Subject: Re: lib/254: uname(1) output breaks scripts, is unreadable
To: None <Mark_Weaver@brown.edu>
From: Chris G. Demetriou <cgd@postgres.Berkeley.EDU>
List: netbsd-bugs
Date: 05/22/1994 21:19:38
> Sure, that's fine if all you want to do is compare with a given
> revision.  But if you want to report the information to the user or
> write it to a file in a reasonably consistent manner, gross hacks will
> be necessary.

In other words, the problem is that the script is trying to parse
the information, and output it in a different format, which is
something that it should _NOT_ do.  In other words, the scripts
which try to do that are inherently broken, if they try to mangle
the 'uname' at all.

> You haven't given any _benefits_ of the new method.  Are there any?
> If there are disadvantages and no advantages, why do it?

Actually, looking at the newvers.sh 'code' again, the duplication of
the "NetBSD" in the revision seems to have been an error that i made
when updating newvers.sh to match the 4.4-Lite "interface."  I've
fixed it, so that the revision is "0.9B" for instance.

Note, however, that uname -v is still multi-word-and-very-ugly, and
is going to stay that way.  uname -v now outputs all relevant kernel
version information, as compared to the previous method, which didn't
print nearly enough.


chris

------------------------------------------------------------------------------