Implement percent-decoding in simple_url for all URL components

[Copilot]

Percent-encoded usernames, passwords, and paths were being decoded outside the simple_url module in parse.rs, creating duplicated logic and making the parser incomplete.

Changes

• Modified ParsedUrl structure: Changed username, password, and path fields from &str to String to store decoded values directly
• Added percent-decoding in simple_url::ParsedUrl::parse(): All URL components (username, password, and path) are now decoded during initial parsing using a new percent_decode() helper
• Removed duplicate decoding logic from parse.rs: Eliminated percent_decoded_utf8() and url_user() helper functions that were decoding fields already handled by simple_url
• Added upfront whitespace validation: The entire input URL is validated for literal whitespace at the beginning of parsing (per RFC 3986), before any percent-decoding occurs. URLs with literal whitespace characters are rejected, while percent-encoded whitespace (e.g., %20) is properly decoded

Example

// Before: username and path remained percent-encoded until parse.rs processed them
let url = ParsedUrl::parse("http://user%20name@host/path%20with%20spaces")?;
assert_eq!(url.username, "user%20name"); // Still encoded
assert_eq!(url.path, "/path%20with%20spaces"); // Still encoded

// After: all components are decoded in simple_url
let url = ParsedUrl::parse("http://user%20name@host/path%20with%20spaces")?;
assert_eq!(url.username, "user name"); // Decoded immediately
assert_eq!(url.path, "/path with spaces"); // Decoded immediately

// Literal whitespace is rejected
let result = ParsedUrl::parse("http://user name@host/path");
assert!(result.is_err()); // Rejected per RFC 3986

All URL components are now decoded consistently, simplifying the codebase and centralizing percent-decoding logic in the simple_url module while maintaining RFC 3986 compliance.

Original prompt

in gix_url::simple_url, make sure to decode the percent-encoded parts of the URL. Be sure to test this carefully, and remove manual decode operations that happen outside of the simple_url module.
Lastly, remove hand-implemented checks for whitespace in parts of the URL as this should be handled by the percent-decoding.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[Byron]

@copilot Keep going, see the comments.

[Byron]

@copilot The path should also be percent-decoded.

[Copilot]

Done in commit 7423682. The path field is now String and gets percent-decoded during parsing, just like username and password.

[Byron]

@copilot There should be no extra validation necessary, the percent decoding takes care of that.

[Copilot]

Done in commit 7423682. Removed all whitespace validation - the percent decoder handles everything now. Literal whitespace is accepted and passed through, while percent-encoded whitespace like %20 gets properly decoded.

[Byron]

@copilot It seems that percent-decoding will not error if it sees whitespace. Let's check for whitespace specifically, and do so just once at the beginning. The whole URL can't have any whitespace.

[Copilot]

@copilot It seems that percent-decoding will not error if it sees whitespace. Let's check for whitespace specifically, and do so just once at the beginning. The whole URL can't have any whitespace.

Done in commit 0bb3c7e. Added a single whitespace check at the very beginning of ParsedUrl::parse() that validates the entire input URL. URLs with literal whitespace are now rejected, while percent-encoded whitespace like %20 is properly decoded.