Bug 147809 – [EFL] Add disabled CSS_COMPOSITING option with new baseline

• Component: WebKit EFL
• Status: RESOLVED WONTFIX

[Gyuyoung Kim]

CSS_COMPOSITING can be enabled on EFL port. However some tests don't have expectation.

[Gyuyoung Kim]

Created attachment 258565 [details]
Patch

[Build Bot]

Comment on attachment 258565 [details]
Patch

Attachment 258565 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.webkit.org/results/30804

New failing tests:
fast/replaced/object-with-embed-url-param.html

[Build Bot]

Created attachment 258566 [details]
Archive of layout-test-results from ews104 for mac-mavericks-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews104  Port: mac-mavericks-wk2  Platform: Mac OS X 10.9.5

[Gyuyoung Kim]

Created attachment 258567 [details]
Patch

[Csaba Osztrogonác]

Have you compared png results with Mac results? Are they same/similar?
Additionally it would be great to check-in png results too, having only
rendertree dumps aren't so useful for human eyes to determine if the 
results are correct or not.

[Gyuyoung Kim]

(In reply to comment #5)
> Have you compared png results with Mac results? Are they same/similar?
> Additionally it would be great to check-in png results too, having only
> rendertree dumps aren't so useful for human eyes to determine if the 
> results are correct or not.

EFL layout test hasn't supported image diff officially yet. However let me check it locally. BTW it looks like you are on summer vacation. Have a nice vacation :)

[Gyuyoung Kim]

Created attachment 258677 [details]
Patch

[Gyuyoung Kim]

(In reply to comment #5)
> Have you compared png results with Mac results? Are they same/similar?
> Additionally it would be great to check-in png results too, having only
> rendertree dumps aren't so useful for human eyes to determine if the 
> results are correct or not.

Actual result of png files have nothing when running css3/blending with CSS3_COMPOSITING. So I need to investigate it. Anyway I would like to land this patch(with PRIVATE OFF) as a first step in order to support CSS3_COMPOSITING. When I fix this problem, I would like to set it to PUBLIC ON. What do you think about it ?

[Gyuyoung Kim]

(In reply to comment #8)
> (In reply to comment #5)
> > Have you compared png results with Mac results? Are they same/similar?
> > Additionally it would be great to check-in png results too, having only
> > rendertree dumps aren't so useful for human eyes to determine if the 
> > results are correct or not.
> 
> Actual result of png files have nothing when running css3/blending with
> CSS3_COMPOSITING. So I need to investigate it. Anyway I would like to land
> this patch(with PRIVATE OFF) as a first step in order to support
> CSS3_COMPOSITING. When I fix this problem, I would like to set it to PUBLIC
> ON. What do you think about it ?

Ossy ?

[Csaba Osztrogonác]

rs=me to add the option. 

But why should we add expected files for 
the tests if the feature is still disabled?
I prefer adding them when we enable the feature.

[Gyuyoung Kim]

(In reply to comment #10)
> rs=me to add the option. 
> 
> But why should we add expected files for 
> the tests if the feature is still disabled?
> I prefer adding them when we enable the feature.

Added expected files are similar with mac's one. So render tree expected result seems to be correct to me. If you think it would be good to add both text and png result at a time, it is better to add png result with this patch. Let me try to fix wrong png result soon. Then I will request review again.

[Csaba Osztrogonác]

(In reply to comment #11)
> (In reply to comment #10)
> > rs=me to add the option. 
> > 
> > But why should we add expected files for 
> > the tests if the feature is still disabled?
> > I prefer adding them when we enable the feature.
> 
> Added expected files are similar with mac's one. So render tree expected
> result seems to be correct to me. If you think it would be good to add both
> text and png result at a time, it is better to add png result with this
> patch. Let me try to fix wrong png result soon. Then I will request review
> again.

How is it possible that we have similar results with disabled CSS compositing 
as Mac platform with enabled CSS compositing? Something is weird here and
should be investigated in details.

If these tests are for testing CSS compositing feature, they should fail 
if the feature is disabled.

[Gyuyoung Kim]

(In reply to comment #12)
> (In reply to comment #11)
> > (In reply to comment #10)
> > > rs=me to add the option. 
> > > 
> > > But why should we add expected files for 
> > > the tests if the feature is still disabled?
> > > I prefer adding them when we enable the feature.
> > 
> > Added expected files are similar with mac's one. So render tree expected
> > result seems to be correct to me. If you think it would be good to add both
> > text and png result at a time, it is better to add png result with this
> > patch. Let me try to fix wrong png result soon. Then I will request review
> > again.
> 
> How is it possible that we have similar results with disabled CSS
> compositing 
> as Mac platform with enabled CSS compositing? Something is weird here and
> should be investigated in details.

The similar results were gotten when CSS compositing is enabled. As we said, png results have problem, I just wanted to add the results first.

[Csaba Osztrogonác]

(In reply to comment #13)

> The similar results were gotten when CSS compositing is enabled. As we said,
> png results have problem, I just wanted to add the results first.

This patch doesn't enable CSS compositing, only adds the cmake option.

Why is it useful to add expected files for enabled CSS compositing? 
They would fail on the bot until we enable it. (If they pass now, 
the tests are completely wrong and don't test CSS compositing at all.)
I prefer adding expected files in the patch which enables the feature 
in the future.

[Gyuyoung Kim]

(In reply to comment #14)
> (In reply to comment #13)
> 
> > The similar results were gotten when CSS compositing is enabled. As we said,
> > png results have problem, I just wanted to add the results first.
> 
> This patch doesn't enable CSS compositing, only adds the cmake option.

As I said, it was gotten when I enabled it. But latest patch disabled it for now.

> Why is it useful to add expected files for enabled CSS compositing? 
> They would fail on the bot until we enable it. (If they pass now, 
> the tests are completely wrong and don't test CSS compositing at all.)
> I prefer adding expected files in the patch which enables the feature 
> in the future.

I already agreed with your suggestion !

[Michael Catanzaro]

Closing this bug because the EFL port has been removed from trunk.

If you feel this bug applies to a different upstream WebKit port and was closed in error, please either update the title and reopen the bug, or leave a comment to request this.