Fix URL scanning in note length validator and preview card fetching (#15827)

* Add tests

* Fix URL scanning in note length validator and preview card fetching
This commit is contained in:
Claire
2021-03-04 00:12:26 +01:00
committed by GitHub
parent 65db262550
commit 5614e6724e
6 changed files with 52 additions and 16 deletions

View File

@ -69,7 +69,7 @@ class LanguageDetector
def simplify_text(text)
new_text = remove_html(text)
new_text.gsub!(FetchLinkCardService::URL_PATTERN, '')
new_text.gsub!(FetchLinkCardService::URL_PATTERN, '\1')
new_text.gsub!(Account::MENTION_RE, '')
new_text.gsub!(Tag::HASHTAG_RE) { |string| string.gsub(/[#_]/, '#' => '', '_' => ' ').gsub(/[a-z][A-Z]|[a-zA-Z][\d]/) { |s| s.insert(1, ' ') }.downcase }
new_text.gsub!(/:#{CustomEmoji::SHORTCODE_RE_FRAGMENT}:/, '')

View File

@ -2,12 +2,13 @@
class FetchLinkCardService < BaseService
URL_PATTERN = %r{
( # $1 URL
(https?:\/\/) # $2 Protocol (required)
(#{Twitter::TwitterText::Regex[:valid_domain]}) # $3 Domain(s)
(?::(#{Twitter::TwitterText::Regex[:valid_port_number]}))? # $4 Port number (optional)
(/#{Twitter::TwitterText::Regex[:valid_url_path]}*)? # $5 URL Path and anchor
(\?#{Twitter::TwitterText::Regex[:valid_url_query_chars]}*#{Twitter::TwitterText::Regex[:valid_url_query_ending_chars]})? # $6 Query String
(#{Twitter::TwitterText::Regex[:valid_url_preceding_chars]}) # $1 preceeding chars
( # $2 URL
(https?:\/\/) # $3 Protocol (required)
(#{Twitter::TwitterText::Regex[:valid_domain]}) # $4 Domain(s)
(?::(#{Twitter::TwitterText::Regex[:valid_port_number]}))? # $5 Port number (optional)
(/#{Twitter::TwitterText::Regex[:valid_url_path]}*)? # $6 URL Path and anchor
(\?#{Twitter::TwitterText::Regex[:valid_url_query_chars]}*#{Twitter::TwitterText::Regex[:valid_url_query_ending_chars]})? # $7 Query String
)
}iox
@ -63,7 +64,7 @@ class FetchLinkCardService < BaseService
def parse_urls
if @status.local?
urls = @status.text.scan(URL_PATTERN).map { |array| Addressable::URI.parse(array[0]).normalize }
urls = @status.text.scan(URL_PATTERN).map { |array| Addressable::URI.parse(array[1]).normalize }
else
html = Nokogiri::HTML(@status.text)
links = html.css('a')

View File

@ -15,7 +15,7 @@ class NoteLengthValidator < ActiveModel::EachValidator
return '' if value.nil?
value.dup.tap do |new_text|
new_text.gsub!(FetchLinkCardService::URL_PATTERN, 'x' * 23)
new_text.gsub!(FetchLinkCardService::URL_PATTERN, StatusLengthValidator::URL_PLACEHOLDER)
new_text.gsub!(Account::MENTION_RE, '@\2')
end
end

View File

@ -2,12 +2,6 @@
class StatusLengthValidator < ActiveModel::Validator
MAX_CHARS = 500
URL_PATTERN = %r{
(?:
(#{Twitter::TwitterText::Regex[:valid_url_preceding_chars]})
(#{FetchLinkCardService::URL_PATTERN})
)
}iox
URL_PLACEHOLDER = "\1#{'x' * 23}"
def validate(status)
@ -35,7 +29,7 @@ class StatusLengthValidator < ActiveModel::Validator
return '' if @status.text.nil?
@status.text.dup.tap do |new_text|
new_text.gsub!(URL_PATTERN, URL_PLACEHOLDER)
new_text.gsub!(FetchLinkCardService::URL_PATTERN, URL_PLACEHOLDER)
new_text.gsub!(Account::MENTION_RE, '@\2')
end
end