[E-Lang] Speaking of Community Process: Memory Model & Threads

[Mark S. Miller]

At 11:57 AM Sunday 7/8/01, Chris Hibbert wrote:
>JavaOne 2000 BOF on revising the Java Thread Spec 
>(http://www.cs.umd.edu/~pugh/java/memoryModel/JavaOneBOF)

Wow.  Java's threading is way more broken than I thought!  I just talked to 
Dean, and it sounds like all E implementations to date may fall into some of 
these problems on shared memory multiprocessor machines (SMPs), and possibly 
even on uni-processor Solaris JDKs as well.  E has never been tested on 
SMPs, and should not be assumed to work there until further notice.  
Unfortunately, we currently have to state the same caution for Solaris.

However, to fix this situation involves only fixing the E implementation in 
a handful of places.  As far as we can tell, neither the current nor the 
proposed Java concurrency semantics should have any effect on the E language 
design, or on the code written by the E programmer.


>I believe you meant to provide this URL for Doug:
>http://www.eros-os.org/mailman/listinfo/e-lang

Yes, thanks.


        Cheers,
        --MarkM