fxruby users - Jun 2006 - Create fails strangely => Fixed resource limits in FXRuby/Fox?

Hi,

I''ve got a problem with my big and sprawling app.  The behaviour
appears to be related to how many widgets i have in use. (I''m creating
100''s and using show/hide at appropriate times.)

What happens when i create a new window is code like this:

    Control.new( owner , ... )
    puts 1
    control.create
    puts 2

will work fine, but at some point will display 1, but not 2, and
the loop it was in will terminate there and then, but the app.
will keep on running, destined to crash oddly later on.

If I put all the creates after all the Control.new calls I get
exactly the same behaviour.  It is completely repeatable.

The create call doesn''t raise any exception.

Can someone shed some light on what''s going on, or tell me
how I can find out more?

I''m using Fox 1.4.2, Ruby 1.8.2.

Thanks you,
Martin.

On Jun 7, 2006, at 5:22 AM, Martin Portman wrote:
> I''ve got a problem with my big and sprawling app.  The behaviour
> appears to be related to how many widgets i have in use. (I''m
creating
> 100''s and using show/hide at appropriate times.)
>
> What happens when i create a new window is code like this:
>
>     Control.new( owner , ... )
>     puts 1
>     control.create
>     puts 2
>
> will work fine, but at some point will display 1, but not 2, and
> the loop it was in will terminate there and then, but the app.
> will keep on running, destined to crash oddly later on.
To address the subject line first: we don''t have any fixed resource 
limits in FOX. The problem that you''re describing, however, is almost 
certainly a GC-related bug in FXRuby. A possible workaround is to add 
the line:

	GC.disable

somewhere near the beginning of your program''s execution. That
won''t
work around all GC-related bugs, but it often does the trick. And I 
stress that it is just a work-around; we need to identify the actual 
underlying bug.
> If I put all the creates after all the Control.new calls I get
> exactly the same behaviour.  It is completely repeatable.
Repeatable bugs are good things. ;)

Can you send me a copy of the program and some instructions on how to 
reproduce the crash? Even better, if you hit on a way to reproduce the 
problem in a pared-down version of the application, that will reduce 
the number of distractions during debugging.