[pro-laxmi]

Summary of Changes

When /etc/os-release is truncated, corrupted, or contains unclosed quotation marks (e.g. NAME="Fedora Linux), shlex.shlex raises ValueError("No closing quotation"). Currently, this unhandled exception crashes the entire program and any application or library importing distro.

Root Cause

_parse_os_release_content called list(lexer) without catching ValueError.

Fix

Wrapped shlex lexer tokenization in a try...except ValueError block with an automatic fallback to robust line-by-line parsing (key, val = line.split("=", 1)), stripping surrounding quotes.

Verification & Testing

Tested with malformed os-release streams and unclosed quotes:

bad_content = io.StringIO('NAME="Fedora Linux\nID=fedora\nVERSION_ID=44\n')
info = distro.LinuxDistribution._parse_os_release_content(bad_content)
assert info['id'] == 'fedora'
assert info['name'] == 'Fedora Linux'

Parsed successfully without crashing.

[HorlogeSkynet]

Hey @pro-laxmi, thanks for you PR.

Could you address the following points :

• code comments in the for loop shouldn't be removed
• io.StringIO wrapping of lines could be replaced by a lines.seek(os.SEEK_SET) if input file ends up being manually parsed the hard way (shouldn't occur much, and these files don't usually contain much data)
• only shlex.shlex call should be wrapped in the try..except ValueError block

Thanks, bye 👋

[pro-laxmi]

Hey @HorlogeSkynet, thanks for the review and suggestions!

I've updated the PR to address all three points:

• Restored the original code comments inside the for loop.
• Replaced the io.StringIO wrapper with lines.seek(os.SEEK_SET) on fallback (and removed the unused import io).
• Narrowed the try...except ValueError block strictly around the shlex.shlex token generation.

Tested locally with both valid and malformed os-release streams and everything is working smoothly. Let me know if there's anything else needed! 😊

[HorlogeSkynet]

Hey @pro-laxmi ! Much better, thanks for the update.

Some more comments :

• do you know a case/platform where lines.seek wouldn't be available ? From documentation, it seems TextIOBase.seek is cross-platform.

• do you mind adding a corresponding test case for this edge case ? This way code coverage would be kept.

Thanks, bye 👋

[pro-laxmi]

Hey @HorlogeSkynet! Thanks for following up.

You're totally right about seek—since lines is typed as TextIO and all standard text streams implement it cross-platform, I removed the hasattr check and simplified it to a direct lines.seek(0).

I also added a unit test case (test_kx_03_unclosed_quotes) in tests/test_distro.py covering unclosed quotes to preserve code coverage. All 282 tests pass cleanly.

Thanks again! 😊