[PATCH 0 of 4] Implement basic sparse op diag and diag op sparse support.

[Jason Riedy]

And Jaroslav Hajek writes:
> Done.

FYI, my silence means I haven't run into any problems.  ;)  There
may be a little fall-out from the sparse LU change, but hopefully
not.  And I haven't updated my external code to use the new
features.  But it appears that nothing *breaks* without changes,
and that's important.

And as I said before, my only use of other ops right now would be
scalar * perm becoming sparse.  It's a nice way of playing in the
permutation polytope.  But that lives in a nice subroutine, so
it's unimportant.

Jason