Logged In: YES
user_id=21627
These changes look good to me, please apply them.
As for the regrtest modification, please change the tests to
provide a skip_expected setting, which is computed depending
on the presence of the test data - see test_normalization.py
for an example.
It would be good if the header files containing large tables
would contain an indication on how these tables have been
created (e.g. what data source have been used, and what
modification had been applied after the tables where created
from the sources).