1. 08 Feb, 2019 1 commit
  2. 07 Feb, 2019 4 commits
  3. 07 Jan, 2019 2 commits
  4. 05 Jan, 2019 4 commits
  5. 20 Dec, 2018 1 commit
  6. 18 Dec, 2018 1 commit
    • Antonio Terceiro's avatar
      config: fix setting debci_data_retention_days · 47c0f37c
      Antonio Terceiro authored
      The defaults are erroneously defined *after* configuration files are
      loaded, so forgetting to check an existing value before settings causes
      settings to not be configurable. In the long run we want to rewrite that
      part to set the default before loading the config files, but this needs
      to be fixed right now.
      47c0f37c
  7. 22 Nov, 2018 1 commit
  8. 04 Nov, 2018 1 commit
  9. 03 Nov, 2018 6 commits
  10. 27 Oct, 2018 1 commit
  11. 23 Oct, 2018 1 commit
  12. 23 Sep, 2018 1 commit
    • Antonio Terceiro's avatar
      lxc/create-testbed: avoid cleaning up valid containers · 2bee4565
      Antonio Terceiro authored
      Sometimes, the creation of a new testbed fails for various reasons (some
      essential package) breaks debootstrap, or something fails to download.
      Those failures are already handled by autopkgtest-build-lxc: the old
      container is only replaced by the new one in case of success.  In that
      case, it's OK for debci to just keep using the older container since it
      will be upgraded before running any tests anyway.
      
      The only case where we indeed want to remove containers is when the
      customizations needed by debci are not successfully applied.
      2bee4565
  13. 13 Sep, 2018 3 commits
  14. 12 Sep, 2018 1 commit
  15. 11 Sep, 2018 2 commits
  16. 09 Sep, 2018 3 commits
  17. 08 Sep, 2018 3 commits
  18. 10 Aug, 2018 3 commits
  19. 07 Aug, 2018 1 commit
    • Antonio Terceiro's avatar
      Escape triggers when enqueuing test jobs · 544586a0
      Antonio Terceiro authored
      Trigger with spaces (or worse, with shell hackery) were being passed to
      workers as is, and then used to compose a shell command line.
      
      Instead, when enqueing the job we CGI-escape them so only ascii
      characters remain untouched. Then, when reading the trigger back from
      the data file, we unescape it.
      
      Closes: #902337
      544586a0