Summary: [Folly] Compute `kSelectInByte` table at compile time rather than hardcoding the resulting precomputed table.

Reviewed By: luciang

Differential Revision: D9627807

fbshipit-source-id: 01fbb07dd14aac4b84ad72bac10a12d3657be589

Summary:
Currently malloc(goodMallocSize(x)) will succeed when x is larger
than the maximum possible allocation supported by jemalloc.  This is
not desirable.  Currently goodMallocSize(x) returns 0 if x is too large,
but 0 is a valid argument for malloc and causes a 1 byte allocation.
This diff causes goodMallocSize(x) to return x in that case, which will
cause the subsequent malloc to fail.  The caller may catch that failure
or crash immediately, but those are both preferrable to returning a
pointer to a valid 1-byte allocation.

Reviewed By: al13n321

Differential Revision: D9778137

fbshipit-source-id: 3eb8da72d6240b28da85483f0d91653f5e04b333

Summary:
[Folly] Refactor `StaticSingletonManager`.

* Move implementation to `.cpp`.
* Remove all `#include`s possible from `.h`.
* Control inlining.

Differential Revision: D9734721

fbshipit-source-id: 82a1cda187674a6bb62e14a80431aeb07c65156c

Summary:
[Folly] A unit-test for the `kSelectInByte` table.

The table essentially contains an encoding of every byte value. So we add a unit-test which decodes from the table, confirming that each encoded byte decodes back to the original.

Also add helpers to the test file to ease viewing disassembly of select64 variants.

Reviewed By: ot

Differential Revision: D9628409

fbshipit-source-id: 61cebaff4e9aff223a06194f7440eb3580a9273d

Summary: [Folly] Test cases for `StaticSingletonManager`.

Differential Revision: D9735443

fbshipit-source-id: e922d7742c12e6ca7725dc7782cc683cf71fb6a7

Summary: Add SpinLock from google/base/spinlock.h to the small lock benchmark

Reviewed By: yfeldblum

Differential Revision: D9771633

fbshipit-source-id: 3cf01866ec6cd4a4a5d8c2fb4274ca20a6f545dd

Summary: getApplicationProtocol should probably be const.

Reviewed By: knekritz, mingtaoy

Differential Revision: D9772305

fbshipit-source-id: c648f9307214258639616e2f4ff3a90dfbf9484b

Summary: This export the 16 times stack size for ASAN TSAN into a config as @[655434592:Andrii Grynenko] suggested

Reviewed By: yfeldblum

Differential Revision: D9685364

fbshipit-source-id: 4c76e99ba10d0fad65dd0dc30eeb003b25514c43

Summary: Use checkedMalloc everywhere in IOBuf code.

Reviewed By: yfeldblum

Differential Revision: D9763962

fbshipit-source-id: d08046b26bf0dd3be5067ad9e0706647fdcf1e83

Summary:
IOBuf::createCombined was derefencing a nullptr in case of allocation failure.
This patches fixes this issue by throwing std::bad_alloc if malloc() returns a
nullptr.

Reviewed By: yfeldblum

Differential Revision: D9763331

fbshipit-source-id: 7d9b9265079e74f5c63c8c8809099de97342f54d

Summary: Typo

Reviewed By: yfeldblum

Differential Revision: D9755749

fbshipit-source-id: b0295c12353a256e4462dfacf1bafce09c09c598

Summary:
Provide an option to force digest updates in BufferedStat. This
is useful in tests when one can't easily mock the clock.

Reviewed By: yfeldblum

Differential Revision: D9732899

fbshipit-source-id: e6a9c01d97a156bdded58fc01d8389bbf9edbbf7

Summary:
4d234b99 introduces `folly/detail/Futex-inl.h`. This patch adds it to `Makefile.am` so it gets installed when building using autotools.
Pull Request resolved: https://github.com/facebook/folly/pull/926

Reviewed By: yfeldblum

Differential Revision: D9757810

Pulled By: Orvid

fbshipit-source-id: 42bdaf292d9ca5e47c01d39381c5ebd4941dcb2a

Summary: Have AsyncSSLSocket use AsyncTransportCertificate interfaces properly

Reviewed By: mingtaoy

Differential Revision: D9752031

fbshipit-source-id: c65c0b808d82843bf1111bb650fe140ac98723b8

Summary:
F14MapTest already had properly relaxed the post-condition for
fallback maps after a move-clear and reserve; this diff applies the same
relaxation to F14SetTest.

Reviewed By: shixiao

Differential Revision: D9756311

fbshipit-source-id: 8285bb05023d923c4a7c4e014c3668bdaed591c4

Summary: Separates variants of Future::then that take differently parameterised lambdas. This will allow separately deprecating and then deleting them without having to solve all at once.

Reviewed By: yfeldblum

Differential Revision: D9512097

fbshipit-source-id: 1b6c0946f0fa38cf8427c66fe1bef47315875e58

Summary:
[Folly] Smaller implementation of `to_ordering`.

Shorter source code, shorter bytecode, no branches/conditionals.

Reviewed By: Orvid

Differential Revision: D9733375

fbshipit-source-id: a2b2a8263b5de2264d3068a424419d63e5b27228

Summary:
[Folly] Add `FOLLY_HAS_EXCEPTIONS` preprocessor symbol.

Add checks for:
* `__cpp_exceptions` and `__has_feature(cxx_exceptions)`.
* MSVC platform.

But avoiding a single unwieldy expression.

And use it in `throw_exception`.

Reviewed By: mnovakovic

Differential Revision: D9732924

fbshipit-source-id: 4ac00e2d84007dfe8307f0e861ab2edfbde80fff

Summary:
[Folly] Assorted partial applications of `clang-format`.

Some parts of the code are explicitly opted out where `clang-format` produces poor results, and for other reasons.

Future passes may remove current opt-outs.

Reviewed By: Orvid, igorsugak

Differential Revision: D9727723

fbshipit-source-id: d191ae51f535335402d5e083c3368d83779b7f6e

Summary: Disable recording highWaterMark for asan builds. I discovered this while using getHighWatermark method on test runs (build with ASAN).  It gave bogus value (>90MB), when stack size was only 24k.

Reviewed By: andriigrynenko

Differential Revision: D9687204

fbshipit-source-id: 7ca311acb6f310694850f065febc3998a11d036a

Summary: [Folly] Split the `FBString.testFixedBugs` test case into a test case per bug.

Reviewed By: Orvid

Differential Revision: D9727569

fbshipit-source-id: b319286031ded00451a1e960c0d450e7cfc4c362

Summary: Fix `throw_exception` with ` -fno-exceptions -Werror=unused-parameter`.

Reviewed By: yfeldblum

Differential Revision: D9693907

fbshipit-source-id: d308147b7d69cf091a9f839fda8be3464aa94b9a

Fix: folly/experimental/coro/test/CoroBenchmarkAllocator.cpp:363:20: error: 'Recursion' has not been declared

Summary: Fix: folly/experimental/coro/test/CoroBenchmarkAllocator.cpp:363:20: error: 'Recursion' has not been declared

Reviewed By: yfeldblum

Differential Revision: D9727088

fbshipit-source-id: 91447187bc6d34b965d5f8c2cedd66950b1fe909

Summary:
This diff adds operator=(initializer_list) to the F14 fallback
implementations, relaxes the guarantees provided by visitAllocationClasses
for fallbacks (so that the current implementation can be considered
correct), enables additional tests, and avoids a bit of logic repetition.
Changes should only be made to the fallback mode and tests, and this diff
should not change the set of platforms for which fallback mode is enabled.

Reviewed By: shixiao

Differential Revision: D9664700

fbshipit-source-id: 1f3fc6f0c48009bb3e5fd249506b42f54d93df43

Summary:
Remove bogus `configure` argument from cmake commands which is likely an
artefact from automake migration. The correct invocation of CMake is

```
cmake [<options>] {<path-to-source> | <path-to-existing-build>}
```

Reviewed By: Orvid

Differential Revision: D9721030

fbshipit-source-id: 95da844d3292cbf2acbec85753499f3603567e08

Summary: The template type here seems unnecessary, since the rest of the APIs refer to `folly::Executor` and not a template typename `Executor`

Reviewed By: LeeHowes

Differential Revision: D9695862

fbshipit-source-id: 26d4ddb6fe09454f037fcf1f5417add062e8371f

Summary:
F14 maps and sets were missing the operator= overload that takes
an initializer list, resulting in map = {} having reset behavior rather
than clear behavior.  Without the overload an empty map is constructed and
then moved onto the lhs, freeing all memory.  With the overload we call
clear(), which retains the existing allocation when it is not too large.
This diff also cleans up the default constructor definitions, adding
noexcept for F14FastMap and F14FastSet and making the others more concise.

Reviewed By: shixiao

Differential Revision: D9661526

fbshipit-source-id: 101d66623bc4d50e7935fa9a73405a825648e6a5

Summary: [Folly] Apply `clang-format` to `folly/futures/`.

Reviewed By: LeeHowes

Differential Revision: D9672492

fbshipit-source-id: c41632c4da6c450096eb249f0708af3db5057a14

Summary: [Folly] Apply `clang-format` to `folly/container/`.

Reviewed By: shixiao

Differential Revision: D9672403

fbshipit-source-id: da8e2a10eafcc8a278c002223cfeec055936d0bc

Summary: [Folly] Apply `clang-format` to `folly/gen/`.

Reviewed By: Orvid

Differential Revision: D9671968

fbshipit-source-id: 1164cf0d236914b5fd7ed35d5d41c785935e876b

Summary: D9624966 deleted FormatArg.cpp but makefile not updated

Reviewed By: knekritz

Differential Revision: D9663460

fbshipit-source-id: 02dc40d4b23de697834aca0247335ee79bebc647

Summary: Add via helpers for KeepAlive argument

Reviewed By: yfeldblum

Differential Revision: D9617303

fbshipit-source-id: 4ace3ea5f48fd8fbed70bcdf7654a1435f5e1d72

Summary:
The format specifier for printf should be %zu to take into account
32bit machines.

Reviewed By: Orvid

Differential Revision: D9665574

fbshipit-source-id: e7019c1deba80addef36ea8f6ebff48b45eceefa

Summary: [Folly] DRY some common code in `folly/gen/test/ParallelBenchmark.cpp`.

Reviewed By: Orvid

Differential Revision: D9664846

fbshipit-source-id: adf198a60ab637e47986ddea9c5c4e7392cf8b27

Summary:
This code fails to compile when FOLLY_HAS_COROUTINES is false. This is
because ExpensiveCopy is defined inside the if scope, and if that macro is
false, then ExpensiveCopy never gets defined. Yet it gets used in the nestedCalls function.

This diff fixes build for that case.

Reviewed By: lewissbaker

Differential Revision: D9663539

fbshipit-source-id: 1c3dd70decab7db955c038cd5c476920bfa6ba53

Summary: [Folly] Refactor `for_each` and `fetch` to fix their unit-tests under MSVC. New code is smaller.

Reviewed By: Orvid

Differential Revision: D9569412

fbshipit-source-id: 87807ddf41f1dbd16244a82c5be30eae08dbdf4c

Summary: [Folly] Assorted fixes to for_each and fetch comments

Reviewed By: aary

Differential Revision: D9632155

fbshipit-source-id: 64932770d3476cdd06c5fa9eb9bd845600f18fff

Reviewed By: lewissbaker

Differential Revision: D9388165

fbshipit-source-id: fa94f008d088c5560d5207c726fabb0e764193fb

Summary: Always using 10000 is no good for concurrent tests.

Reviewed By: Orvid

Differential Revision: D9656798

fbshipit-source-id: 4fb7e48239e42c3b2acd2cc976cb74535c573ef6

Summary: [Folly] Do only single var decl per statement in `folly/test/DynamicConverterTest.cpp`.

Reviewed By: igorsugak

Differential Revision: D9653672

fbshipit-source-id: 08bb529cc90761c69dabd2e00f58f1767f07146a