Blox-Gtk segmentation fault

Hello,
While browsing classes, i get a segmentation fault.
It happens everytime i browse, but not at a particular time.
I can't or reproduce either fix it because i don't know when and where it  
happens.

There's not backtrace for segmentation fault, also when i run gst under  
gdb i get nothing useful.

Does this happen to you too? Any hints?

Bye.

--
http://lethalman.blogspot.com - Thoughts about computer technologies


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

Luca Bruno aka Lethalman wrote

Hello,
While browsing classes, i get a segmentation fault.
It happens everytime i browse, but not at a particular time.
I can't or reproduce either fix it because i don't know when and where it  
happens.

There's not backtrace for segmentation fault, also when i run gst under  
gdb i get nothing useful.

Does this happen to you too? Any hints?

Bye.

Nothing about? No one get Segmentation fault when using BloxGTK?

--
http://lethalman.blogspot.com - Thoughts about computer technologies

Luca Bruno aka Lethalman wrote:
Sorry, I haven't run BloxGTK for a while (hardly ever, in fact). I have
done quite a lot with Gtk, though.

Is it definitely a SEGFAULT, not a Gtk fatal error or an Abort from
Smalltalk?

Mike


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

> >>Bye.
> >
> > Nothing about? No one get Segmentation fault when using BloxGTK?
>

I've tried latest version of GST and BloxGTK and I've got abort
(SIGABRT), not Gtk fatal error nor SEGFAULT.

I've got abort every time I tried to play with BloxGTK browser,
sooner or later. Unfortunately, I cannot isolate the problem,
it seems, that abort comes randomly - sometimes after several seconds,
sometimes after several minutes.

The interesting thing is that no of pure GTK examples crashes...

Jan



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

Jan Vrany wrote:
>From the randomness I would guess that the garbage collector is
collecting something it shouldn't or that a reference is kept to
something that has been destroyed by Gtk.

> The interesting thing is that no of pure GTK examples crashes...

If they aren't running Smalltalk code, I'm not sure that GC will ever
occur (or need to).

Mike


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

In data 05 ottobre 2006 alle ore 15:49:03, Mike Anderson  
<[hidden email]> ha scritto:

Sometimes i get just Segmentation fault, sometimes i get an error from Gtk  
saying something about gslice (just a row) followed by Segmentation fault,  
but seems nothing useful though.



--
http://lethalman.blogspot.com - Thoughts about computer technologies


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

In data 05 ottobre 2006 alle ore 17:09:44, Jan Vrany  
<[hidden email]> ha scritto:

That's the same for me, but Segmentation fault instead...

> The interesting thing is that no of pure GTK examples crashes...
>

The example_tree.st crashes, but changing the model creation by using  
#new:varargs: fixes the problem.



--
http://lethalman.blogspot.com - Thoughts about computer technologies


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

> The example_tree.st crashes, but changing the model creation by using
> #new:varargs: fixes the problem.

Hum, works for me... patch?


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

In data 06 ottobre 2006 alle ore 07:48:34, Paolo Bonzini  
<[hidden email]> ha scritto:

>
>> The example_tree.st crashes, but changing the model creation by using  
>> #new:varargs: fixes the problem.
>
> Hum, works for me... patch?

In #open change:

typs := (CArrayCType elementType: CIntType numberOfElements: 3) new.
typs
        at: 0 put: (GValue typeFromName: 'gint');
       at: 1 put: (GValue typeFromName: 'gchararray');
        at: 2 put: (GValue typeFromName: 'gchararray').

tree := GtkTreeStore newv: 3 types: typs.

to:

tree := GtkTreeStore new: 3 varargs: {
        GValue gTypeInt.
        GValue gTypeString.
        GValue gTypeString}.

I can't make patches from Windows sorry.
GValue>>#typeFromName: seems to work fine, it's the C side that doesn't  
handle typs as well.

Bye.

--
http://lethalman.blogspot.com - Thoughts about computer technologies


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