Constify SkPicture in DrawingDisplayItem (cc side)

Constify SkPicture in DrawingDisplayItem

This initial patch just adds the web-facing methods for cc,
allowing the Blink side web classes to change to the new
form. Yet another patch will move cc completely to the new
form.

The SkPicture is immutable and should be const in a DrawingDisplayItem.
And there is no need for the location point; it is stored in the
cullRect of the picture.

R=ajuma@chromium.org
BUG=458240
Committed: https://crrev.com/0451901e7b5a079ff17240f1b1c797c6946b0735
Cr-Commit-Position: refs/heads/master@{#316116}

[Stephen Chennney, 2015-02-12 19:33:26]

FIXME cleanup. No change in behavior.

[Stephen Chennney, 2015-02-12 19:34:34]

Actually, small problem. Just a minute.

[Stephen Chennney, 2015-02-12 19:37:12]

Problem fixed.

[ajuma, 2015-02-12 19:42:59]

https://codereview.chromium.org/913413005/diff/1/cc/resources/drawing_display...
File cc/resources/drawing_display_item.cc (right):

https://codereview.chromium.org/913413005/diff/1/cc/resources/drawing_display...
cc/resources/drawing_display_item.cc:30:
canvas->translate(picture_->cullRect().x(), picture_->cullRect().y());
Since Blink is currently sending an empty location_, will this change behavior
on some tests? Or is Blink currently setting things up so that the cull rect is
the origin?

https://codereview.chromium.org/913413005/diff/20001/cc/blink/web_display_ite...
File cc/blink/web_display_item_list_impl.h (right):

https://codereview.chromium.org/913413005/diff/20001/cc/blink/web_display_ite...
cc/blink/web_display_item_list_impl.h:40: const blink::WebFloatPoint& location =
gfx::PointF());
I think we can do without the default arguments here, since Blink is currently
calling only the second version with both arguments, and will soon only call the
first version with one argument.

[Stephen Chennney, 2015-02-12 19:47:59]

Yep, it breaks the world. Should have run layout tests before posting. Will
update when I sort out the problems.

https://codereview.chromium.org/913413005/diff/1/cc/resources/drawing_display...
File cc/resources/drawing_display_item.cc (right):

https://codereview.chromium.org/913413005/diff/1/cc/resources/drawing_display...
cc/resources/drawing_display_item.cc:30:
canvas->translate(picture_->cullRect().x(), picture_->cullRect().y());
On 2015/02/12 19:42:59, ajuma wrote:
> Since Blink is currently sending an empty location_, will this change behavior
> on some tests? Or is Blink currently setting things up so that the cull rect
is
> the origin?

It seems to be breaking the world in layout tests, which I have just got around
to testing, so I think Blink needs to be updated. I'll have to restore the old
behavior in the old method. Annoying.

https://codereview.chromium.org/913413005/diff/20001/cc/blink/web_display_ite...
File cc/blink/web_display_item_list_impl.h (right):

https://codereview.chromium.org/913413005/diff/20001/cc/blink/web_display_ite...
cc/blink/web_display_item_list_impl.h:40: const blink::WebFloatPoint& location =
gfx::PointF());
On 2015/02/12 19:42:59, ajuma wrote:
> I think we can do without the default arguments here, since Blink is currently
> calling only the second version with both arguments, and will soon only call
the
> first version with one argument.

Yes, I'll verify.

[Stephen Chennney, 2015-02-12 20:05:15]

Totally minimal now. Verified with and without the associated Blink-side
changes.

[ajuma, 2015-02-12 20:08:54]

lgtm

[ajuma, 2015-02-12 20:10:57]

(Please update the description before landing.)

[Stephen Chennney, 2015-02-12 23:01:22]

The CQ bit was checked by schenney@chromium.org

[commit-bot: I haz the power, 2015-02-12 23:02:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/913413005/60001

[commit-bot: I haz the power, 2015-02-13 01:14:37]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-02-13 01:15:22]

Patchset 4 (id:??) landed as
https://crrev.com/0451901e7b5a079ff17240f1b1c797c6946b0735
Cr-Commit-Position: refs/heads/master@{#316116}

[Stephen Chennney, 2015-02-17 18:29:22]

New patchsets have been uploaded after l-g-t-m from ajuma@chromium.org

[Stephen Chennney, 2015-02-17 18:30:30]

I made a mistake the last time. This works with the Blink side changes, where we
were passing (0,0) as the location, not the cull rect corner.

[ajuma, 2015-02-17 19:05:59]

On 2015/02/17 18:30:30, Stephen Chenney wrote:
> I made a mistake the last time. This works with the Blink side changes, where
we
> were passing (0,0) as the location, not the cull rect corner.

The change lgtm, though I'd suggest opening a separate codereview issue for it
since it isn't a re-landing of the previous patch.