Validation and error handling for data mismatches

I tested this module on a site which is using a dataset of countries and zones that are not Ubercart stock, but I'm sure that some of the same problems would carry over even with a stock set of data. I found a total of 652 provinces which could not be matched throughout the entire dataset (of about 4000 provinces in the cart set) and 3 countries which didn't even have Location datafiles (and would therefore throw PHP errors by trying to include nonexistant files).

Among the areas that failed were all British addresses (Location expects the reserved UK code, while my dataset has the official GB code), and armed forces addresses (such as US zone AP - Armed Forces Pacific). This makes sense actually, because it's perfectly reasonable to want to allow military personnel to purchase from your online store, but not reasonable to expect to geocode and show Google maps and driving directions for their locations. That is, the modules serve different purposes, and we can't expect the data to always match 100%. Of course, there would also be the issue of adding revised or new .inc's to location or .cif's to Ubercart - we can't expect these datasets to always be in synch, even when using stock data.

I would suggest some additional checking in, for instance _uc_location_order_to_location() to validate that the country and province being pulled from the Ubercart data actually match something in the Location module's dataset (and vice-versa in _uc_location_to_uc_address()) before saving or presenting the data. If data from one dataset is invalid in the other, the best course is probably just to log the mismatch and abort the translation. As long as this module is just being used as a convenience for users, having to enter another address is better than having errors or bad data.

In case it's being used for something that requires that the translation happen successfully, it may be useful to have some sort of hook that is called if a mismatch is detected, so that people can create an add-on translation table module (especially useful if you have custom datasets) or some other handler to deal with special cases (ex: if nothing else works, use code 'xx') depending upon their own datasets, use cases, and business needs.