Create a language hierarchy system for Resource Loader to support regional differences better

SAK-40724 recently cleaned up a lot of partial language packs.

During the conversations about cleaning up those language packs, Earle R Nietzel proposed creating a language hierarchy for each language (e.g. English) with a base language (e.g. en_US) and regional differences (e.g. en_UK). The regional differences wouldn't be duplicating the base language; it would only include the few key/value pairs for differences (e.g. "favorite" being "favourite").

This feature request is to implement this change in Resource Loader to understand and look-up hierarchies for each main language that we have supported in Sakai. It would work like:

1. look for the message string for user's or site's language (e.g. en_UK)
2. if the message string doesn't exist in that regional language, then Resource Loader would identify what the base language is (e.g. en_US)
3. look for the message string for the base language (e.g. en_US)
4. if the message string doesn't exist in the base language, then Resource Loader would fallback to the server instance language like it does today

The primary change of this proposed change is to group language differences together.