Bug 2415460

Thanks for a bug report. It looks like it's crashing somewhere in glib2 code, which does not necessarily mean it's a problem in the glib2 itself. Could you install debuginfo packages for the glib2, evolution-data-server and evolution, like:

   sudo dnf install glib2-debuginfo evolution-data-server-debuginfo evolution-debuginfo --enable-repo=updates-debuginfo

only make sure the versions of the debuginfo packages exactly match those installed, aka:

   rpm -q glib2 evolution-data-server evolution


and then run evolution from a terminal under gdb to get the backtrace of the crash, please? You can get the backtrace with a command like this:

   DEBUGINFOD_URLS="" gdb --batch evolution --ex r --ex "t a a bt" &>bt.txt

Please check the bt.txt for any private information, like passwords, email addresses, server addresses,... I usually search for "pass" at least (quotes for clarity only), before sharing it anywhere.

Created attachment 2114909 [details]
gdb backtrace, as requested.

Thanks for a quick update. I see it's crashing in the WebKitGTK code, due to at-spi D-Bus problem. That had been reported in bug #2162351 some time ago, which references an upstream bug and also a link to a workaround:
https://gitlab.gnome.org/GNOME/at-spi2-core/-/issues/180#note_2440753

It's untested by me, I do not face this.

As the path references `lxdm`, it might be a leftover from the enabled accessibility in the login manager (or how it's called) of the LXDE desktop environment, because LXDE is mentioned in the upstream bug as well.

*** This bug has been marked as a duplicate of bug 2162351 ***

Followed your links.

$ NO_AT_BRIDGE=1 evolution : fails.
$ WEBKIT_FORCE_SANDBOX=0 evolution : fails.

In a message from Apr 2023, I found this :

WEBKIT_DISABLE_SANDBOX_THIS_IS_DANGEROUS=1 evolution

and this works. Michael Catanzaro commented on that post :

"Uh, needless to say, I don’t recommend globally adding environment variables with DANGEROUS in the name."

I am surprised this workaround is still necessary. I understand "evolution" is not involved. But you have more authority than I have in negotiating with the webkitgtk team.

I do not have any authority with the WebKitGTK folks, I'm sorry. :) The most I can ping some people, but that's all. They work on things they can. WebKit(GTK) has too many opened bugs, not talking they need to stay current with the technologies, which are constantly evolving and changing.

Anyway, while it's the WebKitGTK code casing the application crash, the problem can be very well elsewhere, at least the root cause of the problem. I know some other distributions have this problem too, apart of those linked here.

I do not recall many details, unfortunately, but from the error message you've got here, the problem is that the lxdm left an at-spi D-Bus service running, while the user was supposed to run its own after he/she logs in. That ensures the logged in user has his/her own at-spi D-Bus service, to which it has access and control of. If I recall correctly, it also happened to some users that the leftover at-spi D-Bus service was closed some time after they logged in, then the new could be started and every thing worked fine again. You can try to find out which process is serving the at-spi D-Bus service and kill it, to imitate the closure of it. Trying to run here `ps ax | grep at-spi` it shows multiple processes running here, but I'm sorry, I do not know which one is "the one", if any at all.