PiperOrigin-RevId: 533121605
Change-Id: I82390045985db5db56e93c40eae24cdcd93e3ab4

libstdc++'s definition of std::launder places it behind a check for
__builtin_launder, which is not available before clang 8.

Fixes: #1309

These have been marked deprecated in the comments for some time,
(since f340f773) but the warnings
were never enabled.

A warning suppression is enabled for Abseil code so that when we
declare types in our code, we don't get a warning.

PiperOrigin-RevId: 532891102
Change-Id: Ife0c5696a061ea44769e02869e4e3d1196e86f9d

PiperOrigin-RevId: 532553508
Change-Id: I813841ff3e5085b64c9b02ca41897bf7f6a8570e

PiperOrigin-RevId: 532199109
Change-Id: I5a2a96b9c8b52feb0107d912b0011911fc1f0f4c

absl::RegisterSymbolizer() has been deprecated for 5 years.
It is being removed following our compatibility policy.
<https://abseil.io/about/compatibility>
PiperOrigin-RevId: 532174866
Change-Id: Id5c3b86698e389099d3d707c4e57f30f1f155d2e

PiperOrigin-RevId: 532174747
Change-Id: I39dc6d5c369e8bfbdde5c826836e42f5baa69dc5

This is a heavily modified version of
https://github.com/abseil/abseil-cpp/pull/1445,
which adds some missing test libraries to the test DLL.

Unlike #1445, this change moves several global variables out of
headers that did not need to be in headers.

For instance, cord_btree_exhaustive_validation was a global
defined/declared in cord_internal, but only used in cord_rep_btree
and its test.

cordz_handle defined a queue in its header even though it wasn't needed,
which also led to ODR problems.

The Spinlock used in CordzHandle is replaced with a Mutex. This was
originally a Mutex, but Chromium asked us to change it to a Spinlock
to avoid a static initializer. After this change, the static
initializer is no longer an issue.

#1407

PiperOrigin-RevId: 531516991
Change-Id: I0e431a193698b20ba03fac6e414c26f153f330a7

PiperOrigin-RevId: 531301442
Change-Id: Ib5a67f827bf89a66a0a8d24d3191f278566af6c8

PiperOrigin-RevId: 530999471
Change-Id: Ifca68cd53be629088244eaa3560d43fd81090747

will just be a copy anyway. Also remove the typedef so that it is
clear from the type at the callsite that moving it isn't necessary.

Fixes #1443

PiperOrigin-RevId: 530596294
Change-Id: I58ffc370bbccc0816bca4c4f85c3bb12c7ee2eb2

PiperOrigin-RevId: 530445615
Change-Id: I0c112773e7f42989840cb8b915167ade8833ef73

on timing as flaky.

This will run them up to 3 times (in the default configuration) and only
consider the test failed if it fails each time.

PiperOrigin-RevId: 530348477
Change-Id: Ib77bd8a166828b82a1b49ee5ef2e8321752e51d0

builds.

PiperOrigin-RevId: 530308301
Change-Id: I7c885156e14d9c8c444c7c6f8d38e7aaac7c01a7

PiperOrigin-RevId: 530238518
Change-Id: I2bfca582c0734f7e6943c5359730a2857809e2d2

Make `absl::StatusOr::AssignStatus` public. Recommend against using it outside of generic programming (attempt #2 after internal fix)

PiperOrigin-RevId: 529796927
Change-Id: I755b7d907f96f4a05d01620503bf0862ce35e847

PiperOrigin-RevId: 529514670
Change-Id: Ia9a19da52f58c91d5131c2a17308a7a322db2714

Make `absl::StatusOr::AssignStatus` public. Recommend against using it outside of generic programming

PiperOrigin-RevId: 529484669
Change-Id: I9889a9f7a7fa7dc6ef8c731f1b583a74e6afd2cf

Also do some minor refactoring in btree.h.

PiperOrigin-RevId: 529460378
Change-Id: I278833ada93bbb7652e149fceed08ce3485e4312

PiperOrigin-RevId: 529423722
Change-Id: Ib1cc427299100956c0f2ae6cb5214467ef5a3790

PiperOrigin-RevId: 529119690
Change-Id: If585274c409e2a344c8d60759da6f8f990023d29

PiperOrigin-RevId: 529086715
Change-Id: Id5462322d5707a1b4475438520fcb70ee7b15450

When we insert a new element, when ASan is enabled, we replace the node that the new element is on in order to try to detect cases of code depending on pointer/reference-stability.

PiperOrigin-RevId: 528826645
Change-Id: Ie5c15c13016a8aa831a0d1edc3ad33c1f5ca4d65

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1433

Some 32-bit configurations may use 64-bit time_t, which leads to different layout of userspace timespec and the one expected by SYS_futex implementation in kernel. In particular the issue occurs when using musl libc which has switched to unconditional 64-bit time_t definition.

This patch introduces custom struct timespec with two longs when old SYS_futex is used to match the kernel timespec definition.
Merge 2eaca415da825b3f31a90f58a35bdef2b6d2a6c5 into f8bf9091

Merging this change closes #1433

COPYBARA_INTEGRATE_REVIEW=https://github.com/abseil/abseil-cpp/pull/1433 from olegartys:futex_time64_bug 2eaca415da825b3f31a90f58a35bdef2b6d2a6c5
PiperOrigin-RevId: 528796119
Change-Id: Idaa952f64bd97c6dc9703a8b44deac43e29ff9ae

on timing as flaky.

This will run them up to 3 times (in the default configuration) and only
consider the test failed if it fails each time.

PiperOrigin-RevId: 528785128
Change-Id: I4aa5b74aa3a6873c83f2a01734c679351b1cd5d0

Instantiate LogMessage::CopyToEncodedBuffer separately for literals and non-literals to save an instruction at each operator<<(string) callsite.

The operator<<s get inlined at the callsite.  The saved instruction is a 'load immediate' to put the StringType argument in a register for passing.

This does make the logging TU slightly larger, but it will be an overall size win for all but the smallest programs.  Smaller, faster code at callsites is worth it anyway.

PiperOrigin-RevId: 528569773
Change-Id: Id8185266a975a12ab6f1a553c2c0846261822396

PiperOrigin-RevId: 528547013
Change-Id: I9b57ee062ef666d6a34696778ff16a46c5ccb17d

Shrink most LOG callsites by one instruction (SysV x86_64 ABI) by dispatching to per-severity constructors and omitting the severity argument.

In that particular ABI, the empty tag structs are NO_CLASS and don't appear in the generated code.

https://godbolt.org/z/crqaPh8Kv

PiperOrigin-RevId: 527989179
Change-Id: I365f11f6ca1de2623690d0a04e6f02f7de82b1fc

These functions are meant to allow for avoiding complex SFINAE pre-C++17.

PiperOrigin-RevId: 527968428
Change-Id: I6524206c9dc50663d7c38abf6d0e8080ca93fc30

PiperOrigin-RevId: 527598377
Change-Id: I0ae1505db595b7f3a420f2f608c6e02f8800f138

PiperOrigin-RevId: 527066823
Change-Id: Ifa1e9a43c7490b34f9f4dbfa12d3acbed6b49777

PiperOrigin-RevId: 526675031
Change-Id: Ib84423ccea2d0183166194a0916a97a7ed32915c

PiperOrigin-RevId: 526653332
Change-Id: I0a20d4ac636da3f1a930f96e0cdb9275527e4688

PiperOrigin-RevId: 526644558
Change-Id: I050162c2c0d667c32f7e11f51201054cca52667d

Currently static analysis tools built using the Clang Dataflow Analysis framework can't prove that code under `switch (0) case 0:` is executed on all paths. The Clang Dataflow Analysis framework should ultimately be improved to handle these cases. In the meantime, to enable the use of such tools in their current state, we add a `default` case back to the `switch` statement in `ABSL_LOG_INTERNAL_STATELESS_CONDITION` to help them understand that the code is executed on all paths.

PiperOrigin-RevId: 526638852
Change-Id: I49490dd477ba777aae2530697b75b583242aebc5

`AbslHashValue()` applied to a C-style array simply decays the array to a pointer and hashes that, which is very unlikely to be what the author intended.

PiperOrigin-RevId: 525923741
Change-Id: I876dbde341f0b0642fa3d10e3f67b0b31d0311bf