Re: About painting-stroke-10-t.svg

Hi Cyril and all,


	I've just analyzed this test and stepped through our stroking
code to verify what it does and Cyril is correct.

	The first two items - are zero length closed paths and so
should be joined with a miter, which is the default join. Hence nothing
should be shown on the first row.

	A circle and square on the 1st row is an error. So IE9 and Bitflash
do this one wrong. They must be detecting the 0 length segment and adding
caps without keeping state regarding the coincident points. Perhaps coincident
points are being discarded before stroking is done, who knows.

	Either way, Cyril is correct - the test is in error. We don't
pass yet but will some day...

Alex

--Original Message--:
>Hi all,
>
>First, apologies for not being very active. Still, on the request from Cameron, I'm trying to update the status of GPAC regarding the SVG 1.1F2 test suite. I have a problem with paint-stroke-10-t.svg, not with the zero-length part (http://www.w3.org/TR/SVG11/implnote.html#PathElementImplementationNotes) but with the use of linecap vs. linejoin. According to http://www.w3.org/TR/SVG11/paths.html#PathDataClosePathCommand, the z in the first row of the tests closes the path and therefore the linejoin should be applicable, not the linecap. Did I miss something in the spec?
>
>Regards,
>
>Cyril  
>-- 
>Cyril Concolato
>Maître de Conférences/Associate Professor
>Groupe Multimedia/Multimedia Group
>Telecom ParisTech
>46 rue Barrault
>75 013 Paris, France
>http://concolato.wp.institut-telecom.fr/
>
>
>

Received on Friday, 4 February 2011 06:41:55 UTC