Bug 157021 - Crash when in Sidebar click on Styles deck (gtk3)
Summary: Crash when in Sidebar click on Styles deck (gtk3)
Status: NEW
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
24.2.0.0 alpha0+
Hardware: All All
: high critical
Assignee: Not Assigned
URL:
Whiteboard:
Keywords: bibisected, bisected, haveBacktrace
Depends on:
Blocks: GTK3 Sidebar-Styles
  Show dependency treegraph
 
Reported: 2023-08-30 18:29 UTC by Timur
Modified: 2023-09-15 10:55 UTC (History)
6 users (show)

See Also:
Crash report or crash signature:


Attachments
bt (19.16 KB, text/plain)
2023-08-31 06:56 UTC, Julien Nabet
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Timur 2023-08-30 18:29:52 UTC
Open Writer new document
Open Sidebar in Navigator.
Click Styles deck. 
Crash 

Version: 24.2.0.0.alpha0+ (X86_64) / LibreOffice Community
Build ID: 7379decd74f9f66011f0023d8903430293dc0532
CPU threads: 16; OS: Linux 5.15; UI render: default; VCL: gtk3
Locale: en-US (en_US.UTF-8); UI: en-US
Calc: threaded

Mentioned in 156978 but different one.
Comment 1 BogdanB 2023-08-30 18:34:34 UTC
Confirm with in
Version: 24.2.0.0.alpha0+ (X86_64) / LibreOffice Community
Build ID: cf1cdc00e1e2d2684cfe57ac002a37c5f3d100c5
CPU threads: 16; OS: Linux 6.2; UI render: default; VCL: gtk3
Locale: ro-RO (ro_RO.UTF-8); UI: en-US
Calc: threaded
Comment 2 Timur 2023-08-30 18:49:24 UTC
commit 6fd755fb36472938757b2581cbe99f5e5fe1ae40	[log]
author	Noel Grandin <noel.grandin@collabora.co.uk>	Tue Aug 22 14:44:38 2023

tdf#100894 lots of Conditional formatting freeze calc Styles sidebar

cc Noel please see
Comment 3 Julien Nabet 2023-08-31 06:56:44 UTC
Created attachment 189273 [details]
bt

On pc Debian x86-64 with master sources updated today + gtk3 rendering, I got an assertion.
Comment 4 Noel Grandin 2023-08-31 12:10:09 UTC
This was working with an older version of gtk, but now it crashes on my machine too, deep inside gtk -> glib -> cairo

The assert that triggers seems to indicate that something is double-releasing a cairo surface, which means we might be tickling a gtk bug.

Caolan, any ideas?
Comment 5 Caolán McNamara 2023-08-31 18:58:48 UTC
I don't know if there is a v. quick fix. I think I suggested using set_text and some explicit column number in the first place to avoid a vcl edge case, but what I think is happening here is that the gtk treeviews, for the other edge case of a custom rendering preview, have a pixmap/cairo-surface at the expected index so the two edge cases show a bit of contradictory poor implementation by me there.

I probably need to do something like remove the pixmap/cairo-surface for custom rendering out of the .ui file, so the column mapping stuff we do works as it normally does for these treeviews in gtk, and bung it at the end at run-time (or something of that nature) when togged into the custom-render more.

Might be easiest to revert for now and if I get a chance then sort it out with a bit of thought.
Comment 6 Timur 2023-09-14 17:50:38 UTC
Caolan, can you please revert? This makes master unusable for testing in GTK, and switching to other is tedious. Also, later bibisects will be trubled.
Comment 7 Caolán McNamara 2023-09-14 19:29:31 UTC
It was reverted as far as I know with https://cgit.freedesktop.org/libreoffice/core/commit/?id=ff66803ec9ec26c918ef652a1478ff9d7da0c97b leaving this as a to-do in order to get it reapplied
Comment 8 Timur 2023-09-15 10:55:44 UTC
Thanks, I did not look, somehow expected a notification here. OK now.