Mime-Version: 1.0
Message-Id: <p0511172ebafd4eefe4e2@[137.78.212.225]>
In-Reply-To: <20030530072339.GA27594@odorn.ics.muni.cz>
References: <200305281343.XAA96732@au.padl.com>
	<20030529130218.GA4296@dsl092-163-146.wdc1.dsl.speakeasy.net>
	<200305291310.XAA14959@au.padl.com>
	<p05111720bafc38450d28@[137.78.212.225]>
	<200305292252.IAA21419@au.padl.com>
	<p05111722bafc494a0a6e@[137.78.212.225]>
	<200305292345.JAA22308@au.padl.com>
	<p05111729bafc56b73023@[137.78.212.225]>
	<20030530072339.GA27594@odorn.ics.muni.cz>
Date: Fri, 30 May 2003 11:42:05 -0700
To: Daniel Kouril <kouril@ics.muni.cz>
From: "Henry B. Hotz" <hotz@jpl.nasa.gov>
Subject: Re: couple socket-connection questions
Cc: lukeh@PADL.COM, buckh@pobox.com, heimdal-discuss@sics.se,
	Dj.Byrne@jpl.nasa.gov, pgsql-docs@postgresql.org
Content-Type: text/plain; charset="us-ascii" ; format="flowed"

At 9:23 AM +0200 5/30/03, Daniel Kouril wrote:
>On Thu, May 29, 2003 at 05:48:44PM -0700, Henry B. Hotz wrote:
>>  Actually that's mod_auth_krb, not kerb.  They are different apache
>>  modules and the one without the 'e' is newer.  I found out about it
>>  from the PostgreSQL manual.  It claims to implement some Microsoft
>>  IETF-draft which is also supported by Mozilla for getting Kerberos
>>  credentials forwarded.  If the connection doesn't support the
>>  IETF-draft then it falls back to getting a name/password via basic
>>  auth.  With the right options apache will retain the (forwarded or
>>  created) tgt which can be used by mod_perl (or presumably PHP) to
>>  authenticate a connection with PostgreSQL.
>
>I don't know which module is meant by the PostgreSQL docs, but you can have
>a look at http://meta.cesnet.cz/software/heimdal/negotiate.en.html
>The page contains implementations of exact the functionality refered above.

Interesting.

I just repeated my Google search for mod-without-e and almost the 
only hits are the PostgreSQL admin manual.  I did find a 
cross-reference to a sourceforge page 
<http://modauthkerb.sourceforge.net/> which looks an awful lot like 
the page I remember for the no-e module --- but it has an e.

My guess is that the module web page had a spelling error once which 
was picked up by the PostgreSQL admin manual.  The spelling error has 
been fixed everywhere now, except at PostgreSQL.

I'm cc'ing the PG doc's list on this reply.  Hopefully someone there 
can tell us what the real situation is.
-- 
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu