Am 21.03.2016 um 01:10 schrieb Joel Kulesza <jkule...@gmail.com>: > > On Sun, Mar 20, 2016 at 4:48 PM, Scott Kostyshak <skost...@lyx.org> wrote: > > #9992 > It is a crash because of a screen management tool. Because it is due to > external software, I don't think we should consider this a rc1 blocker. > There is a regression from the user perspective, and it would be nice to > know whether different Qt versions lead to different behaviors. For > example, if the crash is still experienced with 2.2.0rc1 compiled with > Qt 4.8.6 (which I think is what 2.1.x was compiled with), then perhaps > we can figure out what the change in LyX's code was. > > > Please note that this it was not conclusively determined that this is due to > a screen management tool.
Yes, this was a guess only. I’m not so confident either. > Note also that this was on OSX using the newly implemented Retina support. I don’t think this matters. We had crashes with similar call stacks in the pre-HiDPI-past. IMO, things like that are caused by memory management or thread issues. The biggest problem is the difficulty to reproduce the bug on other systems. > Regardless, it cannot be reproduced to otherwise attribute it to something > else. If/when other Qt versions are built against, I would be happy to > perform testing. Regrettably, it seems I cannot build myself, but I would be > willing to test various installers. In case you’re able to spend some time I think we should find a way to put you on the track to build LyX yourself. It is an important thing to come to a "fail safe“ and redundant environment. Stephan
