#html5spec

2.5.5 URL manipulation and creation

To fragment-escape a string input, a user agent must run the following steps:

Let input be the string to be escaped.Let position point at the first character of input.

Let output be an empty string.

Loop: If position is past the end of input, then jump to the step labeled end.

If the character in input pointed to by position is in the range U+0000 to U+0020 or is one of the following characters:

""" (U+0022) character

"#" (U+0023) character

"%" (U+0025) character

U+003C LESS-THAN SIGN character (<)

">" (U+003E) character

"[" (U+005B) character

"\" (U+005C) character

"]" (U+005D) character

"^" (U+005E) character

"{" (U+007B) character

"|" (U+007C) character

"}" (U+007D) character

...then append the percent-encoded form of the character to output. [RFC3986]

Otherwise, append the character itself to output.

This escapes any ASCII characters that are not valid in the URI <fragment> production without being escaped.

Advance position to the next character in input.

Return to the step labeled loop.

End: Return output.

Implementation:

15: def escape(s) 16: s.to_s.gsub(/([^ a-zA-Z0-9_.-]+)/n) { 17: '%'+$1.unpack('H2'*bytesize($1)).join('%').upcase 18: }.tr(' ', '+') 19: end

23: def unescape(s) 24: s.tr('+', ' ').gsub(/((?:%[0-9a-fA-F]{2})+)/n){ 25: [$1.delete('%')].pack('H*') 26: } 27: end

1.7.1 How to read this specification