Lesson 16 · Case normalization - Build a URL Parser

Case normalization

Two URIs that differ only in letter case for the scheme or host are equivalent, so normalization lowercases them. Today you start the Normalize function with case folding, the first of three syntax-based normalization rules.

The goal

Normalize a URI by lowercasing its scheme and host.

Start here - the target

TO DO

Scenario: Lowercasing scheme and host

Giventhe input "HTTP://EXAMPLE.COM/Path"

Whenit is parsed and then normalized

Thenthe normalized URI has Scheme "http" and Host "example.com"

AndPath keeps its original case as "/Path", since only the scheme and host are case-insensitive

Background

Two URIs can be spelled differently yet identify the same resource, and syntax-based normalization is the process of rewriting a URI into a canonical form so such equivalences become plain string equality. RFC 3986 Section 6.2.2 lists the safe transformations that never change meaning, and the simplest is case normalization. The scheme and the host are defined to be case-insensitive, so HTTP and http, EXAMPLE.COM and example.com, are the same - normalization lowercases both.

The discipline is knowing what not to touch. The path, query, and fragment are case-sensitive: /Path and /path may be different resources, so Normalize must leave their case exactly as parsed. Only the scheme and host get folded. Normalize returns a new URI so the original stays intact, and today it changes just those two fields - the next two lessons add the other two normalization rules on top of this same function before the resolution chapter recomposes a URI back into a string.

Make it work

// scheme and host are case-insensitive; lowercase them.
// path, query, fragment are case-SENSITIVE - leave them alone.
func Normalize(u *URI) *URI {
  out := *u
  out.Scheme = strings.ToLower(out.Scheme)
  out.Host = strings.ToLower(out.Host)
  return &out
}

CheckpointDONE

Normalize now case-folds the scheme and host. Commit and stop here.