Summary:
[Folly] Use simple `exception_wrapper` ctor in the `Promise::setException` test.

And add a new case testing with `exception_ptr`.

Reviewed By: andrewjcg

Differential Revision: D5683035

fbshipit-source-id: e8276166dacbe09a9a745271d636db44c593058c

Summary:
Adding a for_each function that allows generalized indexed and breakable iteration through ranges, these can either be runtime ranges (i.e. entities for which std::begin and std::end work) or compile time ranges (as deemed by the presence of a std::tuple_length<>, get<> (ADL resolved) functions)

The function is made to provide a convenient library based solution to the proposal p0589r0, which aims to generalize the range based for loop even further to work with compile time ranges

A drawback of using range based for loops is that sometimes you do not have access to the index within the range.  This provides easy access to that, even with compile time ranges.

Further this also provides a good way to break out of a loop without any overhead when that is not used.

A simple use case would be when using futures, if the user was doing calls to n servers then they would accept the callback with the futures like this

   auto vec = std::vector<std::future<int>>{request_one(), ...};
   when_all(vec.begin(), vec.end()).then([](auto futures) {
     folly::for_each(futures, [](auto& fut) { ... });
   });

Now when this code switches to use tuples instead of the runtime std::vector, then the loop does not need to change, the code will still work just fine

   when_all(future_one, future_two, future_three).then([](auto futures) {
     folly::for_each(futures, [](auto& fut) { ... });
   });

Reviewed By: yfeldblum

Differential Revision: D5557336

fbshipit-source-id: 79fcbafa7e1671f8856f0dcb7bf7996435dadeaa

Summary:
Even though MockAsyncSSLSocket's initializes AsyncSSLSocket, which should
initialize AsyncSocket, this does not actually happen because AsyncSSLSocket
virtually inherits from AsyncSocket.

Because of this, prior to this diff, the MockAsyncSSLSocket was not properly
setting its EventBase.

Reviewed By: knekritz

Differential Revision: D5676596

fbshipit-source-id: 5f3c0e848179cb5eb4d2dc4921a11e7c04d7c0e0

Summary:
[Folly] Use `exception_wrapper::from_exception_ptr` in `Try` and `Promise`.

It does just what we need; DRY.

Reviewed By: andrewjcg

Differential Revision: D5674663

fbshipit-source-id: 73e46df62c06736c4eaf013d05dfea819cbec515

Summary:
- Create interface to initialize zlib codec using specific parameters
- This enables the raw inflate/deflate and auto inflate options
- Add tests for option initialization

Reviewed By: terrelln, yfeldblum

Differential Revision: D5649980

fbshipit-source-id: fd36e8edc0e8c528cd6c9d8f39e8ef839b6acfef

Summary:
[Folly] `exception_wrapper::from_exception_ptr`.

A handry helper for converting from `std::exception_ptr` to `folly::exception_wrapper`.

Reviewed By: ericniebler

Differential Revision: D5668179

fbshipit-source-id: a81a60cb22a2a697714268e027af62dd8825d2c3

Summary: Fairly minimal API to start with. Refer to the comments for the rationale/usage.

Reviewed By: yfeldblum

Differential Revision: D5554854

fbshipit-source-id: 173bf7c40e70b55bf6afb8c4bc9e83d48f90b6ee

Summary: A SSLOptions recommended for cipher use.

Reviewed By: yfeldblum

Differential Revision: D5614280

fbshipit-source-id: a6b1adfa8d168f35c7bc7d4088c4073c3f4084a5

Summary:
Make SSLContext::setCipherList() and SSLContext::setSignatureAlgorithms()
accept std::array besides std::vector so that cipher lists in SSLOptions
can be initialized POD.

Reviewed By: yfeldblum

Differential Revision: D5578758

fbshipit-source-id: 7e5c2e9a75600e93c89e7b13a9042434a4189384

Summary:
This was added several jemalloc versions ago, and is no longer a win.

From microbenches in folly/test, this codepath is successful less than .5% of the time.

There are a handful of other places directly using this (fbvector, IOBuf, ThreadLocal),
but should probably change them individually.

Reviewed By: interwq

Differential Revision: D5596678

fbshipit-source-id: 354d7204f61df0eace4d98d930d0f6e06a90ffde

Summary:
TAsyncSSLServerSocket was deleted in D1806807, so we should remove
these comments.

Reviewed By: yfeldblum

Differential Revision: D5657431

fbshipit-source-id: ca875293737ad8ba0a8c028c9bfa5651f4a6065f

Summary:
The format is incorrect in [Overview.md](https://github.com/facebook/folly/blob/master/folly/docs/Overview.md)
Closes https://github.com/facebook/folly/pull/661

Reviewed By: eduardo-elizondo

Differential Revision: D5657083

Pulled By: yfeldblum

fbshipit-source-id: a9cedb413cbb3d455f4c1e5f785d05a33f8182c8

Summary: Left shifting `0x7f` by `63` is undefined behavior (i.e. when decoding `0xFFFF...`)

Reviewed By: yfeldblum, terrelln

Differential Revision: D5653353

fbshipit-source-id: c74c9f43a9bc82d15a2223df853dc533cea1478b

Reviewed By: knekritz

Differential Revision: D5632201

fbshipit-source-id: 5db380379ab6cb608922c77dd716bfc4410b2cd8

Summary: Return empty string for `getSecurityProtocol` if the socket is good and the handshake failed, otherwise "TLS"

Reviewed By: knekritz

Differential Revision: D5647637

fbshipit-source-id: b95cbf39e4bb7f89b1ebcbc0238c2becba7ad42a

Summary:
Specifically, the constructor `implicit Range(Iter)` is now declared and defined for all `Range` specializations. It is an error to use it, and we `static_assert`, unless `Iter` is `char const *` or `char *`.

Performance effect
 ---
Measuring compilation-time on a file that just make ~40k StringPieces, compiled with -O3. All compilers produced identical code before/after this change.

* clang-trunk: 4% improvement
* gcc-5: no change
* gcc-4.9: 11% improvement
* gcc-7.1: 5% improvement

Could this possibly break any existing code?
 ---
Yes. If you have a function that's overloaded for both `Range<char const*>` and `Range<X>` and your input object is not `char const*` and is implicitly convertible to both `char const*` and `X`: that is now ambiguous whereas before it would unambiguously pick the `Range<char const*>` overload. I don't consider this scenario likely.

Why should this work?
 ---
Using SFINAE is more expensive at compile time (with some compilation environments) than not using it. It's necessary to use SFINAE when there is an alternative function which will be used in preference when the substitution fails, but when that is not the case it is on average cheaper to make the function always exist and use static_assert to disallow the bad uses of it. A bonus is that the caller gets a more comprehensible error message.

Reviewed By: nbronson

Differential Revision: D5639502

fbshipit-source-id: 13469f2995a487398734f86108087fdc8e32ad71

Summary:
This makes it so that the //folly:config rule can run autoconf to create folly-config

- Uses cxx_genrule get compiler settings and paths to dependent libraries
- Goes around a couple of oddities in the way that buck exports some of its macros
- While it uses buck_cxx_library, that's okay to use outside of facebook's internal repos.

Reviewed By: meyering

Differential Revision: D5620729

fbshipit-source-id: 0d2d8e3bda92182dcdcd3e80cb12a3756b3816ac

Summary:
`prepareSkipTo(x`) allows the client of EliasFanoReader to "hint" that
`skipTo(x)` will be called in the near future. The primary benefit of doing so
is that memory which is needed for `skipTo(x)` can be prefetched to minimize
cache misses incurred when calling `skipTo(x)`.

Reviewed By: ot, philippv

Differential Revision: D5508995

fbshipit-source-id: 4876b566256849f76193db3dc0404768aeeeb30d

Summary:
ConcurrentHashMap is a newly added feature and we wanted to test it in our enviornment, making it installable.

Built & Installed on ubuntu 14 env.
Closes https://github.com/facebook/folly/pull/659

Reviewed By: djwatson

Differential Revision: D5629159

Pulled By: yfeldblum

fbshipit-source-id: 8a40e768f3db7918a3b284059f2efa0b7ecb9aca

Summary: Adding iterateProgramHeaders, similar to iterateSections, just iterates over program headers instead of section headers. I want this to get the size of the first PT_LOAD header.

Differential Revision: D5602575

fbshipit-source-id: f73989cade20214f884571c1099761ecaa4841f7

Summary:
- Added `hazptr.cpp` and `memory_resource.cpp`. Moved singleton code to source.
  - Changed the singleton for the default `hazptr_domain` to a global.
  - Changed hazptr_stats singleton to a global.
  - Moved the thread caching calls from the hazptr_domain functions to the constructor/destructor of hazptr_holder.
  - Changed the TLS singletons to TLS globals.
  - Changed some inlining directives.
  - Leak the hazptr_rec-s in the default domain

Reviewed By: davidtgoldblatt, djwatson

Differential Revision: D5553753

fbshipit-source-id: da69eecec55c0f78fb8ef5591f9aeffee99ff3fa

Summary: add missing source files to Makefile

Reviewed By: andriigrynenko

Differential Revision: D5630758

fbshipit-source-id: 0f04c11b65fd3a6f4f33e34ae57c0fcba8671e97

Summary:
Keep track of which threads belong to which thread pools for use when debugging.
This will be paired with a gdb script that associates threads with their thread pools

Reviewed By: andriigrynenko

Differential Revision: D5514729

fbshipit-source-id: 57ada4dd1aaaec5d7026e4eee05b0ec4e7434c77

Summary:
`std::optional` supports comparing with `std::nullopt`, so
`folly::Optional` should do the same with `folly::none`.
This also involves marking hasValue() noexcept to be the same as `std::optional`.

Reviewed By: yfeldblum, WillerZ

Differential Revision: D5617825

fbshipit-source-id: a4b863dd61c3a86223fb21a5b7759e7c295fd272

Summary:
The added API will be used for storing and retrieving validation result of client cert.

At some point we may change "client" to "peer" to make it generic.

Reviewed By: anirudhvr

Differential Revision: D5610404

fbshipit-source-id: 2feb03abc0594af4c68db33d8885ea214867148a

Summary:
Containers should not tag their default constructor explicit,
because it means that you can't return an empty one with "return {};".
This diff removes the explicit tag from small_vector's default
constructor.

Reviewed By: meyering

Differential Revision: D5602749

fbshipit-source-id: 2a4a356134c69d01e3b040719dc6ffca0fe97e75

Summary: this fixes the case where bulk inserting a range where the smallest element is equal to the largest element of the current set/map fails to deduplicate that element.

Reviewed By: mlogan, yfeldblum

Differential Revision: D5593284

fbshipit-source-id: 487500ee7a5e33f27c24321ad4a3c07a669fc26c

Summary:
[Folly] Explicitly handle gcc in `FOLLY_MAYBE_UNUSED`.

Fixes build break from {D5562220}.

Reviewed By: Orvid

Differential Revision: D5594288

fbshipit-source-id: 5aa2dcb6133d53a0eb20d34b67c8e4407dd2b7ca

Summary: [Folly] Cut unnecessary parens in `IPAddress.h`.

Reviewed By: meyering

Differential Revision: D5589756

fbshipit-source-id: 6499e0db2aeb232356e2808fc59a35869b5dad0a

Summary: [Folly] Let `IPAddressV4::bitCount` and `IPAddressV6::bitCount` be `constexpr`.

Reviewed By: meyering

Differential Revision: D5589954

fbshipit-source-id: f4208efd62e71f47a7c87bdc8ddd421f93413c29

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

Reviewed By: Orvid

Differential Revision: D5581523

fbshipit-source-id: 97b5270e43e279e7deb9606524d5fee844a50649

Summary:
Use it like this:
```
using namespace folly::string_piece_literals;
StringPiece p = "A literal string"_sp;
```

In some compilation environments it can be more efficient than the implicit
conversion from `char const *` to `StringPiece`.

Reviewed By: yfeldblum

Differential Revision: D5562782

fbshipit-source-id: ce715edc65b1510761e127bf89a6936370253a68

Summary: `Future::onError` does not work with a function that takes `exception_wrapper&`, so it should not compile. `exception_wrapper`, `const exception_wrapper&`, and `exception_wrapper&&` are unaffected.

Reviewed By: yfeldblum

Differential Revision: D5556083

fbshipit-source-id: 45d096d9fee00322eeec6b9b9b1dcbef55cced03

Summary:
[Folly] Consistent use of `sformat` in address-related files.

V.s. `format(/*...*/).str()` and v.s. `to<std::string>`.

Reviewed By: meyering

Differential Revision: D5570334

fbshipit-source-id: 83aedf9a694721fb209e62e94f1a5c5ecd355e81

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

Reviewed By: meyering

Differential Revision: D5570553

fbshipit-source-id: 230aa8557c2d402d741aaa541f96f4634dd13d3b

Summary: Make sure server can get client cert using AsyncSSLSocket::getPeerCert() call.

Reviewed By: anirudhvr

Differential Revision: D5557303

fbshipit-source-id: d95696f592e3b3f67acba9f92add32b91c29a000

Summary: ... as titled.

Reviewed By: yfeldblum

Differential Revision: D5558742

fbshipit-source-id: b63b121cde8db93de4cabc80563539297611d600

Summary:
Clang compiler does not define `__CLANG__` macro and
only defines `__clang__`, but we use `__CLANG__` in a
few places in the code which was probably done by mistake.
This should make the code behave in the intended way.

Reviewed By: yfeldblum

Differential Revision: D5542162

fbshipit-source-id: 74d24ba9464d8faec160d73e6c582967d3c80c19

Summary: Adds unique pointer type for X509_EXTENSION

Reviewed By: knekritz

Differential Revision: D5556504

fbshipit-source-id: b84190f20dd0a2ee9f5f07f197caf53a9a063af7

Summary: Remove a bunch of identical code that only exists to provide interfaces for both `const MessageT&` and `MessageT&&` arguments.

Reviewed By: ot

Differential Revision: D5540560

fbshipit-source-id: ae7b9345b64130fc748033be2e0587c4f83345a7