1
0
mirror of https://github.com/mastodon/mastodon synced 2024-11-24 23:26:22 +09:00
mastodon/config/initializers/twitter_regex.rb
Eugen Rochko cefa526c6d
Refactor formatter (#17828)
* Refactor formatter

* Move custom emoji pre-rendering logic to view helpers

* Move more methods out of Formatter

* Fix code style issues

* Remove Formatter

* Add inline poll options to RSS feeds

* Remove unused helper method

* Fix code style issues

* Various fixes and improvements

* Fix test
2022-03-26 02:53:34 +01:00

79 lines
4.2 KiB
Ruby

module Twitter::TwitterText
class Configuration
def emoji_parsing_enabled
false
end
end
class Regex
REGEXEN[:valid_general_url_path_chars] = /[^\p{White_Space}<>\(\)\?]/iou
REGEXEN[:valid_url_path_ending_chars] = /[^\p{White_Space}\(\)\?!\*"'「」<>;:=\,\.\$%\[\]~&\|@]|(?:#{REGEXEN[:valid_url_balanced_parens]})/iou
REGEXEN[:valid_url_balanced_parens] = /
\(
(?:
#{REGEXEN[:valid_general_url_path_chars]}+
|
# allow one nested level of balanced parentheses
(?:
#{REGEXEN[:valid_general_url_path_chars]}*
\(
#{REGEXEN[:valid_general_url_path_chars]}+
\)
#{REGEXEN[:valid_general_url_path_chars]}*
)
)
\)
/iox
UCHARS = '\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}\u{E000}-\u{F8FF}\u{F0000}-\u{FFFFD}\u{100000}-\u{10FFFD}'
REGEXEN[:valid_url_query_chars] = /[a-z0-9!?\*'\(\);:&=\+\$\/%#\[\]\-_\.,~|@#{UCHARS}]/iou
REGEXEN[:valid_url_query_ending_chars] = /[a-z0-9_&=#\/\-#{UCHARS}]/iou
REGEXEN[:valid_url_path] = /(?:
(?:
#{REGEXEN[:valid_general_url_path_chars]}*
(?:#{REGEXEN[:valid_url_balanced_parens]} #{REGEXEN[:valid_general_url_path_chars]}*)*
#{REGEXEN[:valid_url_path_ending_chars]}
)|(?:#{REGEXEN[:valid_general_url_path_chars]}+\/)
)/iox
REGEXEN[:valid_url] = %r{
( # $1 total match
(#{REGEXEN[:valid_url_preceding_chars]}) # $2 Preceding character
( # $3 URL
((?:https?|dat|dweb|ipfs|ipns|ssb|gopher|gemini):\/\/)? # $4 Protocol (optional)
(#{REGEXEN[:valid_domain]}) # $5 Domain(s)
(?::(#{REGEXEN[:valid_port_number]}))? # $6 Port number (optional)
(/#{REGEXEN[:valid_url_path]}*)? # $7 URL Path and anchor
(\?#{REGEXEN[:valid_url_query_chars]}*#{REGEXEN[:valid_url_query_ending_chars]})? # $8 Query String
)
)
}iox
REGEXEN[:validate_nodeid] = /(?:
#{REGEXEN[:validate_url_unreserved]}|
#{REGEXEN[:validate_url_pct_encoded]}|
[!$()*+,;=]
)/iox
REGEXEN[:validate_resid] = /(?:
#{REGEXEN[:validate_url_unreserved]}|
#{REGEXEN[:validate_url_pct_encoded]}|
#{REGEXEN[:validate_url_sub_delims]}
)/iox
REGEXEN[:valid_extended_uri] = %r{
( # $1 total match
(#{REGEXEN[:valid_url_preceding_chars]}) # $2 Preceding character
( # $3 URL
(
(xmpp:) # Protocol
(//#{REGEXEN[:validate_nodeid]}+@#{REGEXEN[:valid_domain]}/)? # Authority (optional)
(#{REGEXEN[:validate_nodeid]}+@)? # Username in path (optional)
(#{REGEXEN[:valid_domain]}) # Domain in path
(/#{REGEXEN[:validate_resid]}+)? # Resource in path (optional)
(\?#{REGEXEN[:valid_url_query_chars]}*#{REGEXEN[:valid_url_query_ending_chars]})? # Query String
) | (
(magnet:) # Protocol
(\?#{REGEXEN[:valid_url_query_chars]}*#{REGEXEN[:valid_url_query_ending_chars]}) # Query String
)
)
)
}iox
end
end