More disentanglement from bug 614732, this time to remove the conflation of the concepts of effects offset and continuation union size in nsSVGIntegrationUtils. I found it terribly confusing trying to understand what on earth was going on when SVG terms such as "user space" and "bbox" were being perverted to mean things that they're not. This doesn't completely fix those issues, but it's a decent first pass.
Created attachment 634423 [details] [diff] [review] patch
Created attachment 634424 [details] [diff] [review] patch Sorry, that was the wrong patch.
nsSVGIntegrationUtils::GetSVGCoordContextForNonSVGFrame might be better called nsSVGIntegrationUtils::GetSVGCoordSizeForNonSVGFrame r=longsonr either way
Pushed https://hg.mozilla.org/integration/mozilla-inbound/rev/d7caec31923e (In reply to Robert Longson from comment #3) > nsSVGIntegrationUtils::GetSVGCoordContextForNonSVGFrame might be better > called nsSVGIntegrationUtils::GetSVGCoordSizeForNonSVGFrame It's more the size of the context than of a coord, at least to my mind, so I left it as it is. Plus I'm really trying to avoid conflicting queue at the moment.