On Tue, 28 Jan 2025 13:53:47 +0100 daniel antoine <univa...@gmail.com> said:
> Hi raster > > I am sorry, I have compiled with the last commit for the cpufreq but the > problem of the hidden window was already gone. I have tried to return back > to find the problem again but without success. > > But I have found the problem of the guru error and I can reproduce it. I > open many terminology windows. In the window terminology list in the dock I > close them with a two finger stroke (I am on a laptop using the touchpad) > by accessing close in the terminology menu window > The problem is that the little square referring to the terminology opened > windows are black grey on a black background. When you close one the > square turns also black but the name of the window remains. As I have a bad > view I click again on the windows I have already closed and get the guru > error.* I try to close the windows that are already closed.* hmm this doesn't happen to me - if i close via any of these: ctrl+d (exits shell) ctl+alt+x (asks wm to close window) right click in terminology for control panel then press close the window disappears from e - from pagers, from the ibar window popup. from the alt-tab window list ... it's gone. pointer focus. click to focus - it's working fine... :( the little if (x & y &z) stack_above(); is a really tiny bit of code added and i can't see how that would cause the window to stay around pointing to junk memory that would be what is causing your problem i guess. i made 2 commits. one that ensures if focusing a new window - the previous focused window is below the new one. this was a little wrong -i should only do this if the unfocused window was forcibly un-fullscreened so i added that to the if case. it's still just a if () stack above(); ... no adding references or deletion etc. etc. so your problem feels orthogonal. but i don't know what/how - i cant reproduce the problem :( > Now this is not part of the problem. I have tried to disable modules during > the efl and enlightenment compilation package by putting them at false ( > build-test for efl ,convertible everything packagekit task vkbd pants > .....for enlightenment) but it was not a good idea . By exemple putting > recognition-gesture at false causes a libmodul error and the loss of all > the management of mouse ,keyboard .....(intrance). Modules look intricate > between them. > Finally I have compiled with the original meson_options.txt for efl and > enlightenment and just activate the modules I need > It's now enlightenment with only X11 ( no wayland) i think the right thing to do here is to try build efl and e with ASAN support and then get the asan log output in ~/.e-log.log when it crashes and see what it says. if i could produce the bug under asan i'd know what to look at - i can't :( > Best regards > > Daniel > > > > > > > > > Le lun. 27 janv. 2025 à 20:41, Carsten Haitzler <ras...@rasterman.com> a > écrit : > > > On Mon, 27 Jan 2025 18:44:56 +0000 Carsten Haitzler <ras...@rasterman.com> > > said: > > > > > On Mon, 27 Jan 2025 16:14:02 +0100 daniel antoine <univa...@gmail.com> > > said: > > > > > > > Hi > > > > > > > > commit 96b08edf36ac4118280fe11800ea0e1f75b8859e When you have two > > > > terminology windows when you maximize one and you want to access the > > > > second window by the terminology window list on the systray bar the > > > > second window remains hidden behind the maximized windows and don't > > > > appears on the screen. You have to minimize the fullscreen windows to > > > > access the second terminology window. > > > > > > > > commit bd7b2d44b273a67d1561ca6e9ffdb2ab70aa8b0d The problem with the > > > > precedent commit still exist but if you open many terminology windows > > > > when you want to close them from the terminology window list you get a > > > > guru error after closing two or three windows > > > > > > I tried this and didn't get any... hmmmmm > > > > i enabled asan (a compile-time memory checker that builds efl and e with > > full > > pointer checking - like valgrind) and i maximize windows and have a bunch > > of > > others on the same screen - close a bunch select them in ibar's window > > selector > > again and again - no complaints from asan. > > > > i found a different bug though :) but it was with cpufreq. > > > > > > Best regards > > > > > > > > The problem have disappeared after compiling without the two last > > commits > > > > > > > > Daniel > > > > > > > > > > > > _______________________________________________ > > > > enlightenment-users mailing list > > > > enlightenment-users@lists.sourceforge.net > > > > https://lists.sourceforge.net/lists/listinfo/enlightenment-users > > > > > > > > > > > > > -- > > > ------------- Codito, ergo sum - "I code, therefore I am" -------------- > > > Carsten Haitzler - ras...@rasterman.com > > > > > > > > > > > > _______________________________________________ > > > enlightenment-users mailing list > > > enlightenment-users@lists.sourceforge.net > > > https://lists.sourceforge.net/lists/listinfo/enlightenment-users > > > > > > > > > -- > > ------------- Codito, ergo sum - "I code, therefore I am" -------------- > > Carsten Haitzler - ras...@rasterman.com > > > > > > > > _______________________________________________ > > enlightenment-users mailing list > > enlightenment-users@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/enlightenment-users > > > > _______________________________________________ > enlightenment-users mailing list > enlightenment-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-users -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- Carsten Haitzler - ras...@rasterman.com _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
