Re: Patch: Thread.holdLock implementation in java

[Steven Augart]

Dalibor Topic wrote:
> Hi all,
>
> I've implemented VMThread.holdLock in java. It's is probably quite slow, 
> and wakes up the ocassional thread without much use, but hey, it's in 
> java, so that's one less method to implement. OK to commit?

I strongly support committing this patch.  I am now convinced that it 
does not violate Java semantics.  I attended a presentation by Doug 
Lea a few months ago (benefit of working at Watson) where he pointed 
out that no Java implementation, including Sun's, implements the Java 
memory model as specified.  He didn't mention this in particular, but 
it fits in with the spirit of his talk.

Dalibor, you did a great job on researching this and on arguing your 
case.  It was informative for me to read the references you carefully 
looked up.

--Steve Augart