From owner-freebsd-questions  Fri Dec 30 13:18:56 1994
Return-Path: questions-owner
Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id NAA00952 for questions-outgoing; Fri, 30 Dec 1994 13:18:56 -0800
Received: from upcoming.dcrt.nih.gov (upcoming.dcrt.nih.gov [156.40.112.44]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id NAA00946; Fri, 30 Dec 1994 13:18:54 -0800
Received: (from crtb@localhost) by upcoming.dcrt.nih.gov (8.6.9/8.6.9) id QAA02761; Fri, 30 Dec 1994 16:18:52 -0500
Date: Fri, 30 Dec 1994 16:18:52 -0500
From: Chuck Bacon <crtb@upcoming.dcrt.nih.gov>
Message-Id: <199412302118.QAA02761@upcoming.dcrt.nih.gov>
To: freebsd-bugs@freebsd.org, freebsd-questions@freebsd.org
Subject: Why does ls report wrong creation date on symlinks?
Sender: questions-owner@freebsd.org
Precedence: bulk

I just discovered that "ls -l" reports the creation date incorrectly
on symlinks.  It reports as the creation date of each symlink, the
modification time of its directory.  Thus, if I "touch foo" in some
directory, a subsequent "ls -l" will report the identical creation
time for both foo and all the symlinks in the directory.  Therefore,
"ls -lt" will position all the symlinks at the top.

I created a perl script, using the "stat" function to report all 3
dates, atime, mtime and ctime; it correctly reports those times, in
contradiction to the mtime reported by "ls -l".

I had previously noted that all directories in a MSDOSFS tree carry
as their creation date, the current moment.  Somebody suggested that
this was a capricious choice, in view of the impossibility of
duplicating the creation date which DOS would report.  I wonder
if there is a link between these two bugs.

	Chuck Bacon -- crtb@helix.nih.gov