RFR: 8239235: Examine SignatureStream performance after consolidation

[coleen.phillimore at oracle.com]

http://cr.openjdk.java.net/~redestad/8239235/open.01/src/hotspot/share/runtime/signature.cpp.udiff.html

+ _previous_name = vmSymbols::java_lang_Object();


Can you add a short comment above this explaining why making this a 
permanent symbol eliminates null checks and refcounting is a nop? It 
took me a minute to realize why you did this.

http://cr.openjdk.java.net/~redestad/8239235/open.01/src/hotspot/share/classfile/symbolTable.cpp.udiff.html

Nice.

I see the bug.  Now it's simpler too.  This change looks good. Thank you 
for adding the gtest.

Coleen



On 2/21/20 9:04 AM, Claes Redestad wrote:
>
>
> On 2020-02-21 11:48, Claes Redestad wrote:
>> Done
>
> Forgot to link to the updated webrev:
>   http://cr.openjdk.java.net/~redestad/8239235/open.01/
>
> /Claes