GNOME Bugzilla – Bug 776977
The headerbar button action are sometimes shifted
Last modified: 2018-05-22 16:58:40 UTC
Hello, Evince can't be closed with the X sometimes. It's like the border of the window is not correctly detected as a resize cursor appear when hovering the X button on the headerbar. I think it happen only when maximized. If you unmaximize / resize the window, the X button work again. Fedora 24 & 25 Evince 3.20 & 3.22 Thanks !
Happen with arc and adwaita themes
In fact it's all the buttons in the toolbar that don't work unless you resize the window.
It seem that the buttons in the headerbar works but their action are shifted !
Does it happen with any other application? If any, I think it would be a bug in GtkHeaderBar itself, and therefore the bug should belong to GTK+.
Or maybe it is related to Header bars densely populated. In such case, this would be related to Bug 732289
(In reply to Germán Poo-Caamaño from comment #4) > Does it happen with any other application? If any, I think it would be a bug > in GtkHeaderBar itself, and therefore the bug should belong to GTK+. I only see this in Evince. (In reply to Germán Poo-Caamaño from comment #5) > Or maybe it is related to Header bars densely populated. In such case, this > would be related to Bug 732289 I don't think it's that the problem.
Created attachment 351853 [details] border action shifted Here you can see that the border action is shifted
Created attachment 351854 [details] close action shifted Here you can see the close action that is shifted
I hope I'm clear enough
Can you reproduce the issue with any PDF document or some documents? I am just speculating what it could be. I have seen the issue with the close button a couple of times in the past, but I cannot reproduce the issue reliably. For example, now I tried 9 different documents, and I could not reproduce the issue, it worked fine on maximized, left pane, and normal windows.
I think it happen with every pdf but not always. I mean, with the same document, sometime it happen and sometime it don't.
It happen very often for me and it happen on all my computers. You really can't reproduce the bug ?
jeremy9856: Please do not comment if there is nothing new to add. Read comment 10. Is your question serious? Why don't you believe what people wrote?
André: Please do not comment if there is nothing new to add. Read comment 12. Is your question serious? Why don't you believe what I wrote? It happen very often for me and it happen on all my computers. Did you try to reproduce the bug before commenting?
With honey with attract bees, not with vinegar. Please Jeremy, stop the rudeness. As I said before, I have seen the issue, but I am unable to reproduce it reliably. If you can give a group of clear steps, with a sample document, and setting on how to reproduce the issue reliably, then please add it explicitly. By reliable I mean, that you can always reproduce the issue following those steps. The first step to fix a bug is to be able to reproduce it. Before that happens, there is nothing we can do about it. Of course I would like to see the issue fixed, but I understand the limitations to do so.
(In reply to Germán Poo-Caamaño from comment #15) > With honey with attract bees, not with vinegar. Please Jeremy, stop the > rudeness. Well it's André that started to be rude but it's me that need to stop ? Whatever I will focus on the bug. > As I said before, I have seen the issue, but I am unable to reproduce it > reliably. That's why I asked to you if it really doesn't happen if you test again. Maybe you can trigger it now (after some update since then) ? > If you can give a group of clear steps, with a sample document, and setting > on how to reproduce the issue reliably, then please add it explicitly. By > reliable I mean, that you can always reproduce the issue following those > steps. Well I just open pdfs with evince. I will try to see if I can found some settings that reliably trigger it. > The first step to fix a bug is to be able to reproduce it. Before that > happens, there is nothing we can do about it. Of course I would like to see > the issue fixed, but I understand the limitations to do so. I understand that and I really want to help. Maybe I don't do it the right way and for that I'm sorry. Thanks !
(In reply to jeremy9856 from comment #16) > (In reply to Germán Poo-Caamaño from comment #15) > > With honey with attract bees, not with vinegar. Please Jeremy, stop the > > rudeness. > > Well it's André that started to be rude but it's me that need to stop ? > Whatever I will focus on the bug. André is one of the bugmasters (people in charge of GNOME's Bugzilla). In his profile, he has 31 points, which means is one of the developers with more experience doing triage. In the best interest of being effective with everyone's time, it is best to add comments only when there is new information that help reproduce the issue. Otherwise, it may be rude. > > As I said before, I have seen the issue, but I am unable to reproduce it > > reliably. > > That's why I asked to you if it really doesn't happen if you test again. > Maybe you can trigger it now (after some update since then) ? I use Evince from master, which I build every other week or so. It is the most recent version. I see the issue once in a while, and when I try to reproduce it, I cannot. > > If you can give a group of clear steps, with a sample document, and setting > > on how to reproduce the issue reliably, then please add it explicitly. By > > reliable I mean, that you can always reproduce the issue following those > > steps. > > Well I just open pdfs with evince. I will try to see if I can found some > settings that reliably trigger it. Yes, please. > > The first step to fix a bug is to be able to reproduce it. Before that > > happens, there is nothing we can do about it. Of course I would like to see > > the issue fixed, but I understand the limitations to do so. > > I understand that and I really want to help. Maybe I don't do it the right > way and for that I'm sorry. Thanks ! Fair enough.
(In reply to jeremy9856 from comment #16) > I understand that and I really want to help. @jeremy9856: Then please follow comment 13 and do not periodically mass-comment on most of your tickets without adding any content that would bring a task closer to resolution, otherwise your account might get disabled at some point. Thanks for your understanding.
Well I switch to Ubuntu 17.10 and the bug was present. I tried from the live session and the problem was here too. So you should be able to reproduce it very easily. It happen on 3 computers with intel igpu. Note that on an other computer with an nvidia gpu it doesn't seem to occur with nouveau or proprietary drivers.
Created attachment 367001 [details] pdf test
I have encountered this bug too, but only in the Gnome Xorg session and not in the Wayland session.
Recently I encountered this bug while running a modified Evince that logs "motion" events. As you can see in the attached screenshots, when this bug arises the events have the wrong coordinates; ordinarily the close button has y coordinate approx 50 and x coordinate approx 950 for a half-wide window. In the last screenshot, right-clicks in the Evince window while sufficiently close to the right border seem to be interpreted as clicks on the desktop.
Created attachment 369311 [details] Wrong event coordinates 1
Created attachment 369312 [details] Right clicking in window brings up desktop context menu
Created attachment 369314 [details] Wrong event coordinates 2 (notice the jump in x)
Created attachment 369315 [details] right click passthrough 2 More generally, events near the right border are passed to windows underneath. Since I haven't seen this bug in the wayland session, maybe this is an X11 bug?
(In reply to Casey Jao from comment #26) > Created attachment 369315 [details] > right click passthrough 2 > > More generally, events near the right border are passed to windows > underneath. Since I haven't seen this bug in the wayland session, maybe this > is an X11 bug? or something in between (in the stack). But, why only Evince? may you attach the 'motion' patch you are using to debug the issue?
Created attachment 369515 [details] [review] Hack to print motion events
Thanks for the patch. I have been trying to reproduce the issue for around 30 minutes, unsuccessfully. I was expecting to use Gtk+ inspector (shift+ctrl+I) to interrogate the window. I can reproduce it when I am doing something else, usually when I do not have time to debug it. Damn bug! ;-)
-- GitLab Migration Automatic Message -- This bug has been migrated to GNOME's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/evince/issues/743.