Dear Maintainer,

After a recent reboot (following a lengthy period during which I did not
launch Lutris because it was already running, and I upgraded many
packages multiple times against Debian testing), lutris now fails to
launch on my system, which with a couple of potentially-important but
hopefully-irrelevant caveats (see below) is tracking Debian testing.

========
$ lutris
2026-02-07 23:01:09,704: Starting Lutris 0.5.19
2026-02-07 23:01:09,739: "card0" is AMD Radeon RX 5700 XT (1002:731f 1462:3810 amdgpu) Driver 25.3.3

(net.lutris.Lutris:8854): Gtk-WARNING **: 23:01:09.883: Could not load a pixbuf from icon theme.
This may indicate that pixbuf loaders or the mime database could not be found.
**
Gtk:ERROR:../../../gtk/gtkiconhelper.c:495:ensure_surface_for_gicon: assertion failed (error == NULL): Failed to load /usr/share/icons/Adwaita/scalable/status/image-missing.svg: XML parse error: Error domain 1 code 1 on line 1 column 14 of data: Parser input data memory error
 (rsvg-error-quark, 0)
Bail out! Gtk:ERROR:../../../gtk/gtkiconhelper.c:495:ensure_surface_for_gicon: assertion failed (error == NULL): Failed to load /usr/share/icons/Adwaita/scalable/status/image-missing.svg: XML parse error: Error domain 1 code 1 on line 1 column 14 of data: Parser input data memory error  (rsvg-error-quark, 0)
Aborted                    lutris
========

I have not managed to find any way to track this any further - not even
to identfiy what data it is that it's complaining about (unless it's
image-missing.svg itself). I have sometimes seen strace produce helpful
leads, but in this case I can't make head or tail of what it shows me. I
am not sure which library packages I would need to install debug symbols
from in order to have a chance of tracing this in a debugger.

I have, thus far, been unable to find any mentions of these error
messages online. I *have* seen some reports of similar-looking
Gtk-WARNING messages, which were tracked down to libgdk-pixbuf2 and I
think also to the gdk glycin transition, but for those the Gtk:ERROR
section was either different or missing entirely.

I have seen bug reports against other packages (#1128079, at
minimum) which suggest that there is currently some type of
(uncoordinate or under-coordinated?) transition going on which would
involve the libgdk-pixbuf packages; it seems possible that this might be
related to that, but I have no way of being certain. I appear to not yet
have installed the glycin-related packages which are involved in that
transition, but that does not necessarily guarantee anything.

Complicating the situation - and leading me to hesitate to even file
this report - is that after the same reboot, one of my other (and much
more day-to-day-essential) nearly-always-running programs started to
segfault on launch, and I wound up downgrading several library packages
from testing to stable in order to get that working again. I do not
think any of those libraries are plausible to be in code paths which
would make this plausible to be related to having older or mismatched
library versions installed, but I currently have no way to entirely rule
that out. The packages involved in that downgrade were primarily
related to fontconfig, with dependency-related branching out into pango.

I am filing this not so much as a request that a fix be provided
(although that would certainly be nice), as so that if someone else
encounters similar errors, there is a record to be found that they have
been seen previously.

I am filing it as 'important' because the issue makes the package
unusable for me, but I have no evidence to indicate that it would break
the package for everyone else, and I presume that if that were the case
there would already be reports about the matter from other people. I
defer to the maintainers' judgment as to whether this severity is
appropriate.

If there is anything I could do to help track this down, please let me
know; I'm not willing to do things that would risk breaking my other
daily-driver program (and I don't yet have available the sandbox
environment I'd need to do the experiments that would let me get out of
the need for that downgrade), but I am at least in principle willing to
try almost anything short of that.

Today I found some time and energy to investigate this further.

The details of how I got there don't seem worth going into, but the
short of it is that after

$ apt-get install libgdk-pixbuf2.0-{bin,common}/stable
libgdk-pixbuf-2.0-{dev,0{,:i386}}/stable
librsvg2-{2,bin,common,dev}/stable librsvg2-{2,common}:i386/stable

lutris now once again launches rather than presenting this error.
Without the librsvg2 downgrades, the error still manifested.


On the first launch after the downgrade, lutris seemed to be spamming
the console with

(net.lutris.Lutris:9818): Gtk-CRITICAL **: 09:10:57.396:
_gtk_css_image_get_concrete_size: assertion 'default_width > 0' failed

as the mouse moved over the window, although at first glance this
didn't seem to break anything else. After an exit and relaunch, these
messages don't seem to be persisting.


This isn't a real solution, but it does get me back into something
resembling operation, and it may point toward the cause and a potential
path toward a proper way to fix things.

I would like to find out whether upgrading libgdk-pixbuf2 again, but
leaving librsvg2 downgraded, makes the problem come back or not.
However, the current testing versions of the libgdk2-pixbuf packages
pull in the glycin transition, which adds enough new moving parts (and
has had enough icon-related issue reports elsewhere, some of them with
similar-seeming error messages) that I'm hesitant to risk installing it
without knowing what it's going to do and how I'd roll back if I needed
to.

If the test would be valuable enough, I can probably find the needed
libgdk-pixbuf2 package versions in my APT cache directory, although
there are enough of them that installing them that way would be a bit of
a pain.