Contributing to libc++

This file contains notes about various tasks and processes specific to contributing to libc++. If this is your first time contributing, please also read this document on general rules for contributing to LLVM.

If you plan on contributing to libc++, it can be useful to join the #libcxx channel on LLVM’s Discord server.

Looking for pre-existing pull requests

Before you start working on any feature, please take a look at the open libc++ pull requests to avoid duplicating someone else’s work. You can do that on GitHub by filtering pull requests tagged with libc++. If you see that your feature is already being worked on, please consider chiming in and helping review the code instead of duplicating work!

RFCs for significant user-affecting changes

Before you start working on a change that can have significant impact on users of the library, please consider creating a RFC on libc++’s Discourse forum. This will ensure that you work in a direction that the project endorses and will ease reviewing your contribution as directional questions can be raised early. Including a WIP patch is not mandatory, but it can be useful to ground the discussion in something concrete.

Coding standards

In general, libc++ follows the LLVM Coding Standards. There are some deviations from these standards.

Libc++ uses __ugly_names. These names are reserved for implementations, so users may not use them in their own applications. When using a name like T, a user may have defined a macro that changes the meaning of T. By using __ugly_names we avoid that problem. Other standard libraries and compilers use these names too. To avoid common clashes with other uglified names used in other implementations (e.g. system headers), the test in libcxx/test/libcxx/ contains the list of reserved names that can’t be used.

Unqualified function calls are susceptible to argument-dependent lookup (ADL). This means calling move(UserType) might not call std::move. Therefore, function calls must use qualified names to avoid ADL. Some functions in the standard library require ADL usage. Names of classes, variables, concepts, and type aliases are not subject to ADL. They don’t need to be qualified.

Function overloading also applies to operators. Using &user_object may call a user-defined operator&. Use std::addressof instead. Similarly, to avoid invoking a user-defined operator,, make sure to cast the result to void when using the ,. For example:

for (; __first1 != __last1; ++__first1, (void)++__first2) {

In general, try to follow the style of existing code. There are a few exceptions:

  • Prefer using foo = int over typedef int foo. The compilers supported by libc++ accept alias declarations in all standard modes.

Other tips are:

  • Keep the number of formatting changes in patches minimal.

  • Provide separate patches for style fixes and for bug fixes or features. Keep in mind that large formatting patches may cause merge conflicts with other patches under review. In general, we prefer to avoid large reformatting patches.

  • Keep patches self-contained. Large and/or complicated patches are harder to review and take a significant amount of time. It’s fine to have multiple patches to implement one feature if the feature can be split into self-contained sub-tasks.


Libc++ specific

  • libcxx/include/__config – this file contains the commonly used macros in libc++. Libc++ supports all C++ language versions. Newer versions of the Standard add new features. For example, making functions constexpr in C++20 is done by using _LIBCPP_CONSTEXPR_SINCE_CXX20. This means the function is constexpr in C++20 and later. The Standard does not allow making this available in C++17 or earlier, so we use a macro to implement this requirement.

  • libcxx/test/support/test_macros.h – similar to the above, but for the test suite.

ISO C++ Standard

Libc++ implements the library part of the ISO C++ standard. The official publication must be bought from ISO or your national body. This is not needed to work on libc++, there are other free resources available.

  • The LaTeX sources used to create the official C++ standard. This can be used to create your own unofficial build of the standard.

  • An HTML rendered version of the draft is available. This is the most commonly used place to look for the wording of the standard.

  • An alternative is available. This link has both recent and historic versions of the standard.

  • When implementing features, there are general requirements. Most papers use this jargon to describe how library functions work.

  • The WG21 redirect service is a tool to quickly locate papers, issues, and wording in the standard.

  • The paper trail of papers is publicly available, including the polls taken. It contains links to the minutes of paper’s discussion. Per ISO rules, these minutes are only accessible by members of the C++ committee.

  • Feature-Test Macros and Policies contains information about feature-test macros in C++. It contains a list with all feature-test macros, their versions, and the paper that introduced them.

  • cppreference is a good resource for the usage of C++ library and language features. It’s easier to read than the C++ Standard, but it lacks details needed to properly implement library features.

Pre-commit check list

Before committing or creating a review, please go through this check-list to make sure you don’t forget anything:

  • Do you have tests for every public class and/or function you’re adding or modifying?

  • Did you update the synopsis of the relevant headers?

  • Did you update the relevant files to track implementation status (in docs/Status/)?

  • Did you mark all functions and type declarations with the proper visibility macro?

  • Did you add all new named declarations to the std module?

  • If you added a header:

    • Did you add it to include/module.modulemap?

    • Did you add it to include/CMakeLists.txt?

    • If it’s a public header, did you update utils/libcxx/

  • Did you add the relevant feature test macro(s) for your feature? Did you update the script with it?

  • Did you run the libcxx-generate-files target and verify its output?

  • If needed, did you add _LIBCPP_PUSH_MACROS and _LIBCPP_POP_MACROS to the relevant headers?

The review process

After uploading your patch, you should see that the “libc++” review group is automatically added as a reviewer for your patch. Once the group is marked as having approved your patch, you can commit it. However, if you get an approval very quickly for a significant patch, please try to wait a couple of business days before committing to give the opportunity for other reviewers to chime in. If you need someone else to commit the patch for you, please mention it and provide your Name <email@domain> for us to attribute the commit properly.

Note that the rule for accepting as the “libc++” review group is to wait for two members of the group to have approved the patch, excluding the patch author. This is not a hard rule – for very simple patches, use your judgement. The “libc++” review group consists of frequent libc++ contributors with a good understanding of the project’s guidelines – if you would like to be added to it, please reach out on Discord.

Exporting new symbols from the library

When exporting new symbols from libc++, you must update the ABI lists located in lib/abi. To test whether the lists are up-to-date, please run the target check-cxx-abilist. To regenerate the lists, use the target generate-cxx-abilist. The ABI lists must be updated for all supported platforms; currently Linux and Apple. If you don’t have access to one of these platforms, you can download an updated list from the failed build at Buildkite. Look for the failed build and select the artifacts tab. There, download the abilist for the platform, e.g.:

  • C++<version>.

  • MacOS X86_64 and MacOS arm64 for the Apple platform.

Pre-commit CI


Unlike most parts of the LLVM project, libc++ uses a pre-commit CI [1]. This CI is hosted on Buildkite and the build results are visible in the review on GitHub. Please make sure the CI is green before committing a patch.

The CI tests libc++ for all supported platforms. The build is started for every commit added to a Pull Request. A complete CI run takes approximately one hour. To reduce the load:

  • The build is cancelled when a new commit is pushed to a PR that is already running CI.

  • The build is done in several stages and cancelled when a stage fails.

Typically, the libc++ jobs use a Ubuntu Docker image. This image contains recent nightly builds of all supported versions of Clang and the current version of the main branch. These versions of Clang are used to build libc++ and execute its tests.

Unless specified otherwise, the configurations:

  • use a nightly build of the main branch of Clang,

  • execute the tests using the language C++<latest>. This is the version “developed” by the C++ committee.


Updating the Clang nightly builds in the Docker image is a manual process and is done at an irregular interval on purpose. When you need to have the latest nightly build to test recent Clang changes, ask in the #libcxx channel on LLVM’s Discord server.


Below is a short description of the most interesting CI builds [2]:

  • Format runs clang-format and uploads its output as an artifact. At the moment this build is a soft error and doesn’t fail the build.

  • Generated output runs the libcxx-generate-files build target and tests for non-ASCII characters in libcxx. Some files are excluded since they use Unicode, mainly tests. The output of these commands are uploaded as artifact.

  • Documentation builds the documentation. (This is done early in the build process since it is cheap to run.)

  • C++<version> these build steps test the various C++ versions, making sure all C++ language versions work with the changes made.

  • Clang <version> these build steps test whether the changes work with all supported Clang versions.

  • Booststrapping build builds Clang using the revision of the patch and uses that Clang version to build and test libc++. This validates the current Clang and lib++ are compatible.

    When a crash occurs in this build, the crash reproducer is available as an artifact.

  • Modular build tests libc++ using Clang modules [3].

  • GCC <version> tests libc++ with the latest stable GCC version. Only C++11 and the latest C++ version are tested.

  • Santitizers tests libc++ using the Clang sanitizers.

  • Parts disabled tests libc++ with certain libc++ features disabled.

  • Windows tests libc++ using MinGW and clang-cl.

  • Apple tests libc++ on MacOS.

  • ARM tests libc++ on various Linux ARM platforms.

  • AIX tests libc++ on AIX.


All files of the CI infrastructure are in the directory libcxx/utils/ci. Note that quite a bit of this infrastructure is heavily Linux focused. This is the platform used by most of libc++’s Buildkite runners and developers.


Contains the Docker image for the Ubuntu CI. Because the same Docker image is used for the main and release branch, it should contain no hard-coded versions. It contains the used versions of Clang, various clang-tools, GCC, and CMake.


This image is pulled from Docker hub and not rebuild when changing the Dockerfile.


Helper script that pulls and runs the Docker image. This image mounts the LLVM monorepo at /llvm. This can be used to test with compilers not available on your system.


Contains the build script executed on Buildkite. This script can be executed locally or inside run-buildbot-container. The script must be called with the target to test. For example, run-buildbot generic-cxx20 will build libc++ and test it using C++20.


This script will overwrite the directory <llvm-root>/build/XX where XX is the target of run-buildbot.

This script contains as little version information as possible. This makes it easy to use the script with a different compiler. This allows testing a combination not in the libc++ CI. It can be used to add a new (temporary) job to the CI. For example, testing the C++17 build with Clang-14 can be done like:

CC=clang-14 CXX=clang++-14 run-buildbot generic-cxx17


Contains the jobs executed in the CI. This file contains the version information of the jobs being executed. Since this script differs between the main and release branch, both branches can use different compiler versions.