Re[2]: [Gnash-commit] gnash ChangeLog server/parser/shape_character_d...

[Udo Giacomozzi]

Hello strk,

Thursday, November 8, 2007, 6:23:31 PM, you wrote:
s> Also, keep the multiplication/addition if possible,
s> your coordinates could be 23,12, becoming 92,148.

Sorry, this was a accidental commit - reverted.

Note the testsuite currently reports a few failures:

FAIL: DrawingApiTestRunner: !inv3.hitTest((20*4), 100 + (10*4), true) 
[./DrawingApiTest.as:297]
FAIL: DrawingApiTestRunner: inv8.hitTest(200 + (25*2), 0 + (15*2), true) 
[./DrawingApiTest.as:499]
FAIL: DrawingApiTestRunner: inv8.hitTest(200 + (35*2), 0 + (15*2), true) 
[./DrawingApiTest.as:500]
FAIL: DrawingApiTestRunner: !inv8.hitTest(200 + (25*2), 0 + (25*2), true) 
[./DrawingApiTest.as:503]
FAIL: DrawingApiTestRunner: inv8.hitTest(200 + (35*2), 0 + (25*2), true) 
[./DrawingApiTest.as:504]
FAIL: DrawingApiTestRunner: !inv8.hitTest(200 + (20*2), 0 + (25*2), true) 
[./DrawingApiTest.as:510]
FAIL: DrawingApiTestRunner: AGG_RGB24 pix:260,160 exp:64,255,64,255 
obt:63,252,63,255 tol:2 [DrawingApiTestRunner.cpp:653]
FAIL: DrawingApiTestRunner: AGG_RGB24 pix:309,160 exp:64,255,64,255 
obt:63,252,63,255 tol:2 [DrawingApiTestRunner.cpp:690]

This is not the fault of the point_test algorithm because these shapes
have unclosed paths (which are implicitely closed in display() only).

We wanted to do implicit path-closing in the point test implementation
but the problem is that I don't know if a path is open or not.
Remember, a path is allowed to be open as long some other path(s)
close the shape as a whole.

This is easier for dynamic shapes because each subshape has it's own,
closed path (since it does not use two fill styles), but the
point_test algorithm does not distinguish between dynamic and
SWF-defined shapes.


Udo