Better URL handling

[N-Coder]

Using urllib's ParseResult as sole data type for representing URL attributes is probably not the best and most flexible way, so we should evaluate ways of letting the user chose how to handle URLs as there is a plethora of different ways and libraries for representing URLs. See e.g. here:

The "mailto" being converted to lowercase probably comes from using python's built-in urlparse to represent the URLs. One could probably use one of many other libraries (rfc3986, rfc3987, hyperlink, furl, ...) as a replacement that might allow customizing this, but I haven't looked into that. The biggest issue there is probably deciding which one to use / how to support multiple of these libraries.

[make-github-pseudonymous-again]

Why not just a string? Can keep parsing for validation.

[make-github-pseudonymous-again]

I mean the part that makes sure the URL has valid format can stay hidden inside the implementation but the return value should be the original slice of input without modifications? What about URL encoded values? Mmmh...

[make-github-pseudonymous-again]

If the URL format is such that the raw string can be pasted into a browser URL bar then I do not see why we would not return the original string slice.