SkData -> SkPicture::SnapshotArray

SkData -> SkPicture::SnapshotArray

Restores type safety with all the same features.

(Also note, less code: 29 insertions, 50 deletions.)

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/b7ee349f108040fbd5cb71d891d5b74a1843b0bc

[mtklein, 2014-11-20 19:13:05]

mtklein@google.com changed reviewers:
+ mtklein@google.com, reed@google.com

[reed1, 2014-11-20 20:36:47]

https://codereview.chromium.org/746553002/diff/20001/include/core/SkPicture.h
File include/core/SkPicture.h (right):

https://codereview.chromium.org/746553002/diff/20001/include/core/SkPicture.h...
include/core/SkPicture.h:210: const SkPicture*& operator[](size_t i)       {
return fArray[i]; }
this looks tricky, since it allows the receiver to modify an entry. Can we
change the constructor to take ownership of an already-built array? If so, that
would more naturally look like a build -> transport pattern.

Worse case, change the method name to something like writable_set(int index,
pic*) which asserts 1 owner?

https://codereview.chromium.org/746553002/diff/20001/src/core/SkRecorder.cpp
File src/core/SkRecorder.cpp (right):

https://codereview.chromium.org/746553002/diff/20001/src/core/SkRecorder.cpp#...
src/core/SkRecorder.cpp:38: SkPicture::SnapshotArray* pics =
SkNEW_ARGS(SkPicture::SnapshotArray, (count));
Seems like this section could turn into:

SkPicture** array = new ...[count];
for (int i = 0; ...) 
    array[i] = newshapshot;

return new SnapshotArray(array, count);

[mtklein, 2014-11-21 16:03:51]

https://codereview.chromium.org/746553002/diff/20001/include/core/SkPicture.h
File include/core/SkPicture.h (right):

https://codereview.chromium.org/746553002/diff/20001/include/core/SkPicture.h...
include/core/SkPicture.h:210: const SkPicture*& operator[](size_t i)       {
return fArray[i]; }
On 2014/11/20 20:36:47, reed1 wrote:
> this looks tricky, since it allows the receiver to modify an entry. Can we
> change the constructor to take ownership of an already-built array? If so,
that
> would more naturally look like a build -> transport pattern.
> 
> Worse case, change the method name to something like writable_set(int index,
> pic*) which asserts 1 owner?

Went with set().  I don't think we can assert single ownership... doesn't that
defeat the point of refcounting the snapshots?

[reed1, 2014-11-21 18:30:02]

lgtm

[mtklein, 2014-11-21 18:54:42]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-11-21 18:55:30]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/746553002/60001

[commit-bot: I haz the power, 2014-11-21 19:06:07]

Committed patchset #4 (id:60001) as
https://skia.googlesource.com/skia/+/b7ee349f108040fbd5cb71d891d5b74a1843b0bc