Export of internal Abseil changes

--
9214ee309fa224a43eb1621bcf826ab9c86a4cd7 by Abseil Team <absl-team@google.com>:

Don't keep constructing/destructing `random_device`.

It causes an open/close pair on /dev/urandom.

This is somewhat expensive on Emscripten, and unnecessary since it is not using
the value anyway.

PiperOrigin-RevId: 332055452

--
4931ab77995b946f3ffba8036e0a00927b0d47e6 by Abseil Team <absl-team@google.com>:

TYPO: Stess->Stress

PiperOrigin-RevId: 332010526

--
95ae1dde505f74dc7fd68447bf27718db223688e by Gennadiy Rozental <rogeeff@google.com>:

Import of CCTZ from GitHub.

PiperOrigin-RevId: 332010411
GitOrigin-RevId: 9214ee309fa224a43eb1621bcf826ab9c86a4cd7
Change-Id: I23c5d6aa5aa488370e1b48f37366d3d99448e7ec

absl/container/internal/hash_generator_testing.cc

@@ -41,8 +41,10 @@ class RandomDeviceSeedSeq {
 }  // namespace
 
 std::mt19937_64* GetSharedRng() {
-  RandomDeviceSeedSeq seed_seq;
-  static auto* rng = new std::mt19937_64(seed_seq);
+  static auto* rng = [] {
+    RandomDeviceSeedSeq seed_seq;
+    return new std::mt19937_64(seed_seq);
+  }();
   return rng;
 }
 

absl/synchronization/mutex_test.cc

@@ -1102,7 +1102,7 @@ TEST(Mutex, DeadlockDetectorBazelWarning) {
 // annotation-based static thread-safety analysis is not currently
 // predicate-aware and cannot tell if the two for-loops that acquire and
 // release the locks have the same predicates.
-TEST(Mutex, DeadlockDetectorStessTest) ABSL_NO_THREAD_SAFETY_ANALYSIS {
+TEST(Mutex, DeadlockDetectorStressTest) ABSL_NO_THREAD_SAFETY_ANALYSIS {
   // Stress test: Here we create a large number of locks and use all of them.
   // If a deadlock detector keeps a full graph of lock acquisition order,
   // it will likely be too slow for this test to pass.

absl/time/internal/cctz/BUILD.bazel

@@ -92,6 +92,11 @@ cc_library(
 
 ### tests
 
+test_suite(
+    name = "all_tests",
+    visibility = ["//visibility:public"],
+)
+
 cc_test(
     name = "civil_time_test",
     size = "small",