Subject: Re: Before I offically call this a time bug can someone else try this? (fwd)
To: None <R.S.Brooks@liverpool.ac.uk>
From: John Kohl <jtk@kolvir.arlington.ma.us>
List: tech-kern
Date: 08/05/1998 13:24:13
>>>>> "Roger" == Roger Brooks <R.S.Brooks@liverpool.ac.uk> writes:

Roger> When I looked at the source for (ufs)dump, sure
Roger> enough a level 0 dump is implemented as a dump of all files for which
Roger> mtime >= 0 || ctime >= 0.  NOT mtime >= LONG_MIN || ctime >= LONG_MIN
Roger> (which is what you'd expect if time_t is signed).  So if you ever set
Roger> the clock before 1 Jan 1970 and create or modify any files, they won't
Roger> EVER get dumped unless you touch them to a non-negative time.  I've
Roger> just had a look at the source for NetBSD dump, and it appears to have
Roger> the same problem (as presumably do all versions of dump derived from
Roger> BSD unix).

I'm not sure I would necessarily call it a "problem."  It's sometimes
useful to keep certain files from ever being backed up, if you don't
trust the backup tapes to be handled with sufficient care and
protection.  e.g. a Kerberos database master key file :)

-- 
==John Kohl <jtk@kolvir.arlington.ma.us>, <john_kohl@alum.mit.edu>
Write a poem, share your heart!
Home page: <URL:http://people.ne.mediaone.net/jtk/>
Note new home zip code as of July 1, 1998: 02476