Hi Raster,
three's a charm, right? - Looks good so far here, too. Despite my best efforts I could not produce another error. :) Thanks again for your quick help! Cheers, Florian "Carsten Haitzler" ras...@rasterman.com – September 7, 2021 1:10 AM > On Mon, 06 Sep 2021 08:23:12 +0000 Florian Schaefer <list...@netego.de> said: > > > hmmm trying again - for me with asan on i don't see an issue with desklock > now... ? i tried to change the code to avoid possibly multiple entries in the > array... > > > Hmm, you might not like it (neither do I BTW) but it crashed again on me > > right after restarting E a second time. > > > > pastebin.com/i1pMSR4d > > > > > > But still, its definitely more stable than before... ;) > > > > > > Sorry about that, > > > > Florian > > > > > > "Florian Schaefer" list...@netego.de – September 6, 2021 5:08 PM > > > "Carsten Haitzler" ras...@rasterman.com – September 6, 2021 4:51 PM > > > > On Mon, 06 Sep 2021 00:00:37 +0000 Florian Schaefer <list...@netego.de> > > > > said: > > > > > > > > fixed it just now. :) > > > > > > Hooray, thanks a bunch! :) > > > > > > With this I am now finally able to also activate/test the modified > > > procstats module. (Before, this bug here would crash the module dialog so > > > that I could not make any changes.) So far so good. Back to work it is and > > > let's see what else turns up. ;-) > > > > > > Cheers > > > Florian > > > > > > > > OK, so it seems that Raster already tried to address this problem. > > > > > Thank you very much! > > > > > > > > > > This morning first thing I installed the changed version. Now it > > > > > crashes in the newly created _e_util_cb_delayed_cancel() instead... > > > > > > > > > > > > > > > pastebin.com/mX6hU40g > > > > > > > > > > > > > > > (This is reliably triggered when I try to set the border style of a > > > > > window to borderless.) > > > > > > > > > > > > > > > Unfortunately my ASAN is currently refusing to give meaningful outputs > > > > > to a logging facility that I can see so I have only the gdb manual and > > > > > the E automatic backtrace (even though the latter one contains the > > > > > ASAN > > > > > information in a reaaaaaly difficult to read way). > > > > > > > > > > > > > > > Cheers, > > > > > > > > > > Florian > > > > > > > > > > > > > > > "Florian Schaefer" list...@netego.de – September 5, 2021 2:32 PM > > > > > > As I mentioned in my last message of the "E crash with Nvidia" > > > > > > thread, > > > > > > in using ASAN I could trigger quite a bit of additional segmentation > > > > > > faults. So I disabled the procstat module for now and went > > > > > > hunting. :-) > > > > > > > > > > > > On closer inspection it now seems to me that the faults are actually > > > > > > all due to an invalid call in e_object_del() of e_object.c. > > > > > > > > > > > > Here are the traces and ASAN outputs of two captured crashes: > > > > > > > > > > > > pastebin.com/3du7GYkS > > > > > > pastebin.com/9V1wCjnw > > > > > > > > > > > > They seem quite identical to me. > > > > > > > > > > > > I did not find a fully reproducible way to trigger this. It mostly > > > > > > happened for me right after a restart of E or when dismissing the > > > > > > initial error dialog (informing me that there was some error with > > > > > > PAM). > > > > > > > > > > > > Anyway, I hope that this is helpful. > > > > > > > > > > > > Cheers > > > > > > Florian _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
