Bug 54745 – -webkit-transform: translate3d(0,0,1) fails to affect z-index of layer

• Component: New Bugs
• Status: RESOLVED INVALID

[W. James MacLean]

-webkit-transform: translate3d(0,0,1) fails to affect z-index of layer

[W. James MacLean]

Created attachment 82956 [details]
Patch

[James Robinson]

This renders correctly without your patch. 3d transforms do not affect the z-index property and do not change the global ordering outside of a -webkit-preserves-3d block.

[W. James MacLean]

(In reply to comment #2)
> This renders correctly without your patch. 3d transforms do not affect the z-index property and do not change the global ordering outside of a -webkit-preserves-3d block.

Hmmm, OK.

1) What platform did you try on? This does not render properly (i.e. the blue box is in front) on Chrome on linux.

2) Since both divs are presumably opaque, should the closer one be on top?

This patch is in response to 

http://code.google.com/p/chromium/issues/detail?id=63781

Is the original bug asking for something incorrect then?

[James Robinson]

Chrome's rendering (blue on top) is correct.  Since the two transformed divs are siblings and do not share any transformed ancestors, they can be thought of as living in separate 3d spaces (see http://www.webkit.org/blog/386/3d-transforms/).  The rendering is determined based on the CSS stacking order.  Since the blue div appears later in the document it paints on top.

There is a bug in Safari here, would you mind checking if that is on file and filing a new bug if not (smfr and cmarrin would be good to CC)?

[W. James MacLean]

(In reply to comment #4)
> Chrome's rendering (blue on top) is correct.  Since the two transformed divs are siblings and do not share any transformed ancestors, they can be thought of as living in separate 3d spaces (see http://www.webkit.org/blog/386/3d-transforms/).  The rendering is determined based on the CSS stacking order.  Since the blue div appears later in the document it paints on top.
> 
> There is a bug in Safari here, would you mind checking if that is on file and filing a new bug if not (smfr and cmarrin would be good to CC)?

Will do ... thanks!