Great Circle Associates Majordomo-Workers
(April 1996)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: Config file commentary.
From: Gene Rackow <rackow @ mcs . anl . gov>
Date: Thu, 25 Apr 1996 06:42:09 -0500
To: Chan Wilson <cwilson @ slurp . neu . sgi . com>
Cc: pdc @ lunch . engr . sgi . com (Paul Close), david_wolfe @ risc . sps . mot . com, majordomo-workers @ GreatCircle . COM, Brent @ GreatCircle . COM (Brent Chapman)
In-reply-to: Your message of "Wed, 24 Apr 1996 23:57:13 MDT." <v02130505ada421fe7a8a@[198.102.244.42]>

I agree with Brent, there needs to be a simple/fast way to correct problems.
I have no problem with the system using a dbm for speed, but the process
should be similar to sendmail's alias file.  You have the real files sitting
there and create the dbm from those.  I don't know how many times there have
been problems because of a broken dbm.  Not having the capability to "use
the source" and "rebuild" would not be good.  I have a feeling few of
you have had the fun of working on NeXT systems where most of the admin
stuff is buried in "netinfo databases" where early releases had a habit
of getting messed up with no way to recover.

I would not be adverse to having the text config file, the dbm, and a 
small "working" file for the digest number and any other things that
need to be updated regularly and rapidly.  Having the digest number similar
to the :include: of the "aliases" file is a possibility.

I don't like the updating of the config file as often as it currently
happens, but I distrust dbm stability even more.

--Gene


References:
Indexed By Date Previous: Re: Config file commentary.
From: "Mr. Erich L. Markert" <1e6244@sunams.usma.edu>
Next: no-reason BOUNCE's?
From: Dave Barr <barr@math.psu.edu>
Indexed By Thread Previous: Re: Config file commentary.
From: Brent@GreatCircle.COM (Brent Chapman)
Next: Re: Config file commentary.
From: bill@celestial.com (Bill Campbell)

Google
 
Search Internet Search www.greatcircle.com