Date: Thu, 1 Mar 2001 15:05:04 -0600 From: "Jacques A. Vidrine" <n@nectar.com> To: Warner Losh <imp@harmony.village.org>, Garrett Wollman <wollman@khavrinen.lcs.mit.edu>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/usr.bin/make arch.c main.c Message-ID: <20010301150504.A20983@hamlet.nectar.com> In-Reply-To: <20010301145142.A20573@hamlet.nectar.com>; from n@nectar.com on Thu, Mar 01, 2001 at 02:51:42PM -0600 References: <200103011743.MAA01724@khavrinen.lcs.mit.edu> <200103010603.f2163IM00616@freefall.freebsd.org> <200103011743.MAA01724@khavrinen.lcs.mit.edu> <200103011817.f21IHZd50132@harmony.village.org> <20010301145142.A20573@hamlet.nectar.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Mar 01, 2001 at 02:51:42PM -0600, Jacques A. Vidrine wrote:
> On Thu, Mar 01, 2001 at 11:17:35AM -0700, Warner Losh wrote:
> > In message <200103011743.MAA01724@khavrinen.lcs.mit.edu> Garrett Wollman writes:
> > : <<On Wed, 28 Feb 2001 22:03:18 -0800 (PST), Warner Losh <imp@FreeBSD.org> said:
> > :
> > : > MAXPATHLEN includes the trailing NUL.
> > :
> > : Should have also changed the code to use PATH_MAX instead of
> > : MAXPATHLEN.
> >
> > OK. Can you give me a pointer to the definition of PATH_MAX in a
> > standards doc so I do it right?
>
> ISO/IEC 9945-1:1996 (POSIX.1) sections 2.8.2, 2.8.5, and tables 2.3,
> 2.6 give us this:
>
> {_POSIX_PATH_MAX} ``The number of bytes in a pathname.'' Must be
> defined as 255 in <limits.h>. We have this already.
>
> {PATH_MAX} ``Maximum number of bytes in a pathname (not a string
> length; count excludes a terminating null).'' Must be at least
> {_POSIX_PATH_MAX} if defined in <limits.h>. We have this already
> (via syslimits.h).
>
> On some systems, {PATH_MAX} may not be defined in <limits.h>, in which
> case you either have to assume {_POSIX_PATH_MAX} or use pathconf() to
> find the ``real'' value. This is to support the case where the
> maximum path length can vary from one filesystem to another, for example.
>
> Note that {PATH_MAX} _does not_ include a terminating null, and that
> in sys/param.h we define MAXPATHLEN = PATH_MAX. A quick look through
> the sources reveals lots & lots of `bugs' due to this.
>
> Hope this helps.
I got the information above from my hardcopy of IEC/ISO 9945-1:1996/
IEEE Std 1003.1 1996. Then I thought it would be a good idea if I
posted a link to an online version, so I surfed off to www.opengroup.org
to get a link to SUSv2.
http://www.opengroup.org/onlinepubs/007908799/xsh/limits.h.html
I'll be damned if SUSv2 doesn't contradict ISO 9945-1: It says
``PATH_MAX Maximum number of bytes in a pathname, including the
terminating null character.''
I think this is SUSv2 bug. Stevens writes in 1993 that PATH_MAX does
not include room for a null character, and his program fragments in
APUE consistently use PATH_MAX + 1.
Cheers,
--
Jacques Vidrine / n@nectar.com / jvidrine@verio.net / nectar@FreeBSD.org
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?20010301150504.A20983>