Re: [Gnash-commit] /srv/bzr/gnash/trunk r10779: Major reorganization of the messy half of Gnash to match the neater half

[strk]

On Tue, Apr 07, 2009 at 09:08:52PM +0200, Benjamin Wolsey wrote:
> Am Dienstag, den 07.04.2009, 20:56 +0200 schrieb strk:
> > On Tue, Apr 07, 2009 at 06:32:11PM +0200, Benjamin Wolsey wrote:
> > 
> > >     revno: 10771.1.12
> > >     committer: Benjamin Wolsey <address@hidden>
> > >     branch nick: work
> > >     timestamp: Mon 2009-04-06 18:21:49 +0200
> > >     message:
> > >       Merge from trunk.
> > >       
> > >       Drop advance() code, as it messes up the morph timing in relation to
> > >       MovieClip's advance. MorphShape isn't a LiveChar anymore; morphing 
> > > is
> > >       done on display() again.
> > 
> > Wouldn't this mess up invalidated bounds computation ?
> > Reminder: we have very few testcases for inv bound computation,
> > altought the supporting code for testing is there.
> > Also, if we skip rendering (gprocessor, -r0 or some heursitic
> > to catch-up with frame rate) wouldn't we miss updating the boundaries ?
> > 
> 
> It was actually an attempt to do bounds properly. Up to now all morphs
> returned the combined bounds of the two shapes in getBounds(), which is
> enough to make sure everything is displayed, but inefficient and not
> really what getBounds is for. The idea to use advance() had other
> problems, so I reverted it to more or less what it was before. Would a
> proper add_invalidated_bounds method be the real fix here?

I'm afraid add_invalidated_bounds is not even called if set_invalidated
isn't. I don't have time to check this now, but I suggest you try to
produce a focused testcase for invalidated bounds and morphs.

Also, note that bounds of a MovieClip (for example containing a 
morph) can be queried by ActionScript, which happens before ::display.


--strk; 

 Free GIS & Flash consultant/developer      ()  ASCII Ribbon Campaign
 http://foo.keybit.net/~strk/services.html  /\  Keep it simple!