[CHANGESET] Add the amd function

[John W. Eaton]

On 21-Mar-2008, David Bateman wrote:

| As the amd code is already used with umfpack, which Octave uses for the
| sparse LU function, and Tim Davis supplies a mex file for the equivalent
| of the amd function in matlab under the GPL, it is trivial to write an
| Octave version of the same code.. Attached is a matlab compatible
| version of the amd function for Octave.

I applied it.

Thanks,

jwe