D12485: Scheduler adjustments and fixes

[Eric Dejouhanet]

TallFurryMan added a comment.


  Well, I think I need an advice: scheduled jobs need a start time, but if multiple jobs start at the same time, the algorithm adding an offset to the start time will trigger.
  Before D12485 <https://phabricator.kde.org/D12485>, jobs scheduled asap would not get a start time, so the algorithm adding the offset would manipulate invalid dates, and in the end be ineffective.
  So incidentally, by scheduling asap jobs to start "now", I properly have the algorithm adding the offset to valid dates, but this causes a regression because this offset now delays the startup of the jobs, which were contiguous before.
  However, this situation is probably due to the incorrect estimation of the duration of the job, which is used as offset. Well, actually, that offset is clamped to the lead time anyway, so jobs are separated by 5min minimum when they go through that part.
  So all in all, for this specific point, the previous behavior was somehow better than what the D12485 <https://phabricator.kde.org/D12485> brings, so I'll probably revert if I don't find a more elegant solution, which might be to drop the algorithm adding offsets.

REPOSITORY
  R321 KStars

REVISION DETAIL
  https://phabricator.kde.org/D12485

To: TallFurryMan, mutlaqja
Cc: #kde_edu, narvaez, apol
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-edu/attachments/20180424/cad4b727/attachment.html>