force to only accept hardware address as dhcp-client-identifier

[John Hascall]

> On Tue, Jan 29, 2008 at 09:50:03AM -0600, John Hascall wrote:
> > Unless they have recently changed their minds,
> > despite numerous requests, ISC refuses to add such
> > an option because it is contrary to the RFC standards
> > documents.

> it is true that i refuse to apply the patch which has been circulating
> that totally removes dhcp client identifier support from the server
> with no option or intervention.  i think it would be irresponsible to
> distribute this as default behaviour.

   Agreed.

> it is not true that i am opposed to developing tools to assist
> operators of dhcp networks in challenging client-identity environs,
> even to the extent of following that up with standards action and
> thus changing the server's default behaviour towards certain things.

   I don't even see any reason to change the default behaviour,
   and I really only care that the standard says "the server MUST
   use the client-identifier if supplied" in so far as that might
   cause ISC to not consider a config option which, if activated,
   caused said non-standard behaviour.  And speaking just for myself,
   even that level of caring is quite small as I can easily make the
   change locally myself.  Others, obviously, have different abilities
   and management environments.

> so, although what is true about me is not necessarily true about ISC,
> in that i just work here, i do not think it can reasonably be said
> that ISC is whitholding or refusing anything.
> 
> for my part, i do what i can when i can.

   Understood and appreciated.  Truly.

> if you feel our priorities are not set straight, there are easy ways
> to do what you can;
> 
> 	mailto:dhcp-suggest at isc.org?subject=my-client-id-patch
> 	mailto:sales at isc.org?subject=custom-development
> 	https://secure.isc.org/index.pl?/store/donation/
> 	https://secure.isc.org/index.pl?/store/t-shirt/

John