From: Beman Dawes (beman_at_[hidden])
Date: 2000-05-09 14:47:08
At 10:46 PM 5/7/00 +0100, Paul Moore wrote:
>In my view, for Boost to be credible, it needs to be of practical
>it or not, this means having working implementations on the
>compilers. For Unix, I assume this means GCC. For Microsoft, GCC is
>(with mingw and cygwin available), but I'd like to see Borland and
>least partially supported. I don't know for Mac or other platforms.
>How about a statement to the effect of
>* Is portable and not restricted to a particular compiler or
>system. Implementations should ideally run on currently available
>(gcc, MSVC and Borland C++ are widely used). Providing limited
>where compiler features are missing is acceptable.
>[[I just spent 15 minutes staring at this posting, trying to word a
>along the lines of "I hate MSVC, but I am forced to use it, I want
>good C++ libraries and I'm sick of finding the good libraries all
>disclaimers along the lines of "we don't support MSVC because it's
>broken"". Thankfully, I decided not to include it. :-) :-(]]
In the update just posted, I avoided naming any specific compiler or
vendor. Most of the compilers seem to be trying to improve
compliance. We will have to wait until July for the beta of MSVC++ 7
to see if that statement applies to Microsoft.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk