[WK2] Use ShareableSurface instead of ShareableBitmap in DrawingAreaImpl

[Igor Trindade Oliveira]

GTK and EFL port have plans to use OpenGL instead of Image backend, as initial step in WebKit2 we should use ShareableSurface instead of ShareableBitmap, so we can use OpenGL surfaces.

[Igor Trindade Oliveira]

Created attachment 137175 [details]
Patch

WIP patch.

[Igor Trindade Oliveira]

Created attachment 137354 [details]
Patch

Proposed patch.

[WebKit Review Bot]

Attachment 137354 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/WebKit2/CMakeLists.txt', u'Source/W..." exit_code: 1
Source/WebKit2/Shared/ShareableSurface.h:67:  The parameter name "hints" adds no information, so it should be removed.  [readability/parameter_name] [5]
Total errors found: 1 in 15 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Build Bot]

Comment on attachment 137354 [details]
Patch

Attachment 137354 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/12416147

[Igor Trindade Oliveira]

Created attachment 137359 [details]
Patch

Fix windows build.

[WebKit Review Bot]

Attachment 137359 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/WebKit2/CMakeLists.txt', u'Source/W..." exit_code: 1
Source/WebKit2/Shared/ShareableSurface.h:67:  The parameter name "hints" adds no information, so it should be removed.  [readability/parameter_name] [5]
Total errors found: 1 in 16 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Build Bot]

Comment on attachment 137359 [details]
Patch

Attachment 137359 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/12412438

[WebKit Review Bot]

Comment on attachment 137359 [details]
Patch

Attachment 137359 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/12414447

New failing tests:
fast/canvas/webgl/shader-precision-format.html

[WebKit Review Bot]

Created attachment 137386 [details]
Archive of layout-test-results from ec2-cr-linux-01

The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: ec2-cr-linux-01  Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'>  Platform: Linux-2.6.35-28-virtual-x86_64-with-Ubuntu-10.10-maverick

[Igor Trindade Oliveira]

Created attachment 137459 [details]
Patch

Trying again.

[Noam Rosenthal]

Comment on attachment 137459 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=137459&action=review

> Source/WebKit2/Shared/ShareableSurface.h:41
> +        SupportsGraphicsSurface = 0x01,
> +        None = 0x0

I'd rather if we put None first, and call it NoHints

[Build Bot]

Comment on attachment 137459 [details]
Patch

Attachment 137459 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/12414631

[Igor Trindade Oliveira]

Created attachment 137471 [details]
Patch

Use NoHints instead of None.

[Build Bot]

Comment on attachment 137471 [details]
Patch

Attachment 137471 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/12416438

[Igor Trindade Oliveira]

Created attachment 137474 [details]
Patch

it is becoming boring, trying again.

[Simon Hausmann]

(In reply to comment #0)
> GTK and EFL port have plans to use OpenGL instead of Image backend, as initial step in WebKit2 we should use ShareableSurface instead of ShareableBitmap, so we can use OpenGL surfaces.

You say initial step, but are you actually going to end up with that design?

I mean, if you go through with this you're also going to have to change the backing stores to be surfaces, and you'll probably want to combine multiple updates into the same surface. And then you'll end up half-way re-implementing what exists already for the WK2 AC implementation.

So wouldn't it be simpler to leave the drawing area with its working approach of using shared memory and software rendering (and update incorporation) and instead focus on the WK2 AC implementation? Because I suspect in the end you're going to want to use that and remain in permanent state of AC.

[Kenneth Rohde Christiansen]

(In reply to comment #16)
> (In reply to comment #0)
> > GTK and EFL port have plans to use OpenGL instead of Image backend, as initial step in WebKit2 we should use ShareableSurface instead of ShareableBitmap, so we can use OpenGL surfaces.
> 
> You say initial step, but are you actually going to end up with that design?
> 
> I mean, if you go through with this you're also going to have to change the backing stores to be surfaces, and you'll probably want to combine multiple updates into the same surface. And then you'll end up half-way re-implementing what exists already for the WK2 AC implementation.
> 
> So wouldn't it be simpler to leave the drawing area with its working approach of using shared memory and software rendering (and update incorporation) and instead focus on the WK2 AC implementation? Because I suspect in the end you're going to want to use that and remain in permanent state of AC.

I totally agree with that. That would be the right approach.

[Simon Hausmann]

Comment on attachment 137474 [details]
Patch

I'm saying r- based on the lack of clarity about where this is going and why it is better than putting an effort into getting WK2 AC working for the ports in question.