Bug 11580
| Summary: | REGRESSION(r16284-r16288): SVGs embedded in an XHTML document do not repaint properly, position wrong, corrupt general HTML rendering | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | dzuelke |
| Component: | SVG | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED WORKSFORME | ||
| Severity: | Normal | CC: | dzuelke, eric, mrowe |
| Priority: | P2 | Keywords: | NeedsReduction, Regression |
| Version: | 420+ | ||
| Hardware: | Mac (PowerPC) | ||
| OS: | OS X 10.4 | ||
| URL: | http://www.bitxtender.com/webkitsvg/ | ||
dzuelke
http://www.bitxtender.com/webkitsvg/ contains description, test cases and screenshots.
As mentioned there, this used to work some time ago, I'd say four weeks or so. If necessary, I could try to find the specific nightly where it broke, but I'm very busy at the moment, sorry, so I couldn't do that today. Let me know if you need any further assistance.
Tested with the latest nightly, of course ;)
P.S: the situation where the SVG graphics are rendered black only happens on localhost for me, not if I load the page from the above web server or via file:///.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Mark Rowe (bdash)
It seems very likely that the problem is caused by r16286 (<http://trac.webkit.org/projects/webkit/changeset/16286>).
I've built both r16285 and r16286, and the problem does not occur in the former. For a reduced test case that demonstrates a repaint issue that appears to have been introduced at the same revision, take a look at <http://trac.webkit.org/projects/webkit/browser/trunk/LayoutTests/svg/custom/simpleCDF.xml?format=raw>. The paragraph saying "You should see an SVG image above this!" does not appear until you make a selection that includes it. It will then disappear when you resize the window, etc.
Eric Seidel (no email)
IIRC, this regressed about a month before I left apple. It's been broken long before the NSGraphicsContext change... I think.
Mark Rowe (bdash)
Eric, if you download r16284.dmg">http://nightly.webkit.org/builds/WebKit-SVN-r16284.dmg and r16288.dmg">http://nightly.webkit.org/builds/WebKit-SVN-r16288.dmg you can quite clearly see that it regressed between r16284 and r16288. I checked out and built r16285 and r16286 manually, and they show the same behaviour: working in the former, broken in the latter.
Eric Seidel (no email)
Sorry, I looked at your comment to quickly. Yes, if the build shows that's the regression... then that's the regression. ;)
dzuelke
I just checked it again with r18159, and it works flawlessly now, this bug may probably be closed.
Eric Seidel (no email)
Looks fixed.