> The notion we have now is to introduce a method called (say) public
> void java.lang.System.runInIsolation(Runnable), which invokes the
> run() method of its parameter. Anything run inside this method is
> treated as a separate thread for the purposes of determining the
> value of final fields.
> What are your thoughts on this? Is it a reasonable compromise? Does
> it have JSR-166 implications (my guess is "yes")? Other names /
> implementations are welcome.
Name-wise it clashes quite nastily with JSR-121 ... I guess there are
some conceptual similarities, but the isolation there is much stronger
than the kind of isolation you want for JSR-133. To avoid confusion,
I'd recommend you pick a name which suggests your methods purpose a
little more precisely. I'm afraid I don't have any very good
suggestions: runWithBarrierForFinalPropagation() is a bit unweildy ;-)
JavaMemoryModel mailing list - http://www.cs.umd.edu/~pugh/java/memoryModel
This archive was generated by hypermail 2b29 : Thu Oct 13 2005 - 07:00:44 EDT