Summary: This fixes a deadlock possible when singleton chain is created concurrently with destroyInstances().

Reviewed By: lbrandy, yfeldblum

Differential Revision: D4329028

fbshipit-source-id: a11b3ff42d164ead2f8e3e77e0e17be43a8ad306

Summary:
GCC has the temerity to insinuate that my code has out-of-array-bounds access. After cross-checking with clang and ubsan, reviewing the code, and running constexpr tests (for which out-of-range errors are caught at compile time), I can say with pretty high confidence that this is an instance of GCC#61971 (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61971). Basically, the gcc-4 series is known to issue spurious -Warray-bounds warnings. The "fix" is to route internal accesses to some helper functions, for which -Warray-bounds has been suppressed.

User code that accesses elements with operator[] will still warn on out-of-bounds access. If this is a problem in practice, we can suppress the warning there, too. Trying this for now since it is less likely to hide real problems.

Reviewed By: yfeldblum

Differential Revision: D4317305

fbshipit-source-id: 7bf92f993ac1a29631463c582c1b64d76f755181

Summary:
A recent diff added [[noreturn]] annotations that cause buck
build to fail for ExceptionTracerTest.cpp, due to our use of -Werror
and -Wunreachable-code-return.  This diff removes the unreachable return
statement.

Differential Revision: D4326888

fbshipit-source-id: 7863c163eedcf2a7e19730c1de194f05c544a2c0

Summary: If your function doesn't return you should be explicit about it.

Reviewed By: meyering

Differential Revision: D4309893

fbshipit-source-id: ce275ec8f42e2cb3253a1e40e263934649f09d9e

Reviewed By: yfeldblum

Differential Revision: D4323918

fbshipit-source-id: 8dcd4ef3629fff061a9f149093ef1fe087f1c7e2

Summary: Find and remove a couple of unnecessary includes

Reviewed By: matbd

Differential Revision: D4322316

fbshipit-source-id: 7b55119bd7566fd62762ace5e35aa536489fafad

Reviewed By: igorsugak

Differential Revision: D4320762

fbshipit-source-id: 2028de6ccc6fb129381b03d694b88677a5aa50d6

Summary: Allow observers to be notified when AsyncSocket attaches and detaches from EVB

Reviewed By: siyengar

Differential Revision: D4256175

fbshipit-source-id: a3ff96811f885e508f20cf11ce52e0f00e5ee461

Summary: The most common place this happened was in tests where it was being used to force the return type of a lambda, but there were a couple of places in the main code that triggered this as well.

Reviewed By: yfeldblum

Differential Revision: D4310187

fbshipit-source-id: e3701cff9827eacaf3be8d28296441466eb2fa11

Summary:
Moving russoue's `JemallocNodumpAllocator` (D3147173) from
`mcrouter` codebase to `folly`, so we can use it in a few other services
that suffer from a huge core dump problem.

Reviewed By: russoue, jmswen, luciang

Differential Revision: D4311394

fbshipit-source-id: 6a13c478b939bd411e0fd37e655125f62434c366

Summary: Adding new `addFunctionOnce` method that executes provided callback only once.

Reviewed By: yfeldblum

Differential Revision: D4315635

fbshipit-source-id: 4819ec30b4f2e4ab3185a37158404e1c7a96758a

Summary: Prior to C++14 these shifts are undefined behavior, but the unsigned version is not, so do the shifts on unsigned values before converting to the final type.

Reviewed By: yfeldblum

Differential Revision: D4309311

fbshipit-source-id: 914b207bac2f77a96c07a8a5df81980c672aa677

Summary: Which warns about the use of a ternary statement with the first operand omitted. This is a GCC extension and MSVC does not support it, so eliminate the last use of it.

Reviewed By: yfeldblum

Differential Revision: D4309989

fbshipit-source-id: 04c968708e47f8cb707fd0892e8780bc676df0de

Summary: They aren't needed and they count as initialized but unused variables.

Reviewed By: yfeldblum

Differential Revision: D4308844

fbshipit-source-id: a6833dbf434ebdefff0b375729a4e62495463ac0

Summary: Split out of an earlier pair of diffs, this focuses exclusively on Folly's networking APIs and works to eliminate places where values were being implicitly truncated.

Reviewed By: yfeldblum

Differential Revision: D4288042

fbshipit-source-id: dd6e19acd319296a45c29c8050dc62f06571e1e6

Summary: Because the `break;` can be dead, and sometimes is indicitive of a different type of error.

Reviewed By: meyering

Differential Revision: D4310056

fbshipit-source-id: ad215eb9b2e5bb4d5c703582203efce893b26a76

Summary: To monitor client support of TLS 1.3.

Reviewed By: ngoyal

Differential Revision: D4308473

fbshipit-source-id: cb6fb444c8b7ced39e6655a0f63b18523c2fb9c5

Summary:
In AsyncSSLSocket tests the shared_ptr structure means that the
AsyncSSLSocket-s outlive the stack-allocated EventBase on which they were
created.  Previously there were scattered calls to detachEventBase on the
last interesting callback, but several calls were missing.  This diff
switches to a model where the SSLServerAcceptCallbackBase is responsible
for detaching the sockets.

This diff also fixes a low-firing race between shutdown
of the TestSSLAsyncCacheServer Main thread and a call to
EmptyReadCallback::readEOF.  Most uses of EmptyReadCallback don't attach
the TCP socket, so they can't actually tolerate a call to readError
or readEOF.

These use-after-destruction problems were discovered by UBSAN.

Reviewed By: djwatson

Differential Revision: D4301416

fbshipit-source-id: 127fb5e506d0694c5ca81d7a38c704d69f4ab3eb

Summary:
Because the extra semicolons are useless and all typos...
GCC doesn't have this warning. Neither does MSVC, but that won't stop me.

Reviewed By: igorsugak

Differential Revision: D4304218

fbshipit-source-id: a06586edc6b9bafd5dd074e9fd4e0ed152181abc

Summary:
This accounts for the places that were triggering warnings 4456, 4457, and 4458, which are all related to shadowing names, be they locals, parameters, or even types.
This doesn't deal with 4459, which is specifically for shadowing global variables, because folly/gen defines globals by the name of `count`, `min`, `max` and a few other similar names.

Reviewed By: meyering

Differential Revision: D4296781

fbshipit-source-id: a2e625095e2c65a53a9226b000aaf0ca95a3a393

Summary:
Do the barrier completion test in a loop. (std::condition_variable has pthreads-style spurious wakeups).

(Sorry for missing this in review).

Reviewed By: djwatson

Differential Revision: D4302035

fbshipit-source-id: 3322d6a0ffba8c47c46bafb1d88034e1a0a9c652

Summary: Add fairness benchmark, contended benchmark

Reviewed By: davidtgoldblatt

Differential Revision: D3706680

fbshipit-source-id: 01814fcd6b0cc3dd12815a2b2c248abc6a3771ec

Summary: It deletes things even when input isn't in the container.

Reviewed By: luciang

Differential Revision: D4298340

fbshipit-source-id: 3e8fc04c2c21eb231dcaf82239ac5f6d25e49e2c

Summary:
`condition_variable::wait_for` takes as second param `std::chrono::duration`
and not a predicate; what we want is `condition_variable::wait`.

Reviewed By: simpkins

Differential Revision: D4295305

fbshipit-source-id: 05f735fe6e7ecb9d8f42cb38a2985b9ce9dad984

Summary: MSVC has the ability to warn about implicit truncations and places where implicit sign coercions are occuring, so do some cleanup to make it possible to compile with the warnings enabled.

Reviewed By: yfeldblum

Differential Revision: D4288028

fbshipit-source-id: f8330c62b2dcb76f696dfc47888f0e3e1eefc21a

Summary: [Folly] Cut some includes from `Executor.h`.

Reviewed By: mzlee

Differential Revision: D4256618

fbshipit-source-id: 1728f48b3c0ec4e0c4c84c44bd8afb583d438129

Summary:
Nothing defines USE_JEMALLOC in the OSS build today and that causes
some portability problems.

Specifically, the homebrew recipe will make libjemalloc available and our
configure script will detect it and add it to the library flags.

Our subsequent check for `malloc_usable_size` then finds this function in
libjemalloc.

When later attempting to build wangle against the homebrew folly we get
compilation failures because the prototype for `malloc_usable_size` is only
available in the jemalloc headers and nothing in the saved configuration for
folly is set up for this to be pulled in as it it guarded by `USE_JEMALLOC`.

This attempts to resolve the situation by forcing on `USE_JEMALLOC` when
we detect the library in configure.

This is made a little more complicated because we cannot set `USE_JEMALLOC`
in the OSS build; it gets rewritten to have a `FOLLY_` prefix.  Since we
have code outside of folly that requires that this symbol be `USE_JEMALLOC`,
I've changed the conditional to check for both flavors of the symbol, with
and without the prefix.

Reviewed By: yfeldblum

Differential Revision: D4289176

fbshipit-source-id: 756bc815c3ef1fac454e603feb72155d98c5aadd

Summary: Inspired by http://wg21.link/p0259, folly::BasicFixedString is a std::string replacement that is usable in constexpr contexts. It uses an internal fixed-size buffer to store up to N characters, where N is a template parameter.

Reviewed By: luciang

Differential Revision: D4249529

fbshipit-source-id: 10f12ea2510a8d21ec60a07a8f15ecc3e35c5431

Summary:
`stdin`, `stdout` and `stderr` are macros that expand to function calls with the MSVC CRT implementation. This is also the case for musl-libc. This means that Subprocess simply cannot be compiled on those platforms without changing the API.
To solve that, we change the API and deprecate the old API.

For more fun, `stdin`, `stdout` and `stderr` are also macros in glibc, they just expand to other identifiers rather than a function call.

Reviewed By: yfeldblum

Differential Revision: D4229544

fbshipit-source-id: 97f1a3b228b83cfdcaffee56d729063ea235e608

Summary: Fix use-after-destroy of NewThreadExecutor.  Detected by opt-ubsan tests.

Reviewed By: yfeldblum

Differential Revision: D4291900

fbshipit-source-id: 206f1d7b80fa04e1570e4b7a20e50b9a8ceb7bcb

Summary:
Freshly opened TCP sockets can have a send buffer size bigger
than wmem_default, because connection parameters are cached in the
route cache.  Some of the async socket tests assumed that writing 8M of
data was enough to guarantee blocking behavior, but since wmem_max is
20M on our systems the resulting tests are flaky.

Reviewed By: marcinpe

Differential Revision: D4292201

fbshipit-source-id: ba5d606d080330e455eee2b17bcae6cf546bf981

Summary:
Add the default move constructor and move operator
to EvictingCacheMap so it can be moved.

Reviewed By: markisaa, luciang

Differential Revision: D4287472

fbshipit-source-id: 130e9d5467d6da14ba95a9e769cf8e8d541a704c

Summary: UBSAN instrumentation results in more stack being used. Applying the same logic we already have for ASAN there.

Reviewed By: igorsugak

Differential Revision: D4287455

fbshipit-source-id: 1a224cb5a3654c23b15fa298bf80476234a1418c

Summary: MacAddress only needs to include ostream and the test header does not need iostream.

Differential Revision: D4286405

fbshipit-source-id: e64fdd4e8428c62279ed7c185188f3dfc14096d3

Summary:
We do not need it and it adds a static initialization

This is the last extra one in liger. The other places actually need iostream to print things

Reviewed By: yfeldblum

Differential Revision: D4286095

fbshipit-source-id: 44a827e48e9fb9106626e784592eb194dff83bf6

Summary:
The benchmark API only supports `unsigned int` as the iteration number, so this being 64-bit means it has to be truncated for no reason.
This change gets us closer to being able to enable MSVC's implicit truncation warnings.

Reviewed By: yfeldblum

Differential Revision: D4281264

fbshipit-source-id: 27397e3f23bace20f3cc457665ea6f21bf994da5

Summary:
I was getting an error when running commands in
https://our.intern.facebook.com/intern/dex/fibers/gdb/

Reviewed By: andriigrynenko

Differential Revision: D4283229

fbshipit-source-id: fe45847d0ec6cadfbce6fe9818066cce3e275a46

Summary: overload constexpr_strlen to work with other character types

Reviewed By: yfeldblum, luciang, Orvid, ot

Differential Revision: D4269654

fbshipit-source-id: 2fcd666669efb51d77e847ff31bda3a301e763da

Reformulate constexpr_min and constexpr_max to achieve stability in sorting as described in http://stepanovpapers.com/notes.pdf

Summary: There is a famous long-standing "bug" in the standard library regarding the semantics of min and max wrt values that are equivalent wrt op< but not equal. Let's not make the same mistake with constexpr_min and constexpr_max.

Reviewed By: yfeldblum, luciang, Orvid, ot

Differential Revision: D4269635

fbshipit-source-id: 19b464c949dc0cf07afb08eaf657ae8b242ca42d

Summary:
Fix potential double close() on exception when SO_REUSEPORT fails.
The fd will be closed twice, once by close() and second by the scope guard above.

Reviewed By: yfeldblum

Differential Revision: D4272300

fbshipit-source-id: a163be822d7522cadb9dc3c3eddca10df453de14