Doug Lea and I are going to be giving a talk at JavaOne entitled
"Correct and Efficient Synchronization of Java Threads"
Amoung other things, we are going to be telling people that the
double-check idiom is broken and should not be used.
Meanwhile, Allen Holub and Craig Larman both have JavaOne talks
scheduled in which they will both be telling people about this cool
trick called the "double-check idiom" -- at least, Allan Holub
covered the double-check idiom in his JavaOne talk last year, and
Craig Larman does in his book.
Does anyone know Holub or Larman? They would probably be more
receptive to being told that the double-checked idiom is invalid from
someone they know.
In addition, we will be organizing a BOF to discuss the JSR to change
the Thread specification. Since these BOFs are restricted to be
fairly short, we won't get into much detailed discussion of the
proposals during the BOF. More, it will just be a chance to discuss
material at the level of the JSR, and get people involved in the
discussion and/or the JSR committee.
JavaMemoryModel mailing list - http://www.cs.umd.edu/~pugh/java/memoryModel
This archive was generated by hypermail 2b29 : Thu Oct 13 2005 - 07:00:25 EDT