D16429: [WIP] Rewrite of Scheduler planning

[Eric Dejouhanet]

TallFurryMan marked an inline comment as done.
TallFurryMan added inline comments.

INLINE COMMENTS

> scheduler.cpp:2279
> +    //ut = geo->LTtoUT(ltWhen);
> +    //KSNumbers ksnum(ut.djd()); // BUG: possibly LT.djd() != UT.djd() because of translation
> +    //LST = geo->GSTtoLST(ut.gst());

This will require deeper investigation. KStarsDateTime possibly has a bug with TimeSpec.
Julian Day must be calculated from Universal Time, which KStarsDateTime doesn't enforce in its constructor.

> mutlaqja wrote in scheduler.cpp:1432
> can such jobs be scheduled later than 24 hours later?

There is no temporal boundary anymore, jobs can be even be scheduled a few years in advance.
So far in the future obviously doesn't make any sense, but a schedule can be now be created spanning several weeks.
If during runtime the Scheduler is robust enough and the setup stable enough, this algorithm is a solid base for a robotic observatory.
I will allow customization of step duration, and add information about "lost" imaging time between jobs.

REPOSITORY
  R321 KStars

REVISION DETAIL
  https://phabricator.kde.org/D16429

To: TallFurryMan, mutlaqja, wreissenberger
Cc: yurchor, kde-edu, narvaez, apol
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-edu/attachments/20181103/06ccc792/attachment-0001.html>