RFR: JDK-8162340: Better class stream parsing

[Coleen Phillimore]

http://cr.openjdk.java.net/~acorn/8162340.hs/webrev/src/share/vm/classfile/systemDictionary.cpp.udiff.html

+// Handles unsafe_DefineAnonymousClass and redefineclasses
+// redefinedclasses do not add to the class hierarchy

Can you change /redefineclasses/RedefineClasses/?  We pattern match on this.

Can you remove the additional blank line at @@ -1418,10 +1393,11 @@

http://cr.openjdk.java.net/~acorn/8162340.hs/webrev/src/share/vm/oops/instanceKlass.cpp.udiff.html

Since check_prohibited_package() is called from instanceKlass, and 
doesn't use any instance variables of either class, it should be a local 
static function defined in instanceKlass.cpp.

http://cr.openjdk.java.net/~acorn/8162340.hs/webrev/src/share/vm/oops/symbol.hpp.udiff.html

Awesome!  I'm glad you found this comment to remove.

Nobody's going to miss parsed_name.  This is a nice change!

Thanks,
Coleen

On 7/21/16 1:23 PM, Karen Kinnear wrote:
> Please review:
> https://bugs.openjdk.java.net/browse/JDK-8162340
>
> http://cr.openjdk.java.net/~acorn/8162340.hs/webrev/
> http://cr.openjdk.java.net/~acorn/8162340.jdk.test/webrev
>
> When moving the check from resolve_from_stream to set_package, I removed
> the no longer need parsed_class (!) which I should have cleaned up when
> tightening the placeholder table cleanup years ago.
>
> Tests run:
> 1. updated VMAnonymousClass test
> 2. jcks
> 3. rbt hs-nightly-runtime linux-x64 in progress
>
> thanks,
> Karen
>
>