Hi All, Am 20.10.2017 um 16:11 schrieb Guillaume Gardet: > >>>>> That's ugly. I found s.th. similar this morning also when deleting >>>>> a cylinder but only in the merged dive I played with. >>>>> Could come from my changes. I found already one bug and fixed it >>>>> (see PRs) but that was not all - crash still there. >>>> >>>> Start Subsurface >>>> >>>> Select file -> New logbook >>>> >>>> Select Log -> Plan dive >>>> >>>> Mine starts with a single D12 AIR twinset being defined and a 20 >>>> min dive at 15m. >>>> >>>> Add an AL80 cylinder with EAN50 >>>> >>>> In nth Dive points table change the dive to 60 minutes at 20m >>>> (results in awful profile) >>>> >>>> Delete EAN50 cylinder >>>> >>>> Crash >>>> >>> Odd. Usually I'm the one seeing all the crashes with my MXE Windows >>> build. But not this one. >>> Any new findings? >>> Can you provide a screenshot of the full planner window with the setup? >>> >>> Meanwhile I will try another build flavor and then also Linux. >> Ok, I have it in front of me now: >> Seems to be nothing were we or I did really change s.th. during the >> last days. But for sure what I did is that I implemented a few more >> triggers for updates between the cylinder table and everything else >> in the planner. That may have revealed an issue. >> >> BT attached below. >> >> From my limited experience: >> Is this s.th. coming from a race condition? >> We do check in planner .c update_cylinder_pressure for "if (!cyl)", >> this doesn't "fail" and then we crash a few lines later doing >> "cyl->gas_used.mliter += gas_used.mliter;" >>
Could you please try: https://github.com/Subsurface-divelog/subsurface/pull/704 -- Stefan Fuchs E-Mail: [email protected] <mailto:[email protected]>
_______________________________________________ subsurface mailing list [email protected] http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
