Summary: | [CSS Regions] clip rectangles are incorrectly computed in some positioned elements | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Mihai Balan <mibalan> | ||||||||
Component: | CSS | Assignee: | Radu Stavila <stavila> | ||||||||
Status: | RESOLVED FIXED | ||||||||||
Severity: | Normal | CC: | commit-queue, johanneswilm, mvujovic, WebkitBugTracker | ||||||||
Priority: | P2 | ||||||||||
Version: | 528+ (Nightly build) | ||||||||||
Hardware: | Unspecified | ||||||||||
OS: | Unspecified | ||||||||||
Bug Depends on: | |||||||||||
Bug Blocks: | 57312 | ||||||||||
Attachments: |
|
This bug doesn't repro for me in today's nightly (r151260). But does the formula renderer? I need to download the attachment and open it locally for the test to work. I can still reproduce this in Chrome Version 29.0.1530.2 dev. (In reply to comment #2) > But does the formula renderer? I need to download the attachment and open it locally for the test to work. I can still reproduce this in Chrome Version 29.0.1530.2 dev. It seems the bug repros in Chrome like you say, but not in Safari. I only tested Safari. Thanks for the info! (In reply to comment #4) > Works for me on today's nightly (r151543) Look at the above comments- We should post a Chromium bug for this issue since it doesn't repro in Safari. I've logged https://code.google.com/p/chromium/issues/detail?id=249715 for Blink. Keeping this open so that we don't forget to import the test when this gets fixed in Blink. Created attachment 216672 [details]
Repaint test
This problem is already fixed in webkit. Added test in fast/repaint.
Comment on attachment 216672 [details] Repaint test View in context: https://bugs.webkit.org/attachment.cgi?id=216672&action=review Looks good but you need to decide whether you want to use repaint.js or text-based-repaint.js for this test. If you want to use repaint.js you need to generate a png file too. > LayoutTests/ChangeLog:3 > + Created test for testing repaint rects for regions and some positioned elements. Please clarify what you mean by "some positioned elements". > LayoutTests/fast/repaint/increasing-region-content-height.html:4 > + <script src="resources/repaint.js" type="text/javascript"></script> If you use repaint.js then you should also have a png file for the mac platform and add lines to skip the test on other non-mac platforms where the test is run. If you decide to use the script resources/text-based-repaint.js, then you will have a text output file without the png. From looking at the comment inside repaint.js, text-based-repaint.js seems the preferred way. > LayoutTests/fast/repaint/increasing-region-content-height.html:22 > + <body onload="runRepaintTest();"> Can you add a comment telling what the success of this test looks like? > LayoutTests/fast/repaint/increasing-region-content-height.html:24 > + <div class="zang"> zang? :) Created attachment 216686 [details]
Patch integrating reviewer feedback
Comment on attachment 216686 [details]
Patch integrating reviewer feedback
r=me
Comment on attachment 216686 [details] Patch integrating reviewer feedback Clearing flags on attachment: 216686 Committed r159132: <http://trac.webkit.org/changeset/159132> All reviewed patches have been landed. Closing bug. |
Created attachment 198476 [details] Test-case Positioned content generated by MathJax gets clipped or rendered invisibly, depending on the position of the region against the viewport. Reproduction ------------ - open attachment - click on the button Actual ------ For each click, the region moves down and some of the equation gets clipped (not all of it) Expected -------- For each click, the region moves down but the equation keeps showing