[OpenJDK Rasterizer] [OpenJDK 2D-Dev] Openjdk java2d rasterizer JEP for pisces (marlin) enhancements ?

[Phil Race]

Hi,

On 03/06/2015 12:37 PM, Laurent Bourgès wrote:
> Phil,
>
> Thanks for your feedback !
>
> 2015-03-06 20:06 GMT+01:00 Phil Race <philip.race at oracle.com 
> <mailto:philip.race at oracle.com>>:
>
>     Hi,
>
>     you placed the test in the java.awt.geom package.
>
>       25 package java.awt.geom;
>
>     and are accessing internals of that package.
>
>     In jigsaw/modular mode that won't even compile.
>
>
> Ok it is annoying:
> as all Path2D fields are package protected, I designed the test using 
> direct access to any fields ...
>
>
>     So the test should go in the anonymous package and avoid accessing
>     internals.
>     It should be possible to use just public API  to verify the
>     arrays  of a shape
>     being cloned are trimmed .
>
>
> No, it is not possible to use Shape API to access arrays nor fields 
> (numTypes ...):
> only getPathIterator() could give me data but it won't tell me if the 
> underlying arrays or fields are correct.

That is true ..

Well, if you need it to be in java.awt.geom, I think even today you'll 
find it won't work
unless you jump through some jtreg hoops to install it on the bootclasspath.
I think its something like "@run main/othervm -Xbootclasspath/a:. "
And later in the modular JDK it will need to be modified again.

I'd say either update the test to work with jtreg today - and test it to 
be sure,
or provide the test without an @test tag, or with an @ignore tag, so 
people can
still manually verify it but the harness won't run it.

>
> Maybe I could use introspection to getDeclaredField(name) and 
> setAccessible(true) to get internal data.

That won't work either. So maybe this is a "noreg-hard" or 
"noreg-cleanup" bug.
We add those labels to the JBS/JIRA bug when something isn't testable.


>
> Any idea or utility class I could use

>
>     Why is it necessary to explicitly add the call to super(); ?
>
>     223             super();
>
>
> I agree it is not necessary but it explicitely says that I use the 
> empty constructor:
>
>     /**
>      * Constructs a new empty {@code Path2D} object.
>      * It is assumed that the package sibling subclass that is
>      * defaulting to this constructor will fill in all values.
>      *
>      * @since 1.6
>      */
>     /* private protected */
>     Path2D() {
>     }

If we all did this, all of the time, there'd be a lot of extra lines in 
the code, that the compiler
would fill in for us anyway.

-phil.

>
> Laurent
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/graphics-rasterizer-dev/attachments/20150306/c595a8f3/attachment-0001.html>