552062 - Black Background on certain Flash Objects

Status: RESOLVED FIXED

(Core Graveyard :: Plug-ins, defect)

Description

[Mike Pesce (:By-Tor)]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.3a3pre) Gecko/20100312 Minefield/3.7a3pre (.NET CLR 3.5.30729)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.3a3pre) Gecko/20100312 Minefield/3.7a3pre (.NET CLR 3.5.30729) ID:20100312160658

Since push http://hg.mozilla.org/mozilla-central/rev/53a85c489708 certain flash objects have black background where once they seemed to be transparent.  

Reproducible: Always

Steps to Reproduce:
1. Go go http://weather.weatherbug.com/NY/New%20York-weather.html?zcode=z5602&zip=10001
2. Look at wind indicator

Actual Results:  
Where once the background of the wind indicator was blue, now it's black as it gets updated.

Expected Results:  
The background should have remained blue.

Comment 2

[u88484]

Reopening this bug as bug 551945 fixed this issue with d2d disabled but the bug now occurs with d2d enabled.

Sometimes while the page is loading you can see the black background instead of seeing through the content.  Sometimes when this happens the black will disappear when the page finishes loading but not always, switching tabs and then back seems to fix it then.

Also always see black on hulu controls even after the page loads.  On the imageshack buttons as mentioned in bug 552052 but that only happens during load.

Comment 3

[Bas Schouten (:bas.schouten)]

It'd be interesting to see if this happens on normal GDI if you make these flash objects partially transparent. If this is the case then OOPP is giving an issue with double pass rendering most likely.

Comment 4

[Jim Mathies [:jimm]]

Attachment: uploader test case

In gdi, I'm seeing artifacts in the initial rendering of the uploader.

Comment 5

[Bas Schouten (:bas.schouten)]

(In reply to comment #4)
> Created an attachment (id=432438) [details]
> uploader test case
> 
> In gdi, I'm seeing artifacts in the initial rendering of the uploader.

Hrm, so this isn't so much a D2D bug. It's just more prevalent in D2D since it -always- uses double pass rendering.

Comment 6

[Jim Mathies [:jimm]]

(In reply to comment #5)
> (In reply to comment #4)
> > Created an attachment (id=432438) [details] [details]
> > uploader test case
> > 
> > In gdi, I'm seeing artifacts in the initial rendering of the uploader.
> 
> Hrm, so this isn't so much a D2D bug. It's just more prevalent in D2D since it
> -always- uses double pass rendering.

Seems that way. A regression range might be handy, we landed the IsDoublePass fixes, plus some race resolution code for paints last week. We likely regressed from those, I'm guessing it was the race resolution code. I'll hopefully have more info here in a bit once I get an updated debug build built.

Comment 7

[Jim Mathies [:jimm]]

hmm, double pass rendering appears to be working correctly. Not sure yet what's causing this.

Comment 8

[Jim Mathies [:jimm]]

w/GDI rendering in today's nightly, I'm not able to reproduce this today. Anyone else care to confirm?

Comment 9

[u88484]

I still can on imageshack at least

Comment 10

[Jim Mathies [:jimm]]

Attachment: fix

Simple way to reproduce this, open the test case, resize the window in, and then drag the window out slowly revealing the uploader.

Fix - reset the clip rect on the shared dib prior to rendering alpha extraction backgrounds and the second pass. Apparently flash mucks with the clipping when it paints.

Comment 11

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Comment on attachment 436009 [details] [diff] [review]
fix

Seems fine to me!

Comment 12

[Jim Mathies [:jimm]]

http://hg.mozilla.org/mozilla-central/rev/d31c87f6d202

Comment 13

[u88484]

Verified fixed using hourly Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.3a4pre) Gecko/20100405 Minefield/3.7a4pre

cset: 67f4c546d403

Verified on the testcase and the link in comment 0

Comment 14

[Mike Beltzner [:beltzner, not reading bugmail]]

Comment on attachment 436009 [details] [diff] [review]
fix

a=beltzner for 1.9.2.4

Comment 15

[Jim Mathies [:jimm]]

http://hg.mozilla.org/releases/mozilla-1.9.2/rev/4312f2140663

Comment 16

[Jim Mathies [:jimm]]

This doesn't appear to be fully addressed yet. While the flash test case and the flash weather site seem to be working ok, the following silverlight site isn't - 

http://www.microsoft.com/surface/en/us/default.aspx

The background underneath the video panel and text is supposed to show through. Sometimes it does, but mousing over the area of the silverlight control tends to flip parts of the background to black.

Bug 558434 is related - in tab previews the background is displayed white.

I'll look into this over the weekend.

Comment 17

[Benjamin Smedberg]

Please file a new bug, this one has already landed on branch and the branch tracking and verification will start getting really confused.

Comment 18

[Al Billings [:abillings - ex-MoCo]]

Verified for 1.9.2 with Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.2.4pre) Gecko/20100413 Namoroka/3.6.4pre using testcase for flash.