[cpp-threads] Update

[Boehm, Hans]

In spite of the fact that this list has been rather quiet, several
things are going on that might be of interest:

There will be an unofficial get-together of a subset of the C++
standards committee to discuss memory model and threads API issues on
August 23-35 in Redmond, WA.  If you are interested in attending, please
let me and Herb Sutter (the host) know asap.

Clark Nelson and I have started to work on turning the memory model
proposal into standardese.  We hope to post something in a few weeks.

I made another pass over the atomic operations proposals that were
previously discussed here, and turned them into N2047
(http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2006/n2047.html).
Lawrence and I had previously discussed making the low level interface
look more like the high level one, and operating only on specially
declared data.  However, it was pointed out to me that this makes it
difficult to convert existing code that uses something like the gcc
__sync_ primitives or the Microsoft InterlockedX primitives.  I ended up
also taking out the volatile qualifiers, for compatibility with the gcc
approach, and in expectation of a possible decision not to use volatile
for anything thread related.

Hans