Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 922937 - mate-base/mate-applets-1.26.1-r2 /usr/libexec/mate-applets/stickynotes-applet fails to spawn sticky notes, then crashes
Summary: mate-base/mate-applets-1.26.1-r2 /usr/libexec/mate-applets/stickynotes-applet...
Status: CONFIRMED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: Current packages (show other bugs)
Hardware: All Linux
: Normal normal
Assignee: immolo
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-01-26 02:22 UTC by tacokoneko
Modified: 2024-02-04 11:31 UTC (History)
4 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments
stdout, stderr and backtrace of stickynotes-applet through (gdb-stickynotes-applet-bt.txt,15.18 KB, text/plain)
2024-01-26 02:22 UTC, tacokoneko
Details
emerge --info (emerge--info.txt,9.91 KB, text/plain)
2024-01-26 02:33 UTC, tacokoneko
Details

Note You need to log in before you can comment on or make changes to this bug.
Description tacokoneko 2024-01-26 02:22:43 UTC
Created attachment 883138 [details]
stdout, stderr and backtrace of stickynotes-applet through

How to reproduce the crash:
right click on a MATE panel->Add to Panel->Sticky Notes->Add->Right click Sticky Notes Icon->New Note->Double click Sticky Notes Icon

How to reproduce the backtrace:
1. compile mate-applets and libX11 with debug symbols and unstripped
2. open a shell and type this:
gdb /usr/libexec/mate-applets/stickynotes-applet
run
3. right click on a MATE panel->Add to Panel->Sticky Notes->Add->Right click Sticky Notes Icon->New Note->Double click Sticky Notes Icon
4. Type this in the shell:
bt

Naive attempt to reproduce the crash in another distro Debian 12 has failed so far
Comment 1 tacokoneko 2024-01-26 02:33:46 UTC
Created attachment 883139 [details]
emerge --info
Comment 2 Sam James archtester Gentoo Infrastructure gentoo-dev Security 2024-01-26 09:18:41 UTC
Note that the window is a null pointer:

#0  0x00007ffff6e9744e in XChangeProperty
    (dpy=dpy@entry=0x0, w=0, property=property@entry=0, type=type@entry=6, format=format@entry=32, mode=mode@entry=0, data=0x7fffffffd140 "\233\002", nelements=4) at /var/tmp/portage/x11-libs/libX11-1.8.7/work/libX11-1.8.7/src/ChProp.c:46
#1  0x000055555555bba4 in set_icon_geometry
    (height=<optimized out>, width=<optimized out>, y=<optimized out>, x=<optimized out>, window=0x0) at stickynotes.c:63
Comment 3 tacokoneko 2024-01-26 14:43:26 UTC
I tested MATE on Debian, Ubuntu, Fedora, Arch Linux and Alpine Linux and among those I have been able to reproduce this only on Alpine Linux. Here are some more issues I opened

https://gitlab.alpinelinux.org/alpine/aports/-/issues/15719
https://github.com/mate-desktop/mate-applets/issues/664
Comment 4 Thibaud CANALE 2024-02-02 06:21:07 UTC
Hello, I also noticed this issue, and I remembered it happened (at least on my system) after updating to the current x11-base/xorg-server, version to 21.1.11 from 21.1.10:
```
% qlop --date '2024-01-01' --date '2024-02-01' --verbose dev-libs/glib x11-libs/libX11 x11-base/xorg-server mate-base/mate-applets
2024-01-19T17:32:36 >>> x11-base/xorg-server-21.1.11: 32s
2024-01-19T17:34:10 >>> mate-base/mate-applets-1.26.1-r1: 29s
2024-01-25T20:37:20 >>> mate-base/mate-applets-1.26.1-r2: 30s
```

I recompiled with debug symbols, and I also meet the same output, the window parameter is a NULL pointer.
However, from a feature point of view, my sticky notes are not visible at beginning, which was not the behavior before, the SEGFAULT only happens when the program tries to hide them. This last is processed by the function `stickynote_set_visible` when `visible` parameter is set to false.
Maybe it explains why the program does SEGFAULT when trying to hide all those notes; hiding can be triggered by simply clicking on the desktop background.

I think we should focus our attention on all those assertion failures reported in tacokoneko’s attachment 883138 [details], because it can explain why the window value is NULL, definitively the SEGFAULT is only the symptom of a wrong doing happening before.
I don’t actually think this is related to the temporary x11-libs/gtksourceview switch to slot 4, and for sure the last revision `mate-base/mate-applets-1.26.1-r2` set back the correct slot, as I can see in the build.log such as `-I/usr/include/gtksourceview-3.0` or `-lgtksourceview-3.0` parameters during compilation.

Meanwhile I put here the link to the commits list between Xorg server 21.1.10 and 21.1.11 releases:
https://gitlab.freedesktop.org/xorg/xserver/-/compare/xorg-server-21.1.10...xorg-server-21.1.11

I have no knowledge in GtkSourceView, I will try to dig more later.
Comment 5 tacokoneko 2024-02-04 11:31:38 UTC
(In reply to Thibaud CANALE from comment #4)
> Hello, I also noticed this issue, and I remembered it happened (at least on
> my system) after updating to the current x11-base/xorg-server, version to
> 21.1.11 from 21.1.10:
> ```
> % qlop --date '2024-01-01' --date '2024-02-01' --verbose dev-libs/glib
> x11-libs/libX11 x11-base/xorg-server mate-base/mate-applets
> 2024-01-19T17:32:36 >>> x11-base/xorg-server-21.1.11: 32s
> 2024-01-19T17:34:10 >>> mate-base/mate-applets-1.26.1-r1: 29s
> 2024-01-25T20:37:20 >>> mate-base/mate-applets-1.26.1-r2: 30s
> ```

> Meanwhile I put here the link to the commits list between Xorg server 21.1.10 > and 21.1.11 releases:
> https://gitlab.freedesktop.org/xorg/xserver/-/compare/xorg-server-21.1.10.> ..xorg-server-21.1.11

It's a good idea, but when I installed this ebuild https://github.com/gentoo/gentoo/blob/d2c6c988cb34001f6d0ce5931c54aed7be494ea4/x11-base/xorg-server/xorg-server-21.1.10-r1.ebuild and rebooted and verified Xorg.0.log displayed "1.21.1.10", the same sticky note problem still happens, so it seems to me that there are more factors involved than just that version change.