Commit graph

6 commits

Author SHA1 Message Date
Abseil Team
33841c5c96 Export of internal Abseil changes.
--
bb92c768e2271ddbebc1b1eb7e16a7b7c86a6e1c by Abseil Team <absl-team@google.com>:

Automated g4 rollback of changelist 244998488.

*** Reason for rollback ***

I'm seeing test failures, rolling this back.

*** Original change description ***

BEGIN_PUBLIC

The default constructor for optional<T> is filling dummy_ with zeros (see https://godbolt.org/z/IVea7X for a reduced example), which has a performance impact for large Ts. This comes from the gcc6 bugfix that made dummy as big as T. Because constexpr constructors are required to initialize all members of a struct, we cannot prevent this in a standard-compliant way as soon as dummy has any members (note that clang will happily accept adding a `constexpr dummy_type() {}` constructor...

***

PiperOrigin-RevId: 245004716

--
6e3ee35af50ffbee604c22300f3260ebc5f6cf52 by Abseil Team <absl-team@google.com>:

The default constructor for optional<T> is filling dummy_ with zeros (see https://godbolt.org/z/IVea7X for a reduced example), which has a performance impact for large Ts. This comes from the gcc6 bugfix that made dummy as big as T. Because constexpr constructors are required to initialize all members of a struct, we cannot prevent this in a standard-compliant way as soon as dummy has any members (note that clang will happily accept adding a `constexpr dummy_type() {}` constructor to dummy_type to prevent zero-initialization, but this is UB AFAICT).

This all stems from the fact that we're constructing an object by using placement new on dummy_. The solution I'm using here is to do the placement new on the actual data_. This creates a new issue in when T is volatile, because we can no longer use `&data_` to do the placement new. The solution I'm using here is to make data_ a non-const and non-volatile T, and only provide fully possibly qualified access through `reference()` accessors. I think this correctly prevents UB.

PiperOrigin-RevId: 244998488

--
4f52e64c4cf6aef8df6360007bcc53d8b00db2b4 by Abseil Team <absl-team@google.com>:

Increase SYMBOL_BUF_SIZE from 2KB to 3KB.

PiperOrigin-RevId: 244954529
GitOrigin-RevId: bb92c768e2271ddbebc1b1eb7e16a7b7c86a6e1c
Change-Id: Iaed9a027064a9ecd194c5c146169c683b77f12ef
2019-04-24 14:35:20 -04:00
nik7273
38b704384c Changed HTTP URLs to HTTPS where possible (#270) 2019-03-08 10:27:53 -05:00
Abseil Team
febc5ee6a9 Export of internal Abseil changes.
--
f9f068aa8a260dc576398e47b8e4540902e41358 by Derek Mauro <dmauro@google.com>:

Fix test string with embedded NUL. Currently parses as octal.

PiperOrigin-RevId: 237088193

--
d271ffdd3f450f817f6d30e98ff39d439aaf3a98 by Abseil Team <absl-team@google.com>:

Make symbolizer examine any mapping with read+exec permission regardless of 'w' bit.

PiperOrigin-RevId: 237056461

--
af315f8306d36a7367a452fd0b58cafdbf20719d by Abseil Team <absl-team@google.com>:

Switch comments referencing base:: CondVar and Mutex to absl::.

PiperOrigin-RevId: 236917884

--
c624d5d1c0bdb917bff5e651ba40599472f84e0e by Gennadiy Rozental <rogeeff@google.com>:

Internal change

PiperOrigin-RevId: 236898300

--
3cdc82429af964846d1152f49148abc61d196a4b by Samuel Benzaquen <sbenza@google.com>:

Make the `long double` overload if AbslHashValue a template to avoid invalid
conversions with implicit operators.

This overload was never meant to capture anything other than `long double` and any current caller to it that wasn't a `long double` is potentially a bug.
In particular, any type with an implicit `bool` conversion is calling this
overload instead of trying to find a hash<> specialization, thus causing
pretty bad hash behavior.

PiperOrigin-RevId: 236877073
GitOrigin-RevId: f9f068aa8a260dc576398e47b8e4540902e41358
Change-Id: If9cc008dd814f0ca06ed881f612c06575f1f7137
2019-03-06 14:49:15 -05:00
Abseil Team
3088e76c59 Export of internal Abseil changes.
--
5278e56bd7d322ecf161eaf29fd7fa3941d7431b by Greg Falcon <gfalcon@google.com>:

internal change

PiperOrigin-RevId: 222078614
GitOrigin-RevId: 5278e56bd7d322ecf161eaf29fd7fa3941d7431b
Change-Id: I1e86bef2e3733c81148a1a42dccd8182fe3f7fae
2018-11-19 15:11:12 -05:00
Abseil Team
a06c4a1d90 Export of internal Abseil changes.
--
5f1ab09522226336830d9ea6ef7276d37f536ac5 by Abseil Team <absl-team@google.com>:

Clarify the documentation of ABSL_MUST_USE_RESULT.

PiperOrigin-RevId: 221663609

--
af4c8359a20d56369fd1dce318220cf3be03ca66 by Greg Falcon <gfalcon@google.com>:

Internal change

PiperOrigin-RevId: 221538448

--
487cd09bd1942bf607080deeae38fee6ce66f294 by Eric Fiselier <ericwf@google.com>:

Work around emscripten bugs and missing features in absl/time:time_test.

The emscripten toolchain has a couple of issues that cause time_test
to fail. Specifically:

1) emscripten doesn't support signals.
2) The javascript implementation of strftime/strptime use different expansions
of '%c' that mean it doesn't round-trip.

PiperOrigin-RevId: 221523701

--
5823652e6a200b97b07334bc47128dfac40e20fc by Xiaoyi Zhang <zhangxy@google.com>:

Fix MSVC compiler warning by explicitly casting to char.
Currently our MSVC build breaks with the following error:
raw_hash_set.h(406): warning C4309: 'argument': truncation of constant value

PiperOrigin-RevId: 221492585

--
c5806358320711a5efbe5c523df13e14ab53a17d by Greg Falcon <gfalcon@google.com>:

Replace calls to getpagesize() with the more portable sysconf(_SC_PAGESIZE); the latter is in POSIX 1.0 and is called out in the Linux `getpagesize` man page as a more portable spelling.

PiperOrigin-RevId: 221492471

--
19ffe82851072229bb7ce73f754ffe4c18e8c575 by Abseil Team <absl-team@google.com>:

Fix -Wundef error in absl/hash/internal/hash.h.

PiperOrigin-RevId: 221444120

--
b30f3d0a848563b6e4ec33f3dc085831dfabb748 by Jon Cohen <cohenjon@google.com>:

Import of CCTZ from GitHub.

PiperOrigin-RevId: 221339736
GitOrigin-RevId: 5f1ab09522226336830d9ea6ef7276d37f536ac5
Change-Id: I96223d522d98bf6616dea88eb047c2d536eeddd0
2018-11-15 15:05:39 -05:00
Abseil Team
7fda099641 Changes imported from Abseil "staging" branch:
- f0a03a750a36dfbd7ab06d2913430ed5f988fd68 Add absl::RegisterSymbolizer() to mutex_nonprod.cc for AP... by Derek Mauro <dmauro@google.com>
  - f34a2ee35b4f6b321c570c51b0c381647176df63 Add the async signal-safe Symbolizer to Abseil. by Derek Mauro <dmauro@google.com>
  - 6a29ec2d6dc080691f6d32e1982201d1d173bdb3 Document preferred placement of ABSL_CONST_INIT attribute... by Abseil Team <absl-team@google.com>
  - 6f04ed6aa9c19bd717f0e8f422a97f3e3368cf30 Internal change. by Abseil Team <absl-team@google.com>
  - 0af9a330aff8fc0b41dcb3fe519930c36b01a9ef Declare absl::raw_logging_internal::SafeWriteToStderr in ... by Abseil Team <absl-team@google.com>
  - 223ff26745d31dfb4b59c36f3dee5441506af3c2 Fix ABSL_ARRAYSIZE() to handle rvalues. by Xiaoyi Zhang <zhangxy@google.com>

GitOrigin-RevId: f0a03a750a36dfbd7ab06d2913430ed5f988fd68
Change-Id: I491f9cc81ca3ee078fb737cbf8fa9bf6a730eee1
2018-02-28 12:52:04 -05:00