Project review: Interposer mechanisms

[Henry B. Hotz]

On Oct 4, 2012, at 9:32 AM, <krbdev-request at mit.edu> <krbdev-request at mit.edu> wrote:

> Date: Thu, 04 Oct 2012 09:45:58 -0400
> From: Simo Sorce <simo at redhat.com>
> Subject: Re: Project review: Interposer mechanisms
> To: Luke Howard <lukeh at padl.com>
> Cc: krbdev at mit.edu
> Message-ID: <1349358358.22373.143.camel at willson.li.ssimo.org>
> Content-Type: text/plain; charset="UTF-8"
> 
> On Thu, 2012-10-04 at 08:42 +1000, Luke Howard wrote:
>> Is there a reason why the interposer has to explicitly enumerate the
>> mechanisms it wishes to interpose? What if it wishes to interpose all
>> mechanisms?
> 
> Good question.
> But in what case would you want to interpose a mechanism you know
> nothing about ?

In the case where you want to replace a standard deployment with a development version or an upgraded version of *everything*.

> Not saying you shouldn't just trying to understand your PoV.
> We could allow the interposer to return a wildcard OID I guess and then
> special case it in the mechglue initialization code, it wouldn't be
> hard.
> 
> Simo.
> 
> -- 
> Simo Sorce * Red Hat, Inc * New York

------------------------------------------------------
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz at jpl.nasa.gov, or hbhotz at oxy.edu