>Date: Mon, 19 May 1997 10:37:23 -0700
>From: "J.T. Conklin" <jtc@NetBSD.ORG>

[lukem wrote:]
>> To do this, I've been converting the use of atoi() to strtoul(), and
>> checking appropriate values to ensure that the parsed field was just a
>> number.
>> 
>> However, on the alpha, the return value of strtoul() is a unsigned 64
>> bit long, which may exceed the size of uid_t (unsigned 32 bit int).
>> 
>> So, I would like to create the following defines in <limits.h> or
>> <machine/limits.h>, and check against them in the relevant areas:
>> 	#define		UID_MAX		4294967294	/* 2^32 - 2 */
>> 	#define		GID_MAX		4294967294
>
>Hmm, if my rusty memory cells aren't failing me, I think that XPG4.2
>may have specified macros for this purpose.  Unfortunately, I don't
>have the documents with me; and the MAXUID I found by grep'ing thru
>Solaris 2.5 headers doesn't ring a bell.
>
>If there is an established standard (or even de facto convention) for
>uid/gid maximum manifest constants, we should use those names.
>Otherwise, UID_MAX and GID_MAX are fine with me.

I took a look at some of the systems we use at ADI.  I found that all
of them have a MAXUID definition in <sys/param.h>, and all but one
(SunOS 4) have an identical UID_MAX definition in <limits.h>.  The
values used were:

OS		MAXUID		UID_MAX
------------------------------------------------
SunOS 4.1.4	0xfffd		(none)
SunOS 5.5.1	2147483647	2147483647
Irix 5.2	60002		60002
HP-UX 9.07	60000		MAXUID
HP-UX B.10.20	0x7fffffff	MAXUID

I was unable to find any evidence of anything like GID_MAX; I think
all of these systems may just use UID_MAX or MAXUID instead.
-- 
Mike Long <mikel@shore.net>                http://www.shore.net/~mikel
"Every normal man must be tempted at times to spit on his hands,
hoist the black flag, and begin slitting throats." -- H.L. Mencken