Re: [g-a-devel] Dangling pointers when an Accessible * window is killed

[Li Yuan <Li Yuan Sun COM>]

Quiring, Sam wrote:
> Greetings,
>  
> We were running an experiment where our app has an Accessible * 
> pointer to a window. 
Your app acts as an AT, right?

> If the window is closed and the app then tries to use the Accessible 
> * pointer, the at-spi system gets a segfault.
How did you get the pointer? Did you ref it? If so there could be a bug 
in GAIL.
> Is there a good way to test the validity of an Accessible * pointer so 
> we can avoid the segfault?  Or is segfault the normal at-spi way of 
> telling the accessible application that a window has disappeared?
The object should not be finalized if you still have reference.

Regards,
Li