[Changeset] Add the hypot function

[John W. Eaton]

On 23-Mar-2008, David Bateman wrote:

| Matlab maps the libm (or in their case the fdlib implementation of libm)
| hypot function into matlab, that calculates sqrt(x.^2+y.^2) but in a way
| that avoids overflows for large numbers. This patch adds the hypot
| function to matlab and fixes a small issue with the atan2 function when
| it is called with sparse matrices.

I applied this changeset.

Thanks,

jwe