Re: New file uploaded to boost

From: williamkempf_at_[hidden]
Date: 2001-06-05 17:16:16

--- In boost_at_y..., <boost_at_y...> wrote:
> 
> Hello,
> 
> This email message is a notification to let you know that
> a file has been uploaded to the Files area of the boost 
> group.
> 
>   File        : /threads/threads2.zip 
>   Uploaded by : williamkempf_at_h... 
>   Description : Temporary refactoring of Boost.Threads.  
> 
> You can access this file at the URL
> 
> http://groups.yahoo.com/group/boost/files/threads/threads2.zip 
> 
> To learn more about file sharing for your group, please visit
> 
> http://help.yahoo.com/help/us/groups/files
> 
> Regards,

* Finished refactoring the mutexes.
* Temporarily(?) removed atomic_t.
* Added support for the Boost.Build system.

I've been discussing issues with the concept of an atomic_t type with 
Alexander Terekhov.  Successfully using a concept like this is going 
to be very problematic, and the main use that people want to put it 
to (ref-counted object management) may well be better done using 
mutexes.  Considering all of this I don't want to include atomic_t in 
Phase I of Boost.Threads and have removed it from the library.  This 
particular concept is going to require a lot of discussion and 
careful design to get "right", i.e. to make it useful for portable 
code.  I don't expect to add the concept back into the library until 
at least Phase III.  Phase II will provide enough support for most MT 
programming and I expect to start work on it soon.

I still haven't settled on which timed locks to support, so currently 
both are included.  I'm leaning towards leaving only the xtime 
variants but really want to hear opinions on this one.

The jamfiles work nicely for me but they've only been tested on MSVC 
and Win2k.  It would be nice to have people attempt to get it to work 
on other compilers/platforms to give the Boost.Build system a real 
world work out.  If changes are required to make it work I'll gladly 
make them.

Bill Kempf