This is a simplification but not a behavior change; we used to choose %u for unsigned ints, but %u and %d generate the same output for these types.

PiperOrigin-RevId: 539104599
Change-Id: I9d7ff561b969a6287889f95063636d6b77a4a78b

PiperOrigin-RevId: 539104398
Change-Id: I6385662e0c5694d40c57887f983b16adc9eced14

This patch makes it use `memcpy` for performance and consistency with other overloads.

PiperOrigin-RevId: 539079130
Change-Id: I5aea9dd9b8a1ce708c787df7d6c9a75ae419c484

in NACL pthread_self() returns a pointer that isn't directly convertible to arithmetic type.

PiperOrigin-RevId: 539023020
Change-Id: I3745ec5565f3a99ccb1d9df12c27a80e57ca4755

PiperOrigin-RevId: 538778978
Change-Id: I0c04b09987eb4d97892825e3c994656d4e300f60

PiperOrigin-RevId: 538651091
Change-Id: Id1aa7ab287beb252f4826f70e26b7e3bfa80c462

PiperOrigin-RevId: 538534137
Change-Id: Ifb2e59c36ffce183d3c93081d9300201757f82d4

Fixes #1469

PiperOrigin-RevId: 538534014
Change-Id: Iab84f73c0a0665355b614e85c62f686dc54c8671

PiperOrigin-RevId: 538488164
Change-Id: I680dcbc2f27dca3beba416b3cd0a55943751cf4c

Update XML output to properly print special characters. Non printable characters are skipped in output. All the "whitespace-like" (including \t, \r, \n) characters printed as plain space ' '.

PiperOrigin-RevId: 538479923
Change-Id: I62dff39bb21f376d00a7b9480f8f1d31d7015e45

PiperOrigin-RevId: 538241594
Change-Id: Ie6f0d913bcf07dea2f33e47198ba952b3800d70e

This is only a workaround, working only with gcc/clang, while waiting
for a better long term fix not using std::aligned_storage, which is
tracked by b/260219225.

Add a unit test that captures the current behavior of formatting of char types and char-backed enum types through StrCat(), StrFormat("%v"), and Substitute().

This test allows us to modify the behavior in this space without introducing undesired changes elsewhere.

PiperOrigin-RevId: 537981963
Change-Id: Icda91b66efcc0dc8c263011b137e130a3db2dc19

PiperOrigin-RevId: 537825067
Change-Id: I3ce8712d5130068fb7d77b563eb502e2e9560810

PiperOrigin-RevId: 537400816
Change-Id: I06794a6b8b6a441e34121047024380190d42869a

PiperOrigin-RevId: 537372070
Change-Id: I46ff6e42856aea2cd8da6ff7105cf58613603dd4

PiperOrigin-RevId: 537120102
Change-Id: I7952e53aca10319eb433e4c4d60cf3d7fe74d19a

PiperOrigin-RevId: 537088042
Change-Id: I6ef219a1a76baf3a751a6084a8fb8639c9bdac51

PiperOrigin-RevId: 537078029
Change-Id: Ibf0be0a8e83f62d492e96f1f1ed7839dbde30fae

Note that RE2 has taken a dependency on Abseil, so the `main` branch
should be used from now on. The `abseil` branch will go away soon...

PiperOrigin-RevId: 536941231
Change-Id: I502fc643a3763de51292396297b0939613473485

 https://google.github.io/styleguide/cppguide.html#Function_Comments

PiperOrigin-RevId: 536785792
Change-Id: I2963dea81a75b01b7275d784f6a2908816d0c7bf

Motivation is for WebAssembly to avoid perror which in turn requires file system emulation.

PiperOrigin-RevId: 536737294
Change-Id: I5177064c9451fb630ec5e9d0c0a0679fabd98afa

This avoids the need to use filesystem APIs just to write to stderr, which emscripten implements the same as this under the hood.

PiperOrigin-RevId: 536525710
Change-Id: I0a647a4593eacfba324505b6e8c4acfb577ac839

the cyclecounter by default, since it may be unsafe in some situations
(for example, if the system may enter a sleep state).

Fixes #1460

PiperOrigin-RevId: 535641718
Change-Id: I41c9cc4bc7a8ae7280ff9df00abd57668205045b

PiperOrigin-RevId: 535534819
Change-Id: Iccf8da3e0b084131e4c0dba205f3e190d3a66f4e

The fallback isn't totally portable, even within POSIX, but we can special case any future platforms where it's not just like this change does for Apple.

PiperOrigin-RevId: 535324103
Change-Id: Ib628925c4946b6c112373678fe37e9bb44259090

hash_test.cc leans heavily on INSTANTIATE_TYPED_TEST_SUITE_P, which is
quite memory- and CPU-hungry. Split a few heavyweight tests into a new
hash_instantiated_test.cc, reducing peak RAM consumption (or, on
multicore systems, compilation time).

PiperOrigin-RevId: 535305679
Change-Id: Ic204da0a47c749c3f7db5f902ade8d74ed3043bb

PiperOrigin-RevId: 535245982
Change-Id: I816f60c8b6476536df6836500f01c9a3ad88ddd4

PiperOrigin-RevId: 534841675
Change-Id: I6fbee6f0fe61d8dd51c9c7d93c3b47684a2f4d9f

PiperOrigin-RevId: 534824761
Change-Id: I0ab78fcb211bc5df756fb581761ed8febc3d18fd

`_LINUX_SOURCE_COMPAT` needs to be defined to fix

    absl/status/status.cc:494:5: error: duplicate case value
      494 |     case ENOTEMPTY:   // Directory not empty
          |     ^~~~
    absl/status/status.cc:480:5: note: previously used here
      480 |     case EEXIST:         // File exists
          |     ^~~~

The AIX assembler doesn't support numeric labels. Use a relative jump
instead to fix

    Assembler:
    /tmp/ccw16WCt.s: line 25: Error In Syntax

`time_zone_impl.cc` uses `std::mutex`.

PiperOrigin-RevId: 534619764
Change-Id: Ied99569176766f9708d5504eac9c7024cdcfd64b

The non-RAW_ versions provide better output but weren't available when most of these tests were written.

There are just a couple spots where RAW_ is actually needed, e.g. signal handlers and malloc hooks.

Also fix a couple warnings in layout_test.cc newly surfaced because the optimizer understands CHECK_XX differently than INTERNAL_CHECK.

PiperOrigin-RevId: 534584435
Change-Id: I8d36fa809ffdaae5a3813064bd602cb8611c1613

Imported from GitHub PR https://github.com/abseil/abseil-cpp/pull/1452

__cpuid is declared in intrin.h, but is excluded on non-Windows platforms.
We add this declaration to compensate.

Fixes #1358

PiperOrigin-RevId: 534449804
Change-Id: I91027f79d8d52c4da428d5c3a53e2cec00825c13