GNOME Bugzilla – Bug 100709
nautilus crashes on FreeBSD in thumbnailing thread due to 64K stack size limit (Was: Crash with [_thread_kern_sched_state_unlock], caused by FreeBSD ABI change)
Last modified: 2009-08-15 18:40:50 UTC
Package: nautilus Severity: normal Version: 2.0.8 Synopsis: Specific JPEG crashes nautilus on FreeBSD Bugzilla-Product: nautilus Bugzilla-Component: Thumbnails BugBuddy-GnomeVersion: 2.0 (2.0.6) Description: Description of Problem: I have a specific image taken from my Canon Powershot S110 camera which causes NAutilus to crash whenever it attempts to thumbnail it or display it. I can't any other application that does it with this same picture. I can't reproduce it with any other picture from my camera. It is this one specific picture. Steps to reproduce the problem: 1. Get http://www.zombie.org/img_0349.jpg 2. Put it in some directory 3. View that directory in Nautilus and watch it crash. Actual Results: Nautilus segfaults. Expected Results: The image should be treated as any other image and display or thumbnail properly, How often does this happen? 100% repeatability. Additional Information: Debugging Information: Backtrace was generated from '/usr/X11R6/bin/nautilus' (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...(no debugging symbols found)... (no debugging symbols found)...[Switching to Process 76277, Thread 1] 0x28c959f3 in poll () from /usr/lib/libc.so.5
+ Trace 31471
Thread 2 (Process 76277, Thread 2)
------- Bug moved to this database by unknown@bugzilla.gnome.org 2002-12-08 22:45 ------- The original reporter (smkelly@zombie.org) of this bug does not have an account here. Reassigning to the exporter, unknown@bugzilla.gnome.org. Reassigning to the default owner of the component, nautilus-maint@bugzilla.gnome.org.
Displays fine here with 2.0.7, debian powerpc Did you compile with any unusual flags? -O>2? If you could recompile with debugging enabled, that might yield a better stack trace. Leaving unconfirmed for other folks to try to reproduce with the image.
I think this was fixed by setting the thread stack size on FreeBSD, can you try the CVS HEAD version?
*** Bug 102219 has been marked as a duplicate of this bug. ***
It turns out this was an accidental FreeBSD ABI change during the import of gcc-3.2.1. div(), a structure returning function, was the culprit.
Ok, closing this bug since it isn't caused by Nautilus.
*** Bug 102367 has been marked as a duplicate of this bug. ***
*** Bug 103052 has been marked as a duplicate of this bug. ***
*** Bug 104956 has been marked as a duplicate of this bug. ***
*** Bug 105551 has been marked as a duplicate of this bug. ***
I apologize for the spam, but in talking to David Kennedy on #bugs we decided that it might be useful for me to add a comment to a lot of bugs that have nearly identical stack traces that aren't very useful. (It took me a while because I gave up on several bugzilla queries that simply hung). Anyway, I may have caught some bugs with this query that aren't relevant, so I apologize again if this bug is one of those. If not: The following comment from 100709 may be relevant: "It turns out this was an accidental FreeBSD ABI change during the import of gcc-3.2.1. div(), a structure returning function, was the culprit."
*** Bug 106683 has been marked as a duplicate of this bug. ***
*** Bug 106244 has been marked as a duplicate of this bug. ***
*** Bug 113875 has been marked as a duplicate of this bug. ***
*** Bug 113825 has been marked as a duplicate of this bug. ***
*** Bug 120601 has been marked as a duplicate of this bug. ***
*** Bug 119757 has been marked as a duplicate of this bug. ***
*** Bug 119572 has been marked as a duplicate of this bug. ***
*** Bug 123544 has been marked as a duplicate of this bug. ***
*** Bug 123650 has been marked as a duplicate of this bug. ***
*** Bug 124290 has been marked as a duplicate of this bug. ***
*** Bug 124281 has been marked as a duplicate of this bug. ***
*** Bug 124916 has been marked as a duplicate of this bug. ***
*** Bug 124990 has been marked as a duplicate of this bug. ***
*** Bug 126586 has been marked as a duplicate of this bug. ***
*** Bug 126902 has been marked as a duplicate of this bug. ***
*** Bug 128631 has been marked as a duplicate of this bug. ***
*** Bug 130124 has been marked as a duplicate of this bug. ***
*** Bug 130670 has been marked as a duplicate of this bug. ***
*** Bug 130751 has been marked as a duplicate of this bug. ***
*** Bug 131243 has been marked as a duplicate of this bug. ***
*** Bug 131241 has been marked as a duplicate of this bug. ***
*** Bug 131949 has been marked as a duplicate of this bug. ***
*** Bug 131990 has been marked as a duplicate of this bug. ***
*** Bug 132779 has been marked as a duplicate of this bug. ***
*** Bug 132931 has been marked as a duplicate of this bug. ***
*** Bug 133325 has been marked as a duplicate of this bug. ***
I'm pretty sure that this bug has nothing to do with the FreeBSD ABI changes smkelly@zombie.org refers to above ( http://tinyurl.com/2hkzr ?). Alex Larson's comment above refers to bug 99917 (a nautilus-2.1 era fix which increased nautilus's default thread stack size limit so it wouldn't crash on FreeBSD). So this should have been fixed in nautilus-2.1 and anyone using a nautilus >= 2.2 should be unaffected. It didn't have anything to do with the "_thread_kern_sched_state_unlock" in the stacktrace and mentioned in the Summary, but was a result of the new thread created to do thumbnailing hitting the stack size limit (i'm assuming the eel_gdk_pixbuf_scale_down() in the stacktrace). Many of the duplicates on this bug are wrong. I'm changing the Summary to reflect this.
*** Bug 163113 has been marked as a duplicate of this bug. ***