Re: [boost] Boost.Threads - thread_desc concept

From: Greg Colvin (gcolvin_at_[hidden])
Date: 2001-06-27 16:26:58

From: Beman Dawes <bdawes_at_[hidden]>
> At 04:43 PM 6/27/2001, Beman Dawes wrote:
> 
>  >On Thu, 28 Jun 2001, John Max Skaller wrote:
>  > > class thread_api
>  > > // lighweight API wrapper
>  > > // purpose: platform independence
>  > >
>  > > class posix_api : public thread_api
>  > > // posix extension
>  > >
>  > > class win32_api: public thread_api
>  > > // win32 extension: boost only
>  > >
>  > > class thread_id
>  > > // copyable thread id wrapper
>  > >
>  > > class my_favourite_thread_manager
>  > > // design competetion on boost to choose
>  > > // one or more alternatives
>  >
>  >I haven't thought in terms of that many layers, but a layered approach is 
> 
>  >very much the approach I'd like to see explored.
> 
> That isn't quite right.  2nd try:  A layered approach is the very much the 
> approach I'd would have explored myself, but developers with threading 
> experience don't agree.  So after much effort making sure their solution is 
> well-thought out and tailored to the peculiarities of threads, I'm becoming 
> convinced we should follow their advice.

I don't mine winning, but I am worried that my recent
threading experience is mostly with implementing the
java thread classes, rather than writing lots of code
that uses threads, so my experience is biased.

That is, the horse might not be dead yet, only sleeping.