From mboxrd@z Thu Jan  1 00:00:00 1970
From: Milan Zamazal <pdm@zamazal.org>
To: Yngve Svendsen <yngve.svendsen@clustra.com>
Cc: gnats-devel@sources.redhat.com
Subject: Re: GNATS password management
Date: Wed, 27 Jun 2001 14:45:00 -0000
Message-id: <87wv5x7ghk.fsf@blackbird.zamazal.org>
References: <5.1.0.14.2.20010626194604.04dcd278@mail.trd.clustra.com>
X-SW-Source: 2001-q2/msg00209.html

>>>>> "YS" == Yngve Svendsen <yngve.svendsen@clustra.com> writes:

    YS> Below is an outline of the two password-management tools I'd
    YS> like to see implemented for GNATS 4.

Thanks.

    YS> gnats-pwconv
    YS> gnats-pwconv [-mp] infile outfile
    YS> gnats-pwconv -n[mp] infile

[...]

    YS> -n, --nocreate	Don't create file; display results on stdout.

This option looks useless to me here.

    YS> gnats-pw
    YS> gnats-pw [-cmdps] passwordfile username accesslevel
    YS> gnats-pw -n[mdps] username accesslevel

I'd call it `gnats-passwd'.

I'm not sure I like specifying `passwordfile' this way (I know htpasswd
does it so).  I'd prefer having an `-f / --file' option instead, thus
making the argument optional, possibly combined with `-d / --database'
for specifying a particular database.  I think it's easier to handle for
GNATS admins -- they don't have to think about the exact names and
locations of the access files.

Using `-d' for database would cause a collision with:

    YS> -d, --crypt	Force CRYPT encryption of the password (default).

but maybe someone can find a better single letter for the --crypt
option.

Regards,

Milan Zamazal

-- 
I think any law that restricts independent use of brainpower is suspect.
                                               -- Kent Pitman in comp.lang.lisp