Bug 112425 – Nautilus in infinite error loop

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 112425 - Nautilus in infinite error loop


Summary:	Nautilus in infinite error loop


Status:	RESOLVED FIXED

Product:	nautilus
Classification:	Core
Component:	general
Version:	2.3.x
Hardware:	Other Solaris

Importance:	High normal
Target Milestone:	---
Assigned To:	Nautilus Maintainers
QA Contact:	Nautilus Maintainers

URL:
Whiteboard:

Duplicates:	112793 (view as bug list)
Depends on:
Blocks:

Reported:	2003-05-06 18:06 UTC by Peter O'Shea
Modified:	2017-07-02 13:48 UTC

See Also:
GNOME target:	---
GNOME version:	2.3/2.4

Attachments
the "at least it doesn't hang" patch (804 bytes, patch) 2003-05-14 00:53 UTC, Samuel Stringham	none	Details \| Review

Description Peter O'Shea 2003-05-06 18:06:14 UTC

Nautilus 2.3.1, when run from the command line, produces an infinite number
of the following error messages:

*** attempt to put segment in horiz list twice
*** attempt to put segment in horiz list twice
*** attempt to put segment in horiz list twice
*** attempt to put segment in horiz list twice
*** attempt to put segment in horiz list twice
....etc ad nauseam.

It also takes up nearly 100% of the cpu and a boatload of memory.

Comment 1 Peter O'Shea 2003-05-07 20:45:16 UTC

After exiting and re-entering gnome, nautilus now starts up with only
two of the above error messages.  A home directory window appears, but
the throbber button keeps rotating and the nautilus process goes up to
nearly 100% cpu.

If I run it under gdb, wait until it hangs, ctrl-C and get a
backtrace, this is what I come up with:

^C
Program received signal SIGINT, Interrupt.
0xfece3e48 in gdk_pixbuf_get_width ()
(gdb) bt

+ Trace 36663

#0 gdk_pixbuf_get_width
#1 nautilus_icon_canvas_item_bounds
#2 eel_canvas_item_get_bounds
#3 icon_get_bounding_box
#4 lay_down_icons_tblr
#5 finish_adding_new_icons
#6 redo_layout_internal
#7 g_cclosure_marshal_VOID__BOXED
#8 g_type_class_meta_marshal
#9 g_closure_invoke
#10 signal_emit_unlocked_R
#11 g_signal_emit_valist
#12 g_signal_emit
#13 gtk_widget_size_allocate
#14 gtk_scrolled_window_size_allocate
#15 g_cclosure_marshal_VOID__BOXED
#16 g_type_class_meta_marshal
#17 g_closure_invoke
#18 signal_emit_unlocked_R
#19 g_signal_emit_valist
#20 g_signal_emit
#21 gtk_widget_size_allocate
#22 gtk_window_size_allocate
#23 gtk_plug_size_allocate
#24 bonobo_plug_size_allocate
#25 g_cclosure_marshal_VOID__BOXED
#26 g_type_class_meta_marshal
#27 g_closure_invoke
#28 signal_emit_unlocked_R
#29 g_signal_emit_valist
#30 g_signal_emit
#31 gtk_widget_size_allocate
#32 gtk_window_show
#33 gtk_plug_show
#34 g_cclosure_marshal_VOID__VOID
#35 g_type_class_meta_marshal
#36 g_closure_invoke
#37 signal_emit_unlocked_R
#38 g_signal_emit_valist
#39 g_signal_emit
#40 gtk_widget_show
#41 impl_Bonobo_Control_getWindowId
#42 Bonobo_Control_getWindowId
#43 bonobo_control_frame_get_remote_window
#44 bonobo_socket_realize
#45 g_cclosure_marshal_VOID__VOID
#46 g_type_class_meta_marshal
#47 g_closure_invoke
#48 signal_emit_unlocked_R
#49 g_signal_emit_valist
#50 g_signal_emit
#51 gtk_widget_realize
#52 gtk_widget_map
#53 gtk_bin_forall
#54 gtk_container_forall
#55 gtk_container_map
#56 nautilus_view_frame_map
#57 g_cclosure_marshal_VOID__VOID
#58 g_type_class_meta_marshal
#59 g_closure_invoke
#60 signal_emit_unlocked_R
#61 g_signal_emit_valist
#62 g_signal_emit
#63 gtk_widget_map
#64 gtk_bin_forall
#65 gtk_container_forall
#66 gtk_container_map
#67 g_cclosure_marshal_VOID__VOID
#68 g_type_class_meta_marshal
#69 g_closure_invoke
#70 signal_emit_unlocked_R
#71 g_signal_emit_valist
#72 g_signal_emit
#73 gtk_widget_map
#74 bonobo_dock_forall
#75 gtk_container_forall
#76 gtk_container_map
#77 bonobo_dock_map
#78 g_cclosure_marshal_VOID__VOID
#79 g_type_class_meta_marshal
#80 g_closure_invoke
#81 signal_emit_unlocked_R
#82 g_signal_emit_valist
#83 g_signal_emit
#84 gtk_widget_map
#85 gtk_box_forall
#86 gtk_container_forall
#87 gtk_container_map
#88 g_cclosure_marshal_VOID__VOID
#89 g_type_class_meta_marshal
#90 g_closure_invoke
#91 signal_emit_unlocked_R
#92 g_signal_emit_valist
#93 g_signal_emit
#94 gtk_widget_map
#95 gtk_window_map
#96 map
#97 g_cclosure_marshal_VOID__VOID
#98 g_type_class_meta_marshal
#99 g_closure_invoke
#100 signal_emit_unlocked_R
#101 g_signal_emit_valist
#102 g_signal_emit
#103 gtk_widget_map
#104 gtk_window_show
#105 nautilus_window_show
#106 g_cclosure_marshal_VOID__VOID
#107 g_type_class_meta_marshal
#108 g_closure_invoke
#109 signal_emit_unlocked_R
#110 g_signal_emit_valist
#111 g_signal_emit
#112 gtk_widget_show
#113 g_cclosure_marshal_VOID__VOID
#114 g_closure_invoke
#115 signal_emit_unlocked_R
#116 g_signal_emit_valist
#117 g_signal_emit
#118 execute_queued_functions
#119 g_idle_dispatch
#120 g_main_dispatch
#121 g_main_context_dispatch
#122 g_main_context_iterate
#123 g_main_loop_run
#124 gtk_main
#125 main

Comment 2 Dave Camp 2003-05-07 20:57:55 UTC

might be from the snap-to-grid patch, I'll take a look.

Comment 3 Samuel Stringham 2003-05-13 17:10:09 UTC

*** Bug 112793 has been marked as a duplicate of this bug. ***

Comment 4 Samuel Stringham 2003-05-13 19:05:11 UTC

if, in src/file-manager/fm-desktop-icon-view.c you change line 708 from:
                                                
NAUTILUS_ICON_LAYOUT_T_B_L_R);

to
                                                
NAUTILUS_ICON_LAYOUT_L_R_T_B);

it stops calling lay_down_icons_tblr() which is the function with the
infinite loop in it.  This is just a hack, but it works.  Also, adding
garnome to the CC, since it is in 0.24.2

Comment 5 Samuel Stringham 2003-05-14 00:28:43 UTC

okay, I have some more information.  This seems related to bug 92333 .
  gtk_plug_size_allocate is getting called more than once, in that bug
it happened to be doing it 3 times.  Our bug is calling the
gtk_plug_size_allocate method 4 times.  The third time it calls it the
data is correct, but the fourth (and final) time the height/width of
the window is somehow corrupted.  I am looking harder.

Comment 6 Samuel Stringham 2003-05-14 00:53:15 UTC

Created attachment 16512 [details] [review]
the "at least it doesn't hang" patch

Comment 7 Peter O'Shea 2003-05-14 15:38:32 UTC

Confirmed that applying the patch makes 2.3.1 run without hanging on
my system.  Thanks for the workaround.

Comment 8 Alexander Larsson 2003-06-12 11:29:56 UTC

A similar patch is in cvs already.