For nested SVG, getScreenCTM() returns wrong position values (.e, .f). Please, see the attachment.
Created attachment 23050 [details] Attachment
I should have been a bit more specific, sorry: The two alerted values should be top and left coordinates of that squate, relative to the browser window.
Created attachment 23626 [details] First attempt The patch fixes the testcase. I hope there are no drawback to using the renderer abs pos and not the parents, I cant see any. Cheers, Rob.
Comment on attachment 23626 [details] First attempt Looks fine. r=me It'd be a lot better if the test case was a text-only test so it could be used cross platform. You should strive to make new tests text-only whenever possible.
(In reply to comment #4) > > It'd be a lot better if the test case was a text-only test so it could be used > cross platform. You should strive to make new tests text-only whenever > possible. I'm sorry, I don't quite understand what do you mean by text only? It is just a text..
(In reply to comment #5) > (In reply to comment #4) > > > > It'd be a lot better if the test case was a text-only test so it could be used > > cross platform. You should strive to make new tests text-only whenever > > possible. > > I'm sorry, I don't quite understand what do you mean by text only? It is just a > text.. Darin, meant a WebKit specific LayoutTest feature "layoutTestController.dumpAsText()". This way the test just logs the actual text output in the test-expected.txt file, instead of the coordinates of the RenderObjects' in the render tree. Greetings, Niko >
Fixed by r36856.