Many headers include `<cstddef>` just for size_t, and pulling in
additional content (e.g. the traits used for std::byte) is unnecessary.
To solve this problem, this patch splits up `<cstddef>` into
subcomponents so that headers can include only the parts that they
actually require.
This has the added benefit of making the modules build a lot stricter
with respect to IWYU, and also providing a canonical location where we
define `std::size_t` and friends (which were previously defined in
multiple headers like `<cstddef>` and `<ctime>`).
After this patch, there's still many places in the codebase where we
include `<cstddef>` when `<__cstddef/size_t.h>` would be sufficient.
This patch focuses on removing `<cstddef>` includes from __type_traits
to make these headers non-circular with `<cstddef>`. Additional
refactorings can be tackled separately.
This makes `std::pair` trivially copyable if its members are and we have
a way to do so. We need either C++20 with requires clauses or support
for `__attribute__((enable_if))`. Only Clang has support for this
attribute, so it's effectively clang or C++20.
Co-authored-by: Christopher Di Bella <cjdb@google.com>
Reverts llvm/llvm-project#96165
The change broke code like
#include <utility>
#include <vector>
struct Test {
std::vector<std::pair<int, Test>> v;
};
std::pair<int, Test> p;
under `-std=c++20`, apparently by triggering certain template
evaluations too eagerly.
Since 83ead2b, std::pair would not be trivially copyable when it holds a
trivially copyable type without an assignment operator. That is because
pair gained an elligible copy-assignment-operator (the const version) in
83ead2b in C++ >= 23.
This means that the trivially copyable property of std::pair for such
types would be inconsistent between C++11/14/17/20 (trivially copyable)
and C++23/26 (not trivially copyable). This patch makes std::pair's
behavior consistent in all Standard modes EXCEPT C++03, which is a
pre-existing condition and we have no way of changing (also, it
shouldn't matter because the std::is_trivially_copyable trait was
introduced in C++11).
While this is not technically an ABI break, in practice we do know that
folks sometimes use a different representation based on whether a type
is trivially copyable. So we're treating 83ead2b as an ABI break and
this patch is fixing said breakage.
This patch also adds tests stolen from #89652 that pin down the ABI of
std::pair with respect to being trivially copyable.
Fixes#95428
This changes the `is_swappable` implementation to use variable templates
first and basing the class templates on that. This avoids instantiating
them when the `_v` versions are used, which are generally less resource
intensive.
This moves the definition of a `pair` constructor for `<tuple>` to
`<__utility/pair.h>` and uses the forward declaration of `pair` in
`<tuple>` instead of including the definition.
The exposition-only type trait `pair-like` includes `ranges::subrange`,
but in every single case excludes `ranges::subrange` from the list. This
patch introduces two new traits `__tuple_like_no_subrange` and
`__pair_like_no_subrange`, which exclude `ranges::subrange` from the
possible matches. `__pair_like` is no longer required, and thus removed.
`__tuple_like` is implemented as `__tuple_like_no_subrange` or a
`ranges::subrange` specialization.
These headers have become very small by using compiler builtins, often
containing only two declarations. This merges these headers, since
there doesn't seem to be much of a benefit keeping them separate.
Specifically, `is_{,_nothrow,_trivially}{assignable,constructible}` are
kept and the `copy`, `move` and `default` versions of these type traits
are moved in to the respective headers.
This patch runs clang-format on all of libcxx/include and libcxx/src, in
accordance with the RFC discussed at [1]. Follow-up patches will format
the benchmarks, the test suite and remaining parts of the code. I'm
splitting this one into its own patch so the diff is a bit easier to
review.
This patch was generated with:
find libcxx/include libcxx/src -type f \
| grep -v 'module.modulemap.in' \
| grep -v 'CMakeLists.txt' \
| grep -v 'README.txt' \
| grep -v 'libcxx.imp' \
| grep -v '__config_site.in' \
| xargs clang-format -i
A Git merge driver is available in libcxx/utils/clang-format-merge-driver.sh
to help resolve merge and rebase issues across these formatting changes.
[1]: https://discourse.llvm.org/t/rfc-clang-formatting-all-of-libc-once-and-for-all
The helper function `__pair_like_explicit_wknd` is only SFINAE-ed with
`tuple_size<remove_cvref_t<_PairLike>>::value == 2`, but its function
body assumes `std::get` being valid.
Fixes#65620
This brings most of the enable_ifs in libc++ to the same style.
Reviewed By: #libc, ldionne
Spies: ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D157753
This brings most of the enable_ifs in libc++ to the same style. It also has the nice side-effect of reducing the size of names of these symbols, since the depedent return type is shorter.
Reviewed By: #libc, ldionne
Spies: ldionne, libcxx-commits
Differential Revision: https://reviews.llvm.org/D157736
The headers that include_next compiler and OS headers need to be in different top level modules in order to avoid module cycles. e.g. libc++'s stdlib.h will #include_next stdlib.h from the compiler and then the C library. Either of those are likely to include stddef.h, which will come back up to the libc++ module map and create a module cycle. Putting stdlib.h and stddef.h (and the rest of the C standard library headers) in top level modules resolves this by letting the order go cxx_stdlib_h -> os_stdlib_h -> cxx_stddef_h -> os_stddef_h.
All of those headers' dependencies then need to be moved into top level modules themselves to avoid module cycles between the new top level level cstd modules. This starts to get complicated, as the libc++ C headers, by standard, have to include many of the C++ headers, which include the private detail headers, which are intertwined. e.g. some `__algorithm` headers include `__memory` headers and vice versa.
Make top level modules for all of the libc++ headers to easily guarantee that the modules aren't cyclic.
Add enough module exports to fix `check-cxx` and `run-buildbot generic-modules`.
`__stop_token/intrusive_shared_ptr.h` uses `__atomic/atomic.h` but has no include path to it. Add that include.
`math.h` absorbs `bits/atomic_wide_counter.h` on some platforms that don't have modules, work around that by including `math.h` in `__threading_support`.
<mutex> doesn't actually require threads, there are a few pieces like once_flag that work without threads. Remove the requirement from its module.
AIX is no longer able to support modular builds.
Reviewed By: ldionne, #libc
Differential Revision: https://reviews.llvm.org/D144322
POSIX allows certain macros to exist with generic names (i.e. refresh(), move(), and erase()) to exist in `curses.h` which conflict with functions found in std::filesystem, among others. This patch undefs the macros in question and adds them to LIBCPP_PUSH_MACROS and LIBCPP_POP_MACROS.
Reviewed By: #libc, philnik, ldionne
Differential Revision: https://reviews.llvm.org/D147356
This adds an extension to std::pair in C++03 mode where we provide an
assignment operator from a pair<U, V>. Previously, any code trying to
trigger such an assignment operator would have tried using the
`operator=(pair const&)` copy assignment operator, which would then
have tried creating a `pair const&` by using the unconstrained
pair<U, V> constructor.
After this patch, pair will instead go through operator= directly if
its member types are assignable. If they are not assignable, the extension
operator= is disabled with SFINAE and the pair(pair<U, V>) constructor
will be used. Since that constructor is unconstrained, that will either
work or give a hard error.
This should be pretty transparent to users, but this is technically a
behavior change in C++03. Indeed, if a type has both a valid cross-type
assignment operator *and* a valid cross-type constructor, the library
will now prefer the cross-type assignment instead of going through the
cross-type constructor and then using the copy-constructor. Since this
is the mandated behavior in C++11, however, one could argue that any user
broken by that needs to change their code.
The motivation for this change is to allow simplifying the definition
of std::map's value_type, which requires handling assignment to a pair
of references properly. This patch will allow removing complexity from
https://llvm.org/D121485 instead of adding complexity in that patch.
Differential Revision: https://reviews.llvm.org/D150119
This patch makes std::pair's constructors and assignment operators
closer to conforming in C++23. The only missing bit I am aware of
now is `reference_constructs_from_temporary_v` checks, which we
don't have the tools for yet.
This patch also refactors a long-standing non-standard extension where
we'd provide constructors for tuple-like types in all standard modes. The
criteria for being a tuple-like type are different from pair-like types
as introduced recently in the standard, leading to a lot of complexity
when trying to implement recent papers that touch the pair constructors.
After this patch, the pre-C++23 extension is provided in a self-contained
block so that we can easily deprecate and eventually remove the extension
in future releases.
Differential Revision: https://reviews.llvm.org/D143914
We already have a clang-tidy check for making sure that `_LIBCPP_HIDE_FROM_ABI` is on free functions. This patch extends this to class members. The places where we don't check for `_LIBCPP_HIDE_FROM_ABI` are classes for which we have an instantiation in the library.
Reviewed By: ldionne, Mordante, #libc
Spies: jplehr, mikhail.ramalho, sstefan1, libcxx-commits, krytarowski, miyuki, smeenai
Differential Revision: https://reviews.llvm.org/D142332
This change is almost fully mechanical. The only interesting change is in `generate_feature_test_macro_components.py` to generate `_LIBCPP_STD_VER >=` instead. To avoid churn in the git-blame this commit should be added to the `.git-blame-ignore-revs` once committed.
Reviewed By: ldionne, var-const, #libc
Spies: jloser, libcxx-commits, arichardson, arphaman, wenlei
Differential Revision: https://reviews.llvm.org/D143962
Rename the `__tuple` directory in libc++ headers to `__tuple_dir`
to avoid file collision when installing. Historically, `__tuple` has
been a file and it has been replaced by a directory
in 2d52c6bfae801b016dd3627b8c0e7c4a99405549. Replacing a regular file
with a directory (or more importantly, the other way around when
downgrading) is not universally supported. Since this is an internal
header, its actual name should not matter, so just rename it to avoid
problems.
Differential Revision: https://reviews.llvm.org/D139270
We require rvalue support anyways, so let's use it.
Reviewed By: ldionne, #libc
Spies: libcxx-commits
Differential Revision: https://reviews.llvm.org/D133013
This was discussed on Discord with the consensus that we should rename the macros.
Reviewed By: ldionne, Mordante, var-const, avogelsgesang, jloser, #libc
Spies: libcxx-commits
Differential Revision: https://reviews.llvm.org/D131498
