Re: [Gnash-commit] gnash ChangeLog server/character.cpp server/cha...

[Sandro Santilli]

On Tue, Feb 13, 2007 at 02:20:16PM +0100, Udo Giacomozzi wrote:
> Hello Sandro,
> 
> Tuesday, February 13, 2007, 2:09:03 PM, you wrote:
> SS> It will be reset right *before* modifications, by set_invalidated()
> SS> I understood that lately.
> 
> SS> You can see I added some notes in the comment about this, it wasn't clear 
> before.
> 
> Ok, I just reviewed some commits to understand what happened on the
> invalidated bounds detection. I think there's been a slight
> regression. Still can't figure out where's the problem but I'm working
> on a solution that will *greatly* improve rendering performance -
> probably for all render handlers.

Excellent. You might be interested in a new support for invalidated
bounds detection testing I added to MovieTester class.
See simple_loop_tester.cpp for an example of use. An excerpt below:

        invalidated = tester.getInvalidatedBounds();
        check( invalidated.isNull() );

        tester.advance():

        invalidated = tester.getInvalidatedBounds();
        check( invalidated.intersects(Bounds(0, 0, 60, 60)) );

        tester.advance():

        invalidated = tester.getInvalidatedBounds();
        check( invalidated.intersects(Bounds(60, 0, 120, 60)) );

simple_loop_tester.cpp is currently the only file actually using the invalidated
bounds tests. Maybe you can add some more *after* verifying they are correct
to other tests ? 

--strk;