RFR: 8178813: Add test for G1 pre-barrier on dereference of weak JNI handles

[Thomas Schatzl]

Hi Kim,

On Wed, 2017-05-10 at 19:39 -0400, Kim Barrett wrote:
> > 
> > On May 9, 2017, at 4:34 AM, Thomas Schatzl <thomas.schatzl at oracle.c
> > om> wrote:
> > 
> > Hi Kim,
> > 
> > On Tue, 2017-05-09 at 10:08 +0200, Mikael Gerdin wrote:
> > > 
> > > Hi Kim,
> > > [...]
> > > On 2017-05-09 07:46, Kim Barrett wrote:
> > > > 
> > > > [...]
> > > > full: http://cr.openjdk.java.net/~kbarrett/8178813/hotspot.01/
> > > > incr: http://cr.openjdk.java.net/~kbarrett/8178813/hotspot.01.i
> > > > nc/
> >   some minor comments:
> Thanks for looking at this.
> 
> > 
> > - I would recommend using whitebox api to issue the correct kind of
> > [...]
> Good point.  Fixed.
> 
> > 
> > - TestJNIWeakG1.java:206: the printed text does not correspond to
> > [...]
> Oops.  Fixed.
> 
> > 
> > - not sure about the need for the error handling using booleans in
> > [...]
> It made it easier to test the test, by backing out the fixes being
> tested and verifying both failure modes.  But I agree it does uglify
> the test itself, and now that we expected it to always pass, that’s
> more important.  So fixed.
> 
> New webrevs:
> full: http://cr.openjdk.java.net/~kbarrett/8178813/hotspot.02/
> incr: http://cr.openjdk.java.net/~kbarrett/8178813/hotspot.02.inc/
> 

  looks good.

Thomas