[OpenJDK 2D-Dev] RFR 8144938: Handle properly coordinate overflow in Marlin Renderer

Jim Graham james.graham at oracle.com
Fri Mar 18 00:44:54 UTC 2016

Hi Laurent,

That looks fine.  If the subPathStarted changes are the only changes 
then it's good to go.  If there was some other change that I missed, let 
me know so I can review it as well.

One curiosity - in the test case, why are there 3 for loops that perform 
exactly one pass?  (lines 134, 221, 282)  I'm guessing that maybe in a 
former life this code may have been part of a performance suite or 
something?  It's not a bug, just an odd bit of code.  You can delete 
them if you want, but I don't need to see a review of that...


On 3/17/16 10:15 AM, Laurent Bourgès wrote:
> Hi Jim,
> Here is the updated webrev using your approach:
> http://cr.openjdk.java.net/~lbourges/marlin/marlin-8144938.4/
> It is more close to 2nd webrev as it does not set subpathStarted to
> false in SEG_CLOSE case.
> Laurent
> 2016-03-17 1:49 GMT+01:00 Jim Graham <james.graham at oracle.com
> <mailto:james.graham at oracle.com>>:
>     Hi Laurent,
>     I'm not sure why you added the pathClosed variable back.  Before
>     your fix we used to rely on all of the downstream consumers
>     correctly implementing the case of CLOSE followed by a non-MOVE
>     command so specifically enforcing a MOVE here should be unnecessary
>     and just complicates the code...?
>                              ...jim

More information about the 2d-dev mailing list