Refactor swisstable iterator debug messages code. The motivations are (a) distinguish between the "likely erased" and "could have rehashed" cases when generations are enabled, (b) suggest running under ASan when generations aren't enabled and doing so would narrow down the possible error cases, and (c) make ABSL_INTERNAL_ASSERT_IS_FULL not be a macro.

PiperOrigin-RevId: 511255275
Change-Id: I5a44a813cd310837d0bd0209d2187b100be201e7

Both e_shentsize and e_shstrndx are uint16, so the product
  elf_header.e_shentsize * elf_header.e_shstrndx
can overflow the promoted type int (MAX_UINT16 * MAX_UINT16 > MAX_INT),
which is undefined behavior. Not sure if it can affect any real cases
or not, though.

Cast e_shentsize to loff_t instead of e_shoff.
This makes both multiplication and addition to use loff_t type.

PiperOrigin-RevId: 511254775
Change-Id: I39c493bfb539cca6742aae807c50718d31e7c001

The `_dll` variable contains the target name. The code
was trying to use `_NAME` when handling ABSL_PROPAGATE_CXX_STD.
Probably a cut&paste error from the other places where we
handle that option.

PiperOrigin-RevId: 510551097
Change-Id: I006435978333dd7a72a300503389a7bf48fb6d0d

PiperOrigin-RevId: 510549728
Change-Id: I4447197b56dc76aaf4dd4e78671de247af2a91e8

PiperOrigin-RevId: 510518832
Change-Id: I086c8f2e34312e2b0384b3c67b9c04814d41ddf0

monotonic clocks on Linux when the implementation uses futexes

After this change, when synchronization methods that wait are passed
an absl::Duration to limit the wait time, these methods will wait for
that interval, even if the system clock is changed (subject to any
limitations with how CLOCK_MONOTONIC keeps track of time). In other
words, an observer measuring the time with a stop watch will now see
the correct interval, even if the system clock is changed. Previously,
the duration was added to the current time, and methods would wait
until that time was reached on the possibly changed realtime system
clock.

The behavior of the synchronization methods that take an absl::Time is
unchanged.  These methods always wait until the absolute point in time
is reached and respect changes to the system clock. In other words, an
observer will always see the timeout occur when a wall clock reaches
that time, even if the clock is manipulated externally.

Note: ABSL_PREDICT_FALSE was removed from the error case in Futex as
timeouts are handled by this case, and timeouts are part of normal
operation.
PiperOrigin-RevId: 510405347
Change-Id: I0b3ea390de97014cfa353079ae2e0c1c637aca69

std::chrono methods used by std::condition_variable.

A followup change will add an implemention of
synchronization_internal::Waiter that can use
std:mutex/std::condition_variable to implement the per-thread
semaphore that absl::Mutex waits on. This implementation may at some
point become the default on platforms such as Windows where there
doesn't seem to be an easy way of supporting real absolute timeouts. In
this case we can defer to their standard library to implement correct
support.
PiperOrigin-RevId: 510204786
Change-Id: Icf4d695013fd060abbd53dae23e71ea36f731565

This also helps a lot with dealing with conversions and data structure creation under the hood.

Change to checking for use_count() >/== 1.

PiperOrigin-RevId: 510125744
Change-Id: I572cca18c3f827f5d3eefb2ec19a1a014c0090ae

instead of absl::ToUnixNanos(absl::Now());

PiperOrigin-RevId: 509829866
Change-Id: Ib34362762304ad6eb7980a1227d717069b84f656

Supporting _m_prefetchw() (officially part of 3DNOW) across various compilers and platforms turns out to be difficult. This change removes the explicit _m_prefetchw call aimed at MSVC compilers given that it causes issues in clang / chromium compilations.

PiperOrigin-RevId: 509632497
Change-Id: Ib1b6b2cf667cbc1af5ed6651cd9aa0294a9265b6

Make default-constructed swisstable iterators use EmptyGroup() for ctrl_ so that we can distinguish between end() iterators and default-constructed iterators in debug mode.

PiperOrigin-RevId: 509606271
Change-Id: I77b68590b3904a4cf7809b75d814d74cb89603b6

PiperOrigin-RevId: 509581648
Change-Id: Ia60053f1172e0e64569e9555524812fc4788e24f

https://bugs.llvm.org/show_bug.cgi?id=38289 has been addressed since 2019.

We just have to check to see if the clang being used to compile with is any version before major version 9.

PiperOrigin-RevId: 509299527
Change-Id: I2f9a801faf87f968f1625711a67d38a528f0740c

The test was getting tripped by EXPECT_DEATH_IF_SUPPORTED keeping a literal
copy of the test statement.

PiperOrigin-RevId: 509299343
Change-Id: I63da29d844cc630d750535fd4bb13a2895d32541

PiperOrigin-RevId: 509236105
Change-Id: I635f521aff106875f06bb93c332d3e437d0ad409

On Linux Kernels >= 5.4 MSan reports a false positive when accessing thread local storage data from loaded libraries.

This was reported on Chromium (which on some build configurations uses absl as a dynamic library). More info here: crbug.com/1414573.

PiperOrigin-RevId: 508645053
Change-Id: I5d5a97e1ee7230cc23f3934a4ec5594b883918b4

APIs that take KernelTimeout as a parameter can now query if an
absolute or relative timeout was requested. If the underlying API can
only use one type of timeout, the code will do a reasonable
conversion.

The goal is to eventually enable the possibility of using wait times
that are based on monotonic clocks that are safe against system clock
steps.

PiperOrigin-RevId: 508541507
Change-Id: Id08bf13515f3e1bfd78d88393cde98a6fd3ef72c

We change AssertSameContainer to be after AssertIsValidForComparison calls so that we can have more specific failure messages.

PiperOrigin-RevId: 508472485
Change-Id: Iff2f7512086948a4aca7fd403596e8d4fde53b2a

PiperOrigin-RevId: 508124592
Change-Id: Ib183e6e241c81b2760e7f849f8af8e7e2c30ea42

Change implementation of OnlyLiteralZero to only fail if the second overload is chosen, not in overload resolution.

PiperOrigin-RevId: 508031321
Change-Id: I6981371fbc6498047babe3468f3343090f8bda47

PiperOrigin-RevId: 507790741
Change-Id: I347357f9a2d698510f29b7d1b065ef73f9289292

Fix a discrepancy between absl::Hash and absl::HashOf for some negative signed integral types and improve the performance of absl::Hash.

PiperOrigin-RevId: 507598042
Change-Id: I96a7bd6b9c360f435f216b2671ae84d9768a46e8

`PrefetchToLocalCacheForWrite` should use `__builtin_prefetch(<addr>, 1, 3)` not `__builtin_prefetch<addr>, 1, 0)`

PiperOrigin-RevId: 507015724
Change-Id: I99f44fe819c27e6dcc0bb7b36f1a37015c6f1987

PiperOrigin-RevId: 506944023
Change-Id: I04ff13dcec89c57ba03707f5698c4198434216b6

PiperOrigin-RevId: 506622658
Change-Id: I17ae2d97a6cadb7bdd8ebd0ec0dd3976568cb7e1

Get rid of tail padding within `absl::Duration`. This reduces memory usage needs when storing duration in containers (e.g. `vector<absl::Duration>` uses 25% less memory), and allows classes with `absl::Duration` fields to fit other stuff in memory previously used by tail padding (e.g. `std::optional<absl::Duration>` is now 16 bytes instead of 24).

PiperOrigin-RevId: 506568782
Change-Id: Ic9e077f02a80da013fb2d312aff77761b970c07a

Get rid of tail padding within `absl::Duration`. This reduces memory usage needs when storing duration in containers (e.g. `vector<absl::Duration>` uses 25% less memory), and allows classes with `absl::Duration` fields to fit other stuff in memory previously used by tail padding (e.g. `std::optional<absl::Duration>` is now 16 bytes instead of 24).

PiperOrigin-RevId: 506543503
Change-Id: Ifeb2397c953a5d3da317a70ab49a3ebb85042344

PiperOrigin-RevId: 506323250
Change-Id: I0f7d4532c19088b011740ceff87ecec55cc34edb

Rollforward: in sanitizer mode, detect when references become invalidated by randomly rehashing on insertions when there is no reserved growth.

Rollforward of ed59f62f

PiperOrigin-RevId: 506314970
Change-Id: I7a654aef36bb169da9ea5c618789ee771f05fe28

PiperOrigin-RevId: 506151912
Change-Id: I388884b0bd80ef0f8a0b3d3c7d0e1b404ddfb742

Rollback in sanitizer mode, detect when references become invalidated by randomly rehashing on insertions when there is no reserved growth.

Rollback of ed59f62f

PiperOrigin-RevId: 506003574
Change-Id: I1766321f279a3226e2821e0390387d5639d28964

In sanitizer mode, detect when references become invalidated by randomly rehashing on insertions when there is no reserved growth.

PiperOrigin-RevId: 505807487
Change-Id: I9051a04f6a75e579d16e9ae8defd404bcc377fba

On some glibcs, this is defined as a call to sysconf(), which returns a long.

PiperOrigin-RevId: 505380003
Change-Id: I53207846d733d3a529630a6aff9bca425cf90a21

PiperOrigin-RevId: 505184961
Change-Id: I64482558a76abda6896bec4b2d323833b6cd7edf

`absl::Hex` and `absl::Dec` are now stringifiable through `AbslStringify`. This means that they can be passed to logging, `absl::StrFormat`, `absl::StrCat` and `absl::Substitute`. Note that this change will break unsupported usages of `absl::AlphaNum`. `absl::AlphaNum` is not intended to be instantiated as a stack variable due to lifetime issues. Unsupported usages include:

* constructing an `absl::AlphaNum` as a local variable.
  * For example, `absl::AlphaNum a = absl::Hex(...);`
* declaring an `absl::AlphaNum` as a member variable

Usage of `absl::AlphaNum` as a function parameter will continue to be supported.

PiperOrigin-RevId: 505158534
Change-Id: Idecdf0b3d137482e86b393c7480229343d68eb36

PiperOrigin-RevId: 504981951
Change-Id: I7357e8e6730ceda458b40aa605cd450e80271dc6

PiperOrigin-RevId: 504941246
Change-Id: I94c1e85afd254e84948477b511d41eeb8285fdae