normalized_path full-width chars japanese

[ignacio-chiazzo]

Using the normalized_path method turns the Japanese １ into 1. This also happens for ２ -> 2 etc. (More about half-width and half-width characters and numbers here)

 url = "http://a.com/%EF%BC%91" # http://a.com/１ 
Addressable::URI.parse(url).normalized_path # "/1"

[152] pry(main)> url = "http://a.com/%EF%BC%91" # http://a.com/１ 
=> "http://a.com/%EF%BC%91"
[153] pry(main)> Addressable::URI.parse(url).normalized_path # 
=> "/1"
[154] pry(main)> 

Looks like it was done on purpose on the lines

addressable/lib/addressable/uri.rb

Lines 1537 to 1541 in 7311daa

	result = path.strip.split(SLASH, -1).map do |segment|
	Addressable::URI.normalize_component(
	segment,
	Addressable::URI::CharacterClasses::PCHAR
	)

Am I missing something from the RFC?

[dentarg]

Addressable is using NFKC normalization, so this is expected.

See #8 (comment) and other issues like #217 and #325.

https://towardsdatascience.com/difference-between-nfd-nfc-nfkd-and-nfkc-explained-with-python-code-e2631f96ae6c has some examples for different forms of normalization.