Dave Watson [Wed, 22 Nov 2017 16:16:04 +0000 (08:16 -0800)]
Add a fast path to folly::ThreadLocal
Summary:
Currently folly::ThreadLocal[Ptr] is pretty heavy-weight for a get():
1) call instance(), take a static init guard, branch
2) call getThreadEntry, check if thread_local is not null, branch
3) check if id < threadEntry->capacity, branch
4) Finally, return threadEntry->elements[id]
If we have real thread_locals, we can do better by caching the capacity directly,
combining all three checks:
1) checkif id < threadLocalCapacityCheck, branch. If not, do slow path.
2) return threadEntry->elements[id]. Threadentry is never null if capacity > 0, and
instance() setup work is called during the first getThreadEntry call when threadlocalcapacity == 0.
Reviewed By: yfeldblum
Differential Revision:
D6379878
fbshipit-source-id:
4fc7564bbb2f319d65875124026aef28d910ef06
Andrii Grynenko [Wed, 22 Nov 2017 04:58:05 +0000 (20:58 -0800)]
Fix ASAN failure in FutureDAG test
Summary: dogscience
Reviewed By: yfeldblum
Differential Revision:
D6387604
fbshipit-source-id:
3e33da3e3c6801d6e2186637c8fb6adaf1eef6f6
Aaron Orenstein [Tue, 21 Nov 2017 22:02:18 +0000 (14:02 -0800)]
Add -Werror to -Wshadow-local and -Wshadow-compatible-local warning check
Summary:
Without -Werror Apple LLVM will print a warning about -Wshadow-local
and -Wshadow-compatible-local but won't fail. This causes problems for
downstream code which uses -Werror.
Reviewed By: meyering
Differential Revision:
D6375771
fbshipit-source-id:
4ac3a4ff46bdb76f1f07b0c8ba479a7d404c8c9b
Adam Simpkins [Tue, 21 Nov 2017 18:44:02 +0000 (10:44 -0800)]
logging: add a LogConfig class and parsing code
Summary:
Add a new LogConfig class to represent the current configuration state of the
LoggerDB. This also includes code to parse config strings in both JSON and a
simpler more human-writable format (intended primarily for use in command line
arguments).
I generally expect the human-writable format to be used mainly to configure log
levels. It also supports configuring log handler details as well, but the
format for this data is slightly more cumbersome and will probably be harder
for people to easily remember.
The parsing code is intentionally kept as part of the 'init' library rather
than the core 'logging' library so that other libraries that simply wish to log
messages do not need to depend on it. For instance, this would allow the folly
JSON library to use the logging library without causing a circular dependency.
Reviewed By: bolinfest
Differential Revision:
D6200560
fbshipit-source-id:
e4e3c7f941808251b6c7bcbbdac0210118675fb0
Dave Watson [Tue, 21 Nov 2017 15:34:15 +0000 (07:34 -0800)]
improve ThreadLocalBenchmark
Summary:
Benchmark appears to be memory bound, and is affected by different cpus.
Add a new benchmark that does more work.
Reviewed By: yfeldblum
Differential Revision:
D6380904
fbshipit-source-id:
5cfbaab70379aa4a2923d957e1147d8486deeff7
Dave Watson [Tue, 21 Nov 2017 15:33:09 +0000 (07:33 -0800)]
fix hazptr_array move
Summary: Broken, found via build failures in
D6259947
Reviewed By: yfeldblum
Differential Revision:
D6374215
fbshipit-source-id:
3a438b5416683c602966c2f029c6b7787acdccaa
Giuseppe Ottaviano [Tue, 21 Nov 2017 07:10:34 +0000 (23:10 -0800)]
Revert
D6366352: [folly] Split get_default() into two for deferred default construction
Summary:
This reverts commit
db55b944ca63e565997094c11b90c4ebe98531ce
bypass-lint
Differential Revision:
D6366352
fbshipit-source-id:
e25906409186b077ef9117aa524cc7c86314ae12
Aaryaman Sagar [Mon, 20 Nov 2017 23:59:34 +0000 (15:59 -0800)]
Split get_default() into two for deferred default construction
Summary:
As it stood currently folly::get_default() would unnecessarily
construct a value into the third parameter, which was unnecessary in the fast
path where the element was found in the map
Reviewed By: yfeldblum
Differential Revision:
D6366352
fbshipit-source-id:
db55b944ca63e565997094c11b90c4ebe98531ce
Dylan Yudaken [Mon, 20 Nov 2017 23:59:02 +0000 (15:59 -0800)]
correct usage of namespace std for coroutines_trait specialization
Summary: When specializing traits you should not use the libcpp inline std namespace, but rather the plain one.
Reviewed By: yfeldblum
Differential Revision:
D6373787
fbshipit-source-id:
d62ef9ccbf90c6f952db5f93f2377c920e68a809
Christopher Dykes [Mon, 20 Nov 2017 21:44:02 +0000 (13:44 -0800)]
Work around a bug in MSVC name lookup within templated friend contexts
Summary: MSVC has a bug that causes it to think that `once_flag` in this context resolves to the `using` statement further up, rather than `detail::once_flag`.
Reviewed By: yfeldblum
Differential Revision:
D6376501
fbshipit-source-id:
481dbd75ed21f3d519bd920258fa743f314668ad
Yedidya Feldblum [Mon, 20 Nov 2017 21:29:48 +0000 (13:29 -0800)]
No need for a wrapping structure for posixTimeToDuration
Summary:
[Folly] No need for a wrapping structure for `posixTimeToDuration`.
We can just use a variant of tag dispatch. In this variant, we pass to `posixTimeToDuration` a default-initialized value of the desired return type and we write overload templates for each possible variant. The argument is used purely for overload resolution and return type deduction, not for its runtime value. It is slightly different from tag dispatch because we do not use separate types which are purely tag types.
Reviewed By: simpkins
Differential Revision:
D6371572
fbshipit-source-id:
1987dee31fceec8733caa61495e96489dbf1ca39
Dave Watson [Mon, 20 Nov 2017 16:18:27 +0000 (08:18 -0800)]
Add free-function retire
Summary:
Adds a hazptr_retire and domain::retire methods to hazptr. They both allocate memory instead of inheriting.
This will make implementation of google's cell proposal trivial, vs.
D4754972 which felt overwraught.
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0561r0.html
Reviewed By: magedm
Differential Revision:
D6361162
fbshipit-source-id:
9f259f434139f960483b2ab7c5190d3807edcf52
Tianjiao Yin [Mon, 20 Nov 2017 02:51:19 +0000 (18:51 -0800)]
fix SingletonTest
Summary: There is no guarantee that this code will finish in 6 seconds (nor other threads will release singleton in 5 seconds), especially in ASAN mode. Though I don't have better idea, this diff will relax the condition in ASAN mode to make the unit-test less flaky.
Reviewed By: yfeldblum
Differential Revision:
D6371692
fbshipit-source-id:
58dd15cc0b3273719314c8b323ba88ee47e8ff61
Adam Simpkins [Sun, 19 Nov 2017 23:14:20 +0000 (15:14 -0800)]
implement to() conversions for std::chrono to timespec/timeval
Summary:
Add folly::to() conversions to convert between std::chrono::duration or
std::chrono::time_point types and struct timespec or struct timeval types.
To conform to the behavior of the existing arithmetic-to-arithmetic
conversions, this code performs proper overflow checking and throws a
`ConversionError` on overflow. This unfortunately does make the code rather
complicated compared to a non-checking implementation.
Conversions between some unusual duration types is not implemented yet, and
will fail at compile time if someone tries to use it. This happens for
durations where neither the numerator nor the denominator of the ratio is 1.
For instance, 7/13ths of a second.
Reviewed By: yfeldblum
Differential Revision:
D6356700
fbshipit-source-id:
9dce8ab8f32d8c18089f32c7176a8abf3c3f11f7
Pingjia Shan [Sun, 19 Nov 2017 05:55:44 +0000 (21:55 -0800)]
get rid of redundant calls to RequestContext::saveContext()
Summary:
In the past, these calls were required.
Used to solve static destruction ordering issue. Any static object that uses RequestContext must call this function in its constructor.
That is when we were using `static folly::ThreadLocal<std::shared_ptr<RequestContext>>`, which was non-leaky.
The problem being addressed is when we have some code of the form:
```lang=c++
void doWork() {
static EventBase eb;
}
```
But now we are using `SingletonThreadLocal<std::shared_ptr<RequestContext>>`, which is leaky.
So the issue that these calls were there to address seems to have been resolved.
Reviewed By: yfeldblum
Differential Revision:
D6332597
fbshipit-source-id:
c6aba6620ef2fb3a344ea20f56c8b9c0cdf42c70
Pádraig Brady [Sat, 18 Nov 2017 22:25:54 +0000 (14:25 -0800)]
folly: support FOLLY_FALLTHROUGH on GCC
Summary:
This is required to avoid new GCC 7 -Wimplict-fallthrough warnings.
We also update to use the C++17 [[fallthrough]] attribute if supported.
Reviewed By: yfeldblum
Differential Revision:
D6367140
fbshipit-source-id:
d5380983cb300f944df9c2885d0faa0155994be7
Tianjiao Yin [Sat, 18 Nov 2017 14:39:09 +0000 (06:39 -0800)]
reduce the number of iteration when test SharedMutexTest in ASAN mode
Summary: This unit-test takes too long to finish in ASAN mode from heavily loaded system (more than 10 minutes).
Reviewed By: yfeldblum
Differential Revision:
D6362111
fbshipit-source-id:
b097eff60f88ace4fb869132598806700804e267
Pingjia Shan [Fri, 17 Nov 2017 17:40:14 +0000 (09:40 -0800)]
Remove unnecessary white line
Summary:
Seems to be causing rendering error:
{
F114885894}
Created from Diffusion's 'Open in Editor' feature.
Reviewed By: lskuff
Differential Revision:
D6358188
fbshipit-source-id:
3c437007b425bcadb79a71807ad92d520a70e8cb
Jim Meyering [Fri, 17 Nov 2017 04:52:38 +0000 (20:52 -0800)]
folly/fibers/test/FibersTest.cpp: accommodate ASAN's detect_stack_use_after_return=1
Summary:
With ASAN enabled (actually, only with ASAN *and* its detect_stack_use_after_return=1 option),
the addTaskFinally test would fail. This adapts to accommodate the larger stack offsets.
Also, use EXPECT_GT and EXPECT_LT rather than EXPECT_TRUE.
Reviewed By: yfeldblum
Differential Revision:
D6353666
fbshipit-source-id:
39e04caffa7b24cde97c749686c7e651a071dcec
Jody Ho [Thu, 16 Nov 2017 07:23:02 +0000 (23:23 -0800)]
Expose the time remaining in HHWheelTimer::Callback
Summary:
We would like to know the time remaining for a scheduled timeout to decide
whether a new event should override the scheduled timeout.
Reviewed By: yfeldblum
Differential Revision:
D6334067
fbshipit-source-id:
f172d5cd7fc804db5fd53a42d06cadfddf857e22
Subodh Iyengar [Thu, 16 Nov 2017 01:59:20 +0000 (17:59 -0800)]
Add unit test for timeout=0
Summary:
Add unit test for immediate timeout since
we're using it in a few places.
Reviewed By: yfeldblum
Differential Revision:
D6327012
fbshipit-source-id:
ea80763d080b06e2a43277118d3147dc8016f348
Yedidya Feldblum [Wed, 15 Nov 2017 23:52:02 +0000 (15:52 -0800)]
Fix wrong source name in tests Makefile
Summary:
[Folly] Fix wrong source name in tests `Makefile.am`.
Closes #714.
Reviewed By: Orvid
Differential Revision:
D6340209
fbshipit-source-id:
88da57fb4d663071eb578efec50c7e162b0c03d1
Benny Chen [Wed, 15 Nov 2017 22:03:27 +0000 (14:03 -0800)]
allow small vector to be storage for sorted_vector_map
Summary:
this is to allow small_vector to be a storage option for sorted_vector_map. Reas
on I want to do this is because in ads there are a lot of small maps where we wo
uld have to allocate separately.
Reviewed By: yfeldblum
Differential Revision:
D6318811
fbshipit-source-id:
b145d1bef2cbbeb946995aa66b55aaadeb6c54f5
Adam Simpkins [Wed, 15 Nov 2017 20:48:56 +0000 (12:48 -0800)]
add EXPECT_THROW_RE() and EXPECT_THROW_ERRNO() test macros
Summary:
Add EXPECT_THROW_RE() and EXPECT_THROW_ERRNO() macros to folly/test/TestUtils.h
These allow more precise checks than the basic EXPECT_THROW() macro provided as
part of gtest.
These macros are being moved into folly from Facebook's eden repository
(https://github.com/facebookexperimental/eden)
This will allow us to use them in folly tests and in other projects that depend
on folly.
Reviewed By: yfeldblum
Differential Revision:
D6301760
fbshipit-source-id:
1f434fb5bc9b7859f763171264fb0b2e1b4bda62
Igor Sugak [Wed, 15 Nov 2017 06:42:28 +0000 (22:42 -0800)]
fix asan build
Reviewed By: yfeldblum
Differential Revision:
D6332075
fbshipit-source-id:
26a279b6ee4253a8be8f9f63e6900c2082a5486d
Yedidya Feldblum [Wed, 15 Nov 2017 06:39:18 +0000 (22:39 -0800)]
Extra expectations for ADL toAppend
Summary: [Folly] Extra expectations for ADL `toAppend`.
Reviewed By: Orvid
Differential Revision:
D6330547
fbshipit-source-id:
d1363280f097f860883ba84d6cfe7caa4e4cf9de
Christopher Dykes [Tue, 14 Nov 2017 22:01:08 +0000 (14:01 -0800)]
Minor tweak to the wording of the vcpkg section
Summary:
The commands as they were previously would attempt to install Folly as 32-bit, which would fail.
Also removes an extra `-` in the link's url.
Closes: https://github.com/facebook/folly/issues/713
Reviewed By: yfeldblum
Differential Revision:
D6327981
fbshipit-source-id:
ced8c86948b8be5c1dd88e14d6c3a77fc783aac9
Jason Fried [Mon, 13 Nov 2017 21:32:49 +0000 (13:32 -0800)]
Add ByteRange/StringPiece Conversion for Cython
Summary:
Moving the cython definition to folly/python.
Adding simple conversion helper to_bytes
This is to cut down on duplicate folly::range cython definitions
Reviewed By: yfeldblum
Differential Revision:
D6291125
fbshipit-source-id:
314b732a1516a03fb5c9a57939552bbabd81970b
Andrew Krieger [Sun, 12 Nov 2017 22:31:02 +0000 (14:31 -0800)]
Fix an ICE in MSVC from functional/Invoke.h
Summary:
Similar to https://github.com/facebook/folly/commit/
29ffcc50981fd50dd2ab1a69f8a262c4b7c27ad4,
(
D6254219), this was manifesting as random ICEs that repro consistently in a given file on a given machine,
but in different files for different machines and sometimes not at all.
Reviewed By: Orvid, yfeldblum
Differential Revision:
D6303186
fbshipit-source-id:
d91ad633352fc5c28640e71fd1539f6a4ba8f70f
Yedidya Feldblum [Sun, 12 Nov 2017 00:16:59 +0000 (16:16 -0800)]
Fix some declared tests not being included in Makefile TESTS list
Summary: [Folly] Fix some declared tests not being included in `Makefile` `TESTS` list.
Reviewed By: pixelb
Differential Revision:
D6307755
fbshipit-source-id:
8853d8fce25768ff17f9b1e53b1fbba0aa213691
Yedidya Feldblum [Sat, 11 Nov 2017 00:03:47 +0000 (16:03 -0800)]
Backport C++17 container access functions: size, empty, data
Summary: [Folly] Backport C++17 container access functions: `size`, `empty`, `data`.
Reviewed By: Orvid
Differential Revision:
D6301986
fbshipit-source-id:
65c31df68b7743d5ed4a2cf2419586b862683c48
Sergey Doroshenko [Fri, 10 Nov 2017 23:27:22 +0000 (15:27 -0800)]
fix typo in Subprocess.h comments
Summary: (Note: this ignores all push blocking failures!)
Reviewed By: yfeldblum
Differential Revision:
D6299313
fbshipit-source-id:
95be2339f4845502e3c446698e54643eeac8055f
Christopher Dykes [Fri, 10 Nov 2017 20:28:17 +0000 (12:28 -0800)]
Minor tweaks to the wording of a couple of errors in the CMake build
Summary:
Tweak the wording of a couple error messages to make them clearer.
Also disable the 'digraphs not supported' warning, as none of our supported compilers actually interpret digraphs in any of our supported build configurations.
Closes: https://github.com/facebook/folly/issues/706
Reviewed By: yfeldblum
Differential Revision:
D6299715
fbshipit-source-id:
7c847ac859e082aea711f6751f626b4b43886da4
Pádraig Brady [Fri, 10 Nov 2017 03:55:39 +0000 (19:55 -0800)]
folly: fix sysMembarrier() with newer kernel headers
Summary:
When __NR_membarrier is defined (on newer kernels),
we don't define the constants, but also by default
don't include the membarrier header to define the necessary constants.
Therefore split the definition of the syscall value and
the constants used with it, to ensure both are defined.
Reviewed By: yfeldblum
Differential Revision:
D6292178
fbshipit-source-id:
31ba9d4a698a4f5e14ae34de0acf8f851d75527d
Pádraig Brady [Fri, 10 Nov 2017 03:26:16 +0000 (19:26 -0800)]
folly: avoid compile warning/failure due to lvalue-to-rvalue conversion
Summary:
With gcc 7.2 we get the warning:
folly/io/async/DelayedDestructionBase.h:252:20:
error: parameter ‘right’ set but not used [-Werror=unused-but-set-parameter]
std::nullptr_t right) {
^~~~~
I presume this is due to the implicit conversion, hence the named parameter is
never assigned. Instead we use an explicit nullptr.
Reviewed By: yfeldblum
Differential Revision:
D6279302
fbshipit-source-id:
ed449601b0410c178777f20e82ed09d9097bd024
Shubhanshu Agrawal [Fri, 10 Nov 2017 01:35:30 +0000 (17:35 -0800)]
adding a fibers compatible once flag
Summary:
The current folly::once_flag is not compatible with folly fibers
and when using it with fibers is inefficient and also cause deadlocks.
This diff makes the once flag's mutex be a templatable paramter
and overrides it in fibers library with a fiber compatible mtuex.
Reviewed By: yfeldblum
Differential Revision:
D6288508
fbshipit-source-id:
6f82e1794d1f417f8d267061f1702a26a7b4ff12
Yedidya Feldblum [Thu, 9 Nov 2017 21:34:35 +0000 (13:34 -0800)]
Add makeSemiFuture declarations to helpers.h
Summary:
[Folly] Add `makeSemiFuture` declarations to `helpers.h`.
For consistency with the `makeFuture` declarations that are also there. Definitions for both are found in `Future-inl.h`.
Reviewed By: LeeHowes
Differential Revision:
D6281826
fbshipit-source-id:
4b22dd9086d05dbdebba358c6f569a772017949a
Phil Willoughby [Thu, 9 Nov 2017 19:39:48 +0000 (11:39 -0800)]
Make ColdClassTest work on ancient compilers
Summary:
Some older versions of GCC/glibc/etc do not have the
std::is_trivially*_constructible or std::is_trivially*_assignable traits.
Reviewed By: yfeldblum
Differential Revision:
D6285887
fbshipit-source-id:
1eb4ae4f899dc1f528321f9f087390291687aca3
Dan Melnic [Thu, 9 Nov 2017 19:28:07 +0000 (11:28 -0800)]
Remove the zerocopy write threshold support, add support for ENOBUFS
Summary: Remove the zerocopy write threshold support since it is a little bit confusing
Reviewed By: djwatson
Differential Revision:
D6256854
fbshipit-source-id:
1c992f93d7b04c4ede2fbefebde7a7ae89de3764
Teng Qin [Thu, 9 Nov 2017 06:11:33 +0000 (22:11 -0800)]
Add parsing for indirect functions
Summary:
Currently `folly::symbolizer`'s `getDefinitionByAddress` and `getSymbolByName` only parses `STT_OBJECT` and `STT_FUNC`. There are some standar library functions that uses the GNU indirect function feature that would have been missed:
==== For libpthread-2.23.so:
====== Symbol system Addr 119d0 Size 8 is a STT_GNU_IFUNC
====== Symbol siglongjmp Addr 10700 Size 8 is a STT_GNU_IFUNC
====== Symbol longjmp Addr 10700 Size 8 is a STT_GNU_IFUNC
====== Symbol __vfork Addr 10af0 Size 8 is a STT_GNU_IFUNC
====== Symbol vfork Addr 10af0 Size 8 is a STT_GNU_IFUNC
====== Symbol system_ifunc Addr 119d0 Size 8 is a STT_GNU_IFUNC
====== Symbol longjmp_ifunc Addr 10700 Size 8 is a STT_GNU_IFUNC
====== Symbol vfork_ifunc Addr 10af0 Size 8 is a STT_GNU_IFUNC
====== Symbol siglongjmp_ifunc Addr 10700 Size 8 is a STT_GNU_IFUNC
====== Symbol __vfork_ifunc Addr 10af0 Size 8 is a STT_GNU_IFUNC
====== Symbol __vfork@GLIBC_2.2.5 Addr 10af0 Size 8 is a STT_GNU_IFUNC
====== Symbol siglongjmp@GLIBC_2.2.5 Addr 10700 Size 8 is a STT_GNU_IFUNC
====== Symbol vfork@GLIBC_2.2.5 Addr 10af0 Size 8 is a STT_GNU_IFUNC
====== Symbol system@GLIBC_2.2.5 Addr 119d0 Size 8 is a STT_GNU_IFUNC
====== Symbol longjmp@GLIBC_2.2.5 Addr 10700 Size 8 is a STT_GNU_IFUNC
==== For libc-2.23.so:
====== Symbol __gettimeofday Addr c05e0 Size a8 is a STT_GNU_IFUNC
====== Symbol strcpy Addr 8e150 Size 35 is a STT_GNU_IFUNC
====== Symbol wmemcmp Addr afb50 Size 37 is a STT_GNU_IFUNC
====== Symbol strncmp Addr 8eb30 Size 41 is a STT_GNU_IFUNC
====== Symbol stpncpy Addr 929f0 Size 35 is a STT_GNU_IFUNC
====== Symbol __mempcpy_chk Addr 11cec0 Size 68 is a STT_GNU_IFUNC
====== Symbol strncpy Addr 903d0 Size 35 is a STT_GNU_IFUNC
====== Symbol time Addr c0500 Size a8 is a STT_GNU_IFUNC
====== Symbol strpbrk Addr 90700 Size 22 is a STT_GNU_IFUNC
====== Symbol strspn Addr 90a80 Size 22 is a STT_GNU_IFUNC
====== Symbol __stpncpy Addr 929f0 Size 35 is a STT_GNU_IFUNC
====== Symbol __strcasecmp Addr 92a80 Size 54 is a STT_GNU_IFUNC
====== Symbol memset Addr 92230 Size 41 is a STT_GNU_IFUNC
====== Symbol strstr Addr 916b0 Size 21 is a STT_GNU_IFUNC
====== Symbol strcspn Addr 8e270 Size 22 is a STT_GNU_IFUNC
====== Symbol memcmp Addr 91c40 Size 37 is a STT_GNU_IFUNC
====== Symbol mempcpy Addr 923b0 Size 68 is a STT_GNU_IFUNC
And 80 more...
This Diff adds parsing for `STT_GNU_IFUNC` symbols as well
Reviewed By: yfeldblum
Differential Revision:
D6282727
fbshipit-source-id:
71b7c44831e4ddfdccf1e794cb86e049e14227bc
Teng Qin [Thu, 9 Nov 2017 05:43:58 +0000 (21:43 -0800)]
Add a multi-type version of iterateSectionsWithType
Summary:
When using `folly::symbolizer`, it's very often that we want to use `iterateSectionsWithType` iterate through sections of a few types using the same callback. Current approach would require iterating the section header multiple times.
This Diff add `iterateSectionsWithTypes`, which is basically just `iterateSectionsWithType` but accepts multiple section types.
It is very similar to
D6279651. However, in this Diff we did not change implementation of `getDefinitionByAddress` and `getSymbolByName`, since going through `.dynsym` separately would improve the efficiency of single-address or single-symbol lookup. However, for the use cases that we want to iterate through all symbols of an ELF file, this new interface would be useful.
Reviewed By: anakryiko, yfeldblum
Differential Revision:
D6281449
fbshipit-source-id:
f9afe0a0e95d9fafcd041014abad8ca86d1a882f
Yedidya Feldblum [Thu, 9 Nov 2017 05:19:56 +0000 (21:19 -0800)]
Fix folly/test/Makefile.am
Summary:
[Folly] Fix `folly/test/Makefile.am`.
Closes #709.
Reviewed By: meyering
Differential Revision:
D6282009
fbshipit-source-id:
0f2a992e92d4b94a535f29341920c9f2959819d7
Teng Qin [Thu, 9 Nov 2017 04:04:18 +0000 (20:04 -0800)]
Add a multi-type version of iterateSymbolsWithType
Summary:
When using `folly::symbolizer`, it's very often that we want to use `iterateSymbolsWithType` iterate through symbols of a few types using the same callback. Current approach would require iterating the section multiple times.
This Diff adds `iterateSymbolsWithTypes`, which is basically just `iterateSymbolsWithType` but accepts symbol types.
This Diff also updated implementation of `getDefinitionByAddress` and `getSymbolByName` which currently does two iterations for `STT_OBJECT` and `STT_FUNC`.
Reviewed By: yfeldblum
Differential Revision:
D6279651
fbshipit-source-id:
a661dd15f18e4f2f63dbcca615f5a86d92e528ea
Yedidya Feldblum [Thu, 9 Nov 2017 02:55:00 +0000 (18:55 -0800)]
In SemiFuture members, fix DCHECK of pointer types
Summary:
[Folly] In `SemiFuture` members, fix `DCHECK` of pointer types.
Use `nullptr !=` to avoid ` error: ISO C++ forbids comparison between pointer and integer`.
Closes #708.
Reviewed By: Orvid
Differential Revision:
D6277832
fbshipit-source-id:
8f65065d5347c6ac407b99cb780c38935e901362
Phil Willoughby [Thu, 9 Nov 2017 00:38:47 +0000 (16:38 -0800)]
Actually mark Unexpected as cold
Summary:
Testing indicates that GCC ignores the cold attribute when the function
is available for inlining. Because Unexpected is a template class we
can't make the constructors non-inline, but we can make it derive from a
class with a cold constructor, which has the effect of making all the
Unexpected constructors implicitly cold.
Reviewed By: yfeldblum
Differential Revision:
D6261013
fbshipit-source-id:
482e49253d5b104742018133c53fb60279dd9f9b
Christopher Dykes [Thu, 9 Nov 2017 00:09:09 +0000 (16:09 -0800)]
Mark the base class of NoFutureInSplitter as public
Summary: Because every other exception is already marked as public.
Reviewed By: yfeldblum
Differential Revision:
D6275414
fbshipit-source-id:
8e1fc4ceafedbdb44733ab57aecb6050c4160994
Arkady Shapkin [Wed, 8 Nov 2017 19:30:38 +0000 (11:30 -0800)]
Fix static linking gflags and glog
Summary:
gflags and glog require some compile definitions when linking statically
/cc Orvid
Closes https://github.com/facebook/folly/pull/693
Reviewed By: yfeldblum
Differential Revision:
D6264382
Pulled By: Orvid
fbshipit-source-id:
60c8af429f10e778d9c313f40227892238829f63
Yedidya Feldblum [Wed, 8 Nov 2017 17:22:17 +0000 (09:22 -0800)]
Heterogeneous lookups for sorted_vector types
Summary:
[Folly] Heterogeneous lookups for `sorted_vector` types.
When the `Compare` type has member type or alias `is_transparent`, enable template overloads of `count`, `find`, `lower_bound`, `upper_bound`, and `equal_range` on both `sorted_vector_set` and `sorted_vector_map`.
This is the protocol found in the equivalent `std::set` and `std::map` member functions.
> This overload only participates in overload resolution if the qualified-id `Compare::is_transparent` is valid and denotes a type. They allow calling this function without constructing an instance of `Key`.
>
> http://en.cppreference.com/w/cpp/container/set/count (same wording in all 10 cases)
Reviewed By: nbronson
Differential Revision:
D6256989
fbshipit-source-id:
a40a181453a019564e8f7674e1e07e241d5ab068
Andre Pinto [Wed, 8 Nov 2017 16:30:36 +0000 (08:30 -0800)]
Avoid implicitly dropping const modifier
Summary: Use const_cast instead of implicitly dropping const modifier.
Reviewed By: reanimus
Differential Revision:
D6269200
fbshipit-source-id:
61e1708c88a4139d3fdd9cf89f4ff778d0354bb2
Andre Pinto [Wed, 8 Nov 2017 01:26:56 +0000 (17:26 -0800)]
Fix oss build
Summary:
Include folly/Portability.h in GenerateFingerprintTables, so that
gflags namespace gets defined.
Reviewed By: yfeldblum
Differential Revision:
D6265226
fbshipit-source-id:
d509dc163564420151de7007ad4336d7e5ef9625
Igor Sugak [Tue, 7 Nov 2017 23:59:54 +0000 (15:59 -0800)]
fix -Wunused-variable
Summary:
Exposed by the latest clang:
```lang=bash
folly/test/FixedStringTest.cpp:52:36: error: unused variable 's' [-Werror,-Wunused-variable]
constexpr folly::FixedString<42> s{"hello world"};
^
folly/gen/test/StringBenchmark.cpp:176:15: error: unused variable 'line' [-Werror,-Wunused-variable]
StringPiece line(kLine);
^
```
Reviewed By: yfeldblum
Differential Revision:
D6264853
fbshipit-source-id:
5cf4b0a8c99eaa31c2499746c70ddc49fdd73074
Christopher Dykes [Tue, 7 Nov 2017 21:53:30 +0000 (13:53 -0800)]
Workaround a bug in MSVC
Summary:
It was broken, now it is fixed.
https://developercommunity.visualstudio.com/content/problem/145407/incorrect-c3520-within-noexcept-expression-in-alia.html
Reviewed By: yfeldblum
Differential Revision:
D6254219
fbshipit-source-id:
a03961db97d7ac211103655229c1ea703405826a
Adam Simpkins [Tue, 7 Nov 2017 20:11:23 +0000 (12:11 -0800)]
logging: set the thread name for the AsyncFileWriter thread
Summary:
AsyncFileWriter uses a separate thread to write messages to the file
descriptor. This diff makes us call setThreadName() to set the name of this
thread to `log_writer`.
Reviewed By: bolinfest
Differential Revision:
D6238229
fbshipit-source-id:
9c93d80e7ac989e03bc3160bb2f135d67e15c8be
Dave Watson [Tue, 7 Nov 2017 18:27:36 +0000 (10:27 -0800)]
Support movable keys
Summary: Use the same trick as the values, so that non-copyable keys can be used in ConcurrentHashMap.
Reviewed By: yfeldblum
Differential Revision:
D6252711
fbshipit-source-id:
f0f168c4eb361d372bdfc3417f32222d66c11aaf
Dave Watson [Tue, 7 Nov 2017 15:38:44 +0000 (07:38 -0800)]
Remove incorrect DCHECKS
Summary:
While the returned iterators are always 'valid' and can be read and iterated from,
they may still be concurrently erased from the map. Current erase(iter) has DCHECKS that assert we can always
erase an iterator, but it may have already been removed.
Reviewed By: davidtgoldblatt
Differential Revision:
D6221382
fbshipit-source-id:
70b21f53e2fc3daa126df4fb60bc5d3ecb253c71
Giuseppe Ottaviano [Tue, 7 Nov 2017 06:41:52 +0000 (22:41 -0800)]
Allow to specify per-priority capacities in PriorityLifoSemMPMCQueue
Summary: The `THROW` behavior of `LifoSemMPMCQueue` is unsafe when calling `join()`, because the queue may be full and `join()` will fail to enqueue the poisons. To work around this we can use `PriorityLifoSemMPMCQueue` and dedicate `LO_PRI` to the poisons, but there's no reason that the low priority queue should have the same size as the normal priority. Add a constructor to be able to specify different sizes.
Reviewed By: yfeldblum
Differential Revision:
D6257017
fbshipit-source-id:
c75f33c38fcdad646ba1499bcd434ab65711250c
Yedidya Feldblum [Tue, 7 Nov 2017 01:29:12 +0000 (17:29 -0800)]
crange, and range const overloads
Summary:
[Folly] `crange`, and `range` `const` overloads.
Instead of using universal reference for `range` overloads, bifurcate explicitly between `&` and `const&` overloads. The `&` overloads return `Range<T*>` while the `const&` overloads return `Range<T const*>`.
Add `crange` overloads, which may accept non-`const` arguments but will return `Range<T const*>` results anyway.
Reviewed By: ot
Differential Revision:
D6242038
fbshipit-source-id:
bc373c3288ea88792f04b49a372262d12204b586
Yedidya Feldblum [Mon, 6 Nov 2017 23:22:31 +0000 (15:22 -0800)]
Move folly/LifoSem.h
Summary: [Folly] Move `folly/LifoSem.h` to `folly/synchronization/`.
Reviewed By: meyering
Differential Revision:
D6245444
fbshipit-source-id:
14ffa012fa92b8c6aaf5900c930156894a492003
Eric Niebler [Mon, 6 Nov 2017 19:42:48 +0000 (11:42 -0800)]
Value-initialize the Data union member of folly::Function
Summary: The Data field was not getting value-initialized, leading compilers to complain when value-initializing const folly::Function objects.
Reviewed By: yfeldblum, ot
Differential Revision:
D6241712
fbshipit-source-id:
99ce7f6016f6e7d16b1cff7aa51b7bef53ec592d
Phil Willoughby [Mon, 6 Nov 2017 17:06:27 +0000 (09:06 -0800)]
Mark constructing an Unexpected as cold
Summary:
By marking the main constructors of `Unexpected` cold we are telling the compiler that we don't expect to execute them often. This can improve the locality and performance of the `Expected` branch, and reduce the code-size for the `Unexpected` branch.
The constructors we have marked cold here are only the ones used to originate an `Unexpected`. We don't mark the copy- move- or conversion-constructors as `cold` because they will typically be used in error-handling branches and it is not correct to say that they are colder than their surroundings in that case. Note that the entire error-handling branch is likely to be deduced to be cold because we told the compiler that it is likely that hasValue() is true and hasError() is false.
Because `makeUnexpected` unconditionally invokes one of these cold constructors it will inherit the property of being cold.
The GCC function attribute reference, which describes what `cold` means, is here: https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#Common-Function-Attributes
Reviewed By: yfeldblum
Differential Revision:
D6234305
fbshipit-source-id:
6876073e92ce54936ba7152175d76be653f5f463
Yedidya Feldblum [Sun, 5 Nov 2017 04:34:13 +0000 (21:34 -0700)]
Move folly/Functional.h
Summary: [Folly] Move `folly/Functional.h` to `folly/lang/RValueReferenceWrapper.h`.
Reviewed By: Orvid
Differential Revision:
D6242160
fbshipit-source-id:
a8ec42106cfe0ea617c87d382694d5eb2f0f61a0
Yedidya Feldblum [Sat, 4 Nov 2017 22:34:00 +0000 (15:34 -0700)]
constexpr_clamp
Summary:
[Folly] `constexpr_clamp`.
Like `std::clamp` (C++17).
Reviewed By: Orvid
Differential Revision:
D6236825
fbshipit-source-id:
0f6c5dc9a955b148021ee6ed3e86201b53ae090c
Lee Howes [Sat, 4 Nov 2017 02:51:04 +0000 (19:51 -0700)]
Make SemiFuture::via throw on nullptr executor.
Summary: Make SemiFuture throw if no executor provided to via because in that situation the deferred work will never run.
Reviewed By: yfeldblum
Differential Revision:
D6233233
fbshipit-source-id:
43b642c46cc0be80b1f13c03bdaf20b8038acec2
Yedidya Feldblum [Fri, 3 Nov 2017 23:53:23 +0000 (16:53 -0700)]
Move folly/SafeAssert.h
Summary: [Folly] Move `folly/SafeAssert.h` to `folly/lang/`.
Reviewed By: Orvid
Differential Revision:
D6230421
fbshipit-source-id:
0086cd6fedd4ce0e7a4d5302a41153ec1a502e74
Lee Howes [Fri, 3 Nov 2017 23:12:29 +0000 (16:12 -0700)]
Parameter order
Summary: Fix for function parameter order that shows up in opt build on later diff.
Reviewed By: andriigrynenko
Differential Revision:
D6237125
fbshipit-source-id:
fbb7be2c70b32203c658fc239cd74164e01fa1ca
Louis Dionne [Fri, 3 Nov 2017 22:17:09 +0000 (15:17 -0700)]
Nitpick: fix typo in name
Summary: Closes https://github.com/facebook/folly/pull/705
Reviewed By: yfeldblum
Differential Revision:
D6234351
Pulled By: Orvid
fbshipit-source-id:
e71bb3882c783a47ace4d08f134b2e450aaabb6b
Dan Melnic [Fri, 3 Nov 2017 18:16:10 +0000 (11:16 -0700)]
Expose the zerocopy buf ID, change the AsyncSocket fd constructor to accept the id, Buff->Buf, make the cmsghdr methods private
Summary: Expose the zerocopy buf ID, change the AsyncScokey fd constructor to accept the id, Buff->Buf, make the cmsghdr methods private
Reviewed By: djwatson
Differential Revision:
D6221324
fbshipit-source-id:
d0fc4937adf6cf5790d11e406ffd3ec64c558b9c
Lee Howes [Fri, 3 Nov 2017 17:16:04 +0000 (10:16 -0700)]
Missing Future/SemiFuture->Value conversion check
Summary: Conversion check was lost in an earlier refactor. This meant that SemiFuture could be accidentally converted to Future through the value constructor. This should be disabled.
Reviewed By: yfeldblum
Differential Revision:
D6214526
fbshipit-source-id:
3fc2d026ec6062b38b9500c8adf3eee12c0f2693
Yedidya Feldblum [Fri, 3 Nov 2017 02:44:08 +0000 (19:44 -0700)]
Alias std::launder when it is available
Summary: [Folly] Alias `std::launder` when it is available.
Reviewed By: Orvid
Differential Revision:
D6221443
fbshipit-source-id:
33136a8744a39db01fb05513d5ed5476ea67559a
Eric Niebler [Thu, 2 Nov 2017 23:58:07 +0000 (16:58 -0700)]
don't try to run the poly tests on gcc-4.9. they will fail.
Summary: gcc-4.9 does not support features that Poly needs. Disable the tests on that platform.
Reviewed By: meyering
Differential Revision:
D6211674
fbshipit-source-id:
289f029122a45b0f9ec740c62b1faaafb51dcab5
Phil Willoughby [Thu, 2 Nov 2017 22:38:42 +0000 (15:38 -0700)]
Make Expected presume it has a value and not an Error
Summary: This only affects instruction ordering in GCC-compatible compilers to make the value-having branch preferred.
Reviewed By: davidtgoldblatt, nbronson, yfeldblum
Differential Revision:
D6223188
fbshipit-source-id:
57c69b88eda7ee769912874921c45b47ec7a38de
Dan Melnic [Thu, 2 Nov 2017 20:50:33 +0000 (13:50 -0700)]
Change kDefaultZeroCopyThreshold to 0 to avoid a regression and avoid a failure while running as not root
Summary:
Change kDefaultZeroCopyThreshold to 0 to avoid a regression when using a buffer chain that exceeds 32K but each buffer is small.
Change the benchmark to set it's own threshold. Also use calloc vs malloc (in the benchmark only) to get around some weird kernel interaction on non zero copy enabled systems - 2 back to back tests report very different results.
Reviewed By: djwatson
Differential Revision:
D6112299
fbshipit-source-id:
3895d3ece2925c4626284ff364495708293edc3e
Andrey Ignatov [Thu, 2 Nov 2017 17:22:09 +0000 (10:22 -0700)]
Introduce non-throwing try* methods for IPAddress{,V4,V6}.
Summary:
Now there is no interface to create `IPAddress{,V4,V6}` from a string or
`ByteRange` that doesn't throw. All available static methods throw
`IPAddressFormatException`.
It has a few disadvantages:
== 1. It's unsafe ==
Caller is not forced to catch exception, it's very easy to write
`IPAddress(potentiallyInvalidString)` and discover that it can throw when it's
already in prod.
== 2. It's inconvenient ==
if caller is aware about exception, (s)he's forced to write `try {} catch` that
is inconvenient and leads to code like this:
folly::IPAddress taskIp;
try {
taskIp = folly::IPAddress(kv.second.taskIp);
} catch (const folly::IPAddressFormatException&) {
// Error handling ..
}
// Use IP ..
== 3. It's expensive ==
Amended benchmark shows that `IPAddress` constructor is ~10 times slower when a
string with invalid IP is passed to it.
---
The diff introduces two non-throwing interfaces for all tree flavors of `IPAddress`:
`tryFromString()` tries to create IP address from string and returns either
corresponding IP address or `enum class IPAddressFormatError` using
`folly::Expected`.
`tryFromBinary()` does same thing but for `ByteRange` input.
The code can be as short as:
if (auto maybeIp = IPAddress::tryFromString(ipStr)) {
// Use maybeIp.value() ..
}
The `try` prefix is chosen to be consistent with other interfaces in folly,
e.g. `to` and `tryTo`.
Reviewed By: yfeldblum
Differential Revision:
D6211182
fbshipit-source-id:
f27cf90997c100a5fd42138e66ff9bb172204c20
Yedidya Feldblum [Thu, 2 Nov 2017 16:55:29 +0000 (09:55 -0700)]
Refactor is_simple_allocator and callers
Summary:
[Folly] Refactor `is_simple_allocator` and callers.
* Swap order of template parameters.
* Do decaying in the callers instead.
* Do a direct invocability test, rather than an indirect test of whether the allocator has a method `destroy` with the expected signature.
Reviewed By: ericniebler
Differential Revision:
D6184062
fbshipit-source-id:
aec32e6e323b8c6023b94c258ab2bcddd8c53e09
Christopher Dykes [Thu, 2 Nov 2017 03:33:39 +0000 (20:33 -0700)]
Fix folly::Function under C++17 exception specifier rules
Summary: Under C++17's exception specifier rules (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/p0012r1.html) we need additional specializations to achieve the desired effect.
Reviewed By: yfeldblum
Differential Revision:
D6214359
fbshipit-source-id:
a007c2fcea1496bdfe4fdf923b39c1611c6ad9bc
Ognjen Dragoljevic [Wed, 1 Nov 2017 22:54:38 +0000 (15:54 -0700)]
Make associative container out-of-range exception provide missing key
Summary:
When the key is missing, the standard associative containers throw an exception like `unordered_map::at: key not found`. That's nice, but it would be even better if we would actually know what key is missing. This is not an issue when one is accessing the container directly, but when that access happens several levels deep, such as here within folly formatting logic, we have no way of knowing what key was missing. This poses some difficulties in presenting error to the user.
This change makes folly format throw a subclass of `std::out_of_range` exception on a missing key when a string keyed associative container is used for providing parameters. That subclass stores the actual key used so it can be accessed in the exception handler. Existing callers can still catch `std::out_of_range` so they should be unaffected by this change.
Reviewed By: ot, yfeldblum
Differential Revision:
D6202184
fbshipit-source-id:
b8a6740aaccc5d8914ad7d099c8b901427f00083
Lee Howes [Wed, 1 Nov 2017 22:46:19 +0000 (15:46 -0700)]
Quick comment on DE to clarify use cases and lack of thread safety.
Summary: (Note: this ignores all push blocking failures!)
Reviewed By: yfeldblum
Differential Revision:
D6212919
fbshipit-source-id:
b066d99e13e97cbab489132bcb91872ed4407f81
Yedidya Feldblum [Wed, 1 Nov 2017 21:11:27 +0000 (14:11 -0700)]
Fix unsynchronized accesses in IOThreadPoolExecutor::getEventBase
Summary:
[Folly] Fix unsynchronized accesses in `IOThreadPoolExecutor::getEventBase`.
`getEventBase` may be invoked concurrently from two threads - RMWs to `nextThread_` must be synchronized with each other.
`getEventBase` may be invoked concurrently with `setNumThreads` - the former's reads of `threadList_.vec_` must be synchronized with the latter's writes to it.
Reviewed By: kennyyu
Differential Revision:
D6206916
fbshipit-source-id:
8bfae158effb5896ab478d0c20310293b037c892
Eric Niebler [Wed, 1 Nov 2017 19:50:41 +0000 (12:50 -0700)]
add a complete example to poly's docs, fix typos and think-o's
Summary: Address doc criticisms
Reviewed By: Orvid
Differential Revision:
D6210376
fbshipit-source-id:
fee11cef1c407b092f891a97f94271a81d3718b8
Alisson Gusatti Azzolini [Wed, 1 Nov 2017 19:11:16 +0000 (12:11 -0700)]
Fix OSS's "make check"
Reviewed By: yfeldblum
Differential Revision:
D6182541
fbshipit-source-id:
31d255819df1f97b13e475903c0627a1ac96b516
Maged Michael [Wed, 1 Nov 2017 14:47:41 +0000 (07:47 -0700)]
Add integrated reference counting
Summary:
Add support for reference counting integrated with the internal structures and operations of the hazard pointer library. The operations are wait-free.
The advantages of this approach over combining reference counting with hazard pointers externally are:
(1) A long list of linked objects that protected by one reference can all be reclaimed together instead of going through a potentially long series of alternating reclamation and calls to retire() for descendants.
(2) Support for iterative deletion as opposed to potential deep recursion of alternating calls to release reference count and object destructors.
Reviewed By: djwatson
Differential Revision:
D6142066
fbshipit-source-id:
02bdfcbd5a2c2d5486d937bb2f9cfb6f192f5e1a
Andrii Grynenko [Wed, 1 Nov 2017 06:06:22 +0000 (23:06 -0700)]
Clear frame cache when activating a fiber
Reviewed By: yfeldblum
Differential Revision:
D6207160
fbshipit-source-id:
57468c9d05cdb3ee6e1d10a3a254a5d1bfddc36f
Yedidya Feldblum [Wed, 1 Nov 2017 05:12:53 +0000 (22:12 -0700)]
A macro for creating member-invoke traits
Summary:
[Folly] A macro for creating member-invoke traits.
The macro creates a specialized traits container with member types and aliases mimicking `std::invoke_result` and the related traits types and aliases.
Reviewed By: aary
Differential Revision:
D6195087
fbshipit-source-id:
07c2bbab6cccb04dc8ff12e20923351e8f38abfd
Lee Howes [Tue, 31 Oct 2017 21:20:08 +0000 (14:20 -0700)]
Adding DeferredExecutor to support deferred execution of tasks on a future returned from an interface.
Summary: This adds a DeferredExecutor type that is boostable, which means that it follows the expectation we expect for C++20 that .then and get will trigger boost-blocking behaviour and ensure work makes progress. Unlike discussions for C++ this adds boost blocking to folly only in the specific case of deferring work to run on the caller's executor, to avoid the necessity to pass an executor into a library purely to ensure that finalisation work and future completion occor on a well-defined exewcutor.
Reviewed By: yfeldblum
Differential Revision:
D5828743
fbshipit-source-id:
9a4b69d7deaa33c3cecd6546651b99cc99f0c286
Martin Martin [Tue, 31 Oct 2017 21:09:59 +0000 (14:09 -0700)]
Remove unused field or local var.
Summary: Remove unused field or local var.
Reviewed By: terrelln
Differential Revision:
D6199120
fbshipit-source-id:
616a2b2549c37bcb57d2f8c530b26089f24c2973
Teng Qin [Tue, 31 Oct 2017 20:49:15 +0000 (13:49 -0700)]
Improve folly::RequestContext onSet and onUnset efficiency
Summary:
In previous discussions, it has been pointed out that `folly::RequestContext::setContext` consumes considerable amount of CPU cycles in production environments. After some investigation, we thought that might be caused by looping over all `RequestData` instances can calling the virtual `onSet` and `onUnset` callback of them. Both the iteration and invoking virtual methods are not cheap.
As you can see from this change, most of the derived classes of `RequestData` don't override the `onSet` and `onUnset` methods. Mostly likely they are only used for per-Request tracking. So the natural idea is to skip those instances when iterating and avoid the unnecessary virtual method invoke.
I have explored the solution to dynamically examine if the `RequestData` instance added has `onSet` and `onUnset` method overridden. That is possible with GCC's PMF extension, but not [[ http://lists.llvm.org/pipermail/llvm-bugs/2015-July/041164.html | for Clang ]] and probably many other compilers. This definitely won't be very good for `folly`'s probability, even if we gate it by compiler flags.
Therefore, this Diff adds the `hasCallback` method to `RequestData` class indicating whether the instance would have `onSet` and `onUnset` overridden. To make it clear to users that they need to correctly override it in order for their `onSet` and `onUnset` callback to work, making it abstract so that user must override it to something and would aware of that.
Also made some improvements on documentation.
Reviewed By: myreg
Differential Revision:
D6144049
fbshipit-source-id:
4c9fd72e9efaeb6763d55f63760eaf582ee4839e
Teng Qin [Tue, 31 Oct 2017 20:49:14 +0000 (13:49 -0700)]
Improve folly::RequestContext::get()
Summary:
Since `folly::RequestContext::get()` returns raw pointer, it could directly use the reference returned by `getStaticContext()`
I don't expect this to make much of a difference, just tiny improvements
Reviewed By: yfeldblum
Differential Revision:
D6153353
fbshipit-source-id:
1c41d4fc259aa5cb3e69e50ed24bed1ba9caf6c3
Alex Guzman [Tue, 31 Oct 2017 06:38:41 +0000 (23:38 -0700)]
Add utility function for loading certificates from a buffer
Summary: Adds a function that reads certificates in from a buffer and returns them as a vector of X509 pointers.
Reviewed By: yfeldblum
Differential Revision:
D6133332
fbshipit-source-id:
eaaaffcbd4d03f37d9d5b4c99a52b0d968b163ba
Yedidya Feldblum [Tue, 31 Oct 2017 05:00:27 +0000 (22:00 -0700)]
Outline most throw expressions in Expected
Summary:
[Folly] Outline most `throw` expressions in `Expected`.
They are definitionally cold, but in-line `throw` statements can expand code size more than is desirable.
* Inline `throw` statement: https://godbolt.org/g/LPaf7V.
* Outline `throw` statement: https://godbolt.org/g/HZBXn6.
Reviewed By: Orvid
Differential Revision:
D6183613
fbshipit-source-id:
28240bb4aa40790d99da783a3c368db81fded124
Yedidya Feldblum [Tue, 31 Oct 2017 05:00:25 +0000 (22:00 -0700)]
Remove dep on Format.h from GenerateFingerprintTables.cpp
Summary:
[Folly] Remove dep on `Format.h` from `GenerateFingerprintTables.cpp`.
`GenerateFingerprintTables.cpp` is a tool used to generate other sources which get built as part of the main library when using the autotools build, so it must be as free of other Folly dependencies as possible.
Reviewed By: ot, Orvid
Differential Revision:
D6183725
fbshipit-source-id:
f12b18553c78e085599a5505ae57f12bc0cd44b0
Yedidya Feldblum [Tue, 31 Oct 2017 05:00:24 +0000 (22:00 -0700)]
Remove folly/ContainerTraits.h
Summary:
[Folly] Remove `folly/ContainerTraits.h`.
It has some handly helpers, but it is not a real abstraction or utility library. Within Folly, only `folly/Padded.h` uses it, so just rewrite the bit that needs it.
Reviewed By: LeeHowes
Differential Revision:
D6183066
fbshipit-source-id:
24a223fe517d21ff531e0fa80172f15d4f963e51
Yedidya Feldblum [Tue, 31 Oct 2017 00:57:54 +0000 (17:57 -0700)]
Move folly/experimental/AsymmetricMemoryBarrier.h
Summary: [Folly] Move `folly/experimental/AsymmetricMemoryBarrier.h` to `folly/synchronization/AsymmetricMemoryBarrier.h`.
Reviewed By: Orvid
Differential Revision:
D6180676
fbshipit-source-id:
f4833318cd365181e202d5f379815e728fba168b
Christopher Dykes [Mon, 30 Oct 2017 23:01:31 +0000 (16:01 -0700)]
Fix the CMake build
Summary: By excluding the tools directory, excluding poly, and moving all the tests that have moved around to their new homes.
Reviewed By: yfeldblum
Differential Revision:
D6191644
fbshipit-source-id:
bdb39d01a796c1e52257200c0d411a4cb44116ce
Felix Leupold [Mon, 30 Oct 2017 22:26:55 +0000 (15:26 -0700)]
Allow to pass ObjC blocks into folly::Function
Summary:
In iOS blocks are initially allocated on the stack and only lazily copied to the heap (e.g when they are assigned to a variable). This means, if you pass a block as an rvalue to a C++ method that keeps moving it around instead of copy assigning it at some point, the block remains on the stack and will get freed once the original method is done (leading to use after free if the block is executed later).
This was mitigated by deleting the conversion from ObjC functions to folly functions. Given that all we need is to make sure that the block is allocated on the heap (that is it is an instance of NSMallocBlock rather than NSStackBlock), it seems drastic to ban the conversion. ObjC developers tend to be more familiar with ObjC blocks and will find it convenient to use this conversion.
This diff insteads implements the constructor and assignment operator by wrapping the ObjC block in a c++ lambda and capturing it by copy. ARC keeps track of the reference count and automatically releases the block when the lambda is deallocated. Moreover, copy only increase the retain count (instead of doing an actual copy) if the block was already stored on the heap (https://www.cocoawithlove.com/2009/10/how-blocks-are-implemented-and.html section NSMallocBlock never actually copies).
Reviewed By: ericniebler
Differential Revision:
D6109932
fbshipit-source-id:
48bb446d3a66f46affba774cfe1cfb8a60c661de
Yedidya Feldblum [Mon, 30 Oct 2017 21:52:50 +0000 (14:52 -0700)]
Simplify type_t
Summary:
[Folly] Simplify `type_t` by lifting the type to be aliased into the structure template parameter list.
May also fix curious build failures in some compilers.
Reviewed By: akrieger
Differential Revision:
D6188953
fbshipit-source-id:
96e1c3af9c11959c0899c092933158922efa7e60
Lee Howes [Mon, 30 Oct 2017 21:28:33 +0000 (14:28 -0700)]
Split SemiFuture and Future into separate types. Add BasicFuture shared between them.
Summary:
To avoid the risk of bugs caused by a Future being cast to a SemiFuture, and losing some of the properties in the process, this splits SemiFuture and Future into unrelated types, sharing a private superclass for code reuse.
* Add BasicFuture in futures::detail
* Make superclass privately inherited.
* Unset executor when constructing SemiFuture from Future.
Reviewed By: yfeldblum
Differential Revision:
D6177780
fbshipit-source-id:
dea3116aeec0572bb973c2a561e17785199e86f2
Arkady Shapkin [Mon, 30 Oct 2017 19:15:00 +0000 (12:15 -0700)]
Mention Windows (Vcpkg) build in README.md
Summary: Closes https://github.com/facebook/folly/pull/697
Reviewed By: yfeldblum
Differential Revision:
D6144274
Pulled By: Orvid
fbshipit-source-id:
a79a2e36e8fcf271925e97ece2a6adbb3c074216
Eric Niebler [Mon, 30 Oct 2017 16:26:13 +0000 (09:26 -0700)]
Folly.Poly: a library for creating type-erasing polymorphic wrappers
Summary:
`Poly` is a class template that makes it relatively easy to define a type-erasing polymorphic object wrapper.
== Type-erasure
`std::function` is one example of a type-erasing polymorphic object wrapper;
`folly::exception_wrapper` is another. Type-erasure is often used as an
alternative to dynamic polymorphism via inheritance-based virtual dispatch.
The distinguishing characteristic of type-erasing wrappers are:
* **Duck typing:** Types do not need to inherit from an abstract base
class in order to be assignable to a type-erasing wrapper; they merely
need to satisfy a particular interface.
* **Value semantics:** Type-erasing wrappers are objects that can be
passed around _by value_. This is in contrast to abstract base classes
which must be passed by reference or by pointer or else suffer from
_slicing_, which causes them to lose their polymorphic behaviors.
Reference semantics make it difficult to reason locally about code.
* **Automatic memory management:** When dealing with inheritance-based
dynamic polymorphism, it is often necessary to allocate and manage
objects on the heap. This leads to a proliferation of `shared_ptr`s and
`unique_ptr`s in APIs, complicating their point-of-use. APIs that take
type-erasing wrappers, on the other hand, can often store small objects
in-situ, with no dynamic allocation. The memory management, if any, is
handled for you, and leads to cleaner APIs: consumers of your API don't
need to pass `shared_ptr<AbstractBase>`; they can simply pass any object
that satisfies the interface you require. (`std::function` is a
particularly compelling example of this benefit. Far worse would be an
inheritance-based callable solution like
`shared_ptr<ICallable<void(int)>>`. )
== Example: Defining a type-erasing function wrapper with `folly::Poly`
Defining a polymorphic wrapper with `Poly` is a matter of defining two
things:
* An *interface*, consisting of public member functions, and
* A *mapping* from a concrete type to a set of member function bindings.
Below is a (heavily commented) example of a simple implementation of a
`std::function`-like polymorphic wrapper. Its interface has only a single
member function: `operator()`
lang=c++
// An interface for a callable object of a particular signature, Fun
// (most interfaces don't need to be templates, FWIW).
template <class Fun>
struct IFunction;
template <class R, class... As>
struct IFunction<R(As...)> {
// An interface is defined as a nested class template called
// Interface that takes a single template parameter, Base, from
// which it inherits.
template <class Base>
struct Interface : Base {
// The Interface has public member functions. These become the
// public interface of the resulting Poly instantiation.
// (Implementation note: Poly<IFunction<Sig>> will publicly
// inherit from this struct, which is what gives it the right
// member functions.)
R operator()(As... as) const {
// The definition of each member function in your interface will
// always consist of a single line dispatching to folly::call<N>.
// The "N" corresponds to the N-th member function in the
// list of member function bindings, Members, defined below.
// The first argument will always be *this, and the rest of the
// arguments should simply forward (if necessary) the member
// function's arguments.
return static_cast<R>(
folly::poly_call<0>(*this, std::forward<As>(as)...));
}
};
// The "Members" alias template is a comma-separated list of bound
// member functions for a given concrete type "T". The
// "FOLLY_POLY_MEMBERS" macro accepts a comma-separated list, and the
// (optional) "FOLLY_POLY_MEMBER" macro lets you disambiguate overloads
// by explicitly specifying the function signature the target member
// function should have. In this case, we require "T" to have a
// function call operator with the signature `R(As...) const`.
//
// If you are using a C++17-compatible compiler, you can do away with
// the macros and write this as:
//
// template <class T>
// using Members =
// folly::PolyMembers<folly::sig<R(As...) const>(&T::operator())>;
//
// And since `folly::sig` is only needed for disambiguation in case of
// overloads, if you are not concerned about objects with overloaded
// function call operators, it could be further simplified to:
//
// template <class T> using Members = folly::PolyMembers<&T::operator()>;
//
template <class T>
using Members = FOLLY_POLY_MEMBERS(
FOLLY_POLY_MEMBER(R(As...) const, &T::operator()));
};
// Now that we have defined the interface, we can pass it to Poly to
// create our type-erasing wrapper:
template <class Fun>
using Function = Poly<IFunction<Fun>>;
Given the above definition of `Function`, users can now initialize instances
of (say) `Function<int(int, int)>` with function objects like
`std::plus<int>` and `std::multiplies<int>`, as below:
lang=c++
Function<int(int, int)> fun = std::plus<int>{};
assert(5 == fun(2, 3));
fun = std::multiplies<int>{};
assert(6 = fun(2, 3));
== Defining an interface with C++17
With C++17, defining an interface to be used with `Poly` is fairly
straightforward. As in the `Function` example above, there is a struct with
a nested `Interface` class template and a nested `Members` alias template.
No macros are needed with C++17.
Imagine we were defining something like a Java-style iterator. If we are
using a C++17 compiler, our interface would look something like this:
lang=c++
template <class Value>
struct IJavaIterator {
template <class Base>
struct Interface : Base {
bool Done() const { return folly::poly_call<0>(*this); }
Value Current() const { return folly::poly_call<1>(*this); }
void Next() { folly::poly_call<2>(*this); }
};
// NOTE: This works in C++17 only:
template <class T>
using Members = folly::PolyMembers<&T::Done, &T::Current, &T::Next>;
};
template <class Value>
using JavaIterator = Poly<IJavaIterator>;
Given the above definition, `JavaIterator<int>` can be used to hold instances
of any type that has `Done`, `Current`, and `Next` member functions with the
correct (or compatible) signatures.
The presence of overloaded member functions complicates this picture. Often,
property members are faked in C++ with `const` and non-`const` member
function overloads, like in the interface specified below:
lang=c++
struct IIntProperty {
template <class Base>
struct Interface : Base {
int Value() const { return folly::poly_call<0>(*this); }
void Value(int i) { folly::poly_call<1>(*this, i); }
};
// NOTE: This works in C++17 only:
template <class T>
using Members = folly::PolyMembers<
folly::sig<int() const>(&T::Value),
folly::sig<void(int)>(&T::Value)>;
};
using IntProperty = Poly<IIntProperty>;
Now, any object that has `Value` members of compatible signatures can be
assigned to instances of `IntProperty` object. Note how `folly::sig` is used
to disambiguate the overloads of `&T::Value`.
== Defining an interface with C++14
In C++14, the nice syntax above doesn't work, so we have to resort to macros.
The two examples above would look like this:
lang=c++
template <class Value>
struct IJavaIterator {
template <class Base>
struct Interface : Base {
bool Done() const { return folly::poly_call<0>(*this); }
Value Current() const { return folly::poly_call<1>(*this); }
void Next() { folly::poly_call<2>(*this); }
};
// NOTE: This works in C++14 and C++17:
template <class T>
using Members = FOLLY_POLY_MEMBERS(&T::Done, &T::Current, &T::Next);
};
template <class Value>
using JavaIterator = Poly<IJavaIterator>;
and
lang=c++
struct IIntProperty {
template <class Base>
struct Interface : Base {
int Value() const { return folly::poly_call<0>(*this); }
void Value(int i) { return folly::poly_call<1>(*this, i); }
};
// NOTE: This works in C++14 and C++17:
template <class T>
using Members = FOLLY_POLY_MEMBERS(
FOLLY_POLY_MEMBER(int() const, &T::Value),
FOLLY_POLY_MEMBER(void(int), &T::Value));
};
using IntProperty = Poly<IIntProperty>;
== Extending interfaces
One typical advantage of inheritance-based solutions to runtime polymorphism
is that one polymorphic interface could extend another through inheritance.
The same can be accomplished with type-erasing polymorphic wrappers. In
the `Poly` library, you can use `folly::PolyExtends` to say that one interface
extends another.
lang=c++
struct IFoo {
template <class Base>
struct Interface : Base {
void Foo() const { return folly::poly_call<0>(*this); }
};
template <class T>
using Members = FOLLY_POLY_MEMBERS(&T::Foo);
};
// The IFooBar interface extends the IFoo interface
struct IFooBar : PolyExtends<IFoo> {
template <class Base>
struct Interface : Base {
void Bar() const { return folly::poly_call<0>(*this); }
};
template <class T>
using Members = FOLLY_POLY_MEMBERS(&T::Bar);
};
using FooBar = Poly<IFooBar>;
Given the above definition, instances of type `FooBar` have both `Foo()` and
`Bar()` member functions.
The sensible conversions exist between a wrapped derived type and a wrapped
base type. For instance, assuming `IDerived` extends `IBase` with `Extends`:
lang=c++
Poly<IDerived> derived = ...;
Poly<IBase> base = derived; // This conversion is OK.
As you would expect, there is no conversion in the other direction, and at
present there is no `Poly` equivalent to `dynamic_cast`.
== Type-erasing polymorphic reference wrappers
Sometimes you don't need to own a copy of an object; a reference will do. For
that you can use `Poly` to capture a //reference// to an object satisfying an
interface rather than the whole object itself. The syntax is intuitive.
lang=c++
int i = 42;
// Capture a mutable reference to an object of any IRegular type:
Poly<IRegular &> intRef = i;
assert(42 == folly::poly_cast<int>(intRef));
// Assert that we captured the address of "i":
assert(&i == &folly::poly_cast<int>(intRef));
A reference-like `Poly` has a different interface than a value-like `Poly`.
Rather than calling member functions with the `obj.fun()` syntax, you would
use the `obj->fun()` syntax. This is for the sake of `const`-correctness.
For example, consider the code below:
lang=c++
struct IFoo {
template <class Base>
struct Interface {
void Foo() { folly::poly_call<0>(*this); }
};
template <class T>
using Members = folly::PolyMembers<&T::Foo>;
};
struct SomeFoo {
void Foo() { std::printf("SomeFoo::Foo\n"); }
};
SomeFoo foo;
Poly<IFoo &> const anyFoo = foo;
anyFoo->Foo(); // prints "SomeFoo::Foo"
Notice in the above code that the `Foo` member function is non-`const`.
Notice also that the `anyFoo` object is `const`. However, since it has
captured a non-`const` reference to the `foo` object, it should still be
possible to dispatch to the non-`const` `Foo` member function. When
instantiated with a reference type, `Poly` has an overloaded `operator->`
member that returns a pointer to the `IFoo` interface with the correct
`const`-ness, which makes this work.
The same mechanism also prevents users from calling non-`const` member
functions on `Poly` objects that have captured `const` references, which
would violate `const`-correctness.
Sensible conversions exist between non-reference and reference `Poly`s. For
instance:
lang=c++
Poly<IRegular> value = 42;
Poly<IRegular &> mutable_ref = value;
Poly<IRegular const &> const_ref = mutable_ref;
assert(&poly_cast<int>(value) == &poly_cast<int>(mutable_ref));
assert(&poly_cast<int>(value) == &poly_cast<int>(const_ref));
== Non-member functions (C++17)
If you wanted to write the interface `ILogicallyNegatable`, which captures
all types that can be negated with unary `operator!`, you could do it
as we've shown above, by binding `&T::operator!` in the nested `Members`
alias template, but that has the problem that it won't work for types that
have defined unary `operator!` as a free function. To handle this case,
the `Poly` library lets you use a free function instead of a member function
when creating a binding.
With C++17 you may use a lambda to create a binding, as shown in the example
below:
lang=c++
struct ILogicallyNegatable {
template <class Base>
struct Interface : Base {
bool operator!() const { return folly::poly_call<0>(*this); }
};
template <class T>
using Members = folly::PolyMembers<
+[](T const& t) -> decltype(!t) { return !t; }>;
};
This requires some explanation. The unary `operator+` in front of the lambda
is necessary! It causes the lambda to decay to a C-style function pointer,
which is one of the types that `folly::PolyMembers` accepts. The `decltype` in
the lambda return type is also necessary. Through the magic of SFINAE, it
will cause `Poly<ILogicallyNegatable>` to reject any types that don't support
unary `operator!`.
If you are using a free function to create a binding, the first parameter is
implicitly the `this` parameter. It will receive the type-erased object.
== Non-member functions (C++14)
If you are using a C++14 compiler, the definition of `ILogicallyNegatable`
above will fail because lambdas are not `constexpr`. We can get the same
effect by writing the lambda as a named free function, as show below:
lang=c++
struct ILogicallyNegatable {
template <class Base>
struct Interface : Base {
bool operator!() const { return folly::poly_call<0>(*this); }
};
template <class T>
static auto negate(T const& t) -> decltype(!t) { return !t; }
template <class T>
using Members = FOLLY_POLY_MEMBERS(&negate<T>);
};
As with the example that uses the lambda in the preceding section, the first
parameter is implicitly the `this` parameter. It will receive the type-erased
object.
== Multi-dispatch
What if you want to create an `IAddable` interface for things that can be
added? Adding requires //two// objects, both of which are type-erased. This
interface requires dispatching on both objects, doing the addition only
if the types are the same. For this we make use of the `Self` template
alias to define an interface that takes more than one object of the the
erased type.
lang=c++
struct IAddable {
template <class Base>
struct Interface : Base {
friend Self<Base>
operator+(Self<Base> const& a, Self<Base> const& b) const {
return folly::poly_call<0>(a, b);
}
};
template <class T>
using Members = folly::PolyMembers<
+[](T const& a, T const& b) -> decltype(a + b) { return a + b; }>;
};
Given the above defintion of `IAddable` we would be able to do the following:
lang=c++
Poly<IAddable> a = 2, b = 3;
Poly<IAddable> c = a + b;
assert(poly_cast<int>(c) == 5);
If `a` and `b` stored objects of different types, a `BadPolyCast` exception
would be thrown.
== Move-only types
If you want to store move-only types, then your interface should extend the
`IMoveOnly` interface.
== Implementation notes
`Poly` will store "small" objects in an internal buffer, avoiding the cost of
of dynamic allocations. At present, this size is not configurable; it is
pegged at the size of two `double`s.
`Poly` objects are always nothrow movable. If you store an object in one that
has a potentially throwing move contructor, the object will be stored on the
heap, even if it could fit in the internal storage of the `Poly` object.
(So be sure to give your objects nothrow move constructors!)
`Poly` implements type-erasure in a manner very similar to how the compiler
accomplishes virtual dispatch. Every `Poly` object contains a pointer to a
table of function pointers. Member function calls involve a double-
indirection: once through the v-pointer, and other indirect function call
through the function pointer.
Reviewed By: yfeldblum
Differential Revision:
D4897112
fbshipit-source-id:
ff1c1156316bfbdd8f2205c4f57932c0067cacac
Alexander Pronchenkov [Mon, 30 Oct 2017 12:13:02 +0000 (05:13 -0700)]
Remove a few memory allocations in ThreadWheelTimekeeper.after()
Summary:
This diff reduces number of memory allocation in folly::ThreadWheelTimekeeper.after() method for a bit.
* std::shared_ptr(new T) is replaced with std::make_shared<T>()
* folly::Promise is stored by value
Reviewed By: yfeldblum
Differential Revision:
D6172017
fbshipit-source-id:
41bf123f10570c76d64eaac1800b7e65fe381110
Murali Vilayannur [Mon, 30 Oct 2017 05:07:46 +0000 (22:07 -0700)]
Override TemporaryFile's default move constructor
Summary:
A default move constructor/move assignment operator caused a bunch
of bugs in my test that was depending on the move constructor to either
not close the underlying file descriptor or to ensure that the TemporaryFile
object doesn't have a dangling reference to a file descriptor that has already
been closed. The current implementation caused both the moved' from and to
object to share the same underlying file descriptor and when the former object
is destroyed it ends up closing the file descriptor leaving the latter
with a dangling file descriptor which could be reused for some other
socket(s)/file descriptor by the kernel and ends up causing seg-faults
when the latter object is eventually destroyed due to it releasing
a file descriptor that now belongs to some other resource.
I changed the move constructor/move assignment operator
to have the former semantics to not close the underlying file descriptor of
the move'd from object (by releasing it and assigning it to the move'd to
object). I am not sure if anyone would ever want the alternative
semantics because the TemporaryFile object would not be usable
without a valid underlying file descriptor
Reviewed By: yfeldblum
Differential Revision:
D6182075
fbshipit-source-id:
bc809d704449c1c1182d76cdfa2f7d17b38a719a