URLs without // don't parse correctly

[sethmlarson]

URLs like: mailto:johndoe@example.com don't parse correctly due to our logic on when to add a scheme delimeter to a URL.

[shazow]

Hm, what should that parse to?

[sethmlarson]

Per RFC3986 should be

Scheme: mailto
Userinfo: johndoe
Hostname: example.com

We have the same problem with javascript:

[sigmavirus24]

@sethmlarson I think you may be confused.

: marks the end of the scheme, so you're correct that mailto is the scheme. // delinates the beginning of the authority, however, and the authority includes the userinfo and hostname and port. Effectively, without // the text following the : is actually part of the path if I remember correctly.

[sigmavirus24]

rfc3986 has specific tests covering this, by the way, https://github.com/python-hyper/rfc3986/blob/0d82fee4b21482ae6cf4337dc8216e4d1ce507ea/tests/test_uri.py#L13

You can check everywhere that that particular fixture is used.

[sigmavirus24]

https://github.com/python-hyper/rfc3986/blob/0d82fee4b21482ae6cf4337dc8216e4d1ce507ea/tests/test_uri.py#L47

[sethmlarson]

You're right @sigmavirus24, our scheme detecting regex is wrong but the above example shouldn't parse that way. Thanks for correcting me before I go too off the rails. :)

[sigmavirus24]

It's unfortunate that I know the RFC as well as I do, but I guess also inevitable.