[Webkit-unassigned] [Bug 15443] SVGImage does not support sub-resource loading

Thu Jan 27 01:05:13 PST 2011

https://bugs.webkit.org/show_bug.cgi?id=15443


--- Comment #11 from Dirk Schulze <krit at webkit.org>  2011-01-27 01:05:13 PST ---
(In reply to comment #9)
> There are two problems to solve here:
> (1) Load all subresources at all levels. (Currently, only the first level is loaded.)
> (2) Do not render images that reference themselves cyclically, either directly or indirectly. (Currently, cyclic references are rendered.)
IIRC ie9 is rendering cycles up to 6 to 7 times. The point is, that we have to detect the cycle on a certain point and end the rendering at a certain point in time.


> To solve (2), Dirk suggested me to use Niko's SVGResourcesCache and SVGResourcesCycleSolver. As far as I understood from examining these classes and the surrounding code, the infrastructure already exists:
> - RenderSVGModelObject::updateFromElement calls the cache / cycle solver, by calling SVGResourcesCache::clientUpdatedFromElement.
> - RenderSVGImage is a subclass of RenderSVGModelObject, therefore the cycle solver should automagically work for SVG image elements.
> 
> However, the even the simplest case of self-recursion, involving one <image> referencing itself, fails to go through this path. See the test case below: WebKit shows two circles instead of one. (Morley Abbot also posted a similar test case in attachment 59154 [details]; see his comment #2 above.)
> 
I did not take a look to the code of RenderSVGModelObject. But, if it checks the subresources, we should see two circles, since self-cycle.svg is the source, the first <image> loads the first subresource and should be drawn. IMHO it's important that we don't draw the circle a third time. Can you check this please?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.