twitter-text 1.5.0 → 1.6.0

data/README.rdoc

@@ -1,4 +1,4 @@
-== twitter-text {<img src="https://secure.travis-ci.org/twitter/twitter-text-rb.png" />}[http://travis-ci.org/twitter/twitter-text-rb]
+== twitter-text {<img src="https://secure.travis-ci.org/twitter/twitter-text-rb.png" />}[http://travis-ci.org/twitter/twitter-text-rb] {<img src="https://codeclimate.com/badge.png" />}[https://codeclimate.com/github/twitter/twitter-text-rb]
 
 A gem that provides text processing routines for Twitter Tweets. The major
 reason for this is to unify the various auto-linking and extraction of

data/lib/twitter-text/autolink.rb

@@ -207,7 +207,7 @@ module Twitter
     OPTIONS_NOT_ATTRIBUTES = Set.new([
       :url_class, :list_class, :username_class, :hashtag_class, :cashtag_class,
       :username_url_base, :list_url_base, :hashtag_url_base, :cashtag_url_base,
-      :username_url_block, :list_url_block, :hashtag_url_block, :link_url_block,
+      :username_url_block, :list_url_block, :hashtag_url_block, :cashtag_url_block, :link_url_block,
       :username_include_symbol, :suppress_lists, :suppress_no_follow, :url_entities,
       :invisible_tag_attrs, :symbol_tag, :text_with_symbol_tag, :url_target,
       :link_attribute_block, :link_text_block
@@ -326,6 +326,11 @@ module Twitter
       hash = chars[entity[:indices].first]
       hashtag = entity[:hashtag]
       hashtag = yield(hashtag) if block_given?
+      hashtag_class = options[:hashtag_class]
+
+      if hashtag.match Twitter::Regex::REGEXEN[:rtl_chars]
+        hashtag_class += ' rtl'
+      end
 
       href = if options[:hashtag_url_block]
         options[:hashtag_url_block].call(hashtag)
@@ -334,7 +339,7 @@ module Twitter
       end
 
       html_attrs = {
-        :class => "#{options[:hashtag_class]}",
+        :class => hashtag_class,
         # FIXME As our conformance test, hash in title should be half-width,
         # this should be bug of conformance data.
         :title => "##{hashtag}"

data/lib/twitter-text/extractor.rb

@@ -311,13 +311,12 @@ module Twitter
       return [] unless text =~ /\$/
 
       tags = []
-      text.scan(Twitter::Regex[:valid_cashtag]) do |cash_text|
+      text.scan(Twitter::Regex[:valid_cashtag]) do |before, dollar, cash_text|
         match_data = $~
-        # cash_text doesn't contain $ symbol, so need to decrement start_position by one
-        start_position = match_data.char_begin(1) - 1
-        end_position = match_data.char_end(1)
+        start_position = match_data.char_begin(2)
+        end_position = match_data.char_end(3)
         tags << {
-          :cashtag => cash_text[0],
+          :cashtag => cash_text,
           :indices => [start_position, end_position]
         }
       end

data/lib/twitter-text/regex.rb

@@ -2,7 +2,7 @@
 
 module Twitter
   # A collection of regular expressions for parsing Tweet text. The regular expression
-  # list is frozen at load time to ensure immutability. These reular expressions are
+  # list is frozen at load time to ensure immutability. These regular expressions are
   # used throughout the <tt>Twitter</tt> classes. Special care has been taken to make
   # sure these reular expressions work with Tweets in all languages.
   class Regex
@@ -82,6 +82,14 @@ module Twitter
           regex_range(0x1e00, 0x1eff)
     ].join('').freeze
 
+    RTL_CHARACTERS = [
+      regex_range(0x0600,0x06FF),
+      regex_range(0x0750,0x077F),
+      regex_range(0x0590,0x05FF),
+      regex_range(0xFE70,0xFEFF)
+    ].join('').freeze
+
+
     NON_LATIN_HASHTAG_CHARS = [
       # Cyrillic (Russian, Ukrainian, etc.)
       regex_range(0x0400, 0x04ff), # Cyrillic
@@ -212,7 +220,7 @@ module Twitter
 
     REGEXEN[:valid_port_number] = /[0-9]+/
 
-    REGEXEN[:valid_general_url_path_chars] = /[a-z0-9!\*';:=\+\,\.\$\/%#\[\]\-_~&|#{LATIN_ACCENTS}]/io
+    REGEXEN[:valid_general_url_path_chars] = /[a-z0-9!\*';:=\+\,\.\$\/%#\[\]\-_~&|@#{LATIN_ACCENTS}]/io
     # Allow URL paths to contain balanced parens
     #  1. Used in Wikipedia URLs like /Primer_(film)
     #  2. Used in IIS sessions like /S(dfd346)/
@@ -220,16 +228,15 @@ module Twitter
     # Valid end-of-path chracters (so /foo. does not gobble the period).
     #   1. Allow =&# for empty URL parameters and other URL-join artifacts
     REGEXEN[:valid_url_path_ending_chars] = /[a-z0-9=_#\/\+\-#{LATIN_ACCENTS}]|(?:#{REGEXEN[:valid_url_balanced_parens]})/io
-    # Allow @ in a url, but only in the middle. Catch things like http://example.com/@user/
     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]}+\/)
+      )|(?:#{REGEXEN[:valid_general_url_path_chars]}+\/)
     )/iox
 
-    REGEXEN[:valid_url_query_chars] = /[a-z0-9!?\*'\(\);:&=\+\$\/%#\[\]\-_\.,~|]/i
+    REGEXEN[:valid_url_query_chars] = /[a-z0-9!?\*'\(\);:&=\+\$\/%#\[\]\-_\.,~|@]/i
     REGEXEN[:valid_url_query_ending_chars] = /[a-z0-9_&=#\/]/i
     REGEXEN[:valid_url] = %r{
       (                                                                                     #   $1 total match
@@ -245,7 +252,7 @@ module Twitter
     }iox;
 
     REGEXEN[:cashtag] = /[a-z]{1,6}(?:[._][a-z]{1,2})?/i
-    REGEXEN[:valid_cashtag] = /(?:^|#{REGEXEN[:spaces]})\$(#{REGEXEN[:cashtag]})(?=$|\s|[#{PUNCTUATION_CHARS}])/i
+    REGEXEN[:valid_cashtag] = /(^|#{REGEXEN[:spaces]})(\$)(#{REGEXEN[:cashtag]})(?=$|\s|[#{PUNCTUATION_CHARS}])/i
 
     # These URL validation pattern strings are based on the ABNF from RFC 3986
     REGEXEN[:validate_url_unreserved] = /[a-z0-9\-._~]/i
@@ -342,6 +349,8 @@ module Twitter
       )?\Z
     }ix
 
+    REGEXEN[:rtl_chars] = /[#{RTL_CHARACTERS}]/io
+
     REGEXEN.each_pair{|k,v| v.freeze }
 
     # Return the regular expression for a given <tt>key</tt>. If the <tt>key</tt>

data/lib/twitter-text/validation.rb

@@ -3,8 +3,9 @@ module Twitter
     MAX_LENGTH = 140
 
     DEFAULT_TCO_URL_LENGTHS = {
-      :short_url_length => 20,
-      :short_url_length_https => 21
+      :short_url_length => 22,
+      :short_url_length_https => 23,
+      :characters_reserved_per_media => 22
     }.freeze
 
     # Returns the length of the string as it would be displayed. This is equivilent to the length of the Unicode NFC

data/spec/autolinking_spec.rb

@@ -488,9 +488,8 @@ describe Twitter::Autolink do
         context "with a username not followed by a /" do
           def original_text; 'http://example.com/@foobar'; end
 
-          it "should link small url and username" do
-            @autolinked_text.should have_autolinked_url('http://example.com/')
-            @autolinked_text.should link_to_screen_name('foobar')
+          it "should link url" do
+            @autolinked_text.should have_autolinked_url('http://example.com/@foobar')
           end
         end
 
@@ -678,6 +677,11 @@ describe Twitter::Autolink do
       linked.should have_autolinked_url('dummy', '#hashtag')
     end
 
+    it "should customize href by cashtag_url_block option" do
+      linked = @linker.auto_link("$CASH", :cashtag_url_block => lambda{|a| "dummy"})
+      linked.should have_autolinked_url('dummy', '$CASH')
+    end
+
     it "should customize href by link_url_block option" do
       linked = @linker.auto_link("http://example.com/", :link_url_block => lambda{|a| "dummy"})
       linked.should have_autolinked_url('dummy', 'http://example.com/')

data/spec/rewriter_spec.rb

@@ -527,9 +527,9 @@ describe Twitter::Rewriter do
       context "with a username not followed by a /" do
         def original_text; "http://example.com/@foobar"; end
 
-        it "should link small url and username" do
-          @block_args.should == ["http://example.com/"]
-          @rewritten_text.should == "[rewritten]@foobar"
+        it "should link url" do
+          @block_args.should == ["http://example.com/@foobar"]
+          @rewritten_text.should == "[rewritten]"
         end
       end
 

data/test/conformance_test.rb

@@ -33,7 +33,7 @@ class ConformanceTest < Test::Unit::TestCase
   def equal_nodes?(expected, actual)
     return false unless expected.name == actual.name
     return false unless ordered_attributes(expected) == ordered_attributes(actual)
-    return false if expected.text? && actual.text? && !(expected.content= actual.content)
+    return false if expected.text? && actual.text? && expected.content != actual.content
 
     expected.children.each_with_index do |child, index|
       return false unless equal_nodes?(child, actual.children[index])

data/twitter-text.gemspec

@@ -2,11 +2,11 @@
 
 Gem::Specification.new do |s|
   s.name = "twitter-text"
-  s.version = "1.5.0"
+  s.version = "1.6.0"
   s.authors = ["Matt Sanford", "Patrick Ewing", "Ben Cherry", "Britt Selvitelle",
-               "Raffi Krikorian", "J.P. Cummins", "Yoshimasa Niwa", "Keita Fujii"]
+               "Raffi Krikorian", "J.P. Cummins", "Yoshimasa Niwa", "Keita Fujii", "James Koval"]
   s.email = ["matt@twitter.com", "patrick.henry.ewing@gmail.com", "bcherry@gmail.com", "bs@brittspace.com",
-             "raffi@twitter.com", "jcummins@twitter.com", "niw@niw.at", "keita@twitter.com"]
+             "raffi@twitter.com", "jcummins@twitter.com", "niw@niw.at", "keita@twitter.com", "jkoval@twitter.com"]
   s.homepage = "http://twitter.com"
   s.description = s.summary = "A gem that provides text handling for Twitter"
 

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: twitter-text
 version: !ruby/object:Gem::Version 
-  hash: 3
+  hash: 15
   prerelease: 
   segments: 
   - 1
-  - 5
+  - 6
   - 0
-  version: 1.5.0
+  version: 1.6.0
 platform: ruby
 authors: 
 - Matt Sanford
@@ -18,12 +18,12 @@ authors:
 - J.P. Cummins
 - Yoshimasa Niwa
 - Keita Fujii
+- James Koval
 autorequire: 
 bindir: bin
 cert_chain: []
 
-date: 2012-06-18 00:00:00 -07:00
-default_executable: 
+date: 2013-03-01 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: nokogiri
@@ -119,6 +119,7 @@ email:
 - jcummins@twitter.com
 - niw@niw.at
 - keita@twitter.com
+- jkoval@twitter.com
 executables: []
 
 extensions: []
@@ -159,7 +160,6 @@ files:
 - spec/validation_spec.rb
 - test/conformance_test.rb
 - twitter-text.gemspec
-has_rdoc: true
 homepage: http://twitter.com
 licenses: []
 
@@ -189,7 +189,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.4.1
+rubygems_version: 1.8.24
 signing_key: 
 specification_version: 3
 summary: Twitter text handling library
@@ -205,3 +205,4 @@ test_files:
 - spec/unicode_spec.rb
 - spec/validation_spec.rb
 - test/conformance_test.rb
+has_rdoc: true