Re: [g-a-devel] Problem with fatal critical warnings and a11y

[Bill Haneman <Bill Haneman Sun COM>]

Kjartan Maraas wrote:

> man, 09,.01.2006 kl. 14.48 +0000, skrev Bill Haneman:
>  
>
> > This suggests to me that something has changed in the g-type system or 
> > elsewhere, maybe in gnome-program?
> >
> >    
> Are you 100% sure that the original assert is safe? This problem makes
> it impossible to test an a11y enabled desktop from a CVS build since
> we've now made all critical warnings fatal to get people to fix those.
> Would be nice if we could meet that goal.
>
> When I look at atk_get_root() it's clear that it can return NULL, 
If it does, however, that is indeed a critical desktop error.

Bill

> which
> in turn will cause spi_application_new() to be called with a NULL arg,
> and that will cause the assert to trigger from what I can see. Looks
> like spi_accessible_construct() should handle that case?
>
> Cheers
> Kjartan
>
>  
>
> > Bill
> >
> > Kjartan Maraas wrote:
> >
> >    
> >
> > > man, 09,.01.2006 kl. 13.31 +0000, skrev Bill Haneman:
> > >
> > >
> > >      
> > >
> > > > Kjartan Maraas wrote
> > > >
> > > >   
> > > >
> > > >        
> > > >
> > > > > But this is gnome-session crashing so then we have to make sure
> > > > > gnome-session gets it right at least.
> > > > >
> > > > >
> > > > >     
> > > > >
> > > > >          
> > > > Kjartan:
> > > >
> > > > The bugzilla bug, and our version of the patch (which still has not been 
> > > > applied to CVS), is 155917.
> > > > http://bugzilla.gnome.org/show_bug.cgi?id=155917
> > > >
> > > > I don't quite understand why the atk-bridge is being loaded into your 
> > > > version of gnome-session, since AFAIK HEAD doesn't have code to do this 
> > > > yet (?).  In any case the code doing it on your system at the moment 
> > > > seems pretty broken.  Perhaps some gnome-session code was added without 
> > > > the bugzilla bug being updated, which accounts for the new behavior?
> > > >
> > > > Looks like the bugzilla patch should be applied to HEAD to fix this.
> > > >
> > > >   
> > > >
> > > >        
> > > I tried this now and it still fails in the same place for me:
> > >
> > > Valgrind backtrace with current HEAD at-spi and gnome-session + the
> > > patch from the report above.
> > >
> > > ==32417== Process terminating with default action of signal 4 (SIGILL)
> > > ==32417==  Illegal opcode at address 0x4B043A8
> > > ==32417==    at 0x4B043A8: g_logv (gmessages.c:493)
> > > ==32417==    by 0x4B04427: g_log (gmessages.c:517)
> > > ==32417==    by 0x4B044E4: g_assert_warning (gmessages.c:552)
> > > ==32417==    by 0x4F1B398: spi_accessible_construct (accessible.c:495)
> > > ==32417==    by 0x4F1C5E2: spi_application_new (application.c:293)
> > > ==32417==    by 0x4F48257: atk_bridge_init (bridge.c:180)
> > > ==32417==    by 0x4F48E02: gnome_accessibility_module_init
> > > (bridge.c:533)
> > > ==32417==    by 0x416A409: accessibility_invoke_module
> > > (gnome-program.c:1601)
> > > ==32417==    by 0x416A511: accessibility_invoke (gnome-program.c:1632)
> > > ==32417==    by 0x416A64F: accessibility_init (gnome-program.c:1673)
> > > ==32417==    by 0x416A762: gnome_program_postinit (gnome-program.c:1713)
> > > ==32417==    by 0x416ABFC: gnome_program_init_common
> > > (gnome-program.c:1928)
> > > ==32417==    by 0x416AC34: gnome_program_initv (gnome-program.c:1965)
> > > ==32417==    by 0x416A7B6: gnome_program_init (gnome-program.c:1746)
> > > ==32417==    by 0x805387D: main (main.c:406)
> > > --32417-- Caught __NR_exit; running __libc_freeres()
> > >
> > > Cheers
> > > Kjartan
> > >
> > >
> > >
> > >
> > >      
>
> _______________________________________________
> Gnome-accessibility-devel mailing list
> Gnome-accessibility-devel gnome org
> http://mail.gnome.org/mailman/listinfo/gnome-accessibility-devel
>