'normalize' needs to be applied to all input text fields
The new datapackage wizard in Morpho 1.5 adds a number of text input fields.
Previous experience has indicated that we need to scan these inputs for illegal
characters (using a method like 'normalize') to create character entities in the
XML for special characters sometime pasted from Word or PDF files.
A quick check indicates that we have not been doing this. We need to check all
the input fields to make sure that illegal characters are handled properly