Arch Linux

Also happened to me, had to downgrade to shared-mime-info-2.3-1.
I encountered this with Dolphin not starting up, stuck in a loop with a fast infinite memory leak. Better reproduction was Thunar crashing every time I opened a directory which contained a JS file, or a GTK file picker crashing any app using it (say of VS Code) on visiting such directory.
Simplest reproduction: `xdg-mime query filetype 0test/1.js`
The JS file has to contain a function declaration or something, empty file doesn't cause issues.

I troubleshooted this on Matrix with users @olly1240 (who tried to reproduce it on a fresh new VM install and couldn't so probably not an upstream issue) and @post-factum.

Stack trace:
#0 0x00007fcb8b754c99 in cache_alias_lookup (alias=alias@entry=0x7fcb87c9b58c "application/ecmascript") at ../glib/gio/xdgmime/xdgmimecache.c:308
#1 0x00007fcb8b7550a2 in __gio_xdg_cache_unalias_mime_type (mime=0x7fcb87c9b58c "application/ecmascript") at ../glib/gio/xdgmime/xdgmimecache.c:988
#2 __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb87c9b58c "application/ecmascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:911
#3 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb879db46c "text/javascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#4 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb87c9b58c "application/ecmascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#5 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb879db46c "text/javascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#6 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb87c9b58c "application/ecmascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#7 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb879db46c "text/javascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#8 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb87c9b58c "application/ecmascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#9 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb879db46c "text/javascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971
#10 0x00007fcb8b7552ff in __gio_xdg_cache_mime_type_subclass (mime=mime@entry=0x7fcb87c9b58c "application/ecmascript", base=base@entry=0x7fcb87c98e48 "text/x-matlab")
at ../glib/gio/xdgmime/xdgmimecache.c:971

Have you tried applying [1], which is referenced in [2]?
Edit:
I suspect a similar change is needed in glib to guard against infinite recursion in _xdg_mime_cache_mime_type_subclass [3].

[1]: https://code.qt.io/cgit/qt/qtbase.git/commit/?id=54656da9ace06caf4a0eeb1832989c0ab211a4a0
[2]: https://gitlab.freedesktop.org/xdg/shared-mime-info/-/merge_requests/258
[3]: https://gitlab.gnome.org/GNOME/glib/-/blob/2.78.1/gio/xdgmime/xdgmimecache.c?ref_type=tags#L971

In my case, the cycle was triggered by a .local/share/mime/packages/freedesktop.org.xml file from 2021, which contains the lines

<mime-type type="application/javascript">
<sub-class-of type="application/ecmascript"/>

Deleting this file solves the issue, even with shared-mime-info 2.4-1.
So I guess that there is no way to assure that the different definitions are, over all, cycle free. The cycle breaking needs to be done by the reading applications.

I think I found my issue, I'm using Nix packages in my Archlinux system, so I have this directory with another mime packages with freedesktop.org.xml, which is from Nix shared-mime-info 2.2 package:

$ ls -al /home/ojirio/.nix-profile/share/mime/packages/
lrwxrwxrwx 1 root root 104 Jan 1 1970 freedesktop.org.xml -> /nix/store/hklmsy2ilprrywn1ppyad7aaglx2zvg4-shared-mime-info-2.2/share/mime/packages/freedesktop.org.xml

Clearly the new 2.4 shared-mime-info Arch package conflicts with that.
This directory is referenced by XDG_DATA_DIRS which I assume means is used in the mime type resolution:

$ export | grep XDG_DATA_DIRS
declare -x XDG_DATA_DIRS="/nix/store/1605y2yxxh4v85nlppv7al28d1rl73ib-mate-terminal-1.26.1/share:/nix/store/drdhg1pghyl0dvmngkadv6viv8jrm0zc-gsettings-desktop-schemas-44.0/share/gsettings-schemas/gsettings-desktop-schemas-44.0:/nix/store/6qp789r70ka5ahxfzrr9l3h6szclsgsr-gtk+3-3.24.38/share/gsettings-schemas/gtk+3-3.24.38:/nix/store/47vwyf8dfi0ny4bd31khq9ah0nqvw504-mate-desktop-1.26.2/share/gsettings-schemas/mate-desktop-1.26.2:/nix/store/1605y2yxxh4v85nlppv7al28d1rl73ib-mate-terminal-1.26.1/share/gsettings-schemas/mate-terminal-1.26.1:/nix/var/nix/profiles/default/share:/home/ojirio/.nix-profile/share:/usr/share/ubuntu:/usr/local/share:/usr/share:/var/lib/snapd/desktop:/usr/share"

So I think until Nix updates shared-mime-info to 2.4 I'll need to stay on shared-mime-info 2.3 in Arch packages.

> So I think until Nix updates shared-mime-info to 2.4 I'll need to stay on shared-mime-info 2.3 in Arch packages.
Or until glib2 can handle cyclic references.

Patch sent: https://gitlab.gnome.org/GNOME/glib/-/merge_requests/3714

So IIUC, this is an upstream bug and not an Arch packaging issue? This is also not an issue with "shared-mime-info" but rather "xdgmime" which is ultimately part of "glib2"? Please let us know when upstream commits a fix.

Maybe it's a different issue (libxml2 update?), but I'm now starting to see stuff like this on a fresh install. Anyone else?

( 4/13) Updating the MIME type database...
Error in type 'application/x-core' (in /usr/share/mime/packages/freedesktop.org.xml): Error in <match> element: Mask is longer than value.
Error in type 'image/jp2' (in /usr/share/mime/packages/freedesktop.org.xml): Error in <match> element: Mask is longer than value.
Error in type 'image/jpx' (in /usr/share/mime/packages/freedesktop.org.xml): Error in <match> element: Mask is longer than value.
Error in type 'image/jpm' (in /usr/share/mime/packages/freedesktop.org.xml): Error in <match> element: Mask is longer than value.
Error in type 'video/mj2' (in /usr/share/mime/packages/freedesktop.org.xml): Error in <match> element: Mask is longer than value.
Error in type 'image/vnd.adobe.photoshop' (in /usr/share/mime/packages/freedesktop.org.xml): Error in <match> element: Mask is longer than value.

Edit: To make matters worse, it doesn't happen every time i.e. it's intermittent, there is some randomness involved.

> So IIUC, this is an upstream bug and not an Arch packaging issue? This is also not an issue with "shared-mime-info" but rather "xdgmime" which is ultimately part of "glib2"? Please let us know when upstream commits a fix.
@ojirio reported encountering the issue using dolphin which I believe does not use glib2 but does use qt5-base and a fix has been committed [1] at least for qt6-base.

[1]: https://code.qt.io/cgit/qt/qtbase.git/commit/?id=54656da9ace06caf4a0eeb1832989c0ab211a4a0

Upstream glib2 has committed a fix [1][2], it has been back-ported to glib-2-78 as [3][4].

[1]: https://gitlab.gnome.org/GNOME/glib/-/commit/d9c50cac5d160798498e42d5b190ad89920a9a6a
[2]: https://gitlab.gnome.org/GNOME/glib/-/commit/77902e1b1c8ff906a560270656a51a28787b9492
[3]: https://gitlab.gnome.org/GNOME/glib/-/commit/e03abe74572618a58b7ac0d17c9480639e48f080
[4]: https://gitlab.gnome.org/GNOME/glib/-/commit/b382d059eec47567f693098d20bc5b936aa53704