Bug#739194: blender: Segfaults at startup on armhf

[Gunnar Wolf]

found 739194 2.69-4
thanks

I did a clean Sid reinstall, and the bug is still present in the
current version.

Blender now attempts to generate a file to aid debugging, although it
crashed too early for this file ot be of use:

    $ blender -noaudio -b
    Color management: using fallback mode for management
    Writing: /tmp/blender.crash.txt
    Segmentation fault
    $ cat /tmp/blender.crash.txt 
    # Blender 2.69 (sub 0), Revision: unknown

    # backtrace
    $

The output from gdb is quite different, so I'm including it again:

    $ gdb --args blender -noaudio -b
    GNU gdb (GDB) 7.6.2 (Debian 7.6.2-1)
    Copyright (C) 2013 Free Software Foundation, Inc.
    License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
    This is free software: you are free to change and redistribute it.
    There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
    and "show warranty" for details.
    This GDB was configured as "arm-linux-gnueabihf".
    For bug reporting instructions, please see:
    <http://www.gnu.org/software/gdb/bugs/>...
    Reading symbols from /usr/bin/blender...Reading symbols from /usr/lib/debug/.build-id/98/59f5bff4ae8dd86cf6c93855cbfde406dd6d98.debug...done.
    done.
    (gdb) run
    Starting program: /usr/bin/blender -noaudio -b
    [Thread debugging using libthread_db enabled]
    Using host libthread_db library "/lib/arm-linux-gnueabihf/libthread_db.so.1".
    [New Thread 0x306fd240 (LWP 3002)]
    [New Thread 0x30efd240 (LWP 3003)]
    [New Thread 0x316fd240 (LWP 3004)]
    [New Thread 0x31efd240 (LWP 3005)]
    [New Thread 0x326fd240 (LWP 3006)]
    Color management: using fallback mode for management
    
    Program received signal SIGSEGV, Segmentation fault.
    0x2acd8cce in PyErr_SetObject ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    (gdb) bt
    #0  0x2acd8cce in PyErr_SetObject ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    #1  0x2acd8c9a in PyErr_Format ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    #2  0x2ac8262c in PyType_Ready ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    #3  0x2ac55052 in _PyExc_Init ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    #4  0x2ace95e2 in _Py_InitializeEx_Private ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    #5  0x00697898 in BPY_python_start (argc=3, argv=0x7efff8d4)
        at /build/blender-dPxPUD/blender-2.69/source/blender/python/intern/bpy_interface.c:274
    #6  0x0044a1ca in WM_init (C=C at entry=0x1d83220, argc=argc at entry=3, 
        argv=argv at entry=0x7efff8d4)
        at /build/blender-dPxPUD/blender-2.69/source/blender/windowmanager/intern/wm_init_exit.c:176
    #7  0x0042e4de in main (argc=3, argv=0x7efff8d4)
        at /build/blender-dPxPUD/blender-2.69/source/creator/creator.c:1597
    (gdb) 

The segfault no longer follows an error in ./Python/errors.c (but it
still happens in PyErr_SetObject).

I am a real gdb newbie, so I'm just repeating what I've been told to
do ;-) But a full backtrace gives:

    (gdb) bt full
    #0  0x2acd8cce in PyErr_SetObject ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    No symbol table info available.
    #1  0x2acd8c9a in PyErr_Format ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    No symbol table info available.
    #2  0x2ac8262c in PyType_Ready ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    No symbol table info available.
    #3  0x2ac55052 in _PyExc_Init ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    No symbol table info available.
    #4  0x2ace95e2 in _Py_InitializeEx_Private ()
       from /usr/lib/arm-linux-gnueabihf/libpython3.3m.so.1.0
    No symbol table info available.
    #5  0x00697898 in BPY_python_start (argc=3, argv=0x7efff8d4)
        at /build/blender-dPxPUD/blender-2.69/source/blender/python/intern/bpy_interface.c:274
            py_tstate = 0x0
            py_path_bundle = 0x0
            program_path_wchar = L"/usr/bin/blender", '\000' <repeats 1007 times>
    #6  0x0044a1ca in WM_init (C=C at entry=0x1d83220, argc=argc at entry=3, 
        argv=argv at entry=0x7efff8d4)
        at /build/blender-dPxPUD/blender-2.69/source/blender/windowmanager/intern/wm_init_exit.c:176
    No locals.
    #7  0x0042e4de in main (argc=3, argv=0x7efff8d4)
        at /build/blender-dPxPUD/blender-2.69/source/creator/creator.c:1597
            C = 0x1d83220
            syshandle = 0x1d8a338
            ba = 0x1d8a840
    (gdb) 

#5 looks very odd to me. Why would "/usr/bin/blender" be repeated 1007
times? That does look like a stack overflow...