RESOLVED FIXED 47181
SVGFEMergeNodeElement doesn't support dynamic invalidation
https://bugs.webkit.org/show_bug.cgi?id=47181
Summary SVGFEMergeNodeElement doesn't support dynamic invalidation
Renata Hodovan
Reported 2010-10-05 08:34:36 PDT
SVGFEMergeNodeElement doesn't support dynamic invalidation
Attachments
Patch for feMergeNode properties (135.60 KB, patch)
2010-10-05 08:46 PDT, Renata Hodovan
krit: review-
krit: commit-queue-
Patch for feMergeNode properties (137.75 KB, patch)
2010-10-06 03:38 PDT, Renata Hodovan
zimmermann: review-
Patch for feMergeNode properties (137.74 KB, patch)
2010-10-06 04:32 PDT, Renata Hodovan
zimmermann: review-
Patch for feMergeNode properties (137.77 KB, patch)
2010-10-06 05:15 PDT, Renata Hodovan
no flags
Renata Hodovan
Comment 1 2010-10-05 08:46:08 PDT
Created attachment 69792 [details] Patch for feMergeNode properties
Dirk Schulze
Comment 2 2010-10-05 10:00:47 PDT
Comment on attachment 69792 [details] Patch for feMergeNode properties Not sure, why you need feComposite on the tests. A test like this should be as simple as possible :-) Nevertheless, the patch itself looks great, but you should describe in the ChangeLog what you did to fix this issue. For example why did you ask the parent? why doesn't feMergeNode have a renderer, so that you have to call invalidation via the parent? And so on. As long as you don't have commit privileges, you have to upload a second patch. :-( r- because of the ChangeLog.
Renata Hodovan
Comment 3 2010-10-06 03:38:09 PDT
Created attachment 69913 [details] Patch for feMergeNode properties
Nikolas Zimmermann
Comment 4 2010-10-06 04:04:01 PDT
Comment on attachment 69913 [details] Patch for feMergeNode properties View in context: https://bugs.webkit.org/attachment.cgi?id=69913&action=review Looks good other than that, though as you can't commit yet on your own, you probably need to upload a new patch, hence r-. > WebCore/ChangeLog:8 > + Since feMergeNode doesn't have own renderer, we have to call the invalidation via it's parent. s/it's/its/ > WebCore/svg/SVGFEMergeNodeElement.cpp:55 > + if (attrName == SVGNames::inAttr) { If there's only one attribute to consider, you may want to use early exit style here. if (attrName != SVGNames::inAttr) return;
Renata Hodovan
Comment 5 2010-10-06 04:32:35 PDT
Created attachment 69921 [details] Patch for feMergeNode properties
Nikolas Zimmermann
Comment 6 2010-10-06 04:50:53 PDT
Comment on attachment 69921 [details] Patch for feMergeNode properties View in context: https://bugs.webkit.org/attachment.cgi?id=69921&action=review Other than that it looks good... > WebCore/svg/SVGFEMergeNodeElement.cpp:62 > + if (attrName != SVGNames::inAttr) > + return; > + if (Node* parentNode = parent()) { > + RenderObject* renderer = parentNode->renderer(); > + if (renderer && renderer->isSVGResourceFilterPrimitive()) > + RenderSVGResource::markForLayoutAndParentResourceInvalidation(renderer); > + } > +} Sorry, for not being clear. I meant: if (attrName != SVGNames::inAttr) return; Node* parentNode = parent(); if (!parentNode) return; RenderObject* renderer = parentNode->renderer(); if (!renderer || !renderer->isSVGResourceFilterPrimitive()) return; RenderSVGResource::mark...
Renata Hodovan
Comment 7 2010-10-06 05:15:27 PDT
Created attachment 69927 [details] Patch for feMergeNode properties
Nikolas Zimmermann
Comment 8 2010-10-06 05:19:56 PDT
Comment on attachment 69927 [details] Patch for feMergeNode properties Looks good, r=me.
WebKit Commit Bot
Comment 9 2010-10-06 10:10:02 PDT
Comment on attachment 69927 [details] Patch for feMergeNode properties Clearing flags on attachment: 69927 Committed r69206: <http://trac.webkit.org/changeset/69206>
WebKit Commit Bot
Comment 10 2010-10-06 10:10:09 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.