The current pseudopotential schema, v.0.99, is implemented in the upflib/ library. It requires some testing of course, but it should more or less work both in writing (write_upf_new.f90) and in reading (read_upf_new.f90) codes. The upfconv.f90 utility should also work. The schema for the GIPAW section should however be slightly modified (see also the FIXME's in read_upf_new.f90):
pp_gipaw_core_orbital
pp_gipaw_orbital
should be preceded by, or included into, a tag containing the information on the number or respective orbitals. This would make reading the xml file in a SAX style, as done by the xmltools.f90 utilities, much easier.
the two sections were already in the schema, just I omitted to write them, probably planning to remove them from the schema.
to spare two tags we could turn these two elements as attributes. What do you think ?
The current pseudopotential schema, v.0.99, is implemented in the upflib/ library. It requires some testing of course, but it should more or less work both in writing (write_upf_new.f90) and in reading (read_upf_new.f90) codes. The upfconv.f90 utility should also work. The schema for the GIPAW section should however be slightly modified (see also the FIXME's in read_upf_new.f90):
should be preceded by, or included into, a tag containing the information on the number or respective orbitals. This would make reading the xml file in a SAX style, as done by the xmltools.f90 utilities, much easier.