Add new skyline Rectanizer

Add new skyline Rectanizer

The skyline Rectanizer tracks the leading silhouette of the
currently packed rects and tries to fit the next rectangle
on top of that. This gives slightly better packing behavior
than the current log2 scheme.

Also shrank the size of the plots, which increases their
number from 18 to 32 and gives better cache behavior (fewer
purges and uploads).

Committed: http://code.google.com/p/skia/source/detail?r=11577

[bsalomon, 2013-10-02 16:40:06]

lgtm with some nits

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
File src/gpu/GrRectanizer_skyline.cpp (right):

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
src/gpu/GrRectanizer_skyline.cpp:71: if (rectangleFits(i, width, height, &y)) {
this->

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
src/gpu/GrRectanizer_skyline.cpp:84: addSkylineLevel(bestIndex, bestX, bestY,
width, height);
this->

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
src/gpu/GrRectanizer_skyline.cpp:99: if (x + width > this->width())
{}

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
src/gpu/GrRectanizer_skyline.cpp:106: {
brace goes on prev line (a bunch in the next function as well)

[tfarina, 2013-10-02 16:49:58]

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
File src/gpu/GrRectanizer_skyline.cpp (right):

https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
src/gpu/GrRectanizer_skyline.cpp:71: if (rectangleFits(i, width, height, &y)) {
On 2013/10/02 16:40:06, bsalomon wrote:
> this->
there may be a better place to ask, but what is the reason to use "this->" all
over the place for member functions? IMO they just add noise, but there may be
some hidden benefit that I'm not aware. Certainly "this->" isn't used in
Chromium and isn't encouraged there.

[tomhudson, 2013-10-02 16:52:20]

On 2013/10/02 16:49:58, tfarina wrote:
>
https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
> File src/gpu/GrRectanizer_skyline.cpp (right):
> 
>
https://codereview.chromium.org/25584002/diff/11001/src/gpu/GrRectanizer_skyl...
> src/gpu/GrRectanizer_skyline.cpp:71: if (rectangleFits(i, width, height, &y))
{
> On 2013/10/02 16:40:06, bsalomon wrote:
> > this->
> there may be a better place to ask, but what is the reason to use "this->" all
> over the place for member functions? IMO they just add noise, but there may be
> some hidden benefit that I'm not aware. Certainly "this->" isn't used in
> Chromium and isn't encouraged there.

The Skia coding standard isn't the same as the Chrome coding standard.
The justification isn't really discussed at
https://sites.google.com/site/skiadocs/developer-documentation/contributing-c...,
but my understanding was always that it made it particularly clear when you're
calling a member function. I've found it useful, and not just noise, when I've
been doing a lot of work in Skia.

[commit-bot: I haz the power, 2013-10-02 17:28:52]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/jvanverth@google.com/25584002/21001

[commit-bot: I haz the power, 2013-10-02 17:38:02]

Change committed as 11577