URL-interop and path parameters

[joakime]

Here's a bit of trivial for you.

From https://tools.ietf.org/html/rfc3986#section-3.3

A path like /path;a=b/with;c=foo/path;d=parameters

Path parameters support has existed in the IETF specs like this ...

• RFC 7230 - HTTP/1.1 spec (June 2014) - https://datatracker.ietf.org/doc/html/rfc7230#section-2.7

• Delegates to RFC 3986 URI Spec - https://datatracker.ietf.org/doc/html/rfc3986#section-3.3 - which explains path parameters

• RFC 2616 - HTTP/1.1 spec (June 1999) - https://datatracker.ietf.org/doc/html/rfc2616#section-3.2

• Delegates to RFC 2396 (Aug 1998) - https://datatracker.ietf.org/doc/html/rfc2396#section-3.3 - which also explains path parameters

• RFC 2068 - HTTP/1.1 spec (January 1997) - https://datatracker.ietf.org/doc/html/rfc2068#section-3.2 - has details about path parameters

The most recent RFC3986 is at odds with RFC7230 when it comes to path normalization when path parameters are in the mix.
The two sides don't agree, and I bet the w3c whatwg specs are also either at odds with, or have punted on, this feature of URI paths.

See https://datatracker.ietf.org/doc/html/rfc3986#section-6

Since path parameters start with a ; and don't have to contain a = or a key or a value.
The following are all allowed.

/a/;/
/a/.;/
/a/..;/
/a/..;b/
/a/..;b=/
/a/..;b=c/

What are you supposed to do when you have to normalize these?

Some examples:

http://machine.com/path;a=foo/..;z=y/what;c=bar
http://machine.com/path/..;?a=b
http://machine.com/path;a=b/../path;a=z
http://machine.com/;?a=b
http://machine.com/a/b/..;/c;/d

Some servers will treat ; as a path parameter and strip it out (to hold separately) and then normalize the non-path parameter version.
Some servers will treat ; as just a character in a path, so that /a/..;b/c is actually a dir called a with a directory ..;b inside of it.