Bug 348170 – Totem eats memory like there's no tomorrow

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 348170 - Totem eats memory like there's no tomorrow


Summary:	Totem eats memory like there's no tomorrow


Status:	RESOLVED DUPLICATE of bug 348222

Product:	totem
Classification:	Core
Component:	xine-lib backend
Version:	1.5.x
Hardware:	Other Linux

Importance:	Normal normal
Target Milestone:	---
Assigned To:	Maintainer alias for xine-lib component of Totem
QA Contact:	Maintainer alias for xine-lib component of Totem

URL:
Whiteboard:

Depends on:
Blocks:

Reported:	2006-07-20 19:13 UTC by Fabio Bonelli
Modified:	2006-07-21 10:53 UTC

See Also:
GNOME target:	---
GNOME version:	2.15/2.16

Description Fabio Bonelli 2006-07-20 19:13:26 UTC

totem 1.5.4-0ubuntu1, xine-lib 1.1.2-2ubuntu2

Start totem and press just "1". The results at second 8 (!) is:

fabio    19768  4.8 59.8 1562568 266740 pts/1  Tl   21:02   0:08 /usr/bin/totem

Stack trace from the thread responsable:

+ Trace 69483

#0 gdk_pixbuf_alpha_mode_get_type
from /usr/lib/libgdk_pixbuf-2.0.so.0
#1 _pixops_scale
from /usr/lib/libgdk_pixbuf-2.0.so.0
#2 gdk_pixbuf_scale
from /usr/lib/libgdk_pixbuf-2.0.so.0
#3 gdk_pixbuf_scale_simple
from /usr/lib/libgdk_pixbuf-2.0.so.0
#4 bacon_video_widget_dvd_event
#5 _gtk_marshal_BOOLEAN__BOXED
from /usr/lib/libgtk-x11-2.0.so.0
#6 g_closure_invoke
from /usr/lib/libgobject-2.0.so.0
#7 g_signal_chain_from_overridden
from /usr/lib/libgobject-2.0.so.0
#8 g_signal_emit_valist
from /usr/lib/libgobject-2.0.so.0
#9 g_signal_emit
from /usr/lib/libgobject-2.0.so.0
#10 gtk_widget_get_default_style
from /usr/lib/libgtk-x11-2.0.so.0
#11 gtk_main_do_event
from /usr/lib/libgtk-x11-2.0.so.0
#12 gdk_window_is_viewable
from /usr/lib/libgdk-x11-2.0.so.0
#13 gdk_window_process_all_updates
from /usr/lib/libgdk-x11-2.0.so.0
#14 gtk_window_new
from /usr/lib/libgtk-x11-2.0.so.0
#15 g_closure_invoke
from /usr/lib/libgobject-2.0.so.0
#16 g_signal_chain_from_overridden
from /usr/lib/libgobject-2.0.so.0
#17 g_signal_emit_valist
from /usr/lib/libgobject-2.0.so.0
#18 g_signal_emit
from /usr/lib/libgobject-2.0.so.0
#19 gtk_container_check_resize
from /usr/lib/libgtk-x11-2.0.so.0
#20 g_main_context_dispatch
from /usr/lib/libglib-2.0.so.0
#21 g_main_context_check
from /usr/lib/libglib-2.0.so.0
#22 g_main_loop_run
from /usr/lib/libglib-2.0.so.0
#23 gtk_main
from /usr/lib/libgtk-x11-2.0.so.0
#24 main

Comment 1 Fabio Bonelli 2006-07-20 19:17:01 UTC

Doesn't happen with xine-gstreamer: when pressing "1" the window doesn't even try to resize itself.

Comment 2 Bastien Nocera 2006-07-20 20:31:04 UTC

(In reply to comment #1)
> Doesn't happen with xine-gstreamer: when pressing "1" the window doesn't even
> try to resize itself.
 
Huh? "xine-gstreamer" ?

Comment 3 Bastien Nocera 2006-07-20 20:55:14 UTC

BTW, I can't reproduce that with the current CVS.

Comment 4 Fabio Bonelli 2006-07-21 00:40:38 UTC

Err... totem-gstreamer.

Better stacktrace:

+ Trace 69488

#0 pixops_process
at pixops.c line 1026
#1 _pixops_scale
at pixops.c line 1623
#2 IA__gdk_pixbuf_scale
at gdk-pixbuf-scale.c line 77
#3 IA__gdk_pixbuf_scale_simple
at gdk-pixbuf-scale.c line 256
#4 bacon_video_widget_dvd_event
#5 _gtk_marshal_BOOLEAN__BOXED
at gtkmarshalers.c line 83
#6 g_closure_invoke
from /usr/lib/libgobject-2.0.so.0
#7 g_signal_chain_from_overridden
from /usr/lib/libgobject-2.0.so.0
#8 g_signal_emit_valist
from /usr/lib/libgobject-2.0.so.0
#9 g_signal_emit
from /usr/lib/libgobject-2.0.so.0
#10 gtk_widget_event_internal
at gtkwidget.c line 3901
#11 IA__gtk_main_do_event
at gtkmain.c line 1406
#12 gdk_window_process_updates_internal
at gdkwindow.c line 2324
#13 IA__gdk_window_process_all_updates
at gdkwindow.c line 2387
#14 gtk_window_check_resize
at gtkwindow.c line 5700
#15 g_closure_invoke
from /usr/lib/libgobject-2.0.so.0
#16 g_signal_chain_from_overridden
from /usr/lib/libgobject-2.0.so.0
#17 g_signal_emit_valist
from /usr/lib/libgobject-2.0.so.0
#18 g_signal_emit
from /usr/lib/libgobject-2.0.so.0
#19 gtk_container_idle_sizer
at gtkcontainer.c line 1110
#20 g_main_context_dispatch
from /usr/lib/libglib-2.0.so.0
#21 g_main_context_check
from /usr/lib/libglib-2.0.so.0
#22 g_main_loop_run
from /usr/lib/libglib-2.0.so.0
#23 IA__gtk_main
at gtkmain.c line 1023
#24 main

Comment 5 Bastien Nocera 2006-07-21 07:29:30 UTC

I don't understand what this backtrace is supposed to be.
The only time gdk_pixbuf_scale_simple() is called, the result is unref 3 lines down.

Please try with the current HEAD, and be a bit more precise as to how much is leaked, and use something like valgrind to show the leak.

Comment 6 Fabio Bonelli 2006-07-21 10:14:38 UTC

There isn't a memory leak, at first sight I was thinking about an allocation inside an infinite loop, but it's just a bug in gdk_pixbuf_scale for which I file a bug 348222.

This is reproducible with totem-xine HEAD and GTK 2.10.0.

This bug now turns to: "totem tries to resize totem_logo.png to 1x1 when hitting '1'"

Comment 7 Bastien Nocera 2006-07-21 10:18:34 UTC

Ha, it's not me that's done it then :)

The second part is bug 348184 that I'll be looking at soon.

*** This bug has been marked as a duplicate of 348222 ***

Comment 8 Fabio Bonelli 2006-07-21 10:53:46 UTC

It seems so. :-)

BTW, thanks for your always-blazing-fast response time!