Summary: To make sure RequestContext is properly unset when we stop processing request on a thread. This changes the API in Fibers, NotificationQueue, HHWheelTimer and AsyncTimeout, and fixes RequestContext handling in Futures (reset RC after the callback is done)

Reviewed By: andriigrynenko

Differential Revision: D3279644

fbshipit-source-id: a6a1c2840cdce179334aa1a3b28fa514cd5358c1

Summary:
This diff standardizes `HHWheelTimer` construction. It's not that safe to allow anyone to call `new HHWheelTimer` because the caller needs to remember to call `DelayedDestruction::destroy`.

Once callers are made to be safer, I'll be able to change the `HHWheelTimer` to use standard smart pointers instead of `DelayedDestruction`.

This picks up some work that I started in D2627941 but had to postpone.

This was mostly a mechanical change:

  ( fbgs -ls 'new HHWheelTimer' ; fbgs -ls 'new folly::HHWheelTimer' ) | xargs perl -pi -e 's/\bnew\s+((?:folly::)?HHWheelTimer)\b/$1::newTimer/g'

Then I manually inspected the code. I reverted `folly/io/async/HHWheelTimer.h`, since the `newTimer` factory method is the one place that we still want to call `new HHWheelTimer`. I manually edited `proxygen/lib/utils/SharedWheelTimer.cpp`, since that was using a `shared_ptr` with a custom destructor, which isn't needed anymore. I reverted `common/io/async/d` since the D shim is meant to pass around only raw pointers. I had to replace each instance of `foo.reset(…)` with `foo = …` . Then I made manual edits to `common/clientpool/ClientPool2-inl.h` because that code wants to store the timer in a `ThreadLocalPtr`.

Reviewed By: yfeldblum

Differential Revision: D3237530

fbshipit-source-id: 96bfb6987098618ad5430c21b1314f385f04a93d

Summary: This -Wbool-compare is apparently new in GCC 5.  We should ignore it, similar to the existing suppression of -Wsign-compare.

Reviewed By: pixelb, yfeldblum

Differential Revision: D3293675

fbshipit-source-id: 5f90b0dbf049a06c0f6c3df2cdfcc8055358e367

Summary:
Error message might be < 256 bytes
in which case buf.size() would be greater
than the size of the message written out
and we might have garbled bytes at the end.

ERR_error_string_n null terminates the string
so just use that.

Reviewed By: anirudhvr

Differential Revision: D3297320

fbshipit-source-id: 2ae626ce4c49ca478806c0bcc40a390e5a84f24b

Summary: This updates FiberManager to automatically register custom stack for signal handlers. This ensures signal handlers will work correctly not only in case of stack overflow, but also if fiber stack size is just not enough for them.

Reviewed By: yfeldblum

Differential Revision: D3289912

fbshipit-source-id: 44959b727f70e1f1748d837b1a06c7414433e5ec

Summary:
This `ThreadLocalPtr::reset` overload will accept a `unique_ptr`. It's actually not totally exception safe, simply because `ElementWrapper::set` is not exception safe. The best I can say is that my additional code is exactly as safe as the underlying implemenation.

liketolivedangerously

Reviewed By: ericniebler

Differential Revision: D3271563

fbshipit-source-id: 774bcf31924b1ed4e29a6cb1c0a36ad710ab6034

Summary: While making an unrelated change, (D3271563, which was needed from an unrelated change, (D3237530)), I noticed a lack of exception safety here. If `std::bad_alloc` were thrown, then we don't want to leak memory.

Reviewed By: ericniebler

Differential Revision: D3271911

fbshipit-source-id: 0d316c0fa865a7d64622c1d62160bb0c2b061d78

Summary:
According to the POSIX definition, IOV_MAX is defined in
limits.h. Given that IOVec.h is used to provide IOV_MAX and similar,
it seems like we should add limits.h as well.

Differential Revision: D3291996

fbshipit-source-id: 238140ebef7b1b90a28f3a67a25c5f5863538fc7

Summary: Windows doesn't have sys/uio.h, so implement it. This doesn't adjust any includes of it yet, I'll leave that for a separate diff.

Reviewed By: yfeldblum

Differential Revision: D2978042

fbshipit-source-id: 16d398a856de50b6c662f739eb11dfb851096592

Summary: Add server support for TFO to wangle.

Reviewed By: Orvid

Differential Revision: D3275427

fbshipit-source-id: 466baa342e2e0c4751e9c96d1e660cffb90e0634

Summary:
abs(-INT_MIN) is UB. Avoid it by first casting the number to
its unsigned equivalent and then negating that.

Reviewed By: yfeldblum

Differential Revision: D3280835

fbshipit-source-id: 40922d686173e6467e15d5a6dc2b62ad718349b5

Summary:
MSVC builds of libevent expect `evutil_socket_t` to be `HANDLE` values, but Folly, HHVM, and Thrift all use them as file descriptors.
This adds a `folly_event_set` function that always expects a file descriptor rather than a socket. This also changes the places in Folly that use it.

Reviewed By: mzlee, yfeldblum

Differential Revision: D2874655

fbshipit-source-id: 66cfd86fd69a9fbac30d150445e4814cd5ca799e

Summary:
We shouldn't rely on the internal buffer size of openssl
for errors.

This makes it more explicit.

Reviewed By: knekritz

Differential Revision: D3285918

fbshipit-source-id: 452bf16a3de151d0aa79c774f2fdfc1e08ee6f9c

Summary: Add NETWOR_ERROR to switch so it doesn't fall into Invalid Exception block.

Differential Revision: D3282448

fbshipit-source-id: 2cc5d21e436e6ac8f2f6581d99341ac47cf5b023

Reviewed By: siyengar

Differential Revision: D3264727

fbshipit-source-id: fb10558096e73a3cbb6f2fb93296524c0206910d

Summary:
These are pretty trivial. Debian's lint tool (lintian) warns about those
in its verbose mode, and since we saw them when packaging for Debian, we
thought of fixing them and forwarding them upstream.
Closes https://github.com/facebook/folly/pull/345

Reviewed By: Orvid

Differential Revision: D2769801

Pulled By: elliottneilclark

fbshipit-source-id: ad37cc53792f21aae6558ba256f20cb56c6caab5

Summary:
These didn't work properly in GitHub.
Closes https://github.com/facebook/folly/pull/375

Differential Revision: D3281872

Pulled By: elliottneilclark

fbshipit-source-id: b5398267ff5856bb6a7c4a3b963ae5cb47859bc3

Summary:
It would be nice to fix, but it's going to be easier to do when folly
support is for GCC>=5 which adds `__builtin_add_overflow`.

For now disable the check as the Histgram code handles overflow
correctly and has tests for these cases.

Fixes need to be done for both float and integer types as Histogram is used with both.

Reviewed By: meyering

Differential Revision: D3280260

fbshipit-source-id: 56e524c517566a4547346859be7770eda2acee96

Summary:
The result of E1 << E2 is E1 left-shifted E2 bit positions; vacated
  bits are filled with zeros.

  If E1 has a signed type and non-negative value, and E1 × 2E2 is
  representable in the result type, then that is the resulting value;
  otherwise, the behavior is undefined.

The code assumed the signed left shift worked like an unsigned left
shift, so make that explicit.

Reviewed By: meyering

Differential Revision: D3280325

fbshipit-source-id: 178b95ee36b7a1126a79bb825c2ad2ffa6fa9464

Summary:
third-party-buck/build/libgcc/include/c++/4.9.x/bits/stl_vector.h:866:9: runtime error: reference binding to null pointer of type 'char'

```
Breakpoint 2, 0x0000000000494894 in __ubsan_handle_type_mismatch_abort ()
(gdb) bt
```

Reviewed By: meyering, yfeldblum

Differential Revision: D3279234

fbshipit-source-id: 63a761587e5b3f79ece09fc99b9a593da0e44b75

Summary: MSVC's preprocessor is different in how it deals with passing arguments expanded from `__VA_ARGS__`, so we have to add a stub to force it to expand the arguments.

Reviewed By: yfeldblum

Differential Revision: D3256328

fbshipit-source-id: 551434833d40e55498a9ab352eb74acdfd094835

Summary: MSVC doesn't define `__cplusplus` as a high enough value to trigger this, so add an explicit check for MSVC instead.

Reviewed By: yfeldblum

Differential Revision: D3271647

fbshipit-source-id: a1e5a5a7eb75dce066dfc7fae8b2086880dc4c3d

Summary:
before:

  [ RUN      ] RangeFunc.Vector
  bits/stl_vector.h:866:9: runtime error: reference binding to null pointer of type 'int'

Reviewed By: yfeldblum

Differential Revision: D3279253

fbshipit-source-id: 957a8ec050f5c3b27febf7e6cd16ad1d307da169

Summary:
Running the folly/io/test:compression_test with a UBSAN-enabled binary would fail with this:

  folly/io/test/CompressionTest.cpp:96:36: runtime error: shift exponent 16777216 is too large for 64-bit type 'size_t' (aka 'unsigned long')

That exposed the unwanted left-shift.
That made it so the actual "start" offsets were in the range [0..numThreads-1], rather than being
e.g., with numThreads == 8, [0, 16777216, 33554432, ...].  Oops.

Reviewed By: luciang

Differential Revision: D3277606

fbshipit-source-id: 4bc48f84b0065de070b62a810f08b96e36ea7136

Summary: Because it is the standard (well, Intel's standard) location for __m128i, and GCC, Clang, MSVC and ICC all implement it in that location.

Reviewed By: yfeldblum

Differential Revision: D3270872

fbshipit-source-id: fa0f0410670cdb310e656ca55ee0f568ca5ee06f

Reviewed By: meyering

Differential Revision: D3277138

fbshipit-source-id: 7050d18c256ac15c8122d4af59aa31208b1466fd

Summary: Due to some fun with the order MSVC does name lookup in templates, it ends up complaining that these are ambigious between folly::detail and folly::recordio_helpers::detail. The solution is to simply reference folly::detail explicitly.

Reviewed By: yfeldblum

Differential Revision: D3271606

fbshipit-source-id: e599ad8d8fe33b1e4ec0b838ddb5dfacfdf60159

Summary:
This fixes a few issues in the portability headers for MSVC.
Not entirely sure how I managed to cause these in the first place, but this fixes them anyways.

Reviewed By: yfeldblum

Differential Revision: D3271859

fbshipit-source-id: 11a5d35246f29112563ee9079525aa37ced206ff

Summary:
I would switch these to just use the intrinsic functions, but GCC 4.8 doesn't support them.
MSVC supports the intrinsics, which is the primary reason for the switch.

Reviewed By: yfeldblum

Differential Revision: D3265572

fb-gh-sync-id: 3268f90d5234bdf77b3579504efd3dcd43f38aff
fbshipit-source-id: 3268f90d5234bdf77b3579504efd3dcd43f38aff

Summary: It's very unlikely that caller will handle such exceptions and FiberManager can proceed even if one task failed. DFATAL will ensure we still fail on such exceptions in debug builds.

Reviewed By: yfeldblum

Differential Revision: D3273573

fb-gh-sync-id: 55eba19c16429a3c89a2fdf5363fd4b55b3998a2
fbshipit-source-id: 55eba19c16429a3c89a2fdf5363fd4b55b3998a2

Summary: Having singleton registered after registrationComplete() is called is not a hard failure. Replacing exception with ERROR, since we probably still want to avoid such situation.

Reviewed By: yfeldblum

Differential Revision: D3266376

fb-gh-sync-id: 8130d279775348c835265987c95189bd3d8106f9
fbshipit-source-id: 8130d279775348c835265987c95189bd3d8106f9

Summary:
I would switch these to just use the intrinsic functions, but GCC 4.8 doesn't support them.
MSVC supports the intrinsics, which is the primary reason for the switch.

Reviewed By: yfeldblum

Differential Revision: D3265572

fb-gh-sync-id: f12d16b8f64cf48adcb97958d7ae88d56689a96f
fbshipit-source-id: f12d16b8f64cf48adcb97958d7ae88d56689a96f

Summary: Windows doesn't define a couple of functions.

Reviewed By: yfeldblum

Differential Revision: D2990676

fb-gh-sync-id: 1a1354bd7976cb8b4747436d6907d65210ba506c
fbshipit-source-id: 1a1354bd7976cb8b4747436d6907d65210ba506c

Summary:
test console reports "address already in use" failure so add some
randomness to it.

Reviewed By: yfeldblum

Differential Revision: D3265329

fb-gh-sync-id: a825cce787394783bb8d18d92385d2497b163385
fbshipit-source-id: a825cce787394783bb8d18d92385d2497b163385

Summary: Windows doesn't have it, and it's used in a lot of places.

Reviewed By: yfeldblum

Differential Revision: D2997499

fb-gh-sync-id: fbf8f5f087b8f39197d5bf2177e4d34efd74631f
fbshipit-source-id: fbf8f5f087b8f39197d5bf2177e4d34efd74631f

Summary: Grouping all folly::ThreadLocal atfork hooks in one. This allows proper ordering between folly::ThreadLocal hooks and hooks added by other libraries (we always want folly::ThreadLocal hooks to be registered first).

Reviewed By: ericniebler

Differential Revision: D3262666

fb-gh-sync-id: 1aa928b9ddd484580f4a4a7b320e9e64164bb04a
fbshipit-source-id: 1aa928b9ddd484580f4a4a7b320e9e64164bb04a

Summary: LoopKeepAlive can be useful to tell EventBase, that loop() shouldn't return even if there are no events registered, because some tasks will later be added via runInEventBaseThread. While at least one LoopKeepAlive is alive - EventBase::loop() behaves like EventBase::loopForever().

Reviewed By: yfeldblum

Differential Revision: D3261706

fb-gh-sync-id: d91424d3d12cae11abd72cffdbd57f136f628dae
fbshipit-source-id: d91424d3d12cae11abd72cffdbd57f136f628dae

Reviewed By: yfeldblum

Differential Revision: D3244669

fb-gh-sync-id: 73fa4ecb0432a802e67ef922255a896d96f32374
fbshipit-source-id: 73fa4ecb0432a802e67ef922255a896d96f32374

Summary:
There are two places that still includes gflags/gflags.h:
portability/GFlags.h and futures/exercises/Koan.h. Also removing
gflags entirely from a couple of places which never reference it at
all.

Reviewed By: yfeldblum

Differential Revision: D3203187

fb-gh-sync-id: e7725fbf2c4b3c475c0035e1b0e8c6e1227f277f
fbshipit-source-id: e7725fbf2c4b3c475c0035e1b0e8c6e1227f277f

Summary: Windows doesn't have these functions under these names, so be nice to the code using it.

Reviewed By: yfeldblum

Differential Revision: D3003641

fb-gh-sync-id: 1657545af6975c96aea4a48c359250828b5921e0
fbshipit-source-id: 1657545af6975c96aea4a48c359250828b5921e0