Re: [boost] Re: Re: VC7 and wchar_t in Boost

From: Mattias Flodin (flodin_at_[hidden])
Date: 2002-10-25 00:57:46

On Thu, Oct 24, 2002 at 01:38:34PM -0400, Edward Diener wrote:
> "Beman Dawes" <bdawes_at_[hidden]> wrote in message
> > The default options might be preferred where an option is truly optional.
> >
> > But for options which increase standards conformance, I think we should
> > turn the option on. Particularly if Boost libraries fail without the
> > option. That isn't an option in my mind, its a requirement:-)
> 
> So essentially VC7 now adds the ability to support C++ wchar_t but
> defaults to the previous VC6 situation.

Same thing with the for-scope option. I strongly agree with Beman that
these options are not really optional. For the for scope, it is
difficult to write code that compiles both with and without the option
-- you have to make sure never to reuse a variable name in a second for
loop. I don't think the library writer should be required to keep this
in mind all the time. Trying to stick to the standard takes enough
effort.

Someone ought to write a syntax checker for C++ that checks for ANSI/ISO
compliance of the code (proper use of typename etc). That would make it
so much easier to write portable code. Perhaps something the (to-be?)
C++ parser example in spirit could be used for? :)

-- 
Mattias Flodin <flodin_at_[hidden]>  -  http://www.cs.umu.se/~flodin/
Room NADV 102
Department of Computing Science
Umeå University
S-901 87 Umeå, Sweden
--
"Dew knot trussed yore spell chequer two fined awl yore mistakes."
  -- Brendan Hills