Re: review requiest: config_06.zip

From: williamkempf_at_[hidden]
Date: 2001-08-02 08:42:07

--- In boost_at_y..., John Maddock <John_Maddock_at_c...> wrote:
> 
> >BOOST_HAS_THREADS and BOOST_NO_THREADS conflict.  The 
Boost.Threads 
> >library only uses BOOST_HAS_THREADS and any need to check for the 
> >opposite only requires a boolean not operation.  One or the other 
> >should go away, and I think you know which I'd vote to keep ;).
> 
> I take your point - however something I've been asked for in the 
regex lib
> is the ability to turn off threading support, even if the compiler 
thinks
> it's in multithread mode.  In other words BOOST_HAS_THREADS gets 
turned on
> automatically if one of: _MT __MT__, _REENTRANT or _PTHREADS is 
defined,
> *unless* BOOST_NO_THREADS is defined.  The logic is also in place 
in the
> config headers to make sure that these two are mutually exclusive, 
with
> BOOST_NO_THREADS taking precedence.  BTW this is very similar to 
the way in
> which STLport does things...

Ahh... I get the reason behind this now.  Dangerous, but *possibly* 
useful.  If we're to keep this, however, I'd suggest a different name 
to indicate what this macro is really for.  Maybe 
BOOST_DISABLE_THREADS?

Bill Kempf