Fix the problem of rendering closePath not properly after a moveTo call in

Fix the problem of rendering closePath not properly after a moveTo call in
canvas 2D interface. If there is a polyline, followed by a moveTo and a
closePath, both the moveTo and the closePath should be ignored for the purposes
of drawing, and the polyline should not be closed (unless force closed is true
(for filling, for instance).

Tested for path with both valid and degenerate content, when asked to
consume degenerates and not, force closed and not.

This patch also includes a uni test refactoring to reduce the amount of code
to test path iteration and zero length paths.

BUG=6297049

TEST=tests/PathTest.cpp, testIter method.
Committed: https://code.google.com/p/skia/source/detail?r=4247

[reed1, 2012-06-12 14:31:20]

Is there a more compact, data-drive way to structure the unit-tests? That is a
huge amount of vertical real estate to have to read and manually confirm that it
matches.

[Stephen Chenney, 2012-06-12 14:34:25]

On 2012/06/12 14:31:20, reed1 wrote:
> Is there a more compact, data-drive way to structure the unit-tests? That is a
> huge amount of vertical real estate to have to read and manually confirm that
it
> matches.

Yeah. It's not pretty. I'll put some time into improving all of the iteration
tests.

[reed1, 2012-06-13 16:09:04]

much more gooder.

Would some local macros help condense those long lines?

e.g.

#define M  SkPath::kMove_Verb

not required, but might help.

lgtm

[reed1, 2012-06-13 16:09:43]

can we restore the { } braces?

https://codereview.appspot.com/6300086/diff/3001/src/core/SkPath.cpp
File src/core/SkPath.cpp (right):

https://codereview.appspot.com/6300086/diff/3001/src/core/SkPath.cpp#newcode1489
src/core/SkPath.cpp:1489: // follows a move which follows a segment.
Why did we lose the { } braces?

[Stephen Chenney, 2012-06-13 16:24:06]

I'll leave the test as it is for now. Things are confusing enough as it is
without adding another layer of indirection.

http://codereview.appspot.com/6300086/diff/3001/src/core/SkPath.cpp
File src/core/SkPath.cpp (right):

http://codereview.appspot.com/6300086/diff/3001/src/core/SkPath.cpp#newcode1489
src/core/SkPath.cpp:1489: // follows a move which follows a segment.
On 2012/06/13 16:09:43, reed1 wrote:
> Why did we lose the { } braces?

My brain is locked into WebKit standards. I'll re-add the braces.

[Stephen Chenney, 2012-06-13 17:06:14]

Committed revision 4247.

[yunchao.he, 2012-06-14 02:54:33]

On 2012/06/13 17:06:14, Stephen Chenney wrote:
> Committed revision 4247.

Hi, stephen, I thought you are a committer, then I CC to you in skia community.
But you commit a new patch as your own according to the
problem I state clearly, and exactly the code location I provide, even you code
are exactly the same meaning as mine--just simplified my code. As a committer,
It is the right way? I am a new comer, don't know the way it is.

If you are not a committer, it is my fault.

[yunchao.he, 2012-06-14 08:28:49]

On 2012/06/14 02:54:33, yunchao.he wrote:
> On 2012/06/13 17:06:14, Stephen Chenney wrote:
> > Committed revision 4247.
> 
> Hi, stephen, I thought you are a committer, then I CC to you in skia
community.
> But you commit a new patch as your own according to the
> problem I state clearly, and exactly the code location I provide, even you
code
> are exactly the same meaning as mine--just simplified my code. As a committer,
> It is the right way? I am a new comer, don't know the way it is.
> 
> If you are not a committer, it is my fault.

I wonder whether I didn't write test case, I found that you wrote a test case
for it, in addition to simplified my code.

[Stephen Chenney, 2012-06-14 15:12:25]

On 2012/06/14 08:28:49, yunchao.he wrote:
> On 2012/06/14 02:54:33, yunchao.he wrote:
> > On 2012/06/13 17:06:14, Stephen Chenney wrote:
> > > Committed revision 4247.
> > 
> > Hi, stephen, I thought you are a committer, then I CC to you in skia
> community.
> > But you commit a new patch as your own according to the
> > problem I state clearly, and exactly the code location I provide, even you
> code
> > are exactly the same meaning as mine--just simplified my code. As a
committer,
> > It is the right way? I am a new comer, don't know the way it is.
> > 
> > If you are not a committer, it is my fault.
> 
> I wonder whether I didn't write test case, I found that you wrote a test case
> for it, in addition to simplified my code.

Hi Yunchao, sorry I didn't originally add you as a reviewer on this patch. It
was oversight on my part. I should also have added your name in the Changelog.
My apologies.

You did the right thing in initially investigating and proposing a patch. I
grabbed it because the testing was quite complicated and needed to cover all of
the ways in which the path iterator can be used. That testing led to the code
changes I made.