RFR: 8064846: Lazy-init thread safety problems in core reflection

[Aleksey Shipilev]

Hi Martin,

On 11/14/2014 03:03 AM, Martin Buchholz wrote:
> I'd like you to do a code review.
> 
> http://cr.openjdk.java.net/~martin/webrevs/openjdk9/core-reflection-volatile/
> https://bugs.openjdk.java.net/browse/JDK-8064846

Looks good and sane, thanks for taking care of it.

Not a big fan of chained assignments though. It looks like you may reuse
the same local variable "superinterfaces" here:

  97     public Type[] getSuperInterfaces() {
  98         Type[] superInterfaces = this.superInterfaces;
  99         if (superInterfaces == null) { // lazily initialize super
interfaces
 100             // first, extract super interface subtree(s) from AST
 101             TypeTree[] ts  = getTree().getSuperInterfaces();
 102             // create array to store reified subtree(s)
 103             superInterfaces = new Type[ts.length];
 104             // reify all subtrees
 105             for (int i = 0; i < ts.length; i++) {
 106                 Reifier r = getReifier(); // obtain visitor
 107                 ts[i].accept(r);// reify subtree
 108                 // extract result from visitor and store it
 109                 superInterfaces[i] = r.getResult();
 110             }
 111             this.superInterfaces = superInterfaces;
 112         }
 113         return superInterfaces.clone(); // return cached result
 114     }


Thanks,
-Aleksey.