Enlightenment Error

[ssweeny]

Every so often when using the 2008.8 image I get an Enlightenment window pop up with the title "Enlightenment Error" and the text:

This is very bad. Enlightenment SIGABRT'd.

And the option to recover (which restarts X) or exit (which quits X).

I can't seem to find a pattern that triggers it reliably.

[ssweeny]

copy of /var/tm/x.log

[raster]

installing gdb and getting a backtrace when the white box is up would be helpful... other than that i'd need a RELIABLE way to reproduce... i have yet to see the white box of death :(

[shashank]

I'm using FDOM (version 20080913)
and I'm getting this *white box of death* too

I'm guessing that it occurs whenever a qtopia application is launched. (be it Dialer, Calender...)
But it's not consistent. Sometimes it launches without errors also.

[john_lee]

I have seen this on my neo as well. will look into this later.

[john_lee]

don't run qtopia in Xsession, start illume first, attach gdb, then run qtopia.

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x408a6ba0 (LWP 2456)]
0x407db684 in _int_malloc () from /lib/libc.so.6
(gdb) bt
#0  0x407db684 in _int_malloc () from /lib/libc.so.6
#1  0x407dc3ac in malloc () from /lib/libc.so.6
#2  0x4028057c in _evas_common_rgba_image_surface_alloc (ie=<value optimized out>, w=0, h=1006291193) at evas_image_main.c:164
#3  0x40253290 in _evas_cache_image_entry_surface_alloc (cache=0x156a60, ie=0x28f9f0, w=<value optimized out>, h=<value optimized out>)
    at evas_cache_image.c:250
#4  0x40253484 in evas_cache_image_surface_alloc (im=0x28f9f0, w=1006632960, h=1076180100) at evas_cache_image.c:795
#5  0x40a590e8 in eng_font_draw (data=<value optimized out>, context=0x1ac1b8, surface=0x3c97b8, font=0x1b14e8, x=69, y=157, w=98, h=32, ow=98, oh=32, 
    text=0x3ca13c "_ignore_") at evas_engine.c:771
#6  0x4023a954 in evas_object_text_render (obj=0x1ade40, output=0x1ac0e8, context=0x1ac1b8, surface=0x3c97b8, x=0, y=0) at evas_object_text.c:1534
#7  0x4024f3d4 in evas_render_updates_internal (e=0x1abf00, make_updates=1 '\001', do_draw=1 '\001') at evas_render.c:517
#8  0x4024f814 in evas_render_updates (e=0x1abf00) at evas_render.c:626
#9  0x4002f95c in _ecore_evas_x_render (ee=0x1abde0) at ecore_evas_x.c:330
#10 0x40031c14 in _ecore_evas_x_idle_enter (data=<value optimized out>) at ecore_evas_x.c:1320
#11 0x406422fc in _ecore_idle_enterer_call () at ecore_idle_enterer.c:101
#12 0x406461e0 in _ecore_main_loop_iterate_internal (once_only=0) at ecore_main.c:693
#13 0x40646260 in ecore_main_loop_begin () at ecore_main.c:87
#14 0x0002a3dc in main ()

[john_lee]

raster suggested I give software_x engine a try instead of software_16_x. The SIGSEGV/SIGABRT is gone under software_x, which means this might be a bug in software_16_x.

[gromgull]

(no comment just CC me)

[Benson]

Any news on this? Getting the same SIGABRT message on om-testing! I get it even when doing nothing or entering a menu in the illume settings...

{{{root@om-gta02:~# /etc/X11/Xserver
*** glibc detected *** enlightenment: realloc(): invalid next size: 0x003be640 ***}}}

Switching to software_x engine solves the problem, but it's much slower than software_16_x. Is there no way you can reproduce it? Might

[john_lee]

upstream mainly works on software_x and illume-theme-illume, so this combination is known to work. if you prefer to use upstream illume theme, please switch to software_x engine. since asu theme seems to work with software_16, fixing this is low priority at the moment.

[rmn30]

For those of you (like me) who didn't know how to switch 'engine' you can do it in the settings dialog (wrench) in illume.

[WRXIzumi]

Someone on the community mailing list did a trace of the sigabrt and found that it was dying in the drop shadow module. I unloaded that module and have found there are a lot less errors when using SOFTWARE_16 engine. I went through all the items on the wrench settings application and found a few more that caused the error when starting (e.g. Themes). I unloaded all that caused an error. Now the only time I will get an error is occasionally when I am switching from the home screen to an already running applications.

[TimoJyrinki]

How to actually unload drop shadow module remotely? (or switch the engine). I cannot get enlightenment_remote to work via ssh. I'd like to test the various options, though.

[kmeaw]

root@om-gta02:/usr/share/openmoko/scenarios# export DISPLAY=:0
root@om-gta02:/usr/share/openmoko/scenarios# grep E_IPC_SO /proc/*/environ
/proc/725/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/748/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/760/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/780/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/785/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/798/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/799/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
/proc/805/environ:E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
root@om-gta02:/usr/share/openmoko/scenarios# export E_IPC_SOCKET=/tmp/enlightenment-root/disp-:0.0-725
root@om-gta02:/usr/share/openmoko/scenarios# enlightenment_remote -module-disable dropshadow
root@om-gta02:/usr/share/openmoko/scenarios# enlightenment_remote -module-unload dropshadow