[LORENE] d'Alembert operator

[Jerome Novak]

Dear LOREN-ers,

The wave equation solver (avance_dalembert) now works only with Scalar's
(and no more with Cmp's as before). if this is a prolem to anybody,
please tell me and I will put back all the files linked with the Cmp
version of this solver. I have also modified this solver so that it now
accepts a grid (and a a mapping) that have a compactified external
domain (CED). As you may know, the change of variable of the type u=1/r
(which is used in these domains) is not compatible with hyperbolic PDEs.
So, the solver DOES NOT solve the wave equation in the CED, but simply
makes a copy of the field values in the CED at the last time-step to the
next.

All this is not fully tested yet but should at least compile... :-)

Bon code!

--
Jerome Novak
Laboratoire de l'Univers et de ses THeories (LUTH)
Observatoire de Meudon       92195 Meudon FRANCE
tel: (33) 01 45 07 75 82     fax: (33) 01 45 07 79 71