Gtk Blox tip

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Gtk Blox tip

MrGwen
Hi,

Here is a simple tip if you've tried to launch gtk-blox more than 5 min ;)
without a crash you can launch like this :

G_SLICE=always-malloc gst-blox

Cheers,
Gwenael


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk
Reply | Threaded
Open this post in threaded view
|

Re: Gtk Blox tip

Paolo Bonzini-2
Gwenael Casaccio wrote:
> Hi,
>
> Here is a simple tip if you've tried to launch gtk-blox more than 5 min ;)
> without a crash you can launch like this :

Interesting.  I never could reproduce the failure, this gives a first
hint.  BTW, possibly configuring with --disable-generational-gc will
work around the valgrind bug.

Paolo



_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk
Reply | Threaded
Open this post in threaded view
|

Re: Gtk Blox tip

Paolo Bonzini-2
In reply to this post by MrGwen
Gwenael Casaccio wrote:
> Hi,
>
> Here is a simple tip if you've tried to launch gtk-blox more than 5 min ;)
> without a crash you can launch like this :
>
> G_SLICE=always-malloc gst-blox

Hum, maybe G_SLICE=debug-blocks will help debugging it?  The example in
the documentation sounds like a likely case...

"An example of a memory corruption scenario that cannot be reproduced
with G_SLICE=always-malloc, but will be caught by G_SLICE=debug-blocks
is as follows:

  void *slist = g_slist_alloc(); /* void* gives up type-safety */
  g_list_free (slist);           /* sizeof (GSList) != sizeof (GList) */

Paolo


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk
Reply | Threaded
Open this post in threaded view
|

Re: Gtk Blox tip

MrGwen
On Wednesday 25 February 2009 16:28:45 you wrote:

> Gwenael Casaccio wrote:
> > Hi,
> >
> > Here is a simple tip if you've tried to launch gtk-blox more than 5 min
> > ;) without a crash you can launch like this :
> >
> > G_SLICE=always-malloc gst-blox
>
> Hum, maybe G_SLICE=debug-blocks will help debugging it?  The example in
> the documentation sounds like a likely case...
>
> "An example of a memory corruption scenario that cannot be reproduced
> with G_SLICE=always-malloc, but will be caught by G_SLICE=debug-blocks
> is as follows:
>
>   void *slist = g_slist_alloc(); /* void* gives up type-safety */
>   g_list_free (slist);           /* sizeof (GSList) != sizeof (GList) */
>
> Paolo

When I try to debug gst with gdb I've got this segfault :

Program received signal SIGSEGV, Segmentation fault.
0x00007fb251f16f0b in _gst_dictionary_add (dictionaryOOP=0x2b203f3d2030,
associationOOP=0x2b203f3dc3d0) at dict.c:1877
1877          SET_ASSOCIATION_VALUE (associationOOP, value);

Gwenael


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk
Reply | Threaded
Open this post in threaded view
|

Re: Gtk Blox tip

Paolo Bonzini-2

> When I try to debug gst with gdb I've got this segfault :
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x00007fb251f16f0b in _gst_dictionary_add (dictionaryOOP=0x2b203f3d2030,
> associationOOP=0x2b203f3dc3d0) at dict.c:1877
> 1877          SET_ASSOCIATION_VALUE (associationOOP, value);

Oops, I should really document this somewhere (or maybe it is :-)
documented).  You need to do

  handle SIGSEGV noprint

at the beginning of the session, so that the SIGSEGV's are passed
properly to the generational GC.

This is done automatically if you're in the build tree (see .gdbinit
file), but if you debug an installed gst you need to do it manually.

But I think it's easier to use G_SLICE=debug-blocks and see what happens.

Oh, by the way, did you know that I have an "omnibrowser" branch in my
repository with OB-Core and OB-GTK already ported?

Paolo


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk