There is a bug in urlparse.urljoin which means that a relative
URL which is just a query string is not correctly added to the
base. This is a workaround for that.

Once we have Python 2.6 it can all go back to how it was before.

The database table which says which scrapers are where will now be filled in automatically, 
which should result in rather less in the way of manual editing errors.

I've also redone the python PublicAccess scraper and set all the PublicAccess sites to use 
it (removing the PHP PublicAccess scrapers).

Add all publicaccess sites to the python scraper.

Make display method on a planningapplication work out the postcode if it isn't set.

Factor out CookieAddingHTTPRedirectHandler.

before).