bazel/generate_gnulib_srcs.py

"""Generates Starlark source lists for Gnulib license tiers using pygnulib."""

import argparse
import pathlib
import sys

from pygnulib import constants
from pygnulib.GLConfig import GLConfig
from pygnulib.GLModuleSystem import GLModule, GLModuleSystem

sys.dont_write_bytecode = True

ALL_COMPATIBLE_LICENSES = frozenset({
    "GPLv2+ build tool",
    "GPLed build tool",
    "public domain",
    "unlimited",
    "unmodifiable license text",
})

LGPL2_COMPATIBLE_LICENSES = ALL_COMPATIBLE_LICENSES | frozenset({
    "LGPLv2+",
})

LGPL3_COMPATIBLE_LICENSES = ALL_COMPATIBLE_LICENSES | frozenset({
    "LGPLv2+",
    "LGPLv3+ or GPLv2+",
    "LGPLv3+",
    "LGPL",
})

INFRASTRUCTURE_PREFIXES = (
    "pygnulib/",
    "build-aux/",
    "top/",
    "config/",
)

INFRASTRUCTURE_EXACT_FILES = frozenset({
    "gnulib-tool",
    "gnulib-tool.py",
    ".gnulib-tool.py",
    "gnulib-tool.sh",
    "m4/gnulib-tool.m4",
    "doc/INSTALL",
    "doc/COPYING.LESSERv2",
    "doc/COPYING.LESSERv3",
    "doc/COPYINGv2",
    "doc/COPYINGv3",
})


def _find_gnulib_dir(inputs: list[pathlib.Path]) -> pathlib.Path:
    matches = [input_path for input_path in inputs if input_path.name == "gnulib-tool.py"]
    if len(matches) != 1:
        raise ValueError(
            f"Expected exactly 1 'gnulib-tool.py' among inputs, found {len(matches)}: {matches}"
        )
    return matches[0].parent


def _format_starlark_tuple(name: str, items: list[str]) -> str:
    lines = [f"{name} = ("]
    for item in items:
        lines.append(f'    "{item}",')
    lines.append(")")
    return "\n".join(lines)


def generate_gnulib_srcs(inputs: list[pathlib.Path], output: pathlib.Path) -> None:
    gnulib_dir = _find_gnulib_dir(inputs)
    constants.init_DIRS(str(gnulib_dir))

    existing_files: set[str] = set()
    for input_path in inputs:
        relative_path = input_path.relative_to(gnulib_dir).as_posix()
        existing_files.add(relative_path)

    config = GLConfig()
    module_system = GLModuleSystem(config)

    non_test_modules: list[GLModule] = []
    for relative_path in sorted(existing_files):
        if not relative_path.startswith("modules/"):
            continue
        module_name = relative_path.removeprefix("modules/")
        if not module_system.file_is_module(module_name):
            continue
        module = module_system.find(module_name)
        if module is not None and module.isNonTests():
            non_test_modules.append(module)

    all_module_files: set[str] = set()
    for module in non_test_modules:
        all_module_files.update(module.getFiles())

    infrastructure_files = {
        relative_path
        for relative_path in existing_files - all_module_files
        if relative_path.startswith(INFRASTRUCTURE_PREFIXES)
        or relative_path in INFRASTRUCTURE_EXACT_FILES
    }

    def collect_tier_files(tier_name: str, allowed_licenses: frozenset[str]) -> list[str]:
        allowed_modules = [
            module for module in non_test_modules if module.getLicense() in allowed_licenses
        ]
        allowed_module_set = set(allowed_modules)

        for module in allowed_modules:
            for dependency in module.getDependenciesWithoutConditions():
                if dependency not in allowed_module_set:
                    raise ValueError(
                        f"Module '{module.name}' (license: '{module.getLicense()}') in {tier_name} "
                        f"depends on '{dependency.name}' (license: '{dependency.getLicense()}'), "
                        f"which is outside the allowed module set for {tier_name}."
                    )

        tier_files = set(infrastructure_files)
        for module in allowed_modules:
            module_files = {f"modules/{module.name}", *module.getFiles()}
            tier_files |= module_files & existing_files

        return sorted(tier_files)

    lgpl2_srcs = collect_tier_files("LGPL2_SRCS", LGPL2_COMPATIBLE_LICENSES)
    lgpl3_srcs = collect_tier_files("LGPL3_SRCS", LGPL3_COMPATIBLE_LICENSES)

    header = "\n".join([
        "# Generated by //third_party/gnulib:write_gnulib_srcs. DO NOT EDIT.",
        "# Regenerate with: bazel run //third_party/gnulib:write_gnulib_srcs",
        '"""Generated Gnulib license-scoped source file lists."""',
    ])
    content = "\n\n".join([
        header,
        _format_starlark_tuple("LGPL2_SRCS", lgpl2_srcs),
        _format_starlark_tuple("LGPL3_SRCS", lgpl3_srcs),
    ]) + "\n"

    output.parent.mkdir(parents=True, exist_ok=True)
    output.write_text(content, encoding="utf-8")


def main() -> None:
    parser = argparse.ArgumentParser(
        description="Generate Starlark source lists for Gnulib license tiers.",
        fromfile_prefix_chars="@",
    )
    parser.add_argument(
        "--output",
        "-o",
        required=True,
        type=pathlib.Path,
        help="Output .bzl file path",
    )
    parser.add_argument(
        "inputs",
        nargs="*",
        type=pathlib.Path,
        help="Input files from the Gnulib repository",
    )
    args = parser.parse_args()
    generate_gnulib_srcs(args.inputs, args.output)


if __name__ == "__main__":
    main()