Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!elroy.jpl.nasa.gov!ames!amdahl!pacbell!att!chinet!les
From: les@chinet.chi.il.us (Leslie Mikesell)
Newsgroups: comp.unix.questions
Subject: Re: Why does "file" change the creation time on some Unix systems?
Summary: a more serious problem
Message-ID: <8877@chinet.chi.il.us>
Date: 6 Jul 89 13:56:15 GMT
References: <95@anasaz.UUCP> <2268@faline.bellcore.com> <1850@auspex.auspex.com> <4217@tekcrl.LABS.TEK.COM> <1878@auspex.auspex.com>
Reply-To: les@chinet.chi.il.us (Leslie Mikesell)
Organization: Chinet - Public Access Unix
Lines: 15


> >>It doesn't refer to *ANY* "creation" time, it refers to the inode change
> >>time.

If anyone has mentioned this I must have missed it.
This discussion started about find(1) changing the atime of files
to pretend it didn't read them, thus unavoidably changing the ctime.
The real problem with this is that if you do incremental backups based
on ctime (which you should if you do incremental backups), anything
that file(1) looked at is going to be copied unnecessarily.  The
same goes for the -a flag of cpio.  Perhaps we need a real "archived"
flag in the inodes since ctime is hopelessly overloaded and mtime (rightly)
doesn't change when a file is renamed.

Les Mikesell