meson/mesonbuild/mdist.py

# SPDX-License-Identifier: Apache-2.0
# Copyright 2017 The Meson development team
# Copyright © 2023-2024 Intel Corporation

from __future__ import annotations


import abc
import argparse
import gzip
import os
import sys
import shlex
import shutil
import subprocess
import tarfile
import tempfile
import hashlib
import typing as T

from dataclasses import dataclass
from glob import glob
from pathlib import Path
from mesonbuild.environment import Environment, detect_ninja
from mesonbuild.mesonlib import (GIT, MesonException, RealPathAction, get_meson_command, quiet_git,
                                 windows_proof_rmtree, setup_vsenv)
from .options import OptionKey
from mesonbuild.msetup import add_arguments as msetup_argparse
from mesonbuild.wrap import wrap
from mesonbuild import mlog, build, coredata
from .scripts.meson_exe import run_exe

if T.TYPE_CHECKING:
    from ._typing import ImmutableListProtocol
    from .mesonlib import ExecutableSerialisation

archive_choices = ['bztar', 'gztar', 'xztar', 'zip']

archive_extension = {'bztar': '.tar.bz2',
                     'gztar': '.tar.gz',
                     'xztar': '.tar.xz',
                     'zip': '.zip'}

# Note: when adding arguments, please also add them to the completion
# scripts in $MESONSRC/data/shell-completions/
def add_arguments(parser: argparse.ArgumentParser) -> None:
    parser.add_argument('-C', dest='wd', action=RealPathAction,
                        help='directory to cd into before running')
    parser.add_argument('--allow-dirty', action='store_true',
                        help='Allow even when repository contains uncommitted changes.')
    parser.add_argument('--formats', default='xztar',
                        help='Comma separated list of archive types to create. Supports xztar (default), bztar, gztar, and zip.')
    parser.add_argument('--include-subprojects', action='store_true',
                        help='Include source code of subprojects that have been used for the build.')
    parser.add_argument('--no-tests', action='store_true',
                        help='Do not build and test generated packages.')


def create_hash(fname: str) -> None:
    hashname = fname + '.sha256sum'
    m = hashlib.sha256()
    m.update(open(fname, 'rb').read())
    with open(hashname, 'w', encoding='utf-8') as f:
        # A space and an asterisk because that is the format defined by GNU coreutils
        # and accepted by busybox and the Perl shasum tool.
        f.write('{} *{}\n'.format(m.hexdigest(), os.path.basename(fname)))


msg_uncommitted_changes = 'Repository has uncommitted changes that will not be included in the dist tarball'

def handle_dirty_opt(msg: str, allow_dirty: bool) -> None:
    if allow_dirty:
        mlog.warning(msg)
    else:
        mlog.error(msg + '\n' + 'Use --allow-dirty to ignore the warning and proceed anyway')
        sys.exit(1)

def is_git(src_root: str) -> bool:
    '''
    Checks if meson.build file at the root source directory is tracked by git.
    It could be a subproject part of the parent project git repository.
    '''
    if quiet_git(['ls-files', '--error-unmatch', 'meson.build'], src_root)[0]:
        return True

    if os.path.exists(os.path.join(src_root, '.git')):
        msg = 'Source tree looks like it may be a git repo, '
        if not GIT:
            msg += 'but git is not installed!'
            if 'GITLAB_CI' in os.environ:
                msg += ' This is a gitlab bug.'
        else:
            msg += 'but git returned a failure. '
            p, oe = quiet_git(['status'], src_root)
            if 'dubious ownership' in oe:
                # For a few years now, git has absolved itself of the responsibility to implement
                # robust, safe software. Instead of detecting the signs of a problematic scenario,
                # they have chosen to consider many legitimate and reasonable use cases as "dangerous",
                # and implemented the number one threat to security worldwide: alert fatigue. Having
                # done so, they then washed their hands of the matter and permanently tabled the
                # notion of adding fine-grained detection. This is not just useless, it is *worse*
                # than useless.
                #
                # In our case, the error is triply meaningless since we are already executing build
                # system commands from the same directory. Either way, reject the notion that git is
                # well designed or that its error messaging is a valid approach to the problem space.
                msg += 'This is a bug in git itself, please set `git config --global safe.directory "*"`'
            else:
                msg += 'meson.build may not have been committed to git?'
        mlog.warning(msg)
    return False


def is_hg(src_root: str) -> bool:
    return os.path.isdir(os.path.join(src_root, '.hg'))


@dataclass
class Dist(metaclass=abc.ABCMeta):
    dist_name: str
    src_root: str
    bld_root: str
    dist_scripts: T.List[ExecutableSerialisation]
    subprojects: T.Dict[str, str]
    options: argparse.Namespace

    def __post_init__(self) -> None:
        self.dist_sub = os.path.join(self.bld_root, 'meson-dist')
        self.distdir = os.path.join(self.dist_sub, self.dist_name)

    @abc.abstractmethod
    def create_dist(self, archives: T.List[str]) -> T.List[str]:
        pass

    def run_dist_scripts(self) -> None:
        assert os.path.isabs(self.distdir)
        mesonrewrite = Environment.get_build_command() + ['rewrite']
        env = {'MESON_DIST_ROOT': self.distdir,
               'MESON_SOURCE_ROOT': self.src_root,
               'MESON_BUILD_ROOT': self.bld_root,
               'MESONREWRITE': ' '.join(shlex.quote(x) for x in mesonrewrite),
               }
        for d in self.dist_scripts:
            if d.subproject and d.subproject not in self.subprojects:
                continue
            subdir = self.subprojects.get(d.subproject, '')
            env['MESON_PROJECT_DIST_ROOT'] = os.path.join(self.distdir, subdir)
            env['MESON_PROJECT_SOURCE_ROOT'] = os.path.join(self.src_root, subdir)
            env['MESON_PROJECT_BUILD_ROOT'] = os.path.join(self.bld_root, subdir)
            name = ' '.join(d.cmd_args)
            print(f'Running custom dist script {name!r}')
            try:
                rc = run_exe(d, env)
                if rc != 0:
                    sys.exit('Dist script errored out')
            except OSError:
                print(f'Failed to run dist script {name!r}')
                sys.exit(1)


class GitDist(Dist):
    def git_root(self, dir_: str) -> Path:
        # Cannot use --show-toplevel here because git in our CI prints cygwin paths
        # that python cannot resolve. Workaround this by taking parent of src_root.
        prefix = quiet_git(['rev-parse', '--show-prefix'], dir_, check=True)[1].strip()
        if not prefix:
            return Path(dir_)
        prefix_level = len(Path(prefix).parents)
        return Path(dir_).parents[prefix_level - 1]

    def have_dirty_index(self) -> bool:
        '''Check whether there are uncommitted changes in git'''
        # Optimistically call update-index, and disregard its return value. It could be read-only,
        # and only the output of diff-index matters.
        subprocess.call(['git', '-C', self.src_root, 'update-index', '-q', '--refresh'])
        ret = subprocess.call(['git', '-C', self.src_root, 'diff-index', '--quiet', 'HEAD'])
        return ret == 1

    def copy_git(self, src: T.Union[str, os.PathLike], distdir: str, revision: str = 'HEAD',
                 prefix: T.Optional[str] = None, subdir: T.Optional[str] = None) -> None:
        cmd = ['git', 'archive', '--format', 'tar', revision]
        if prefix is not None:
            cmd.insert(2, f'--prefix={prefix}/')
        if subdir is not None:
            cmd.extend(['--', subdir])
        with tempfile.TemporaryFile() as f:
            subprocess.check_call(cmd, cwd=src, stdout=f)
            f.seek(0)
            t = tarfile.open(fileobj=f) # [ignore encoding]
            t.extractall(path=distdir)

    def process_git_project(self, src_root: str, distdir: str) -> None:
        if self.have_dirty_index():
            handle_dirty_opt(msg_uncommitted_changes, self.options.allow_dirty)
        if os.path.exists(distdir):
            windows_proof_rmtree(distdir)
        repo_root = self.git_root(src_root)
        if repo_root.samefile(src_root):
            os.makedirs(distdir)
            self.copy_git(src_root, distdir)
        else:
            subdir = Path(src_root).relative_to(repo_root)
            tmp_distdir = distdir + '-tmp'
            if os.path.exists(tmp_distdir):
                windows_proof_rmtree(tmp_distdir)
            os.makedirs(tmp_distdir)
            self.copy_git(repo_root, tmp_distdir, subdir=str(subdir))
            Path(tmp_distdir, subdir).rename(distdir)
            windows_proof_rmtree(tmp_distdir)
        self.process_submodules(src_root, distdir)

    def process_submodules(self, src: str, distdir: str) -> None:
        module_file = os.path.join(src, '.gitmodules')
        if not os.path.exists(module_file):
            return
        cmd = ['git', 'submodule', 'status', '--cached', '--recursive']
        modlist = subprocess.check_output(cmd, cwd=src, universal_newlines=True).splitlines()
        for submodule in modlist:
            status = submodule[:1]
            sha1, rest = submodule[1:].split(' ', 1)
            subpath = rest.rsplit(' ', 1)[0]

            if status == '-':
                mlog.warning(f'Submodule {subpath!r} is not checked out and cannot be added to the dist')
                continue
            elif status in {'+', 'U'}:
                handle_dirty_opt(f'Submodule {subpath!r} has uncommitted changes that will not be included in the dist tarball', self.options.allow_dirty)

            self.copy_git(os.path.join(src, subpath), distdir, revision=sha1, prefix=subpath)

    def create_dist(self, archives: T.List[str]) -> T.List[str]:
        self.process_git_project(self.src_root, self.distdir)
        for path in self.subprojects.values():
            sub_src_root = os.path.join(self.src_root, path)
            sub_distdir = os.path.join(self.distdir, path)
            if os.path.exists(sub_distdir):
                continue
            if is_git(sub_src_root):
                self.process_git_project(sub_src_root, sub_distdir)
            else:
                shutil.copytree(sub_src_root, sub_distdir)
        self.run_dist_scripts()
        output_names = []
        for a in archives:
            compressed_name = self.distdir + archive_extension[a]
            shutil.make_archive(self.distdir, a, root_dir=self.dist_sub, base_dir=self.dist_name)
            output_names.append(compressed_name)
        windows_proof_rmtree(self.distdir)
        return output_names


class HgDist(Dist):
    def have_dirty_index(self) -> bool:
        '''Check whether there are uncommitted changes in hg'''
        env = os.environ.copy()
        env['LC_ALL'] = 'C'
        # cpython's gettext has a bug and uses LANGUAGE to override LC_ALL,
        # contrary to the gettext spec
        env.pop('LANGUAGE', None)
        out = subprocess.check_output(['hg', '-R', self.src_root, 'summary'], env=env)
        return b'commit: (clean)' not in out

    def create_dist(self, archives: T.List[str]) -> T.List[str]:
        if self.have_dirty_index():
            handle_dirty_opt(msg_uncommitted_changes, self.options.allow_dirty)
        if self.dist_scripts:
            mlog.warning('dist scripts are not supported in Mercurial projects')

        os.makedirs(self.dist_sub, exist_ok=True)
        tarname = os.path.join(self.dist_sub, self.dist_name + '.tar')
        xzname = tarname + '.xz'
        bz2name = tarname + '.bz2'
        gzname = tarname + '.gz'
        zipname = os.path.join(self.dist_sub, self.dist_name + '.zip')
        # Note that -X interprets relative paths using the current working
        # directory, not the repository root, so this must be an absolute path:
        # https://bz.mercurial-scm.org/show_bug.cgi?id=6267
        #
        # .hg[a-z]* is used instead of .hg* to keep .hg_archival.txt, which may
        # be useful to link the tarball to the Mercurial revision for either
        # manual inspection or in case any code interprets it for a --version or
        # similar.
        subprocess.check_call(['hg', 'archive', '-R', self.src_root, '-S', '-t', 'tar',
                               '-X', self.src_root + '/.hg[a-z]*', tarname])
        output_names = []
        if 'xztar' in archives:
            import lzma
            with lzma.open(xzname, 'wb') as xf, open(tarname, 'rb') as tf:
                shutil.copyfileobj(tf, xf)
            output_names.append(xzname)
        if 'bztar' in archives:
            import bz2
            with bz2.open(bz2name, 'wb') as bf, open(tarname, 'rb') as tf:
                shutil.copyfileobj(tf, bf)
            output_names.append(bz2name)
        if 'gztar' in archives:
            with gzip.open(gzname, 'wb') as zf, open(tarname, 'rb') as tf:
                shutil.copyfileobj(tf, zf)
            output_names.append(gzname)
        os.unlink(tarname)
        if 'zip' in archives:
            subprocess.check_call(['hg', 'archive', '-R', self.src_root, '-S', '-t', 'zip', zipname])
            output_names.append(zipname)
        return output_names


def run_dist_steps(meson_command: T.List[str], unpacked_src_dir: str, builddir: str, installdir: str, ninja_args: T.List[str]) -> int:
    if subprocess.call(meson_command + ['--backend=ninja', unpacked_src_dir, builddir]) != 0:
        print('Running Meson on distribution package failed')
        return 1
    if subprocess.call(ninja_args, cwd=builddir) != 0:
        print('Compiling the distribution package failed')
        return 1
    if subprocess.call(ninja_args + ['test'], cwd=builddir) != 0:
        print('Running unit tests on the distribution package failed')
        return 1
    myenv = os.environ.copy()
    myenv['DESTDIR'] = installdir
    if subprocess.call(ninja_args + ['install'], cwd=builddir, env=myenv) != 0:
        print('Installing the distribution package failed')
        return 1
    return 0

def check_dist(packagename: str, _meson_command: ImmutableListProtocol[str], extra_meson_args: T.List[str], bld_root: str, privdir: str) -> int:
    print(f'Testing distribution package {packagename}')
    unpackdir = os.path.join(privdir, 'dist-unpack')
    builddir = os.path.join(privdir, 'dist-build')
    installdir = os.path.join(privdir, 'dist-install')
    for p in (unpackdir, builddir, installdir):
        if os.path.exists(p):
            windows_proof_rmtree(p)
        os.mkdir(p)
    ninja_args = detect_ninja()
    shutil.unpack_archive(packagename, unpackdir)
    unpacked_files = glob(os.path.join(unpackdir, '*'))
    assert len(unpacked_files) == 1
    unpacked_src_dir = unpacked_files[0]
    meson_command = _meson_command.copy()
    meson_command += ['setup']
    meson_command += create_cmdline_args(bld_root)
    meson_command += extra_meson_args

    ret = run_dist_steps(meson_command, unpacked_src_dir, builddir, installdir, ninja_args)
    if ret > 0:
        print(f'Dist check build directory was {builddir}')
    else:
        windows_proof_rmtree(unpackdir)
        windows_proof_rmtree(builddir)
        windows_proof_rmtree(installdir)
        print(f'Distribution package {packagename} tested')
    return ret

def create_cmdline_args(bld_root: str) -> T.List[str]:
    parser = argparse.ArgumentParser()
    msetup_argparse(parser)
    args = T.cast('coredata.SharedCMDOptions', parser.parse_args([]))
    coredata.parse_cmd_line_options(args)
    coredata.read_cmd_line_file(bld_root, args)
    args.cmd_line_options.pop(OptionKey('backend'), '')
    return shlex.split(coredata.format_cmd_line_options(args))

def determine_archives_to_generate(options: argparse.Namespace) -> T.List[str]:
    result = []
    for i in options.formats.split(','):
        if i not in archive_choices:
            sys.exit(f'Value "{i}" not one of permitted values {archive_choices}.')
        result.append(i)
    if len(i) == 0:
        sys.exit('No archive types specified.')
    return result

def run(options: argparse.Namespace) -> int:
    buildfile = Path(options.wd) / 'meson-private' / 'build.dat'
    if not buildfile.is_file():
        raise MesonException(f'Directory {options.wd!r} does not seem to be a Meson build directory.')
    b = build.load(options.wd)
    need_vsenv = T.cast('bool', b.environment.coredata.get_option(OptionKey('vsenv')))
    setup_vsenv(need_vsenv)
    src_root = b.environment.source_dir
    bld_root = b.environment.build_dir
    priv_dir = os.path.join(bld_root, 'meson-private')

    dist_name = b.project_name + '-' + b.project_version

    archives = determine_archives_to_generate(options)

    subprojects = {}
    extra_meson_args = []
    if options.include_subprojects:
        subproject_dir = os.path.join(src_root, b.subproject_dir)
        for sub in b.subprojects:
            directory = wrap.get_directory(subproject_dir, sub)
            subprojects[sub] = os.path.join(b.subproject_dir, directory)
        extra_meson_args.append('-Dwrap_mode=nodownload')

    cls: T.Type[Dist]
    if is_git(src_root):
        cls = GitDist
    elif is_hg(src_root):
        if subprojects:
            print('--include-subprojects option currently not supported with Mercurial')
            return 1
        cls = HgDist
    else:
        print('Dist currently only works with Git or Mercurial repos')
        return 1

    project = cls(dist_name, src_root, bld_root, b.dist_scripts, subprojects, options)
    names = project.create_dist(archives)

    if names is None:
        return 1
    rc = 0
    if not options.no_tests:
        # Check only one.
        rc = check_dist(names[0], get_meson_command(), extra_meson_args, bld_root, priv_dir)
    if rc == 0:
        for name in names:
            create_hash(name)
            print('Created', name)
    return rc