Bug 49866 – Assert hit in WebCore::invalidateResourcesOfChildren in SVGRenderSupport.cpp

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: SVG

[Rob Buis]

The following testcase triggers this assert:

<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
<rect id="rect" x="0" y="100" width="50" height="50" fill="white" mask="url(#mask4)"/> 
<defs>
     <mask id="mask4">
    </mask>
</defs>
</svg>

Though this unusual SVG, I am pretty sure it is valid.
Cheers,

Rob.

[Rob Buis]

*** This bug has been marked as a duplicate of bug 47682 ***

[Dirk Schulze]

(In reply to comment #0)
> The following testcase triggers this assert:
> 
> <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
> <rect id="rect" x="0" y="100" width="50" height="50" fill="white" mask="url(#mask4)"/> 
> <defs>
>      <mask id="mask4">
>     </mask>
> </defs>
> </svg>
> 
> Though this unusual SVG, I am pretty sure it is valid.
> Cheers,
> 
> Rob.

If I reference the rect after the <defs> section by a <use> element, the test works, but if it is before this section, it asserts as well. So it might be a problem that the parsing of the document did not finished and the resources are unknown. Note, this only happens for mask and filter, but not for clipper, gradient or pattern.

[Nikolas Zimmermann]

(In reply to comment #2)
> (In reply to comment #0)
> > The following testcase triggers this assert:
> > 
> > <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
> > <rect id="rect" x="0" y="100" width="50" height="50" fill="white" mask="url(#mask4)"/> 
> > <defs>
> >      <mask id="mask4">
> >     </mask>
> > </defs>
> > </svg>
> > 
> > Though this unusual SVG, I am pretty sure it is valid.
> > Cheers,
> > 
> > Rob.
> 
> If I reference the rect after the <defs> section by a <use> element, the test works, but if it is before this section, it asserts as well. So it might be a problem that the parsing of the document did not finished and the resources are unknown. Note, this only happens for mask and filter, but not for clipper, gradient or pattern.

<use> elements are way trickier to understand. The attach time is different, the <use> rendering subtree is built delayed, so the layout flow is completly different than using no <use> element.

[Dirk Schulze]

(In reply to comment #3)
> (In reply to comment #2)
> > (In reply to comment #0)
> > > The following testcase triggers this assert:
> > > 
> > > <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
> > > <rect id="rect" x="0" y="100" width="50" height="50" fill="white" mask="url(#mask4)"/> 
> > > <defs>
> > >      <mask id="mask4">
> > >     </mask>
> > > </defs>
> > > </svg>
> > > 
> > > Though this unusual SVG, I am pretty sure it is valid.
> > > Cheers,
> > > 
> > > Rob.
> > 
> > If I reference the rect after the <defs> section by a <use> element, the test works, but if it is before this section, it asserts as well. So it might be a problem that the parsing of the document did not finished and the resources are unknown. Note, this only happens for mask and filter, but not for clipper, gradient or pattern.
> 
> <use> elements are way trickier to understand. The attach time is different, the <use> rendering subtree is built delayed, so the layout flow is completly different than using no <use> element.

Nevertheless, if you move the rect after the resource, it works.

[Nikolas Zimmermann]

(In reply to comment #4)
> (In reply to comment #3)
> > (In reply to comment #2)
> > > (In reply to comment #0)
> > > > The following testcase triggers this assert:
> > > > 
> > > > <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
> > > > <rect id="rect" x="0" y="100" width="50" height="50" fill="white" mask="url(#mask4)"/> 
> > > > <defs>
> > > >      <mask id="mask4">
> > > >     </mask>
> > > > </defs>
> > > > </svg>
> > > > 
> > > > Though this unusual SVG, I am pretty sure it is valid.
> > > > Cheers,
> > > > 
> > > > Rob.
> > > 
> > > If I reference the rect after the <defs> section by a <use> element, the test works, but if it is before this section, it asserts as well. So it might be a problem that the parsing of the document did not finished and the resources are unknown. Note, this only happens for mask and filter, but not for clipper, gradient or pattern.
> > 
> > <use> elements are way trickier to understand. The attach time is different, the <use> rendering subtree is built delayed, so the layout flow is completly different than using no <use> element.
> 
> Nevertheless, if you move the rect after the resource, it works.
Sure, the order matters. I suspect we have a problem with the calls to RenderSVGResource:* in RenderSVGPath and other renderers.