Re: C standard, wint_t, and stddef.h

[Luc Morin <luc_m@videotron.ca>]

Ulrich Drepper wrote:
> 
> "Mark E." <snowball3@usa.net> writes:
> 
> > Yes, DJGPP does use its own libc.
> 
> I'm not talking about the libc, I'm talking about gcc's stddef.h.  And
> this has to contain wint_t and wchar_t.  There is nothing to reconsider.
> 
> --
> ---------------.      drepper at gnu.org  ,-.   1325 Chesapeake Terrace
> Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
> Cygnus Solutions `--' drepper at cygnus.com   `------------------------


I was just browsing through old messages on the list when I came across
this (must have missed it when it was originally posted).

You say stddef.h "has to contain" win_t and wchar_t. If this is the
case, how can it be missing from my version of stddef.h, since my
system's compiler is gcc (2.7.2.1) ? I use this gcc to compile egcs, and
still, no win_t and wchar_t in stddef.h. BTW, this is on FreeBSD
3.1-stable, with egcs-1.1.2.

When I asked the FreeBSD egcs port maintainer about this, here's what he
replied:

"While I look into this, please ask Bejamin Koznik which ISO/ANSI/IEEE
standard addresses ``wint_t'' and (if he is willing) what does it say
about it."

So far, all I can see is that the problem won't get adressed because
each party seems to be pointing the other for the reason why libstdc++v3
won't build on FreeBSD.

Please, all I'm trying to do is find out how to solve the problem, but
my skills and time resources won't allow me to dig into the problem as
much as I'd like to. That's why I'm asking for help. 

I wish someone from your group would contact the FreeBSD egcs
maintainer, Mr. David O'Brien <obrien@NUXI.com>, to see if there's
anything that can be done to solve the problem.

Thank you.

-- 
Luc Morin    <luc_m@videotron.ca>
Electrical Engineering Technologist