JavaMemoryModel: Javasoft bug reports

From: Bill Pugh (pugh@cs.umd.edu)
Date: Sat Apr 15 2000 - 08:49:57 EDT


One of the issues we've been discussing is the fact the final fields
can be changed by either reflection or native code.

The current spec for reflection states that reflection should _not_
be able to change final fields.
Bug# 4250960
        http://developer.java.sun.com/developer/bugParade/bugs/4250960.html
states that this is now enforced in version 1.3. So the only issue we
need to worry about is
native code, and the back door for System.in, System.out and System.err.

Bug report 4242244, we states that Sun's JVM was in violation of the
Spec (because of the reads kill issue) was closed a month or so,
because the bug report was classified as only applying to their old
JIT, but not HotSpot.

I've entered a new bug report 4329831 that HotSpot is in violation of
the Spec because of the reads kill issue. If you cast votes in the
Bug parade, please change your votes from 4242244 to 4329831.

        Bill
-------------------------------
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