Also default to HTML entity output, and replace --contains with
--starts, which does the same thing but checks only the start of
the line and removes the matched item.
Currently we rely on any N: field with a Change: subnotice being the first
N: field. Add --contains field to abcfield.py to remove this restriction.
Also use it to extract new subfield Credit: and add that if present
to the tune information in print and on web.
This is done by
a) reducing the page margins
b) compressing the heading and removing all but Notes: fields.
The latter produced vertically smaller tune PDFs that will scale wider.
In the process, modify abctitle.py to extract the first of any header field, and
move formatting for a single tune into a .fmt file.