Date: Tue, 18 Jul 1995 13:31:32 -0400 (EDT) From: Charles Henrich <henrich@crh.cl.msu.edu> To: terry@cs.weber.edu (Terry Lambert) Cc: freebsd-hackers@freebsd.org Subject: Re: UID > 65536 works ? Message-ID: <199507181731.NAA08760@crh.cl.msu.edu> In-Reply-To: <9507181714.AA24069@cs.weber.edu> from "Terry Lambert" at Jul 18, 95 11:14:44 am
next in thread | previous in thread | raw e-mail | index | archive | help
> The check is there for interoperability with FreeBSD as an NIS master
> or slave (not client) with a non-FreeBSD slave or client (not master).
>
> This is because in order for the entries to be useful on most older
> systems, you have to fit their idea of uid_t and gid_t.
>
> On FreeBSD, these are unsigned longs; on older systems, they are unsigned
> (or even signed) shorts.
>
> By moving this restriction, you are breaking interoperability.
>
> Potentially, the line transfer representation in NIS would cause the
> values to be truncated before they are sent out (instead of being
> truncated by the recieving system). I haven't checked the NIS code
> to be sure ...but you can while you are in there putting it back 8-).
>
> Probably, the code wants comments to this effect so it doesn't happen
> again.
It goes against every other piece of FreeBSD to keep out-dated defaults so
things work (e.g. the LINEMODE in telnetd, or the tcp options). Having > 65535
UIDS is more useful to a fair number of folk than working with outdated NIS
clients. Why not make an option so that you can enable the current broken
behavior if need be, and default to what most modern operating systems now use,
or are moving to rapidly.
-Crh
Charles Henrich Michigan State University henrich@crh.cl.msu.edu
http://rs560.msu.edu/~henrich/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199507181731.NAA08760>