dtype size error for large h5 data

[dschick]

when reading data from h5 sources, e.g. a NeXuS file, the following error happens for rather large datasets:
ValueError: invalid shape in fixed-type tuple: dtype size in bytes must fit into a C int.
A possible solution is proposed here

This would mean to change the reading of raw data from fromarrays to append_fields in

pyEvalData/pyEvalData/io/sardana_nexus.py

Line 172 in fa72dad

scan.data = fromarrays(data_list, dtype=dtype_list)

thanks to @amolodts for finding this bug