Date: Thu, 20 Jun 1996 20:00:45 -0700 (PDT) From: Tom Samplonius <tom@sdf.com> To: Michael Smith <msmith@atrad.adelaide.edu.au> Cc: hackers@freebsd.org Subject: Re: Longer usernames? Message-ID: <Pine.NEB.3.93.960620194412.2813B-100000@misery.sdf.com> In-Reply-To: <199606210218.LAA08107@genesis.atrad.adelaide.edu.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 21 Jun 1996, Michael Smith wrote: > Tom Samplonius stands accused of saying: > > > > > > > > What is the general consensus on usernames longer than 8 chars? I've > > > > been doing some testing on a -stable system with 16 character usernames, > > > > and everything seems to work ok. > > > > > > NIS won't. > > > > Really? NIS supports databases of arbitrary key-data pairs. Keys can > > be longer than 8 characters. > > > > On FreeBSD systems, long usernames in /var/yp/master.passwd propogate to > > slaves just fine, and if the NIS clients have been built to handle longer > > usernames, those users can login too. > > That's the problem though; other hosts that _don't_ support longer usernames > won't be able to interoperate with FreeBSD's NIS. We might as well call it > NIS++ then 8) The whole idea behind doing NIS at all is to be able to > work in a mixed environment. It's a little more subtle that than. Using long username in NIS does work with a non-FreeBSD NIS slave server (SunOS 4.1) that I tried. It's just that that non-FreeBSD NIS clients that have a smaller username length, fail to work with these long usernames as predicted. I would presume that other os'es that support longer usernames would work just fine. Don't some SYSV systems have 12 character usernames? Doesn't Solaris 2.x? How about IRIX and Digital UNIX? The whole idea behind NIS is to centralize user and network settings. If that means working in a mixed environment, so be it. > I appreciate your basic idea though; it would be nice to come up with a > solution that didn't involve a complete rebuild to swap from one to the > other. Why don't we make the global change to 16, similar to BSDI, and just print warnings in pwd_mkdb on reading a longer username? Also, adding a check to NIS map build process should adequately warn those who cross the line. This way you don't need to rebuild the world, and several of the packages in order to change this limit. Instead we'd just a impose a soft limit. > > > > Also, I recently noticed that BSDI 2.1 supports 16 character usernames > > > > too (UT_NAMESIZE is 16). This means that BSDI 2.1 bins that access wtmp, > > > > utmp, etc will not work under FreeBSD. > > > > > > What do they do about NIS? Truncate the usernames? Bad bad bad. > > > > BSDI 2.1 doesn't have NIS. > > Ah. This is major oversight in my opinion. Tom
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.93.960620194412.2813B-100000>