[Webkit-unassigned] [Bug 45750] Move backing-store contents on slow scrolls

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=45750


João Paulo Rechi Vita <jprvita at profusion.mobi> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hyatt at apple.com




--- Comment #8 from João Paulo Rechi Vita <jprvita at profusion.mobi>  2010-09-30 13:44:08 PST ---
(In reply to comment #5)
> (From update of attachment 67551 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=67551&action=review
> 
> > WebCore/ChangeLog:9
> > +        With this information is possible to scroll the backing-store contents
> > +        without redrawing the whole scrolled area.
> 
> Without this change we can acomplish this for Qt. You are also introducing behaviour changes elsewhere and there are no tests, thus r-.
> 

I can't imagine how you can move the pixels without having the direction and delta information, but from your latter comment I think you have misunderstood what I'm using the information for due to my bad explanation on the ChangeLog. But if you can really do that, please point me to where you do so on the Qt port.

> > WebCore/page/ChromeClient.h:138
> > +        virtual void invalidateContentsForSlowScroll(const IntSize&, const IntRect&, bool) = 0;
> 
> it is not obvious what the size here is for so I would write out the variable name... is it for a delta? The original method is not much better though, I wonder what the bool is for!
> 

Yes, it's for a delta. I've just followed the WebCore policy of putting only the types on the header files.

> > WebCore/platform/ScrollView.cpp:544
> > -       hostWindow()->invalidateContentsForSlowScroll(updateRect, false);
> > +       hostWindow()->invalidateContentsForSlowScroll(-scrollDelta, updateRect, false);
> 
> There seems to be behaviour changes here, where are the layout tests?

There are no behavior changes, just the additional parameter is passed on, as you also stated on comment #6. I've updated this function declaration/implementation on the other ports but the parameter is ignored. Only the EFL port uses this information so far.

I'll send a updated version with a better description on the ChangeLog (sorry, wasn't sure about how much info I should put there) and also CC dhyatt on this bug so he can let us know his opinion.

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