Bug 97212 – [Qt] Don't render at scale 1.0 when doing a pinch unzoom

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Jocelyn Turcotte]

[Qt] Don't render at scale 1.0 when doing a pinch unzoom

[Kenneth Rohde Christiansen]

What is an unzoom?

[Jocelyn Turcotte]

Created attachment 164917 [details]
Patch

[Kenneth Rohde Christiansen]

Comment on attachment 164917 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=164917&action=review

> Source/WebKit2/ChangeLog:13
> +        This creates a couple of issues:
> +        - Create extra rendering when the user starts pinching that is bound to the
> +          page size and can be considerably big, reducing performance and peaking tile
> +          memory usage.
> +        - Request a new visible rect for each pinch update which in turn update all fixed layers.
> +

So why update at all states? Wont that make pinch zoom slow?

[Andras Becsi]

Comment on attachment 164917 [details]
Patch

LGTM.
Although other approaches to fix the mentioned tiling artifacts might increase the needed complexity considerably, this needs to be explored further but I think it is not high priority for Qt5.

[Jocelyn Turcotte]

(In reply to comment #3)
> (From update of attachment 164917 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=164917&action=review
> 
> > Source/WebKit2/ChangeLog:13
> > +        This creates a couple of issues:
> > +        - Create extra rendering when the user starts pinching that is bound to the
> > +          page size and can be considerably big, reducing performance and peaking tile
> > +          memory usage.
> > +        - Request a new visible rect for each pinch update which in turn update all fixed layers.
> > +
> 
> So why update at all states? Wont that make pinch zoom slow?

Humm I'll remove that comment, it's misleading. The problem is that we request a new visible rect but always with a 1.0 scale. For some reason that would relayout fixed layers, that might be the intended behavior for some other port.
This could be fixed easily but this patch rather reverts it for the first reason.

[Jocelyn Turcotte]

Created attachment 164932 [details]
Patch

Update the ChangeLog

[Kenneth Rohde Christiansen]

Comment on attachment 164932 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=164932&action=review

> Source/WebKit2/UIProcess/qt/PageViewportControllerClientQt.cpp:110
> +    // are rendered at the final destination while we perform the animation.

during the animation?

[Jocelyn Turcotte]

Created attachment 165596 [details]
Patch

Patch to commit

[WebKit Review Bot]

Comment on attachment 165596 [details]
Patch

Clearing flags on attachment: 165596

Committed r129497: <http://trac.webkit.org/changeset/129497>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.