Hi Robert, hi Tomaz,
Am 30.08.2017 um 19:45 schrieb Robert Helling: > >> On 30. Aug 2017, at 10:18, Tomaz Canabrava <[email protected] >> <mailto:[email protected]>> wrote: >> >> Robert, >> >> What is still missing: >> A ui element to turn this expensive calculation off. >> >> >> This is easy to add, wouldn't take me much time. > > Yes, that is something even I can do, just haven’t had time so far. >> >> Finally, thanks to signals/slots flying in all directions, I do >> not understand when exactly the planner is called. I am under the >> impression it is called three times upon each change of a dive >> parameter in the UI. By understanding this, there is a possible >> 200% performance gain in sight. >> >> >> I need to look into this, can you tell me the entry point so I can debug? > > Just place a breakpoint on the beginning of plan() in planner.c. > Besides my new code and the tests this gets called in two places (and > even there it is not obvious why those two places are necessary). Then > enter the planner and move a waypoint around. For example using the > cursor keys (as then you have control over how often you change > things) and see how often the breakpoint gets hit. > I have a finding here which may be related to this discussion about long calculation times for Roberts new feature and the questions about plan() being called multiple times: Since Roberts change here on my side not only the calculation of the plan takes much longer but also this happens: If I increase or decrease any QSpinBox value in the planner preferences by clicking on the up or down arrow the QSpinBox value is increased/decreased by two steps instead of one. This happens with any QSpinBox based value like gas options, ascent rates, VPM conservatism,... I don't believe this really comes from Roberts change but it was maybe revealed by it. Can you reproduce this? Best regards Stefan -- Stefan Fuchs E-Mail: [email protected] <mailto:[email protected]>
_______________________________________________ subsurface mailing list [email protected] http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
