D14684: Fixes for the Scheduler and Capture modules

Eric Dejouhanet noreply at phabricator.kde.org
Mon Aug 13 20:49:42 BST 2018

TallFurryMan added a comment.

  In D14684#307375 <https://phabricator.kde.org/D14684#307375>, @mutlaqja wrote:
  > 3. I stopped KStars on 2nd scheduler just before H_Alpha, so I expected the capture job to pick it up there but instead it started at Lum like last time.
  >   F6190091: log_09-48-38.txt <https://phabricator.kde.org/F6190091>
  I'll try to clarify the log, because it's not clear enough for the end-user: inside the running batch, the sequence has to provide 6xL, and there are 2 batches to run.
  Therefore to complete the L item of the sequence, the Capture module must take 6xL right now.
  Even if there are already, say, 4xL in storage, right now the Capture module has to take 6xL to complete the sequence, not 2xL (I'm not considering the LRGB one for the sake of clarity here).
  On the next batch, there will be 10xL in storage, thus the Capture module will have to take 2xL to comply with what the Scheduler requested, that is, two batches of 6xL.
  This behavior is my last change to the Scheduler/Capture pair, it actually fixes the last completion issue that was remaining.
  I must admit this last bit was quite hard to diagnose and fix properly: it's counter-intuitive to consider that only the last batch of capture does require a partial capture.
  Actually, that's quite logical: in order to complete the sequence optimally, each capture item must gather a maximum of frame types on the first run.
  But in fact, the root cause for this is that once the capture file is in the storage, the information about which sequence item provided it is lost.
  I've got two additional fixes being prepared (the image update int the scheduler list, and removal of the manual scheduled focus message) but I strongly suggest we land this differential as soon as possible as it's getting really large.
  @wreissenberger were you able to test on your side?

  R321 KStars


To: TallFurryMan, mutlaqja, wreissenberger
Cc: kde-edu, narvaez, apol
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-edu/attachments/20180813/02582254/attachment.html>

More information about the kde-edu mailing list