Handle convex paths with degeneracies in cheap direction computation

[reed1, 2012-07-10 17:59:45]

http://codereview.appspot.com/6349085/diff/2001/src/core/SkPath.cpp
File src/core/SkPath.cpp (right):

http://codereview.appspot.com/6349085/diff/2001/src/core/SkPath.cpp#newcode2140
src/core/SkPath.cpp:2140: // the same as the orientation of the path).
Is your loop the same as?

    for (++i; i < n; ++i) {
        ...
    }
    return false;

http://codereview.appspot.com/6349085/diff/2001/src/core/SkPath.cpp#newcode2157
src/core/SkPath.cpp:2157: // for use with convex_dir_test
Can we move the two definitions of mul above the template, and then not need to
pass it as a template-parameter?

[bsalomon, 2012-07-10 18:10:30]

http://codereview.appspot.com/6349085/diff/2001/src/core/SkPath.cpp
File src/core/SkPath.cpp (right):

http://codereview.appspot.com/6349085/diff/2001/src/core/SkPath.cpp#newcode2140
src/core/SkPath.cpp:2140: // the same as the orientation of the path).
On 2012/07/10 17:59:45, reed1 wrote:
> Is your loop the same as?
> 
>     for (++i; i < n; ++i) {
>         ...
>     }
>     return false;

Done.

http://codereview.appspot.com/6349085/diff/2001/src/core/SkPath.cpp#newcode2157
src/core/SkPath.cpp:2157: // for use with convex_dir_test
On 2012/07/10 17:59:45, reed1 wrote:
> Can we move the two definitions of mul above the template, and then not need
to
> pass it as a template-parameter?

Done.

[reed1, 2012-07-10 18:25:45]

lgtm

[bsalomon, 2012-07-11 12:40:49]

Landed in 4515