1696 Commits

Author SHA1 Message Date
William Tran-Viet
1c51886920
[libc++] Implement P3168R2: Give optional range support (#149441)
Resolves #105430

- Implement all required pieces of P3168R2
- Leverage existing `wrap_iter` and `bounded_iter` classes to implement
the `optional` regular and hardened iterator type, respectively
- Update documentation to match
2025-08-18 18:04:45 +03:00
Aiden Grossman
36d31b0c00 [CI][Github] Bump Github Runner Version to 2.328.0
Only some minor changes, but we might as well upgrade now.
2025-08-13 19:37:27 +00:00
Aiden Grossman
4fd41c4afb
[libcxx] Bump runner version in container image (#153339)
This patch bumps the runner version in the libcxx container image. Here
we update the runner image without updating the rest of the container.
We would probably be fine updating everything, but this lets us test the
new workflow for updating the runner binary independently.
2025-08-13 08:29:41 -07:00
Aiden Grossman
77cb0980bc
[libcxx] Fix container build and bump clang version (#153035)
This patch fixes the libcxx container build. There was a missing env
variable that would cause the containers to not be built. Now that clang
22 is out, we also want to bump the LLVM head version to 22.
2025-08-11 12:15:50 -07:00
A. Jiang
4a509f853f
[libc++] Implement comparison operators for tuple added in C++23 (#148799)
And constrain the new `operator==` since C++26.

This patch implements parts of P2165R4, P2944R3, and a possibly improved
resolution of LWG3882. Currently, libstdc++ and MSVC STL constrain the
new overloads in the same way.

Also set feature-test macro `__cpp_lib_constrained_equality` and add
related release note, as P2944R3 will completed with this patch.

Fixes #136765
Fixes #136770
Fixes #105424
2025-08-01 11:53:33 -04:00
Konstantin Varlamov
3eee9fc2c4
[libc++][hardening] Introduce assertion semantics. (#149459)
Assertion semantics closely mimic C++26 Contracts evaluation semantics.
This brings our implementation closer in line with C++26 Library Hardening
(one particular benefit is that using the `observe` semantic makes adopting
hardening easier for projects).
2025-07-29 00:19:15 -07:00
Aiden Grossman
b33f9f64c7
[libcxx] Enable installing new runner binary on existing container
This patch does some refactoring to enable installing a new GHA runner binary
into an existing libcxx image. We achieve this by pushing the base image to the
registry and enabling control over the base image used for building the actions
image. This will always build and push both images even if an existing image is
being used for the actions image, but this should not impact anything as the
SHAs are pinned everywhere and space/build time is not a large concern.

Reviewers: ldionne, EricWF, #reviewers-libcxx

Reviewed By: ldionne

Pull Request: https://github.com/llvm/llvm-project/pull/148073
2025-07-28 12:35:40 -07:00
Aiden Grossman
6107e3aa22
[libcxx] Install runner last when building CI containers
This patch changes when we install the GHA runner in the CI containers. Instead
of having it in the base image, we install it last. This will enable a follow up
patch that will do some setup enabling building the full container image with an
existing base image, thus enabling updating the GHA runner without modifying the
important bits.

Reviewers: EricWF, ldionne

Reviewed By: ldionne

Pull Request: https://github.com/llvm/llvm-project/pull/148072
2025-07-28 12:32:24 -07:00
Konstantin Varlamov
d750c6de8a
[libc++][hardening] Add an experimental function to log hardening errors (#149452)
Unlike `verbose_abort`, this function merely logs the error but does not
terminate execution. It is intended to make it possible to implement the
`observe` semantic for Hardening.
2025-07-24 13:39:48 -04:00
Louis Dionne
1a0dd5a67e
[libc++] Avoid duplicate LWGXYZ prefixes in status tables (#148874)
When synchronizing the status tables with Github issues, we use the
title of the Github issue as the name of the paper in the status table.
However, the Github issue titles are prefixed with PXYZ or LWGXYZ (which
is useful to quickly find papers), and that is redundant in the context
of status tables. This patch ensures that we don't add that redundant
PXYZ or LWGXYZ prefix.

As a drive-by, also specify the encoding for opening files explicitly,
which fixes issues on Windows.
2025-07-21 10:02:54 -04:00
Ryan Prichard
038e80cfd3
[libc++][Android] Update compiler and sysroot (#148998)
* Upgrade from r536225 to r563880.

* Upgrade from ab/12644632 to f8b85cc5262c6e5cbc9a92c1bab2b18b32a4c63f,
the current HEAD commit of
https://android.googlesource.com/platform/prebuilts/ndk/+/refs/heads/mirror-goog-main-ndk

The previous source of sysroots (ci.android.com), deleted its artifacts
after a short period of time, and is currently out-of-date because of
the aosp-main turndown.

Updating the Docker image also fixes two tests.
2025-07-16 15:57:13 -07:00
Steve O'Brien
a89e6f6672
Minor formatting fix in 'generate_feature_test_macro_components' (#148889)
Fixes a small annoyance where generated files have a format which does
not agree with the one checked during `code-formatter` in CI.

For example `libcxx-generate-files` updates (among possibly others) the
`*.version.compile.pass.cpp` files. Previously these files contained an
extra newline which would fail the code format check. If you update that
file manually to remove just that extra trailing newline, then
`check-generated-output` will fail due to the file's contents differing
from what's expected.

Contains a number of changes: one actual change to the py script, and
lots of resulting whitespace changes.

My process for this was:
* Update `generate_feature_test_macro_components`: just remove an extra
newline which causes the code-format step to fail
* Run `$NINJA libcxx-generate-files` to rebuild all these
`.version.pass.cpp`'s
* Watch this PR's CI run to ensure things pass (i.e. this didn't break
things worse)
2025-07-16 17:19:54 +02:00
Nikolas Klauser
339a1f2e8f
Revert "[libc++][hardening] Introduce assertion semantics" (#148822)
Reverts llvm/llvm-project#148268

It looks like this was based on #148266, which I reverted in #148787.
2025-07-15 12:43:37 +02:00
Nikolas Klauser
fda3fbee6f
Revert "[libc++][hardening] Introduce a dylib function to log hardening errors." (#148787)
Reverts llvm/llvm-project#148266

I'm reverting this temporarily, since the release branch is today and
this is ABI sensitive. Let's wait until after the branch so that we have
plenty time to discuss the patch.
2025-07-15 12:12:41 +02:00
Konstantin Varlamov
7345508c6f
[libc++][hardening] Introduce assertion semantics (#148268)
Assertion semantics closely mimic C++26 Contracts evaluation semantics.
This brings our implementation closer in line with C++26 Library
Hardening (one particular benefit is that using the `observe` semantic
makes adopting hardening easier for projects).
2025-07-15 02:14:30 -07:00
Konstantin Varlamov
49d2b5f1cd
[libc++][hardening] Introduce a dylib function to log hardening errors. (#148266)
Unlike `verbose_abort`, this function merely logs the error but does not
terminate execution. It is intended to make it possible to implement the
`observe` semantic for Hardening.
2025-07-14 17:04:33 -07:00
Aiden Grossman
b12fcff4ff
[libcxx] Bump Container Runner Version (#147831)
This patch bumps the runner version from v3.222.0 to v3.226.0 as
v3.222.0 is too old at this point to connect to Github. This is needed
for the new premerge system given we are directly using this container.
This did not impact the existing libc++ CI as the runner was contained
in a separate container image.
2025-07-09 14:56:11 -07:00
Hui
0d1e5ab2fd
[libc++] P2655R3 common_reference_t of reference_wrapper Should Be a Reference Type (#141408)
Fixes #105260

This patch applies the change as a DR to C++20. The rationale is that
the paper is more like a bug fix. It does not introduce new features, it
simply changes an existing behaviour (as a bug fix). MSVC STL DRed this
paper to C++20 as well.
2025-07-06 17:32:59 +01:00
Nikolas Klauser
22f8ceded4
[libc++] Add the thread safety annotations unconditionally (#117497)
For these annotations to do anything you need `-Wthread-safety`, in
which case users most likely enable them anyways. This avoids that users
have to explictly define a macro just to use the feature they already
had to opt-in to.
2025-07-05 18:07:07 +02:00
Raul Tambre
431507db52
[libcxxabi][cmake] Account for LIBCXXABI_TARGET_SUBDIR in test config (#138527)
This makes the logic and code structure match that of libc++, which handles this case (i.e. the target subdirectory being changed).
The `%{target}` substitution from libc++ is removed as libc++abi's config seems to be the only place it's used.
2025-07-03 12:07:25 +03:00
Louis Dionne
05ebb363c9
[libc++] Re-add a script to trigger the libc++ Buildkite pipeline (#146730)
This was removed in 99cdc26c94 when the LLVM monorepo Buildkite setup
was removed. Libc++ previously relied on the LLVM monorepo pipeline to
trigger its own pipeline when needed. Since there is no LLVM monorepo
pipeline anymore, we must trigger the libc++ Buildkite pipeline on every
pull request and use this script to determine which jobs to actually
start based on the files that are touched by the PR.
2025-07-02 11:43:19 -04:00
Ashwin Banwari
2599a9aeb5
[clang] [modules] Implement P3618R0: Allow attaching main to the global module (#146461)
Remove the prior warning for attaching extern "C++" to main.
2025-07-02 09:52:10 +08:00
Nikolas Klauser
4c8fab399b
[libc++] Avoid type-punning between __hash_value_type and pair (#143501)
This patch is very similar to #134819 in nature. Before this patch, we
were dereferencing pointers to objects which were never constructed. Now
we always assume that nodes store `pair<const KeyT, ValueT>` for
unordered_maps instead, as they actually do.
2025-06-26 19:43:59 +02:00
Jakub Mazurkiewicz
1bb2328fd3
[libc++] Implement views::join_with (#65536)
* Implement "P2441R2 `views::join_with`" (https://wg21.link/P2441R2),
closes #105185
* Implement LWG4074 (https://wg21.link/LWG4074), closes #105346
* Complete implementation of "P2711R1 Making multi-param constructors of
views explicit" (https://wg21.link/P2711R1), closes #105252
* Complete implementation of "P2770R0 Stashing stashing iterators for
proper flattening" (https://wg21.link/P2770R0), closes #105250
2025-06-21 10:54:50 +01:00
Hristo Hristov
945ce1aa3d
[libc++] Update the value of __cpp_lib_constrained_equality after P3379R0 (#144553)
https://wg21.link/P3379R0 updated the value of __cpp_lib_constrained_equality,
but we forgot to update it when we implemented the paper.
2025-06-20 12:36:46 -04:00
Peng Liu
13510c0736
[libc++] Make list constexpr as part of P3372R3 (#129799)
This patch makes `std::list` constexpr as part of P3372R3.

Fixes #128659.
2025-06-18 12:13:50 -04:00
kotborealis
a53003fe23
[libc++] Update GDB pretty-printer to work with GDB 17 (#142106)
This patch fixes an issue in libcxx/utils/gdb/libcxx/printers.py.

With gdb 17 (binutils 2_44) pretty-printers do not work anymore because
calls to `gdb.printing` requires `import gdb.printing` statement, which
was missing from the `printers.py`.

This was broken after commit https://github.com/bminor/binutils-gdb/commit/fc14343205d3a
and `import gdb.printing` was first referenced in https://github.com/bminor/binutils-gdb/commit/ee06c79b0f.

Co-authored-by: Dmitry Chestnykh <dm.chestnykh@gmail.com>
2025-06-12 11:51:22 -04:00
Peng Liu
38fb0117ab
[libc++] Make forward_list constexpr as part of P3372R3 (#129435)
Fixes #128658
2025-06-11 12:13:36 -04:00
Louis Dionne
baceb7923f
[libc++] Don't skip localization-related headers in header tests (#134877)
When localization is disabled, we used to skip testing a lot of headers.
However, these headers are now "no-ops" when localization is disabled,
so they can actually be included. As such, we should test their
inclusion in our usual header inclusion tests.
2025-06-10 14:12:46 -04:00
Peng Liu
3e5fd77d32
[libc++] constexpr priority_queue (#140634)
This patch makes `priority_queue` constexpr as part of P3372R3.

Fixes #128671.
2025-06-04 19:58:25 +08:00
Nikolas Klauser
afd53edefa
[libc++] Bump the docker image hash (#140517)
This updates GCC 15 to the release version to allow us to upgrade to GCC
15.
2025-05-27 10:58:45 +02:00
Nikolas Klauser
b060022103 [libc++] Bump GCC_HEAD_VERSION to 16
We need to update or GCC version to be able to upgrade to GCC 15.
2025-05-19 10:50:18 +02:00
Mark de Wever
b12d68e0b2
[libc++] Implements the new FTM header test generator. (#134542)
This generator has almost identical output to the existing script.
Notable differences are
- conditionally include headers that are not implemented yet
- removes the synopsis
- uses 2 spaces indent in `# if`

There are a few more test macros added that triggered bugs in existing
FTM.
2025-05-18 18:21:20 +02:00
Nikolas Klauser
e1ca2c5b60
[libc++] Reject abilist if it contains an ABI tag (#139030)
We currently don't have any ABI tags in our dylib symbols, and this is
unlikely to change in the future. By diagnosing this we avoid
accidentally adding one through e.g. having `_LIBCPP_HIDE_FROM_ABI` on
an exported symbol.
2025-05-16 15:41:53 +02:00
cor3ntin
381a649fb9
[Clang] Add warnings when mixing different charN_t types (#138708)
charN_t represent code units of different UTF encodings. Therefore the
values of 2 different charN_t objects do not represent the same
characters.

In order to avoid comparing apples and oranges, we add new warnings to
warn on:
  - Implicit conversions
  - Comparisons
  - Other cases involving arithmetic conversions

We only produce the warning if we cannot establish the comparison would
be safe through constant evaluation.

The new `-Wimplicit-unicode-conversion` warning is enabled by default.

Note that this PR intentionally doesn;t touches char/wchar_t, but it
would be worth considering also warning on extending the new warnings to
these types (in a follow up)

Additionally most arithmetic operations on charN_t don't really make
sense (ie what does it mean to addition code units), so we could add
warnings for that.

Fixes #138526
2025-05-15 18:16:05 +02:00
Nikolas Klauser
fbdf38ed43
[libc++] run clang-tidy on src/ in the CI (#121198)
This adds a new test to run clang-tidy on the `src/` directory and
temporarily disables and clang-tidy checks that currently fail. They
will be enabled in follow-up patches.
2025-05-15 10:34:09 +02:00
Nikolas Klauser
53f11dd4d6
[libc++] Avoid type-punning between __value_type and pair (#134819)
Before this patch, we were dereferencing pointers to objects which were
never constructed. Now we always assume that nodes store `pair<const
KeyT, ValueT>` for maps instead, as they actually do. This patch also
allows for significant follow-up simplifications, since
`__node_value_type` and `__container_value_type` are the same type now.
2025-05-15 09:01:56 +02:00
Aaron Ballman
7548cec16f
[www][docs] Remove last mentions of IRC (#139076)
It's the end of an era. The IRC channel was previously where the
community gathered to discuss technical topics but is now a ghost town
where the primary activity is moderators (me) kickbanning the same
individual dozens of times a day for CoC violations and the secondary
activity is telling the occasional person to come to Discord for help.
The number of people engaging on IRC for the community's intended
purposes seems to be roughly one person a month.

So this removes all remaining mentions of IRC from our documentation so
that it no longer appears to be an "official" channel for communicating
with the community. It also removes IRC handles from the various
maintainers lists, since those would stand out as confusing
anachronisms.

The IRC channel topic already recommends people come to the Discord
server. There is no way to "shut down" an IRC channel such that it no
longer exists, so the channel will continue to exist on OFTC, but will
be unmoderated.

(This was previously discussed in https://discourse.llvm.org/c/llvm/5
but some mentions persisted.)
2025-05-08 09:40:33 -04:00
Louis Dionne
3ea0754d9a
[libc++] Enable sized deallocation when building operator new shims on Apple (#134152)
This ensures that we are able to build the shims on compilers that
enable sized deallocation by default and those that don't, regardless.
2025-05-06 15:21:23 -04:00
Mark de Wever
46c730656a
[libc++][CI] Use latest Docker image. (#132539)
This image was updated in #132271.
Due to issues with upstream packages, it uses #134497 instead.
2025-05-04 18:57:52 +02:00
Mark de Wever
2b57ebb50b
[libc++][CI] Improves updating Docker image. (#134497)
This makes it easier to build a new Docker image in the CI. Since the
new image is not used automatically it's safe to commit these changes
directly to main. Then use a PR to test the new image.
2025-04-28 18:51:32 +02:00
Louis Dionne
860e88411d
[libc++] Make __config_site modular (#134699)
This patch makes the __config_site header modular, which solves various
problems with non-modular headers. This requires going back to
generating the modulemap file, since we only know how to make
__config_site modular when we're not using the per-target runtime dir.

The patch also adds a test that we support
-Wnon-modular-include-in-module, which warns about non-modular includes
from modules.

---------

Co-authored-by: Konstantin Varlamov <varconst@apple.com>
2025-04-18 06:06:25 -07:00
Damien L-G
557e931d95
[libc++] Implement P2897R7 aligned_accessor: An mdspan accessor expressing pointer over-alignment (#122603)
Closes #118372
2025-04-14 17:33:57 -04:00
Mark de Wever
6de15379f4
[NFC][libc++][test] Minor updates to generated header version test. (#134543)
Adjusting the existing script to match the new output makes it easy to
review the new script works correctly.

This adjusts the tests to match the changes in the new tests, Notably
- removes the synopsis uses 2 spaces indent in `# if`

It does not implement the conditional include part. This would be quite
some effort and these diffs are easy to review manually.

Note there are no tests for the changes; the existing script will be
phased out when the next generators are complete.
2025-04-12 14:20:01 +02:00
Mark de Wever
df579ce4b6
[libc++] Adds is_implemented function for new ftm generator. (#134538)
At the moment the ftm macro for __cpp_lib_to_chars will have the
following values:

standard_ftms: {
    "c++17": "201611L",
    "c++20": "201611L",
    "c++23": "201611L",
    "c++26": "201611L",
}

implemented_ftms: {
    "c++17": None,
}

This is an issue with the test whether the FTM is implemented it does:
  self.implemented_ftms[ftm][std] == self.standard_ftms[ftm][std]
This will fail in C++20 since implemented_ftms[ftm] does not have the
key c++20. This adds a new helper function and removes the None entries
when a FTM is not implemented.

---------

Co-authored-by: Louis Dionne <ldionne.2@gmail.com>
2025-04-11 20:27:29 +02:00
Mark de Wever
88e15b7815
[libc++] A quick-fix for the Apple CI. (#135202)
The Apple runners automatically pick up newer XCode versions breaking
the CI. This disables the test to get the CI green which allows us to
investigate the issue properly later on.
2025-04-11 17:05:30 +02:00
Mark de Wever
1847b00330
[libc++][ci] Removes C++2b selection option. (#88557)
Since Clang 16 is no longer supported all compilers support C++23.
2025-04-07 18:17:11 +02:00
Hui
7013b51548
[libc++] Implement std::flat_multiset (#128363)
fixes https://github.com/llvm/llvm-project/issues/105193
2025-04-06 10:50:55 +01:00
James E T Smith
475cbf0ad6
[libc++] Implement ranges::iota (#68494)
# Overview

As a disclaimer, this is my first PR to LLVM and while I've tried to
ensure I've followed the LLVM and libc++ contributing guidelines,
there's probably a good chance I missed something. If I have, just let
me know and I'll try to correct it as soon as I can.

This PR implements `std::ranges::iota` and
`std::ranges::out_value_result` outlined in
[P2440r1](https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2021/p2440r1.html).

As outlined in the paper above, I've:
- Implemented `out_value_result` and added to `<algorithm>`
- Added `out_value_result`, `iota_result`, and two overloads of `iota`
to `std::ranges` in `<numeric>`
- Updated the version macro `__cpp_lib_ranges_iota` in `<version>`

I've also added tests for `ranges::iota` and `ranges::out_value_result`.
Lastly, I added those structs to the appropriate module files.

Partially implements #105184

EDIT: Forgot to mention in the original post, thanks to @hawkinsw for
taking a look at a preliminary version of this PR!

# TODOs
- [x] Updating the range [status
doc](https://github.com/jamesETsmith/llvm-project/blob/main/libcxx/docs/Status/RangesMajorFeatures.csv)
- [x] Ensure all comments from https://reviews.llvm.org/D121436 are
addressed here
- [X] EDIT (I'll do this in a separate PR). ~~I'm open to implementing
the rest of P2440r1 (`ranges::shift_left` and `ranges::shift_right`) if
that's ok, I just wanted to get feedback on `ranges::iota` first~~
- [x] I've been having trouble building the modules locally and want to
make sure that's working properly

Closes: #134060
2025-04-05 13:46:11 +02:00
David Spickett
1f90a88b80
[libcxx] Remove clang-18 workaround in picolib build (#133254)
clang-19 changed how Arm triples were normalised and so while we
supported 18 and 19, we could not hard code the path here.

Now that Linaro's bots are running clang-19, and libcxx is going to drop
clang-18 support (https://github.com/llvm/llvm-project/pull/130142) I
have simplified it by hard coding the path again.

I also looked into why this exists in the first place. It was added in
https://reviews.llvm.org/D154246 but not questioned at the time.

It is due to the way we build compiler-rt, which is due to the final
layout we need in the install:
1. The builtins library must be called libclang_rt.builtins.a for clang
to find it. There must not be an architecture name in the filename.
2. That builtins library must be directly in lib/, next to picolib's
installed files.

To achieve number 1 we must set LLVM_ENABLE_PER_TARGET_RUNTIME_DIR=ON.
However, that causes the file to be installed in a per-target dir which
breaks number 2. So to fix that, we move the builtins library up one
level into lib/.

The alternative is to turn off per-target dirs, which results in a
builtin file with an arch in the name, then rename and move that file
(since it gets installed into lib/generic/).

So in the end, it's the same amount of hacks. I think it's best to keep
the one that uses LLVM_ENABLE_PER_TARGET_RUNTIME_DIR=ON, as this is the
recommended way to built these days.
2025-03-28 14:30:38 +00:00