Diagnostics about compatibility of elevation data at boundary

[ole]

It'd be handy to provide diagnostics and perhaps a warning in case the elevation data in files used by File_boundary (and Field_boundary) aren't matching the elevation data used by the domain being evolved. In some cases small discrepancies can be safely ignored e.g. if they are small relative to the water depth, but in other cases, a small discrepancy can cause spurious flows. This may be the problem highlighted by Joaquim Luis in postings January 2008.

Currently, File_boundary works only with conserved quantities, so elevation has to be obtained in order to do this.

See changeset:4925 for a possible example of this