emacs should launch without the referenced warning and render properly without causing a segfault.
Actual behaviour
When starting emacs or emacs -Q under a Wayland compositor, specifically Sway, Emacs displays the following warning message:
(emacs:8670): GdkPixbuf-WARNING **: 13:38:44.606: Error loading XPM image loader: Image type “xpm” is not supported
And then proceeds to segfault (see backtrace emacs.gdb.txt for more information, but the culprit function should be a cairo_surface_create_similar_image() call, as mentioned in the upstream report) as shown by this message:
I spoke with the core maintainer and they asked that I verify that the XPM library is installed at the distribution-level. I have it installed:
~ $ xbps-query -Rs xpm
[*] libXpm-3.5.17_1 X PixMap Library from modular Xorg X11
What makes that warning out of place is that invoking the following code to check if the xpm image loader is supported actually returns t; thankfully the batch processor which doesn't require opening a client window and rendering a surface works (and so does emacs -nw of course).
Is this a new report?
Yes
System Info
Void 6.6.61_1 x86_64-musl GenuineIntel uptodate rrFFFF
Package(s) Affected
emacs-pgtk-29.4_2
Does a report exist for this bug with the project's home (upstream) and/or another distro?
https://lists.gnu.org/archive/html/bug-gnu-emacs/2024-11/msg00817.html
Expected behaviour
emacs
should launch without the referenced warning and render properly without causing a segfault.Actual behaviour
When starting
emacs
oremacs -Q
under a Wayland compositor, specifically Sway, Emacs displays the following warning message:And then proceeds to segfault (see backtrace emacs.gdb.txt for more information, but the culprit function should be a
cairo_surface_create_similar_image()
call, as mentioned in the upstream report) as shown by this message:I spoke with the core maintainer and they asked that I verify that the XPM library is installed at the distribution-level. I have it installed:
What makes that warning out of place is that invoking the following code to check if the
xpm
image loader is supported actually returnst
; thankfully the batch processor which doesn't require opening a client window and rendering a surface works (and so doesemacs -nw
of course).Eli, the core maintainer, suggested that this problem may be due to a GTK issue. What do you think?
Steps to reproduce
emacs-pgtk
emacs
oremacs -Q