D18041: Fixed START_AT jobs disregarded while evaluating, and fixed the update of their score.

[Eric Dejouhanet]

TallFurryMan added a comment.


  In D18041#388253 <https://phabricator.kde.org/D18041#388253>, @wreissenberger wrote:
  
  > I found some minor things while playing around, but nothing serious:
  >
  > - How does priority and fix startup time correlate? Currently, a job with a lower priority needs to have a earlier startup time. Otherwise, the other one will be marked invalid. To be discussed, I would expect that such a schedule is valid.
  > - A culmination determined job, that precedes but overlaps a fix startup time job, is set to aborted. I would expect that the fix startup time job is invalid.
  > - A job with fix startup time and duration set to a "repeat until" time that is earlier than the startup time leads to status "completed". Hm, formally correct, but I would consider this as an invalid configuration.
  
  
  Your remarks are indeed valid.
  Right now, priority will only affect two jobs being scheduled when the sort option is enabled. In the future, I'd like Scheduler to be able to alternate between multiple looping jobs.
  About the culmination being set to invalid instead of the fixed-startup job, I had to make a choice and chose the most general implementation, that is, fixed-startup retain priority.
  About the inverted startup/completion time, I ended up with this behavior when repeating a job. The only way to complete a fixed-completion job is to exceed the completion time. It happened the case you refer to was handled gracefully by the implementation of this rule, so I left it as is, nearly surprised that additional code would for once not induce a crippling side effect somewhere else :)

REPOSITORY
  R321 KStars

REVISION DETAIL
  https://phabricator.kde.org/D18041

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/20190107/8958a038/attachment.html>