Summary: This will ensure SingletonVault is always available. It matters for singletons, not managed by folly::Singleton. Singletons in it will actually be destroyed via static SingletonVaultDestructor.

Test Plan: unit test

Reviewed By: chip@fb.com

Subscribers: njormrod

FB internal diff: D1591270

Summary: For great speed

Test Plan:
$ fbmake opt && _bin/folly/wangle/wangle-bench
============================================================================
folly/wangle/test/Benchmark.cpp                 relative  time/iter  iters/s
============================================================================
constantFuture                                             235.92ns    4.24M
promiseAndFuture                                 100.37%   235.04ns    4.25M
withThen                                          41.97%   562.17ns    1.78M
----------------------------------------------------------------------------
oneThen                                                    539.03ns    1.86M
twoThens                                          64.01%   842.14ns    1.19M
fourThens                                         39.27%     1.37us  728.62K
hundredThens                                       1.82%    29.63us   33.75K
============================================================================

Reviewed By: davejwatson@fb.com

Subscribers: trunkagent, net-systems@, fugalh, exa, njormrod

FB internal diff: D1587871

Summary:
I'm not thrilled with this implementation, but it mostly works. The big bummer is enforcing that Observables are held in shared_ptrs, or rather that enforcing that condition is impossible.
The protected constructor / friended dumb make_shared() pattern is clunky, and it'd be really easy for a subclasser to shoot themselves in the foot or even in the face.

It does seem like maybe Observable should be made an interface again, and all these details should live in a subclass (FanoutObservable?) where the restriction are super obvious. For instance, the langtech AudioStream object doesn't need all this crap because it overrides subscribe() without using the observer list, but it subclasses anyways.

I'm noodling another approach that (if it works) will not require the shared_ptr dancing, but will probably have some additional overhead... the observable would have to keep track of the subscriptions itself.

I like the RAII subscriptions, though perhaps subscriptions should be optional as long as it's clear that your subscription will last forever it you opt out of them.

Thoughts?

Test Plan: added unit

Reviewed By: hans@fb.com

Subscribers: rushix, hannesr, trunkagent, fugalh, mwa, jgehring, fuegen, njormrod, bmatheny

FB internal diff: D1580443

Summary:
This adds versions of stringPrintf() and stringAppendf() that accept
va_list arguments.

This also fixes the existing stringPrintf() functions to properly call
va_end() even when an exception is thrown in stringPrintfImpl().

Test Plan: Included new unit tests for stringVPrintf() and stringVAppendf().

Reviewed By: davejwatson@fb.com

Subscribers: trunkagent, doug, net-systems@, exa, njormrod

FB internal diff: D1583675

Summary: needed for thrift server. if you have some sort of stateful/functional factory and you want to hold on to it, you should be allowed to

Test Plan: compiles with forthcoming thrift diff

Reviewed By: davejwatson@fb.com

Subscribers: trunkagent, fugalh, njormrod

FB internal diff: D1584587

Summary:
I'm not 100% sure this is the best way to go about this but I don't hate it either.
I'm going to start seeing how it might fit into tserver - my guess is that some sort Cpp2WorkerFactory which also manages those objects would get plugged in as the thread factory
Haven't fleshed out how this would relate to TEventBaseManager

Test Plan: added unit, starting to play with this in Thrift2 server

Reviewed By: davejwatson@fb.com

Subscribers: alandau, bmatheny, trunkagent, fugalh, njormrod

FB internal diff: D1574660

Summary:
Could the problem be that via continues the existing chain of futures,
whereas we actually want to start a new chain?

Is there any particular reason this wasn't implemented like this originally?

Test Plan:
Ran all the unit tests. I hope to try to reproduce the thread issue and
see if this improves things.

Reviewed By: davejwatson@fb.com

Subscribers: trunkagent, net-systems@, exa, njormrod, fugalh

FB internal diff: D1500225

Tasks: 4920689

Summary:

Test Plan: fbmake runtests, OBC no longer segfaults

Reviewed By: mshneer@fb.com

Subscribers: fbcode-common-diffs@, trunkagent, mcduff, marccelani, hitesh, mshneer, njormrod, lins

FB internal diff: D1573880

Summary: I'm going to have hhvm depend on Subprocess, found this build error in open source.  I figured the intent was to set r to the return code.  I could also delete the whole thing since it's unused.

Test Plan: Unit tests

Reviewed By: tudorb@fb.com

Subscribers: trunkagent, mwilliams, doug, njormrod

FB internal diff: D1583626

Summary: Some minor bug and a failed test caused by D1578466

Test Plan:
fbconfig thrift/lib/cpp/test:HHWheelTimerTest && fbmake
runtests

Reviewed By: davejwatson@fb.com

Subscribers: trunkagent, alandau, bmatheny, njormrod

FB internal diff: D1581949

Summary: We only need exclusive lock when we add items to singletons_. Each SingletonEntry has its own mutex, so it's safe to rely on it for any modifications within individual entries.

Test Plan: Applied D1573880 and ran fbconfig -r servicerouter/client/cpp2 && fbmake runtests

Reviewed By: chip@fb.com

Subscribers: trunkagent, njormrod, hitesh, mshneer

FB internal diff: D1579877

Summary:
Because we put FBString and Malloc into libgcc, we have to be
careful which dependencies they each have. We cannot include ScopeGuard.
Reorganize the code a bit to avoid the ScopeGuard.

Test Plan: fbconfig -r folly && fbmake runtests

Reviewed By: je@fb.com

Subscribers: trunkagent, sdwilsh, njormrod

FB internal diff: D1580957

Summary:
By the birthday paradox, this will fail about once every 8,100
runs. Dropping to 100 with cut that to 1 in 860,000.

Test Plan: fbconfig -r folly && fbmake runtests

Reviewed By: sdoroshenko@fb.com

Subscribers: sdwilsh, njormrod

FB internal diff: D1581238

Summary: This should be the last test abusing sleeps.

Test Plan: ran

Reviewed By: hannesr@fb.com

Subscribers: fugalh, njormrod

FB internal diff: D1580830

Tasks: 5225808

Summary:
Avoid a clang-caught heap-buffer-overrun
and document that the ssse3 decoder requires at least 17 bytes
of readable memory starting at the source pointer.
* folly/GroupVarint.h (decode_simple): Comment out dead code.
(decode)[__SSSE3__]: Add a comment describing this limitation.
* folly/test/GroupVarintTest.cpp (testGroupVarint32):
Include <algorithm> for use of std::max.
Ensure that the buffer we use has length at least 17.

Test Plan:
Now, all tests succeed with clang/asan

Reviewed By: simpkins@fb.com

Subscribers: trunkagent, mathieubaudet, njormrod

FB internal diff: D1579109

Tasks: 5241437

Summary:
not necessarily for commit but i was playing around a bit with this today to see what it might look like, so i figured i'd put it up
i assume this shenanigan isn't remotely safe (threadwise) without some extra magic... @fugalh we chatted about this last time i was in mpk a bit.
maybe addFuture could take an executor to make sure the promise is fulfilled on the correct thread, or something.

you'd have to reimplement timeouts via this executor or via futures if you wanted to get them via the futures channel, which makes sense anyway as this could be used with any Executor and not just ThreadPoolExecutor which owns that implementation

also specialized makeFutureTry to take functions that return futures but treat them like they return the contained types, so fulfil() can be used liked then()
this specialization could just as easily be done on fulfil() itself if we wanted.

Test Plan: unit

Reviewed By: davejwatson@fb.com

Subscribers: hannesr, trunkagent, craffert, njormrod, fugalh, bmatheny

FB internal diff: D1566369

Summary: fixed the one brought up in the task, and tweaked another that could theoretically cause problems

Test Plan: ran, though I have not been able to repro the

Reviewed By: njormrod@fb.com

Subscribers: fugalh, njormrod

FB internal diff: D1575632

Tasks: 5225808

Summary:
when event base is busy we're timing out tasks too early.
BIG SCARY NOTE: I don't have the thrift knowledge to know if there's sideeffects of this, please let
me know if this change is harmful.

Test Plan:
tested internally

Reviewed By: davejwatson@fb.com

Subscribers: njormrod

FB internal diff: D1578466

Tasks: 4752162

Summary: check external_ vs. AF_UNIX.  Also fix reset() to reset storage_.addr

Test Plan: fbconfig folly/test:network_address_test; fbmake runtests

Reviewed By: soren@fb.com

Subscribers: tudorb, trunkagent, doug, njormrod

FB internal diff: D1575098

Summary:
This gets almost all of folly building with buck.  There are a few
cases we don't support yet (custom test wrappers for spooky test
binaries, dlopen-enabled exception tracer binary, etc), but the
vast majority is covered.

Test Plan: built all build targets and ran all tests buck

Reviewed By: sdwilsh@fb.com

Subscribers: tjackson, sdwilsh, fugalh, njormrod

FB internal diff: D1577256

Tasks: 5055879

Summary:
Add a class_name method, which doesn't need to throw/catch
regardless how the exception is captured.

Test Plan: exception_wrapper_test

Reviewed By: mshneer@fb.com

Subscribers: njormrod

FB internal diff: D1543596

Tasks: 5025089

Blame Revision:

Summary:
Under moderate load, this part of the test would fail. When
there is resource contention, it is quite possible that the timeout will
be exceeded. A further reduction of the time spread will render this
test next to useless.

Moderate load can be as little as running all of folly's tests at once,
which is something that our testing infrastructure does quite
frequently.

fbconfig -r folly
fbmake runtests

Test Plan:
fbconfig -r folly && fbmake runtests

Reviewed By: hannesr@fb.com

Subscribers: sdwilsh, fugalh, njormrod

FB internal diff: D1576570

Tasks: 5225808

Summary:
Reverts the revert of the small_vector and test parts of D1569246.
I'm saving the SocketAddress fix for later, because it caused problems
and needs more research.

Test Plan:
unit tests with clang:dev/asan

Reviewed By: njormrod@fb.com

Subscribers: mathieubaudet, philipp, meyering, njormrod

FB internal diff: D1575574

Blame Revision: D1575190

Summary:
Use the new managed connection code in folly.  Cleans things up a fair bit

The only catch was the current ConnectionManager code uses raw pointers, thrift expects shared_ptr so the request object can keep the channel around until the requests are finished

Test Plan: fbconfig -r thrift/lib/cpp2; fbmake runtests

Reviewed By: afrind@fb.com

Subscribers: trunkagent, njormrod, doug, fugalh, alandau, bmatheny, dcsommer, afrind

FB internal diff: D1519923

Tasks: 5002343

Summary: This reverts commit 2c726dcf86bb176fb3695739a3d8bca2f95c41e6.

Test Plan: build it

Reviewed By: soren@fb.com

Subscribers: woo, mathieubaudet, njormrod

FB internal diff: D1575190

Summary: make these more serialized / event based so they don't get flakey under high load

Test Plan:
ran under load - caveat: i was not able to repro the flakiness @njormrod reported
but by inspection these should be fine

Reviewed By: njormrod@fb.com

Subscribers: fugalh, njormrod

FB internal diff: D1574640

Tasks: 5225808

Summary: I tried running folly/test with clang:dev and asan and found these leaks.

Test Plan:
unit tests with clang:dev/asan

Reviewed By: davejwatson@fb.com

Subscribers: davejwatson, trunkagent, mathieubaudet, njormrod, meyering, philipp

FB internal diff: D1569246

Summary:
Couple of notes:
1. is it a bummer not to have per-task callbacks of some kind? the interfaces set up here only tell you that some task expired, not which one expired. TM calls back with the Runnable object. is that useful?
2. std::chrono::* business is frustratingly verbose, but the safety/explicitness is nice. Not sure how I feel overall.
3. perhaps expirations should be given in microseconds even if we don't think we can accurately accomplish that

Test Plan: added unit

Reviewed By: hans@fb.com

Subscribers: fugalh, njormrod, bmatheny

FB internal diff: D1563520

Summary: Mcrouter open source build is failing because Codel is missing in fbthrift.

Test Plan: visual

Reviewed By: davejwatson@fb.com

Subscribers: fugalh, njormrod

FB internal diff: D1572122

Blame Revision: D1566128

Summary: Looks like the test explicitly looks for InvalidAddressFamilyException.  I don't think any users are depending on this particular error type either way, I just switched it to InvalidAddressFamilyException

Test Plan:
unit tests

Reviewed By: ldbrandy@fb.com

Subscribers: doug, njormrod

FB internal diff: D1570363

Tasks: 5153819

Summary: this way we can subscribe to an observable and blast data through it simultaneously from different threads

Test Plan: not much... the one client of rx compiles

Reviewed By: davejwatson@fb.com

Subscribers: fugalh, njormrod, bmatheny

FB internal diff: D1560647

Summary:
pool-wide stats via a call on the pool, and per-task stats (e.g. to be funneled into a histogram) via an rx subscription
rx needs a little work before this diff is safe - e.g. synchronization around the subscriber list, and perhaps exposing whether there are any subscribers so we can skip stat tracking if no one is listening
won't commit this without moving rx into folly/experimental of course

the idea is that timeout/expiration notifications can also go through the same subscription channel

haven't run the benchmarks yet and have to leave for the evening but tmrw i'll commit changes to the benchmark and get this stuff into windtunnel so i don't have to do so much manual output inspection on future diffs

Test Plan: added unit

Reviewed By: davejwatson@fb.com

Subscribers: fugalh, njormrod, bmatheny

FB internal diff: D1558424

Tasks: 5002392, 5002425

Summary: As above. I want to use this for the thread pools and it probably belongs in folly long-term anywya (if we stick with it)

Test Plan: compiled the one user

Reviewed By: hans@fb.com

Subscribers: fugalh, mwa, jgehring, fuegen, njormrod

FB internal diff: D1560578

Summary: under folly::wangle::

Test Plan: ran unit for experimental/wangle, thrift/lib/cpp/concurrency, and a user of FbThreadManager

Reviewed By: hans@fb.com

Subscribers: trunkagent, fbcode-common-diffs@, fugalh, alandau, bmatheny, everstore-dev@, njormrod

FB internal diff: D1566128

Summary: Pass const StringPieces by value instead of reference.

Test Plan:
fbconfig -r folly && fbmake runtests

Reviewed By: robbert@fb.com

Subscribers: trunkagent, sdwilsh, njormrod

FB internal diff: D1569488

Summary:
This is not a fix but me seeking advice here. The context here is that in liger (our common network
stack for mobile platforms), we saw bugs due to truncation and want to enable these warnings. To find those in our code,
I had to first resolve the folly ones. I just got around by making the truncation explicit so that I can get to real errors.
Ideally I would like owners to examine each case and fix the original/accept the explicit truncate if its intended. Let me
know what you think. The last (least preferred) is for us to keep this as a liger only change. We have a couple of ways to
do that but it would be nice to fix.

N.B : this covers only liger dependencies

Test Plan: errors resolved after these changes.

Reviewed By: njormrod@fb.com

Subscribers: trunkagent, doug, shilin, njormrod, seanc, pgriess

FB internal diff: D1509355

Summary:
MPMCQueueTest's intrusive reference count test implementation was
incorrect, so it simultaneously didn't test what it should and had a leak.

Test Plan: tests, plus new __thread to track lifetimes

Reviewed By: meyering@fb.com

Subscribers: njormrod, soren, meyering

FB internal diff: D1569448

Summary: Lines exceeding 80 characters have been restructured.

Test Plan:
fbconfig -r folly && fbmake runtests

Reviewed By: robbert@fb.com

Subscribers: trunkagent, sdwilsh, fugalh, njormrod

FB internal diff: D1568086