531871 - Clipping problem with some flash

Status: VERIFIED FIXED

(Firefox for Android Graveyard :: General, defect, critical)

Description

[Doug Turner (:dougt)]

Steps to reproduce:

1) load http://people.mozilla.org/~dougt/flash_test.html
2) pan right and up such that the plugin is partially clipped by the url bar
3) pan down such that the plugin is completely in view.

Expected result:

The plugin to draw completely and not be clipped

Actual result:

When panning down such that the plugin is in view, we maintain the clip and do not draw the upper part of the plugin.


Here are the calls to setAbsoluteScreenPosition:

loading the page:

*** setAbsoluteScreenPosition  [8,80,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [8,80,0,0] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [262,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,80,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [262,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,80,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [262,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,80,250,100] ,  [0,0,250,100] 
*** setAbsoluteScreenPosition  [262,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,80,250,100] ,  [0,0,250,100] 


panning the plugin offscreen:

*** setAbsoluteScreenPosition  [262,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,80,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [262,112,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,12,250,100] ,  [0,0,250,100] 
*** setAbsoluteScreenPosition  [262,112,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [8,12,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [409,112,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [155,12,250,100] ,  [0,64,250,36] 


panning the plugin back in view:

*** setAbsoluteScreenPosition  [409,112,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [155,12,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [409,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [155,80,250,100] ,  [0,64,250,36] 
*** setAbsoluteScreenPosition  [409,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [155,80,250,100] ,  [-8,-8,0,0] 
*** setAbsoluteScreenPosition  [409,180,0,0] ,  [-262,-108,0,0] 
*** setAbsoluteScreenPosition  [155,80,250,100] ,  [0,64,250,36]

Comment 1

[Benjamin Stover (:stechz)]

Attachment: Do urlbar subtraction after mapping into client coordinates

The math was just wrong.

The problem boiled down to this: the critical region is represented in viewport coordinates (origin is at tile-container top left).  crit is (0, 0, ..., ...) whether the urlbar is being shown or not.

After mapping to client coordinates, it then makes sense to subtract the urlbar area in client coordinates.  I added a comment that explains what we are conceptually doing with the Math.max magic.

Comment 2

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

Comment on attachment 415253 [details] [diff] [review]
Do urlbar subtraction after mapping into client coordinates

>diff --git a/chrome/content/browser.js b/chrome/content/browser.js

>   updateEmbedRegions: function updateEmbedRegions(objects, crit) {

>       oprivate = objects[i].QueryInterface(nsIObjectLoadingContent);

Just realized that you might be able to QI this when you add to the objects array. That way you save some time when updating.

Comment 3

[Benjamin Stover (:stechz)]

Pushed: http://hg.mozilla.org/mobile-browser/rev/d002eaaf4528

Comment 4

[Aakash Desai [:aakashd]]

verified FIXED on build:

Mozilla/5.0 (X11; U; Linux armv7l; Nokia N900; en-US; rv:1.9.2b5pre) Gecko/20091202 Firefox/3.6b5pre Fennec/1.0b6pre