Created attachment 1852395 [details] journal Description of problem: As shown in the attached screencast,gnome-software gets stuck every time when I try to search an app in gnome-software,and as shown in the attached journal, there are tons of "localhost-live gnome-software[3884]: gtk_widget_measure: assertion 'for_size >= -1' failed" in journal, 3 mins journal file size is more than 20M Version-Release number of selected component (if applicable): gnome-software-42~alpha-1.fc36.x86_64 How reproducible: always Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Created attachment 1852396 [details] screencast
Thanks for a bug report. This is tracked upstream as a gtk4 bug [1], thus I move this there. [1] https://gitlab.gnome.org/GNOME/gtk/-/issues/4517
This bug appears to have been reported against 'rawhide' during the Fedora 36 development cycle. Changing version to 36.
Still happens with gnome-software-42~beta-1.fc36.x86_64,and it seems get worse,Software hangs when I searched "hexc":as shown in the attached screenshot from f35,only hexchat match that string.I thought the hang will only happen when there are many results to be displayed.
Created attachment 1861382 [details] screenshot
Created attachment 1861383 [details] screencast
Proposed as a Blocker for 36-final by Fedora user lnie using the blocker tracking app because: This is a violation of the criteria: All applications that can be launched using the standard graphical mechanism after a default installation of Fedora Workstation on the x86_64 architecture must start successfully and withstand a basic functionality test.
The hang will be triggered when install some apps (for example, firefox or chrome)
Created attachment 1861386 [details] screencast
It's not about the amount of apps being found, but rather about the texts being shown for them. I've been playing with this yesterday and it resulted in a fill of a crash-variant of this bug, just by disabling wrapping for some labels. More is filled upstream at: https://gitlab.gnome.org/GNOME/gtk/-/issues/4700
I got the impression that "the hang will only happen when there are many results to be displayed" because as shown in the attached screencast, when I search"hexc" or "hex" it will not hang,but when I delete "x" ie,when the search string is "he",the hang will be triggered.I looked into the ticket you mentioned in #Comment 2,neither "testing" nor "esting" will trigger the hang for me.
Created attachment 1861434 [details] screencast
Just found out that when hexchat is not installed,the hang will not be trigger whenever you search "hexc" or "hex",but will be trigger when you search"he",however after you install hexchat,the hang will be triggered if you search string shorter than "hexch",
Created attachment 1861437 [details] before install
Created attachment 1861438 [details] after install
This needs to wait for the upstream bug resolution. Maybe there are workarounds, but they may or may not work (as can be seen within the upstream report).
I also opened a bug under gnome-software [1], whose purpose is to open a discussion about possible temporary workaround on the gnome-software side. [1] https://gitlab.gnome.org/GNOME/gnome-software/-/issues/1646
(In reply to Milan Crha from comment #17) > I also opened a bug under gnome-software [1], whose purpose is to open a > discussion about possible temporary workaround on the gnome-software side. > > [1] https://gitlab.gnome.org/GNOME/gnome-software/-/issues/1646 Even that discussion is not over yet, I added the temporary workaround from that upstream bug report into gnome-software-42~beta-2.fc36.
The decision to classify this bug as an "AcceptedBlocker (Final)" was made as it violates the following criterion: All applications that can be launched using the standard graphical mechanism after a default installation of Fedora Workstation on the x86_64 architecture must start successfully and withstand a basic functionality test. https://pagure.io/fedora-qa/blocker-review/issue/613
(In reply to Milan Crha from comment #18) > Even that discussion is not over yet, I added the temporary workaround from > that upstream bug report into gnome-software-42~beta-2.fc36. Lili, can you please confirm whether the issue is resolved with that build (or later)? Thanks.
Checked with gnome-software-42~rc-2.fc36.x86_64,didn't see this bug anymore.
That's stable, so we can close this.
(In reply to Adam Williamson from comment #22) > That's stable, so we can close this. I disagree. There is no fix for this on the gtk4 side (where this bug is filled for), the gnome-software contains only a workaround, which degrades the user experience. Any other application can face the same problem. Once the workaround is removed this one is back (or when they fix another bug [1]). There is a (significant) difference between a workaround and a real fix. I gave some details on the faulty gtk4 code upstream [2] four weeks ago, but no response from there so far. [1] https://gitlab.gnome.org/GNOME/gtk/-/issues/4714 [2] https://gitlab.gnome.org/GNOME/gtk/-/issues/4517#note_1397148
There's limited value to a downstream report if we don't have an actual concrete bug that Fedora users are hitting, since upstream doesn't really do anything much with downstream reports. But if you want to keep it open, we at least need to strip the blocker status.