545687 - Mouse coordinates incorrect with Flash Player 10.x only in Firefox 3.6

Status: RESOLVED INCOMPLETE

(Core Graveyard :: Plug-ins, defect)

Description

[me]

User-Agent:       Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_6_2; en-us) AppleWebKit/531.21.8 (KHTML, like Gecko) Version/4.0.4 Safari/531.21.10
Build Identifier: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2) Gecko/20100115 Firefox/3.6

When running a Flash site/application inside Firefox 3.6 a click on the Flash area immediately reports two new mouse coordinates, the first varying greatly from the current mouse position. We can only recreate this issue with Firefox 3.6 and have only tested on OSX 10.6.2.

The bug has also been reported to Adobe:
http://bugs.adobe.com/jira/browse/FP-3888

Reproducible: Always

Steps to Reproduce:
How to reproduce with Flash:
1. Add MouseEvent.MOUSE_MOVE listener to stage. 
2. Press and release the stage. 

  Code: 
package { 
   import flash.events.MouseEvent; 

   public class Main extends Sprite 
   { 
      public function Main() 
      { 
         this.stage.addEventListener(MouseEvent.MOUSE_MOVE, trackMouseMove); 
      } 
   
      private function trackMouseMove( $evt:MouseEvent ):void 
      { 
         trace( "X:",$evt.stageX," Y:",$evt.stageY ); 
      } 
   } 
} 
Actual Results:  
On mouse up, Flash throws two MouseEvent.MOUSE_MOVE events. The first event's X & Y are greatly less than the current mouse position. The second events X & Y are at the actual mouse position. 

Expected Results:  
On mouse up no MouseEvent.MOUSE_MOVE events should be fired. This works correctly in previous versions of Firefox on Mac as well as Safari on Mac. This bug has not been tested on any other platforms. 

This bug breaks most navigation that relies on Flash's native MOUSE_OVER and MOUSE_OUT events.

Comment 1

[Charles]

Hi thanks for filing this and the jira bug.  I am  not sure I follow.  When I set up the situation I cannot reproduce what you are saying.  The only coordinates I see are the mouse move ones I am listening for.  Mouse down had no effect on the latest shipping version of Flash Player.

Comment 2

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

Dupe of bug 542068?

Comment 3

[Steven Michaud [:smichaud] (Retired)]

> Dupe of bug 542068?

Possibly.

I'll post a patch for bug 542068 soon (probably later today).  Then we can see if it fixes this bug.

Comment 4

[Steven Michaud [:smichaud] (Retired)]

There's not really enough information here to reproduce this bug.

But I now have patches for bug 542068, and tryserver builds made from
them (see bug 542068 comment #39).  Please test with them and let us
know your results.

Comment 5

[Kevin Brosnan]

Reporter would you test the build at ftp://ftp.mozilla.org/pub/firefox/nightly/3.6.4-candidates/build6/mac/en-US/ or ftp://ftp.mozilla.org/pub/firefox/nightly/latest-trunk/ either should have the fix for bug 542068.

Comment 6

[Kevin Brosnan]

Need a conformation that this is not fixed in recent builds and a working testcase to reopen this issue.

Comment 7

[Alexander Casassovici]

The bug is still there ... on all firefox versions up to 10 and with flash player 11
this only happens on the mac though (windows has been tested OK)

here's a video to show it : http://www.youtube.com/watch?v=D6MqWaB4tco

Comment 8

[Alexander Casassovici]

The bug is still there ... on all firefox versions up to 10 and with flash player 10 & 11
this only happens on the mac though (windows has been tested OK)

here's a video to show it : http://www.youtube.com/watch?v=D6MqWaB4tco

Comment 9

[koenpunt]

I found a solution for this bug.

The incorrect mouse position happens when the surrounding element is absolute positioned has an offset and has the transform property set to translate(0, 0) (or alike), which is commonly used to enable hardware acceleration. 

Removing this line from my css fixed my problem.

See the difference in the following examples:

With: transform
http://jsfiddle.net/koenpunt/jHgJh/

Without: transform
http://jsfiddle.net/koenpunt/jHgJh/1/

Comment 10

[Georg Fritzsche [:gfritzsche]]

(In reply to koenpunt from comment #9)
> I found a solution for this bug.

This looks like an issue specific to the transform, while the original was probably not limited to that. Can you please file a new bug on this?