Missing parameter from the dynamic update of FESpecularLighting filter
Created attachment 82314 [details] Patch
Comment on attachment 82314 [details] Patch I'm unsure about lightning-color. This is a CSS property and should be be handled via SVGStyledElement, RenderStyle and the Renderer of primitives. Also you checked in a reference image with fail messages. Guess this is not expected.
(In reply to comment #2) > (From update of attachment 82314 [details]) > Also you checked in a reference image with fail messages. Guess this is not expected. Yes, you are right. My mistake. :$ > I'm unsure about lightning-color. This is a CSS property and should be be handled via SVGStyledElement, RenderStyle and the Renderer of primitives. I don't know, but FEDiffuseLighting handles the update of lighting-color the same way.
(In reply to comment #2) > (From update of attachment 82314 [details]) > I'm unsure about lightning-color. This is a CSS property and should be be handled via SVGStyledElement, RenderStyle and the Renderer of primitives. Also you checked in a reference image with fail messages. Guess this is not expected. There is also a SVG _attribute_ "lighting-color", despite the CSS property, I think handling the attr update this way is just fine, no? Or is your concern that "lighting-color" attribute should be mapped to a CSS property, and thus updates should be handled differently? Reni, can you check wheter the lighting-color attribute is properly mapped to it's corresponding CSS attribute? (Are changes to the attribute value, immediately visible in the computed style, even _without_ your patch?)
Created attachment 82426 [details] Patch
Dirk is right, SVGStyledElement properly maps diffuse_lightingAttr to it's corresponding CSS property (mapToEntry & friends). That means that the RenderSVGResourceFilterPrimitive (the renderer created for SVGFEDiffuseLighting/SVGFESpecLighting) should receive styleDidChange/styleWillChange calls, where you can catch those style changes.
The dynamic update of CSS related SVG props are handled in #56906. *** This bug has been marked as a duplicate of bug 56906 ***