Update changelog

Updated pcre2s

.clang-format

@@ -0,0 +1,22 @@
+BasedOnStyle: LLVM
+UseTab: AlignWithSpaces
+IndentWidth: 4
+TabWidth: 4
+BreakBeforeBraces: Allman
+ColumnLimit: 0
+NamespaceIndentation: Inner
+FixNamespaceComments: true
+AccessModifierOffset: -2
+AllowShortCaseLabelsOnASingleLine: true
+IndentCaseLabels: true
+BreakConstructorInitializers: BeforeComma
+BraceWrapping:
+  BeforeLambdaBody: false
+AlignAfterOpenBracket: DontAlign
+Cpp11BracedListStyle: false
+IncludeBlocks: Regroup
+LambdaBodyIndentation: Signature
+AllowShortLambdasOnASingleLine: Inline
+EmptyLineBeforeAccessModifier: LogicalBlock
+IndentPPDirectives: AfterHash
+PPIndentWidth: 1

.github/workflows/cmake-multi-platform.yml

@@ -33,13 +33,18 @@ jobs:
 
     - name: Install dependencies Ubuntu
       if: matrix.os == 'ubuntu-latest'
-      run: sudo apt-get update && sudo apt-get install mrc
+      run: sudo apt-get update && sudo apt-get install mrc catch2
 
     - name: Install dependencies Window
       if: matrix.os == 'windows-latest'
       run: ./tools/depends.cmd
       shell: cmd
 
+    - name: Install Catch2 macOS
+      if: matrix.os == 'macos-latest'
+      run: >
+        brew install catch2
+
     - name: Configure CMake
       run: >
         cmake -B ${{ steps.strings.outputs.build-output-dir }}

.gitignore

@@ -13,3 +13,6 @@ docs/api
 docs/conf.py
 build_ci/
 data/components.cif
+perf.data*
+.cache/
+

CMakeLists.txt

@@ -32,26 +32,17 @@ endif()
 # set the project name
 project(
 	libcifpp
-	VERSION 9.0.0
-	LANGUAGES CXX)
+	VERSION 9.0.2
+	LANGUAGES CXX C)
 
 list(PREPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
 
 include(FindAtomic)
-include(CheckFunctionExists)
-include(CheckIncludeFiles)
-include(CheckLibraryExists)
 include(CMakePackageConfigHelpers)
-include(CheckCXXSourceCompiles)
 include(GenerateExportHeader)
 include(CTest)
-include(FetchContent)
 include(ExternalProject)
-
-# FindBoost, take care of it now.
-if(CMAKE_VERSION VERSION_GREATER_EQUAL 3.30)
-	cmake_policy(SET CMP0167 NEW)
-endif()
+include(FetchContent)
 
 # When building with ninja-multiconfig, build both debug and release by default
 if(CMAKE_GENERATOR STREQUAL "Ninja Multi-Config")
@@ -149,51 +140,8 @@ 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 "cifpp: Testing for known regex bug, since you're using GNU libstdc++")
-
-	try_run(STD_REGEX_RUNNING STD_REGEX_COMPILING
-		${CMAKE_CURRENT_BINARY_DIR}/test
-		${CMAKE_CURRENT_SOURCE_DIR}/cmake/test-rx.cpp)
-
-	if(STD_REGEX_RUNNING STREQUAL FAILED_TO_RUN)
-		message(
-			STATUS
-			"cifpp: You are probably trying to compile using the g++ standard library which contains a crashing std::regex implementation. Will use boost::regex instead"
-		)
-
-		find_package(Boost 1.80 QUIET COMPONENTS regex)
-
-		if(NOT Boost_FOUND)
-			set(BOOST_REGEX_STANDALONE ON)
-
-			FetchContent_Declare(
-				boost-rx
-				GIT_REPOSITORY https://github.com/boostorg/regex
-				GIT_TAG boost-1.83.0)
-
-			FetchContent_MakeAvailable(boost-rx)
-		endif()
-
-		set(BOOST_REGEX ON)
-	endif()
-endif()
-
 if(MSVC)
-	# Avoid linking the shared library of zlib Search ZLIB_ROOT first if it is
+	# 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)
@@ -220,13 +168,37 @@ if(MSVC)
 	endforeach()
 endif()
 
-find_package(ZLIB QUIET)
+# First check if <format> is available
+find_file(FMT NAME format)
+
+if(FMT EQUAL "FMT-NOTFOUND")
+	if(NOT (fmt_FOUND OR TARGET fmt))
+		find_package(fmt REQUIRED)
+		message(FATAL_ERROR "fmt not found, compiler too old, you're out of luck")
+	endif()
+endif()
+
 find_package(Threads)
+find_package(ZLIB QUIET)
 
 if(NOT ZLIB_FOUND)
 	message(FATAL_ERROR "cifpp: The zlib development files were not found you this system, please install them and try again (hint: on debian/ubuntu use apt-get install zlib1g-dev)")
 endif()
 
+include(FindPkgConfig)
+
+if(PKG_CONFIG_FOUND)
+	pkg_check_modules(PCRE2 IMPORTED_TARGET libpcre2-8)
+
+	if(PCRE2_FOUND)
+		message(STATUS "Using pcre2 found using pkg-config")
+	endif()
+endif()
+
+if(NOT PCRE2_FOUND)
+	add_subdirectory(pcre2-simple)
+endif()
+
 # Using Eigen3 is a bit of a thing. We don't want to build it completely since
 # we only need a couple of header files. Nothing special. But often, eigen3 is
 # already installed and then we prefer that.
@@ -238,16 +210,16 @@ if(Eigen3_FOUND AND TARGET Eigen3::Eigen)
 else()
 	# Use ExternalProject since FetchContent always tries to install the result...
 	ExternalProject_Add(my-eigen3
-		GIT_REPOSITORY https://gitlab.com/libeigen/eigen.git
-		GIT_TAG 3.4.0
+		URL https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.zip
+		DOWNLOAD_EXTRACT_TIMESTAMP TRUE
+		CONFIGURE_COMMAND ""
+		BUILD_COMMAND ""
 		INSTALL_COMMAND "")
 
 	ExternalProject_Get_Property(my-eigen3 SOURCE_DIR)
 	set(EIGEN_INCLUDE_DIR ${SOURCE_DIR})
 endif()
 
-message(STATUS "cifpp: Eigen include dir is ${EIGEN_INCLUDE_DIR}")
-
 # Create a revision file, containing the current git version info
 include(VersionString)
 write_version_header(${CMAKE_CURRENT_SOURCE_DIR}/src/ LIB_NAME "LibCIFPP")
@@ -350,19 +322,9 @@ target_sources(cifpp
 # The code now really requires C++20
 target_compile_features(cifpp PUBLIC cxx_std_20)
 
-set(CMAKE_DEBUG_POSTFIX d)
-set_target_properties(cifpp PROPERTIES DEBUG_POSTFIX "d")
-
 generate_export_header(cifpp EXPORT_FILE_NAME
 	${CMAKE_CURRENT_SOURCE_DIR}/include/cif++/exports.hpp)
 
-if(BOOST_REGEX)
-	target_compile_definitions(cifpp PRIVATE USE_BOOST_REGEX=1
-		BOOST_REGEX_STANDALONE=1)
-	get_target_property(BOOST_REGEX_INCLUDE_DIR Boost::regex
-		INTERFACE_INCLUDE_DIRECTORIES)
-endif()
-
 if(MSVC)
 	target_compile_definitions(cifpp PUBLIC NOMINMAX=1)
 endif()
@@ -373,9 +335,21 @@ target_include_directories(
 	cifpp
 	PUBLIC "$<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/include>"
 	"$<INSTALL_INTERFACE:${CMAKE_INSTALL_INCLUDEDIR}>"
-	PRIVATE "${BOOST_REGEX_INCLUDE_DIR}" "${EIGEN_INCLUDE_DIR}")
+	PRIVATE "${EIGEN_INCLUDE_DIR}")
 
-target_link_libraries(cifpp PUBLIC Threads::Threads ZLIB::ZLIB $<$<TARGET_EXISTS:std::atomic>:std::atomic>)
+target_link_libraries(cifpp
+	PUBLIC Threads::Threads ZLIB::ZLIB $<$<TARGET_EXISTS:std::atomic>:std::atomic>)
+
+if(PCRE2_FOUND)
+	target_include_directories(cifpp PRIVATE ${PCRE2_INCLUDE_DIRS})
+	target_link_libraries(cifpp PRIVATE ${PCRE2_LINK_LIBRARIES})
+else()
+	target_link_libraries(cifpp PRIVATE $<BUILD_INTERFACE:pcre2s>)
+endif()
+
+if(fmt_FOUND)
+	target_link_libraries(cifpp PUBLIC fmt)
+endif()
 
 if(CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
 	target_link_options(cifpp PRIVATE -undefined dynamic_lookup)
@@ -513,6 +487,7 @@ if(CIFPP_DATA_DIR AND CIFPP_DOWNLOAD_CCD)
 endif()
 
 set(CONFIG_TEMPLATE_FILE ${CMAKE_CURRENT_SOURCE_DIR}/cmake/cifpp-config.cmake.in)
+set(REQUIRE_FMT ${fmt_FOUND})
 
 configure_package_config_file(
 	${CONFIG_TEMPLATE_FILE} ${CMAKE_CURRENT_BINARY_DIR}/cifpp/cifpp-config.cmake

README.md

@@ -117,12 +117,8 @@ Other libraries you might want to install beforehand are:
   `libeigen3-dev`
 - [zlib](https://github.com/madler/zlib), the development version of this
   library. On Debian/Ubuntu this is the package `zlib1g-dev`.
-- [boost](https://www.boost.org), in Debian/Ubuntu this is `libboost-dev`.
-  
-  The Boost libraries are only needed in case you are using GCC due to a long
-  standing bug in GNU's implementation of std::regex. It simply crashes
-  on the regular expressions used in the mmcif_pdbx dictionary and so
-  we use the boost regex implementation instead.
+- [pcre2](https://www.pcre.org/), the Perl Compatible Regular Expression
+  library. On Debian/Ubuntu this is the package `libpcre2-dev`.
 
 ### Building
 

changelog

@@ -1,3 +1,11 @@
+Version 9.0.2
+- Fix code that reconstructs sequences, could throw a map::at
+- Many optimisations in validation and reconstruction code.
+
+Version 9.0.1
+- Use pcre2 from pkg-config if available, if not
+  build a version from the original code.
+
 Version 9.0.0
 - Rename fields of cif::mm::polymer to match the naming
   in mmcif_pdbx.dic. Also, related, fix building mm::structure

cmake/FindPCRE2.cmake

@@ -0,0 +1,12 @@
+# The problem is, find_package(PCRE2) does not work
+# and using pkg-config results in linking to a shared library
+# causing all kinds of trouble later on
+
+find_path(PCRE2_INCLUDEDIR NAMES pcre2.h HINTS "C:/Program Files (x86)/PCRE2/include" REQUIRED)
+find_library(PCRE2_LIBRARY NAMES pcre2-8-static libpcre2-8.a HINTS "C:/Program Files (x86)/PCRE2/lib" REQUIRED)
+
+add_library(pcre2-8 IMPORTED STATIC)
+target_include_directories(pcre2-8 INTERFACE ${PCRE2_INCLUDEDIR})
+target_compile_definitions(pcre2-8 INTERFACE PCRE2_STATIC)
+set_target_properties(pcre2-8 PROPERTIES IMPORTED_LOCATION ${PCRE2_LIBRARY})
+set_target_properties(pcre2-8 PROPERTIES IMPORTED_IMPLIB ${PCRE2_LIBRARY})

cmake/cifpp-config.cmake.in

@@ -8,5 +8,8 @@ include(CMakeFindDependencyMacro)
 
 find_dependency(Threads)
 find_dependency(ZLIB REQUIRED)
+if(@REQUIRE_FMT@)
+    find_dependency(fmt REQUIRED)
+endif()
 
 check_required_components(cifpp)

cmake/test-rx.cpp

@@ -1,18 +0,0 @@
-// 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) << '\n';
-	std::cout << m.str(1).substr(0, 10) << '\n';
-
-	return 0;
-}

include/cif++/category.hpp

@@ -157,7 +157,7 @@ class category
 		emplace(std::forward<row_initializer>(rows));
 	}
 
-	category(const category &rhs);   ///< Copy constructor
+	category(const category &rhs); ///< Copy constructor
 
 	category(category &&rhs) noexcept ///< Move constructor
 	{
@@ -332,8 +332,16 @@ class category
 	// --------------------------------------------------------------------
 	// A category can have a key, as defined by the validator/dictionary
 
+	/// @brief The type of an element of the key_type
+	struct key_element_type
+	{
+		std::string name;         ///< Name of the item
+		std::string value;        ///< Value to be found
+		bool may_be_null = false; ///< If true, value should be same or empty
+	};
+
 	/// @brief The key type
-	using key_type = row_initializer;
+	using key_type = std::vector<key_element_type>;
 
 	/// @brief Return a row_handle for the row specified by \a key
 	/// @param key The value for the key, items specified in the dictionary should have a value
@@ -1249,7 +1257,7 @@ class category
 		{
 		}
 
-// TODO: NEED TO FIX THIS!
+		// TODO: NEED TO FIX THIS!
 		category *linked;
 		const link_validator *v;
 	};

include/cif++/compound.hpp

@@ -180,7 +180,7 @@ class compound
 	friend class local_compound_factory_impl;
 
 	compound(cif::datablock &db);
-	
+
 	std::string m_id;
 	std::string m_name;
 	std::string m_type;
@@ -290,6 +290,13 @@ class compound_factory
 
 	void report_missing_compound(std::string_view compound_id);
 
+	bool get_report_missing() const { return m_report_missing; }
+
+	void set_report_missing(bool report)
+	{
+		m_report_missing = report;
+	}
+
   private:
 	compound_factory();
 
@@ -301,6 +308,7 @@ class compound_factory
 	static bool s_use_thread_local_instance;
 
 	std::shared_ptr<compound_factory_impl> m_impl;
+	bool m_report_missing = true;
 };
 
 // --------------------------------------------------------------------

include/cif++/condition.hpp

@@ -27,6 +27,7 @@
 #pragma once
 
 #include "cif++/row.hpp"
+#include "cif++/format.hpp"
 
 #include <cassert>
 #include <concepts>
@@ -49,49 +50,49 @@
  * @code {.cpp}
  * cif::condition c = cif::key("id") == 1;
  * @endcode
- * 
+ *
  * That will find rows where the ID item contains the number 1. If
  * using cif::key is a bit too much typing, you can also write:
- * 
+ *
  * @code{.cpp}
  * using namespace cif::literals;
- * 
+ *
  * cif::condition c2 = "id"_key == 1;
  * @endcode
- * 
+ *
  * Now if you want both ID = 1 and ID = 2 in the result:
- * 
+ *
  * @code{.cpp}
  * auto c3 = "id"_key == 1 or "id"_key == 2;
  * @endcode
- * 
+ *
  * There are some special values you can use. To find rows with item that
  * do not have a value:
- * 
+ *
  * @code{.cpp}
  * auto c4 = "type"_key == cif::null;
- * @endcode 
- * 
+ * @endcode
+ *
  * Of if it should not be NULL:
- * 
+ *
  * @code{.cpp}
  * auto c5 = "type"_key != cif::null;
- * @endcode 
- * 
+ * @endcode
+ *
  * There's even a way to find all records:
- * 
+ *
  * @code{.cpp}
  * auto c6 = cif::all;
  * @endcode
- * 
+ *
  * And when you want to search for any item containing the value 'foo':
- * 
+ *
  * @code{.cpp}
  * auto c7 = cif::any == "foo";
- * @endcode 
- * 
+ * @endcode
+ *
  * All these conditions can be chained together again:
- * 
+ *
  * @code{.cpp}
  * auto c8 = std::move(c3) and std::move(c5);
  * @endcode
@@ -106,7 +107,7 @@ namespace cif
 
 /**
  * @brief Get the items that can be used as key in conditions for a category
- * 
+ *
  * @param cat The category whose items to return
  * @return iset The set of key item names
  */
@@ -115,7 +116,7 @@ iset get_category_fields(const category &cat);
 
 /**
  * @brief Get the items that can be used as key in conditions for a category
- * 
+ *
  * @param cat The category whose items to return
  * @return iset The set of key field names
  */
@@ -123,7 +124,7 @@ iset get_category_items(const category &cat);
 
 /**
  * @brief Get the item index for item @a col in category @a cat
- * 
+ *
  * @param cat The category
  * @param col The name of the item
  * @return uint16_t The index
@@ -132,7 +133,7 @@ uint16_t get_item_ix(const category &cat, std::string_view col);
 
 /**
  * @brief Return whether the item @a col in category @a cat has a primitive type of *uchar*
- * 
+ *
  * @param cat The category
  * @param col The item name
  * @return true If the primitive type is of type *uchar*
@@ -175,14 +176,13 @@ namespace detail
 class condition
 {
   public:
-
 	/** @cond */
 	using condition_impl = detail::condition_impl;
 	/** @endcond */
 
 	/**
 	 * @brief Construct a new, empty condition object
-	 * 
+	 *
 	 */
 	condition()
 		: m_impl(nullptr)
@@ -191,7 +191,7 @@ class condition
 
 	/**
 	 * @brief Construct a new condition object with implementation @a impl
-	 * 
+	 *
 	 * @param impl The implementation to use
 	 */
 	explicit condition(condition_impl *impl)
@@ -230,15 +230,15 @@ class condition
 	/**
 	 * @brief Prepare the condition to be used on category @a c. This will
 	 * take care of setting the correct indices for items e.g.
-	 * 
+	 *
 	 * @param c The category this query should act upon
 	 */
 	void prepare(const category &c);
 
 	/**
-	 * @brief This operator returns true if the row referenced by @a r is 
+	 * @brief This operator returns true if the row referenced by @a r is
 	 * a match for this condition.
-	 * 
+	 *
 	 * @param r The reference to a row.
 	 * @return true If there is a match
 	 * @return false If there is no match
@@ -263,7 +263,7 @@ class condition
 	/**
 	 * @brief If the prepare step found out there is only one hit
 	 * this single hit can be returned by this method.
-	 * 
+	 *
 	 * @return std::optional<row_handle> The result will contain
 	 * a row reference if there is a single hit, it will be empty otherwise
 	 */
@@ -292,7 +292,7 @@ class condition
 
 	/**
 	 * @brief Operator to use to write out a condition to @a os, for debugging purposes
-	 * 
+	 *
 	 * @param os The std::ostream to write to
 	 * @param cond The condition to write
 	 * @return std::ostream& The same as @a os
@@ -752,28 +752,9 @@ namespace detail
 				delete sub;
 		}
 
-		condition_impl *prepare(const category &c) override
-		{
-			for (auto &sub : m_sub)
-				sub = sub->prepare(c);
-			return this;
-		}
+		condition_impl *prepare(const category &c) override;
 
-		bool test(row_handle r) const override
-		{
-			bool result = true;
-
-			for (auto sub : m_sub)
-			{
-				if (sub->test(r))
-					continue;
-
-				result = false;
-				break;
-			}
-
-			return result;
-		}
+		bool test(row_handle r) const override;
 
 		void str(std::ostream &os) const override
 		{
@@ -820,6 +801,7 @@ namespace detail
 		static condition_impl *combine_equal(std::vector<and_condition_impl *> &subs, or_condition_impl *oc);
 
 		std::vector<condition_impl *> m_sub;
+		std::optional<row_handle> m_single; // Potential result of index lookup
 	};
 
 	struct or_condition_impl : public condition_impl
@@ -977,9 +959,9 @@ inline condition operator or(condition &&a, condition &&b)
 			if (ci->m_item_name == ce->m_item_name)
 				return condition(new detail::key_equals_or_empty_condition_impl(ci));
 		}
-		
+
 		if (typeid(*b.m_impl) == typeid(detail::key_equals_condition_impl) and
-				 typeid(*a.m_impl) == typeid(detail::key_is_empty_condition_impl))
+			typeid(*a.m_impl) == typeid(detail::key_is_empty_condition_impl))
 		{
 			auto ci = static_cast<detail::key_equals_condition_impl *>(b.m_impl);
 			auto ce = static_cast<detail::key_is_empty_condition_impl *>(a.m_impl);
@@ -997,9 +979,9 @@ inline condition operator or(condition &&a, condition &&b)
 			if (ci->m_item_name == ce->m_item_name)
 				return condition(new detail::key_equals_number_or_empty_condition_impl(ci));
 		}
-		
+
 		if (typeid(*b.m_impl) == typeid(detail::key_equals_number_condition_impl) and
-				 typeid(*a.m_impl) == typeid(detail::key_is_empty_condition_impl))
+			typeid(*a.m_impl) == typeid(detail::key_is_empty_condition_impl))
 		{
 			auto ci = static_cast<detail::key_equals_number_condition_impl *>(b.m_impl);
 			auto ce = static_cast<detail::key_is_empty_condition_impl *>(a.m_impl);
@@ -1019,7 +1001,7 @@ inline condition operator or(condition &&a, condition &&b)
 
 /**
  * @brief A helper class to make it possible to search for empty items (NULL)
- * 
+ *
  * @code{.cpp}
  * "id"_key == cif::empty_type();
  * @endcode
@@ -1031,7 +1013,7 @@ struct empty_type
 
 /**
  * @brief A helper to make it possible to have conditions like
- * 
+ *
  * @code{.cpp}
  * "id"_key == cif::null;
  * @endcode
@@ -1041,14 +1023,14 @@ inline constexpr empty_type null = empty_type();
 
 /**
  * @brief Class to use in creating conditions, creates a reference to a item or item
- * 
+ *
  */
 struct key
 {
 	/**
 	 * @brief Construct a new key object using @a item_name as name
-	 * 
-	 * @param item_name 
+	 *
+	 * @param item_name
 	 */
 	explicit key(const std::string &item_name)
 		: m_item_name(item_name)
@@ -1057,8 +1039,8 @@ struct key
 
 	/**
 	 * @brief Construct a new key object using @a item_name as name
-	 * 
-	 * @param item_name 
+	 *
+	 * @param item_name
 	 */
 	explicit key(const char *item_name)
 		: m_item_name(item_name)
@@ -1067,8 +1049,8 @@ struct key
 
 	/**
 	 * @brief Construct a new key object using @a item_name as name
-	 * 
-	 * @param item_name 
+	 *
+	 * @param item_name
 	 */
 	explicit key(std::string_view item_name)
 		: m_item_name(item_name)
@@ -1090,7 +1072,8 @@ concept Numeric = ((std::is_floating_point_v<T> or std::is_integral_v<T>) and no
 template <Numeric T>
 condition operator==(const key &key, const T &v)
 {
-	return condition(new detail::key_equals_number_condition_impl(key.m_item_name, v));
+	// TODO: change key_equals_etc... to use std::variant<double,int64_t> or something
+	return condition(new detail::key_equals_number_condition_impl(key.m_item_name, static_cast<double>(v)));
 }
 
 /**
@@ -1137,13 +1120,10 @@ inline condition operator!=(const key &key, std::string_view value)
 template <Numeric 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_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v) > 0; },
-		s.str()));
+		cif::format(" > {}", v)));
 }
 
 /**
@@ -1152,13 +1132,10 @@ condition operator>(const key &key, const T &v)
 template <Numeric 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_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v) >= 0; },
-		s.str()));
+		cif::format(" >= {}", v)));
 }
 
 /**
@@ -1167,13 +1144,10 @@ condition operator>=(const key &key, const T &v)
 template <Numeric 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_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v) < 0; },
-		s.str()));
+		cif::format(" < {}", v)));
 }
 
 /**
@@ -1182,13 +1156,10 @@ condition operator<(const key &key, const T &v)
 template <Numeric 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_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v) <= 0; },
-		s.str()));
+		cif::format(" <= {}", v)));
 }
 
 /**
@@ -1196,13 +1167,10 @@ condition operator<=(const key &key, const T &v)
  */
 inline condition operator>(const key &key, std::string_view v)
 {
-	std::ostringstream s;
-	s << " > " << v;
-
 	return condition(new detail::key_compare_condition_impl(
 		key.m_item_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v, icase) > 0; },
-		s.str()));
+		cif::format(" > {}", v)));
 }
 
 /**
@@ -1210,13 +1178,10 @@ inline condition operator>(const key &key, std::string_view v)
  */
 inline condition operator>=(const key &key, std::string_view v)
 {
-	std::ostringstream s;
-	s << " >= " << v;
-
 	return condition(new detail::key_compare_condition_impl(
 		key.m_item_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v, icase) >= 0; },
-		s.str()));
+		cif::format(" >= {}", v)));
 }
 
 /**
@@ -1224,13 +1189,10 @@ inline condition operator>=(const key &key, std::string_view v)
  */
 inline condition operator<(const key &key, std::string_view v)
 {
-	std::ostringstream s;
-	s << " < " << v;
-
 	return condition(new detail::key_compare_condition_impl(
 		key.m_item_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v, icase) < 0; },
-		s.str()));
+		cif::format(" < {}", v)));
 }
 
 /**
@@ -1238,13 +1200,10 @@ inline condition operator<(const key &key, std::string_view v)
  */
 inline condition operator<=(const key &key, std::string_view v)
 {
-	std::ostringstream s;
-	s << " <= " << v;
-
 	return condition(new detail::key_compare_condition_impl(
 		key.m_item_name, [item_name = key.m_item_name, v](row_handle r, bool icase)
 		{ return r[item_name].compare(v, icase) <= 0; },
-		s.str()));
+		cif::format(" <= {}", v)));
 }
 
 /**
@@ -1345,7 +1304,7 @@ namespace literals
 {
 	/**
 	 * @brief Return a cif::key for the item name @a text
-	 * 
+	 *
 	 * @param text The name of the item
 	 * @param length The length of @a text
 	 * @return key The cif::key created

include/cif++/format.hpp

@@ -26,138 +26,28 @@
 
 #pragma once
 
+#if __has_include(<format>)
+#include <format>
+#define USE_STD_FORMAT 1
+#else
+#include <fmt/format.h>
+#endif
+
 #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.
+ * Now using cif::format instead of a home grown rip off
  */
 
 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
-
-/** @cond */
-
-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 <std::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;
-};
-
-/** @endcond */
-
-/**
- * @brief A simplistic reimplementation of boost::format, in fact it is
- * actually a way to call the C function snprintf to format the arguments
- * in @a args into the format string @a fmt
- * 
- * The string in @a fmt should thus be a C style format string.
- * 
- * TODO: Move to C++23 style of printing.
- * 
- * @tparam Args The types of the arguments
- * @param fmt The format string
- * @param args The arguments
- * @return An object that can be written out to a std::ostream using operator<<
- */
-
-template <typename... Args>
-constexpr auto format(std::string_view fmt, Args... args)
-{
-	return format_plus_arg(fmt, std::forward<Args>(args)...);
-}
+#if USE_STD_FORMAT
+using std::format;
+#else
+using fmt::format;
+#endif
 
 // --------------------------------------------------------------------
 /// A streambuf that fills out lines with spaces up until a specified width

include/cif++/gzio.hpp

@@ -1,7 +1,33 @@
-//          Copyright Maarten L. Hekkelman, 2022
-// 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)
+/*-
+ * SPDX-License-Identifier: BSD-2-Clause
+ * 
+ * Copyright (c) 2025 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.
+ */
+
+/*
+	Original code comes from libgxrio at https://github.com/mhekkel/gxrio
+	This is a stripped down version.
+*/
 
 #pragma once
 

include/cif++/item.hpp

@@ -250,6 +250,8 @@ class item
 			return value();
 	}
 
+	auto operator<=>(const item &rhs) const = default;
+
   private:
 	std::string_view m_name;
 	std::string m_value;

include/cif++/utilities.hpp

@@ -53,6 +53,7 @@
 #pragma warning(disable : 4068) // unknown pragma
 #pragma warning(disable : 4100) // unreferenced formal parameter
 #pragma warning(disable : 4101) // unreferenced local variable
+#pragma warning(disable : 4702) // unreachable code (too bad, this one. Happens in for loops)
 #define _SILENCE_CXX17_CODECVT_HEADER_DEPRECATION_WARNING 1
 #endif
 

pcre2-simple/CMakeLists.txt

@@ -0,0 +1,316 @@
+# SPDX-License-Identifier: BSD-2-Clause
+# 
+# Copyright (c) 2025 Maarten L. Hekkelman
+# 
+# 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.
+
+# A simplified wrapper CMakeLists.txt file for PCRE2
+#
+# This will generate an OBJECT library so it can be linked into another library
+
+cmake_minimum_required(VERSION 3.25)
+
+include(FetchContent)
+
+project(pcre2s VERSION 1.0.0 LANGUAGES C CXX)
+
+# The original code:
+
+file(DOWNLOAD https://github.com/PCRE2Project/pcre2/releases/download/pcre2-10.46/pcre2-10.46.tar.gz
+    ${CMAKE_CURRENT_BINARY_DIR}/pcre2-code.tgz
+    EXPECTED_HASH SHA256=8d28d7f2c3b970c3a4bf3776bcbb5adfc923183ce74bc8df1ebaad8c1985bd07)
+file(ARCHIVE_EXTRACT INPUT ${CMAKE_CURRENT_BINARY_DIR}/pcre2-code.tgz
+    DESTINATION ${CMAKE_CURRENT_BINARY_DIR})
+set(PCRE2_SOURCE_DIR ${CMAKE_CURRENT_BINARY_DIR}/pcre2-10.46)
+set(PCRE2_MAJOR 10)
+set(PCRE2_MINOR 46)
+set(PCRE2_VERSION "${PCRE2_MAJOR}.${PCRE2_MINOR}")
+set(PCRE2_DATE "2024-06-09")
+
+# Some needed configuration options
+
+# option(PCRE2_BUILD_PCRE2_8 "Build 8 bit PCRE2 library" ON)
+# option(PCRE2_BUILD_PCRE2_16 "Build 16 bit PCRE2 library" OFF)
+# option(PCRE2_BUILD_PCRE2_32 "Build 32 bit PCRE2 library" OFF)
+
+option(PCRE2_STATIC_PIC "Build the static library with the option position independent code enabled." OFF)
+
+set(PCRE2_NEWLINE "LF" CACHE STRING "What to recognize as a newline (one of CR, LF, CRLF, ANY, ANYCRLF, NUL)." FORCE)
+set_property(CACHE PCRE2_NEWLINE PROPERTY STRINGS "CR" "LF" "CRLF" "ANY" "ANYCRLF" "NUL")
+
+set(PCRE2_LINK_SIZE "2" CACHE STRING "Internal link size (2, 3 or 4 allowed). See LINK_SIZE in config.h.in for details.")
+set_property(CACHE PCRE2_LINK_SIZE PROPERTY STRINGS "2" "3" "4")
+
+set(PCRE2_PARENS_NEST_LIMIT "250" CACHE STRING "Default nested parentheses limit. See PARENS_NEST_LIMIT in config.h.in for details.")
+set(PCRE2_HEAP_LIMIT "20000000" CACHE STRING "Default limit on heap memory (kibibytes). See HEAP_LIMIT in config.h.in for details.")
+set(PCRE2_MAX_VARLOOKBEHIND "255" CACHE STRING "Default limit on variable lookbehinds.")
+set(PCRE2_MATCH_LIMIT "10000000" CACHE STRING "Default limit on internal looping. See MATCH_LIMIT in config.h.in for details.")
+set(PCRE2_MATCH_LIMIT_DEPTH "MATCH_LIMIT" CACHE STRING "Default limit on internal depth of search. See MATCH_LIMIT_DEPTH in config.h.in for details.")
+set(PCRE2GREP_BUFSIZE "20480" CACHE STRING "Buffer starting size parameter for pcre2grep. See PCRE2GREP_BUFSIZE in config.h.in for details.")
+set(PCRE2GREP_MAX_BUFSIZE "1048576" CACHE STRING "Buffer maximum size parameter for pcre2grep. See PCRE2GREP_MAX_BUFSIZE in config.h.in for details.")
+set(PCRE2_SUPPORT_JIT OFF CACHE BOOL "Enable support for Just-in-time compiling.")
+
+if(${CMAKE_SYSTEM_NAME} MATCHES Linux|NetBSD)
+    set(PCRE2_SUPPORT_JIT_SEALLOC OFF CACHE BOOL "Enable SELinux compatible execmem allocator in JIT (experimental).")
+else()
+    set(PCRE2_SUPPORT_JIT_SEALLOC IGNORE)
+endif()
+
+set(PCRE2GREP_SUPPORT_JIT ON CACHE BOOL "Enable use of Just-in-time compiling in pcre2grep.")
+set(PCRE2GREP_SUPPORT_CALLOUT ON CACHE BOOL "Enable callout string support in pcre2grep.")
+set(PCRE2GREP_SUPPORT_CALLOUT_FORK ON CACHE BOOL "Enable callout string fork support in pcre2grep.")
+set(PCRE2_SUPPORT_UNICODE ON CACHE BOOL "Enable support for Unicode and UTF-8/UTF-16/UTF-32 encoding.")
+set(PCRE2_SUPPORT_BSR_ANYCRLF OFF CACHE BOOL "ON=Backslash-R matches only LF CR and CRLF, OFF=Backslash-R matches all Unicode Linebreaks")
+set(PCRE2_NEVER_BACKSLASH_C OFF CACHE BOOL "If ON, backslash-C (upper case C) is locked out.")
+set(PCRE2_SUPPORT_VALGRIND OFF CACHE BOOL "Enable Valgrind support.")
+
+if(MINGW)
+    option(NON_STANDARD_LIB_PREFIX "ON=Shared libraries built in mingw will be named pcre2.dll, etc., instead of libpcre2.dll, etc." OFF)
+    option(NON_STANDARD_LIB_SUFFIX "ON=Shared libraries built in mingw will be named libpcre2-0.dll, etc., instead of libpcre2.dll, etc." OFF)
+endif()
+
+# 
+
+set(NEWLINE_DEFAULT "")
+
+if(PCRE2_NEWLINE STREQUAL "CR")
+    set(NEWLINE_DEFAULT "1")
+elseif(PCRE2_NEWLINE STREQUAL "LF")
+    set(NEWLINE_DEFAULT "2")
+elseif(PCRE2_NEWLINE STREQUAL "CRLF")
+    set(NEWLINE_DEFAULT "3")
+elseif(PCRE2_NEWLINE STREQUAL "ANY")
+    set(NEWLINE_DEFAULT "4")
+elseif(PCRE2_NEWLINE STREQUAL "ANYCRLF")
+    set(NEWLINE_DEFAULT "5")
+elseif(PCRE2_NEWLINE STREQUAL "NUL")
+    set(NEWLINE_DEFAULT "6")
+else()
+    message(FATAL_ERROR "The PCRE2_NEWLINE variable must be set to one of the following values: \"LF\", \"CR\", \"CRLF\", \"ANY\", \"ANYCRLF\".")
+endif()
+
+# Some tests
+
+include(CheckCSourceCompiles)
+include(CheckFunctionExists)
+include(CheckSymbolExists)
+include(CheckIncludeFile)
+
+check_include_file(assert.h HAVE_ASSERT_H)
+check_include_file(dirent.h HAVE_DIRENT_H)
+check_include_file(sys/stat.h HAVE_SYS_STAT_H)
+check_include_file(sys/types.h HAVE_SYS_TYPES_H)
+check_include_file(unistd.h HAVE_UNISTD_H)
+check_include_file(windows.h HAVE_WINDOWS_H)
+
+check_symbol_exists(bcopy "strings.h" HAVE_BCOPY)
+check_symbol_exists(memfd_create "sys/mman.h" HAVE_MEMFD_CREATE)
+check_symbol_exists(memmove "string.h" HAVE_MEMMOVE)
+check_symbol_exists(secure_getenv "stdlib.h" HAVE_SECURE_GETENV)
+check_symbol_exists(strerror "string.h" HAVE_STRERROR)
+
+check_c_source_compiles(
+  "int main(void) { char buf[128] __attribute__((uninitialized)); (void)buf; return 0; }"
+  HAVE_ATTRIBUTE_UNINITIALIZED
+)
+
+check_c_source_compiles(
+  [=[
+  extern __attribute__ ((visibility ("default"))) int f(void);
+  int main(void) { return f(); }
+  int f(void) { return 42; }
+  ]=]
+  HAVE_VISIBILITY
+)
+
+if(HAVE_VISIBILITY)
+  set(PCRE2_EXPORT [=[__attribute__ ((visibility ("default")))]=])
+else()
+  set(PCRE2_EXPORT)
+endif()
+
+check_c_source_compiles("int main(void) { __assume(1); return 0; }" HAVE_BUILTIN_ASSUME)
+
+check_c_source_compiles(
+  [=[
+  #include <stddef.h>
+  int main(void) { int a,b; size_t m; __builtin_mul_overflow(a,b,&m); return 0; }
+  ]=]
+  HAVE_BUILTIN_MUL_OVERFLOW
+)
+
+check_c_source_compiles(
+  "int main(int c, char *v[]) { if (c) __builtin_unreachable(); return (int)(*v[0]); }"
+  HAVE_BUILTIN_UNREACHABLE
+)
+
+# # Check whether Intel CET is enabled, and if so, adjust compiler flags. This
+# # code was written by PH, trying to imitate the logic from the autotools
+# # configuration.
+
+# check_c_source_compiles(
+#   [=[
+#   #ifndef __CET__
+#   #error CET is not enabled
+#   #endif
+#   int main() { return 0; }
+#   ]=]
+#   INTEL_CET_ENABLED
+# )
+
+# if(INTEL_CET_ENABLED)
+#   set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mshstk")
+# endif()
+
+# Set up some dependencies first
+
+configure_file(
+    ${PCRE2_SOURCE_DIR}/src/pcre2_chartables.c.dist
+    ${CMAKE_CURRENT_BINARY_DIR}/pcre2_chartables.c
+    COPYONLY
+)
+
+configure_file(
+    ${PCRE2_SOURCE_DIR}/config-cmake.h.in
+    ${CMAKE_CURRENT_BINARY_DIR}/interface/config.h
+    @ONLY
+)
+
+configure_file(
+    ${PCRE2_SOURCE_DIR}/src/pcre2.h.in
+    ${CMAKE_CURRENT_BINARY_DIR}/interface/pcre2.h
+    @ONLY
+)
+
+# Define our library
+
+list(APPEND PCRE2_HEADERS
+    ${CMAKE_CURRENT_BINARY_DIR}/interface/pcre2.h)
+
+list(APPEND PCRE2_SOURCES
+    ${PCRE2_SOURCE_DIR}/src/pcre2_auto_possess.c
+    ${CMAKE_CURRENT_BINARY_DIR}/pcre2_chartables.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_chkdint.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_compile.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_compile_class.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_config.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_context.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_convert.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_dfa_match.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_error.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_extuni.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_find_bracket.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_jit_compile.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_maketables.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_match.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_match_data.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_newline.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_ord2utf.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_pattern_info.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_script_run.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_serialize.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_string_utils.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_study.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_substitute.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_substring.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_tables.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_ucd.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_valid_utf.c
+    ${PCRE2_SOURCE_DIR}/src/pcre2_xclass.c
+)
+
+add_library(pcre2s OBJECT)
+
+target_sources(pcre2s
+    PRIVATE ${PCRE2_SOURCES}
+    PUBLIC
+    FILE_SET pcre2_headers TYPE HEADERS
+    BASE_DIRS ${PCRE2_SOURCE_DIR}/include ${CMAKE_CURRENT_BINARY_DIR}/interface
+    FILES ${PCRE2_HEADERS}
+)
+
+target_compile_definitions(pcre2s PUBLIC PCRE2_CODE_UNIT_WIDTH=8 HAVE_CONFIG_H)
+if(NOT BUILD_SHARED_LIBS)
+    target_compile_definitions(pcre2s PUBLIC PCRE2_STATIC)
+endif()
+
+target_include_directories(pcre2s PRIVATE ${CMAKE_CURRENT_BINARY_DIR}/interface ${PCRE2_SOURCE_DIR}/src)
+
+if(PCRE2_STATIC_PIC)
+    set_target_properties(pcre2s PROPERTIES POSITION_INDEPENDENT_CODE 1)
+endif()
+
+# # Installation and config files
+
+# include(CMakePackageConfigHelpers)
+# include(GenerateExportHeader)
+
+# # Install rules
+# install(TARGETS pcre2s
+#     EXPORT pcre2s
+#     FILE_SET pcre2_headers DESTINATION ${CMAKE_INSTALL_INCLUDEDIR})
+
+# if(MSVC AND BUILD_SHARED_LIBS)
+#     install(
+#         FILES $<TARGET_PDB_FILE:pcre2s>
+#         DESTINATION ${CMAKE_INSTALL_LIBDIR}
+#         OPTIONAL)
+# endif()
+
+# install(EXPORT pcre2s
+#     NAMESPACE pcre2s::
+#     FILE "pcre2s-targets.cmake"
+#     DESTINATION lib/cmake/pcre2s)
+
+# configure_package_config_file(
+#     ${CMAKE_CURRENT_SOURCE_DIR}/pcre2s-config.cmake.in ${CMAKE_CURRENT_BINARY_DIR}/pcre2s/pcre2s-config.cmake
+#     INSTALL_DESTINATION lib/cmake/pcre2s)
+
+# install(
+#     FILES "${CMAKE_CURRENT_BINARY_DIR}/pcre2s/pcre2s-config.cmake"
+#     "${CMAKE_CURRENT_BINARY_DIR}/pcre2s/pcre2s-config-version.cmake"
+#     DESTINATION lib/cmake/pcre2s)
+
+# set_target_properties(
+#     pcre2s
+#     PROPERTIES VERSION ${PCRE2_VERSION}
+#     SOVERSION ${PCRE2_VERSION}
+#     INTERFACE_pcre2s_MAJOR_VERSION ${PCRE2_MAJOR})
+
+# set_property(
+#     TARGET pcre2s
+#     APPEND
+#     PROPERTY COMPATIBLE_INTERFACE_STRING pcre2s_MAJOR_VERSION)
+
+# write_basic_package_version_file(
+#     "${CMAKE_CURRENT_BINARY_DIR}/pcre2s/pcre2s-config-version.cmake"
+#     VERSION "${PCRE2_VERSION}"
+#     COMPATIBILITY AnyNewerVersion)
+
+# # Testing
+
+# if(PROJECT_IS_TOP_LEVEL)
+#     include(CTest)
+
+#     if(BUILD_TESTING)
+#         add_subdirectory(test)
+#     endif()
+# endif()

src/category.cpp

@@ -92,7 +92,7 @@ class row_comparator
 		return d;
 	}
 
-	int operator()(const category &cat, const row_initializer &a, const row *b) const
+	int operator()(const category &cat, const category::key_type &a, const row *b) const
 	{
 		assert(b);
 
@@ -105,10 +105,11 @@ class row_comparator
 		{
 			assert(ai != a.end());
 
-			std::string_view ka = ai->value();
+			std::string_view ka = ai->value;
 			std::string_view kb = rhb[k].text();
 
-			d = f(ka, kb);
+			if (not (ai->may_be_null and rhb[k].empty()))
+				d = f(ka, kb);
 
 			if (d != 0)
 				break;
@@ -142,7 +143,7 @@ class category_index
 	}
 
 	row *find(const category &cat, row *k) const;
-	row *find_by_value(const category &cat, row_initializer k) const;
+	row *find_by_value(const category &cat, const category::key_type &k) const;
 
 	void insert(category &cat, row *r);
 	void erase(category &cat, row *r);
@@ -352,19 +353,19 @@ row *category_index::find(const category &cat, row *k) const
 	return r ? r->m_row : nullptr;
 }
 
-row *category_index::find_by_value(const category &cat, row_initializer k) const
+row *category_index::find_by_value(const category &cat, const category::key_type &k) const
 {
 	// sort the values in k first
 
-	row_initializer k2;
+	category::key_type k2;
 	for (auto &f : cat.key_item_indices())
 	{
 		auto fld = cat.get_item_name(f);
 
 		auto ki = find_if(k.begin(), k.end(), [&fld](auto &i)
-			{ return i.name() == fld; });
+			{ return i.name == fld; });
 		if (ki == k.end())
-			k2.emplace_back(fld, "");
+			k2.emplace_back(std::string{ fld }, "");
 		else
 			k2.emplace_back(*ki);
 	}
@@ -1338,8 +1339,7 @@ std::string category::get_unique_value(std::string_view item_name)
 		// brain-dead implementation
 		for (std::size_t ix = 0; ix < size(); ++ix)
 		{
-			// result = m_name + "-" + std::to_string(ix);
-			result = cif_id_for_number(ix);
+			result = cif_id_for_number(static_cast<int>(ix));
 			if (not contains(key(item_name) == result))
 				break;
 		}

src/compound.cpp

@@ -480,10 +480,13 @@ compound *local_compound_factory_impl::create(const std::string &id)
 
 			try
 			{
-				const auto &[id, name, threeLetterCode, group] =
+				const auto &[id2, name, threeLetterCode, group] =
 					chem_comp->front().get<std::string, std::string, std::string, std::string>("id", "name", "three_letter_code", "group");
 
-				result = construct_compound(db, id, name, threeLetterCode, group);
+				if (id == id2)
+					result = construct_compound(db, id, name, threeLetterCode, group);
+				else
+					throw std::runtime_error("Compound ID's don't match: id 1=" + id + ", id 2=" + id2);
 			}
 			catch (const std::exception &ex)
 			{
@@ -757,8 +760,7 @@ bool compound_factory::is_monomer(std::string_view res_name) const
 
 void compound_factory::report_missing_compound(std::string_view compound_id)
 {
-	static bool s_reported = false;
-	if (std::exchange(s_reported, true) == false)
+	if (std::exchange(m_report_missing, false))
 	{
 		using namespace cif::colour;
 

src/condition.cpp

@@ -24,8 +24,8 @@
  * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
-#include "cif++/category.hpp"
 #include "cif++/condition.hpp"
+#include "cif++/category.hpp"
 #include "cif++/validate.hpp"
 
 namespace cif
@@ -61,6 +61,52 @@ bool is_item_type_uchar(const category &cat, std::string_view col)
 
 namespace detail
 {
+	// 	// index lookup
+	// 	struct index_lookup_condition_impl : public condition_impl
+	// 	{
+	// 		index_lookup_condition_impl(row_initializer &&key_values)
+	// 			: m_key_values(std::move(key_values))
+	// 		{
+	// 		}
+	//
+	// 		condition_impl *prepare(const category &c) override
+	// 		{
+	// 			m_single_hit = c[m_key_values];
+	// 			return this;
+	// 		}
+	//
+	// 		bool test(row_handle r) const override
+	// 		{
+	// 			return m_single_hit == r;
+	// 		}
+	//
+	// 		void str(std::ostream &os) const override
+	// 		{
+	// 			os << "index scan";
+	// 		}
+	//
+	// 		virtual std::optional<row_handle> single() const override
+	// 		{
+	// 			return m_single_hit;
+	// 		}
+	//
+	// 		virtual bool equals(const condition_impl *rhs) const override
+	// 		{
+	// 			if (typeid(*rhs) == typeid(index_lookup_condition_impl))
+	// 			{
+	// 				auto ri = static_cast<const index_lookup_condition_impl *>(rhs);
+	// 				if (m_single_hit or ri->m_single_hit)
+	// 					return m_single_hit == ri->m_single_hit;
+	// 				else
+	// 					// watch out, both m_item_ix might be the same while item_names might be diffent (in case they both do not exist in the category)
+	// 					return m_key_values == ri->m_key_values;
+	// 			}
+	// 			return this == rhs;
+	// 		}
+	//
+	// 		row_initializer m_key_values;
+	// 		row_handle m_single_hit;
+	// 	};
 
 	condition_impl *key_equals_condition_impl::prepare(const category &c)
 	{
@@ -85,7 +131,8 @@ namespace detail
 			c.key_item_indices().contains(m_item_ix) and
 			c.key_item_indices().size() == 1)
 		{
-			m_single_hit = c[{ { m_item_name, m_value } }];
+			item v(m_item_name, m_value);
+			m_single_hit = c[{ { m_item_name, std::string{ v.value() }, false } }];
 		}
 
 		return this;
@@ -99,7 +146,8 @@ namespace detail
 		{
 			auto &cs = (*s)->m_sub;
 
-			if (find_if(cs.begin(), cs.end(), [c](const condition_impl *i) { return i->equals(c); }) == cs.end())
+			if (find_if(cs.begin(), cs.end(), [c](const condition_impl *i)
+					{ return i->equals(c); }) == cs.end())
 			{
 				result = false;
 				break;
@@ -119,7 +167,8 @@ namespace detail
 		for (size_t fc_i = 0; fc_i < fc.size();)
 		{
 			auto c = fc[fc_i];
-			if (not found_in_range(c, subs.begin() + 1, subs.end())) {
+			if (not found_in_range(c, subs.begin() + 1, subs.end()))
+			{
 				++fc_i;
 				continue;
 			}
@@ -137,11 +186,12 @@ namespace detail
 				for (size_t ssub_i = 0; ssub_i < ssub.size();)
 				{
 					auto sc = ssub[ssub_i];
-					if (not sc->equals(c)) {
+					if (not sc->equals(c))
+					{
 						++ssub_i;
 						continue;
 					}
-					
+
 					ssub.erase(ssub.begin() + ssub_i);
 					delete sc;
 					break;
@@ -158,6 +208,99 @@ namespace detail
 		return oc;
 	}
 
+	condition_impl *and_condition_impl::prepare(const category &c)
+	{
+		for (auto &sub : m_sub)
+			sub = sub->prepare(c);
+
+		if (auto cv = c.get_cat_validator(); cv != nullptr)
+		{
+			// See if we can collapse a search part of this and_condition into a single index lookup
+
+			cif::iset keys{ cv->m_keys.begin(), cv->m_keys.end() };
+			category::key_type lookup;
+			std::vector<condition_impl *> subs;
+			std::vector<std::string> may_be_empty;
+
+			for (auto &sub : m_sub)
+			{
+				if (auto s = dynamic_cast<const key_equals_condition_impl *>(sub); s != nullptr)
+				{
+					if (keys.contains(s->m_item_name))
+					{
+						lookup.emplace_back(s->m_item_name, s->m_value);
+						subs.emplace_back(sub);
+					}
+					continue;
+				}
+
+				if (auto s = dynamic_cast<const key_equals_number_condition_impl *>(sub); s != nullptr)
+				{
+					if (keys.contains(s->m_item_name))
+					{
+						item v{ s->m_item_name, s->m_value };
+						lookup.emplace_back(s->m_item_name, std::string{ v.value() } );
+						subs.emplace_back(sub);
+					}
+					continue;
+				}
+
+				if (auto s = dynamic_cast<const key_equals_or_empty_condition_impl *>(sub); s != nullptr)
+				{
+					if (keys.contains(s->m_item_name))
+					{
+						lookup.emplace_back(s->m_item_name, s->m_value, true);
+						subs.emplace_back(sub);
+						may_be_empty.emplace_back(s->m_item_name);
+					}
+					continue;
+				}
+
+				if (auto s = dynamic_cast<const key_equals_number_or_empty_condition_impl *>(sub); s != nullptr)
+				{
+					if (keys.contains(s->m_item_name))
+					{
+						item v{ s->m_item_name, s->m_value };
+						lookup.emplace_back(s->m_item_name, std::string{ v.value() }, true );
+						subs.emplace_back(sub);
+					}
+					continue;
+				}
+			}
+
+			if (lookup.size() == keys.size())
+			{
+				m_single = c[lookup];
+
+				for (auto s : subs)
+					m_sub.erase(std::remove(m_sub.begin(), m_sub.end(), s), m_sub.end());
+			}
+		}
+
+		return this;
+	}
+
+	bool and_condition_impl::test(row_handle r) const
+	{
+		bool result = true;
+
+		if (m_single.has_value() and *m_single != r)
+			result = false;
+		else
+		{
+			for (auto sub : m_sub)
+			{
+				if (sub->test(r))
+					continue;
+
+				result = false;
+				break;
+			}
+		}
+
+		return result;
+	}
+
 	condition_impl *or_condition_impl::prepare(const category &c)
 	{
 		std::vector<and_condition_impl *> and_conditions;
@@ -181,7 +324,7 @@ void condition::prepare(const category &c)
 {
 	if (m_impl)
 		m_impl = m_impl->prepare(c);
-	
+
 	m_prepared = true;
 }
 

src/file.cpp

@@ -25,6 +25,7 @@
  */
 
 #include "cif++/file.hpp"
+#include "cif++/condition.hpp"
 #include "cif++/gzio.hpp"
 
 namespace cif
@@ -46,8 +47,16 @@ bool file::is_valid()
 {
 	bool result = not empty();
 
-	for (auto &d : *this)
-		result = d.is_valid() and result;
+	for (bool first = true; auto &d : *this)
+	{
+		if (first)
+		{
+			result = d.is_valid() and result;
+			first = false;
+		}
+		else if (d.get_validator() != nullptr)
+			result = d.is_valid() and result;
+	}
 
 	if (result)
 		result = validate_links();

src/model.cpp

@@ -47,15 +47,10 @@ void atom::atom_impl::moveTo(const point &p)
 
 	auto r = row();
 
-#if __cpp_lib_format
-	r.assign("Cartn_x", std::format("{:.3f}", p.m_x), false, false);
-	r.assign("Cartn_y", std::format("{:.3f}", p.m_y), false, false);
-	r.assign("Cartn_z", std::format("{:.3f}", p.m_z), false, false);
-#else
-	r.assign("Cartn_x", cif::format("%.3f", p.m_x).str(), false, false);
-	r.assign("Cartn_y", cif::format("%.3f", p.m_y).str(), false, false);
-	r.assign("Cartn_z", cif::format("%.3f", p.m_z).str(), false, false);
-#endif
+	r.assign("Cartn_x", cif::format("{:.3f}", p.m_x), false, false);
+	r.assign("Cartn_y", cif::format("{:.3f}", p.m_y), false, false);
+	r.assign("Cartn_z", cif::format("{:.3f}", p.m_z), false, false);
+
 	m_location = p;
 }
 
@@ -1300,7 +1295,7 @@ void structure::load_atoms_for_model(structure_open_options options)
 	
 		for (auto &&[key, value] : alts)
 		{
-			const auto &[asym_id, seq_id] = key;
+			// const auto &[asym_id, seq_id] = key;
 	
 			// select highest occupancy for this residue's alternates
 			std::string alt_id;
@@ -2874,7 +2869,7 @@ static int compare_numbers(std::string_view a, std::string_view b)
 
 int compare_cif_id(const std::string &a, const std::string &b)
 {
-	int d = a.length() - b.length();
+	int d = static_cast<int>(a.length() - b.length());
 	if (d == 0)
 		d = a.compare(b);
 	return d;

src/pdb/cif2pdb.cpp

@@ -33,7 +33,6 @@
 #include <regex>
 #include <set>
 
-
 namespace cif::pdb
 {
 
@@ -58,9 +57,9 @@ std::string cif2pdbDate(const std::string &d)
 		int month = std::stoi(m[2].str());
 
 		if (m[3].matched)
-			result = cif::format("%02.2d-%3.3s-%02.2d", stoi(m[3].str()), kMonths[month - 1], (year % 100)).str();
+			result = cif::format("{:02}-{:3.3}-{:02}", stoi(m[3].str()), kMonths[month - 1], (year % 100));
 		else
-			result = cif::format("%3.3s-%02.2d", kMonths[month - 1], (year % 100)).str();
+			result = cif::format("{:3.3}-{:02}", kMonths[month - 1], (year % 100));
 	}
 
 	return result;
@@ -258,16 +257,14 @@ std::size_t WriteCitation(std::ostream &pdbFile, const datablock &db, row_handle
 	{
 		to_upper(pubname);
 
-		const std::string kRefHeader = s1 + "REF %2.2s %-28.28s  %2.2s%4.4s %5.5s %4.4s";
-		pdbFile << cif::format(kRefHeader, "" /* continuation */, pubname, (volume.empty() ? "" : "V."), volume, pageFirst, year)
+		pdbFile << s1 << cif::format("REF {:2.2s} {:<28.28s}  {:2.2s}{:>4.4s} {:>5.5s} {:4.4s}", "" /* continuation */, pubname, (volume.empty() ? "" : "V."), volume, pageFirst, year)
 				<< '\n';
 		++result;
 	}
 
 	if (not issn.empty())
 	{
-		const std::string kRefHeader = s1 + "REFN                   ISSN %-25.25s";
-		pdbFile << cif::format(kRefHeader, issn) << '\n';
+		pdbFile << s1 << cif::format("REFN                   ISSN {:<25.25s}", issn) << '\n';
 		++result;
 	}
 
@@ -276,27 +273,25 @@ std::size_t WriteCitation(std::ostream &pdbFile, const datablock &db, row_handle
 	////    0         1         2         3         4         5         6         7         8
 	////    HEADER    xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxDDDDDDDDD   IIII
 	// const char kRefHeader[] =
-	//          "REMARK   1  REFN    %4.4s %-6.6s  %2.2s %-25.25s";
+	//          "REMARK   1  REFN    {:4.4s} {:<6.6s}  {:2.2s} {:<25.25s}";
 	//
 	//			pdbFile << (boost::cif::format(kRefHeader)
 	//						% (astm.empty() ? "" : "ASTN")
 	//						% astm
 	//						% country
-	//						% issn).str()
+	//						% issn)
 	//					<< '\n';
 	//		}
 
 	if (not pmid.empty())
 	{
-		const std::string kPMID = s1 + "PMID   %-60.60s ";
-		pdbFile << cif::format(kPMID, pmid) << '\n';
+		pdbFile << s1 << cif::format("PMID   {:<60.60s} ", pmid) << '\n';
 		++result;
 	}
 
 	if (not doi.empty())
 	{
-		const std::string kDOI = s1 + "DOI    %-60.60s ";
-		pdbFile << cif::format(kDOI, doi) << '\n';
+		pdbFile << s1 << cif::format("DOI    {:<60.60s} ", doi) << '\n';
 		++result;
 	}
 
@@ -307,10 +302,10 @@ void write_header_lines(std::ostream &pdbFile, const datablock &db)
 {
 	//    0         1         2         3         4         5         6         7         8
 	//    HEADER    xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxDDDDDDDDD   IIII
-	const char kHeader[] =
-		"HEADER    %-40.40s"
-		"%-9.9s"
-		"   %-4.4s";
+	// const char kHeader[] =
+	// 	"HEADER    {:<40.40s}"
+	// 	"{:<9.9s}"
+	// 	"   {:<4.4s}";
 
 	// HEADER
 
@@ -345,7 +340,12 @@ void write_header_lines(std::ostream &pdbFile, const datablock &db)
 		}
 	}
 
-	pdbFile << cif::format(kHeader, keywords, date, db.name()) << '\n';
+	pdbFile << cif::format(/* kHeader */
+		"HEADER    {:<40.40s}"
+		"{:<9.9s}"
+		"   {:<4.4s}"
+	
+	, keywords, date, db.name()) << '\n';
 
 	// TODO: implement
 	// OBSLTE (skip for now)
@@ -535,7 +535,6 @@ void WriteTitle(std::ostream &pdbFile, const datablock &db)
 	write_header_lines(pdbFile, db);
 
 	// REVDAT
-	const char kRevDatFmt[] = "REVDAT %3d%2.2s %9.9s %4.4s    %1d      ";
 	auto &cat2 = db["database_PDB_rev"];
 	std::vector<row_handle> rev(cat2.begin(), cat2.end());
 	sort(rev.begin(), rev.end(), [](row_handle a, row_handle b) -> bool
@@ -559,9 +558,9 @@ void WriteTitle(std::ostream &pdbFile, const datablock &db)
 		{
 			std::string cs = ++continuation > 1 ? std::to_string(continuation) : std::string();
 
-			pdbFile << cif::format(kRevDatFmt, revNum, cs, date, db.name(), modType);
+			pdbFile << cif::format("REVDAT {:3}{:2.2s} {:9.9s} {:4.4s}    {:1}      ", revNum, cs, date, db.name(), modType);
 			for (std::size_t i = 0; i < 4; ++i)
-				pdbFile << cif::format(" %-6.6s", (i < types.size() ? types[i] : std::string()));
+				pdbFile << cif::format(" {:<6.6s}", (i < types.size() ? types[i] : std::string()));
 			pdbFile << '\n';
 
 			if (types.size() > 4)
@@ -614,7 +613,7 @@ void WriteRemark2(std::ostream &pdbFile, const datablock &db)
 		{
 			float resHigh = refine.front()["ls_d_res_high"].as<float>();
 			pdbFile << "REMARK   2\n"
-					<< cif::format("REMARK   2 RESOLUTION. %7.2f ANGSTROMS.", resHigh) << '\n';
+					<< cif::format("REMARK   2 RESOLUTION. {:7.2f} ANGSTROMS.", resHigh) << '\n';
 		}
 		catch (...)
 		{ /* skip it */
@@ -761,10 +760,7 @@ class Fs : public FBase
 		else
 		{
 			os << '\n';
-
-			std::stringstream ss;
-			ss << "REMARK " << std::setw(3) << std::right << mNr << ' ';
-			WriteOneContinuedLine(os, ss.str(), 0, s);
+			WriteOneContinuedLine(os, cif::format("REMARK {:3} ", mNr), 0, s);
 		}
 	}
 
@@ -1617,7 +1613,7 @@ void WriteRemark3Phenix(std::ostream &pdbFile, const datablock &db)
 
 		percent_reflns_obs /= 100;
 
-		pdbFile << RM3("  ") << cif::format("%3d %7.4f - %7.4f    %4.2f %8d %5d  %6.4f %6.4f", bin++, d_res_low, d_res_high, percent_reflns_obs, number_reflns_R_work, number_reflns_R_free, R_factor_R_work, R_factor_R_free) << '\n';
+		pdbFile << RM3("  ") << cif::format("{:3} {:7.4f} - {:7.4f}    {:4.2f} {:8} {:5}  {:6.4f} {:6.4f}", bin++, d_res_low, d_res_high, percent_reflns_obs, number_reflns_R_work, number_reflns_R_free, R_factor_R_work, R_factor_R_free) << '\n';
 	}
 
 	pdbFile << RM3("") << '\n'
@@ -2585,7 +2581,7 @@ void WriteRemark465(std::ostream &pdbFile, const datablock &db)
 		cif::tie(modelNr, resName, chainID, iCode, seqNr) =
 			r.get("PDB_model_num", "auth_comp_id", "auth_asym_id", "PDB_ins_code", "auth_seq_id");
 
-		pdbFile << cif::format("REMARK 465 %3.3s %3.3s %1.1s %5d%1.1s", modelNr, resName, chainID, seqNr, iCode) << '\n';
+		pdbFile << cif::format("REMARK 465 {:3.3s} {:3.3s} {:1.1s} {:5}{:1.1s}", modelNr, resName, chainID, seqNr, iCode) << '\n';
 	}
 }
 
@@ -2632,7 +2628,7 @@ void WriteRemark470(std::ostream &pdbFile, const datablock &db)
 
 			while (not a.second.empty())
 			{
-				pdbFile << cif::format("REMARK 470 %3.3s %3.3s %1.1s%4d%1.1s  ", modelNr, resName, chainID, seqNr, iCode) << "  ";
+				pdbFile << cif::format("REMARK 470 {:>3.3s} {:3.3s} {:1.1s}{:4}{:1.1s}  ", modelNr, resName, chainID, seqNr, iCode) << "  ";
 
 				for (std::size_t i = 0; i < 6 and not a.second.empty(); ++i)
 				{
@@ -2730,16 +2726,16 @@ int WritePrimaryStructure(std::ostream &pdbFile, const datablock &db)
 
 			if (dbAccession.length() > 8 or db_code.length() > 12 or atoi(dbseqEnd.c_str()) >= 100000)
 				pdbFile << cif::format(
-							   "DBREF1 %4.4s %1.1s %4.4s%1.1s %4.4s%1.1s %-6.6s               %-20.20s",
+							   "DBREF1 {:>4.4s} {:1.1s} {:>4.4s}{:1.1s} {:>4.4s}{:1.1s} {:<6.6s}               {:<20.20s}",
 							   idCode, chainID, seqBegin, insertBegin, seqEnd, insertEnd, db_name, db_code)
 						<< '\n'
 						<< cif::format(
-							   "DBREF2 %4.4s %1.1s     %-22.22s     %10.10s  %10.10s",
+							   "DBREF2 {:>4.4s} {:1.1s}     {:<22.22s}     {:10.10s}  {:10.10s}",
 							   idCode, chainID, dbAccession, dbseqBegin, dbseqEnd)
 						<< '\n';
 			else
 				pdbFile << cif::format(
-							   "DBREF  %4.4s %1.1s %4.4s%1.1s %4.4s%1.1s %-6.6s %-8.8s %-12.12s %5.5s%1.1s %5.5s%1.1s",
+							   "DBREF  {:>4.4s} {:1.1s} {:>4.4s}{:1.1s} {:>4.4s}{:1.1s} {:<6.6s} {:<8.8s} {:<12.12s} {:>5.5s}{:1.1s} {:>5.5s}{:1.1s}",
 							   idCode, chainID, seqBegin, insertBegin, seqEnd, insertEnd, db_name, dbAccession, db_code, dbseqBegin, dbinsBeg, dbseqEnd, dbinsEnd)
 						<< '\n';
 		}
@@ -2758,9 +2754,8 @@ int WritePrimaryStructure(std::ostream &pdbFile, const datablock &db)
 		to_upper(conflict);
 
 		pdbFile << cif::format(
-					   "SEQADV %4.4s %3.3s %1.1s %4.4s%1.1s %-4.4s %-9.9s %3.3s %5.5s %-21.21s",
+					   "SEQADV {:4.4s} {:3.3s} {:1.1s} {:>4.4s}{:1.1s} {:<4.4s} {:<9.9s} {:3.3s} {:>5.5s} {:<21.21s}",
 					   idCode, resName, chainID, seqNum, iCode, database, dbAccession, dbRes, dbSeq, conflict)
-					   .str()
 				<< '\n';
 	}
 
@@ -2788,7 +2783,7 @@ int WritePrimaryStructure(std::ostream &pdbFile, const datablock &db)
 				t = 13;
 
 			pdbFile << cif::format(
-						   "SEQRES %3d %1.1s %4d  %-51.51s          ",
+						   "SEQRES {:3} {:1.1s} {:4}  {:<51.51s}          ",
 						   n++, std::string{ chainID }, seqresl[chainID], join(seq.begin(), seq.begin() + t, " "))
 					<< '\n';
 
@@ -2808,9 +2803,8 @@ int WritePrimaryStructure(std::ostream &pdbFile, const datablock &db)
 			r.get("auth_asym_id", "auth_seq_id", "auth_comp_id", "PDB_ins_code", "parent_comp_id", "details");
 
 		pdbFile << cif::format(
-					   "MODRES %4.4s %3.3s %1.1s %4.4s%1.1s %3.3s  %-41.41s",
+					   "MODRES {:4.4s} {:3.3s} {:1.1s} {:4.4s}{:1.1s} {:3.3s}  {:<41.41s}",
 					   db.name(), resName, chainID, seqNum, iCode, stdRes, comment)
-					   .str()
 				<< '\n';
 	}
 
@@ -2925,7 +2919,7 @@ int WriteHeterogen(std::ostream &pdbFile, const datablock &db)
 	{
 		if (h.water)
 			continue;
-		pdbFile << cif::format("HET    %3.3s  %c%4d%c  %5d", h.hetID, h.chainID, h.seqNum, h.iCode, h.numHetAtoms) << '\n';
+		pdbFile << cif::format("HET    {:3.3s}  {:1c}{:4}{:1c}  {:5}", h.hetID, h.chainID, h.seqNum, h.iCode, h.numHetAtoms) << '\n';
 		++numHet;
 	}
 
@@ -2940,7 +2934,7 @@ int WriteHeterogen(std::ostream &pdbFile, const datablock &db)
 
 		for (;;)
 		{
-			pdbFile << cif::format("HETNAM  %2.2s %3.3s ", (c > 1 ? std::to_string(c) : std::string()), id);
+			pdbFile << cif::format("HETNAM  {:2.2s} {:3.3s} ", (c > 1 ? std::to_string(c) : std::string()), id);
 			++c;
 
 			if (name.length() > 55)
@@ -3032,7 +3026,7 @@ int WriteHeterogen(std::ostream &pdbFile, const datablock &db)
 			{
 				std::stringstream fs;
 
-				fs << cif::format("FORMUL  %2d  %3.3s %2.2s%c", componentNr, hetID, (c > 1 ? std::to_string(c) : std::string()), (hetID == water_comp_id ? '*' : ' '));
+				fs << cif::format("FORMUL  {:2}  {:3.3s} {:2.2s}{:1c}", componentNr, hetID, (c > 1 ? std::to_string(c) : std::string()), (hetID == water_comp_id ? '*' : ' '));
 				++c;
 
 				if (formula.length() > 51)
@@ -3099,7 +3093,7 @@ std::tuple<int, int> WriteSecondaryStructure(std::ostream &pdbFile, const databl
 				"pdbx_PDB_helix_class", "pdbx_PDB_helix_length", "beg_auth_seq_id", "end_auth_seq_id");
 
 		++numHelix;
-		pdbFile << cif::format("HELIX  %3d %3.3s %3.3s %1.1s %4d%1.1s %3.3s %1.1s %4d%1.1s%2d%-30.30s %5d",
+		pdbFile << cif::format("HELIX  {:3} {:>3.3s} {:3.3s} {:1.1s} {:4}{:1.1s} {:3.3s} {:1.1s} {:4}{:1.1s}{:2}{:<30.30s} {:5}",
 					   numHelix, pdbx_PDB_helix_id, beg_label_comp_id, beg_auth_asym_id, beg_auth_seq_id, pdbx_beg_PDB_ins_code, end_label_comp_id, end_auth_asym_id, end_auth_seq_id, pdbx_end_PDB_ins_code, pdbx_PDB_helix_class, details, pdbx_PDB_helix_length)
 				<< '\n';
 	}
@@ -3136,7 +3130,7 @@ std::tuple<int, int> WriteSecondaryStructure(std::ostream &pdbFile, const databl
 					"pdbx_end_PDB_ins_code", "beg_auth_comp_id", "beg_auth_asym_id", "beg_auth_seq_id",
 					"end_auth_comp_id", "end_auth_asym_id", "end_auth_seq_id");
 
-				pdbFile << cif::format("SHEET  %3.3s %3.3s%2d %3.3s %1.1s%4d%1.1s %3.3s %1.1s%4d%1.1s%2d", rangeID1, sheetID, numStrands, initResName, initChainID, initSeqNum, initICode, endResName, endChainID, endSeqNum, endICode, 0) << '\n';
+				pdbFile << cif::format("SHEET  {:>3.3s} {:>3.3s}{:2} {:3.3s} {:1.1s}{:4}{:1.1s} {:3.3s} {:1.1s}{:4}{:1.1s}{:2}", rangeID1, sheetID, numStrands, initResName, initChainID, initSeqNum, initICode, endResName, endChainID, endSeqNum, endICode, 0) << '\n';
 
 				first = false;
 			}
@@ -3155,7 +3149,7 @@ std::tuple<int, int> WriteSecondaryStructure(std::ostream &pdbFile, const databl
 
 			if (h.empty())
 			{
-				pdbFile << cif::format("SHEET  %3.3s %3.3s%2d %3.3s %1.1s%4d%1.1s %3.3s %1.1s%4d%1.1s%2d", rangeID2, sheetID, numStrands, initResName, initChainID, initSeqNum, initICode, endResName, endChainID, endSeqNum, endICode, sense) << '\n';
+				pdbFile << cif::format("SHEET  {:>3.3s} {:>3.3s}{:2} {:3.3s} {:1.1s}{:4}{:1.1s} {:3.3s} {:1.1s}{:4}{:1.1s}{:2}", rangeID2, sheetID, numStrands, initResName, initChainID, initSeqNum, initICode, endResName, endChainID, endSeqNum, endICode, sense) << '\n';
 			}
 			else
 			{
@@ -3168,8 +3162,8 @@ std::tuple<int, int> WriteSecondaryStructure(std::ostream &pdbFile, const databl
 				curAtom = cif2pdbAtomName(curAtom, compID[0], db);
 				prevAtom = cif2pdbAtomName(prevAtom, compID[1], db);
 
-				pdbFile << cif::format("SHEET  %3.3s %3.3s%2d %3.3s %1.1s%4d%1.1s %3.3s %1.1s%4d%1.1s%2d "
-										"%-4.4s%3.3s %1.1s%4d%1.1s %-4.4s%3.3s %1.1s%4d%1.1s",
+				pdbFile << cif::format("SHEET  {:>3.3s} {:>3.3s}{:2} {:3.3s} {:1.1s}{:4}{:1.1s} {:3.3s} {:1.1s}{:4}{:1.1s}{:2} "
+										"{:<4.4s}{:3.3s} {:1.1s}{:4}{:1.1s} {:<4.4s}{:3.3s} {:1.1s}{:4}{:1.1s}",
 							   rangeID2, sheetID, numStrands, initResName, initChainID, initSeqNum, initICode, endResName, endChainID, endSeqNum, endICode, sense, curAtom, curResName, curChainID, curResSeq, curICode, prevAtom, prevResName, prevChainID, prevResSeq, prevICode)
 						<< '\n';
 			}
@@ -3207,7 +3201,7 @@ void WriteConnectivity(std::ostream &pdbFile, const datablock &db)
 		sym1 = cif2pdbSymmetry(sym1);
 		sym2 = cif2pdbSymmetry(sym2);
 
-		pdbFile << cif::format("SSBOND %3d CYS %1.1s %4d%1.1s   CYS %1.1s %4d%1.1s                       %6.6s %6.6s %5.2f", nr, chainID1, seqNum1, icode1, chainID2, seqNum2, icode2, sym1, sym2, Length) << '\n';
+		pdbFile << cif::format("SSBOND {:3} CYS {:1.1s} {:4}{:1.1s}   CYS {:1.1s} {:4}{:1.1s}                       {:6.6s} {:6.6s} {:5.2f}", nr, chainID1, seqNum1, icode1, chainID2, seqNum2, icode2, sym1, sym2, Length) << '\n';
 
 		++nr;
 	}
@@ -3234,10 +3228,10 @@ void WriteConnectivity(std::ostream &pdbFile, const datablock &db)
 		sym1 = cif2pdbSymmetry(sym1);
 		sym2 = cif2pdbSymmetry(sym2);
 
-		pdbFile << cif::format("LINK        %-4.4s%1.1s%3.3s %1.1s%4d%1.1s               %-4.4s%1.1s%3.3s %1.1s%4d%1.1s  %6.6s %6.6s", name1, altLoc1, resName1, chainID1, resSeq1, iCode1, name2, altLoc2, resName2, chainID2, resSeq2, iCode2, sym1, sym2);
+		pdbFile << cif::format("LINK        {:<4.4s}{:1.1s}{:3.3s} {:1.1s}{:4}{:1.1s}               {:<4.4s}{:1.1s}{:3.3s} {:1.1s}{:4}{:1.1s}  {:>6.6s} {:>6.6s}", name1, altLoc1, resName1, chainID1, resSeq1, iCode1, name2, altLoc2, resName2, chainID2, resSeq2, iCode2, sym1, sym2);
 
 		if (not Length.empty())
-			pdbFile << cif::format(" %5.2f", stod(Length));
+			pdbFile << cif::format(" {:5.2f}", stod(Length));
 
 		pdbFile << '\n';
 	}
@@ -3255,7 +3249,7 @@ void WriteConnectivity(std::ostream &pdbFile, const datablock &db)
 				"pdbx_label_comp_id_2", "pdbx_auth_asym_id_2", "pdbx_auth_seq_id_2", "pdbx_PDB_ins_code_2",
 				"pdbx_PDB_model_num", "pdbx_omega_angle");
 
-		pdbFile << cif::format("CISPEP %3.3s %3.3s %1.1s %4d%1.1s   %3.3s %1.1s %4d%1.1s       %3.3s       %6.2f",
+		pdbFile << cif::format("CISPEP {:3.3s} {:3.3s} {:1.1s} {:4}{:1.1s}   {:3.3s} {:1.1s} {:4}{:1.1s}       {:3.3s}       {:6.2f}",
 			serNum, pep1, chainID1, seqNum1, icode1, pep2, chainID2, seqNum2, icode2, modNum, measure) << '\n';
 	}
 }
@@ -3276,7 +3270,7 @@ int WriteMiscellaneousFeatures(std::ostream &pdbFile, const datablock &db)
 		cif::tie(siteID, resName, chainID, seq, iCode) =
 			r.get("site_id", "auth_comp_id", "auth_asym_id", "auth_seq_id", "pdbx_auth_ins_code");
 
-		sites[siteID].push_back(cif::format("%3.3s %1.1s%4d%1.1s ", resName, chainID, seq, iCode).str());
+		sites[siteID].push_back(cif::format("{:3.3s} {:1.1s}{:4}{:1.1s} ", resName, chainID, seq, iCode));
 	}
 
 	for (auto s : sites)
@@ -3289,7 +3283,7 @@ int WriteMiscellaneousFeatures(std::ostream &pdbFile, const datablock &db)
 		int nr = 1;
 		while (res.empty() == false)
 		{
-			pdbFile << cif::format("SITE   %3d %3.3s %2d ", nr, siteID, numRes);
+			pdbFile << cif::format("SITE   {:3} {:3.3s} {:2} ", nr, siteID, numRes);
 
 			for (int i = 0; i < 4; ++i)
 			{
@@ -3318,7 +3312,7 @@ void WriteCrystallographic(std::ostream &pdbFile, const datablock &db)
 
 	r = db["cell"].find_first(key("entry_id") == db.name());
 
-	pdbFile << cif::format("CRYST1%9.3f%9.3f%9.3f%7.2f%7.2f%7.2f %-11.11s%4d", r["length_a"].as<double>(), r["length_b"].as<double>(), r["length_c"].as<double>(), r["angle_alpha"].as<double>(), r["angle_beta"].as<double>(), r["angle_gamma"].as<double>(), symmetry, r["Z_PDB"].as<int>()) << '\n';
+	pdbFile << cif::format("CRYST1{:9.3f}{:9.3f}{:9.3f}{:7.2f}{:7.2f}{:7.2f} {:<11.11s}{:4}", r["length_a"].as<double>(), r["length_b"].as<double>(), r["length_c"].as<double>(), r["angle_alpha"].as<double>(), r["angle_beta"].as<double>(), r["angle_gamma"].as<double>(), symmetry, r["Z_PDB"].as<int>()) << '\n';
 }
 
 int WriteCoordinateTransformation(std::ostream &pdbFile, const datablock &db)
@@ -3327,18 +3321,18 @@ int WriteCoordinateTransformation(std::ostream &pdbFile, const datablock &db)
 
 	for (auto r : db["database_PDB_matrix"])
 	{
-		pdbFile << cif::format("ORIGX%1d    %10.6f%10.6f%10.6f     %10.5f", 1, r["origx[1][1]"].as<float>(), r["origx[1][2]"].as<float>(), r["origx[1][3]"].as<float>(), r["origx_vector[1]"].as<float>()) << '\n';
-		pdbFile << cif::format("ORIGX%1d    %10.6f%10.6f%10.6f     %10.5f", 2, r["origx[2][1]"].as<float>(), r["origx[2][2]"].as<float>(), r["origx[2][3]"].as<float>(), r["origx_vector[2]"].as<float>()) << '\n';
-		pdbFile << cif::format("ORIGX%1d    %10.6f%10.6f%10.6f     %10.5f", 3, r["origx[3][1]"].as<float>(), r["origx[3][2]"].as<float>(), r["origx[3][3]"].as<float>(), r["origx_vector[3]"].as<float>()) << '\n';
+		pdbFile << cif::format("ORIGX{:1}    {:10.6f}{:10.6f}{:10.6f}     {:10.5f}", 1, r["origx[1][1]"].as<float>(), r["origx[1][2]"].as<float>(), r["origx[1][3]"].as<float>(), r["origx_vector[1]"].as<float>()) << '\n';
+		pdbFile << cif::format("ORIGX{:1}    {:10.6f}{:10.6f}{:10.6f}     {:10.5f}", 2, r["origx[2][1]"].as<float>(), r["origx[2][2]"].as<float>(), r["origx[2][3]"].as<float>(), r["origx_vector[2]"].as<float>()) << '\n';
+		pdbFile << cif::format("ORIGX{:1}    {:10.6f}{:10.6f}{:10.6f}     {:10.5f}", 3, r["origx[3][1]"].as<float>(), r["origx[3][2]"].as<float>(), r["origx[3][3]"].as<float>(), r["origx_vector[3]"].as<float>()) << '\n';
 		result += 3;
 		break;
 	}
 
 	for (auto r : db["atom_sites"])
 	{
-		pdbFile << cif::format("SCALE%1d    %10.6f%10.6f%10.6f     %10.5f", 1, r["fract_transf_matrix[1][1]"].as<float>(), r["fract_transf_matrix[1][2]"].as<float>(), r["fract_transf_matrix[1][3]"].as<float>(), r["fract_transf_vector[1]"].as<float>()) << '\n';
-		pdbFile << cif::format("SCALE%1d    %10.6f%10.6f%10.6f     %10.5f", 2, r["fract_transf_matrix[2][1]"].as<float>(), r["fract_transf_matrix[2][2]"].as<float>(), r["fract_transf_matrix[2][3]"].as<float>(), r["fract_transf_vector[2]"].as<float>()) << '\n';
-		pdbFile << cif::format("SCALE%1d    %10.6f%10.6f%10.6f     %10.5f", 3, r["fract_transf_matrix[3][1]"].as<float>(), r["fract_transf_matrix[3][2]"].as<float>(), r["fract_transf_matrix[3][3]"].as<float>(), r["fract_transf_vector[3]"].as<float>()) << '\n';
+		pdbFile << cif::format("SCALE{:1}    {:10.6f}{:10.6f}{:10.6f}     {:10.5f}", 1, r["fract_transf_matrix[1][1]"].as<float>(), r["fract_transf_matrix[1][2]"].as<float>(), r["fract_transf_matrix[1][3]"].as<float>(), r["fract_transf_vector[1]"].as<float>()) << '\n';
+		pdbFile << cif::format("SCALE{:1}    {:10.6f}{:10.6f}{:10.6f}     {:10.5f}", 2, r["fract_transf_matrix[2][1]"].as<float>(), r["fract_transf_matrix[2][2]"].as<float>(), r["fract_transf_matrix[2][3]"].as<float>(), r["fract_transf_vector[2]"].as<float>()) << '\n';
+		pdbFile << cif::format("SCALE{:1}    {:10.6f}{:10.6f}{:10.6f}     {:10.5f}", 3, r["fract_transf_matrix[3][1]"].as<float>(), r["fract_transf_matrix[3][2]"].as<float>(), r["fract_transf_matrix[3][3]"].as<float>(), r["fract_transf_vector[3]"].as<float>()) << '\n';
 		result += 3;
 		break;
 	}
@@ -3348,9 +3342,9 @@ int WriteCoordinateTransformation(std::ostream &pdbFile, const datablock &db)
 	{
 		std::string given = r["code"] == "given" ? "1" : "";
 
-		pdbFile << cif::format("MTRIX%1d %3d%10.6f%10.6f%10.6f     %10.5f    %1.1s", 1, nr, r["matrix[1][1]"].as<float>(), r["matrix[1][2]"].as<float>(), r["matrix[1][3]"].as<float>(), r["vector[1]"].as<float>(), given) << '\n';
-		pdbFile << cif::format("MTRIX%1d %3d%10.6f%10.6f%10.6f     %10.5f    %1.1s", 2, nr, r["matrix[2][1]"].as<float>(), r["matrix[2][2]"].as<float>(), r["matrix[2][3]"].as<float>(), r["vector[2]"].as<float>(), given) << '\n';
-		pdbFile << cif::format("MTRIX%1d %3d%10.6f%10.6f%10.6f     %10.5f    %1.1s", 3, nr, r["matrix[3][1]"].as<float>(), r["matrix[3][2]"].as<float>(), r["matrix[3][3]"].as<float>(), r["vector[3]"].as<float>(), given) << '\n';
+		pdbFile << cif::format("MTRIX{:1} {:3}{:10.6f}{:10.6f}{:10.6f}     {:10.5f}    {:1.1s}", 1, nr, r["matrix[1][1]"].as<float>(), r["matrix[1][2]"].as<float>(), r["matrix[1][3]"].as<float>(), r["vector[1]"].as<float>(), given) << '\n';
+		pdbFile << cif::format("MTRIX{:1} {:3}{:10.6f}{:10.6f}{:10.6f}     {:10.5f}    {:1.1s}", 2, nr, r["matrix[2][1]"].as<float>(), r["matrix[2][2]"].as<float>(), r["matrix[2][3]"].as<float>(), r["vector[2]"].as<float>(), given) << '\n';
+		pdbFile << cif::format("MTRIX{:1} {:3}{:10.6f}{:10.6f}{:10.6f}     {:10.5f}    {:1.1s}", 3, nr, r["matrix[3][1]"].as<float>(), r["matrix[3][2]"].as<float>(), r["matrix[3][3]"].as<float>(), r["vector[3]"].as<float>(), given) << '\n';
 
 		++nr;
 		result += 3;
@@ -3369,10 +3363,6 @@ std::tuple<int, int> WriteCoordinatesForModel(std::ostream &pdbFile, const datab
 
 	auto &atom_site = db["atom_site"];
 	auto &atom_site_anisotrop = db["atom_site_anisotrop"];
-	auto &entity = db["entity"];
-	// auto &pdbx_poly_seq_scheme = db["pdbx_poly_seq_scheme"];
-	// auto &pdbx_nonpoly_scheme = db["pdbx_nonpoly_scheme"];
-	auto &pdbx_branch_scheme = db["pdbx_branch_scheme"];
 
 	int serial = 1;
 	auto ri = atom_site.begin();
@@ -3417,7 +3407,7 @@ std::tuple<int, int> WriteCoordinatesForModel(std::ostream &pdbFile, const datab
 
 			if (terminate)
 			{
-				pdbFile << cif::format("TER   %5d      %3.3s %1.1s%4d%1.1s",  serial,  resName,  chainID,  resSeq,  iCode) << '\n';
+				pdbFile << cif::format("TER   {:5}      {:3.3s} {:1.1s}{:4}{:1.1s}",  serial,  resName,  chainID,  resSeq,  iCode) << '\n';
 
 				++serial;
 				terminatedChains.insert(chainID);
@@ -3446,26 +3436,6 @@ std::tuple<int, int> WriteCoordinatesForModel(std::ostream &pdbFile, const datab
 			r.get("id", "group_PDB", "label_atom_id", "label_alt_id", "auth_comp_id", "auth_asym_id", "auth_seq_id",
 				"pdbx_PDB_ins_code", "Cartn_x", "Cartn_y", "Cartn_z", "occupancy", "B_iso_or_equiv", "type_symbol", "pdbx_formal_charge");
 
-		if (resName != "HOH")
-		{
-			int entity_id = r.get<int>("label_entity_id");
-			try
-			{
-				auto type = entity.find1<std::string>("id"_key == entity_id, "type");
-
-				if (type == "branched")	// find the real auth_seq_num, since sugars have their auth_seq_num reused as sugar number... sigh.
-					resSeq = pdbx_branch_scheme.find1<int>("asym_id"_key == r.get<std::string>("label_asym_id") and "pdb_seq_num"_key == resSeq, "auth_seq_num");
-				// else if (type == "non-polymer")	// same for non-polymers
-				// 	resSeq = pdbx_nonpoly_scheme.find1<int>("asym_id"_key == r.get<std::string>("label_asym_id") and "pdb_seq_num"_key == resSeq, "auth_seq_num");
-				// else if (type == "polymer")
-				// 	resSeq = pdbx_poly_seq_scheme.find1<int>("asym_id"_key == r.get<std::string>("label_asym_id") and "pdb_seq_num"_key == resSeq, "auth_seq_num");
-			}
-			catch (const std::exception &ex)
-			{
-				std::cerr << "Oops, there was not exactly one entity with id " << entity_id << '\n';
-			}
-		}
-		
 		if (chainID.length() > 1)
 			throw std::runtime_error("Chain ID " + chainID + " won't fit into a PDB file");
 
@@ -3476,7 +3446,8 @@ std::tuple<int, int> WriteCoordinatesForModel(std::ostream &pdbFile, const datab
 		if (charge != 0)
 			sCharge = std::to_string(charge) + (charge > 0 ? '+' : '-');
 
-		pdbFile << cif::format("%-6.6s%5d %-4.4s%1.1s%3.3s %1.1s%4d%1.1s   %8.3f%8.3f%8.3f%6.2f%6.2f          %2.2s%2.2s", group, serial, name, altLoc, resName, chainID, resSeq, iCode, x, y, z, occupancy, tempFactor, element, sCharge) << '\n';
+		pdbFile << cif::format("{:<6.6s}{:5} {:<4.4s}{:1.1s}{:3.3s} {:1.1s}{:4}{:1.1s}   {:8.3f}{:8.3f}{:8.3f}{:6.2f}{:6.2f}          {:>2.2s}{:2.2s}",
+			group, serial, name, altLoc, resName, chainID, resSeq, iCode, x, y, z, occupancy, tempFactor, element, sCharge) << '\n';
 
 		++numCoord;
 
@@ -3491,7 +3462,7 @@ std::tuple<int, int> WriteCoordinatesForModel(std::ostream &pdbFile, const datab
 			tie(u11, u22, u33, u12, u13, u23) =
 				ai.get("U[1][1]", "U[2][2]", "U[3][3]", "U[1][2]", "U[1][3]", "U[2][3]");
 
-			pdbFile << cif::format("ANISOU%5d %-4.4s%1.1s%3.3s %1.1s%4d%1.1s %7d%7d%7d%7d%7d%7d      %2.2s%2.2s", serial, name, altLoc, resName, chainID, resSeq, iCode, std::lrintf(u11 * 10000), std::lrintf(u22 * 10000), std::lrintf(u33 * 10000), std::lrintf(u12 * 10000), std::lrintf(u13 * 10000), std::lrintf(u23 * 10000), element, sCharge) << '\n';
+			pdbFile << cif::format("ANISOU{:5} {:<4.4s}{:1.1s}{:3.3s} {:1.1s}{:4}{:1.1s} {:7}{:7}{:7}{:7}{:7}{:7}      {:2.2s}{:2.2s}", serial, name, altLoc, resName, chainID, resSeq, iCode, std::lrintf(u11 * 10000), std::lrintf(u22 * 10000), std::lrintf(u33 * 10000), std::lrintf(u12 * 10000), std::lrintf(u13 * 10000), std::lrintf(u23 * 10000), element, sCharge) << '\n';
 		}
 
 		++serial;
@@ -3543,7 +3514,7 @@ std::tuple<int, int> WriteCoordinate(std::ostream &pdbFile, const datablock &db)
 		for (int model_nr : models)
 		{
 			if (models.size() > 1)
-				pdbFile << cif::format("MODEL     %4d",  model_nr) << '\n';
+				pdbFile << cif::format("MODEL     {:4}",  model_nr) << '\n';
 
 			std::set<std::string> TERminatedChains;
 			auto n = WriteCoordinatesForModel(pdbFile, db, last_resseq_for_chain_map, TERminatedChains, model_nr);
@@ -3615,7 +3586,7 @@ std::string get_HEADER_line(const datablock &db, std::string::size_type truncate
 		}
 	}
 
-	return FixStringLength(cif::format("HEADER    %-40.40s%-9.9s   %-4.4s", keywords, date, db.name()).str(), truncate_at);
+	return FixStringLength(cif::format("HEADER    {:<40.40s}{:<9.9s}   {:<4.4s}", keywords, date, db.name()), truncate_at);
 }
 
 std::string get_COMPND_line(const datablock &db, std::string::size_type truncate_at)
@@ -3788,7 +3759,7 @@ void write(std::ostream &os, const datablock &db)
 	numXform = WriteCoordinateTransformation(os, db);
 	std::tie(numCoord, numTer) = WriteCoordinate(os, db);
 
-	os << cif::format("MASTER    %5d    0%5d%5d%5d%5d%5d%5d%5d%5d%5d%5d",  numRemark,  numHet,  numHelix,  numSheet,  numTurn,  numSite,  numXform,  numCoord,  numTer,  numConect,  numSeq) << '\n'
+	os << cif::format("MASTER    {:5}    0{:5}{:5}{:5}{:5}{:5}{:5}{:5}{:5}{:5}{:5}",  numRemark,  numHet,  numHelix,  numSheet,  numTurn,  numSite,  numXform,  numCoord,  numTer,  numConect,  numSeq) << '\n'
 			<< "END\n";
 }
 

src/pdb/pdb2cif.cpp

@@ -32,6 +32,7 @@
 #include <map>
 #include <set>
 #include <stack>
+#include <stdexcept>
 
 using cif::category;
 using cif::datablock;
@@ -895,12 +896,7 @@ class PDBFileParser
 				if (year < 1950)
 					year += 100;
 
-				std::stringstream ss;
-				ss << std::setw(4) << std::setfill('0') << year << '-'
-				   << std::setw(2) << std::setfill('0') << month << '-'
-				   << std::setw(2) << std::setfill('0') << day;
-
-				s = ss.str();
+				s = cif::format("{:04}-{:02}-{:02}", year, month, day);
 			}
 			else if (regex_match(s, m, rx2))
 			{
@@ -912,7 +908,7 @@ class PDBFileParser
 				if (year < 1950)
 					year += 100;
 
-				s = cif::format("%04d-%02d", year, month).str();
+				s = cif::format("{:04}-{:02}", year, month);
 			}
 			else
 				ec = error::make_error_code(error::pdbErrors::invalidDate);
@@ -3341,18 +3337,18 @@ void PDBFileParser::ParseRemark350()
 								{ "type", type },
 								// { "name", "" },
 							    // { "symmetryOperation", "" },
-								{ "matrix[1][1]", cif::format("%12.10f", mat[0]).str() },
-								{ "matrix[1][2]", cif::format("%12.10f", mat[1]).str() },
-								{ "matrix[1][3]", cif::format("%12.10f", mat[2]).str() },
-								{ "vector[1]", cif::format("%12.10f", vec[0]).str() },
-								{ "matrix[2][1]", cif::format("%12.10f", mat[3]).str() },
-								{ "matrix[2][2]", cif::format("%12.10f", mat[4]).str() },
-								{ "matrix[2][3]", cif::format("%12.10f", mat[5]).str() },
-								{ "vector[2]", cif::format("%12.10f", vec[1]).str() },
-								{ "matrix[3][1]", cif::format("%12.10f", mat[6]).str() },
-								{ "matrix[3][2]", cif::format("%12.10f", mat[7]).str() },
-								{ "matrix[3][3]", cif::format("%12.10f", mat[8]).str() },
-								{ "vector[3]", cif::format("%12.10f", vec[2]).str() }
+								{ "matrix[1][1]", cif::format("{:12.10f}", mat[0]) },
+								{ "matrix[1][2]", cif::format("{:12.10f}", mat[1]) },
+								{ "matrix[1][3]", cif::format("{:12.10f}", mat[2]) },
+								{ "vector[1]", cif::format("{:12.10f}", vec[0]) },
+								{ "matrix[2][1]", cif::format("{:12.10f}", mat[3]) },
+								{ "matrix[2][2]", cif::format("{:12.10f}", mat[4]) },
+								{ "matrix[2][3]", cif::format("{:12.10f}", mat[5]) },
+								{ "vector[2]", cif::format("{:12.10f}", vec[1]) },
+								{ "matrix[3][1]", cif::format("{:12.10f}", mat[6]) },
+								{ "matrix[3][2]", cif::format("{:12.10f}", mat[7]) },
+								{ "matrix[3][3]", cif::format("{:12.10f}", mat[8]) },
+								{ "vector[3]", cif::format("{:12.10f}", vec[2]) }
 							});
 																			// clang-format on
 
@@ -4318,7 +4314,7 @@ void PDBFileParser::ConstructEntities()
 	}
 
 	// build sugar trees first
-	ConstructSugarTrees(asymNr);
+	// ConstructSugarTrees(asymNr);
 
 	// done with the sugar, resume operation as before
 
@@ -5772,6 +5768,9 @@ void PDBFileParser::ParseCoordinate(int modelNr)
 		std::string element = vS(77, 78);    //	77 - 78        LString(2)    element      Element symbol, right-justified.
 		std::string charge = vS(79, 80);     //	79 - 80        LString(2)    charge       Charge  on the atom.
 
+		if (element.empty())
+			throw std::runtime_error("Empty element column in PDB file at line " + std::to_string(mRec->mLineNr));
+
 		std::string entityID = mAsymID2EntityID[asymID];
 
 		charge = pdb2cifCharge(charge);
@@ -5850,7 +5849,7 @@ void PDBFileParser::ParseCoordinate(int modelNr)
 
 			auto f = [](float f) -> std::string
 			{
-				return cif::format("%6.4f", f).str();
+				return cif::format("{:6.4f}", f);
 			};
 
 			// clang-format off
@@ -6406,7 +6405,7 @@ file read(std::istream &is)
 		if (std::isalpha(ch) and std::toupper(ch) != 'D')
 		{
 			read_pdb_file(is, result);
-			reconstruct_pdbx(result);
+			fixup_pdbx(result);
 		}
 		else
 		{
@@ -6442,8 +6441,14 @@ file read(std::istream &is)
 	}
 
 	// Must be a PDB like file, right?
-	if (not result.empty() and result.front().get_validator() == nullptr)
-		result.front().set_validator(&validator_factory::instance().get("mmcif_pdbx.dic"));
+	if (not result.empty())
+	{
+		auto &db = result.front();
+		if (db.get_validator() == nullptr)
+			db.set_validator(&validator_factory::instance().get("mmcif_pdbx.dic"));
+		if (db.is_valid())
+			db.get_validator()->fill_audit_conform(db["audit_conform"]);
+	}
 
 	return result;
 }

src/pdb/reconstruct.cpp

@@ -25,6 +25,8 @@
  */
 
 #include "cif++.hpp"
+#include "cif++/compound.hpp"
+#include "cif++/row.hpp"
 
 // --------------------------------------------------------------------
 
@@ -138,15 +140,15 @@ void checkEntities(datablock &db)
 				auto compound = cf.create(comp_id);
 				if (compound)
 					formula_weight += compound->formula_weight();
-				else if (cif::VERBOSE > 0)
-					std::clog << "missing information for compound " + comp_id << '\n';
+				// else if (cif::VERBOSE > 0)
+				// 	std::clog << "missing information for compound " + comp_id << '\n';
 				++n;
 			}
 
-			formula_weight -= (n - 1) * 18.015;
+			formula_weight -= (n - 1) * 18.015f;
 		}
 		else if (type == "water")
-			formula_weight = 18.015;
+			formula_weight = 18.015f;
 		else if (type == "branched")
 		{
 			int n = 0;
@@ -156,12 +158,12 @@ void checkEntities(datablock &db)
 				auto compound = cf.create(comp_id);
 				if (compound)
 					formula_weight += compound->formula_weight();
-				else if (cif::VERBOSE > 0)
-					std::clog << "missing information for compound " + comp_id << '\n';
+				// else if (cif::VERBOSE > 0)
+				// 	std::clog << "missing information for compound " + comp_id << '\n';
 				++n;
 			}
 
-			formula_weight -= (n - 1) * 18.015;
+			formula_weight -= (n - 1) * 18.015f;
 		}
 		else if (type == "non-polymer")
 		{
@@ -171,7 +173,7 @@ void checkEntities(datablock &db)
 				auto compound = cf.create(*comp_id);
 				if (not compound)
 				{
-					std::cerr << "missing information for compound " << *comp_id << "\n";
+					// std::cerr << "missing information for compound " << *comp_id << "\n";
 					continue;
 				}
 				formula_weight = compound->formula_weight();
@@ -198,26 +200,28 @@ void createEntityIDs(datablock &db)
 	auto &atom_site = db["atom_site"];
 	auto &cf = compound_factory::instance();
 
-	std::vector<std::vector<residue_key_type>> entities;
+	std::vector<std::vector<row_handle>> entities;
 
 	std::string lastAsymID;
 	int lastSeqID = -1;
-	std::vector<residue_key_type> waters;
+	std::vector<row_handle> waters;
 
-	for (residue_key_type k : atom_site.rows<std::optional<std::string>,
+	for (auto rh : atom_site)
+	{
+		residue_key_type k = rh.get<std::optional<std::string>,
 			 std::optional<int>,
 			 std::optional<std::string>,
 			 std::optional<std::string>,
 			 std::optional<int>,
 			 std::optional<std::string>>(
 			 "auth_asym_id", "auth_seq_id", "auth_comp_id",
-			 "label_asym_id", "label_seq_id", "label_comp_id"))
-	{
+			 "label_asym_id", "label_seq_id", "label_comp_id");
+
 		std::string comp_id = get_comp_id(k);
 
 		if (cf.is_water(comp_id))
 		{
-			waters.emplace_back(k);
+			waters.emplace_back(rh);
 			continue;
 		}
 
@@ -232,7 +236,7 @@ void createEntityIDs(datablock &db)
 		if (asym_id != lastAsymID or (not is_monomer and lastSeqID != seq_id))
 			entities.push_back({});
 
-		entities.back().emplace_back(k);
+		entities.back().emplace_back(rh);
 
 		lastAsymID = asym_id;
 		lastSeqID = seq_id;
@@ -259,20 +263,17 @@ void createEntityIDs(datablock &db)
 
 	for (std::size_t ix = 0; auto &e : entities)
 	{
-		auto k = e.front();
 		const auto &entity_id = entity_ids[ix++];
 
-		std::string comp_id = get_comp_id(k);
-
-		for (auto &k : e)
-			atom_site.update_value(get_condition(k), "label_entity_id", entity_id);
+		for (auto rh : e)
+			rh["label_entity_id"] = entity_id;
 	}
 
 	if (not waters.empty())
 	{
 		std::string waterEntityID = std::to_string(entities.size() + 1);
-		for (auto &k : waters)
-			atom_site.update_value(get_condition(k), "label_entity_id", waterEntityID);
+		for (auto rh : waters)
+			rh["label_entity_id"] = waterEntityID;
 	}
 }
 
@@ -319,7 +320,7 @@ void fillLabelAsymID(category &atom_site)
 		{
 			if (not mapAuthAsymIDAndEntityToLabelAsymID.contains(key))
 			{
-				std::string asym_id = cif_id_for_number(mapAuthAsymIDAndEntityToLabelAsymID.size());
+				std::string asym_id = cif_id_for_number(static_cast<int>(mapAuthAsymIDAndEntityToLabelAsymID.size()));
 				mapAuthAsymIDAndEntityToLabelAsymID[key] = asym_id;
 			}
 		}
@@ -438,10 +439,39 @@ void checkAtomRecords(datablock &db)
 	// And negative seq_id values
 	if (atom_site.contains(key("label_seq_id") < 0))
 		fixNegativeSeqID(atom_site);
+	
+	std::set<std::string> polymer_entities;
+	if (db["entity"].empty())
+	{
+		// No entity, so we have to guess the types based on the content of atom_site
 
-	std::set<int> polymer_entities;
-	for (int id : db["entity"].find<int>("type"_key == "polymer", "id"))
-		polymer_entities.insert(id);
+		std::string last_entity_id;
+		std::optional<int> last_label_seq_id, last_auth_seq_id;
+
+		std::set<std::string> entityIDs;
+		for (auto &[entity_id, label_comp_id, label_seq_id, auth_comp_id, auth_seq_id] :
+			atom_site.rows<std::string, std::string, std::optional<int>, std::string, std::optional<int>>(
+				"label_entity_id", "label_comp_id", "label_seq_id", "auth_comp_id", "auth_seq_id"))
+		{
+			if (cf.is_water(label_comp_id) or cf.is_water(auth_comp_id))
+				continue;
+
+			if (polymer_entities.contains(entity_id))
+				continue;
+
+			if (last_entity_id == entity_id and (last_label_seq_id != label_seq_id or last_auth_seq_id != auth_seq_id))
+				polymer_entities.emplace(entity_id);
+
+			last_entity_id = entity_id;
+			last_label_seq_id = label_seq_id;
+			last_auth_seq_id = auth_seq_id;
+		}
+	}
+	else
+	{
+		for (std::string id : db["entity"].find<std::string>("type"_key == "polymer", "id"))
+			polymer_entities.insert(id);
+	}
 
 	std::set<std::string> missingCompounds;
 
@@ -478,13 +508,14 @@ void checkAtomRecords(datablock &db)
 		if (missingCompounds.contains(comp_id))
 			continue;
 
-		bool is_polymer = polymer_entities.contains(row["label_entity_id"].as<int>());
+		bool is_polymer = polymer_entities.contains(row["label_entity_id"].as<std::string>());
 		auto compound = cf.create(comp_id);
 
 		if (not compound)
 		{
 			missingCompounds.insert(comp_id);
-			std::cerr << "Missing compound information for " << comp_id << "\n";
+			// if (cif::VERBOSE > 0)
+			// 	std::cerr << "Missing compound information for " << comp_id << "\n";
 			continue;
 		}
 
@@ -531,18 +562,24 @@ void checkAtomRecords(datablock &db)
 		if (is_polymer and row["label_seq_id"].empty() and cf.is_monomer(comp_id))
 			row["label_seq_id"] = std::to_string(seq_id);
 
-		if (row["label_atom_id"].empty())
-			row["label_atom_id"] = row["auth_atom_id"].text();
 		if (row["label_asym_id"].empty())
 			row["label_asym_id"] = row["auth_asym_id"].text();
+		else if (row["auth_asym_id"].empty())
+			row["auth_asym_id"] = row["label_asym_id"].text();
+
 		if (row["label_comp_id"].empty())
 			row["label_comp_id"] = row["auth_comp_id"].text();
+		else if (row["auth_comp_id"].empty())
+			row["auth_comp_id"] = row["label_comp_id"].text();
+
 		if (row["label_atom_id"].empty())
 			row["label_atom_id"] = row["auth_atom_id"].text();
-
+		else if (row["auth_atom_id"].empty())
+			row["auth_atom_id"] = row["label_atom_id"].text();
+		
 		// Rewrite the coordinates and other items that look better in a fixed format
 		// Be careful not to nuke invalidly formatted data here
-		for (auto [item_name, prec] : std::vector<std::tuple<std::string_view, std::string::size_type>>{
+		for (auto [item_name, prec] : std::vector<std::tuple<std::string_view, int>>{
 				 { "cartn_x", 3 },
 				 { "cartn_y", 3 },
 				 { "cartn_z", 3 },
@@ -557,11 +594,11 @@ void checkAtomRecords(datablock &db)
 			if (auto [ptr, ec] = cif::from_chars(s.data(), s.data() + s.length(), v); (bool)ec)
 				continue;
 
-			if (s.length() < prec + 1 or s[s.length() - prec - 1] != '.')
+			if (s.length() < prec + 1UL or s[s.length() - prec - 1] != '.')
 			{
 				char b[12];
 
-				if (auto [ptr, ec] = cif::to_chars(b, b + sizeof(b), v, cif::chars_format::fixed, prec); (bool)ec)
+				if (auto [ptr, ec] = cif::to_chars(b, b + sizeof(b), v, cif::chars_format::fixed, prec); ec == std::errc{})
 					row.assign(item_name, { b, static_cast<std::string::size_type>(ptr - b) }, false, false);
 			}
 		}
@@ -603,19 +640,24 @@ void checkAtomAnisotropRecords(datablock &db)
 
 	std::vector<row_handle> to_be_deleted;
 
+	std::map<int, row_handle> atoms;
+	for (auto rh : atom_site)
+		atoms[rh.get<int>("id")] = rh;
+
 	bool warnReplaceTypeSymbol = true;
 	for (auto row : atom_site_anisotrop)
 	{
-		auto parents = atom_site_anisotrop.get_parents(row, atom_site);
-		if (parents.size() != 1)
+		auto ai = atoms.find(row.get<int>("id"));
+
+		if (ai == atoms.end())
 		{
 			to_be_deleted.emplace_back(row);
 			continue;
 		}
 
-		// this happens sometimes (Phenix):
+		auto parent = ai->second;
 
-		auto parent = parents.front();
+		// this happens sometimes (Phenix):
 
 		if (row["type_symbol"].empty())
 			row["type_symbol"] = parent["type_symbol"].text();
@@ -628,16 +670,14 @@ void checkAtomAnisotropRecords(datablock &db)
 
 		if (row["pdbx_auth_alt_id"].empty() and not parent["pdbx_auth_alt_id"].empty())
 			row["pdbx_auth_alt_id"] = parent["pdbx_auth_alt_id"].text();
-		if (row["pdbx_label_seq_id"].empty() and not parent["pdbx_label_seq_id"].empty())
+		if (row["pdbx_label_seq_id"].empty() and not parent["label_seq_id"].empty())
 			row["pdbx_label_seq_id"] = parent["label_seq_id"].text();
-		if (row["pdbx_label_asym_id"].empty() and not parent["pdbx_label_asym_id"].empty())
+		if (row["pdbx_label_asym_id"].empty() and not parent["label_asym_id"].empty())
 			row["pdbx_label_asym_id"] = parent["label_asym_id"].text();
-		if (row["pdbx_label_atom_id"].empty() and not parent["pdbx_label_atom_id"].empty())
+		if (row["pdbx_label_atom_id"].empty() and not parent["label_atom_id"].empty())
 			row["pdbx_label_atom_id"] = parent["label_atom_id"].text();
-		if (row["pdbx_label_comp_id"].empty() and not parent["pdbx_label_comp_id"].empty())
+		if (row["pdbx_label_comp_id"].empty() and not parent["label_comp_id"].empty())
 			row["pdbx_label_comp_id"] = parent["label_comp_id"].text();
-		// if (row["pdbx_PDB_model_num"].empty() and not parent["pdbx_PDB_model_num"].empty())
-		// 	row["pdbx_PDB_model_num"] = parent["pdbx_PDB_model_num"].text();
 	}
 
 	if (not to_be_deleted.empty())
@@ -722,7 +762,7 @@ void createEntity(datablock &db)
 
 		std::string type, desc;
 		float weight = 0;
-		int count = 0;
+		size_t count = 0;
 
 		auto first_comp_id = std::get<0>(content.front());
 
@@ -737,8 +777,11 @@ void createEntity(datablock &db)
 			auto c = cf.create(first_comp_id);
 
 			type = "non-polymer";
-			desc = c->name();
-			weight = c->formula_weight();
+			if (c)
+			{
+				desc = c->name();
+				weight = c->formula_weight();
+			}
 		}
 		else
 		{
@@ -809,28 +852,28 @@ void createEntityPoly(datablock &db)
 			if (type != "other")
 			{
 				std::string c_type;
-				if (cf.is_base(comp_id))
+				if (auto i = compound_factory::kBaseMap.find(comp_id); i != compound_factory::kBaseMap.end())
 				{
 					c_type = "polydeoxyribonucleotide";
-					letter_can = compound_factory::kBaseMap.at(comp_id);
+
+					letter_can = i->second;
+
 					if (comp_id.length() == 1)
 						letter = letter_can;
 					else
-						letter = '(' + letter_can + ')';
+						letter = '(' + comp_id + ')';
 				}
-				else if (cf.is_peptide(comp_id))
+				else if (auto i = compound_factory::kAAMap.find(comp_id); i != compound_factory::kAAMap.end())
 				{
 					c_type = "polypeptide(L)";
-					letter = letter_can = compound_factory::kAAMap.at(comp_id);
+
+					letter = letter_can = i->second;
 				}
 				else if (iequals(c->type(), "D-PEPTIDE LINKING"))
 				{
 					c_type = "polypeptide(D)";
 
 					letter_can = c->one_letter_code();
-					if (letter_can == 0)
-						letter_can = 'X';
-
 					letter = '(' + comp_id + ')';
 
 					non_std_linkage = true;
@@ -841,9 +884,6 @@ void createEntityPoly(datablock &db)
 					c_type = "polypeptide(L)";
 
 					letter_can = c->one_letter_code();
-					if (letter_can == 0)
-						letter_can = 'X';
-
 					letter = '(' + comp_id + ')';
 
 					non_std_monomer = true;
@@ -853,9 +893,6 @@ void createEntityPoly(datablock &db)
 					// c_type = "other";
 
 					letter_can = c->one_letter_code();
-					if (letter_can == 0)
-						letter_can = 'X';
-
 					letter = '(' + comp_id + ')';
 
 					non_std_monomer = true;
@@ -868,7 +905,7 @@ void createEntityPoly(datablock &db)
 			}
 
 			seq[auth_asym_id] += letter;
-			seq_can[auth_asym_id] += letter_can;
+			seq_can[auth_asym_id] += letter_can ? letter_can : 'X';
 
 			if (find(pdb_strand_ids.begin(), pdb_strand_ids.end(), auth_asym_id) == pdb_strand_ids.end())
 				pdb_strand_ids.emplace_back(auth_asym_id);
@@ -1167,7 +1204,7 @@ void createPdbxNonpolyScheme(datablock &db)
 		for (int ndb_nr = 1; auto row : atom_site.find("label_entity_id"_key == entity_id and "label_comp_id"_key == comp_id))
 		{
 			// Skip existing records
-			auto linked = atom_site.get_linked(row, pdbx_nonpoly_scheme);
+			auto linked = atom_site.get_children(row, pdbx_nonpoly_scheme);
 			if (not linked.empty())
 				continue;
 
@@ -1242,6 +1279,101 @@ void createPdbxBranchScheme(datablock &db)
 	}
 }
 
+void reconstruct_index_for_category(const validator &validator, category &cat, datablock &db)
+{
+	auto cv = validator.get_validator_for_category(cat.name());
+
+	enum class State
+	{
+		Start,
+		MissingKeys,
+		DuplicateKeys
+	} state = State::Start;
+
+	for (;;)
+	{
+		// See if we can build an index
+		try
+		{
+			cat.set_validator(&validator, db);
+		}
+		catch (const missing_key_error &ex)
+		{
+			if (state == State::MissingKeys)
+			{
+				if (cif::VERBOSE > 0)
+					std::clog << "Repairing failed for category " << cat.name() << ", missing keys remain: " << ex.what() << '\n';
+
+				throw;
+			}
+
+			state = State::MissingKeys;
+
+			auto key = ex.get_key();
+
+			if (cif::VERBOSE > 1)
+				std::clog << "Need to add key " << key << " to category " << cat.name() << '\n';
+
+			for (auto row : cat)
+			{
+				auto ord = row.get<std::string>(key.c_str());
+				if (ord.empty())
+					row.assign({ //
+						{ key, cat.get_unique_value(key) } });
+			}
+
+			continue;
+		}
+		catch (const duplicate_key_error &ex)
+		{
+			if (state == State::DuplicateKeys)
+			{
+				if (cif::VERBOSE > 0)
+					std::clog << "Repairing failed for category " << cat.name() << ", duplicate keys remain: " << ex.what() << '\n';
+
+				throw;
+			}
+
+			state = State::DuplicateKeys;
+
+			if (cif::VERBOSE > 0)
+				std::clog << "Attempt to fix " << cat.name() << " failed: " << ex.what() << '\n';
+
+			// replace items that do not define a relation to a parent
+
+			std::set<std::string> replaceableKeys;
+			for (auto key : cv->m_keys)
+			{
+				bool replaceable = true;
+				for (auto lv : validator.get_links_for_child(cat.name()))
+				{
+					if (find(lv->m_child_keys.begin(), lv->m_child_keys.end(), key) != lv->m_child_keys.end())
+					{
+						replaceable = false;
+						break;
+					}
+				}
+
+				if (replaceable)
+					replaceableKeys.insert(key);
+			}
+
+			if (replaceableKeys.empty())
+				throw std::runtime_error("Cannot repair category " + cat.name() + " since it contains duplicate keys that cannot be replaced");
+
+			for (auto key : replaceableKeys)
+			{
+				for (auto row : cat)
+					row.assign(key, cat.get_unique_value(key), false, false);
+			}
+
+			continue;
+		}
+
+		break;
+	}
+}
+
 bool reconstruct_pdbx(file &file)
 {
 	if (file.empty())
@@ -1337,7 +1469,7 @@ bool reconstruct_pdbx(file &file, const validator &validator)
 				              iv->m_type != nullptr and
 				              iv->m_type->m_primitive_type == cif::DDL_PrimitiveType::Numb;
 
-				for (std::size_t ix = 0; auto row : cat)
+				for (int ix = 0; auto row : cat)
 				{
 					if (number)
 						row.assign(key, std::to_string(++ix), false, false);
@@ -1408,95 +1540,7 @@ bool reconstruct_pdbx(file &file, const validator &validator)
 				}
 			}
 
-			enum class State
-			{
-				Start,
-				MissingKeys,
-				DuplicateKeys
-			} state = State::Start;
-
-			for (;;)
-			{
-				// See if we can build an index
-				try
-				{
-					cat.set_validator(&validator, db);
-				}
-				catch (const missing_key_error &ex)
-				{
-					if (state == State::MissingKeys)
-					{
-						if (cif::VERBOSE > 0)
-							std::clog << "Repairing failed for category " << cat.name() << ", missing keys remain: " << ex.what() << '\n';
-
-						throw;
-					}
-
-					state = State::MissingKeys;
-
-					auto key = ex.get_key();
-
-					if (cif::VERBOSE > 0)
-						std::clog << "Need to add key " << key << " to category " << cat.name() << '\n';
-
-					for (auto row : cat)
-					{
-						auto ord = row.get<std::string>(key.c_str());
-						if (ord.empty())
-							row.assign({ //
-								{ key, cat.get_unique_value(key) } });
-					}
-
-					continue;
-				}
-				catch (const duplicate_key_error &ex)
-				{
-					if (state == State::DuplicateKeys)
-					{
-						if (cif::VERBOSE > 0)
-							std::clog << "Repairing failed for category " << cat.name() << ", duplicate keys remain: " << ex.what() << '\n';
-
-						throw;
-					}
-
-					state = State::DuplicateKeys;
-
-					if (cif::VERBOSE > 0)
-						std::clog << "Attempt to fix " << cat.name() << " failed: " << ex.what() << '\n';
-
-					// replace items that do not define a relation to a parent
-
-					std::set<std::string> replaceableKeys;
-					for (auto key : cv->m_keys)
-					{
-						bool replaceable = true;
-						for (auto lv : validator.get_links_for_child(cat.name()))
-						{
-							if (find(lv->m_child_keys.begin(), lv->m_child_keys.end(), key) != lv->m_child_keys.end())
-							{
-								replaceable = false;
-								break;
-							}
-						}
-
-						if (replaceable)
-							replaceableKeys.insert(key);
-					}
-
-					if (replaceableKeys.empty())
-						throw std::runtime_error("Cannot repair category " + cat.name() + " since it contains duplicate keys that cannot be replaced");
-
-					for (auto key : replaceableKeys)
-					{
-						for (auto row : cat)
-							row.assign(key, cat.get_unique_value(key), false, false);
-					}
-
-					continue;
-				}
-
-				break;
-			}
+			reconstruct_index_for_category(validator, cat, db);
 		}
 		catch (const std::exception &ex)
 		{
@@ -1537,7 +1581,7 @@ bool reconstruct_pdbx(file &file, const validator &validator)
 
 	if (auto cat = db.get("ndb_poly_seq_scheme"); cat == nullptr or cat->empty())
 		comparePolySeqSchemes(db);
-	
+
 	createPdbxNonpolyScheme(db);
 
 	// Create a minimal set of branch records
@@ -1577,11 +1621,16 @@ void fixup_pdbx(file &file, const validator &validator)
 	// assuming the first datablock contains the entry ...
 	auto &db = file.front();
 
+	if (auto cat = db.get("atom_site"); cat == nullptr or cat->empty())
+		throw std::runtime_error("Cannot reconstruct PDBx file, atom data missing");
+
 	// ... and any additional datablock will contain compound information
 	cif::compound_source cs(file);
 
-	if (auto cat = db.get("atom_site"); cat == nullptr or cat->empty())
-		throw std::runtime_error("Cannot reconstruct PDBx file, atom data missing");
+	// Be silent about missing compound info in fixup
+	auto &cf = compound_factory::instance();
+	bool save_report = cf.get_report_missing();
+	cf.set_report_missing(cif::VERBOSE > 1);
 
 	std::string entry_id;
 
@@ -1610,11 +1659,24 @@ void fixup_pdbx(file &file, const validator &validator)
 	if (not db["atom_site"].find_first(key("label_entity_id") != null))
 		createEntityIDs(db);
 
-	// Now see if atom records make sense at all
+	// Now see if atom records make sense at all, but in a silent way, this time
 	checkAtomRecords(db);
 
 	db["chem_comp"].reorder_by_index();
 
+	// See if we can easily reconstruct missing data fields in order to create an index
+	for (auto &cat : db)
+	{
+		try
+		{
+			cat.set_validator(&validator, db);
+		}
+		catch (const missing_key_error &)
+		{
+			reconstruct_index_for_category(validator, cat, db);
+		}
+	}
+
 	db.set_validator(&validator);
 
 	// Now create any missing categories
@@ -1630,7 +1692,7 @@ void fixup_pdbx(file &file, const validator &validator)
 
 	if (auto cat = db.get("ndb_poly_seq_scheme"); cat == nullptr or cat->empty())
 		comparePolySeqSchemes(db);
-	
+
 	createPdbxNonpolyScheme(db);
 
 	// Create a minimal set of branch records
@@ -1640,6 +1702,7 @@ void fixup_pdbx(file &file, const validator &validator)
 	checkEntities(db);
 
 	// That's it
+	cf.set_report_missing(save_report);
 }
 
 } // namespace cif::pdb

src/pdb/validate-pdbx.cpp

@@ -61,8 +61,6 @@ condition get_parents_condition(const validator &validator, row_handle rh, const
 			result = std::move(result) or std::move(cond);
 		}
 	}
-	else if (cif::VERBOSE > 0)
-		std::cerr << "warning: no child to parent links were found for child " << childName << " and parent " << parentName << '\n';
 
 	return result;
 }
@@ -71,7 +69,7 @@ bool is_valid_pdbx_file(const file &file, const validator &v)
 {
 	std::error_code ec;
 	bool result = is_valid_pdbx_file(file, v, ec);
-	return result and not (bool)ec;
+	return result and not(bool) ec;
 }
 
 bool is_valid_pdbx_file(const file &file, std::error_code &ec)
@@ -84,7 +82,7 @@ bool is_valid_pdbx_file(const file &file, std::error_code &ec)
 		result = is_valid_pdbx_file(file, validator_factory::instance().get(*ac), ec);
 	else
 		result = is_valid_pdbx_file(file, validator_factory::instance().get("mmcif_pdbx.dic"), ec);
-	
+
 	return result;
 }
 
@@ -92,7 +90,7 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 {
 	using namespace cif::literals;
 
-	bool result = true;
+	bool result = true, warned_missing_parents = false;
 
 	try
 	{
@@ -129,10 +127,18 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 			if (not cf.is_monomer(comp_id))
 				continue;
 
-			auto p = pdbx_poly_seq_scheme.find(get_parents_condition(validator, r, pdbx_poly_seq_scheme));
+			auto cond = get_parents_condition(validator, r, pdbx_poly_seq_scheme);
+			if (not cond)
+			{
+				if (VERBOSE > 0 and std::exchange(warned_missing_parents, true) == false)
+					std::cerr << "warning: missing links for atom_site/pdbx_poly_seq_scheme\n";
+				continue;
+			}
+
+			auto p = pdbx_poly_seq_scheme.find(std::move(cond));
 			if (p.size() != 1)
 			{
-				if (cif::VERBOSE > 0)
+				if (VERBOSE > 0)
 					std::clog << "In atom_site record: " << r["id"].text() << '\n';
 				throw std::runtime_error("For each monomer in atom_site there should be exactly one pdbx_poly_seq_scheme record");
 			}
@@ -161,7 +167,7 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 
 			const auto entity_poly_type = entity_poly.find1<std::string>("entity_id"_key == entity_id, "type");
 
-			std::map<int,std::set<std::string>> mon_per_seq_id;
+			std::map<int, std::set<std::string>> mon_per_seq_id;
 
 			for (const auto &[num, mon_id, hetero] : entity_poly_seq.find<int, std::string, bool>("entity_id"_key == entity_id, "num", "mon_id", "hetero"))
 			{
@@ -196,28 +202,37 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 					throw std::runtime_error("Mismatch between the hetero flag in the poly seq schemes and the number residues per seq_id");
 			}
 
-			for (const auto &[seq_id, mon_ids] : mon_per_seq_id)
-			{
-				for (auto asym_id : struct_asym.find<std::string>("entity_id"_key == entity_id, "id"))
-				{
-					condition cond;
-					
-					for (auto mon_id : mon_ids)
-						cond = std::move(cond) or "label_comp_id"_key == mon_id;
+			// This code proved to take too much time ...
 
-					cond = "label_entity_id"_key == entity_id and
-						"label_asym_id"_key == asym_id and
-						"label_seq_id"_key == seq_id and not std::move(cond);
-					
-					if (atom_site.contains(std::move(cond)))
-						throw std::runtime_error("An atom_site record exists that has no parent in the poly seq scheme categories");
-				}
+			// for (const auto &[seq_id, mon_ids] : mon_per_seq_id)
+			// {
+			// 	for (auto asym_id : struct_asym.find<std::string>("entity_id"_key == entity_id, "id"))
+			// 	{
+			// 		condition cond;
+
+			// 		for (auto mon_id : mon_ids)
+			// 			cond = std::move(cond) or "label_comp_id"_key == mon_id;
+
+			// 		cond = "label_entity_id"_key == entity_id and
+			// 			"label_asym_id"_key == asym_id and
+			// 			"label_seq_id"_key == seq_id and not std::move(cond);
+
+			// 		if (atom_site.contains(std::move(cond)))
+			// 			throw std::runtime_error("An atom_site record exists that has no parent in the poly seq scheme categories");
+			// 	}
+			// }
+
+			// ... so we're using this instead, should be almost the same...
+
+			for (const auto &[comp_id, seq_id] :
+				atom_site.find<std::string, int>("label_entity_id"_key == entity_id, "label_comp_id", "label_seq_id"))
+			{
+				if (not mon_per_seq_id[seq_id].contains(comp_id))
+					throw std::runtime_error("An atom_site record exists that has no parent in the poly seq scheme categories");
 			}
 
 			auto &&[seq, seq_can] = entity_poly.find1<std::optional<std::string>, std::optional<std::string>>("entity_id"_key == entity_id,
 				"pdbx_seq_one_letter_code", "pdbx_seq_one_letter_code_can");
-			
-			std::string::const_iterator si, sci, se, sce;
 
 			auto seq_match = [&](bool can, std::string::const_iterator si, std::string::const_iterator se)
 			{
@@ -254,8 +269,8 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 							else
 								letter = '(' + comp_id + ')';
 						}
-						
-						if (iequals(std::string{si, si + letter.length()}, letter))
+
+						if (iequals(std::string{ si, si + letter.length() }, letter))
 						{
 							match = true;
 							si += letter.length();
@@ -274,12 +289,14 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 
 			if (not seq.has_value())
 			{
-				if (cif::VERBOSE > 0)
+				if (VERBOSE > 0)
 					std::clog << "Warning: entity_poly has no sequence for entity_id " << entity_id << '\n';
 			}
 			else
 			{
-				seq->erase(std::remove_if(seq->begin(), seq->end(), [](char ch) { return std::isspace(ch); }), seq->end());
+				seq->erase(std::remove_if(seq->begin(), seq->end(), [](char ch)
+							   { return std::isspace(ch); }),
+					seq->end());
 
 				if (not seq_match(false, seq->begin(), seq->end()))
 					throw std::runtime_error("Sequences do not match for entity " + entity_id);
@@ -287,12 +304,14 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 
 			if (not seq_can.has_value())
 			{
-				if (cif::VERBOSE > 1)
+				if (VERBOSE > 1)
 					std::clog << "Warning: entity_poly has no canonical sequence for entity_id " << entity_id << '\n';
 			}
 			else
 			{
-				seq_can->erase(std::remove_if(seq_can->begin(), seq_can->end(), [](char ch) { return std::isspace(ch); }), seq_can->end());
+				seq_can->erase(std::remove_if(seq_can->begin(), seq_can->end(), [](char ch)
+								   { return std::isspace(ch); }),
+					seq_can->end());
 
 				if (not seq_match(true, seq_can->begin(), seq_can->end()))
 					throw std::runtime_error("Canonical sequences do not match for entity " + entity_id);
@@ -304,16 +323,15 @@ bool is_valid_pdbx_file(const file &file, const validator &validator, std::error
 	catch (const std::exception &ex)
 	{
 		result = false;
-		if (cif::VERBOSE > 0)
+		if (VERBOSE > 0)
 			std::clog << ex.what() << '\n';
 		ec = make_error_code(validation_error::not_valid_pdbx);
 	}
 
-	if (not result and (bool)ec)
+	if (not result and (bool) ec)
 		ec = make_error_code(validation_error::not_valid_pdbx);
 
 	return result;
 }
 
 } // namespace cif::pdb
-  

src/symmetry.cpp

@@ -32,6 +32,11 @@
 
 #include "symop_table_data.hpp"
 
+#if defined(_MSC_VER)
+#pragma warning (disable : 5054)	// warning C5054: operator '&': deprecated between enumerations of different types
+#pragma warning (disable : 4127)	// conditional expression is constant
+#endif
+
 #include <Eigen/Eigen>
 
 namespace cif
@@ -90,10 +95,10 @@ float cell::get_volume() const
 	auto cos_beta = std::cos(beta);
 	auto cos_gamma = std::cos(gamma);
 
-	auto vol = m_a * m_b * m_c;
+	double vol = m_a * m_b * m_c;
 	vol *= std::sqrt(1.0f - cos_alpha * cos_alpha - cos_beta * cos_beta - cos_gamma * cos_gamma + 2.0f * cos_alpha * cos_beta * cos_gamma);
 
-	return vol;
+	return static_cast<float>(vol);
 }
 
 // --------------------------------------------------------------------

src/validate.cpp

@@ -36,16 +36,11 @@
 
 // The validator depends on regular expressions. Unfortunately,
 // the implementation of std::regex in g++ is buggy and crashes
-// on reading the pdbx dictionary. Therefore, in case g++ is used
-// the code will use boost::regex instead.
+// on reading the pdbx dictionary. We used to use boost regex
+// instead but using pcre2 is even easier and faster.
 
-#if USE_BOOST_REGEX
-# include <boost/regex.hpp>
-using boost::regex;
-#else
-# include <regex>
-using std::regex;
-#endif
+#define PCRE2_CODE_UNIT_WIDTH 8
+#include <pcre2.h>
 
 namespace cif
 {
@@ -67,14 +62,58 @@ validation_exception::validation_exception(std::error_code ec, std::string_view
 
 // --------------------------------------------------------------------
 
-struct regex_impl : public regex
+struct regex_impl
 {
-	regex_impl(std::string_view rx)
-		: regex(rx.begin(), rx.end(), regex::extended | regex::optimize)
-	{
-	}
+	regex_impl(std::string_view rx);
+	~regex_impl();
+
+	regex_impl(const regex_impl &) = delete;
+	regex_impl &operator=(const regex_impl &) = delete;
+
+	bool match(std::string_view v) const;
+
+  private:
+	pcre2_code *m_rx = nullptr;
+	pcre2_match_data *m_data = nullptr;
 };
 
+regex_impl::regex_impl(std::string_view rx)
+{
+	int err_code;
+	size_t err_offset;
+	m_rx = pcre2_compile((PCRE2_SPTR)rx.data(), rx.length(), 0, &err_code, &err_offset, nullptr);
+	if (m_rx == nullptr)
+	{
+		PCRE2_UCHAR buffer[256];
+		int n = pcre2_get_error_message(err_code, buffer, sizeof(buffer));
+
+		throw std::runtime_error(std::string("PCRE2 compilation failed: ") + std::string{ (char *)buffer, (char *)buffer + n });
+	}
+
+	m_data = pcre2_match_data_create_from_pattern(m_rx, nullptr);
+}
+
+regex_impl::~regex_impl()
+{
+	if (m_data)
+		pcre2_match_data_free(m_data);
+
+	if (m_rx)
+		pcre2_code_free(m_rx);
+}
+
+bool regex_impl::match(std::string_view v) const
+{
+	bool result = false;
+
+	if (int rc = pcre2_match(m_rx, (PCRE2_SPTR)v.data(), v.length(), 0, 0, m_data, nullptr); rc >= 0)
+		result = true;
+	else if (rc != PCRE2_ERROR_NOMATCH)
+		std::cerr << "Error matching with pcre\n";
+
+	return result;
+}
+
 // --------------------------------------------------------------------
 
 DDL_PrimitiveType map_to_primitive_type(std::string_view s, std::error_code &ec) noexcept
@@ -233,7 +272,7 @@ bool item_validator::validate_value(std::string_view value, std::error_code &ec)
 
 	if (not value.empty() and value != "?" and value != ".")
 	{
-		if (m_type != nullptr and not regex_match(value.begin(), value.end(), *m_type->m_rx))
+		if (m_type != nullptr and not m_type->m_rx->match(value))
 			ec = make_error_code(validation_error::value_does_not_match_rx);
 		else if (not m_enums.empty() and m_enums.count(std::string{ value }) == 0)
 			ec = make_error_code(validation_error::value_is_not_in_enumeration_list);

test/CMakeLists.txt

@@ -22,12 +22,12 @@ list(
 	CIFPP_tests
 	unit-v2
 	unit-3d
-	format
 	model
+	query
 	rename-compound
 	sugar
 	spinner
-	# reconstruction
+	reconstruction
 	validate-pdbx
 	)
 

test/query-test.cpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- *
- * Copyright (c) 2020 NKI/AVL, Netherlands Cancer Institute
- *
+ * 
+ * Copyright (c) 2025 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
@@ -24,38 +24,31 @@
  * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
-
 #include "test-main.hpp"
 
-#include <stdexcept>
-
 #include <cif++.hpp>
 
-// --------------------------------------------------------------------
+#include <iostream>
+#include <fstream>
 
-TEST_CASE("fmt_1")
+TEST_CASE("q-1")
 {
-	std::ostringstream os;
+	using namespace cif::literals;
 
-	std::string world("world");
-	os << cif::format("Hello, %-10.10s, the magic number is %d and pi is %g", world, 42, cif::kPI);
-	REQUIRE(os.str() == "Hello, world     , the magic number is 42 and pi is 3.14159");
+	cif::compound_factory::instance().push_dictionary(gTestDir / "REA.cif");
+
+	cif::file a = cif::pdb::read(gTestDir / "pdb1cbs.ent.gz");
+	auto &pdbx_poly_seq_scheme = a.front()["pdbx_poly_seq_scheme"];
+	REQUIRE_FALSE(pdbx_poly_seq_scheme.empty());
+
+	SECTION("s-11")
+	{
+		CHECK(pdbx_poly_seq_scheme.count("asym_id"_key == "A") == 137);
+
+		CHECK(pdbx_poly_seq_scheme.count("asym_id"_key == "A" and "entity_id"_key == 1 and "seq_id"_key == 1 and "mon_id"_key == "PRO") == 1);
+
+		CHECK(pdbx_poly_seq_scheme.count("asym_id"_key == "A" and "entity_id"_key == 1 and "seq_id"_key == 1 and "mon_id"_key == "PRO" and "hetero"_key == false) == 1);
+	}
+	
 
-	REQUIRE(cif::format("Hello, %-10.10s, the magic number is %d and pi is %g", world, 42, cif::kPI).str() ==
-		"Hello, world     , the magic number is 42 and pi is 3.14159");
 }
-
-// --------------------------------------------------------------------
-
-TEST_CASE("clr_1")
-{
-	using namespace cif::colour;
-
-	std::cout << "Hello, " << cif::coloured("world!", white, red, cif::colour::regular) << '\n'
-			  << "Hello, " << cif::coloured("world!", white, red, bold) << '\n'
-			  << "Hello, " << cif::coloured("world!", black, red) << '\n'
-			  << "Hello, " << cif::coloured("world!", white, green) << '\n'
-			  << "Hello, " << cif::coloured("world!", white, blue) << '\n'
-			  << "Hello, " << cif::coloured("world!", blue, white) << '\n'
-			  << "Hello, " << cif::coloured("world!", red, white, bold) << '\n';
-}

test/reconstruction-test.cpp

@@ -28,6 +28,7 @@
 
 #include <cif++.hpp>
 
+#include <filesystem>
 #include <iostream>
 #include <fstream>
 

test/test-main.hpp

@@ -1,17 +1,17 @@
 /*-
  * SPDX-License-Identifier: BSD-2-Clause
- * 
+ *
  * Copyright (c) 2024 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

test/unit-3d-test.cpp

@@ -30,6 +30,11 @@
 
 #include <cif++.hpp>
 
+#if defined(_MSC_VER)
+#pragma warning (disable : 5054)	// warning C5054: operator '&': deprecated between enumerations of different types
+#pragma warning (disable : 4127)	// conditional expression is constant
+#endif
+
 #include <Eigen/Eigenvalues>
 
 // --------------------------------------------------------------------
@@ -296,7 +301,7 @@ TEST_CASE("m2q_0a")
 		auto d = cif::kSymopNrTable[i].symop().data();
 
 		Eigen::Matrix3f rot;
-		rot << d[0], d[1], d[2], d[3], d[4], d[5], d[6], d[7], d[8];
+		rot << static_cast<float>(d[0]), static_cast<float>(d[1]), static_cast<float>(d[2]), static_cast<float>(d[3]), static_cast<float>(d[4]), static_cast<float>(d[5]), static_cast<float>(d[6]), static_cast<float>(d[7]), static_cast<float>(d[8]);
 
 		// check to see if this matrix contains a true rotation
 		if (rot * rot.transpose() != Eigen::Matrix3f::Identity() or rot.determinant() != 1)
@@ -435,11 +440,11 @@ TEST_CASE("symm_4")
 
 	// based on 2b8h
 	auto sg = cif::spacegroup(154); // p 32 2 1
-	auto c = cif::cell(107.516, 107.516, 338.487, 90.00, 90.00, 120.00);
+	auto c = cif::cell(107.516f, 107.516f, 338.487f, 90.00f, 90.00f, 120.00f);
 
-	cif::point a{ -8.688, 79.351, 10.439 };  // O6 NAG A 500
-	cif::point b{ -35.356, 33.693, -3.236 }; // CG2 THR D 400
-	cif::point sb(-6.916, 79.34, 3.236);     // 4_565 copy of b
+	cif::point a{ -8.688f, 79.351f, 10.439f };  // O6 NAG A 500
+	cif::point b{ -35.356f, 33.693f, -3.236f }; // CG2 THR D 400
+	cif::point sb(-6.916f, 79.34f, 3.236f);     // 4_565 copy of b
 
 	CHECK_THAT(distance(a, sg(a, c, "1_455"_symop)), Catch::Matchers::WithinRel(static_cast<float>(c.get_a()), 0.01f));
 	CHECK_THAT(distance(a, sg(a, c, "1_545"_symop)), Catch::Matchers::WithinRel(static_cast<float>(c.get_b()), 0.01f));
@@ -466,7 +471,7 @@ TEST_CASE("symm_4wvp_1")
 
 	cif::crystal c(db);
 
-	cif::point p{ -78.722, 98.528, 11.994 };
+	cif::point p{ -78.722f, 98.528f, 11.994f };
 	auto a = s.get_residue("A", 10, "").get_atom_by_atom_id("O");
 
 	auto sp1 = c.symmetry_copy(a.get_location(), "2_565"_symop);

tools/depends.cmd

@@ -5,6 +5,8 @@ IF NOT EXIST build_ci\libs (
   MKDIR build_ci\libs
 )
 CD build_ci\libs
+
+@REM Install ZLib
 IF NOT EXIST zlib-%ZLIB_VERSION%.zip (
   ECHO Downloading https://github.com/libarchive/zlib/archive/v%ZLIB_VERSION%.zip
   curl -L -o zlib-%ZLIB_VERSION%.zip https://github.com/libarchive/zlib/archive/v%ZLIB_VERSION%.zip || EXIT /b 1
@@ -14,9 +16,9 @@ IF NOT EXIST zlib-%ZLIB_VERSION% (
   C:\windows\system32\tar.exe -x -f zlib-%ZLIB_VERSION%.zip || EXIT /b 1
 )
 CD zlib-%ZLIB_VERSION%
-cmake -G "Visual Studio 17 2022" . || EXIT /b 1
-cmake --build . --target ALL_BUILD --config Release || EXIT /b 1
-cmake --build . --target RUN_TESTS --config Release || EXIT /b 1
-cmake --build . --target INSTALL --config Release || EXIT /b 1
+cmake -B build || EXIT /b 1
+cmake --build build --target ALL_BUILD --config Release || EXIT /b 1
+cmake --build build --target RUN_TESTS --config Release || EXIT /b 1
+cmake --build build --target INSTALL --config Release || EXIT /b 1
 
 @EXIT /b 0