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 775274 - evolution composer occasionally hangs for up to 15 seconds
evolution composer occasionally hangs for up to 15 seconds
Status: RESOLVED DUPLICATE of bug 777086
Product: evolution
Classification: Applications
Component: Composer
3.23.x (obsolete)
Other Linux
: Normal normal
: ---
Assigned To: Tomas Popela
Evolution QA team
Depends on:
Blocks:
 
 
Reported: 2016-11-28 19:10 UTC by Hussam Al-Tayeb
Modified: 2017-01-18 08:22 UTC
See Also:
GNOME target: ---
GNOME version: ---


Description Hussam Al-Tayeb 2016-11-28 19:10:32 UTC 
Evolution composer occasionally hangs for up to 15 seconds especially when there are highlighted words for spellchecking.
I can't type more words or right click in the composer area in the meantime.
e-d-s and evolution are git master branch and webkit is 2.15.2
Comment 1 Hussam Al-Tayeb 2016-11-28 19:45:12 UTC 
Not sure if helpful but this is the backtrace I got:

[New LWP 26607]
[New LWP 26608]
[New LWP 26609]
[New LWP 26610]
[New LWP 26612]
[New LWP 26628]
[New LWP 26629]
[New LWP 26630]
[New LWP 26633]
[New LWP 26652]
[New LWP 26653]
[New LWP 26664]
[New LWP 26676]
[New LWP 26677]
[New LWP 26678]
[New LWP 26687]
[New LWP 26688]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/libthread_db.so.1".
0x00007f47922a813d in poll () from /usr/lib/libc.so.6

+ Trace 236894







Comment 2


Hussam Al-Tayeb





          2016-11-30 13:00:58 UTC
        



This looks fixed after https://git.gnome.org/browse/evolution/commit/?id=05bad42e6a09ea429e4aa41ae106280946018f5d






Comment 3


Hussam Al-Tayeb





          2016-11-30 14:46:13 UTC
        



Still happening.






Comment 4


Milan Crha





          2016-12-06 14:07:29 UTC
        



Thanks for a bug report. The backtrace doesn't show anything unusual, the evolution itself is simply idle, waiting for your orders. Try to get backtrace of the WebKitWebProcess too. there will be at least two, one is for the message preview, another is for the composers. The one for the composers stops as soon as all the instances of the composer are closed (and freed).






Comment 5


Hussam Al-Tayeb





          2016-12-07 12:00:37 UTC
        



Hi. I am trying but gdb keeps wanting to restart webkitwebprocess. The one time I didn't, I saw:

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/libthread_db.so.1".
0x00007fe39fb8813d in poll () from /usr/lib/libc.so.6
(gdb) bt full




+
Trace
    236938






    

  • 
#0
poll

    • 

  • 
#1
g_main_context_poll

    
                at gmain.c
                  line
                  4228

    

    • 

  • 
#2
g_main_context_iterate

    
                at gmain.c
                  line
                  3924

    

    • 

  • 
#3
g_main_loop_run

    
                at gmain.c
                  line
                  4125

    

    • 

  • 
#4
WTF::RunLoop::run()

    
                at /home/hussam/cache/webkit2gtk/src/webkitgtk-2.15.2/Source/WTF/wtf/glib/RunLoopGLib.cpp
                  line
                  94

    

    • 

  • 
#5
WebKit::ChildProcessMain<WebKit::WebProcess, WebKit::WebProcessMain>(int, char**)

    
                at /home/hussam/cache/webkit2gtk/src/webkitgtk-2.15.2/Source/WebKit2/Shared/unix/ChildProcessMain.h
                  line
                  61

    

    • 

  • 
#6
__libc_start_main

    • 

  • 
#7
_start

    • 













Comment 6


Milan Crha





          2016-12-07 14:43:59 UTC
        



It depends how you capture the backtrace. One command which works for me is:

   $ gdb --batch --ex "t a a bt" -pid=`pidof evolution` &>bt.txt

change the -pid argument to other process ID to get a different process backtrace.






Comment 7


Hussam Al-Tayeb





          2017-01-17 11:19:57 UTC
        



It's too early to tell but I think applying the patch from https://trac.webkit.org/changeset/210670 maybe have fixed this.
I haven't encountered lag so far today.






Comment 8


Hussam Al-Tayeb





          2017-01-17 19:42:41 UTC
        



Nevermind, it happened again.

I don't have a debug webkit installed as it was very unstable the last time I tried.

[New LWP 9611]
[New LWP 9612]
[New LWP 9613]
[New LWP 9614]
[New LWP 9620]
[New LWP 9657]
[New LWP 9658]
[New LWP 9659]
[New LWP 9662]
[New LWP 9694]
[New LWP 9695]
[New LWP 9731]
[New LWP 10289]
[New LWP 10449]
[New LWP 10450]
[New LWP 10451]
[New LWP 10469]
[New LWP 10470]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/libthread_db.so.1".
0x00007f5096a0a11d in poll () from /usr/lib/libc.so.6





+
Trace
    237065






Thread 1
          (Thread 0x7f509b9ca400 (LWP 9609))


    

  • 
#0
poll

    • 

  • 
#1
g_main_context_poll

    
                at gmain.c
                  line
                  4228

    

    • 

  • 
#2
g_main_context_iterate

    
                at gmain.c
                  line
                  3924

    

    • 

  • 
#3
g_main_context_iteration

    
                at gmain.c
                  line
                  3990

    

    • 

  • 
#4
e_util_invoke_g_dbus_proxy_call_sync_wrapper_full

    • 

  • 
#5
e_util_invoke_g_dbus_proxy_call_sync_wrapper_with_error_check

    • 

  • 
#6
0x00007f506cac9644 in

    • 

  • 
#7
0x00007f5096db6e38 in

    • 

  • 
#11
<emit signal ??? on instance 0x266d960 [EHTMLEditor]>

    
                at gsignal.c
                  line
                  3447

    

    • 

  • 
#12
0x00007f5096db79f1 in

    • 

  • 
#13
ffi_call_unix64

    • 

  • 
#14
ffi_call

    • 

  • 
#19
<emit signal ??? on instance 0x2f4cac0 [EWebKitEditor]>

    
                at gsignal.c
                  line
                  3447

    

    • 

  • 
#20
e_content_editor_emit_context_menu_requested

    • 

  • 
#21
webkit_marshal_BOOLEAN__OBJECT_BOXED_OBJECT

    • 

  • 
#25
<emit signal ??? on instance 0x2f4cac0 [EWebKitEditor]>

    
                at gsignal.c
                  line
                  3447

    

    • 

  • 
#26
0x00007f50978c03ae in

    • 

  • 
#27
0x00007f50978933aa in

    • 

  • 
#28
0x00007f50978edcff in

    • 

  • 
#29
0x00007f50976b8223 in

    • 

  • 
#30
0x00007f50976b846b in

    • 

  • 
#31
0x00007f50979368b5 in

    • 

  • 
#32
0x00007f509792fd98 in

    • 

  • 
#33
0x00007f509761fd29 in

    • 

  • 
#34
0x00007f50976ddc72 in

    • 

  • 
#35
0x00007f509761baf6 in

    • 

  • 
#36
0x00007f509761cb68 in

    • 

  • 
#37
WTF::RunLoop::performWork()

    • 

  • 
#38
0x00007f509075f819 in

    • 

  • 
#39
g_main_dispatch

    
                at gmain.c
                  line
                  3203

    

    • 

  • 
#40
g_main_context_dispatch

    
                at gmain.c
                  line
                  3856

    

    • 

  • 
#41
g_main_context_iterate

    
                at gmain.c
                  line
                  3929

    

    • 

  • 
#42
g_main_loop_run

    
                at gmain.c
                  line
                  4125

    

    • 

  • 
#43
gtk_main

    
                at gtkmain.c
                  line
                  1312

    

    • 

  • 
#44
main

    • 













Comment 9


Milan Crha





          2017-01-18 08:22:50 UTC
        



Aha, that's when you right-click and ideally on a misspelled word. I see. It's bug #777086.

*** This bug has been marked as a duplicate of bug 777086 ***