Bug 95238 - [skia] Switch FEComponentTransfer to use skia's SkColorFilter
Summary: [skia] Switch FEComponentTransfer to use skia's SkColorFilter
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: SVG (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Stephen White
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-08-28 13:22 PDT by Stephen White
Modified: 2012-09-11 13:53 PDT (History)
8 users (show)

See Also:


Attachments
Patch (9.65 KB, patch)
2012-08-28 13:46 PDT, Stephen White
no flags Details | Formatted Diff | Diff
Archive of layout-test-results from gce-cr-linux-02 (643.49 KB, application/zip)
2012-08-28 15:41 PDT, WebKit Review Bot
no flags Details
Patch (9.89 KB, patch)
2012-08-29 07:32 PDT, Stephen White
no flags Details | Formatted Diff | Diff
Patch (7.15 KB, patch)
2012-09-11 11:37 PDT, Stephen White
no flags Details | Formatted Diff | Diff
Patch (9.83 KB, patch)
2012-09-11 11:43 PDT, Stephen White
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Stephen White 2012-08-28 13:22:35 PDT
Switch FEComponentTransfer to use skia's SkColorFilter when compiling for Skia.
Comment 1 Stephen White 2012-08-28 13:46:27 PDT
Created attachment 161049 [details]
Patch
Comment 2 WebKit Review Bot 2012-08-28 15:41:25 PDT
Comment on attachment 161049 [details]
Patch

Attachment 161049 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/13652573

New failing tests:
css3/filters/effect-reference-ordering.html
css3/filters/effect-combined.html
css3/filters/effect-opacity.html
Comment 3 WebKit Review Bot 2012-08-28 15:41:28 PDT
Created attachment 161074 [details]
Archive of layout-test-results from gce-cr-linux-02

The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: gce-cr-linux-02  Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'>  Platform: Linux-2.6.39-gcg-201203291735-x86_64-with-Ubuntu-10.04-lucid
Comment 4 Stephen White 2012-08-29 07:32:37 PDT
Created attachment 161223 [details]
Patch
Comment 5 Dirk Schulze 2012-08-29 18:00:59 PDT
I would like to have Philip or Stephen take a look at this first.
Comment 6 Stephen White 2012-09-11 09:23:46 PDT
schenney, pdr:  ping?
Comment 7 Stephen Chenney 2012-09-11 10:46:02 PDT
Comment on attachment 161223 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=161223&action=review

> Source/WebCore/platform/graphics/filters/FEComponentTransfer.h:83
> +    virtual bool platformApplySkia();

Where is this actually called? Did the patch fail to pick up a changed file? If this is a preparatory patch then the test expectations shouldn't be updated.

> Source/WebCore/platform/graphics/filters/skia/FEComponentTransferSkia.cpp:39
> +    if (!in) 

This is only invoked if there is an inputEffect, so no need to check this.
Comment 8 Florin Malita 2012-09-11 10:49:09 PDT
(In reply to comment #7)
> > Source/WebCore/platform/graphics/filters/FEComponentTransfer.h:83
> > +    virtual bool platformApplySkia();
> 
> Where is this actually called? 

FilterEffect::apply()
Comment 9 Stephen Chenney 2012-09-11 10:53:34 PDT
(In reply to comment #8)
> (In reply to comment #7)
> > > Source/WebCore/platform/graphics/filters/FEComponentTransfer.h:83
> > > +    virtual bool platformApplySkia();
> > 
> > Where is this actually called? 
> 
> FilterEffect::apply()

Not in this patch. :-)
Comment 10 Stephen White 2012-09-11 10:56:35 PDT
(In reply to comment #9)
> (In reply to comment #8)
> > (In reply to comment #7)
> > > > Source/WebCore/platform/graphics/filters/FEComponentTransfer.h:83
> > > > +    virtual bool platformApplySkia();
> > > 
> > > Where is this actually called? 
> > 
> > FilterEffect::apply()
> 
> Not in this patch. :-)

Florin is correct.  It's virtual, and the base class version returns false.  This patch picks up the override for FEComponentTransfer.
Comment 11 Stephen Chenney 2012-09-11 11:01:01 PDT
Comment on attachment 161223 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=161223&action=review

>> Source/WebCore/platform/graphics/filters/FEComponentTransfer.h:83
>> +    virtual bool platformApplySkia();
> 
> Where is this actually called? Did the patch fail to pick up a changed file? If this is a preparatory patch then the test expectations shouldn't be updated.

Ah, now I remember how this stupid thing works. Everything defines platformApplySkia but the default returns false. Sorry.

>> Source/WebCore/platform/graphics/filters/skia/FEComponentTransferSkia.cpp:39
>> +    if (!in) 
> 
> This is only invoked if there is an inputEffect, so no need to check this.

This is a nit. If a reviewer doesn't care then I don't care, but if I were a reviewer I would care having gotten grief about this kind of thing in the past.
Comment 12 Stephen White 2012-09-11 11:37:50 PDT
Created attachment 163412 [details]
Patch
Comment 13 Stephen White 2012-09-11 11:43:52 PDT
Created attachment 163414 [details]
Patch
Comment 14 Stephen White 2012-09-11 12:11:00 PDT
Comment on attachment 161223 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=161223&action=review

Thanks for the review.

>>> Source/WebCore/platform/graphics/filters/skia/FEComponentTransferSkia.cpp:39
>>> +    if (!in) 
>> 
>> This is only invoked if there is an inputEffect, so no need to check this.
> 
> This is a nit. If a reviewer doesn't care then I don't care, but if I were a reviewer I would care having gotten grief about this kind of thing in the past.

Fixed.
Comment 15 Stephen Chenney 2012-09-11 13:26:50 PDT
Comment on attachment 163414 [details]
Patch

I would certainly R+ this. Dirk, over to you, 'cause you have the power.
Comment 16 Dirk Schulze 2012-09-11 13:44:37 PDT
Comment on attachment 163414 [details]
Patch

r=me
Comment 17 WebKit Review Bot 2012-09-11 13:53:43 PDT
Comment on attachment 163414 [details]
Patch

Clearing flags on attachment: 163414

Committed r128223: <http://trac.webkit.org/changeset/128223>
Comment 18 WebKit Review Bot 2012-09-11 13:53:47 PDT
All reviewed patches have been landed.  Closing bug.