Summary:
When a user counter is added to benchmarks, the counter's value is written as a bare integer.
This change gives chance to user to either still keep it reported as bare integer, a time value or a metric value.

Differential Revision: D14451341

fbshipit-source-id: 95e64a89c201bb385d9a403f3442fa70c2ac7e2a

Summary: Make AsyncIOTest.cpp conditionally depend on $LIBAIO_FOUND in CMakeLists.txt.

Reviewed By: yfeldblum

Differential Revision: D14365682

fbshipit-source-id: 5ac7dc25174ab48ed1de8846956e69c7759fc9bd

Summary: convert all executors in pushmi and all executor concepts and usage in pushmi to use `schedule()`

Reviewed By: ericniebler

Differential Revision: D13463902

fbshipit-source-id: 01eac4fe0b5cd49afbd95118db769db59a3b968c

Summary: It's dead.

Reviewed By: yfeldblum

Differential Revision: D14425042

fbshipit-source-id: 098de614955eedd7ccc733858db85b997d01d9e0

Summary: At least make it possible to do things correctly.

Reviewed By: yfeldblum, LeeHowes

Differential Revision: D14426753

fbshipit-source-id: 63b3572175224dbbda828981a5c9f067a5e35f33

Summary: Fix typo

Reviewed By: aary

Differential Revision: D14411090

fbshipit-source-id: 4ca0273201facebc815db26d23662d0d192c831f

Summary:
[Folly] Move `FOLLY_DEFINE_CPO` to `folly/lang/CustomizationPoint.h`.

And move `StaticConst` to top-level `folly` namespace and into `folly/lang/`.

And Americanize the spellings of words in the comments.

Reviewed By: ericniebler, lewissbaker

Differential Revision: D14402315

fbshipit-source-id: e13dbc9d471ee73ba0a289d375c235734334b920

Summary:
future.then([] -> T)  currently returns Future<T>
future.then([] -> Try<T>)  currently returns Future<Try<T>>

This changes that behavior, such that future.then([] -> Try<T>) will
return Future<T>.

This is a more desirable interface for futures. It also simplifies
implementation of a separate task (propagating exceptions more
efficiently through chained callbacks) because it enables us to internally
substitute lambdas returning T with lambdas returning Try<T>.

Reviewed By: yfeldblum, ryantimwilson

Differential Revision: D14318624

fbshipit-source-id: 51363e82ceb86bfecf87ae661188be9ca3dd8002

Summary: Instead prefer the NetworkSocket version

Reviewed By: yfeldblum

Differential Revision: D14371075

fbshipit-source-id: 5dfa34e21b360cae987b1ffb1dd5f5ee7b105623

Summary:
- Instead of defining our own version of whether a type is nothrow
  swappable, reuse the existing definition from `folly/Traits.h` which
  is spelled `IsNothrowSwappable`.
Pull Request resolved: https://github.com/facebook/folly/pull/1049

Reviewed By: Orvid

Differential Revision: D14402098

Pulled By: yfeldblum

fbshipit-source-id: 5b61bd6d43db939b4499e908b26a3eb61f8929a3

Summary:
- Remove `remove_cvref_t` alias template from `pusmhi/traits.h` in favor of
  `folly::remove_cvref_t` from `folly/Traits.h` which is the same
  definition.
Pull Request resolved: https://github.com/facebook/folly/pull/1048

Reviewed By: ericniebler

Differential Revision: D14402040

Pulled By: yfeldblum

fbshipit-source-id: baba640e5fa410e0c48e666ea004de61b7f12a12

Summary:
- Replace `Not` template alias from `PolyDetail.h` with `folly::Negation` from
 `folly/Traits.h` which does the same thing.
Pull Request resolved: https://github.com/facebook/folly/pull/1045

Reviewed By: ericniebler

Differential Revision: D14402105

Pulled By: yfeldblum

fbshipit-source-id: 100347dbdc159c5217c7452b202650d58d80850c

Summary: Add timerfd based timers

Reviewed By: djwatson

Differential Revision: D13809490

fbshipit-source-id: 9ae70d9cc4d481245623efbc3e74ce5ff8c8f3ca

fbshipit-source-id: 4c8ab4cdad047a4311c84b302b28160dd2bc9ed4

Summary:
- In `c2e51a8e`, i.e.
  Differential Revision: D9213429
  fbshipit-source-id: 849c0fce01d4903d3f8a7b6d196037bf950e5921
  a `libfolly.pc.in` file was placed in the `CMake` directory so that
  the `libfolly.pc` file is installed during CMake configure time.
- Note that the `libfolly.pc.in` in the `folly` directory is not
  referenced anywhere in the build tree, and it has been superseded with
  the one in the `CMake` directory. As such, remove it.
Pull Request resolved: https://github.com/facebook/folly/pull/1050

Reviewed By: calebmarchent

Differential Revision: D14400016

Pulled By: yfeldblum

fbshipit-source-id: 4c8ab4cdad047a4311c84b302b28160dd2bc9ed4

Summary: [Folly] Cut glog include in `folly/Range.h`.

Reviewed By: meyering

Differential Revision: D14402324

fbshipit-source-id: ed37939d69b2de4016483e7a2564d73892a64b3b

Summary:
Some tests fail to compile because gflags.h is not included:

    .../folly/test/FormatOtherTest.cpp: In function 'int main(int, char**)':
    .../folly/test/FormatOtherTest.cpp:113:3: error: 'gflags' has not been declared
       gflags::ParseCommandLineFlags(&argc, &argv, true);
       ^~~~~~

The tests compile without issue on Travis CI without this patch, but I
don't see how that's possible.

Include gflags.h explicitly to fix the compilation of these tests.
Pull Request resolved: https://github.com/facebook/folly/pull/1051

Reviewed By: aary

Differential Revision: D14400709

Pulled By: yfeldblum

fbshipit-source-id: 73999a940f45511373d3922ea8eb782a27af0008

Summary:
When a socket is configured to read from `MSG_ERRQUEUE`, the corresponding callback handler for `errMessageCallback_->errMessage(*cmsg)` may close the socket when processing a message.

If this happens, `fd_` will be set to -1 (reinitialized) and subsequent calls to `netops::recvmsg(fd_, &msg, MSG_ERRQUEUE)` will fail. On the read failure, an error will be propagated via `failErrMessageRead(__func__, ex)`. This quickly becomes confusing -- you end up with an error message `::recvmsg exited with code 9` (EBADF) with little context.

This diff adds logic to check if the socket is still open before each call to `netops::recvmsg(fd_, &msg, MSG_ERRQUEUE)`. If not, we exit out.

It's arguable that the application is at fault in this scenario instead of the library. However, I've found this problem and given the complexity, it seems like a scenario worth guarding against in the library.

Reviewed By: yfeldblum

Differential Revision: D14354115

fbshipit-source-id: 49e1efbbf6cecbe3bf6db7ee6bdfcac4c8aaa8c2

Summary: [GDB] Migrate `gdb/scripts/` to `platform007`.

Reviewed By: pixelb

Differential Revision: D14400956

fbshipit-source-id: 59f800c8770fae78ffd51d33161bb56dc6c9ff12

Summary:
- Replace `Uncvref` alias template with `folly::remove_cvref_t` from
  `folly/Traits.h` which does the same thing.
Pull Request resolved: https://github.com/facebook/folly/pull/1047

Reviewed By: Orvid

Differential Revision: D14395552

Pulled By: yfeldblum

fbshipit-source-id: 10ac4fbadbd20876fda4957f9a85454ad6c8a994

Summary: Fix SignalHandlerTest signal handler output parsing since we do not need to print the "maybe from PID" message

Reviewed By: yfeldblum, mjoras

Differential Revision: D14399778

fbshipit-source-id: a5192c0427fa79479218ebb4671dee4164448111

Summary:
- the test was disabled in {D2299981} (2015)
- philippv enabled `-gdwarf-aranges` in fbcode {D4013485} (2016)
- `-gdwarf-aranges` is available in clang since https://reviews.llvm.org/D1649 (2013)

Reviewed By: philippv

Differential Revision: D14398382

fbshipit-source-id: 7b6b4e7ebf86268d8bc34089b0e77dcaa5badf92

Summary: [Folly] `is_transparent`, for testing whether a hash or equal-to type marks itself as transparent.

Reviewed By: shixiao

Differential Revision: D14355772

fbshipit-source-id: 9bb0ce6f7fc177e2445d094796ae4e56d9a00e76

Summary: It should be dead now.

Reviewed By: yfeldblum

Differential Revision: D14350795

fbshipit-source-id: 30a021c0180b73b9d76e34722353c4e65e1c98f8

Summary:
CMake allows the user to add additional compilation options using
CXXFLAGS= or -DCMAKE_CXX_FLAGS=. For example, each of the following
commands would enable GCC security features in generated code:

    $ CXXFLAGS=-fstack-protector-strong cmake .
    $ cmake -DCMAKE_CXX_FLAGS=-fstack-protector-strong .

Unfortunately, this doesn't work with folly's build system; CXXFLAGS and
CMAKE_CXX_FLAGS are ignored. This makes it harder for package managers
to tweak folly's compilation.

Teach folly to respect user-supplied compiler flags instead of
overriding them.
Pull Request resolved: https://github.com/facebook/folly/pull/1046

Reviewed By: wez

Differential Revision: D14383310

Pulled By: yfeldblum

fbshipit-source-id: a6c527fa1659b93bb88a4e0be617a5d9022489dd

Summary: We are getting SIGILLs that end up not printing a code because it's not in the cases in `signal_reason`, which isn't particularly helpful. Additionally, the check for kernel-sourced si_codes is incorrect. Any code > 0 is from the kernel. This causes folly to log nonsense PIDs and UIDs.

Differential Revision: D14388102

fbshipit-source-id: 3201a16bbe9eb9038ebc5601044406f5333ec9e9

Summary:
folly dynamic "double / int" (division, same goes for subtraction) results in the operands switching places (always becomes "int / double") due to int and non-int checking logic.

Same applies to addition and multiplication, but order doesn't matter for those.

If both dynamics are doubles or both are ints, then the problematic logic is bypassed.

So this is only when you have one int and one double, and the int comes second, and the operand is subtraction or division.

Reviewed By: yfeldblum

Differential Revision: D14379896

fbshipit-source-id: e76f1fe677fc4ea69856649ab046a5080a902b0d

Summary: Exposing this counter allows us to fairly divide the memory usage of shared IOBuf buffers among all the owners in memory-usage tracking code.

Reviewed By: simpkins

Differential Revision: D14361976

fbshipit-source-id: 73c63c473b117198786143b421c6fcb8aed684d5

Summary: The file descriptor backed version will be going away.

Reviewed By: yfeldblum

Differential Revision: D14235700

fbshipit-source-id: 990b0276b085e6bb88bf3bdb83cdeba332f75ac6

Summary: The tests were accidentally checked in without the cmake build rules to build them.

Reviewed By: kevinlewi

Differential Revision: D14277667

fbshipit-source-id: 5e0ff305341772232579bbde2dfc60477b4f1937

Summary:
ssize_t is not standard, so include <sys/types.h>. (I considered
switching to intptr_t but decided against it.)

Also, on macOS, the function appears to be called backtrace, not
unw_backtrace.

Reviewed By: yfeldblum

Differential Revision: D14307252

fbshipit-source-id: d919bd37fea8f68b5d610e771e577c8ce385f6ec

Summary: They are dead.

Reviewed By: yfeldblum

Differential Revision: D14192072

fbshipit-source-id: dcb95be39614d9968db9d63cd9f06396f4779a0a

Summary:
- Update Ubuntu 14.04 build script to use CMake instead of Autotools.
- Rename Ubuntu 14.04 build script to use hyphens rather than underscores to match POSIX standard for bash scripts.
- The Ubuntu 14.04 build script is now meant to be run from the top-level rather than from inside the `folly` source tree.
- Remove the Ubuntu 12.04 build scripts. It was previously only needed for Travis CI, but Travis CI isn't using it anymore.
Pull Request resolved: https://github.com/facebook/folly/pull/1042

Reviewed By: Orvid

Differential Revision: D14298995

Pulled By: yfeldblum

fbshipit-source-id: e3151620c0f0fa73147ab258b8750480c868c44a

Summary:
- Change folly::coro::co_viaIfAsync to be a customization-point object.
- Modify co_viaIfAsync() to call a .viaIfAsync() member function if one exists.
  Otherwise it calls the co_viaIfAsync() function found by ADL.
- Add FOLLY_DEFINE_CPO() helper macro to folly/Portability.h
  Borrows some techniques from range-v3 to support CPOs in different
  compilers and C++ versions.
- Redefined co_viaIfAsync() CPO in terms of this macro.
  This allows it to be callable as folly::coro::co_viaIfAsync() while
  still allowing other types in folly::coro namespace to define friend
  functions that customise the behaviour of this customisation point
  without causing conflicts.
- Add folly::coro::is_semi_awaitable<T> metafunction.
- Add folly::coro::semi_await_result_t<T> metafunction.

Reviewed By: yfeldblum

Differential Revision: D14016389

fbshipit-source-id: 63998993b2fa8560700f81666a910dee31f3ba72

Summary:
The behavior of fiber Baton is currently inconsistent around posting of a timed out baton, and subsequent calls to try_wait().  As stated in the post, only when the waiter is a thread will a post overwrite a timeout internally, and this is observable externally because try_wait returns true only if a baton is in the POSTED state, but not the timed out state.  Example code:

```
folly::fibers::Baton b;

b.try_wait_for(std::chrono::milliseconds(1));
b.try_wait(); // returns false

b.post();
b.try_wait(); // Returns true on threads, false on fibers
```

Other options we could consider for the fix:
 * Update postThread to leave baton in a timeout state, and have waitThread move the baton into the timeout state on a failed wait.
 * Update try_wait() to return true if the state is either POSTED or TIMEOUT.

Reviewed By: andriigrynenko

Differential Revision: D14324532

fbshipit-source-id: 51b993a6153cba0fc3cf5701f765fff8ad1c8e3c

Summary:
Some codecs may not want counters enabled. Right now that will
`std::terminate`. Allow disabled counters, and add a test.

Reviewed By: felixhandte

Differential Revision: D14327170

fbshipit-source-id: 37858518c80e192e351e8578e088b59732d2a0cb

Summary:
- Some constant expressions are wrapped in a type carrying a bool, but
  are only used in value contexts. Remove the type wrapper
  (bool_constant) use and just use `constexpr bool` directly. This makes the
  usages not have to use `::value` everywhere.
Pull Request resolved: https://github.com/facebook/folly/pull/1039

Reviewed By: Orvid

Differential Revision: D14293828

Pulled By: yfeldblum

fbshipit-source-id: 78399e4fb401ab680a4f1ecef63fb4c2a82927c2

Summary: [Folly] Remove extraneous uses of keyword `typename` in fibers `collectN`.

Reviewed By: ericniebler

Differential Revision: D14317102

fbshipit-source-id: 1ae04081258bf2398865a1cc5461618648623158

Summary: [Folly] Canonical SFINAE in fibers, where the `enable_if` is on the outside.

Reviewed By: andriigrynenko

Differential Revision: D14294993

fbshipit-source-id: 498c9bfcd346edb5e4b5b65941e2a09c93e870b6

Summary:
- On some platforms, `CMAKE_SYSTEM_ARCHITECTURE` may resolve to an empty
  string.
- When this value is used without quotes to the `string` command, it is
  ill-formed with error:

```
  CMake Error at CMakeLists.txt:174 (string):
  string sub-command FIND requires 3 or 4 parameters.
```

- Explicitly wrap the value to the `string` operation in quotes to treat
  `CMAKE_SYSTEM_ARCHITECTURE` as a string, which is allowed to be empty.
  The result is that the `string` operation will not result in a hard
  error.
Pull Request resolved: https://github.com/facebook/folly/pull/1040

Reviewed By: calebmarchent

Differential Revision: D14293784

Pulled By: yfeldblum

fbshipit-source-id: cd5924fa62277aa07e930c7b088197ef596be977