results<BidiIterator, match_allocator_type> m2(m);
+         matcher.setf(match_not_null | match_continuous);
+         if(matcher.find())
+         {
+            ++count;
+            if(0 == foo(m))
+               return count;
+         }
+         else
+         {
+            // reset match back to where it was:
+            m = m2;
+         }
+         matcher.unsetf((match_not_null | match_continuous) & ~flags);
+      }
+   }
+   return count;
+}
+
+//
+// regex_grep convenience interfaces:
+//
+template <class Predicate, class charT, class traits>
+inline unsigned int regex_grep(Predicate foo, const charT* str, 
+                        const basic_regex<charT, traits>& e, 
+                        match_flag_type flags = match_default)
+{
+   return regex_grep(foo, str, str + traits::length(str), e, flags);
+}
+
+template <class Predicate, class ST, class SA, class charT, class traits>
+inline unsigned int regex_grep(Predicate foo, const std::basic_string<charT, ST, SA>& s, 
+                 const basic_regex<charT, traits>& e, 
+                 match_flag_type flags = match_default)
+{
+   return regex_grep(foo, s.begin(), s.end(), e, flags);
+}
+
+} // namespace boost
+
+#endif  // BOOST_REGEX_V5_REGEX_GREP_HPP
+

regex/include/boost/regex/v5/regex_iterator.hpp

@@ -0,0 +1,173 @@
+/*
+ *
+ * Copyright (c) 2003
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_iterator.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Provides regex_iterator implementation.
+  */
+
+#ifndef BOOST_REGEX_V5_REGEX_ITERATOR_HPP
+#define BOOST_REGEX_V5_REGEX_ITERATOR_HPP
+
+#include <memory>
+
+namespace boost{
+
+template <class BidirectionalIterator, 
+          class charT,
+          class traits>
+class regex_iterator_implementation 
+{
+   typedef basic_regex<charT, traits> regex_type;
+
+   match_results<BidirectionalIterator> what;  // current match
+   BidirectionalIterator                base;  // start of sequence
+   BidirectionalIterator                end;   // end of sequence
+   const regex_type                     re;   // the expression
+   match_flag_type                      flags; // flags for matching
+
+public:
+   regex_iterator_implementation(const regex_type* p, BidirectionalIterator last, match_flag_type f)
+      : base(), end(last), re(*p), flags(f){}
+   regex_iterator_implementation(const regex_iterator_implementation& other)
+      :what(other.what), base(other.base), end(other.end), re(other.re), flags(other.flags){}
+   bool init(BidirectionalIterator first)
+   {
+      base = first;
+      return regex_search(first, end, what, re, flags);
+   }
+   bool compare(const regex_iterator_implementation& that)
+   {
+      if(this == &that) return true;
+      return (&re.get_data() == &that.re.get_data()) && (end == that.end) && (flags == that.flags) && (what[0].first == that.what[0].first) && (what[0].second == that.what[0].second);
+   }
+   const match_results<BidirectionalIterator>& get()
+   { return what; }
+   bool next()
+   {
+      //if(what.prefix().first != what[0].second)
+      //   flags |= match_prev_avail;
+      BidirectionalIterator next_start = what[0].second;
+      match_flag_type f(flags);
+      if(!what.length() || (f & regex_constants::match_posix))
+         f |= regex_constants::match_not_initial_null;
+      //if(base != next_start)
+      //   f |= regex_constants::match_not_bob;
+      bool result = regex_search(next_start, end, what, re, f, base);
+      if(result)
+         what.set_base(base);
+      return result;
+   }
+private:
+   regex_iterator_implementation& operator=(const regex_iterator_implementation&);
+};
+
+template <class BidirectionalIterator, 
+          class charT = typename std::iterator_traits<BidirectionalIterator>::value_type,
+          class traits = regex_traits<charT> >
+class regex_iterator 
+{
+private:
+   typedef regex_iterator_implementation<BidirectionalIterator, charT, traits> impl;
+   typedef std::shared_ptr<impl> pimpl;
+public:
+   typedef          basic_regex<charT, traits>                   regex_type;
+   typedef          match_results<BidirectionalIterator>                    value_type;
+   typedef typename std::iterator_traits<BidirectionalIterator>::difference_type 
+                                                                            difference_type;
+   typedef          const value_type*                                       pointer;
+   typedef          const value_type&                                       reference; 
+   typedef          std::forward_iterator_tag                               iterator_category;
+   
+   regex_iterator(){}
+   regex_iterator(BidirectionalIterator a, BidirectionalIterator b, 
+                  const regex_type& re, 
+                  match_flag_type m = match_default)
+                  : pdata(new impl(&re, b, m))
+   {
+      if(!pdata->init(a))
+      {
+         pdata.reset();
+      }
+   }
+   regex_iterator(const regex_iterator& that)
+      : pdata(that.pdata) {}
+   regex_iterator& operator=(const regex_iterator& that)
+   {
+      pdata = that.pdata;
+      return *this;
+   }
+   bool operator==(const regex_iterator& that)const
+   { 
+      if((pdata.get() == 0) || (that.pdata.get() == 0))
+         return pdata.get() == that.pdata.get();
+      return pdata->compare(*(that.pdata.get())); 
+   }
+   bool operator!=(const regex_iterator& that)const
+   { return !(*this == that); }
+   const value_type& operator*()const
+   { return pdata->get(); }
+   const value_type* operator->()const
+   { return &(pdata->get()); }
+   regex_iterator& operator++()
+   {
+      cow();
+      if(0 == pdata->next())
+      {
+         pdata.reset();
+      }
+      return *this;
+   }
+   regex_iterator operator++(int)
+   {
+      regex_iterator result(*this);
+      ++(*this);
+      return result;
+   }
+private:
+
+   pimpl pdata;
+
+   void cow()
+   {
+      // copy-on-write
+      if(pdata.get() && (pdata.use_count() > 1))
+      {
+         pdata.reset(new impl(*(pdata.get())));
+      }
+   }
+};
+
+typedef regex_iterator<const char*> cregex_iterator;
+typedef regex_iterator<std::string::const_iterator> sregex_iterator;
+#ifndef BOOST_NO_WREGEX
+typedef regex_iterator<const wchar_t*> wcregex_iterator;
+typedef regex_iterator<std::wstring::const_iterator> wsregex_iterator;
+#endif
+
+// make_regex_iterator:
+template <class charT, class traits>
+inline regex_iterator<const charT*, charT, traits> make_regex_iterator(const charT* p, const basic_regex<charT, traits>& e, regex_constants::match_flag_type m = regex_constants::match_default)
+{
+   return regex_iterator<const charT*, charT, traits>(p, p+traits::length(p), e, m);
+}
+template <class charT, class traits, class ST, class SA>
+inline regex_iterator<typename std::basic_string<charT, ST, SA>::const_iterator, charT, traits> make_regex_iterator(const std::basic_string<charT, ST, SA>& p, const basic_regex<charT, traits>& e, regex_constants::match_flag_type m = regex_constants::match_default)
+{
+   return regex_iterator<typename std::basic_string<charT, ST, SA>::const_iterator, charT, traits>(p.begin(), p.end(), e, m);
+}
+
+} // namespace boost
+
+#endif // BOOST_REGEX_V5_REGEX_ITERATOR_HPP
+

regex/include/boost/regex/v5/regex_match.hpp

@@ -0,0 +1,92 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_match.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Regular expression matching algorithms.
+  *                Note this is an internal header file included
+  *                by regex.hpp, do not include on its own.
+  */
+
+
+#ifndef BOOST_REGEX_MATCH_HPP
+#define BOOST_REGEX_MATCH_HPP
+
+namespace boost{
+
+//
+// proc regex_match
+// returns true if the specified regular expression matches
+// the whole of the input.  Fills in what matched in m.
+//
+template <class BidiIterator, class Allocator, class charT, class traits>
+bool regex_match(BidiIterator first, BidiIterator last, 
+                 match_results<BidiIterator, Allocator>& m, 
+                 const basic_regex<charT, traits>& e, 
+                 match_flag_type flags = match_default)
+{
+   BOOST_REGEX_DETAIL_NS::perl_matcher<BidiIterator, Allocator, traits> matcher(first, last, m, e, flags, first);
+   return matcher.match();
+}
+template <class iterator, class charT, class traits>
+bool regex_match(iterator first, iterator last, 
+                 const basic_regex<charT, traits>& e, 
+                 match_flag_type flags = match_default)
+{
+   match_results<iterator> m;
+   return regex_match(first, last, m, e, flags | regex_constants::match_any);
+}
+//
+// query_match convenience interfaces:
+//
+template <class charT, class Allocator, class traits>
+inline bool regex_match(const charT* str, 
+                        match_results<const charT*, Allocator>& m, 
+                        const basic_regex<charT, traits>& e, 
+                        match_flag_type flags = match_default)
+{
+   return regex_match(str, str + traits::length(str), m, e, flags);
+}
+
+template <class ST, class SA, class Allocator, class charT, class traits>
+inline bool regex_match(const std::basic_string<charT, ST, SA>& s, 
+                 match_results<typename std::basic_string<charT, ST, SA>::const_iterator, Allocator>& m, 
+                 const basic_regex<charT, traits>& e, 
+                 match_flag_type flags = match_default)
+{
+   return regex_match(s.begin(), s.end(), m, e, flags);
+}
+template <class charT, class traits>
+inline bool regex_match(const charT* str, 
+                        const basic_regex<charT, traits>& e, 
+                        match_flag_type flags = match_default)
+{
+   match_results<const charT*> m;
+   return regex_match(str, str + traits::length(str), m, e, flags | regex_constants::match_any);
+}
+
+template <class ST, class SA, class charT, class traits>
+inline bool regex_match(const std::basic_string<charT, ST, SA>& s, 
+                 const basic_regex<charT, traits>& e, 
+                 match_flag_type flags = match_default)
+{
+   typedef typename std::basic_string<charT, ST, SA>::const_iterator iterator;
+   match_results<iterator> m;
+   return regex_match(s.begin(), s.end(), m, e, flags | regex_constants::match_any);
+}
+
+
+} // namespace boost
+
+#endif   // BOOST_REGEX_MATCH_HPP
+

regex/include/boost/regex/v5/regex_merge.hpp

@@ -0,0 +1,71 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_format.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Provides formatting output routines for search and replace
+  *                operations.  Note this is an internal header file included
+  *                by regex.hpp, do not include on its own.
+  */
+
+#ifndef BOOST_REGEX_V5_REGEX_MERGE_HPP
+#define BOOST_REGEX_V5_REGEX_MERGE_HPP
+
+
+namespace boost{
+
+template <class OutputIterator, class Iterator, class traits, class charT>
+inline OutputIterator regex_merge(OutputIterator out,
+                         Iterator first,
+                         Iterator last,
+                         const basic_regex<charT, traits>& e, 
+                         const charT* fmt, 
+                         match_flag_type flags = match_default)
+{
+   return regex_replace(out, first, last, e, fmt, flags);
+}
+
+template <class OutputIterator, class Iterator, class traits, class charT>
+inline OutputIterator regex_merge(OutputIterator out,
+                         Iterator first,
+                         Iterator last,
+                         const basic_regex<charT, traits>& e, 
+                         const std::basic_string<charT>& fmt,
+                         match_flag_type flags = match_default)
+{
+   return regex_merge(out, first, last, e, fmt.c_str(), flags);
+}
+
+template <class traits, class charT>
+inline std::basic_string<charT> regex_merge(const std::basic_string<charT>& s,
+                         const basic_regex<charT, traits>& e, 
+                         const charT* fmt,
+                         match_flag_type flags = match_default)
+{
+   return regex_replace(s, e, fmt, flags);
+}
+
+template <class traits, class charT>
+inline std::basic_string<charT> regex_merge(const std::basic_string<charT>& s,
+                         const basic_regex<charT, traits>& e, 
+                         const std::basic_string<charT>& fmt,
+                         match_flag_type flags = match_default)
+{
+   return regex_replace(s, e, fmt, flags);
+}
+
+} // namespace boost
+
+#endif  // BOOST_REGEX_V5_REGEX_MERGE_HPP
+
+

regex/include/boost/regex/v5/regex_raw_buffer.hpp

@@ -0,0 +1,213 @@
+/*
+ *
+ * Copyright (c) 1998-2002
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_raw_buffer.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Raw character buffer for regex code.
+  *                Note this is an internal header file included
+  *                by regex.hpp, do not include on its own.
+  */
+
+#ifndef BOOST_REGEX_RAW_BUFFER_HPP
+#define BOOST_REGEX_RAW_BUFFER_HPP
+
+#ifndef BOOST_REGEX_CONFIG_HPP
+#include <boost/regex/config.hpp>
+#endif
+
+#include <algorithm>
+#include <cstddef>
+
+namespace boost{
+   namespace BOOST_REGEX_DETAIL_NS{
+
+struct empty_padding{};
+
+union padding
+{
+   void* p;
+   unsigned int i;
+};
+
+template <int N>
+struct padding3
+{
+   enum{
+      padding_size = 8,
+      padding_mask = 7
+   };
+};
+
+template<>
+struct padding3<2>
+{
+   enum{
+      padding_size = 2,
+      padding_mask = 1
+   };
+};
+
+template<>
+struct padding3<4>
+{
+   enum{
+      padding_size = 4,
+      padding_mask = 3
+   };
+};
+
+template<>
+struct padding3<8>
+{
+   enum{
+      padding_size = 8,
+      padding_mask = 7
+   };
+};
+
+template<>
+struct padding3<16>
+{
+   enum{
+      padding_size = 16,
+      padding_mask = 15
+   };
+};
+
+enum{
+   padding_size = padding3<sizeof(padding)>::padding_size,
+   padding_mask = padding3<sizeof(padding)>::padding_mask
+};
+
+//
+// class raw_storage
+// basically this is a simplified vector<unsigned char>
+// this is used by basic_regex for expression storage
+//
+
+class raw_storage
+{
+public:
+   typedef std::size_t           size_type;
+   typedef unsigned char*        pointer;
+private:
+   pointer last, start, end;
+public:
+
+   raw_storage();
+   raw_storage(size_type n);
+
+   ~raw_storage()
+   {
+      ::operator delete(start);
+   }
+
+   void  resize(size_type n)
+   {
+      size_type newsize = start ? last - start : 1024;
+      while (newsize < n)
+         newsize *= 2;
+      size_type datasize = end - start;
+      // extend newsize to WORD/DWORD boundary:
+      newsize = (newsize + padding_mask) & ~(padding_mask);
+
+      // allocate and copy data:
+      pointer ptr = static_cast<pointer>(::operator new(newsize));
+      BOOST_REGEX_NOEH_ASSERT(ptr)
+         if (start)
+            std::memcpy(ptr, start, datasize);
+
+      // get rid of old buffer:
+      ::operator delete(start);
+
+      // and set up pointers:
+      start = ptr;
+      end = ptr + datasize;
+      last = ptr + newsize;
+   }
+
+   void*  extend(size_type n)
+   {
+      if(size_type(last - end) < n)
+         resize(n + (end - start));
+      pointer result = end;
+      end += n;
+      return result;
+   }
+
+   void*  insert(size_type pos, size_type n)
+   {
+      BOOST_REGEX_ASSERT(pos <= size_type(end - start));
+      if (size_type(last - end) < n)
+         resize(n + (end - start));
+      void* result = start + pos;
+      std::memmove(start + pos + n, start + pos, (end - start) - pos);
+      end += n;
+      return result;
+   }
+
+   size_type  size()
+   {
+      return size_type(end - start);
+   }
+
+   size_type  capacity()
+   {
+      return size_type(last - start);
+   }
+
+   void*  data()const
+   {
+      return start;
+   }
+
+   size_type  index(void* ptr)
+   {
+      return size_type(static_cast<pointer>(ptr) - static_cast<pointer>(data()));
+   }
+
+   void  clear()
+   {
+      end = start;
+   }
+
+   void  align()
+   {
+      // move end up to a boundary:
+      end = start + (((end - start) + padding_mask) & ~padding_mask);
+   }
+   void swap(raw_storage& that)
+   {
+      std::swap(start, that.start);
+      std::swap(end, that.end);
+      std::swap(last, that.last);
+  }
+};
+
+inline raw_storage::raw_storage()
+{
+   last = start = end = 0;
+}
+
+inline raw_storage::raw_storage(size_type n)
+{
+   start = end = static_cast<pointer>(::operator new(n));
+   BOOST_REGEX_NOEH_ASSERT(start)
+   last = start + n;
+}
+
+} // namespace BOOST_REGEX_DETAIL_NS
+} // namespace boost
+
+#endif
+

regex/include/boost/regex/v5/regex_replace.hpp

@@ -0,0 +1,77 @@
+/*
+ *
+ * Copyright (c) 1998-2009
+ * John Maddock
+ *
+ * Use, modification and distribution are subject to the 
+ * Boost Software License, Version 1.0. (See accompanying file 
+ * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
+ *
+ */
+
+ /*
+  *   LOCATION:    see http://www.boost.org for most recent version.
+  *   FILE         regex_format.hpp
+  *   VERSION      see <boost/version.hpp>
+  *   DESCRIPTION: Provides formatting output routines for search and replace
+  *                operations.  Note this is an internal header file included
+  *                by regex.hpp, do not include on its own.
+  */
+
+#ifndef BOOST_REGEX_V5_REGEX_REPLACE_HPP
+#define BOOST_REGEX_V5_REGEX_REPLACE_HPP
+
+
+namespace boost{
+
+template <class OutputIterator, class BidirectionalIterator, class traits, class charT, class Formatter>
+OutputIterator regex_replace(OutputIterator out,
+                         BidirectionalIterator first,
+                         BidirectionalIterator last,
+                         const basic_regex<charT, traits>& e, 
+                         Formatter fmt, 
+                         match_flag_type flags = match_default)
+{
+   regex_iterator<BidirectionalIterator, charT, traits> i(first, last, e, flags);
+   regex_iterator<BidirectionalIterator, charT, traits> j;
+   if(i == j)
+   {
+      if(!(flags & regex_constants::format_no_copy))
+         out = BOOST_REGEX_DETAIL_NS::copy(first, last, out);
+   }
+   else
+   {
+      BidirectionalIterator last_m(first);
+      while(i != j)
+      {
+         if(!(flags & regex_constants::format_no_copy))
+            out = BOOST_REGEX_DETAIL_NS::copy(i->prefix().first, i->prefix().second, out);
+         out = i->format(out, fmt, flags, e);
+         last_m = (*i)[0].second;
+         if(flags & regex_constants::format_first_only)
+            break;
+         ++i;
+      }
+      if(!(flags & regex_constants::format_no_copy))
+         out = BOOST_REGEX_DETAIL_NS::copy(last_m, last, out);
+   }
+   return out;
+}
+
+template <class traits, class charT, class Formatter>
+std::basic_string<charT> regex_replace(const std::basic_string<charT>& s,
+                         const basic_regex<charT, traits>& e, 
+                         Formatter fmt,
+                         match_flag_type flags = match_default)
+{
+   std::basic_string<charT> result;
+   BOOST_REGEX_DETAIL_NS::string_out_iterator<std::basic_string<charT> > i(result);
+   regex_replace(i, s.begin(), s.end(), e, fmt, flags);
+   return result;
+}
+
+} // namespace boost
+
+#endif  // BOOST_REGEX_V5_REGEX_REPLACE_HPP
+
+