RFR (M): 8000740: remove LinkWellKnownClasses

[Christian Thalinger]

On Oct 11, 2012, at 12:00 PM, John Rose <john.r.rose at oracle.com> wrote:

> The change in methodHandles.cpp is not under -XX:+LinkWellKnownClasses.
> As the comment says, it is intended to be an optional optimization for 292 users.
> The short-cut for Object (and maybe Class) is probably necessary for bootstrapping.
> I don't think there will be a regression there, but let's be watchful.

I ran the JDK 292 tests and Nashorn stuff and it looked good.  Thanks for the review.

-- Chris

> 
> Code removal is good...  Reviewed.
> 
> — John
> 
> On Oct 10, 2012, at 5:10 PM, Christian Thalinger wrote:
> 
>> http://cr.openjdk.java.net/~twisti/8000740/
>> 
>> 8000740: remove LinkWellKnownClasses
>> Reviewed-by:
>> 
>> With 8000263 LinkWellKnownClasses was turned on by default to see if
>> it works. Nightly testing showed that various JDI and JVMTI tests fail
>> with LinkWellKnownClasses turned on, e.g.:
>> 
>> jdk/test/com/sun/jdi/NewInstanceTest.java 
>> 
>> To avoid future problems we remove the code altogether.
>> 
>> Additionally I resurrected some 292 code that is required when
>> building with JDK 6.
>> 
>> src/share/vm/classfile/classFileParser.cpp
>> src/share/vm/classfile/classFileParser.hpp
>> src/share/vm/classfile/systemDictionary.cpp
>> src/share/vm/classfile/systemDictionary.hpp
>> src/share/vm/prims/methodHandles.cpp
>> src/share/vm/runtime/globals.hpp
>> 
>