[Differential] [Commented On] D4411: UnknownDeclarationProblem: scan headers on a background thread to preserve GUI responsiveness (WIP)

[René J.V. Bertin]

rjvbb added a comment.


  In https://phabricator.kde.org/D4411#82688, @mwolff wrote:
  
  > sorry, but nested eventloops are a no-go.
  >
  > the api has to be restructured to allow async assistants. until that is not done, shoehorning it in this way is a clear -2 from my side
  
  
  Is it worth anyone's time then that I try to deal with the other comments?
  Given the way the API is implemented in multiple places over 2 projects and the amount of work it will thus be to rewrite it I'm going to leave that to someone who knows the code much better. Or maybe it's a nice GSOC project (deadline for proposals coming up).
  Just be aware that the UI freezes are a real blocker; having to wait 2-3s (or more) *each* time you stumble upon an unknown variable is just not acceptable and can put people off very quickly.
  
  This was intended as an acceptable/reasonable workaround, which is also why I have changed as little as possible to the actual payload code.
  I just have 1 request: If this won't go in please don't break the patch (too much) before a real fix is committed, so this workaround remains available as an optional local patch for those who want to use it.
  
  Quick reactions to the comments:
  
  - the `interrupted` global var. is scheduled to go into the `UDPWorkerThread` class, hasn't been done yet because:
  - the #if 0 blocks are still there to make it easier to compare behaviour under old and new implementations.
  - UDP = Unknown Declaration Problem, no confusion intended (or expected...) with the networking protocol.

REPOSITORY
  R782 KDevelop Clang Integration

REVISION DETAIL
  https://phabricator.kde.org/D4411

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: rjvbb, #kdevelop, mwolff
Cc: mwolff, kdevelop-devel, Pilzschaf, akshaydeo, surgenight, arrowdodger
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kdevelop-devel/attachments/20170203/9c801dce/attachment.html>