Currently contrib_archive.py retrieves some pages that are 404 responses for missing library properties files.

These may then be automatically committed by update.sh.

It would be preferable if they were not checked in, as they obscure the last known working copy -- and some 404s contain elements like dates and thus change every single time they are pinged.

update.sh contain read lines to pause the script, but these appear to not work correctly when run from a virtualenv -- this could be fixed. Another solution would be to verify files (both .txt and .zip) as valid.