Re: [Gnash-commit] gnash/server shape.cpp

[strk]

On Sat, Mar 03, 2007 at 05:54:25PM +0100, Udo Giacomozzi wrote:
> Hello strk,
> 
> Saturday, March 3, 2007, 4:58:25 PM, you wrote:
> s> The problem is that even shapes with *NO* fill , but only lines
> s> should be checked for containing a point.
> s> See DrawingApiTest.swf under misc-ming.all
> 
> It should only be checked if the point is on the (arbitrary sized)
> outline of a path. With other words, the test for m_fill0==0 *is*
> relevant.

> However, a function path::point_test() is completely nonsense as a
> path does *not* to be closed. A straight horizontal line (only one
> edge) is a fully valid path, regardless of m_fill0 and m_fill1.
> 

Argh, you're right !
So, shape_character_def should find closed paths itself... being
closed path sequences with *all*  paths having either a left a fill specified.
Sounds as we'd need more testcases for this.


--strk;