Summary:
- Fix all cases of "segfault if consumed/moved-out"
  - In Future, SemiFuture, Promise.
  - Those now throw NoState instead.

Reviewed By: yfeldblum

Differential Revision: D7721921

fbshipit-source-id: 76fafcbd30e23d63150196d3d7e404f348b220f2

Summary: There isn't a good way to access ThreadPoolExecutors to add instrumentation, such as grabbing basic stats on all executors to better understand pool capacity. This provides a way to do that.

Reviewed By: yfeldblum, djwatson

Differential Revision: D7658144

fbshipit-source-id: 3e72acdccf25742b76d438d2895d586049a0cc17

Summary:
MicroSpinLock was doing a non-relaxed load before cas, which seems odd because typically we expect spin locks to be unlocked. This diff flips the order and assumes the lock is unlocked before doing a load-loop.

While the benchmark results for SmallLocksBenchmark are a little mixed, the results in a more real world use case in stats/DigestBuilderBenchmark show that this is a real winner.

Reviewed By: djwatson

Differential Revision: D7828158

fbshipit-source-id: e195b9739ce931a67a39ffb9715a228aef59f6bc

Summary: Some infrastructure sends `SIGQUIT` to stuck processes and we would like to be able to log stacktraces for these.

Reviewed By: yfeldblum

Differential Revision: D7819980

fbshipit-source-id: d99a1232ca259bd95aed82505751e6591af748ff

Summary:
Just re-running it on newer hardware:

32 processors
model name      : Intel(R) Xeon(R) CPU E5-2660 0 @ 2.20GHz

Reviewed By: yfeldblum

Differential Revision: D7828138

fbshipit-source-id: 30a25c6bf7bd0294f32f7598c701b49f045ab235

Summary: Small change so that a template class `ExtractFirstArg` can do the same thing without relying on an expensive header, `boost/type_traits.hpp`.

Reviewed By: yfeldblum

Differential Revision: D7781675

fbshipit-source-id: 1e5f431217da6cdd31ba1d77a8fc2b19a43808aa

Summary: In a test build `<iterator/iterator_facade>` showed up as the most heavyweight include.  Remove this include, dependency on the boost lib, and implement whatever's needed for the bidirectional iterator interface.

Reviewed By: yfeldblum

Differential Revision: D7781659

fbshipit-source-id: 380e0ae85b18c245845f479b2d268fdd3e16d54a

Summary:
Promote the folly logging code out of the experimental subdirectory.
We have been using this for several months in a few projects and are pretty
happy with it so far.

After moving it out of the experimental/ subdirectory I plan to update
folly::Init() to automatically support configuring it via a `--logging` command
line flag (similar to the initialization it already does today for glog).

Reviewed By: yfeldblum, chadaustin

Differential Revision: D7755455

fbshipit-source-id: 052db34c97f7516728f7cbb1a5ad959def2f6efb

Summary: As title. We were special casing q == 1 or q == 0 improperly before. This led to some strange results where p100 < p99.9. Track min and max (useful anyway), and use those for boundaries.

Reviewed By: anakryiko

Differential Revision: D7800645

fbshipit-source-id: 5fa3e1d0d286d4b31442196ba74b072030706786

Summary: This should fix running tests in a vcpkg environment.

Reviewed By: yfeldblum

Differential Revision: D7822368

fbshipit-source-id: a549febeb0bbea7208386cbed8a6ecb316b98661

Summary: Because Windows doesn't have `uint` as it's non-standard.

Reviewed By: djwatson

Differential Revision: D7821774

fbshipit-source-id: 4f7ff4468fc56052e28995f386e194fe70c6c513

Summary: Because it makes Visual Studio's intellisense happier, and it's always better to match exactly, rather than matching implicitly.

Reviewed By: yfeldblum

Differential Revision: D7821865

fbshipit-source-id: 9e2380884b2d8d12944e3a3fc921e532208b1791

Summary:
And move it to folly::python namespace.
We are going to make the python bridge work with fibers. This will require to make changes to this executor that are quite specific for the python use case.
It doesn't hurt to remove YetAnotherExecutor choice from the main executor directory.

Reviewed By: yfeldblum

Differential Revision: D7811125

fbshipit-source-id: eec21ccbbbad58aaa795df661ba58f67fa177645

Summary:
As per discussions on D7784214, to be consistent with the functions used by the
integral_hasher, mark these `noexcept`. Not much benefit for this diff other
than consistency.

Reviewed By: yfeldblum

Differential Revision: D7797826

fbshipit-source-id: 382c82556efd261841f3f6fb2ae892d26f3ceb0b

Summary:
Currently we'll are getting incorrect Idle&Busy time counters for loops when there were no LoopCallbacks. This is because we don't properly refresh the `startWork_` and use the one from the last loop that had LoopCallbacks.

This diff fixes that by properly refreshing `startWork_` even if there're no LoopCallbacks.

Reviewed By: yfeldblum

Differential Revision: D7811127

fbshipit-source-id: 94ded08963b1bddc7b7859484f076d4c8a1e897f

Summary: Fix shadown warning for xplat builds

Reviewed By: yfeldblum

Differential Revision: D7811636

fbshipit-source-id: 609b5acaca3cfdd11059ed1ebbaa969e281487a2

Summary:
Calling connect on the UDP socket can avoid route lookups
on the write path. This would be an optimization that would be
useful for clients.

There are several caveats of using connect() which we discovered
during testing. These are documented and unit tests have been written
for each of them.

Reviewed By: yfeldblum

Differential Revision: D7802389

fbshipit-source-id: 09b71373a3a95c5dab73ee9345db0dbbf66d4ec5

Summary: Some links are complaining of multiple definitions. Move to the cpp file to resolve.

Reviewed By: yfeldblum

Differential Revision: D7806874

fbshipit-source-id: f25342d647505b676ba3659400c63d6a59344024

Summary: This header is no longer needed since D7687476.

Reviewed By: yfeldblum

Differential Revision: D7804756

fbshipit-source-id: 6596f5296965b593ab5b0decf4c67203e19b03c8

Summary: As title

Reviewed By: tageorgiou

Differential Revision: D7804647

fbshipit-source-id: 9a58de258cd49a4d15ded05163975304c0e7e83b

Summary: This worked in clang but not in gcc. Quick fix

Differential Revision: D7800576

fbshipit-source-id: fcd04c0779f8a176043f3dede61289830a1ef607

Summary: Introduces two QuantileEstimators. Both estimators use the TDigest for estimating quantiles, and giving accurate sum, count, and mean. One estimator uses a single digest, and the other uses a SlidingWindow of 60 digests. All writes to the data structure are buffered for 1 second.

Reviewed By: anakryiko

Differential Revision: D7645824

fbshipit-source-id: 4f1f2c25388c3280f148591502449f857c6467d3

Summary: [Folly] Use `invoke_result` v.s. `std::result_of`, which is deprecated.

Reviewed By: aary, ericniebler

Differential Revision: D7771480

fbshipit-source-id: 1d89d489be98bfbe2014f2d6b5fc037f02147876

Summary: [Folly] Cut unused alias `isFutureResult`.

Reviewed By: Orvid

Differential Revision: D7787146

fbshipit-source-id: 44ed51e9e04c71530f67fc6dceb76b247dcadb55

Summary:
The test was missing a `main` meaning that it would fail even if libc++ was in use, making all the rest fail on macos for example.
Closes https://github.com/facebook/folly/pull/831

Reviewed By: Orvid

Differential Revision: D7787115

Pulled By: yfeldblum

fbshipit-source-id: 4a10323298be3f26f783fa1257cbde2483fb75b9

Summary:
The libstdc++ implementation of `std::unordered_map|set` determines whether to
cache a per-node hash code based on (1) whether the supplied hasher is fast
(true by default) and (2) whether the hashing operation will throw.

Mark the `operator()` of hashers of numeric types (integral, float, enum)
`noexcept`, such that stl unordered containers won't cache a hash code. This is
essentially free memomy savings, as comparing the cached hash code is as
expensive as comparing the numerical types directly.

Reviewed By: luciang

Differential Revision: D7784214

fbshipit-source-id: 020e1c38265f33a482b6ba913a11f1a383e840f8

Reviewed By: yfeldblum

Differential Revision: D7765404

fbshipit-source-id: 3f09d6806d0fb98cb59529adc6ea2d51fa7ccbbb

Summary:
BufferedStat hooks up the DigestBuilder and a Digest or a SlidingWindow

Every windowDuration time, the DigestBuilder builds a new digest that will be merged into the central data structure (either a digest or a sliding window of digests).

This diff removes the shared buffer from DigestBuilder, as the concurrency model was getting overly complex with it.

Reviewed By: simpkins

Differential Revision: D7629082

fbshipit-source-id: 54027363d37c0fdced287550eea5b12caf6b8138

Summary: [Folly] Use `is_invocable_r` in `MapUtil` v.s. reimplementing it.

Reviewed By: aary

Differential Revision: D7771481

fbshipit-source-id: 41565a06198ed2289206f01a088c3c1b3a01bc8e

Summary: [Folly] Use `is_invocable_r` in `FunctionRef` v.s. reimplementing it.

Reviewed By: ot

Differential Revision: D7769745

fbshipit-source-id: 3cb7bf9908cc6e93c7d19ace9a91c69ba14cdfae

Summary:
OpenSSL 1.1.0 renamed and deprecated X509_CRL_get_lastUpdate -> X509_CRL_get0_lastUpdate
along with X509_CRL_get_nextUpdate -> X509_CRL_get0_nextUpdate

Reviewed By: ngoyal, anirudhvr, Orvid

Differential Revision: D7782339

fbshipit-source-id: 63f04284bb54da87d4235a02890107db8abaa4b4

Summary:
This was exposed by the latest clang with -Wreturn-std-move enabled:
```lang=bash
folly/FBString.h:2487:12: error: local variable 'rhs' will be copied despite being returned by name [-Werror,-Wreturn-std-move]
    return rhs;
           ^~~
folly/test/FBStringTest.cpp:1294:23: note: in instantiation of function template specialization 'folly::operator+<char, std::char_traits<char>, std::allocator<char>, folly::fbstring_core<char> >' requested here
    auto test1 = '\0' + std::move(s1);
                      ^
folly/FBString.h:2487:12: note: call 'std::move' explicitly to avoid copying
    return rhs;
           ^~~
           std::move(rhs)
2 errors generated.
```

Reviewed By: Orvid

Differential Revision: D7776805

fbshipit-source-id: f11ae5a1fc416b8838b8ea74d7f6b488f07cab38

Summary: Need to also allow calling this function with a writeLock.

Reviewed By: magedm

Differential Revision: D7477326

fbshipit-source-id: 403e2667a3f77be4c1d7d06dbd5f36218a659abc

Summary: The test is intended to verify the handling of trailing space.

Reviewed By: yfeldblum

Differential Revision: D7761229

fbshipit-source-id: 0196f57e6c76562562ef1b724f6dff1c7bfd5d81

Summary:
Rename folly/Logging.h to folly/GLog.h.  This header primarily exists to just
include <glog/logging.h>, and it also defines a single `FB_LOG_EVERY_MS()`
macro.

I plan to move the folly/experimental/logging/ directory to folly/logging/
soon, and I want to avoid confusing between this `Logging.h` header and the
separate `folly/logging/` code.

Reviewed By: yfeldblum

Differential Revision: D7754543

fbshipit-source-id: c91b8016d303cd0fae9e688e8782b22d095a6554

Summary:
D7642870 accidentally broke linking of folly when libunwind is available.
It change the code to use the contents of `${FOLLY_LINK_LIBRARIES}` in
folly-deps.cmake, but libunwind was added to `${FOLLY_LINK_LIBRARIES}` after
folly-deps.cmake was included, in FollyConfigChecks.cmake.

This moves the check for libunwind to folly-deps.cmake to fix this issue.  All
of the other checks for external libraries are already in folly-deps.cmake.

Reviewed By: yfeldblum

Differential Revision: D7764762

fbshipit-source-id: 922459a7eefcbd92d3d77c02e9c215bb1a5467cc

Summary: This allows using a keep-alive token in places where Executor* was previously used.

Reviewed By: yfeldblum

Differential Revision: D7751516

fbshipit-source-id: af2cb31191a78306439408dbee78dcd923492e30

Summary:
`PRETTY_TIME_HMS` extends `PRETTY_TIME` by adding additional minutes
and hours units which are more useful for human-reporting of time intervals.

Reviewed By: yfeldblum

Differential Revision: D7762524

fbshipit-source-id: 881936916db182c163a3df14196f4f11fb649527

Summary: [Folly] Avoid using gmock private macros - instead, use only the public interface.

Reviewed By: knekritz

Differential Revision: D7752463

fbshipit-source-id: 625465e7ec93613869e259fe1fe0b08aee7bb7d0

Summary:
This diff adds a method to calculate the amount of memory allocated by F14
sets. It mimics the node structure in `std::unordered_map` to calculate the
results. For certain keys and hashers, the libstdc++ implementation of
`std::unordered_map` caches a hash code in the node. libcpp seems to do this
for everything.

https://github.com/gcc-mirror/gcc/blob/9ce5a00d2d3b8b107cb1ca0e23bf1e31a095f9b6/libstdc%2B%2B-v3/include/bits/hashtable.h#L45-L49
https://github.com/llvm-mirror/libcxx/blob/58bcf28c63e0cf2ee5828e7811829c98bd244dda/include/__hash_table#L117

Depends on D7715573.

Reviewed By: nbronson

Differential Revision: D7728281

fbshipit-source-id: 0aa3ab60c746d3ab6bdb7c879f2f987110e0329d