Bug 32874 - Typed text in table not correctly redrawn on page with compositing layers
Summary: Typed text in table not correctly redrawn on page with compositing layers
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Layout and Rendering (show other bugs)
Version: 528+ (Nightly build)
Hardware: Macintosh OS X 10.5
: P2 Normal
Assignee: Simon Fraser (smfr)
URL: http://www.iquilezles.org/apps/shader...
Keywords:
Depends on:
Blocks:
 
Reported: 2009-12-22 11:39 PST by Simon Fraser (smfr)
Modified: 2009-12-29 09:19 PST (History)
1 user (show)

See Also:


Attachments
Reduced testcase (410 bytes, text/html)
2009-12-22 18:46 PST, Simon Fraser (smfr)
no flags Details
Patch (1.55 KB, patch)
2009-12-28 20:48 PST, Simon Fraser (smfr)
mitz: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Simon Fraser (smfr) 2009-12-22 11:39:44 PST
Visit http://www.iquilezles.org/apps/shadertoy/ with WebGL enabled, and try typing into Source text box. Note that your typing doesn't show correctly (but you do see some weird pixel dust).
Comment 1 Simon Fraser (smfr) 2009-12-22 18:46:23 PST
Created attachment 45414 [details]
Reduced testcase

This is related to table redraw in compositing mode, but, oddly, even when the table is not in a compositing layer.
Comment 2 Simon Fraser (smfr) 2009-12-22 19:45:12 PST
This is related to the code that walks up the container() chain in computeRectForRepaint(). RenderTableRow and RenderTableSection are both adding offsets?
Comment 3 mitz 2009-12-22 19:51:10 PST
Hmm… perhaps something needs to be done to compensate for <http://trac.webkit.org/changeset/51864>? Or just more of the same?
Comment 4 mitz 2009-12-22 19:54:17 PST
Actually RenderTableCell::computeRectForRepaint() appears to be doing the right thing. Are we starting the computation at the table row?
Comment 5 Simon Fraser (smfr) 2009-12-28 20:06:41 PST
Ah, we're getting a difference answer when RenderBox::computeRectForRepaint() goes through the LayoutState code path.
Comment 6 Simon Fraser (smfr) 2009-12-28 20:48:05 PST
Created attachment 45586 [details]
Patch

I wasn't able to create a repaint layoutTest that shows the problem, so no test here, alas.
Comment 7 Simon Fraser (smfr) 2009-12-29 09:19:13 PST
http://trac.webkit.org/changeset/52624