Several components in libc++ aren't defending against overloaded
`operator,(T, Iter)` currently. Existing deleted overloads in
`test_iterators.h` are insufficient for such cases.
This PR adds corresponding deleted overloads with reversed order and
fixes these libc++ components.
- `piecewise_linear_distribution`'s iterator pair constructor,
- `piecewise_linear_distribution::param_type`'s iterator pair
constructor,
- `piecewise_constant_distribution`'s iterator pair constructor,
- `piecewise_constant_distribution::param_type`'s iterator pair
constructor,
- `money_get::do_get`,
- `money_put::do_put`, and
- `num_put::do_put`.
P0513R0 is essentially a collective resolution paper of LWG2543,
LWG2791, LWG2809, and LWG2817.
Among these LWG issues, LWG2543 (conditionally enabled `hash`) and
LWG2817 (`hash<nullptr_t>`) affect pre-C++17 utilities. We generally
backport changes from LWG issues, so this patch backports the relevant
parts of P0513R0.
Although we provide `hash<unique_ptr>` as an extension in C++03 mode, as
C++03 mode isn't encouraged now, this patch leaves `hash<unique_ptr>`
unchanged in C++03 mode.
This ensures that run-buildbot can find the appropriate build tools
without having to go through a bunch of hoops inside run-buildbot
itself. When we eventually run the bootstrapping build on macOS, it also
allows the bootstrapped Clang to find the SDK headers it requires.
This also allows simplifying run-buildbot a bunch since we should be
able to find build tools in the current $PATH.
This fixes incompleteness and inconsistency for test files added in
adc7932461, by
- renaming
`libcxx/test/std/time/time.traits.is.clock/trait.is.clock.compile.pass.cpp`
to `libcxx/test/std/time/time.traits/is.clock.compile.pass.cpp`,
- renaming
`libcxx/test/libcxx/time/time.traits.is.clock/trait.is.clock.compile.verify.cpp`
to `libcxx/test/libcxx/time/time.traits/is.clock.verify.cpp` , and
- adding comments clarifying what are being tested.
This patch does two things.
(1) It replaces SFINAE with `if constexpr`, avoiding some overload
resolution and unnecessary boilerplate.
(2) It removes an overload from `__for_each_n` to forward to
`__for_each`, since `__for_each` doesn't provide any further
optimizations.
This doesn't improve performance (except with optimizations disabled),
since the compiler is able to fold our current implementation. However,
it does significantly reduce the amount of code the compiler has to sift
through, reducing compile times a bit.
This commit adds move constructor, move assignment and `swap`
to `exception_ptr`. Adding those operators allows us to avoid
unnecessary calls to `__cxa_{inc,dec}rement_refcount`.
Performance results (from libc++'s CI):
```
Benchmark Baseline Candidate Difference % Difference
------------------------------------ ---------- ----------- ------------ --------------
bm_exception_ptr_copy_assign_nonnull 9.77 9.94 0.18 1.79%
bm_exception_ptr_copy_assign_null 10.29 10.65 0.35 3.42%
bm_exception_ptr_copy_ctor_nonnull 7.02 7.01 -0.01 -0.13%
bm_exception_ptr_copy_ctor_null 10.54 10.60 0.06 0.56%
bm_exception_ptr_move_assign_nonnull 16.92 13.76 -3.16 -18.70%
bm_exception_ptr_move_assign_null 10.61 10.76 0.14 1.36%
bm_exception_ptr_move_ctor_nonnull 13.31 10.25 -3.06 -23.02%
bm_exception_ptr_move_ctor_null 10.28 7.30 -2.98 -28.95%
bm_exception_ptr_swap_nonnull 19.22 0.63 -18.59 -96.74%
bm_exception_ptr_swap_null 20.02 7.79 -12.23 -61.07%
```
As expected, the `bm_exception_ptr_copy_*` benchmarks are not influenced by
this change. `bm_exception_ptr_move_*` benefits between 18% and 30%. The
`bm_exception_ptr_swap_*` tests show the biggest improvements since multiple
calls to the copy constructor are replaced by a simple pointer swap.
While `bm_exception_ptr_move_assign_null` did not show a regression in the CI
measurements, local measurements showed a regression from 3.98 to 4.71, i.e. by
18%. This is due to the additional `__tmp` inside `operator=`. The destructor
of `__other` is a no-op after the move because `__other.__ptr` will be a
nullptr. However, the compiler does not realize this, since the destructor is
not inlined and is lacking a fast-path. As such, the swap-based implementation
leads to an additional destructor call. `bm_exception_ptr_move_assign_nonnull`
still benefits because the swap-based move constructor avoids unnecessary
__cxa_{in,de}crement_refcount calls. As soon as we inline the destructor, this
regression should disappear again.
Works towards #44892
8dae17be29 refactors basic_string for more
code reuse. This makes sense in most cases, but has performance overhead
in the case of ~basic_string. The refactoring of ~basic_string to call
__reset_internal_buffer() added a redundant (inside the destructor)
reset of the object, which the optimizer is unable to optimize away in
many cases. This patch prevents a ~1% regression we observed on an
internal workload when applying the original refactoring. This does
slightly pessimize the code readability, but I think this change is
worth it given the performance impact.
I'm hoping to add a benchmark(s) to the upstream libc++ benchmark suite
around string construction/destruction to ensure that this case does not
regress as it seems common in real world applications. I will put up a
separate PR for that when I figure out a reasonable way to write it.
Previously, I was getting the following error when attempting to compile
libc++ on windows with modules enabled.
```
While building module 'std':
In file included from <module-includes>:1:
In file included from gen/third_party/libc++/src/include/algorithm:1865:
In file included from gen/third_party/libc++/src/include/__algorithm/inplace_merge.h:28:
In file included from gen/third_party/libc++/src/include/__memory/unique_temporary_buffer.h:17:
In file included from gen/third_party/libc++/src/include/__memory/allocator.h:19:
gen/third_party/libc++/src/include/__new/allocate.h(40,73): error:
declaration of 'align_val_t' must be imported from module
'sys_stage1.sysroot_vcruntime_new_h' before it is required
40 | return static_cast<_Tp*>(__builtin_operator_new(__size, static_cast<align_val...
| ^
../../third_party/depot_tools/win_toolchain/vs_files/e4305f407e/VC/Tools/MSVC/14.44.35207/include/vcruntime_new.h(27,33): note:
declaration here is not visible
27 | _VCRT_EXPORT_STD enum class align_val_t : size_t {};
| ^
```
An LLVM 19-aligned libc++ was released with macOS 15.4 (and corresponding OSes),
and LLVM-20 aligned with macOS 26.0. This patch adds availability markup to
reflect that.
<https://wg21.link/P2641R4>
Implements the C++26 function in `<type_traits>` [meta.const.eval] (and
the corresponding feature test macro `__cpp_lib_is_within_lifetime`)
```c++
template<class T>
consteval bool is_within_lifetime(const T*) noexcept;
```
This is done with the `__builtin_is_within_lifetime` builtin added to
Clang 20 by #91895 / 2a07509c8d. This is
not (currently) available with GCC.
This implementation has provisions for LWG4138
<https://cplusplus.github.io/LWG/issue4138> where it is ill-formed to
instantiate `is_within_lifetime<T>` with a function type `T`.
Closes#105381
Co-authored-by: Mital Ashok <mital@mitalashok.co.uk>
This patch contains several changes to `deallocate_size.pass.cpp`:
1. `static_cast`-ing some parameters to `size_t` to avoid narrowing.
2. Changing the type of loop variable `i` to `unsigned int` avoid
signedness mismatch with the constructor parameter.
3. Separately counting allocations and deallocations in variables
`allocated_` and `deallocated_`, and changing their type to `uint64_t`.
4. Avoiding `assert`-ing count of allocations when a `basic_string` is
allocated, just `assert`-ing after destruction instead.
This patch reverts e15025dd and 88e15b781 which were temporary measures
until we had figured out the underlying issues. It turns out that recent
OSes updated localization data, removing the need for several Apple-specific
workarounds in the tests.
Fixes#135385
To my knowledge, NetBSD is mostly like other BSDs, but doesn't have
`xlocale.h`. I think c664a7f may have inadvertently broken this.
With this change, I was able to run
[zig-bootstrap](https://github.com/ziglang/zig-bootstrap) to completion
for `x86_64-netbsd10.1-none`.
Enabling modules makes the CI quite a bit faster with basically no
downsides. The non-modules build is still tested through the
`generic-cxxab` configurations, but most of the other CI runs on
platforms with modules support now use modules.
Define `_LIBCPP_HAS_C8RTOMB_MBRTOC8` to `1` if compiling with clang.
Some tests involving functionality from `uchar.h`/`cuchar` fail when the
platform or the supporting C library does not provide support for the
corresponding features. These have been xfailed.
This patch will enable the adoption of newer picolibc versions.
`include_next` doesn't work very well with the C++03 headers and
modules. Since these specific headers are very self-contained there
isn't much of a reason to split them into C++03/non-C++03 headers, so
let's just remove them. The few C wrapper headers that aren't as
self-contained will be refactored in a separate patch.
We started to list GitHub issues in the status CSV tables/pages due to
e0d0b90d52. However, the changes made
existing column header insufficient. We probably need to add new column
headers for GitHub issues.
This patch does two things:
1) The branches for `sizeof(_Dp) == sizeof(int32_t)` and `sizeof(_Dp) ==
sizeof(int64_t)` are merged, since these are basically identical.
2) The branch where `numeric_limits<_Dp>::radix != 2` is removed, since
no platform we support (nor any I'm aware of) have a floating point type
with `radix != 2`. This means that the branch is basically dead code.
There are a few cases where test only fail with the C++03 frozen headers
because some implementation details are asserted. This patch introduces
new assertion macros to differentiate between frozen and non-frozen
headers.
