Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 26 Feb 2002 01:25:49 -0500
From:      Mike Barcroft <mike@FreeBSD.org>
To:        Bruce Evans <bde@zeta.org.au>
Cc:        Mark Murray <mark@grondar.za>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/include grp.h
Message-ID:  <20020226012549.H31007@espresso.q9media.com>
In-Reply-To: <20020226162341.O43512-100000@gamplex.bde.org>; from bde@zeta.org.au on Tue, Feb 26, 2002 at 04:37:44PM %2B1100
References:  <200202260015.g1Q0FgPA005896@grimreaper.grondar.org> <20020226162341.O43512-100000@gamplex.bde.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Bruce Evans <bde@zeta.org.au> writes:
> No.  It is very old namespace pollution (it has been there since rev.1.1
> like that in <unistd.h>), so it is too late to fix it in ways that are

In my tree, <unistd.h> only includes <sys/types.h> in the non-
_POSIX_SOURCE case, which is somewhat better than what we have now.
I'm interested in fixing everything in our tree that depends on this
brokenness and freeing <unistd.h> of <sys/types.h>.

> wrong in current standards.  Even removing the pollution will cause a
> few problems.  Hopefully not as many as in <sys/stat.h>, since <sys/types.h>
> is not so polluted.  AFAIR, its pollution consists mainly of:
> - select macros (POSIX.1-2001 does something about this)
> - ntoh* and endian functions and macros (mostly already fixed in -current by
>   mike)
> - too many typedefs for type names ending in _t.  Not a bug, but it is a bug
>   if applications depend on the nonstandard ones.
> - typedefs for type names not ending in _t.  Mostly compatibility cruft for
>   old BSD and SYSV types, but a large number of things depend on the old
>   BSD types (u_char, etc.  We have an even sillier problem for the not-so-old
>   BSD fixed-width types.  They are spelled u_int8_t, etc. in FreeBSD but
>   uint8_t etc. in C99.  I expect the old spellings will take longer to go
>   away than __P(()) :-(.

This could probably be incorporated into the __P(()) removal effort.
It's a very mechanical change in nature.

Best regards,
Mike Barcroft

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020226012549.H31007>