use NewUninitialized, cheaper than NewWithMalloc

use NewUninitialized, cheaper than NewWithMalloc

TBR=miguelg@chromium.org

Committed: https://crrev.com/a351adc8dba1513c4df1e947f176f189473c763e
Cr-Commit-Position: refs/heads/master@{#294646}

[reed2, 2014-09-12 00:46:10]

reed@chromium.org changed reviewers:
+ fmalita@chromium.org

[f(malita), 2014-09-12 11:49:59]

Nice, we were previously sk-freeing new-allocated data :)

https://codereview.chromium.org/569483002/diff/1/chrome/browser/android/thumb...
File chrome/browser/android/thumbnail/thumbnail_store.cc (right):

https://codereview.chromium.org/569483002/diff/1/chrome/browser/android/thumb...
chrome/browser/android/thumbnail/thumbnail_store.cc:574:
SkMallocPixelRef::NewWithData(info, 0, NULL, etc1_pixel_data.get()));
Is this safe? Looks like SkMallocPixelRef is ref-ing etc1_pixel_data before we
write data, but per SkData::NewUninitialized() docs

"The caller should call writable_data()  to write into the buffer, but this must
be done before another ref() is made"

[f(malita), 2014-09-12 11:54:34]

https://codereview.chromium.org/569483002/diff/1/chrome/browser/android/thumb...
File chrome/browser/android/thumbnail/thumbnail_store.cc (right):

https://codereview.chromium.org/569483002/diff/1/chrome/browser/android/thumb...
chrome/browser/android/thumbnail/thumbnail_store.cc:574:
SkMallocPixelRef::NewWithData(info, 0, NULL, etc1_pixel_data.get()));
On 2014/09/12 11:49:59, Florin Malita wrote:
> Is this safe? Looks like SkMallocPixelRef is ref-ing etc1_pixel_data before we
> write data, but per SkData::NewUninitialized() docs
> 
> "The caller should call writable_data()  to write into the buffer, but this
must
> be done before another ref() is made"
> 

I see it's just writable_data() asserting it gets called before ref-ing, so if
we're not using that at all it should be OK.

non-owner lgtm

[reed1, 2014-09-12 18:00:08]

reed@google.com changed reviewers:
+ miguelg@chromium.org, reed@google.com

[reed1, 2014-09-12 18:00:09]

need owner for chrome/browser/android

https://codereview.chromium.org/569483002/diff/1/chrome/browser/android/thumb...
File chrome/browser/android/thumbnail/thumbnail_store.cc (right):

https://codereview.chromium.org/569483002/diff/1/chrome/browser/android/thumb...
chrome/browser/android/thumbnail/thumbnail_store.cc:574:
SkMallocPixelRef::NewWithData(info, 0, NULL, etc1_pixel_data.get()));
On 2014/09/12 11:54:34, Florin Malita wrote:
> On 2014/09/12 11:49:59, Florin Malita wrote:
> > Is this safe? Looks like SkMallocPixelRef is ref-ing etc1_pixel_data before
we
> > write data, but per SkData::NewUninitialized() docs
> > 
> > "The caller should call writable_data()  to write into the buffer, but this
> must
> > be done before another ref() is made"
> > 
> 
> I see it's just writable_data() asserting it gets called before ref-ing, so if
> we're not using that at all it should be OK.
> 
> non-owner lgtm

Yes, the existing code is tricky that way, but since they're not calling
data->data() we won't get an assert.

[reed1, 2014-09-12 18:29:45]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2014-09-12 18:50:10]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patchset/569483002/1

[commit-bot: I haz the power, 2014-09-12 20:21:30]

Committed patchset #1 (id:1) as 01ebdb993f01dbe81fe3d1428d8ef92dc0267c18

[commit-bot: I haz the power, 2014-09-12 20:26:45]

Patchset 1 (id:??) landed as
https://crrev.com/a351adc8dba1513c4df1e947f176f189473c763e
Cr-Commit-Position: refs/heads/master@{#294646}