RESOLVED WONTFIX 55675
Changing svg transform attribute to the same value triggers repaint
https://bugs.webkit.org/show_bug.cgi?id=55675
Summary Changing svg transform attribute to the same value triggers repaint
Patrick R. Gansterer
Reported 2011-03-03 07:42:21 PST
When you have <rect fill="red" transform="scale(2)" /> and set the _fill_ attribute to "red" again it doesn't trigger a repaint. If you set the _transform_ attribute of the rect to "scale(2)" again, it requests a repaint.
Attachments
Patch (3.12 KB, patch)
2011-03-04 10:07 PST, Patrick R. Gansterer
aroben: review-
Alternative Patch (depends on bug 55829) (1.31 KB, patch)
2011-03-05 15:12 PST, Patrick R. Gansterer
eric: review-
Patrick R. Gansterer
Comment 1 2011-03-04 10:07:56 PST
Nikolas Zimmermann
Comment 2 2011-03-05 04:31:58 PST
Comment on attachment 84772 [details] Patch Are you sure it's worth the gain for the common-case, where the attribute indeed changed to a new value? Looks like an expensive test for a corner case...
Patrick R. Gansterer
Comment 3 2011-03-05 15:12:14 PST
Created attachment 84879 [details] Alternative Patch (depends on bug 55829) (In reply to comment #2) > Are you sure it's worth the gain for the common-case, where the attribute indeed changed to a new value? Looks like an expensive test for a corner case... I'm not 100% happy with the patch either, but I was my first try to fix this problem. I also added a a change comparison in the setAttribute directly, but many unit LayoutTest failed. This patch disables the whole svgAttributeChanged method, if the attribute stays the same but requires bug 55829.
WebKit Review Bot
Comment 4 2011-03-05 15:16:39 PST
Build Bot
Comment 5 2011-03-05 15:34:37 PST
Collabora GTK+ EWS bot
Comment 6 2011-03-05 16:01:26 PST
Early Warning System Bot
Comment 7 2011-03-05 16:05:48 PST
WebKit Review Bot
Comment 8 2011-03-05 16:46:30 PST
WebKit Review Bot
Comment 9 2011-03-05 17:11:46 PST
Eric Seidel (no email)
Comment 10 2011-04-10 16:22:48 PDT
svgAttributeChanged is kinda a disaster anyway. I can't remember why it's needed.
Nikolas Zimmermann
Comment 11 2011-04-11 08:01:08 PDT
(In reply to comment #10) > svgAttributeChanged is kinda a disaster anyway. I can't remember why it's needed. You're saying this since ages, it's not a disaster in anyway, but just fine. I explained several times why it's needed (unify SVG DOM _and_ XML DOM changes).
Adam Roben (:aroben)
Comment 12 2011-04-26 15:38:57 PDT
Comment on attachment 84772 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=84772&action=review I don't understand why the code was moved from svgAttributeChanged to parseMappedAttribute. Your ChangeLog should probably explain this. > Source/WebCore/ChangeLog:5 > + Chaning svg transform attribute to the same value triggers repaint Typo: Chaning
Adam Roben (:aroben)
Comment 13 2011-04-26 16:09:33 PDT
Comment on attachment 84879 [details] Alternative Patch (depends on bug 55829) View in context: https://bugs.webkit.org/attachment.cgi?id=84879&action=review > Source/WebCore/svg/SVGElement.cpp:386 > - if (attr->name() != HTMLNames::styleAttr) > + if (type != ChangedAttributeToSameValue && attr->name() != HTMLNames::styleAttr) Does HTMLElement or Element do this?
Eric Seidel (no email)
Comment 14 2012-01-30 15:02:51 PST
Just noticed this nearly 12-month-old bug. Nico: Thoughts on this change? You're much more familiar with SVG these days than I.
Eric Seidel (no email)
Comment 15 2012-03-01 13:53:40 PST
Comment on attachment 84879 [details] Alternative Patch (depends on bug 55829) With no response, and no answer to Adam's question, marking this r-. Feel free to mark this r? again with questions answered.
Dirk Schulze
Comment 16 2014-05-12 06:23:44 PDT
We do no longer have AttributeChangedType. Transforms should be handled by CSS in the future anyway.
Note You need to log in before you can comment on or make changes to this bug.