fetch-crx3.py - mozsearch

Enable keyboard shortcuts

#!/usr/bin/python3 -u

# This Source Code Form is subject to the terms of the Mozilla Public

# License, v. 2.0. If a copy of the MPL was not distributed with this

# file, You can obtain one at http://mozilla.org/MPL/2.0/.

"""

This script downloads a Chrome component, as distributed in the CRX3 format,

and repacks it in the layout Gecko expects of a GMP plugin directory.

A CRX3 is a zip with a header prepended. Zip readers find the central

directory from the end of the file and skip over the header, so it can be

read without any special handling, which is how GMPInstallManager consumes

these in the browser.

"""

import argparse

import hashlib

import io

import os

import pathlib

import tarfile

import zipfile

import requests

import zstandard

# Mirrors GMPExtractor.worker.js, which keeps only the binaries and metadata.

# Also flattens away the platform specific subdirectory they are packaged in.

LIB_SUFFIXES = (".dll", ".dylib", ".so")

KEEP_SUFFIXES = LIB_SUFFIXES + (".info", ".sig", ".txt")

KEEP_NAMES = ("LICENSE", "manifest.json")

def fetch(url, sha256, size):

    response = requests.get(url, timeout=120)

    response.raise_for_status()

    data = response.content

    if len(data) != size:

        raise Exception(f"{url} is {len(data)} bytes, expected {size}")

    digest = hashlib.sha256(data).hexdigest()

    if digest != sha256:

        raise Exception(f"{url} hashes to {digest}, expected {sha256}")

    return data

def repack(data, dest, prefix):

    dest.parent.mkdir(parents=True, exist_ok=True)

    kept = []

    libs = []

    with zipfile.ZipFile(io.BytesIO(data)) as zf, open(dest, "wb") as fh:

        compressor = zstandard.ZstdCompressor()

        with compressor.stream_writer(fh) as stream, tarfile.open(

            fileobj=stream, mode="w:"

        ) as tar:

            for entry in zf.infolist():

                if entry.is_dir():

                    continue

                name = os.path.basename(entry.filename)

                if not name.endswith(KEEP_SUFFIXES) and name not in KEEP_NAMES:

                    continue

                info = tarfile.TarInfo(prefix + name)

                if info.name in kept:

                    raise Exception(

                        f"Flattening {entry.filename} would collide with an "

                        f"entry already packed as {info.name}"

                info.size = entry.file_size

                info.mode = 0o755 if name.endswith(LIB_SUFFIXES) else 0o644

                # The zip holds local times with no zone, so use a fixed

                # timestamp rather than something that varies per download.

                info.mtime = 0

                with zf.open(entry) as src:

                    tar.addfile(info, src)

                kept.append(info.name)

                if name.endswith(LIB_SUFFIXES):

                    libs.append(info.name)

    if not libs:

        raise Exception(f"No plugin library found in the archive, kept {kept}")

    for name in kept:

        print(f"Packed {name}")

def main():

    parser = argparse.ArgumentParser()

    parser.add_argument("--sha256", required=True, help="SHA-256 of the download")

    parser.add_argument("--size", required=True, type=int, help="size of the download")

    parser.add_argument(

        "--add-prefix",

        default="",

        help="prefix to give each file in the resulting archive",

    parser.add_argument("url", help="URL of the CRX3 to download")

    parser.add_argument("dest", help="path of the .tar.zst to produce")

    args = parser.parse_args()

    data = fetch(args.url, args.sha256, args.size)

    repack(data, pathlib.Path(args.dest), args.add_prefix)

if __name__ == "__main__":

    main()