[webkit-reviews] review denied: [Bug 15135] REGRESSION (r19843-r19850): Changing a flexbox's contents makes its container scroll to the top : [Attachment 31284] Patch v3

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Jun 15 23:16:35 PDT 2009


Dave Hyatt <hyatt at apple.com> has denied Shinichiro Hamaji
<hamaji at chromium.org>'s request for review:
Bug 15135: REGRESSION (r19843-r19850): Changing a flexbox's contents makes its
container scroll to the top
https://bugs.webkit.org/show_bug.cgi?id=15135

Attachment 31284: Patch v3
https://bugs.webkit.org/attachment.cgi?id=31284&action=review

------- Additional Comments from Dave Hyatt <hyatt at apple.com>
You've picked a very challenging bug to work on here.

I'm not sure I'd worry so much about repaint being wrong.  If the flexing block
has kids that move during the first layout, your patch isn't going to help that
case.  They are still going to repaint too much.  I think you should just
concentrate on the scrolling issue and ignore repaint for now.

As for scrolling, trying to stop just the immediate child block from updating
its scroll info during the first layout doesn't really generally solve the
problem.  What about a scrollable object inside a flexing block that
incorrectly adjusts its scroll position in response to the first layout?

I think a simpler fix for this bug is to just make the scroll position updating
code immune to flex box layout thrashing.


More information about the webkit-reviews mailing list