[Q] 'Quit Squeak without saving?' dialog

Ned Konz ned at bike-nomad.com
Mon May 19 17:43:26 UTC 2003


On Monday 19 May 2003 10:32 am, Andreas Raab wrote:
> I don't really know what to do here. Perhaps we might hand Squeak
> some extra event which is recognized and 'acknowledged' so that we
> may have a sequence like:
>
> * User indicates to 'close Squeak'
> * VM posts an event to Squeak telling it to close
>   - Squeak acknowledges the receipt of the event by some primitive
> invokation
>
> The reason why I'd like to acknowledge the event is that if the
> system is somehow locked up one could click *again* on this button
> - this time the VM could notify the user that the last request to
> close wasn't received so that we can either try again or 'just
> quit'.

Well, perhaps it would make sense to have a semaphore that would be 
signaled on the attempt-to-close event.

That way we could have a very high priority Process that waits on the 
semaphore and can gain control upon that signal.

And can then call a primitive to allow closing.

This way we're clear of the event processing loop, and can set up any 
kind of policy we want at the Squeak side.

For instance, I'd like to be able to suspend processes that were hung 
or deeply recursed and drop into a debugger.

-- 
Ned Konz
http://bike-nomad.com
GPG key ID: BEEA7EFE



More information about the Squeak-dev mailing list