utils/test-at-commit

#!/usr/bin/env python3

import argparse
import os
import pathlib
import subprocess
import sys
import tempfile

PARENT_DIR = pathlib.Path(os.path.dirname(os.path.abspath(__file__)))

# Unofficial list of directories required to build libc++. This is a best guess that should work
# when checking out the monorepo at most commits, but it's technically not guaranteed to work
# (especially for much older commits).
LIBCXX_REQUIRED_DIRECTORIES = [
    'libcxx',
    'libcxxabi',
    'llvm/cmake',
    'llvm/utils/llvm-lit',
    'llvm/utils/lit',
    'runtimes',
    'cmake',
    'third-party/benchmark',
    'libc'
]

def checkout_subdirectories(git_repo, commit, paths, destination):
    """
    Produce a copy of the specified Git-tracked files/directories at the given commit.
    The resulting files and directories at placed at the given location.
    """
    with tempfile.TemporaryDirectory() as tmp:
        tmpfile = os.path.join(tmp, 'archive.tar.gz')
        git_archive = ['git', '-C', git_repo, 'archive', '--format', 'tar.gz', '--output', tmpfile, commit, '--'] + list(paths)
        subprocess.check_call(git_archive)
        os.makedirs(destination, exist_ok=True)
        subprocess.check_call(['tar', '-x', '-z', '-f', tmpfile, '-C', destination])

def exists_in_commit(git_repo, commit, path):
    """
    Return whether the given path (file or directory) existed at the given commit.
    """
    cmd = ['git', '-C', git_repo, 'show', f'{commit}:{path}']
    result = subprocess.call(cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
    return result == 0

def directory_path(string):
    if os.path.isdir(string):
        return pathlib.Path(string)
    else:
        raise NotADirectoryError(string)

def main(argv):
    parser = argparse.ArgumentParser(
        prog='test-at-commit',
        description='Test a Standard Library against the test suite at the specified commit (or the currently '
                    'checked-out sources by default). This makes it easier to perform historical analyses of '
                    'libc++ behavior, gather historical performance data, bisect issues, and so on.')
    parser.add_argument('--build-dir', '-B', type=pathlib.Path, required=True,
        help='Path to create the build directory for running the test suite at. The results of the tests '
             'are located in that directory after the run.')
    parser.add_argument('--test-config', type=str, required=True,
        help='The testing configuration to use, which determines the Standard Library being tested. This has '
             'the same meaning as the LIBCXX_TEST_CONFIG CMake option: either an absolute path to a Lit '
             'configuration file, or a path relative to the libcxx/test/configs directory of the test suite '
             'being used. Note that when --test-suite-commit is used, a relative path is resolved against '
             'that older version of the test suite, so an absolute path is usually what you want.')
    parser.add_argument('--compiler', type=pathlib.Path, required=True,
        help='Path to the compiler to use to run the tests.')
    parser.add_argument('--test-suite-commit', type=str, required=False,
        help='Commit to use for the test suite. If left unspecified, the currently checked-out version of the '
             'test suite is used. Otherwise, the requested version is checked out in a separate directory and '
             'that version of the test suite is used.')
    parser.add_argument('lit_options', nargs=argparse.REMAINDER,
        help='Optional arguments passed to lit when running the tests. Should be provided last and '
             'separated from other arguments with a `--`.')
    parser.add_argument('--git-repo', type=directory_path, default=pathlib.Path(os.getcwd()),
        help='Optional path to the Git repository to use. By default, the current working directory is used.')
    parser.add_argument('--tmp-src-dir', type=pathlib.Path, required=False,
        help='Optional path to use for the ephemeral checkout of the test suite. By default, a temporary '
             'directory is used and it is cleaned up after the run. If a custom directory is specified, it '
             'is not cleaned up automatically. Note that the build directory refers to that checkout, so '
             'keeping it around is required in order to use the build directory after the run.')
    args = parser.parse_args(argv)

    args.build_dir = args.build_dir.resolve()

    # Gather lit options
    lit_options = []
    if args.lit_options is not None:
        if args.lit_options[0] != '--':
            raise ArgumentError('For clarity, Lit options must be separated from other options by --')
        lit_options = args.lit_options[1:]

    # This is the list of directories that must be cleaned up before we return
    tempdirs = []
    try:
        # If needed, check out the test suite at the commit we're going to use for the suite
        if args.test_suite_commit is None:
            test_suite_sources = args.git_repo
        else:
            if args.tmp_src_dir is not None:
                test_suite_sources = args.tmp_src_dir
            else:
                tempdirs.append(tempfile.TemporaryDirectory())
                test_suite_sources = pathlib.Path(tempdirs[-1].name)
            checkout_dirs = [d for d in LIBCXX_REQUIRED_DIRECTORIES if exists_in_commit(args.git_repo, args.test_suite_commit, d)]
            checkout_subdirectories(args.git_repo, args.test_suite_commit, checkout_dirs, test_suite_sources)

        # Configure the test suite in the specified build directory.
        args.build_dir.mkdir(parents=True, exist_ok=True)
        test_suite_cmd = ['cmake', '-B', args.build_dir, '-S', test_suite_sources / 'runtimes', '-G', 'Ninja']
        test_suite_cmd += ['-D', f'CMAKE_CXX_COMPILER={args.compiler}']
        test_suite_cmd += ['-D', 'LLVM_ENABLE_RUNTIMES=libcxx;libcxxabi']
        test_suite_cmd += ['-D', 'LIBCXXABI_USE_LLVM_UNWINDER=OFF']
        test_suite_cmd += ['-D', f'LIBCXX_TEST_CONFIG={args.test_config}']
        subprocess.check_call(test_suite_cmd)

        # Run the specified tests against the built library
        lit_cmd = [PARENT_DIR / 'libcxx-lit', args.build_dir] + lit_options
        subprocess.check_call(lit_cmd)
    finally:
        for d in tempdirs:
            d.cleanup()


if __name__ == '__main__':
    main(sys.argv[1:])