URL userinfo parser does not follow RFC3986

Hi, I appreciate that 99.99% of use cases are "user:pass" for the 
userinfo portion of the URL, but RFC3986 allows for a much broader list 
of characters, and colons are not necessarily delimiters.

https://www.rfc-editor.org/rfc/rfc3986#section-3.2.1 
<https://www.rfc-editor.org/rfc/rfc3986#section-3.2.1>

For example: myscheme://a:1=foo&a:[email protected]/

The CURL URL parser is useful on its own for parsing arbitrary URI. 
Especially when curl is imported for its more general use case in other 
aspects of the program.

I propose that a CURLUPART_USERINFO CURLUPart be added such that 
curl_url_get/set returns the complete string rather than trying to parse 
the ':' assuming a specific scheme.

https://github.com/curl/curl/blob/dbff3eec456d86386be4cebb89cbe4dfc21a06b6/include/curl/urlapi.h#L73

Thoughts?

-- 
Unsubscribe: https://lists.haxx.se/mailman/listinfo/curl-library
Etiquette:   https://curl.se/mail/etiquette.html