RFR: 8191870: Remove badJNIHandle

[Kim Barrett]

> On Nov 28, 2017, at 9:58 AM, coleen.phillimore at oracle.com wrote:
> 
> http://cr.openjdk.java.net/~kbarrett/8191870/open.01/src/hotspot/share/runtime/jniHandles.hpp.udiff.html
> 
> I think you need PRODUCT_RETURN for zap() in case the optimized target still builds.  Otherwise, looks good.
> 
> Coleen

I think PRODUCT_RETURN would break optimize builds.

The definition of zap in the .cpp file is wrapped in #ifdef ASSERT.
NOT_DEBUG_RETURN's definition is conditional on ASSERT.  So those
match.

Using PRODUCT_RETURN would require the .cpp definition to be wrapped
in #ifndef PRODUCT.  But that combination would make zapping happen for
optimize builds, which I think is wrong.