Version-Release number of selected component: evolution-3.26.6-1.fc27 Additional info: reporter: libreport-2.9.3 backtrace_rating: 4 cmdline: evolution crash_function: gtk_action_group_set_visible executable: /usr/bin/evolution journald_cursor: s=e783bba16abc4b429c85d1aa4fbb46aa;i=73125;b=674a72949cd34f258b9d799535305466;m=d2f5e904;t=5684fe1c14732;x=6a45c874303ee4d8 kernel: 4.15.10-300.fc27.x86_64 rootdir: / runlevel: N 5 type: CCpp uid: 1000 Truncated backtrace: Thread no. 1 (10 frames) #0 gtk_action_group_set_visible at deprecated/gtkactiongroup.c:810 #1 action_mode_cb at /usr/src/debug/evolution-3.26.6-1.fc27.x86_64/src/e-util/e-html-editor-actions.c:516 #6 gtk_radio_action_activate at deprecated/gtkradioaction.c:381 #11 _gtk_action_emit_activate at deprecated/gtkaction.c:909 #12 gtk_toggle_action_set_active at deprecated/gtktoggleaction.c:299 #13 action_combo_box_changed at /usr/src/debug/evolution-3.26.6-1.fc27.x86_64/src/e-util/e-action-combo-box.c:418 #18 gtk_combo_box_set_active_internal at gtkcombobox.c:3847 #19 gtk_combo_box_set_active_iter at gtkcombobox.c:3909 #20 gtk_combo_box_menu_activate at gtkcombobox.c:2821 #21 g_cclosure_marshal_VOID__STRINGv at gmarshal.c:1794
Created attachment 1413123 [details] File: backtrace
Created attachment 1413124 [details] File: cgroup
Created attachment 1413125 [details] File: core_backtrace
Created attachment 1413126 [details] File: cpuinfo
Created attachment 1413127 [details] File: dso_list
Created attachment 1413128 [details] File: environ
Created attachment 1413129 [details] File: exploitable
Created attachment 1413130 [details] File: limits
Created attachment 1413131 [details] File: maps
Created attachment 1413132 [details] File: mountinfo
Created attachment 1413133 [details] File: open_fds
Created attachment 1413134 [details] File: proc_pid_status
Created attachment 1413135 [details] File: var_log_messages
Thanks for a bug report. According to var_log_messages there happened something to the background WebKitWebProcess, it had been probably locked or busy with something or even crashed, that it caused the calls to it timeout. It seems you've been changing something in the message composer window, whose body part might be frozen, I guess. The thing is that the changes in the composer window itself are not that important, unless they caused the WebKitWebProcess lock up. Do you remember what you did before the crash, please? It can be you replied to a message, or you had been constructing a new message with certain changes, then one of the changes caused the freeze of the message body? Maybe you changed from HTML to Plain Text mode, or vice versa?
Indeed the bug happened after trying to reply to a message (with HTML formatting enabled by default) which made the composer hang, and after that Evolution had all sorts of issues. I've been able to crash Evolution again by replying to that message, typing/removing some text until the composer hangs, and then doing a few things (new message, etc.). I've just sent you the offending message by private e-mail. I think the reason is that the original message has a large HTML attachment, which is rendered direcly in the body when replying (which doesn't make sense to me).
Thanks for the message. I can reproduce the hangs, the WebKitWebProcess is busy, mostly with spell checking here, which causes freezes of the composer body and high CPU usage. I checked the content of the message and it has no attachments, it's really that long HTML formatted message. The message structure is: multipart/alternative text/plain text/html I could not reproduce the crash though. I saw some runtime warnings on the console of evolution, but that all was just about the WebKitWebProcess being busy and unresponsive to the calls evolution does to it. One such is: (evolution:14123): evolution-util-WARNING **: Failed to call a DBus Proxy method org.gnome.Evolution.WebExtension.EWebKitEditor.WC0x599feb0:: ElementSetAttributeBySelector: Timeout was reached I did not get the crash itself, unfortunately, though I've been able to reproduce some runtime warnings, but none seemed related to the place where it crashed for you (when I've been lucky, I made the composer busy before it had been trying to auto-save the message and I closed the composer before the auto-save finished, which caused some runtime warnings on the console. That would be similar to bug #1538865.
> I checked the content of the message and it has no attachments, it's really that long HTML formatted message. Hmm, that's weird, it definitely has an attachment here. But indeed I see in my Sent box that when transferring it, Evo changed the attached HTML file to inline HTML content. That must be the same problem. I've sent you the HTML attachment from a Webmail to ensure it's not inlined. This is an annoying bug that I've seen for a long time.
I see. The new message has an attachment, clearly marked as such also for the code, but it's included in the replied message body. I agree it's a wrong behaviour. Would you mind if I change this bug report from a crash report to "Reply includes HTML attachment in message body" and fix that, instead of the crash, please?
No, that's fine. Anyway that's the only way to trigger the hang.
Until you receive reeeeealy long mail. I see that the second part is under an attachment node and it's set as hidden when the list of parts is generated, but when the quotation is using the part it is marked as not being hidden, thus it's included in the message. I'm currently searching what and why unmarked the hidden flag.
I figured out I caused it with this change: https://bugzilla.gnome.org/show_bug.cgi?id=681279#c14 I made a better fix here: Created commit d8a5e6267f in evo master (3.29.1+) [1] Created commit b7b0a1f975 in evo gnome-3-28 (3.28.1+) [1] https://git.gnome.org/browse/evolution/commit/?id=d8a5e6267f
Thanks!