Fallback to other librex instances

[davidovski]

Supersedes #260

When a google request fails for whatever reason, rather than redirecting to another instance on the user's end, librex will pick another instance at random and use it's api to fetch search results.

Todo:

• Fetch results from randomly chosen librex instance
• Create config parameter to disable fallback functionality
• Don't return results from "broken" instances
• Ensure that search settings are preserved when making api request
• Do not allow requests to be made to the current instance (recursive request)

[davidovski]

Generally this seems to work well, though occasionally some instances will return results in the wrong language (perhaps an outdated instance?) The main drawback is the response time, since not only does google need to tell librex it hasn't got anything, but it needs to check other instances which slow down the speed at which results are returned to the user.

However despite these minor issues, results are (nearly) always returned to the user when librex cannot get any results from google itself (tested by blocking via hostsfile).

Hopefuly this resolves #251 and can be merged into main soon.