543201 - [OOPP] Shockwave position is not correct

Status: VERIFIED FIXED

(Core Graveyard :: Plug-ins, defect)

Description

[pal-moz]

Attachment: screenshot with ipc OFF and ON

1. go http://www.miniclip.com/games/ice-racer/en/

Comment 1

[Gary [:streetwolf52]]

confimed.

Comment 2

[Benjamin Smedberg]

Windowed plugin... I'm having trouble using Spy to figure out whether the window is in the correct position but it's painting incorrectly, or whether one of the windows is positioned incorrectly.

Blocks 1.9.3a1, perhaps only for adding shockwave to the OOPP blacklist.

Comment 3

[Jim Mathies [:jimm]]

(In reply to comment #2)
> Windowed plugin... I'm having trouble using Spy to figure out whether the
> window is in the correct position but it's painting incorrectly, or whether one
> of the windows is positioned incorrectly.
> 
> Blocks 1.9.3a1, perhaps only for adding shockwave to the OOPP blacklist.

Looks to be an issue with windowed controls. Spy indicates a chain of child windows (class="ImlWinCls") parented to the main GeckoPluginWindow. The final window in the chain which shockwave renders to is offset down and to the right of the root GeckoPluginWindow for some reason.

Comment 4

[Jim Mathies [:jimm]]

Attachment: fix

Comment 5

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

Comment on attachment 426711 [details] [diff] [review]
fix

So we could do this, or we could just always set our x&y to 0 and our width and height equal to parent. I think that might be smarter, actually, because it doesn't really make sense for us to allow a plugin to resize its window.

Comment 6

[Jim Mathies [:jimm]]

(In reply to comment #5)
> (From update of attachment 426711 [details] [diff] [review])
> So we could do this, or we could just always set our x&y to 0 and our width and
> height equal to parent. I think that might be smarter, actually, because it
> doesn't really make sense for us to allow a plugin to resize its window.

Sounds good, I really don't see any issues with completely locking it down for all content. Plugins should not be modifying containing window traits.

Comment 7

[Jim Mathies [:jimm]]

Attachment: fix

I had to cache width and height, GetClientRect in the changing event of the child on the parent retrieves the old dimensions.

Comment 8

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

Comment on attachment 426843 [details] [diff] [review]
fix

This looks good except I'm worried that plugins may expect to receive this message somehow. Can we pass the message to the plugins procedure and then fiddle with it? Or do you think that is overkill?

Comment 9

[Jim Mathies [:jimm]]

Attachment: patch

letting it fall through seems to work as well.

Comment 10

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

Comment on attachment 427472 [details] [diff] [review]
patch

r+ if you reset after the plugin handles it!

Comment 11

[[not reading bugmail]]

Attachment: screenshot: webpage leaking through tabs and scrolling tearing

Here is a screenshot of what looks like web page content leaking between tab switches as well as scrolling tearing

Comment 12

[Jim Mathies [:jimm]]

Attachment: final

Comment 13

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

Comment on attachment 427601 [details] [diff] [review]
final

r=me

Comment 14

[Jim Mathies [:jimm]]

http://hg.mozilla.org/mozilla-central/rev/c254eea884cb

Comment 15

[Marcia Knous [:marcia]]

https://litmus.mozilla.org/show_test.cgi?id=11636 added to Litmus.

Comment 16

[Simona Badau, Desktop QA]

Verified fixed on Firefox Beta 5:
Mozilla/5.0 (Windows NT 5.1; rv:5.0) Gecko/20100101 Firefox/5.0

*Note: The flash content (game) loads and displays correctly; it does not shift its location