1. 07 Dec, 2017 2 commits
    • Karel Koci's avatar
      Add system test replan-no-effect · 92f3cce4
      Karel Koci authored
      This tests that even if won't do any changes after replan that we will
      rin postupdate hooks.
      92f3cce4
    • Karel Koci's avatar
      Run postupdate hooks as part of cleanup when --replan is used · 554d9c7b
      Karel Koci authored
      There is situation when we are replanning when we end up with no new
      changes to be done and this way we exit early (going to cleanup). That
      is problematic as that means that we won't run postupdate hooks in such
      situation. This fixes it. When we go to cleanup and we have --replan we
      also run postupdate hooks.
      554d9c7b
  2. 28 Nov, 2017 5 commits
    • Karel Koci's avatar
      Make random sleep in updater.sh longer · 552ee0bf
      Karel Koci authored
      From half an hour to two hours. We are running updater every fourth hour
      so two hours should be good enough margin and should give us well
      enough distributed users.
      552ee0bf
    • Karel Koci's avatar
      Disable valgrind test, they are not passing because of slow server · 394e8d9a
      Karel Koci authored
      Gitlab runner is set to be running multiple parallel jobs but that means
      that everyone of those jobs takes too long. When we combine network and
      vagrind then those tests almost never pass unless there is almost no one
      at work. I just don't want to increase timeout more so just disable
      these tests on gitlab ci. Maybe try to re-enable them some time in
      future.
      394e8d9a
    • Karel Koci's avatar
      Drop restrict option of Script · 70aaae4f
      Karel Koci authored
      Restricting accessible uris for script at first look looks like an very
      good feature but in reality it cripples scripts ability and adds no
      additional security because if we use https and signatures then there
      should be no worry about script badness. And if it's then this is just
      small garden wall anyone exploiting https and signatures could go
      around just by using same https exploit again and again so this really
      doesn't make any sense.
      70aaae4f
    • Karel Koci's avatar
      Previous commit dropped support for early uri checks so don't test it · 49af043c
      Karel Koci authored
      Yes this allows us to pass sub-script a invalid uri but it doesn't
      matter if it won't use if and if it will then it will fail on its own.
      So just do that check when we need it not always.
      49af043c
    • Karel Koci's avatar
      Don't validate uri options when they are inserted to context · 63a7ae41
      Karel Koci authored
      It makes little to no sense to validate extra arguments when they are
      pushed to sub-script context as if they would be invalid the target
      script would already failed. This is just useless check and would add
      complexity as it makes table from any argument passed to it so this way
      it discards uri.system_cas and uri.no_crl and creates bugged system.
      63a7ae41
  3. 23 Nov, 2017 3 commits
  4. 07 Nov, 2017 1 commit
  5. 06 Nov, 2017 2 commits
    • Karel Koci's avatar
      Allow files access functions in Local restriction level · 1770357d
      Karel Koci authored
      There is no reason why we should deny access to these functions in local
      restriction level. Those scripts are running on local machine and has
      access to Run function so there is no sense in protection.
      1770357d
    • Karel Koci's avatar
      Fix l10n variable content · bf8746aa
      Karel Koci authored
      During cleanup the code loading content of l10n was moved to unrelated
      file. This moves it back and this way fixes languages support.
      bf8746aa
  6. 02 Nov, 2017 4 commits
  7. 01 Nov, 2017 2 commits
  8. 30 Oct, 2017 3 commits
    • Karel Koci's avatar
      Split all functionality off entry.lua and do some cleanups · 666f6814
      Karel Koci authored
      In next updater release will localrepo and opkg wrapper be in separate
      packages and because we now have conf.d directory we can easily add any
      additional configuration code trough it and package it in separate
      package. This makes much more clean entry script.
      
      Major changes with this is that auto.lua and user.lua are no longer
      loaded and we have to migrate them to /etc/updater/conf.d. So following
      move will be done:
      /etc/updater/user.lua -> /etc/updater/conf.d/user.lua
      /etc/updater/auto.lua -> /etc/updater/conf.d/opkg-auto.lua
      666f6814
    • Karel Koci's avatar
      Propagate root_dir to configuration · 2dd0fa41
      Karel Koci authored
      2dd0fa41
    • Karel Koci's avatar
      Allow DIE command to be called from configuration · 0b60f417
      Karel Koci authored
      There is no reason why configuration shouldn't be allowed to signal full
      updater failure. It's even highly desired. If configuration script
      detects unsolvable problem updater shouldn't be continuing.
      0b60f417
  9. 25 Oct, 2017 9 commits
  10. 10 Oct, 2017 2 commits
  11. 03 Oct, 2017 2 commits
  12. 26 Sep, 2017 1 commit
  13. 25 Sep, 2017 1 commit
  14. 22 Sep, 2017 1 commit
  15. 21 Sep, 2017 2 commits