Fix ß bug in regexp for mentions and tags (#31122)

This commit is contained in:
Adam Niedzielski 2024-07-25 16:24:19 +02:00 committed by GitHub
parent e881a59671
commit 337a7fcf30
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 18 additions and 2 deletions

View file

@ -68,7 +68,7 @@ class Account < ApplicationRecord
INSTANCE_ACTOR_ID = -99 INSTANCE_ACTOR_ID = -99
USERNAME_RE = /[a-z0-9_]+([a-z0-9_.-]+[a-z0-9_]+)?/i USERNAME_RE = /[a-z0-9_]+([a-z0-9_.-]+[a-z0-9_]+)?/i
MENTION_RE = %r{(?<![=/[:word:]])@((#{USERNAME_RE})(?:@[[:word:].-]+[[:word:]]+)?)}i MENTION_RE = %r{(?<![=/[:word:]])@((#{USERNAME_RE})(?:@[[:word:].-]+[[:word:]]+)?)}
URL_PREFIX_RE = %r{\Ahttp(s?)://[^/]+} URL_PREFIX_RE = %r{\Ahttp(s?)://[^/]+}
USERNAME_ONLY_RE = /\A#{USERNAME_RE}\z/i USERNAME_ONLY_RE = /\A#{USERNAME_RE}\z/i
USERNAME_LENGTH_LIMIT = 30 USERNAME_LENGTH_LIMIT = 30

View file

@ -37,7 +37,7 @@ class Tag < ApplicationRecord
HASHTAG_LAST_SEQUENCE = '([[:word:]_]*[[:alpha:]][[:word:]_]*)' HASHTAG_LAST_SEQUENCE = '([[:word:]_]*[[:alpha:]][[:word:]_]*)'
HASHTAG_NAME_PAT = "#{HASHTAG_FIRST_SEQUENCE}|#{HASHTAG_LAST_SEQUENCE}" HASHTAG_NAME_PAT = "#{HASHTAG_FIRST_SEQUENCE}|#{HASHTAG_LAST_SEQUENCE}"
HASHTAG_RE = %r{(?<![=/)\p{Alnum}])#(#{HASHTAG_NAME_PAT})}i HASHTAG_RE = %r{(?<![=/)\p{Alnum}])#(#{HASHTAG_NAME_PAT})}
HASHTAG_NAME_RE = /\A(#{HASHTAG_NAME_PAT})\z/i HASHTAG_NAME_RE = /\A(#{HASHTAG_NAME_PAT})\z/i
HASHTAG_INVALID_CHARS_RE = /[^[:alnum:]\u0E47-\u0E4E#{HASHTAG_SEPARATORS}]/ HASHTAG_INVALID_CHARS_RE = /[^[:alnum:]\u0E47-\u0E4E#{HASHTAG_SEPARATORS}]/

View file

@ -711,6 +711,14 @@ RSpec.describe Account do
it 'does not match URL query string' do it 'does not match URL query string' do
expect(subject.match('https://example.com/?x=@alice')).to be_nil expect(subject.match('https://example.com/?x=@alice')).to be_nil
end end
it 'matches usernames immediately following the letter ß' do
expect(subject.match('Hello toß @alice from me')[1]).to eq 'alice'
end
it 'matches usernames containing uppercase characters' do
expect(subject.match('Hello to @aLice@Example.com from me')[1]).to eq 'aLice@Example.com'
end
end end
describe 'validations' do describe 'validations' do

View file

@ -95,6 +95,14 @@ RSpec.describe Tag do
it 'does not match purely-numeric hashtags' do it 'does not match purely-numeric hashtags' do
expect(subject.match('hello #0123456')).to be_nil expect(subject.match('hello #0123456')).to be_nil
end end
it 'matches hashtags immediately following the letter ß' do
expect(subject.match('Hello toß #ruby').to_s).to eq '#ruby'
end
it 'matches hashtags containing uppercase characters' do
expect(subject.match('Hello #rubyOnRails').to_s).to eq '#rubyOnRails'
end
end end
describe '#to_param' do describe '#to_param' do