[cpp-threads] Draft Plan for Threads in C++

[Boehm, Hans]

Lawrence -

I suggested earlier in our discussion that we separate out the
different issues, and not try to keep them in lock-step.  Keeping
the discussion on a single list is probably good though, so we
all keep up on other relevant pieces.

I'm most interested in the memory model piece, and will continue
to push for that.  At least for that piece, I think
we should have a basic proposal (i.e. agreement on major issues,
no standardese, but some idea what changes are needed) by the
Quebec meeting.  I'm sure we'll discover new and interesting
issues after that, though.

I'm planning on regularly updating the strawman proposal in place.
Attaching a date to it is probably not a good idea.

As far as atomic operations are concerned, I separated out
some relevant background references on my page at

http://www.hpl.hp.com/personal/Hans_Boehm/c++mm

I think we're also in tolerable shape there.  Doug led the effort to
do this for Java, and both Alexander and I have also thought about
interfaces that provide finer ordering control.  What we really
need next is an overall framework for the interface.  Should it
provide specific "atomic" classes like the Java one?  Should it
be a template on arbitrary types, which can just say
"not implemented" if you ask for atomic 276 byte structs
with overloaded assignment?  Do we agree that the operations
should be (op, ordering_constraint) pairs?  Is the ordering_constraint
a template parameter?  What about optional hardware-dependent
functionality?

The piece for which in my mind it might be tricky to get
agreement on time is the threads API.

Hans


> -----Original Message-----
> From: 
> Cpp-threads_decadentplace.org.uk-bounces at decadentplace.org.uk 
> [mailto:Cpp-threads_decadentplace.org.uk-bounces at decadentplace
> .org.uk] On Behalf Of Lawrence Crowl
> Sent: Wednesday, April 27, 2005 1:22 PM
> To: cpp-threads at decadentplace.org.uk
> Subject: [cpp-threads] Draft Plan for Threads in C++
> 
> 
> The recent Lillehammer meeting of the ISO C++ standards 
> committee showed significant support for standardizing 
> threads within C++. Unfortunately, there seems to be 
> unrelated proposals for related issues.  The attached paper 
> is a draft attempt to organize the thinking and goals for 
> full thread support in C++.  Please send me your comments by 
> Thursday.  I will collect those comments and submit a revised 
> version in time for the post-Lillehammer mailing on Friday.
> 
> This email has rather broad circulation.  If you get two 
> copies, I apologize.
> 
>   Lawrence Crowl             650-786-6146   Sun Microsystems, Inc.
>                    Lawrence.Crowl at Sun.com   901 San Antonio 
> Road, UMPK16-303
>            http://www.Crowl.org/Lawrence/   Palo Alto, 
> California, 94303
>