RFR 8027137: Merge ScriptFunction and ScriptFunctionImpl

Attila Szegedi attila.szegedi at oracle.com
Wed Sep 9 11:37:32 UTC 2015


Excellent, +1.

> On Sep 9, 2015, at 1:22 PM, Sundararajan Athijegannathan <sundararajan.athijegannathan at oracle.com> wrote:
> 
> Thanks for the reviews Hannes & Attila.
> 
> I've incorporated the changes suggested. Updated webrev: http://cr.openjdk.java.net/~sundar/8027137/webrev.01/
> 
> PS. I've changed all counters (used from NativeDebug) to use LongAdder.
> 
> -Sundar
> 
> On 9/9/2015 2:57 PM, Attila Szegedi wrote:
>> Great work! I’m very happy that ScriptFunctionImpl is going away.
>> 
>> Some implementation remarks:
>> 
>> - The newly public methods in Global.java, getFunctionPrototype and getTypeErrorThrower don’t have Javadoc.
>> - The newly public two constructors of PrototypeObject also don’t hava Javadoc.
>> -  while you're at it, how about we replace ScriptFunction.constructorCount/invokes/allocations fields with LongAdder objects? We should gradually be replacing all such static counters.
>> - can we create an explicit subclass for bound functions? It can be declared as "private static class Bound extends ScriptFunction { ... }" within the ScriptFunction class. When we debug, we'll then see its name as ScriptFunction$Bound instead of ScriptFunction$1.
>> 
>> Attila.
>> 
>>> On Sep 9, 2015, at 10:19 AM, Sundararajan Athijegannathan <sundararajan.athijegannathan at oracle.com> wrote:
>>> 
>>> Please review http://cr.openjdk.java.net/~sundar/8027137/ for https://bugs.openjdk.java.net/browse/JDK-8027137
>>> 
>>> Thanks,
>>> -Sundar
> 



More information about the nashorn-dev mailing list