destroystack wrong behavior...

Jim Bufalini jim at visitrieve.com
Thu May 20 08:15:19 EDT 2010


Thierry wrote:

> Another point: this reminds me one thread on this list about someone
> who couldn't start an external. I guess/feel that his problem could be
> that
> his stack was still in memory, and when he restarted it, the normal
> process of
> loading the external didn't work because the stack was already in
> memory.....

There is an issue here. The issue involves destroyStack and externals and
memory usage both in the IDE and in standalones on Windows (probably on Mac
also, but I was never able to positively confirm it on Mac). This issue is
if you start a stack that uses an external that leaks memory, like
revBrowser (revBrowser itself doesn't leak, but the browser control that it
uses does). The known way to release the leaked memory is to close the
external. The way to release the external is to destroy the stack that
opened the external.

However, after destroying the stack (and there is some question as to
whether the stack was actually destroyed or not), the fact is that the
memory is not released. This would indicate that the stack was not destroyed
and the external is still in memory. In any case, the only way to release
the memory is to close the IDE or standalone itself.

My understanding is that this is a big architectural change that will be
addressed as part of other architectural changes in the future.

Aloha from Hawaii,

Jim Bufalini








More information about the use-livecode mailing list