RPMlint at its current state is quite big hog doing various tasks while being very slow.

Plan is to talk to upstream developer and figure out multiple things:

  • how to limit the features and make all the possible things to error while avoiding any warnings (Either something is bad or we do not care, do not mess with warns that everyone ignores)
  • how to paralelize execution to avoid the time hog
  • how to merge in our brp checks keeping stuff only in one place
  • discuss how to merge suse rpmlint checks to upstream ones
  • discuss SUSE specific patches for merging In this HW I at least plan to start with this as it is quite huge.

Looking for mad skills in:

Nothing? Add some keywords!

This project is part of:

Hack Week 17 Hack Week 18

Activity

  • 12 days ago: osukup liked RPMlint cleanups
  • 12 days ago: osukup joined RPMlint cleanups
  • 11 months ago: mkubecek liked RPMlint cleanups
  • 12 months ago: vitezslav_cizek liked RPMlint cleanups
  • 12 months ago: Pharaoh_Atem liked RPMlint cleanups
  • Show History

    Comments

    • simotek
      12 months ago by simotek | Reply

      https://rpmlint.opensuse.org/rules/openSUSE:Factory/x86_64/standard is a useful resource for getting started in terms of warnings we care about the least.

      There are certain checks like the incorrect-fsf-address that maybe worth keeping as info, ie the release team won't enforce and doesn't care about them, but if a packager see's it they might choose to report it upstream.

    • simotek
      12 months ago by simotek | Reply

      Another enhancement would be to make the non-executable-script slightly smarter, currently it just checks for files starting with #!/XXXXX, but there are some languages such as python and ruby (I think) where files may start with #! but don't need to be marked as executable to work (think python modules / libraries) so this test currently generates a lot of false positives.

    • scarabeus_iv
      11 months ago by scarabeus_iv | Reply

      Atm merging SUSE patches from develadd-emojiFactory:rpmlint/rpmlint to the main code https://github.com/scarabeusiv/rpmlint/tree/tests

    • scarabeus_iv
      11 months ago by scarabeus_iv | Reply

      Actually upstream is against python3 codebase, as such it does not make sense to try to paralelize the thing as it is pretty ugly to make it work on python 2.7 as such I will have to find something else to do on hackweek.

    • scarabeus_iv
      12 days ago by scarabeus_iv | Reply

      In last years run we switched to modularized rpmlint but it is still way off from 2.0 release that could be trusted. Thus there is milestone in upstream git and we will work in HW18 to get it in better shape again.

    Similar Projects

    This project is one of its kind!