Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 12 Apr 1997 12:26:52 -0700 (MST)
From:      Terry Lambert <terry@lambert.org>
To:        tom@uniserve.com (Tom Samplonius)
Cc:        langfod@dihelix.com, bradley@dunn.org, freebsd-isp@FreeBSD.ORG, freebsd-current@FreeBSD.ORG
Subject:   Re: longer usernames
Message-ID:  <199704121926.MAA15601@phaeton.artisoft.com>
In-Reply-To: <Pine.NEB.3.95.970411174629.579g-100000@haven.uniserve.com> from "Tom Samplonius" at Apr 11, 97 05:49:56 pm

next in thread | previous in thread | raw e-mail | index | archive | help
>   Misinformation being passed around by the misinformed.  NIS is concerned
> with key-data pairs, the lengths of those pairs is imposed only by the
> underlying database (and a couple of other things).
> 
>   Even the SunOS 4.1 implementation of NIS has no problems with
> longer-usernames, its just that login truncates all user-ids to 8
> characters before doing lookups (I guess, it is hard to verify without
> looking at the code).  I did some real brief testing on this.
> 
>   Solaris 2.x has 16 character username support, and it definitely works
> with NIS.

Yes; the problem is interoperability with systems which have
historically truncated login names to 8 characters; all accounts
more than 8 characters will be useless on these systems when they
are configured as clients of a system which supports more than
8 character names.

So if you support more than 8 character names, and send them out
via NIS, older SGI, MIPS, Pyramid, and SVR3/SVR4 NIS clients will
be unable to log into any account with a long name.

In a homogeneous environment, there is no problem.

In an environment with Solaris 2.x, there is no interoperability
problem as long as you geep the names shorter than 17 characters.

In other words, over 8 characters, the behaviour is undefined in
all but homogeneous environments, with the above documented
exceptions.

Feel free to buy one of each type of UNIX box ever built, and
run each OS release they've ever had, and document the remaining
exceptions in a cannonical list.  8-) 8-).


					Regards,
					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199704121926.MAA15601>