remove subclassing from ImageFilter::Proxy

remove subclassing from ImageFilter::Proxy

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/2c55d7b7f3c2c834085d019bf6b1519b315c8aa1

[reed1, 2015-06-09 14:10:08]

reed@google.com changed reviewers:
+ robertphillips@google.com, senorblanco@google.com

[reed1, 2015-06-09 14:10:16]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2015-06-09 14:10:32]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1172693002/60001

[commit-bot: I haz the power, 2015-06-09 14:10:33]

Note for Reviewers:
The CQ is waiting for an approval.
If you believe that the CL is not ready yet, or if you would like to L-G-T-M
with comments then please uncheck the CQ checkbox.

Waiting for LGTM from valid reviewer(s) till 2015-06-09 20:10 UTC

[Stephen White, 2015-06-09 14:17:52]

senorblanco@chromium.org changed reviewers:
+ senorblanco@chromium.org

[Stephen White, 2015-06-09 14:17:53]

https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
File include/core/SkImageFilter.h (right):

https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
include/core/SkImageFilter.h:82: typedef SkImageFilterProxy Proxy;
Could we just nest the Proxy definition directly into SkImageFilter, use
SkImageFilter::Proxy everywhere (same as we do for ::Context and ::Cache), and
get rid of the typedef? No need to pollute the global scope with something
that's ImageFilter-specific.

https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
include/core/SkImageFilter.h:363: const SkSurfaceProps fProps;
Nit: indent fDevice to match fProps?

[reed1, 2015-06-09 14:19:48]

On 2015/06/09 14:17:53, Stephen White wrote:
>
https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
> File include/core/SkImageFilter.h (right):
> 
>
https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
> include/core/SkImageFilter.h:82: typedef SkImageFilterProxy Proxy;
> Could we just nest the Proxy definition directly into SkImageFilter, use
> SkImageFilter::Proxy everywhere (same as we do for ::Context and ::Cache), and
> get rid of the typedef? No need to pollute the global scope with something
> that's ImageFilter-specific.

Tried that, but I didn't see how to forward-declare-friend it in SkDevice.h

> 
>
https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
> include/core/SkImageFilter.h:363: const SkSurfaceProps fProps;
> Nit: indent fDevice to match fProps?

[reed1, 2015-06-09 14:20:56]

On 2015/06/09 14:19:48, reed1 wrote:
> On 2015/06/09 14:17:53, Stephen White wrote:
> >
>
https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
> > File include/core/SkImageFilter.h (right):
> > 
> >
>
https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
> > include/core/SkImageFilter.h:82: typedef SkImageFilterProxy Proxy;
> > Could we just nest the Proxy definition directly into SkImageFilter, use
> > SkImageFilter::Proxy everywhere (same as we do for ::Context and ::Cache),
and
> > get rid of the typedef? No need to pollute the global scope with something
> > that's ImageFilter-specific.
> 
> Tried that, but I didn't see how to forward-declare-friend it in SkDevice.h

Doh -- I see that device.h already pulls in SkImageFilter... ok, I'll try using
the nested friend.

> 
> > 
> >
>
https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
> > include/core/SkImageFilter.h:363: const SkSurfaceProps fProps;
> > Nit: indent fDevice to match fProps?

[reed1, 2015-06-09 14:22:59]

The CQ bit was unchecked by reed@google.com

[reed1, 2015-06-09 14:23:30]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2015-06-09 14:23:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1172693002/80001

[reed1, 2015-06-09 14:24:24]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2015-06-09 14:24:46]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1172693002/100001

[reed1, 2015-06-09 14:28:21]

https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
File include/core/SkImageFilter.h (right):

https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
include/core/SkImageFilter.h:82: typedef SkImageFilterProxy Proxy;
On 2015/06/09 14:17:53, Stephen White wrote:
> Could we just nest the Proxy definition directly into SkImageFilter, use
> SkImageFilter::Proxy everywhere (same as we do for ::Context and ::Cache), and
> get rid of the typedef? No need to pollute the global scope with something
> that's ImageFilter-specific.

Done.

https://codereview.chromium.org/1172693002/diff/60001/include/core/SkImageFil...
include/core/SkImageFilter.h:363: const SkSurfaceProps fProps;
On 2015/06/09 14:17:53, Stephen White wrote:
> Nit: indent fDevice to match fProps?

Done.

[Stephen White, 2015-06-09 15:17:04]

LGTM

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
File src/core/SkImageFilter.cpp (right):

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
src/core/SkImageFilter.cpp:529: #include "SkBitmapDevice.h"
Nit: move this to top-of-file?

[commit-bot: I haz the power, 2015-06-09 15:18:45]

Committed patchset #6 (id:100001) as
https://skia.googlesource.com/skia/+/2c55d7b7f3c2c834085d019bf6b1519b315c8aa1

[robertphillips, 2015-06-09 17:07:30]

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
File src/core/SkImageFilter.cpp (right):

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
src/core/SkImageFilter.cpp:532: SkBaseDevice::CreateInfo
cinfo(SkImageInfo::MakeN32Premul(w, h),
What about the SkTileImageFilter?

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
src/core/SkImageFilter.cpp:536: SkBaseDevice* dev =
fDevice->onCreateDevice(cinfo, NULL);
!dev ?

[Stephen White, 2015-06-10 14:42:45]

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
File src/core/SkImageFilter.cpp (right):

https://codereview.chromium.org/1172693002/diff/100001/src/core/SkImageFilter...
src/core/SkImageFilter.cpp:532: SkBaseDevice::CreateInfo
cinfo(SkImageInfo::MakeN32Premul(w, h),
On 2015/06/09 17:07:29, robertphillips wrote:
> What about the SkTileImageFilter?

SkTileImageFilter always does an extractSubset() of its input, which I believe
makes it tiling-safe, even on GPU.

At any rate, this code is not new -- Mike just moved it here from
SkDeviceImageFilterProxy.