Coordinated Graphics: Remove an invisible TiledBackingStore of CoordinatedGraphicsLayer.

[Dongseong Hwang]

I found a bug in http://www.satine.org/research/webkit/snowleopard/snowstack.html

When we push space, the selected image zooms in. And then when we push space one more time, we can see two images: one image is zooming out and another lays behind.
We must remove the behind image.

I assume that the owner div of an image element draws the image although the div is not drawsContent.

[Dongseong Hwang]

This bug is not reproduced in WebKit1.
It is a Coordinated Graphics bug.

[Dongseong Hwang]

Created attachment 172737 [details]
Patch

[Noam Rosenthal]

Comment on attachment 172737 [details]
Patch

How is this tested?

[Dongseong Hwang]

(In reply to comment #3)
> (From update of attachment 172737 [details])
> How is this tested?

I'll make a layout test extracting from snowstack.html.

[Dongseong Hwang]

Created attachment 172983 [details]
Patch

[Dongseong Hwang]

(In reply to comment #5)
> Created an attachment (id=172983) [details]
> Patch

I fix this bug in the correct way.

Previous patch fixed the bug by change. In detail, TextureMapperLayer had m_backingStore although m_size is empty. TextureMapperLayer called m_backingStore->paintToTextureMapper(options.textureMapper, layerRect(),) in paintSelf. The previous patch made CoordinatedBackingStore not draw anything when targetRect (layerRect() above) is empty. So the bug disappeared.

The second patch fixes the root.

However, I need CoordinatedBackingStore::paintToTextureMapper to act like TextureMapperSurfaceBackingStore and TextureMapperTiledBackingStore in order for a direct composited image to use CoordinatedBackingStore.

I'll change this in another bug.

[Dongseong Hwang]

(In reply to comment #5)
> Created an attachment (id=172983) [details]
> Patch

Currently, we store an invisible TBS, so we waste heap and video memory.

[WebKit Review Bot]

Comment on attachment 172983 [details]
Patch

Rejecting attachment 172983 [details] from commit-queue.

Failed to run "['/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '-..." exit_code: 1

Last 500 characters of output:
  return self.open(self.click(*args, **kwds))
  File "/mnt/git/webkit-commit-queue/Tools/Scripts/webkitpy/thirdparty/autoinstalled/mechanize/_mechanize.py", line 203, in open
    return self._mech_open(url, data, timeout=timeout)
  File "/mnt/git/webkit-commit-queue/Tools/Scripts/webkitpy/thirdparty/autoinstalled/mechanize/_mechanize.py", line 255, in _mech_open
    raise response
webkitpy.thirdparty.autoinstalled.mechanize._response.httperror_seek_wrapper: HTTP Error 500: Internal Server Error

Full output: http://queues.webkit.org/results/14764621

[Helder Correia]

I think you should add #include <wtf/HashSet.h> at the top of CoordinatedBackingStore.h.

[WebKit Review Bot]

Comment on attachment 172983 [details]
Patch

Attachment 172983 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/14762784

New failing tests:
inspector-protocol/debugger-terminate-dedicated-worker-while-paused.html

[Dongseong Hwang]

(In reply to comment #9)
> I think you should add #include <wtf/HashSet.h> at the top of CoordinatedBackingStore.h.

Thank you very much!

[Dongseong Hwang]

Created attachment 173109 [details]
Patch

[Dongseong Hwang]

(In reply to comment #12)
> Created an attachment (id=173109) [details]
> Patch

wow, so fast. Thank you!

[WebKit Review Bot]

Comment on attachment 173109 [details]
Patch

Clearing flags on attachment: 173109

Committed r133955: <http://trac.webkit.org/changeset/133955>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.