This change adds a new macro, ABSL_HARDENING_ASSERT_SLOW, which can be used for hardening assertions which have too great a performance impact to enable widely.  It also adds an additional supported value for ABSL_OPTION_HARDENED to only enable assertions which are not slow.  ABSL_OPTION_HARDENED=1 will continue to enable all hardening assertions, while ABSL_OPTION_HARDENED=2 will only enable assertions not marked as SLOW.

PiperOrigin-RevId: 668976425
Change-Id: I12cb209715d5b359c4093cc8c62e16794d7729b3

used.

PiperOrigin-RevId: 668612700
Change-Id: I018ce756632154a9e7f2de77e6faa24296d80002

We enable this validation when sanitizers are enabled rather than in debug mode because debug mode can be enabled/disabled separately in different translation units, which can cause bugs for this type of validation. E.g. a hashtable that was moved from in a TU with debug enabled is then destroyed in a TU with debug disabled.

PiperOrigin-RevId: 668492783
Change-Id: Ifab6894de0890aa0a7d8ea338b53b6759b097cb1

indicate the default nullability.
This is a no-op for now (not understood by checkers), but does communicate
intention to human readers.

PiperOrigin-RevId: 668448175
Change-Id: I19802d19be9a310aa5f0dacf8b0567dd768fb830

This change adds support for unwinding stack traces on supported android platforms (NDK 33+) to absl::GetStackTrace().

PiperOrigin-RevId: 668041186
Change-Id: Ie4c4c123686050c6b81683b71fe0246e8fa75f2b

PiperOrigin-RevId: 667621254
Change-Id: Ibff25d275c1c6846f857d90ed905c70f9545d8dd

PiperOrigin-RevId: 666760220
Change-Id: I9e85d4db0442546bdebc13b0a6e6d9c7f8d3c459

PiperOrigin-RevId: 666465929
Change-Id: Idcbed10e4e787f625f45460cd3651a37d06845bc

Migrate ABSL_INTERNAL_ATTRIBUTE_OWNER and ABSL_INTERNAL_ATTRIBUTE_VIEW to ABSL_ATTRIBUTE_OWNER and ABSL_ATTRIBUTE_VIEW

PiperOrigin-RevId: 666390885
Change-Id: I85457d39cf5ecb1c6622c6d59cb1876b154d64c3

due to false positives.

https://godbolt.org/z/evjd73cv6

PiperOrigin-RevId: 666027404
Change-Id: Ib0fe9c7b3d2aeb8710342e9e2aadc6a346e03e60

Deprecates ABSL_INTERNAL_ATTRIBUTE_VIEW and ABSL_INTERNAL_ATTRIBUTE_OWNER

PiperOrigin-RevId: 665523780
Change-Id: Ic4f7e3083a7db9d8fd057d3e002b81c775a4a87e

PiperOrigin-RevId: 664999751
Change-Id: I6f908cc1b8de8a6a11bb1e02cca761df6aae4e07

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1728

When targeting Android NDK r25 (reached EOL now, but still in some use) one gets the following error:
```
$(SOURCE_ROOT)/contrib/restricted/abseil-cpp/absl/time/time.h:1762:37: error: invalid operands to binary expression ('std::strong_ordering' and 'const std::strong_ordering')
  if (auto c = lhs_hi <=> rhs_hi; c != std::strong_ordering::equal) {
```

The error indicates the lack of `operator<=>` between two `std::strong_ordering` items.
I believe that this should be controlled by `__cpp_lib_three_way_comparison` (_Three-way comparison (library support)_, see [this article](https://en.cppreference.com/w/cpp/feature_test)), not by `__cpp_impl_three_way_comparison` (which stands for _Three-way comparison (compiler support)_).

Fixes #1725.
The problem was introduced in a2766235.

Merge 32e1d3d889859ef8a32bf3460349814a2203444c into 69c46839

Merging this change closes #1728

COPYBARA_INTEGRATE_REVIEW=https://github.com/abseil/abseil-cpp/pull/1728 from georgthegreat:fix-time 32e1d3d889859ef8a32bf3460349814a2203444c
PiperOrigin-RevId: 663723253
Change-Id: I29f7d2b562cc5ad8e11cd46b538ba69acee0f314

PiperOrigin-RevId: 663373500
Change-Id: I6b2014f8cc4244e1599c907197cfaadb6739f72b

This fixes the behavior of
```c++
ABSL_RAW_LOG(INFO, "RAW INFO: %s%c%s", "Hello", 0, "World");
```
which would previously truncate at the `\0`. The new behavior is consistent with `printf`.

PiperOrigin-RevId: 663049889
Change-Id: I171dcb8a61b19873b88920e383f03acf7fb112d7

PiperOrigin-RevId: 662635356
Change-Id: I8b3d992147ee3affb92947135f09076c851448e3

PiperOrigin-RevId: 662586240
Change-Id: Ibee60b935da5d8135ac99b6caf00b3c035ecfd89

The motivation for renaming AssertValidCapacity() is to avoid confusion with IsValidCapacity().

PiperOrigin-RevId: 662178137
Change-Id: If4fd15725376a899c4f40afc3db9e300c8c1edc7

Add debug mode checks that element constructors/destructors don't make reentrant calls to raw_hash_set member functions.

PiperOrigin-RevId: 660889825
Change-Id: I02e0e364a5215431eddeeabde66531a95aa03f22

PiperOrigin-RevId: 660539039
Change-Id: I2f3dee7c60ce2234ea81c64b23e50448681fa885

The GetTID() now uses the value of the current thread handle from
zx_thread_self() on Fuchsia for the thread ID. This is better than the default
implementation which uses pthread_self(), since that results in a 64-bit
pointer being truncated into the 32-bit pid_t type.

PiperOrigin-RevId: 660270084
Change-Id: I7c8b4a26cca775be62eb6f6926cb734cbb893585

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1738

Since #1707 and #1710, `-Xarch_` option groups are present as single elements in the libraries' COPTS, and a `SHELL:` prefix is added to each of them. While these addressed certain build issues, they broke the handling of `-Xarch_` option groups in pkgconfig files.

Fix that by taking care of the `SHELL:` prefix in COPTS when generating pkgconfig files. The skip-next-flag mechanism is also removed as the option groups are now present as single elements.

Merge fd1246acbf4052d66fef66ee195cf254bca35b6c into 08850701

Merging this change closes #1738

COPYBARA_INTEGRATE_REVIEW=https://github.com/abseil/abseil-cpp/pull/1738 from ZhongRuoyu:pkgconfig-xarch-handling fd1246acbf4052d66fef66ee195cf254bca35b6c
PiperOrigin-RevId: 660055129
Change-Id: I57a1a51b2eb8983d076a3d5336acef12acae4560

PiperOrigin-RevId: 659987892
Change-Id: I8cb64c7569abe932851ab52d68e3b44c0c008ece

An `operator->` that doesn't return a record type doesn't really serve a useful
purpose, and it is not required by the iterator concepts. I haven't been able
to find any usage in non-test code, so I'm removing it.

Even though the `pointer` type alias is now not being used anywhere, I'm keeping
it because otherwise some standard libraries don't accept `CharIterator` when
passed to `std::string::apend()`, for example.

PiperOrigin-RevId: 659832167
Change-Id: I5de183a01009bfd71c8477f2023973075be8d860

This change marks Abseil's non-modifying sequence operations including
absl::linear_search and absl::c_linear_search as constexpr when building with
C++20.
PiperOrigin-RevId: 659812405
Change-Id: I8dc2cee873f30531b2eb8fb3da12085505a43a1a

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1739

GCC 15 will no longer include <cstdint> by default, resulting in build failures in projects that do not explicitly include it.

Merge faf1b03a591f06933da02976119da5743f428e4f into 9cb5e5d1

Merging this change closes #1739

COPYBARA_INTEGRATE_REVIEW=https://github.com/abseil/abseil-cpp/pull/1739 from csfore:gcc-15-fix faf1b03a591f06933da02976119da5743f428e4f
PiperOrigin-RevId: 659637669
Change-Id: If14cb0e3522774cb700bd5a74abffb75feb7a0f5

warnings in headers". It should fall through to the "don't impose our
warnings on others" case. Do this by matching on "-Wno-*" instead of
"-Wno*".

Fixes #1737

PiperOrigin-RevId: 659548798
Change-Id: I49d7ba7ddcd7be30f946fca90ba9be467181e854

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1732

Fix build on NVIDIA Jetson board. Fix #1665

This patch is already used by the spark project.
I'm fixing this as this break the build of Tensorflow and JAX on Jetson board.
Merge 7db2d2ab9fbed1f0fabad10a6ec73533ba71bfff into 6b8ebb35

Merging this change closes #1732

COPYBARA_INTEGRATE_REVIEW=https://github.com/abseil/abseil-cpp/pull/1732 from nouiz:fix_neon_on_jetson 7db2d2ab9fbed1f0fabad10a6ec73533ba71bfff
PiperOrigin-RevId: 658501520
Change-Id: If502ede4efc8c877fb3fed227eca6dc7622dd181

This updates to a version of GoogleTest that fixes
https://github.com/google/googletest/issues/4581
for downstream users of rules_python

PiperOrigin-RevId: 658398482
Change-Id: I4c6d431fd38d3a108e3c252eb26be9d81ff05a8e

PiperOrigin-RevId: 657719871
Change-Id: If0f16b1d8cfae79f31427a804a5b838599154895

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1735

If asan is enabled with VS, the following warning is issued by the compiler:

1>C:\work\abseil-cpp\absl\debugging\leak_check.cc(46,68): error C4800: Implicit conversion from 'int' to bool. Possible information loss
1>    C:\work\abseil-cpp\absl\debugging\leak_check.cc(46,68):
1>    consider using explicit cast or comparison to 0 to avoid this warning
Merge 9a6a677add58a9cd31b2381f6bdb5328c9b03c02 into 8cdf4821

Merging this change closes #1735

COPYBARA_INTEGRATE_REVIEW=https://github.com/abseil/abseil-cpp/pull/1735 from pps83:master-int-bool-warn 9a6a677add58a9cd31b2381f6bdb5328c9b03c02
PiperOrigin-RevId: 657591557
Change-Id: I709424b5064160860186a42f854a1856465e80d0

We already have `swap(raw_hash_set&, raw_hash_set&)` thus we may expect argument dependent lookup will use that function in following code snippet:
```
absl::flat_hash_map<...> x, y;
using std::swap;
swap(x, y);
```
But in practice `std::swap` will be called because `swap(raw_hash_set&, raw_hash_set&)` requires derived-to-base conversion while `std::swap` doesn't (it is a function template), thus `std::swap` is picked by compiler.

To avoid this, we need a `swap` implementations which accept the exact `*_hash_*` types which is more preferred over `std::swap`.

Note that this will fix issues#1571.

PiperOrigin-RevId: 657419420
Change-Id: Id003b2129187ce2d4583029789ff96247d1490de

PiperOrigin-RevId: 657352849
Change-Id: I443f218b5500492948cd4c974313209907916355

https://github.com/llvm/llvm-project/issues/100639

The second issue is still relevant.

PiperOrigin-RevId: 657298935
Change-Id: Ifa2f1627deee3fc24e23ea78c5b01ed232813390

PiperOrigin-RevId: 656439732
Change-Id: I7f405d8075c57e634d51c11a37fbe3e317abf103

PiperOrigin-RevId: 656188416
Change-Id: I43d901d99f8a4688406b8aab08bc2655a065af46

Many compiler optimizations, even some in Asan, assume that size of
the object on the stack is fixed. Compiler takes into account the lifetime
of the stack object, but it does not expect that size of the object can
change.

This assumption allow to accesses when ever object is alive.

However, this assumption does not work well with custom poisoning
of objects by sanitizers. `Cord` may poison parts of the object
to improve a precision of the bug detection. But compiler still may
load from poisoned parts of the object, assuming it's alive.

So without the patch we can have `false-positive`, because
compiler reordered load from poisoned local before size-check.
https://github.com/llvm/llvm-project/issues/100639

Alternative is to make compiler optimization to know about
possibility of local object resize, but it's not worth of effort,
as it's needed only for sanitizers so far.

Another minor issue fixed with this patch is `false-negative`.
https://github.com/llvm/llvm-project/issues/100640
Asan drops instrumentation of memory accesses proven to be inbound
of local variable. However if we if we take into account `custom
poisoning` we will have to pessimize too often: any random function
call can potentially `resize` a variable by `poisoning`.

We are already using similar `poisoning` workaround in libc++:
https://github.com/llvm/llvm-project/pull/79536/files#diff-534bc2907ddb3b074ded1353d18fd7d578daf1707943b3039bab4ed975aba3b3R772

PiperOrigin-RevId: 656129711
Change-Id: I6d78997da6d31c7ab979a00b84dc9b3b7cffc26f

The POSIX sigaltstack() function is not implemented on Fuchsia.

PiperOrigin-RevId: 656080882
Change-Id: Ibf4892df85f299b888eb7f0016c220f49f975b96

PiperOrigin-RevId: 656010740
Change-Id: I99ef6f18502a0cb2c363c805ae5f9a9a42bd3b41

PiperOrigin-RevId: 655958414
Change-Id: I09b3a86b009cd64e53d27b615c0ec127ed64e8e4