LP#1208875: follow-up to standardize extract fields

LP#1208875: follow-up to standardize extract fields

This patch removes the proposed custom methods for extracting
title, author, and record format in favor of tweaking
->fetch_user_circ_history to invoke unapi.bre and adjusting
the template to use get_marc_attrs. Also, nowadays
->fetch_user_circ_history can flesh what we need it to without
having to rely on the existance of an action.circulation row,
which won't be present if the circ was aged but was otherwise
retained in the user circ history.

The result is slower than the previous approach, but still
retains the core idea of getting A/T out of the equation, and remains
much faster than the A/T approach.

Dropping use of unapi.bre would speed things up a bit more, as it
was added only to match the addition of the record format column
in the CSV output. Drop the column, and we no longer need to worry
about MVFs.

There would also be opportunities to improve caching further.  Bib
display fields, when it comes, will likely help even more, as it
would mean being able to drop a lot of the XML parsing currently used.

This patch also adjusts misc_util.tt2 so that including it doesn't
result in an unwanted blank line.

Signed-off-by: Galen Charlton <gmc@equinoxinitiative.org>