Re: [boost] Re: RFC: Multithreading design constraints

From: Dietmar Kuehl (dietmar_kuehl_at_[hidden])
Date: 2000-06-06 11:10:17

Hi,

--- John Maddock <John_Maddock_at_[hidden]> wrote:
> >As for implementing a version of the library in terms separate from
> the pthreads API: the C++ function overloading mechanism is wonderful
> at handling this sort of this.
> <
> 
> Why not down the iostreams route, and allow the objects contructors
> to be
> overloaded for native types (in  the way the fstreams can sometimes
> be
> constructed from platform specific file handles).

I'm strongly in favour of the idiom which uses automatic releases of
locks in the destructor. The constructor might use objects of the
underlying thread implementation to provide access to the low level
mechanism. However, I would like to have an implementation with no such
extensions to allow portability testing. .. but this can be seen as a
kind of QoI issue when multiple vendors start providing Boost
implementations :-)

=====
<mailto:dietmar_kuehl_at_[hidden]>
<http://www.dietmar-kuehl.de/>

__________________________________________________
Do You Yahoo!?
Yahoo! Photos -- now, 100 FREE prints!
http://photos.yahoo.com