This issue was detected by ASan in one of our tests. This test manually
invokes basic_filebuf::cloe(). fclose(__h.release() returned a non-zero
exit status, so __file_ wasn't set to 0. Later when basic_filebuf
destructor ran, we would enter the if (__file_) block again leading to
heap-use-after-free error.
The POSIX specification for fclose says that independently of the return
value, fclose closes the underlying file descriptor and any further
access (including another call to fclose()) to the stream results in
undefined behavior. This is exactly what happened in our test case.
To avoid this issue, we have to always set __file_ to 0 independently of
the fclose return value.
Differential Revision: https://reviews.llvm.org/D64979
llvm-svn: 366730
Some minor versions of AppleClang 9 appear not to fail the test. It's
such a mess that the only sane thing to do is to mark the test as
UNSUPPORTED.
llvm-svn: 366606
This reverts r366593, which caused unforeseen breakage on the build bots.
I'm reverting until the problems have been figured out and fixed.
llvm-svn: 366603
Summary:
This commit allows specifying LIBCXX_ENABLE_PARALLEL_ALGORITHMS when
configuring libc++ in CMake. When that option is enabled, libc++ will
assume that the PSTL can be found somewhere on the CMake module path,
and it will provide the C++17 parallel algorithms based on the PSTL
(that is assumed to be available).
The commit also adds support for running the PSTL tests as part of
the libc++ test suite.
Reviewers: rodgert, EricWF
Subscribers: mgorny, christof, jkorous, dexonsmith, libcxx-commits, mclow.lists, EricWF
Tags: #libc
Differential Revision: https://reviews.llvm.org/D60480
llvm-svn: 366593
Without the link flags, the test always fails on Linux. For some reason,
however, it works on Darwin -- which is why it wasn't caught at first.
llvm-svn: 366579
Summary:
This is effectively a revert of r344616, which was a partial fix for
PR38964 (compilation of <string> with GCC in C++03 mode). However, that
configuration is explicitly not supported anymore and that partial fix
breaks compilation with Clang when per-TU insulation is provided.
PR42676
rdar://52899715
Reviewers: mclow.lists, EricWF
Subscribers: christof, jkorous, dexonsmith, libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D64941
llvm-svn: 366567
In r361572, we introduced library support for C++20 destroying delete
and decided to only define the library feature-test macro when the
compiler supports the underlying language feature. This patch reworks
the tests to mirror that.
llvm-svn: 366263
The tests for unordered_set and unordered_multiset were missing UNSUPPORTED
markup for Apple Clang 9.1, which is still being used on some CI bots.
llvm-svn: 366259
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
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
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
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
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
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
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
The new meta-programming primitives are lower cost than the old versions. This patch removes those old versions and switches libc++ to use the new ones.
llvm-svn: 364160
Clang provides __is_same that doesn't produce any instantiations
and just returns a bool. It's a lot faster than using std::is_same
I'll follow up with a patch to actually start using it.
llvm-svn: 364148
Using class templates instead of alias templates causes a lot of
instantiations. As part of the move away from C++03, we want to
improve the efficiency of our meta-programming.
This patch lays the groundwork by introducing new _If, _EnableIf,
_And, _Or, and _IsValidExpansion (detect member). Future patches
will replace the existing implementations after verifying there
compile time differences.
llvm-svn: 364114
These functions are key to allowing the use of rvalues and variadics
in C++03 mode. Everything works the same as in C++11, except for one
tangentially related case:
struct T {
T(T &&) = default;
};
In C++11, T has a deleted copy constructor. But in C++03 Clang gives
it both a move and a copy constructor. This seems reasonable enough
given the extensions it's using.
The other changes in this patch were the minimal set required
to keep the tests passing after the move/forward change. Most notably
the removal of the `__rv<unique_ptr>` hack that was present
in an attempt to make unique_ptr move only without language support.
llvm-svn: 364063
This is a re-application of r362986 (which was reverted in r363688) with fixes
for the issue that caused it to be reverted.
Thanks to Arthur O'Dwyer for the patch.
Differential Revision: https://reviews.llvm.org/D58587
llvm-svn: 363968
Summary:
AIX system headers need stdint.h and inttypes.h to be re-enterable when macro _STD_TYPES_T is defined so that limit macro definitions such as UINT32_MAX can be found. This patch attempts to allow that on AIX.
Reviewers: hubert.reinterpretcast, jasonliu, mclow.lists, EricWF
Reviewed by: hubert.reinterpretcast, mclow.lists
Subscribers: jfb, jsji, christof, cfe-commits, libcxx-commits, llvm-commits
Tags: #LLVM, #clang, #libc++
Differential Revision: https://reviews.llvm.org/D59253
llvm-svn: 363939
This was found to be broken on Clang trunk. This is a revert of the
following commits (the subsequent commits added XFAILs to the tests
that were missing from the original submission):
r362986: Implement deduction guides for map/multimap.
r363014: Add some XFAILs
r363097: Add more XFAILs
r363197: Add even more XFAILs
llvm-svn: 363688
As we gear up to drop support for GCC in C++03, we should make clear
what our C++03 mode is, the C++11 extensions it provides,
and the C++11 extensions it depends on.
The section of this document discussing user-facing extensions has
been left blank while the community discusses new directions. For now
it's just a warning to users.
Additionally, the document contains examples of how these extensions
should be used and why. For example, using alias templates over class
templates.
llvm-svn: 363110
Summary:
This is not mandated by the Standard, but it's nonetheless a nice
property to have, especially since it's so easy to implement. It
also shrinks our bug list!
PR41714
Reviewers: mclow.lists, EricWF
Subscribers: christof, jkorous, dexonsmith, libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D62618
llvm-svn: 363075
Summary:
Following the discussion on the libcxx-dev mailing list
(http://lists.llvm.org/pipermail/libcxx-dev/2019-May/000358.html),
this implements the new policy for handling experimental features and
their deprecation. We basically add a deprecation warning for
std::experimental::filesystem, and we remove a bunch of <experimental/*>
headers that were now empty.
Reviewers: mclow.lists, EricWF
Subscribers: mgorny, christof, jkorous, dexonsmith, arphaman, libcxx-commits, jfb
Tags: #libc
Differential Revision: https://reviews.llvm.org/D62428
llvm-svn: 363072
Some tests require `TEST_WORKAROUND_CONSTEXPR_IMPLIES_NOEXCEPT`, but they
did not include the header that defines that macro.
Thanks to Michael Park for the patch.
Differential Revision: https://reviews.llvm.org/D62920
llvm-svn: 362660
This commit adds tests that repeated characters in regular expressions
are within numeric limits, and that a <= b in a regex like `x{a,b}`.
Thanks to Andrey Maksimov for the patch.
Differential Revision: https://reviews.llvm.org/D62816
llvm-svn: 362525
Summary:
This provides the `std::destroying_delete_t` declaration in C++2a and after. (Even when the compiler doesn't support the language feature).
However, the feature test macro `__cpp_lib_destroying_delete` is only defined when we have both language support and C++2a.
Reviewers: ldionne, ckennelly, serge-sans-paille, EricWF
Reviewed By: EricWF
Subscribers: dexonsmith, riccibruno, christof, jwakely, jdoerfert, mclow.lists, ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D55840
llvm-svn: 361572
Some tests assume that iteration through an unordered multiset elements
will return them in the same order as at the container creation. This
assumption is not true since the container is unordered, so that no
specific order of elements is ever guaranteed for such container. This
patch introduces checks verifying that any iteration will return
elements exactly from a set of valid values and without repetition,
but in no particular order.
Thanks to Andrey Maksimov for the patch.
Differential Revision: https://reviews.llvm.org/D56500
llvm-svn: 361494
Some tests assume that iteration through an unordered multimap elements
will return them in the same order as at the container creation. This
assumption is not true since the container is unordered, so that no
specific order of elements is ever guaranteed for such container. This
patch is a continuation of D54838 and introduces checks verifying that
any iteration will return elements exactly from a set of valid values
and without repetition, but in no particular order.
Thanks to Andrey Maksimov for the patch.
Differential Revision: https://reviews.llvm.org/D56498
llvm-svn: 361414
The `using namespace std;` opens us up to ambiguity
when any of the std:: names are also present in the global namespace.
Instead we should properly qualify names we use from std::.
llvm-svn: 361074
Use std::nextafter() instead of std::nexttoward() in midpoint tests.
In the context of this test, this should not cause any difference.
Since nexttowardl() is not implemented on NetBSD 8, the latter function
combined with 'long double' type caused test failure. nextafterl() does
not have this problem.
Differential Revision: https://reviews.llvm.org/D61748
llvm-svn: 360673
Drive-by:
* Fix potential race between check and update of `throw_one` in `operator new`
* Fix latent bug in `operator delete`, which shouldn't decrement `outstanding_new` when passed a null pointer
* Specifically catch the expected `bad_alloc` in `main` instead of `...`
Differential Revision: https://reviews.llvm.org/D50860
llvm-svn: 359827
Linux is failing even though the test runner does report this locale
is available, but the test still isn't expected to work on platforms
without the locale (like Android).
llvm-svn: 359726
This adds explicit support for the WASI platform to libcxx.
WASI libc uses some components from musl, however it's not fully compatible
with musl, so we're planning to stop using _LIBCPP_HAS_MUSL_LIBC and
customize for WASI libc specifically.
Differential Revision: https://reviews.llvm.org/D61336
Reviewers: sbc100, ldionne
llvm-svn: 359703
Summary:
The current implementation of aligned storage was written before we had `alignas`, so it used a list of builtin types to force the alignment. But this doesn't work overaligned requests.
This patch adds a fallback case supporting over-alignment. It only affects case that were previously ill-formed.
Reviewers: rsmith, ldionne, dlj, mclow.lists
Reviewed By: mclow.lists
Subscribers: mclow.lists, dexonsmith, libcxx-commits
Differential Revision: https://reviews.llvm.org/D61301
llvm-svn: 359596
The standard requires the following for the std::regex_constants::error_type
values: "The type error_type is an implementation-defined enumerated type."
The values of this enumerated type are not required to be non-zero.
This patch makes such checks in tests libc++-specific to let the tests
pass for other conforming implementations.
Thanks to Andrey Maksimov for the patch.
Differential Revision: https://reviews.llvm.org/D61195
llvm-svn: 359320
When the arguments to tuple cat were const, the const was incorrectly
propagated into the type of the resulting tuple. For example:
const std::tuple<int> t(42);
auto r = std::tuple_cat(t, t);
// Incorrect! should be std::tuple<int, int>.
static_assert(is_same_v<decltype(r), std::tuple<const int, const int>>);
llvm-svn: 359255
All constant expressions are non-potentially-throwing in C++14, but that is *not* the case in C++17. Change these tests of the `variant`-flavored overloads of `std::get` to expect the correct behavior when the compiler is not GCC or is GCC 9+.
Credit to Jonathan Wakely for providing an improved version of my initial change that validates the incorrect behavior on GCC < 9 as well as validating the correct behavior on other compilers.
Differential Revision: https://reviews.llvm.org/D61033
llvm-svn: 359220
Teach libcxx to stop using various deprecated __has_* type traits, in favor of
the ("modern", C++11 era) __is_* type traits.
This is mostly just a simplification, but fixes at least one bug: _Atomic T
should be considered trivially-destructible, but is not considered to be POD by
Clang, and __has_trivial_destructor is specified in the GCC documentation as
returning false for non-POD non-class types.
Differential Revision: https://reviews.llvm.org/D48292
llvm-svn: 359159
Summary:
All overloads of `::abs` and `std::abs` must be present in both `<cmath>` and `<cstdlib>`. This is problematic to implement because C defines `fabs` in `math.h` and `labs` in `stdlib.h`. This introduces a circular dependency between the two headers.
This patch implements that requirement by moving `abs` into `math.h` and making `stdlib.h` include `math.h`. In order to get the underlying C declarations from the "real" `stdlib.h` inside our `math.h` we need some trickery. Specifically we need to make `stdlib.h` include next itself.
Suggestions for a cleaner implementation are welcome.
Reviewers: mclow.lists, ldionne
Reviewed By: ldionne
Subscribers: krytarowski, fedor.sergeev, dexonsmith, jdoerfert, jsji, libcxx-commits
Differential Revision: https://reviews.llvm.org/D60097
llvm-svn: 359020
Due to MSVC's decision to encode `wchar_t` as UTF-16, it rejects wide
character/string literals that expect a character value greater than
`\xffff`. UTF-16 `wchar_t` is clearly non-conforming, given that the
standard requires wchar_t to be capable of representing all characters
in the supported wide character execution sets, but rejecting e.g.
`\x40003` is a reasonably sane compromise given that encoding choice:
there's an expectation that `\xFOO` produces a single character in the
resulting literal. Consequently `L'\x40003'`/`L"\x40003"` are ill-formed
literals on MSVC. `L'\U00040003'` is a high surrogate (and produces a
warning about ignoring the "second character" in a multi-character
literal), and `L"\U00040003"` is a perfectly-valid `const wchar_t[3]`.
This change updates these tests to use universal-character-names instead
of raw values for the intended character values, which technically makes
them portable even to implementations that don't use a unicode
transformation format encoding for their wide character execution
character set. The two-character literal `L"\u1005e"` is awkward - the
`e` looks like part of the UCN's hex encoding - but necessary to compile
in '03 mode since '03 didn't allow UCNs to be used for members of the
basic execution character set even in character/string literals.
I've also eliminated the extraneous `\x00` "bonus null-terminator" in
some of the string literals which doesn't affect the tested behavior.
I'm sorry about using `*L"\U00040003"` in `conversions.string/to_bytes.pass.cpp`,
but it's correct for platforms with 32-bit wchar_t, *and* doesn't
trigger narrowing warnings as did the prior `CharT(0x40003)`.
Differential Revision: https://reviews.llvm.org/D60950
llvm-svn: 358908
This is a followup to [1] which added a new `__debug_less::operator()` overload.
[2] added `_LIBCPP_CONSTEXPR_AFTER_CXX17` to the original
`__debug_less::operator()` between the time of writing [1] and landing it. This
change adds `_LIBCPP_CONSTEXPR_AFTER_CXX17` to the new overload too.
[1] https://reviews.llvm.org/rL358423
[2] https://reviews.llvm.org/rL358252
Differential Revision: https://reviews.llvm.org/D60724
llvm-svn: 358725
In r358591, I added a test that uses the debug database from multiple
threads and that helped us uncover the problem that was fixed in r355367.
However, the test broke the tsan CI bots, and I think the problem is the
test allocator that was used in the test (which is not thread safe).
I'm committing again without using the test allocator, and in a separate
test file.
llvm-svn: 358610
This reverts r358591, which seems to have uncovered an actual bug and
causes the tsan CI to fail. We need to fix the bug and re-commit the
test.
llvm-svn: 358593
There are many STL algorithms (such as lexicographical_compare) that compare
values pointed to by iterators like so:
__comp(*it1, *it2);
When building with `_LIBCPP_DEBUG=0`, comparators are wrapped in `__debug_less`
which does some additional validation. But `__debug_less::operator()` takes
non-const references, so if the type of `*it1` is int, not int&, then the build
will fail.
This change adds a `const&` overload for `operator()` to fix the build.
Differential Revision: https://reviews.llvm.org/D60592
llvm-svn: 358423
We used to do it against the current system's libc++abi, which is not as
good as doing it with the libc++abi that matches the libc++ we're running
against.
Note that I made sure we were indeed picking up the provided libc++abi
by replacing it by something that doesn't work and watching it burn.
llvm-svn: 358294
Summary:
In r348529, I improved the library-defined diagnostic for using containers
with a non-const comparator/hasher. However, the check is now performed
too early, which leads to the diagnostic being emitted in cases where it
shouldn't. See PR41360 for details.
This patch moves the diagnostic to the destructor of the containers, which
means that the diagnostic will only be emitted when the container is instantiated
at a point where the comparator and the key/value are required to be complete.
We still retain better diagnostics than before r348529, because the diagnostics
are performed in the containers themselves instead of __tree and __hash_table.
As a drive-by fix, I improved the diagnostic to mention that we can't find
a _viable_ const call operator, as suggested by EricWF in PR41360.
Reviewers: EricWF, mclow.lists
Subscribers: christof, jkorous, dexonsmith, libcxx-commits, zoecarver
Tags: #libc
Differential Revision: https://reviews.llvm.org/D60540
llvm-svn: 358189
We fixed incorrect behavior of input streams in r357775 and tests were
added accordingly. However, older versions of macOS don't have the
change in the dylib yet, so the tests fail on those platforms.
llvm-svn: 357794
Summary:
This is a re-application of r357533 and r357531. They had been reverted
because we thought the commits broke the LLDB data formatters, but it
turns out this was because only r357531 had been included in the CI
run.
Before this patch, we would only ever throw an exception if the badbit
was set on the stream. The Standard is currently very unclear on how
exceptions should be propagated and what error flags should be set by
the input stream operations. This commit changes libc++ to behave under
a different (but valid) interpretation of the Standard. This interpretation
of the Standard matches what other implementations are doing.
This effectively implements the wording in p1264r0. It hasn't been voted
into the Standard yet, however there is wide agreement that the fix is
correct and it's just a matter of time before the fix is standardized.
PR21586
PR15949
rdar://problem/15347558
Reviewers: mclow.lists, EricWF
Subscribers: christof, dexonsmith, cfe-commits
Differential Revision: https://reviews.llvm.org/D49863
llvm-svn: 357775
This builds on the work done in r342808 and adds _LIBCPP_NODISCARD_EXT
to 37 more functions, namely:
adjacent_find, all_of, any_of, binary_search, clamp, count_if, count,
equal_range, equal, find_end, find_first_not_of, find_first_of, find_if,
find, includes, is_heap_until, is_heap, is_partitioned, is_permutation,
is_sorted_until, is_sorted, lexicographical_compare, lower_bound,
max_element, max, min_element, min, minmax_element, minmax, mismatch,
none_of, remove_if, remove, search_n, search, unique, upper_bound
The motivation here is that we noticed that find_if is nodiscard with
Visual Studio's standard library, and we deemed that useful
(https://crbug.com/948122).
https://devblogs.microsoft.com/cppblog/c17-progress-in-vs-2017-15-5-and-15-6/
says "Our criteria for emitting the warning are: discarding the return
value is a guaranteed leak [...], discarding the return value is
near-guaranteed to be incorrect (e.g. remove()/remove_if()/unique()), or
the function is essentially a pure observer (e.g. vector::empty() and
std::is_sorted())." so I went through algorithm and tried to apply these
criteria.
Some of these, like vector::empty() are already nodiscard per C++
standard and didn't need changing.
I didn't (yet?) go over std::string::find* methods which should probably
have _LIBCPP_NODISCARD_EXT too (but not as part of this change).
Differential Revision: https://reviews.llvm.org/D60145
llvm-svn: 357619
Eric Fiselier