fix: fallback to old version of login

[Lash-L]

So I believe what is happening is the following:

Accounts that have not logged in with the newer methods have not been fully migrated. So when we get the information, we only get a country code or a country + a base url.

If we get either one, we can trust that it is a good url. If we get both - we should use code_login_v4, if we just get one, we should use v1.

[allenporter]

Possible to reproduce the problem with a test?

[tomtomtomtom1994]

@Lash-L The fall back needs to be moved out of the scope of the login limiter. Other wise it throws rate limit exceptions in the case of a fall back.

request_code_v4() acquires item "login",
then we call request_code() and immediately try and acquire item "login" again which fails.

I've tested the following change and it works but I am unable to commit to this PR.