Fix bug in layer hoisting transition to SkRecord backend

Fix bug in layer hoisting transition to SkRecord backend

Care must be taken when setting up the initial CTM matrix for partial SkRecord playbacks b.c. all the setMatrix calls will concatenate with the initial matrix (which may be different then the CTM that is required to draw correctly).

Committed: https://skia.googlesource.com/skia/+/4815fe5a0a497b676677fb4e4a0f05c511855490

[robertphillips, 2014-09-08 16:48:16]

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

[bsalomon, 2014-09-08 17:52:54]

lgtm

[mtklein, 2014-09-08 18:41:24]

https://codereview.chromium.org/549143003/diff/100001/src/core/SkRecordDraw.h
File src/core/SkRecordDraw.h (right):

https://codereview.chromium.org/549143003/diff/100001/src/core/SkRecordDraw.h...
src/core/SkRecordDraw.h:36: explicit Draw(SkCanvas* canvas)
Would slightly prefer

Draw(SkCanvas* canvas, const SkMatrix* initialCTM = NULL)
  : fInitialCTM(initialCTM ? *initialCTM : canvas->getTotalMatrix())
  , fCanvas(canvas) {}

Then later

PartialDraw(SkCanvas* canvas, const SkRect& rect, const SkMatrix& initalCTM)
  : INHERITED(canvas, &initialCTM), fClearRect(rect) {}

I just don't like protected members.  It can be sometimes tricky to figure out
what's going on when a far-off subclass messes around with things when you don't
expect.  This way fInitialCTM can stay private and const and so a very known
quantity.

[robertphillips, 2014-09-15 18:33:19]

https://codereview.chromium.org/549143003/diff/100001/src/core/SkRecordDraw.h
File src/core/SkRecordDraw.h (right):

https://codereview.chromium.org/549143003/diff/100001/src/core/SkRecordDraw.h...
src/core/SkRecordDraw.h:36: explicit Draw(SkCanvas* canvas)
On 2014/09/08 18:41:23, mtklein wrote:
> Would slightly prefer
> 
> Draw(SkCanvas* canvas, const SkMatrix* initialCTM = NULL)
>   : fInitialCTM(initialCTM ? *initialCTM : canvas->getTotalMatrix())
>   , fCanvas(canvas) {}
> 
> Then later
> 
> PartialDraw(SkCanvas* canvas, const SkRect& rect, const SkMatrix& initalCTM)
>   : INHERITED(canvas, &initialCTM), fClearRect(rect) {}
> 
> I just don't like protected members.  It can be sometimes tricky to figure out
> what's going on when a far-off subclass messes around with things when you
don't
> expect.  This way fInitialCTM can stay private and const and so a very known
> quantity.

Done.

[mtklein, 2014-09-15 19:15:52]

On 2014/09/15 18:33:19, robertphillips wrote:
> https://codereview.chromium.org/549143003/diff/100001/src/core/SkRecordDraw.h
> File src/core/SkRecordDraw.h (right):
> 
>
https://codereview.chromium.org/549143003/diff/100001/src/core/SkRecordDraw.h...
> src/core/SkRecordDraw.h:36: explicit Draw(SkCanvas* canvas)
> On 2014/09/08 18:41:23, mtklein wrote:
> > Would slightly prefer
> > 
> > Draw(SkCanvas* canvas, const SkMatrix* initialCTM = NULL)
> >   : fInitialCTM(initialCTM ? *initialCTM : canvas->getTotalMatrix())
> >   , fCanvas(canvas) {}
> > 
> > Then later
> > 
> > PartialDraw(SkCanvas* canvas, const SkRect& rect, const SkMatrix& initalCTM)
> >   : INHERITED(canvas, &initialCTM), fClearRect(rect) {}
> > 
> > I just don't like protected members.  It can be sometimes tricky to figure
out
> > what's going on when a far-off subclass messes around with things when you
> don't
> > expect.  This way fInitialCTM can stay private and const and so a very known
> > quantity.
> 
> Done.

lgtm

[robertphillips, 2014-09-16 17:24:19]

The CQ bit was checked by robertphillips@google.com

[commit-bot: I haz the power, 2014-09-16 17:24:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patchset/549143003/160001

[commit-bot: I haz the power, 2014-09-16 17:32:48]

Committed patchset #9 (id:160001) as 4815fe5a0a497b676677fb4e4a0f05c511855490