After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 794693 - Crash when trying to view email message source
Crash when trying to view email message source
Status: RESOLVED FIXED
Product: evolution
Classification: Applications
Component: Mailer
3.26.x (obsolete)
Other Linux
: Normal critical
: ---
Assigned To: evolution-mail-maintainers
Evolution QA team
: 795119 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2018-03-26 11:52 UTC by André Klapper
Modified: 2018-04-10 14:14 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
stacktrace (120.54 KB, text/plain)
2018-03-26 11:53 UTC, André Klapper
Details
output of "G_SLICE=always-malloc valgrind --num-callers=30 evolution" (120.54 KB, text/plain)
2018-03-26 13:33 UTC, André Klapper
Details
output of "G_SLICE=always-malloc valgrind --num-callers=30 evolution" (67.33 KB, text/plain)
2018-03-26 13:35 UTC, André Klapper
Details

Description André Klapper 2018-03-26 11:52:39 UTC
webkitgtk4-2.18.6-1.fc27.x86_64
evolution-3.26.6-1.fc27.x86_64
evolution-data-server-3.26.6-1.fc27.x86_64
gtk3-3.22.26-2.fc27.x86_64

(evolution:19369): evolution-mail-CRITICAL **: mail_reader_display_load_changed_cb: assertion 'E_IS_MAIL_READER (reader)' failed

(evolution:19369): Gtk-CRITICAL **: gtk_action_set_sensitive: assertion 'GTK_IS_ACTION (action)' failed

(evolution:19369): Gtk-CRITICAL **: gtk_action_set_sensitive: assertion 'GTK_IS_ACTION (action)' failed

(evolution:19369): GLib-GObject-WARNING **: invalid unclassed pointer in cast to 'GtkWidget'

(evolution:19369): Gtk-CRITICAL **: gtk_widget_get_visible: assertion 'GTK_IS_WIDGET (widget)' failed

(evolution:19369): evolution-util-CRITICAL **: e_web_view_update_highlights: assertion 'E_IS_WEB_VIEW (web_view)' failed

(evolution:19369): GLib-GObject-WARNING **: invalid unclassed pointer in cast to 'GObject'

(evolution:19369): GLib-GObject-CRITICAL **: g_object_get_qdata: assertion 'G_IS_OBJECT (object)' failed

Thread 1 "evolution" received signal SIGSEGV, Segmentation fault.
mail_reader_load_changed_cb (reader=0x55555916c5d0, event=<optimized out>, display=0x555559baf550 [EMailDisplay]) at /usr/src/debug/evolution-3.26.6-1.fc27.x86_64/src/mail/e-mail-reader.c:2920
Comment 1 André Klapper 2018-03-26 11:53:17 UTC
Created attachment 370146 [details]
stacktrace
Comment 2 Milan Crha 2018-03-26 12:19:29 UTC
Thanks for a bug report. I suppose it's an up-to-today F27 without updates-testing in GNOME under Wayland, right?

The first runtime warning might have the clue. The rest can be just due to it. Incorporate G_SLICE=always-malloc ideally.

I cannot run evolution under GNOME with Wayland here, due to
   mesa-dri-drivers-17.3.6-1.fc27.x86_64
update, thus I downgraded it to:
   mesa-dri-drivers.x86_64 17.2.2-4.fc27
which downgraded also
   mesa-filesystem.x86_64 17.2.2-4.fc27

I do not see any such runtime warnings, neither the crash, after pressing View->Message Source from the menu of Evolution when run in GNOME under Wayland with those package versions.
Comment 3 André Klapper 2018-03-26 13:33:19 UTC
Created attachment 370149 [details]
output of "G_SLICE=always-malloc valgrind --num-callers=30 evolution"

Here: mesa-dri-drivers-17.3.6-1.fc27, mesa-filesystem-17.3.6-1.fc27.

"G_SLICE=always-malloc valgrind --num-callers=30 evolution" gets me a SIGSEGV.
Comment 4 André Klapper 2018-03-26 13:35:43 UTC
Created attachment 370150 [details]
output of "G_SLICE=always-malloc valgrind --num-callers=30 evolution"

This time the correct file.
Comment 5 André Klapper 2018-03-27 15:39:39 UTC
This is on Wayland.
Comment 6 André Klapper 2018-03-27 15:43:48 UTC
Interesting. Cannot reproduce anymore.
Comment 7 Milan Crha 2018-03-29 07:47:37 UTC
I've been able to reproduce similar crash. For me, the window which was supposed to show the message source had been closed before it had been populated, with some ongoing operations in the background (like in the idle queue), and when they elapsed they used already freed objects, which could lead at least to some runtime warnings and in some cases also to the crash. I've seen it in the same functions as is shown in your valgrind long, similarly with the runtime warnings (I've got even more of them), thus I believe it's that issue.

Created commit 913472df2c in evo master (3.29.1+)
Created commit c4ada09c6e in evo gnome-3-28 (3.28.1+)
Comment 8 Milan Crha 2018-04-10 14:14:01 UTC
*** Bug 795119 has been marked as a duplicate of this bug. ***