Re: [boost] Boost.Threads - Final Design Decisions

From: John Max Skaller (skaller_at_[hidden])
Date: 2001-08-20 19:02:48

Peter Dimov wrote:

> > Don't be silly. If the design is that close to being
> > able to use shared_ptr on a thread object, the design has
> > succeeded. No one wanted to force you to use the low level
> > non-copyable thread objects, rather, the design _should_
> > make it easy to wrap those objects in a 'standard' ref counting
> > template.
> 
> I thought that it did but I encountered a problem - see my other post.

        And that kind of information is vital.
I read your other post, and I think I agree with you that
you have identified a genuine issue.
 
> I think that I'm close to a prototype thread::ref running over
> boost::thread; I only need to solve the problem with passing the thread proc
> a shared_ptr to the thread object (and to find the time to sit down and
> implement it. Bind reviewers had a rather high standard of acceptance. :-) )

        But now you have a basis for working WITH Bill: if you want
a ref counted thread object, you should be able to build one,
and I'd guess Bill will agree: its a sensible design constraint
on the lower level thread library that it be wrappable.

        Indeed, Felix has a design constraint: only
copyable C++ types may be bound as primitives. So I'm quite
interested in some kind of ref_thread!

-- 
John (Max) Skaller, mailto:skaller_at_[hidden] 
10/1 Toxteth Rd Glebe NSW 2037 Australia voice: 61-2-9660-0850
New generation programming language Felix  http://felix.sourceforge.net
Literate Programming tool Interscript     
http://Interscript.sourceforge.net