[290] in Coldmud discussion meeting

Re: time in the log

daemon@ATHENA.MIT.EDU (Fri May 20 19:37:45 1994 )

From: deforest@netcom.com
To: rayn@q.crossaccess.com (Ray Nawara jr.)
Date: Fri, 20 May 1994 16:30:22 -0700 (PDT)
Cc: coldstuff@MIT.EDU
In-Reply-To: <9405202256.AA23959@q.crossaccess.com> from "Ray Nawara jr." at May 20, 94 03:56:26 pm

I feel compelled to respond to this, for some reason...
> 
>    > I definitely dont think time should be made in-db. That makes it
>    > modifiable from the db, and therefore prettymuch invalid as a logging
>    > tool security-wise.
> 
>    It is a $sys-only builtin at the moment, so security is not an issue.
> 
> That is exactly WHY it's an issue. I'm talking about when an admin
> account is hacked. It may be that your only way of detecting this is
> via the logs, and if time isnt in the logs, then thats one less clue.

If someone hacks an admin account you've got worse problems than what
 time the accuont was hacked...Dare I discuss cluefulness of admins? Would
 that be a 'people in glass houses' thing? Hmmm...

> anything that can be changed in-db is not strictly secure
> serverwise. (ignoring the in db-security, which may be perfectly
> adequate). I dont consider the DB's time function as a trusted thing,
> only the server's call to unix, and if time were moved to the db, i'd
> be miffed ;) on the other hand, I'd still live, so this is just the
> opinon of a currently extremely paranoid individual :)
> 
> 	Ray / Ash
> 
> 
The only part to be moved in-db would be the formatting. More exactly, all
 log output. This was Lynx's idea, but I think I started this instance of the
 thread, and in doing so I mentioned the possibility of logging sender().

The only reason I mention that is that it sounds like you're suggesting that
 time() being moved in-db was an idea. I don't think anyone could argue for
 this and mean it and know what a computer is.

In conclusion, we are in agreement if it's an option in config.h or Makefile.

Crag, probably beating a dead chiwawa(sp?).