Add a makeDrawContextWithFallback that handles config fallback

Add a makeDrawContextWithFallback that handles config fallback

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005

Committed: https://skia.googlesource.com/skia/+/48fde9c4127860ca5851b88ba123169b9889445c

[robertphillips, 2016-09-01 16:42:10]

Description was changed from

==========
Move guard for renderable of A8 targets into GrContext::makeDrawContext
==========

to

==========
Move guard for renderable of A8 targets into GrContext::makeDrawContext
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

[robertphillips, 2016-09-01 17:13:59]

The CQ bit was checked by robertphillips@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-09-01 17:14:09]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[robertphillips, 2016-09-01 17:16:40]

Description was changed from

==========
Move guard for renderable of A8 targets into GrContext::makeDrawContext
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

to

==========
Move guard for renderable of A8 targets into GrContext::makeDrawContext

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. it wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

[robertphillips, 2016-09-01 17:16:41]

robertphillips@google.com changed reviewers:
+ bsalomon@google.com

[robertphillips, 2016-09-01 17:17:02]

Description was changed from

==========
Move guard for renderable of A8 targets into GrContext::makeDrawContext

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. it wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

to

==========
Move guard for renderability of A8 targets into GrContext::makeDrawContext

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. it wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

[robertphillips, 2016-09-01 17:17:25]

Description was changed from

==========
Move guard for renderability of A8 targets into GrContext::makeDrawContext

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. it wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

to

==========
Move guard for renderability of A8 targets into GrContext::makeDrawContext

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

[robertphillips, 2016-09-01 17:19:37]

What do you think about cherry picking this back to M54?

[commit-bot: I haz the power, 2016-09-01 17:49:22]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-01 17:49:23]

Dry run: This issue passed the CQ dry run.

[bsalomon, 2016-09-01 19:47:05]

https://codereview.chromium.org/2296193005/diff/20001/include/gpu/GrContext.h
File include/gpu/GrContext.h (right):

https://codereview.chromium.org/2296193005/diff/20001/include/gpu/GrContext.h...
include/gpu/GrContext.h:187: * This method will check the renderability of
'config' and fallback to
Should we specifically call out that the set of channels may be a superset of
the request but that the data type for the requested channels will be the same?
Main concern is that a caller won't think they might get an 8bit unorm channel
when they asked for a float16 (for example).

https://codereview.chromium.org/2296193005/diff/20001/src/gpu/GrContext.cpp
File src/gpu/GrContext.cpp (right):

https://codereview.chromium.org/2296193005/diff/20001/src/gpu/GrContext.cpp#n...
src/gpu/GrContext.cpp:692: // If the config isn't renderable try converting to
either A8 or an 32 bit config. Otherwise,
I think this should be more narrowly defined as always preserving the
representation for the channels in the requested config (e.g. not allowing 8bit
channels to substitute for 16 bit or 4 bit). We also need to be careful about
losing sRGBness.

https://codereview.chromium.org/2296193005/diff/20001/src/gpu/GrContext.cpp#n...
src/gpu/GrContext.cpp:698: } else if
(this->caps()->isConfigRenderable(kSkia8888_GrPixelConfig, sampleCnt > 0)) {
Let's not use kSkia8888. I'd like to remove it. kRGBA_8888 is renderable more
often than kSkia8888.

[robertphillips, 2016-09-01 20:52:54]

PTAL. I do feel the desire to boost 565 & 4444 to 8888.

https://codereview.chromium.org/2296193005/diff/20001/include/gpu/GrContext.h
File include/gpu/GrContext.h (right):

https://codereview.chromium.org/2296193005/diff/20001/include/gpu/GrContext.h...
include/gpu/GrContext.h:187: * This method will check the renderability of
'config' and fallback to
On 2016/09/01 19:47:05, bsalomon wrote:
> Should we specifically call out that the set of channels may be a superset of
> the request but that the data type for the requested channels will be the
same?
> Main concern is that a caller won't think they might get an 8bit unorm channel
> when they asked for a float16 (for example).

Done.

https://codereview.chromium.org/2296193005/diff/20001/src/gpu/GrContext.cpp
File src/gpu/GrContext.cpp (right):

https://codereview.chromium.org/2296193005/diff/20001/src/gpu/GrContext.cpp#n...
src/gpu/GrContext.cpp:692: // If the config isn't renderable try converting to
either A8 or an 32 bit config. Otherwise,
On 2016/09/01 19:47:05, bsalomon wrote:
> I think this should be more narrowly defined as always preserving the
> representation for the channels in the requested config (e.g. not allowing
8bit
> channels to substitute for 16 bit or 4 bit). We also need to be careful about
> losing sRGBness.

Done.

https://codereview.chromium.org/2296193005/diff/20001/src/gpu/GrContext.cpp#n...
src/gpu/GrContext.cpp:698: } else if
(this->caps()->isConfigRenderable(kSkia8888_GrPixelConfig, sampleCnt > 0)) {
On 2016/09/01 19:47:05, bsalomon wrote:
> Let's not use kSkia8888. I'd like to remove it. kRGBA_8888 is renderable more
> often than kSkia8888.

Done.

[robertphillips, 2016-09-01 22:14:26]

Description was changed from

==========
Move guard for renderability of A8 targets into GrContext::makeDrawContext

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

to

==========
Add a GrContextPriv::makeDrawContext that handles config fallback

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

[robertphillips, 2016-09-01 22:16:01]

PTAL

[bsalomon, 2016-09-02 13:13:05]

On 2016/09/01 22:16:01, robertphillips wrote:
> PTAL

I like this a lot better. However, do you think we could come up with a name to
distinguish it (other than the fact that it is on ContextPriv)?

makeDrawContextRelaxedConfig?

makeDrawContextWithConfigFallback?

Also, I don't mind if this is just on GrC as it'd be ok for someone's custom
effect to use it. But I don't really care either way.

[robertphillips, 2016-09-02 18:38:32]

Description was changed from

==========
Add a GrContextPriv::makeDrawContext that handles config fallback

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

to

==========
Add a makeDrawContextWithFallback that handles config fallback

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

[robertphillips, 2016-09-02 18:40:14]

renamed & now on GrContext

[bsalomon, 2016-09-02 18:58:55]

lgtm

[robertphillips, 2016-09-02 18:59:29]

The CQ bit was checked by robertphillips@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-09-02 18:59:37]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-02 19:33:41]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-02 19:33:42]

Dry run: This issue passed the CQ dry run.

[robertphillips, 2016-09-06 11:54:49]

The CQ bit was checked by robertphillips@google.com

[commit-bot: I haz the power, 2016-09-06 11:54:53]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-06 12:20:22]

Description was changed from

==========
Add a makeDrawContextWithFallback that handles config fallback

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005
==========

to

==========
Add a makeDrawContextWithFallback that handles config fallback

This fixes a bug in find_or_create_rrect_blur_mask where an A8-based drawContext
was desired but creation was failing b.c. A8 wasn't renderable.

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2296193005

Committed:
https://skia.googlesource.com/skia/+/48fde9c4127860ca5851b88ba123169b9889445c
==========

[commit-bot: I haz the power, 2016-09-06 12:20:23]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/48fde9c4127860ca5851b88ba123169b9889445c