1. 20 Sep, 2019 1 commit
  2. 18 Jun, 2019 1 commit
    • Fabian Vogt's avatar
      Don't give up if no results arrive after 500ms · c918c284
      Fabian Vogt authored
      Summary:
      Currently it resets the entire context if no results appear after 500ms,
      which means no new results can appear. This makes no sense.
      
      RunnerManager does not emit any matchChanged signal while the runners update
      the matches too quickly. The way this delay is implemented can result in never
      actually emitting the signal until all runners are done though.
      Thus it's actually likely to hit this situation.
      
      BUG: 389611
      
      Test Plan:
      Test results appear now, even if the first batch of runners is slow.
      List of matches still clears if nothing was found.
      
      Reviewers: #plasma
      
      Subscribers: broulik, plasma-devel
      
      Tags: #plasma
      
      Differential Revision: https://phabricator.kde.org/D21605
      c918c284
  3. 10 Jun, 2019 1 commit
    • Fabian Vogt's avatar
      Enforce use of the reset timer for clearing the model · 9d5921fb
      Fabian Vogt authored
      Summary:
      Currently when the query string changes, krunner sends a signal with no matches
      directly, which circumvents use of the reset timer.
      
      Test Plan:
      Enter =1+1, results are shown. Add another "1", the shown 2 switches to a 12
      instantly without hiding the previous result for a few frames.
      
      Reviewers: #plasma, mart
      
      Reviewed By: #plasma, mart
      
      Subscribers: plasma-devel
      
      Tags: #plasma
      
      Differential Revision: https://phabricator.kde.org/D21670
      9d5921fb
  4. 19 Sep, 2018 1 commit
  5. 14 May, 2018 1 commit
  6. 29 Jan, 2018 1 commit
  7. 10 Aug, 2016 1 commit
    • Kai Uwe Broulik's avatar
      [Milou] Optimize and cleanup sources model · 7bfa1a25
      Kai Uwe Broulik authored
      * Simplify queryString once, rather than during every iteration and use QStringRef where possible
      * Foreach instead of potentially detaching range-for
      * Use reverse iterators (new in Qt 5.6) instead of Java-style iterators
      * Remove traces of begin/endInsert/RemoveRows feature - while it would be nice to not reset
        the model, with all this sorting and merging going on it's pretty difficult to predict
        where an item will end up which is probably why it never got implemented.
        It still calculated tons of indices that went unused, so remove this and go a different
        route in the future
      
      Differential Revision: https://phabricator.kde.org/D2395
      7bfa1a25
  8. 17 Jun, 2016 1 commit
  9. 02 Mar, 2016 1 commit
  10. 11 Feb, 2016 1 commit
  11. 28 Dec, 2015 1 commit
    • Kai Uwe Broulik's avatar
      Support runner actions · f94c18a6
      Kai Uwe Broulik authored
      This brings back support for additional runner actions. A runner can specify
      supplemental actions to a search result.
      
      REVIEW: 126514
      f94c18a6
  12. 10 Mar, 2015 1 commit
    • Vishesh Handa's avatar
      SourcesModel: Expose a 'isDuplicate' role · 65372a29
      Vishesh Handa authored
      The delegate can use this 'isDuplicate' role to determine if the main
      text is duplicated and then optionally show the subtext.
      
      This categorized delegate uses this option while the classical list view
      always shows the subtext.
      65372a29
  13. 07 Jan, 2015 1 commit
  14. 30 Dec, 2014 1 commit
  15. 13 Dec, 2014 1 commit
  16. 28 Oct, 2014 1 commit
  17. 20 Oct, 2014 1 commit
  18. 20 Aug, 2014 2 commits
    • Vishesh Handa's avatar
      SourcesModel: Only consider each types first result for priority · 3e4bb6b1
      Vishesh Handa authored
      Each type is given an additional priority if any of its shown results
      has the string which the user searched. This is not perfect as often the
      top most result of that type doesn't have that string but is still given
      a higher priority. Arguably, the backends should also be improved.
      
      BUG: 338228
      3e4bb6b1
    • Vishesh Handa's avatar
      SourcesModel: Improve the sorting of the types · d1ecd115
      Vishesh Handa authored
      We used to just sort the results based on how we got them from the
      krunner framework. The runner framework doesn't do much. It mostly just
      gives the results as it gets them + some extra weight for results that
      have been run before + some weight based on the type of the result
      (Completion Result, Exact match, etc). This never resulted in
      Applications always being on the top unless the full application name
      was written.
      
      This patch improves this by comparing the user visible text. The type
      where the item's text contains the user visible string are given a
      higher priority.
      d1ecd115
  19. 03 Jul, 2014 1 commit
  20. 17 Jun, 2014 1 commit
    • Vishesh Handa's avatar
      Show the subtext in the case of duplicates · 700d4df8
      Vishesh Handa authored
      This way the user gets some extra context and does not think that the
      same result has been returned multiple times.
      
      Currently, there is a slight eliding problem which I cannot seem to
      figure out.
      700d4df8
  21. 12 Jun, 2014 1 commit
  22. 03 Jun, 2014 6 commits
  23. 30 May, 2014 1 commit
    • Vishesh Handa's avatar
      SourcesModel: Clear the model immediately · dec847de
      Vishesh Handa authored
      We wait 250 msecs, so that the results appear in one go instead of each
      signal being propogated as the results appear in a haphazzard fashion.
      This works well for normal queries, but it also results in us waiting an
      extra 250 msecs when the query is empty, and the data should be cleared.
      
      Since the views typically already have a delay when passing the query
      string, it makes sense to not wait.
      dec847de
  24. 08 May, 2014 1 commit
    • Vishesh Handa's avatar
      SourcesModel: Always keep the applications on the top · 9f6e0ce1
      Vishesh Handa authored
      This is a hack, but I cannot figure out a better way to try and keep
      Applications on top. They are supposed to be ordered based on the type
      (ExactMatch / PartialMatch / etc) and the speed, but it seems like Baloo
      is way faster than the Applications Runner.
      9f6e0ce1
  25. 06 May, 2014 1 commit
  26. 05 May, 2014 1 commit
    • Vishesh Handa's avatar
      SourcesModel: Re-order the types based on the "type" of the result · c2ea48d5
      Vishesh Handa authored
      In this case the type corresponds to the type set by each runner which
      can then be converted into an integer. The most commonly used types are
      "Completion Match" (30) or "Exact Match" (100).
      
      This is not perfect as applications are still not always on top, but it
      is a step in the right direction, given that we're trying our best to be
      all "generic".
      c2ea48d5
  27. 02 May, 2014 1 commit
  28. 25 Apr, 2014 2 commits
  29. 15 Apr, 2014 1 commit
  30. 14 Apr, 2014 2 commits
  31. 02 Apr, 2014 1 commit
  32. 10 Feb, 2014 1 commit