643836 - Invalidation does not always trigger a displayport update

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Benjamin Stover (:stechz)]

The problem is that we were invalidating for the visual overflow rect when we should be accounting for the displayport. This is overshadowed by Fennec redrawing everytime a scroll is finished, but it causes us to unnecessarily checkerboard in some cases until the pan is finished.

Comment 1

[Benjamin Stover (:stechz)]

Attachment: Invalidation does not always trigger a displayport update

Ironically, I forgot to adjust the invalidation for any displayport.

Comment 2

[Timothy Nikkel (:tnikkel)]

Comment on attachment 521005 [details] [diff] [review]
Invalidation does not always trigger a displayport update

A few lines are way past 80 chars long.

Comment 3

[Benjamin Stover (:stechz)]

Will fix that and extra space on landing.

Comment 4

[Benjamin Stover (:stechz)]

Pushed http://hg.mozilla.org/mozilla-central/rev/070a9c6209b3

Should I leave open since this may be a ridealong for any potential RC2?

Comment 5

[Mark Finkle (:mfinkle) (use needinfo?)]

We can move this to branch for 4.0.1

Comment 6

[Doug Turner (:dougt)]

landed on mozilla-2.1 -- 0a6405a47498

Comment 7

[Aaron Train [:aaronmt]]

Any exemplary methods for verifying the invalidation improvement?