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 74374 - Natulis desktop crash when starting Gnome-desktop (startx from a shell)
Natulis desktop crash when starting Gnome-desktop (startx from a shell)
Status: RESOLVED FIXED
Product: nautilus
Classification: Core
Component: general
2.3.x
Other other
: Urgent critical
: 2.4.x
Assigned To: Nautilus Maintainers
Nautilus Maintainers
: 96660 96799 96914 97131 97135 97263 98449 114318 118187 119041 120192 120707 122148 152975 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2002-03-12 17:22 UTC by andewid
Modified: 2004-12-22 21:47 UTC
See Also:
GNOME target: ---
GNOME version: 2.3/2.4

Attachments
Proposed Patch (1005 bytes, patch)
2003-09-18 03:12 UTC, mwehner 		none Details | Review

Description andewid 2002-03-12 15:19:50 UTC 
Package: nautilus
Severity: blocker
Version: 1.1.8
Synopsis: Natulis desktop crash when starting Gnome-desktop (startx from a shell)
Bugzilla-Product: nautilus
Bugzilla-Component: general
BugBuddy-GnomeVersion: 2.0 (1.112.1)

Description:
Description of Problem:

Natilus doesn't start when starting the Gnome 2 desktop. It crashes and
10-15 "program crash" windows come up.

I am running RedHat 7.2 with the latest updates from Red-Carpet.

Steps to reproduce the problem:
1.  Install Gnome 2 desktop from Ximian Red-Carpet
2.  run startx from a shell.
3.  Watch how Gnome is starting, and natilus crashes after a few
seconds..

Actual Results:


Expected Results:


How often does this happen?

Everytime. It seems to be possible (sometimes) to start nautilus from a
X-server without gnome (from an xterm).

Additional Information:



Debugging Information:

[New Thread 1024 (LWP 4257)]
0x40959ca9 in __wait4 () from /lib/i686/libc.so.6

+ Trace 19019

  • #0 __wait4
    from /lib/i686/libc.so.6
  • #1 __DTOR_END__
    from /lib/i686/libc.so.6
  • #2 waitpid
    at wrapsyscall.c line 172
  • #3 libgnomeui_segv_handle
    at gnome-ui-init.c line 598
  • #4 pthread_sighandler
    at signals.c line 97
  • #5 <signal handler called>
  • #6 g_logv
    at gmessages.c line 551
  • #7 g_log
    at gmessages.c line 574
  • #8 die_on_failed_activation
    at nautilus-directory-metafile.c line 90
  • #9 get_factory
    at nautilus-directory-metafile.c line 108
  • #10 open_metafile
    at nautilus-directory-metafile.c line 131
  • #11 get_metafile
    at nautilus-directory-metafile.c line 155
  • #12 nautilus_directory_register_metadata_monitor
    at nautilus-directory-metafile.c line 528
  • #13 nautilus_directory_call_when_ready_internal
    at nautilus-directory-async.c line 1279
  • #14 vfs_file_call_when_ready
    at nautilus-vfs-file.c line 70
  • #15 nautilus_file_call_when_ready
    at nautilus-file.c line 4737
  • #16 nautilus_determine_initial_view
    at nautilus-applicable-views.c line 159
  • #17 begin_location_change
    at nautilus-window-manage-views.c line 1456
  • #18 nautilus_window_open_location
    at nautilus-window-manage-views.c line 741
  • #19 nautilus_desktop_window_update_directory
    at nautilus-desktop-window.c line 94
  • #20 nautilus_desktop_window_new
    at nautilus-desktop-window.c line 118
  • #21 nautilus_application_create_desktop_window
    at nautilus-application.c line 582
  • #22 _ORBIT_skel_small_Nautilus_Shell_start_desktop
    at nautilus-shell-interface-skels.c line 58
  • #23 ORBit_POAObject_invoke
    at poa.c line 891
  • #24 ORBit_OAObject_invoke
    at orbit-adaptor.c line 116
  • #25 ORBit_small_invoke_adaptor
    at orbit-small.c line 772
  • #26 ORBit_POAObject_handle_request
    at poa.c line 1085
  • #27 ORBit_POA_handle_request
    at poa.c line 1172
  • #28 ORBit_handle_request
    at orbit-adaptor.c line 79
  • #29 giop_connection_handle_input
    at giop-recv-buffer.c line 945
  • #30 linc_connection_io_handler
    at linc-connection.c line 840
  • #31 linc_source_dispatch
    at linc-source.c line 56
  • #32 g_main_dispatch
    at gmain.c line 1617
  • #33 g_main_context_dispatch
    at gmain.c line 2161
  • #34 g_main_context_iterate
    at gmain.c line 2242
  • #35 g_main_context_iteration
    at gmain.c line 2301
  • #36 linc_main_iteration
    at linc.c line 125
  • #37 giop_recv_buffer_get
    at giop-recv-buffer.c line 667
  • #38 orbit_small_demarshal
    at orbit-small.c line 396
  • #39 ORBit_small_invoke_stub
    at orbit-small.c line 607
  • #40 ConfigDatabase2_lookup_with_schema_name
Comment 1 Alexander Larsson 2002-03-14 01:53:25 UTC 
Seems it can't activate the metafile factory. This is a
bonobo-activation problem or more probably a busted installation in
some way.
Comment 2 Luis Villa 2002-03-30 19:26:39 UTC 
Should we close, then, Alex?
Comment 3 Dave Bordoley [Not Reading Bug Mail] 2002-04-08 19:09:09 UTC 
I'm going to close this, I did a fresh redhat 7.2 build using the
gnome rpms from today and startx worked fine and dandy.
Comment 4 John Fleck 2002-10-24 01:37:20 UTC 
*** Bug 96660 has been marked as a duplicate of this bug. ***
Comment 5 John Fleck 2002-10-24 01:38:07 UTC 
I'm reopening this, since we have a dup in Nautilus 2.0.7.
Comment 6 Luis Villa 2002-10-24 14:08:47 UTC 
And marking NEW and high/critical.
Comment 7 John Fleck 2002-10-26 03:02:11 UTC 
*** Bug 96799 has been marked as a duplicate of this bug. ***
Comment 8 Vincent Untz 2002-10-26 17:29:46 UTC 
*** Bug 96914 has been marked as a duplicate of this bug. ***
Comment 9 Elijah Newren 2002-10-29 17:09:05 UTC 
*** Bug 97135 has been marked as a duplicate of this bug. ***
Comment 10 Elijah Newren 2002-10-29 17:10:58 UTC 
*** Bug 97131 has been marked as a duplicate of this bug. ***
Comment 11 Elijah Newren 2002-10-31 02:02:24 UTC 
*** Bug 97263 has been marked as a duplicate of this bug. ***
Comment 12 Luis Villa 2002-11-07 19:45:19 UTC 
Apologies for the spam- I'm removing 'bugsquad' from some keywords via the web
interface. This is a one-time only thing before I re-add bugsquad via the SQL
interface, which will generate no mail. Apologies again.
Comment 13 Aschwin van der Woude 2002-11-14 12:53:06 UTC 
*** Bug 98449 has been marked as a duplicate of this bug. ***
Comment 14 Elijah Newren 2003-06-09 05:04:30 UTC 
*** Bug 114318 has been marked as a duplicate of this bug. ***
Comment 15 Elijah Newren 2003-06-09 05:05:40 UTC 
The most recent duplicate was in version 2.2.x of nautilus.  So I'm
setting the version to 2.2.x and updating the keyword
GNOMEVER2.0->GNOMEVER2.2
Comment 16 John Fleck 2003-07-24 22:34:32 UTC 
*** Bug 118187 has been marked as a duplicate of this bug. ***
Comment 17 John Fleck 2003-07-24 22:37:05 UTC 
Now spotted in the wild in 2.3.x....
Comment 18 John Fleck 2003-08-04 13:56:40 UTC 
*** Bug 119041 has been marked as a duplicate of this bug. ***
Comment 19 Elijah Newren 2003-08-19 16:21:15 UTC 
*** Bug 120192 has been marked as a duplicate of this bug. ***
Comment 20 Elijah Newren 2003-08-19 16:22:13 UTC 
The latest duplicate was also in GNOME 2.3 (and specifically, with
nautilus 2.3.8).  I'm changing version from 2.2.x to 2.3.x to reflect
this.
Comment 21 Elijah Newren 2003-08-26 19:46:56 UTC 
*** Bug 120707 has been marked as a duplicate of this bug. ***
Comment 22 Elijah Newren 2003-08-26 19:50:12 UTC 
The latest duplicate was from the same reporter as bug 119041, but
this time he's using nautilus-2.3.9.  I'm updating the target
milestone to 2.4.0 and adding the TARGET2.4.0 keyword.
Comment 23 mwehner 2003-09-18 03:10:36 UTC 
I looked into this and found out a few things:

Although all of the dups crash in die_on_activation (it could not get
the metafile factory from bonobo), they are really two different bugs.
All traces before 2003-07-24 are startup issues (originating in
show_desktop_window etc.), all traces hereafter are shutdown issues
(originating in gnome_vfs_shutdown()).

The later ones are caused by gnome_vfs_shutdown() called in main()
after the application was unrefed from bonobo. The vfs_shutdown call
dispatches pending completion callbacks which in turn call
register_monitor stuff which ends up trying to get a metafile factory
from an already unregistered server. (I'm not into CORBA/Bonobo but
that's how I understand it). This is easily fixed: call
gnome_vfs_shutdown() before bonobo_object_unref(application) in
nautilus-main. I could reliably reproduce the shutdown crash and after
this change it went away.

The former stacks look like a bonobo_activation bug which was fixed
along the way (it wasn't reported for a year with 1 exception). Well,
at least I couldn't reproduce it and the startup seems sane with
respect to these issues.
Comment 24 mwehner 2003-09-18 03:12:50 UTC 
Created attachment 20028 [details] [review]
Proposed Patch
Comment 25 Alexander Larsson 2003-09-18 08:20:17 UTC 
And it doesn't seem like application shutdown will call any gnome vfs
things, so this looks fine. Commiting.
Comment 26 Luis Villa 2003-09-25 12:04:59 UTC 
*** Bug 122148 has been marked as a duplicate of this bug. ***
Comment 27 zeitgeist 2003-09-28 11:49:31 UTC 
That patch does not fix the problem. I just tried it out on a stock
2.4.0 nautilus and it doesn't fix a thing.
Comment 28 Martin Wehner 2004-09-20 12:56:51 UTC 
*** Bug 152975 has been marked as a duplicate of this bug. ***