[OpenJDK Rasterizer] [OpenJDK 2D-Dev] Path2D optimizations

[Phil Race]

Hi,

This looks like something it would be good to push to the
http://hg.openjdk.java.net/jdk9/client/jdk repo first.
That can then be backported to 8u60.

Then we'll sync it into graphics-rasterizer as part of a normal JDK 
forest sync.

The only catch is you aren't a JDK 9 committer so you can prepare
a patch, or even a changeset,  but can't do the actual push although
preparing this patch will help get you there. So Jim, or I, or someone
else who is a committer can push it for you.

-phil.

On 3/31/2015 7:33 AM, Laurent Bourgès wrote:
> Jim,
>
>
>         Test program, line 490 - MOVETO has 2 coordinates associated
>         with it.
>
>
>     Well spotted: I did it too quickly, sorry.
>
>         Test program, line 492 - perhaps we should throw an exception
>         on default since it indicates a problem with the iterator
>
>
>     Ok.
>
>
> Here is the new webrev:
> http://cr.openjdk.java.net/~lbourges/path2D/Path2D.4/ 
> <http://cr.openjdk.java.net/%7Elbourges/path2D/Path2D.4/>
>
> I fixed the getLength(type) method:
>   482     static int getLength(int type) {
>   483         switch(type) {
>   484             case PathIterator.SEG_CUBICTO:
>   485                 return 6;
>   486             case PathIterator.SEG_QUADTO:
>   487                 return 4;
> *  488             case PathIterator.SEG_LINETO:
>   489             case PathIterator.SEG_MOVETO:
>   490                 return 2;
> *  491             case PathIterator.SEG_CLOSE:
>   492                 return 0;
> *  493             default:
>   494                 throw new IllegalStateException("Invalid type: " + type);
> *  495         }
>   496     }
>
> If it is good, could you push this patch in both graphics-rasterizer and java2d repositories ?
>
> or do you prefer I push this patch intographics-rasterizer and later it will be merged into java2d and possibly backported to JDK8 ?
>
> Laurent