Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!husc6!mit-eddie!ll-xn!cit-vax!nike!sri-spam!sri-unix!hplabs!sdcrdcf!ucla-cs!rutgers!seismo!vrdxhq!BASKIN!peter
From: peter@BASKIN.UUCP (Peter Klosky)
Newsgroups: net.unix
Subject: Re: getpwent for indoes?
Message-ID: <2367@curly.ucla-cs.ARPA>
Date: Thu, 23-Oct-86 14:49:52 EST
Article-I.D.: curly.2367
Posted: Thu Oct 23 14:49:52 1986
Date-Received: Sun, 26-Oct-86 04:00:20 EST
Organization: Digital Analysis Corporation
Lines: 12

A little more investigation has shown that du() and ls() may in fact
tell the truth about file size under 4.2 file systems in terms
of the number of blocks used; looks like there are more
fields available from stat() than in historic systems such
as System V.  I wondered why our 4.2 manual did not list the du()
bug present in System V as a "limitation."  Now I know.
So stat() may tell the truth about some things, though perhaps
not the whole truth.

Of course, I still don't know how to cat a file by inode number
under 4.2; in specific, the concept of cylinder groups has me
confused.  Is there just one i-list or many on a given file system?