On Fri, Dec 08, 2006 at 02:25:10PM +0100, strk wrote: > I'm ready to add invalidated bounds testing to our > testsuite, but it seems I can't handle to understand the > mechanism of it. > > What I would expect is that: > > 0) get_invalidated_bounds is misnamed, should be > expand_to_invalidated_bounds() as it actually > expands the given rect to enclose the invalidated bounds. > > 1) Before *any* ::advance() is called, the > get_invalidated_bounds() should add NOTHING > to the provided bounds.
I correct myself, since it isn't clear what updates the "invalidated bounds" state, can't only talk about ::advance() but about anything that somehow changes the internal state (think about mouse events). > 2) After *first* call to advance(), the get_invalidated_bounds() > should turn the bounds into the WORLD bounds (everything changed) Same as above, first call to anything that changes the internal state. > 3) From *second* call on, the get_invalidated_bounds() should > only add the boundaries of things that actually changed > since when ? Last call to ::display ? Last call to > get_invalidated_bounds () ?? It seems this is the central question. --strk; _______________________________________________ Gnash-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnash-dev
