zulip/tools/check-thirdparty

#!/usr/bin/env python3
"""Check the docs/THIRDPARTY file against the DEP-5 copyright format.

https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/

Disclaimer: This script is not a lawyer.  It cannot validate that the
claimed licenses are correct.  It can only check for basic syntactic
issues.
"""
import difflib
import io
import os
import subprocess
import sys

from debian import copyright

COPYRIGHT_FILENAME = "docs/THIRDPARTY"
os.chdir(os.path.join(os.path.dirname(__file__), ".."))
status = 0

*files, empty = subprocess.check_output(
    ["git", "ls-files", "-z"],
    encoding="utf-8",
).split("\0")
assert empty == ""
files += [
    "static/generated/emoji/images/emoji/unicode/ignore-this-path",
]

with open(COPYRIGHT_FILENAME) as f:
    lines = list(f)
c = copyright.Copyright(lines)

if not c.header.known_format():
    print(f"{COPYRIGHT_FILENAME}: Unknown header format {c.header.format}")
    status = 1

defined_licenses = {
    p.license.to_str().split("\n", 1)[0]
    for p in c.all_license_paragraphs()
    if "\n" in p.license.to_str()
}

for p in c.all_files_paragraphs():
    for g in p.files:
        if not any(map(copyright.globs_to_re([g]).fullmatch, files)):
            print(f"{COPYRIGHT_FILENAME}: No such file {g}")
            status = 1

    if "\n" not in p.license.to_str() and p.license.to_str() not in defined_licenses:
        print(f"{COPYRIGHT_FILENAME}: Missing license text for {p.license.to_str()}")
        status = 1

dumped = c.dump()
if dumped != "".join(lines):
    print(f"{COPYRIGHT_FILENAME}: Changes expected:")
    sys.stdout.writelines(
        difflib.unified_diff(
            lines,
            io.StringIO(dumped).readlines(),
            COPYRIGHT_FILENAME,
            COPYRIGHT_FILENAME,
        ),
    )
    status = 1

sys.exit(status)