I see two misrendering svgs on this page, and both due to the same, likely gradient related, bug. Compare the svg and generated png, the differences are quite clear. Please fix :) Cheers, Rob.
I believe this is related to the hackish way in which we currently do style resolution for <stop> elements. However, I see: 2005-12-12 11:26:15.712 Safari[25820] Warning, no gradient stops, gradient (0x13da4be4) will be all black! in the console, which indicates this might not be a style problem at all.
Created attachment 5063 [details] Reduced testcase
Created attachment 5064 [details] Further reduction
Created attachment 5119 [details] Simple patch The referenced gradient (the one containing the stops) was never properly built (ie: addStop never called). Hense the gradient using xlink:href did not inherit any stop.
Rendering of gradients is much improved with this patch. Also fixes display (but not performance) of http://me.in-berlin.de/~darwin/svg/incoming/gearflowers.svg as referred to by 6060.
Comment on attachment 5119 [details] Simple patch Patch looks fine. But there's no ChangeLog and no regression test -- we would like to have both of those.
Comment on attachment 5119 [details] Simple patch This works, but I think we can do this better by removing the whole resourceNotification() system all together. I have a patch for that and will post it. Thanks for such a great diagnosis!
This works again. Although it's still not a nice as it should be.