Continuous update of Amagama server data

[dwaynebailey]

We need to have the latest translations for our users. This is more important while Pootle can't deliver local TM. This way any translation done recently will be suggested to a user of Pootle. This helps Verbatim users. It also helps minority languages.

This should probably become a tracking bug. But issues that should be addressed in this are:

• Automation of scripts - we don't need an user intervention and problems are reported to us.
• We should be able to checkout PO files on the OSUOSL server or their related projects VMs - to prevent us having to maintain or store them on our own servers.
• Dump and load in a way that keeps the server going - we don't want the server to go away if the load breaks or where we load it too much when updating the server. Maybe by having an active vs an loading server and swapping them when the load is ready
• Cron the scripts - we want to keep the data as up-to-date as possible. We should either do a cron to update or be able to respond in some way to post-commit hooks that might be available upstream.
• Include all Mozilla related PO files that are on verbatim
• Do Firefox recovery for inclusion in the server - so that we can include Firefox strings for anyone using Verbatim
• Create generic scripts for pulling translations from standards hosts - making it easy to add a project with little effort. E.g. github or sourceforge.

I realise this is a tonne but I'd like to get some input on these and then break them out into sub-bugs that we can schedule for now or the future.