[Webkit-unassigned] [Bug 66969] Chromium: Add a layer for rubber-band overhang painting to the hardware path.

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Sep 7 17:32:04 PDT 2011


--- Comment #26 from Adrienne Walker <enne at google.com>  2011-09-07 17:32:04 PST ---
(In reply to comment #25)
> Created an attachment (id=106638)
 --> (https://bugs.webkit.org/attachment.cgi?id=106638&action=review) [details]
> Patch

I talked with James a little bit offline (who is pretty busy).  I think part of the problem with this patch is what I mentioned in comment 2.  This continues to not really look like a Chromium-specific patch, even though this overhang logic is only needed for the Chromium port.  So, I think you're going to need some #ifdefs or something.  Also, your patch as written applies to iframes as well as the main frame, so you're going to need some additional logic there.

I worry a little bit about having a full viewport graphics layer.  I still don't like that you have to paint the overhang layers every frame because paintOverhangLayers intermingles paints of both a drop shadow that moves with the layer and a texture background that's pinned to the viewport.  And, even if you only paint part of this layer, you still potentially have to upload the entire viewport's worth of that buffer due to the way invalidation rect unioning happens in TiledLayerChromium.

Can you create a GPU trace of how this overhang painting performs? In particular, on a large monitor, full screen Chrome, open up a page that you can scroll on, open chrome://tracing in another tab, hit record, then go back to your original page.  Do some normal scrolling for a few seconds, then do some scrolling where you hit both the vertical and the horizontal overhang.  Hit the stop tracing button on the tracing tab.  You can then save that trace and send it to me or just attach it here.

Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

More information about the webkit-unassigned mailing list