Commit Graph

5712 Commits

Author SHA1 Message Date
Louis Dionne
dfcd4384cb [libc++] Implement P0433: deduction guides for <unordered_map>
Thanks to Arthur O'Dwyer for the patch.

Differential Revision: https://reviews.llvm.org/D58590

llvm-svn: 366124
2019-07-15 20:06:01 +00:00
Eric Fiselier
c5d7b0c454 Constrain workaround to avoid affecting other buildbots
llvm-svn: 366122
2019-07-15 19:53:42 +00:00
Eric Fiselier
4066978cb7 Improve compile time of variant.
In particular, improve the compile time of the overload set builder
that variant uses to determine which alternative to construct.

Instead of having the __overload type construct itself recursively,
this patch uses a flat construction for the overload set.

llvm-svn: 366033
2019-07-14 21:29:39 +00:00
Eric Fiselier
3c0e2bb0cb Add test for variant construction with duplicate types.
llvm-svn: 366032
2019-07-14 20:59:51 +00:00
Eric Fiselier
aae0cb67ed Cleanup whitespace in <variant>. NFC.
llvm-svn: 366026
2019-07-14 18:31:55 +00:00
Eric Fiselier
273857d1ea Harden variant test added in r366022
The test was brittle since it only went boom for one specific type, when
really it should go boom for all of them.

llvm-svn: 366025
2019-07-14 18:30:34 +00:00
Eric Fiselier
194b337f33 Avoid eager template instantiation caused by the variant narrowing checks.
The standard disallows narrowing conversions when constructing a variant.
This is checked by attempting to perform braced initialization of the
destination type from the argument type. However, braced initialization
can force the compiler (mostly clang) to eagerly instantiate the
constructors of the destintation type -- which can lead to errors in
a non-immediate context.

However, as variant is currently specified, the narrowing checks only
observably apply when the destination type is arithmetic. Meaning we can
skip the check for class types. Hense avoiding the hard errors.

In order to cause fewer build breakages, this patch avoids the narrowing
check except when the destination type is arithmetic.

llvm-svn: 366022
2019-07-14 18:21:15 +00:00
Eric Fiselier
882fdf68b7 Fix non-conformance it std::tuple.
Previously we implemented all one trillion tuple-like constructors using
a single generic overload. This worked fairly well, except that it
differed in behavior from the standard version because it didn't
consider both T&& and T const&. This was observable for certain
types.

This patch addresses that issue by splitting the generic constructor
in two. We now provide both T&& and T const& versions of the
tuple-like constructors (sort of).

llvm-svn: 365973
2019-07-12 23:01:48 +00:00
Eric Fiselier
128af31595 Add option to disable variant narrowing conversion changes.
The paper P0608R3 - "A sane variant converting constructor" disallows
narrowing conversions in variant. It was meant to address this
surprising problem:

  std::variant<std::string, bool> v = "abc";
  assert(v.index() == 1); // constructs a bool.

However, it also disables every potentially narrowing conversion. For
example:

  variant<unsigned> v = 0; // ill-formed
  variant<string, double> v2 = 42; // ill-formed (int -> double narrows)

These latter changes break code. A lot of code. Within Google it broke
on the order of a hundred thousand target with thousands of root causes
responsible for the breakages.

Of the breakages related to the narrowing restrictions, none of them
exposed outstanding bugs. However, the breakages caused by boolean
conversions (~13 root causes), all but one of them were bugs.

For this reasons, I am adding a flag to disable the narrowing conversion
changes but not the boolean conversions one.

One purpose of this flag is to allow users to opt-out of breaking changes
in variant until the offending code can be cleaned up. For non-trivial
variant usages the amount of cleanup may be significant.

This flag is also required to support automated tooling, such as
clang-tidy, that can automatically fix code broken by this change.
In order for clang-tidy to know the correct alternative to construct,
it must know what alternative was being constructed previously, which
means running it over the old version of std::variant.

Because this change breaks so much code, I will be implementing the
aforementioned clang-tidy check in the very near future.

Additionally I'm plan present this new information to the committee so they can
re-consider if this is a breaking change we want to make.

I think libc++ should very seriously consider pulling this change
before the 9.0 release branch is cut. But that's a separate discussion
that I will start on the lists.

For now this is the minimal first step.

llvm-svn: 365960
2019-07-12 21:32:11 +00:00
Louis Dionne
189639b1c3 [libc++] Add XFAILs for CTAD tests on older compilers
llvm-svn: 365923
2019-07-12 17:30:57 +00:00
Eric Fiselier
d49103bd2f Mark destroying delete test as UNSUPPORTED with clang 7
llvm-svn: 365856
2019-07-12 01:16:08 +00:00
Eric Fiselier
a780216fb3 Tolerate import errors in "not.py" implementation
llvm-svn: 365855
2019-07-12 01:13:05 +00:00
Marshall Clow
f3b851f0b8 Reorganize the 'bit' header to make most of the facilities available for internal use pre-C++20. NFC for external users
llvm-svn: 365854
2019-07-12 01:01:55 +00:00
Eric Fiselier
02fdc51c04 Add another buildbot username to the workaround list
llvm-svn: 365848
2019-07-12 00:16:38 +00:00
Eric Fiselier
ce13f67618 Attempt to override broken buildbot config for libc++abi.
The buildbots were changed to pass -DLIBCXX_CXX_ABI=libcxxabi, but
they don't provide an include path for the library, so cxxabi.h is
never found while building libc++.

This is a temporary change until the buildbots are updated or until
D63883 lands in a form that unbreaks the bots

llvm-svn: 365847
2019-07-12 00:02:49 +00:00
Eric Fiselier
41798c05cd Fix memory leak in set and map.
When assigning an initializer list into set/map, libc++ would
leak memory if the initializer list contained equivalent keys
because we failed to check if the insertion was successful.

llvm-svn: 365840
2019-07-11 23:13:38 +00:00
Louis Dionne
296a80102a [libc++] Implement deduction guides for <unordered_set>
Thanks to Arthur O'Dwyer for the patch.
Differential Revision: https://reviews.llvm.org/D58617

llvm-svn: 365788
2019-07-11 15:16:39 +00:00
Saleem Abdulrasool
97d6fcce4e build: use multiple install rather than building up a list
Rather than building up a list to iterate over later, just create multiple
install commands based on the configuration. This makes it easier to see what
is getting installed and allows for the install handling to be centralised. NFC

llvm-svn: 365562
2019-07-09 21:43:01 +00:00
Vitaly Buka
a9d5c186e2 Revert "[libc++] Take 2: Do not cleverly link against libc++abi just because it happens to be there"
r365326 still breaks bots:
http://lab.llvm.org:8011/builders/netbsd-amd64/builds/20712/steps/ninja%20build%20local/logs/stdio
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-autoconf/builds/39477/steps/test%20tsan%20in%20debug%20compiler-rt%20build/logs/stdio

And probably others

This reverts commit 945b9ec069.

llvm-svn: 365359
2019-07-08 17:46:23 +00:00
Louis Dionne
945b9ec069 [libc++] Take 2: Do not cleverly link against libc++abi just because it happens to be there
Summary:
Otherwise, when libcxxabi is not an enabled project in the monorepo, we
get a link error because we try to link against non-existent cxxabi_shared.

More generally, we shouldn't change the behavior of the build based on
implicit things like whether a file happens to be at a specific path or
not.

This is a re-application of r365222 that had been reverted in r365233
because it broke the build bots. However, the build bots now specify
explicitly what ABI library they want to use (libc++abi), so this
commit should now be OK to merge.

Differential Revision: https://reviews.llvm.org/D63883

llvm-svn: 365326
2019-07-08 14:49:35 +00:00
Marshall Clow
f814dcbafb Make forward_list::remove/remove_if/unique all return void before C++20; undoes that bit of D58332. Thanks to Mikhail Maltsev for pointing this out
llvm-svn: 365290
2019-07-08 03:45:28 +00:00
Eric Fiselier
8cedf04a6c Make ~mutex and ~condition_variable trivial on Windows.
The implementations of __libcpp_mutex_destroy and __libcpp_condvar_destroy
are already NOPs, so this optimization is safe to perform.

See r365273 and PR27658 for more information.

llvm-svn: 365281
2019-07-07 17:24:03 +00:00
Eric Fiselier
8baf83839e Fix PR27658 - Make ~mutex trivial when possible.
Currently std::mutex has a constexpr constructor, but a non-trivial
destruction.

The constexpr constructor is required to ensure the construction of a
mutex with static storage duration happens at compile time, during
constant initialization, and not during dynamic initialization.
This means that static mutex's are always initialized and can be used
safely during dynamic initialization without the "static initialization
order fiasco".

A trivial destructor is important for similar reasons. If a mutex is
used during dynamic initialization it might also be used during program
termination. If a static mutex has a non-trivial destructor it will be
invoked during termination. This can introduce the "static
deinitialization order fiasco".

Additionally, function-local statics emit a guard variable around
non-trivially destructible types. This results in horrible codegen and
adds a runtime cost to every call to that function. non-local static's
also result in slightly worse codegen but it's not as big of a problem.

Example codegen can be found here: https://goo.gl/3CSzbM

Note: This optimization is not safe with every pthread implementation.
Some implementations allocate on the first call to pthread_mutex_lock
and free the allocation in pthread_mutex_destroy.

Also, changing the triviality of the destructor is not an ABI break.
At least to the best of my knowledge :-)

llvm-svn: 365273
2019-07-07 01:20:54 +00:00
Marshall Clow
1ab3fe8a7a Make list::remove/remove_if/unique all return void before C++20; undoes that bit of D58332. Thanks to Mikhail Maltsev for pointing this out
llvm-svn: 365261
2019-07-06 06:10:03 +00:00
Zoe Carver
28e0187175 This patch makes swap functions constexpr. Both swap overloads, swap_ranges and iter_swap are updated (with tests).
llvm-svn: 365238
2019-07-05 20:13:34 +00:00
Louis Dionne
51de516486 Revert "[libc++] Do not cleverly link against libc++abi just because it happens to be there"
This reverts r365222, which broke the libc++ build bots.

llvm-svn: 365233
2019-07-05 18:44:25 +00:00
Louis Dionne
4ba29e5fbe [libc++] Do not cleverly link against libc++abi just because it happens to be there
Summary:
Otherwise, when libcxxabi is not an enabled project in the monorepo, we
get a link error because we try to link against non-existent cxxabi_shared.

More generally, we shouldn't change the behavior of the build based on
implicit things like whether a file happens to be at a specific path or
not.

Reviewers: EricWF

Subscribers: mgorny, christof, jkorous, dexonsmith, libcxx-commits

Tags: #libc

Differential Revision: https://reviews.llvm.org/D63883

llvm-svn: 365222
2019-07-05 17:06:23 +00:00
Saleem Abdulrasool
bce9d8b138 docs: add documentation for LIBCXX_INCLUDE_TESTS
Add some missing documentation for the `LIBCXX_INCLUDE_TESTS` option.

Patch by Jean Heyd Meneide!

llvm-svn: 365154
2019-07-04 19:08:16 +00:00
Marshall Clow
7afe9ea96f Add tests for regex_match ambiguity (aka LWG2273). NFC. Reviewed as https://reviews.llvm.org/D63051
llvm-svn: 365080
2019-07-03 20:32:35 +00:00
Eric Fiselier
f1807a7df6 Fix tuple's conditionally explicit constructors for very weird user
types.

It seems some people like to write types that can explicitly convert
to anything, but cannot be used to explicitly construct anything.

This patch makes tuple tolerate such types, as is required
by the standard.

llvm-svn: 365074
2019-07-03 19:21:40 +00:00
Louis Dionne
1c478d6e85 [libc++] Update availability markup for Filesystem on Apple platforms
llvm-svn: 365068
2019-07-03 18:29:02 +00:00
Marshall Clow
2a622b30e3 Update status of papers for upcoming WG21 meeting. NFC
llvm-svn: 364885
2019-07-02 03:23:06 +00:00
Marshall Clow
bd7f84a482 Use new '__libcpp_is_constant_evaluated' call to remove an '#ifdef' from the bit code. NFC
llvm-svn: 364884
2019-07-02 03:21:16 +00:00
Marshall Clow
745379a0af Mark the newly added '__libcpp_is_constant_evaluated' as 'inline', since it can be included multiple times by multiple headers, and we don't want 'duplicate definition' errors.
llvm-svn: 364879
2019-07-02 00:20:06 +00:00
Marshall Clow
491ddc00ae Add a private call '__libcpp_is_constant_evaluated' which 'works' for old language versions and w/o any compiler support. 'Working', in this case, means that it returns false in those cases.
llvm-svn: 364873
2019-07-01 23:16:46 +00:00
Marshall Clow
fc61db5a3e Update status for bit operations
llvm-svn: 364863
2019-07-01 23:00:34 +00:00
Marshall Clow
a5c3485a58 Bit Operations: P0556, P0553 and P1355. Reviewed as: https://reviews.llvm.org/D51262
llvm-svn: 364862
2019-07-01 23:00:32 +00:00
Eric Fiselier
d1523f7a8c Ensure bitset's string constructor doesn't poison the overload set.
llvm-svn: 364842
2019-07-01 19:59:34 +00:00
Marshall Clow
24edf8ef4b Implement P0646R1: Erase-Like Algorithms Should Return size_type. Reviewed as https://reviews.llvm.org/D58332, and then updated because I rewrote a couple of those routines to eliminate some UB. Thanks to Zoe for tghe patch.
llvm-svn: 364840
2019-07-01 19:22:00 +00:00
Marshall Clow
3c125fe821 Implement LWG2221: 'Formatted output for nullptr_t' Reviewed as: https://reviews.llvm.org/D63053
llvm-svn: 364802
2019-07-01 16:20:25 +00:00
Bruce Mitchener
4603460a39 __threading_support: Remove (void) in favor of ().
Summary:
This fixes a clang-tidy warning when building something that uses
this file.

Reviewers: mclow.lists, EricWF

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D43226

llvm-svn: 364799
2019-07-01 16:18:38 +00:00
Bruce Mitchener
ba41d3b1fd Fix -Wdouble-promotion warnings.
Reviewers: mclow.lists

Subscribers: christof, ldionne, cfe-commits, libcxx-commits

Differential Revision: https://reviews.llvm.org/D62782

llvm-svn: 364798
2019-07-01 16:13:31 +00:00
Billy Robert O'Neal III
0384a78054 [libcxx] [test] Add void cast to result of compare_exchange_weak to suppress [[nodiscard]].
llvm-svn: 364732
2019-07-01 08:09:02 +00:00
Marshall Clow
954014a0fa Add a missing '__uncvref_t' to the SFINAE constraints for optional's assignment operator. Fixes PR38638. Thanks to Jonathan Wakely for the report
llvm-svn: 364574
2019-06-27 18:40:55 +00:00
Marshall Clow
7d8274d54d Followup to revision 364545: Turns out that clang issues different errors for C++11 vs c++2a, so I tweaked the 'expected-error' bits that I added to match either of them.
llvm-svn: 364554
2019-06-27 15:37:31 +00:00
Marshall Clow
9318430237 Provide hashers for string_view only if they are using the default char_traits. Seen on SO: test/std/strings/string.view/string.view.hash/char_type.hash.fail.cpp
llvm-svn: 364545
2019-06-27 14:18:32 +00:00
Eric Fiselier
ae9e42f614 Revert "Change the ABI version and ABI namespace to be _LIBCPP_VERSION"
There are some suspicious bot failures that I want to ensure aren't
caused by this patch.

I'll recommit tomorrow.

llvm-svn: 364363
2019-06-26 00:05:14 +00:00
Eric Fiselier
1fa0f4b942 Change the ABI version and ABI namespace to be _LIBCPP_VERSION
when _LIBCPP_ABI_UNSTABLE is defined.

User defined _LIBCPP_ABI_NAMESPACE will still be respected,
but the default version namespace in unstable mode will be the libc++ version
(Currently '__9000').

Previously  `_LIBCPP_ABI_VERSION` and `_LIBCPP_ABI_NAMESPACE` were
`1` and `__1` respectively, whuch conflicted with the stable ABI

llvm-svn: 364354
2019-06-25 22:13:39 +00:00
Richard Smith
4bd4acc969 Fix test failures due to modified wording in Clang diagnostics.
llvm-svn: 364241
2019-06-24 22:01:03 +00:00
Richard Smith
66e6e37447 Fix test failures when using a custom ABI namespace.
llvm-svn: 364239
2019-06-24 21:46:05 +00:00