twitter-text 1.0

data/lib/validation.rb

@@ -0,0 +1,51 @@
+
+module Twitter
+  module Validation
+    MAX_LENGTH = 140
+
+    # Character not allowed in Tweets
+    INVALID_CHARACTERS = [
+      0xFFFE, 0xFEFF, # BOM
+      0xFFFF,         # Special
+      0x202A, 0x202B, 0x202C, 0x202D, 0x202E # Directional change
+    ].map{|cp| [cp].pack('U') }.freeze
+
+    # Returns the length of the string as it would be displayed. This is equivilent to the length of the Unicode NFC
+    # (See: http://www.unicode.org/reports/tr15). This is needed in order to consistently calculate the length of a
+    # string no matter which actual form was transmitted. For example:
+    #
+    #     U+0065  Latin Small Letter E
+    # +   U+0301  Combining Acute Accent
+    # ----------
+    # =   2 bytes, 2 characters, displayed as é (1 visual glyph)
+    #     … The NFC of {U+0065, U+0301} is {U+00E9}, which is a single chracter and a +display_length+ of 1
+    #
+    # The string could also contain U+00E9 already, in which case the canonicalization will not change the value.
+    #
+    def tweet_length(text)
+      ActiveSupport::Multibyte::Chars.new(text).normalize(:c).length
+    end
+
+    # Check the <tt>text</tt> for any reason that it may not be valid as a Tweet. This is meant as a pre-validation
+    # before posting to api.twitter.com. There are several server-side reasons for Tweets to fail but this pre-validation
+    # will allow quicker feedback.
+    #
+    # Returns <tt>false</tt> if this <tt>text</tt> is valid. Otherwise one of the following Symbols will be returned:
+    #
+    #   <tt>:too_long</tt>:: if the <tt>text</tt> is too long
+    #   <tt>:empty</tt>:: if the <tt>text</tt> is nil or empty
+    #   <tt>:invalid_characters</tt>:: if the <tt>text</tt> contains non-Unicode or any of the disallowed Unicode characters
+    def tweet_invalid?(text)
+      return :empty if text.blank?
+      begin
+        return :too_long if tweet_length(text) > MAX_LENGTH
+        return :invalid_characters if INVALID_CHARACTERS.any?{|invalid_char| text.include?(invalid_char) }
+      rescue ArgumentError, ActiveSupport::Multibyte::EncodingError => e
+        # non-Unicode value.
+        return :invalid_characters
+      end
+
+      return false
+    end
+  end
+end

data/spec/autolinking_spec.rb

@@ -0,0 +1,427 @@
+require File.dirname(__FILE__) + '/spec_helper'
+
+class TestAutolink
+  include Twitter::Autolink
+end
+
+describe Twitter::Autolink do
+  def original_text; end
+  def url; end
+
+  describe "auto_link_custom" do
+    before do
+      @autolinked_text = TestAutolink.new.auto_link(original_text) if original_text
+    end
+
+    describe "username autolinking" do
+      context "username preceded by a space" do
+        def original_text; "hello @jacob"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('jacob')
+        end
+      end
+
+      context "username at beginning of line" do
+        def original_text; "@jacob you're cool"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('jacob')
+        end
+      end
+
+      context "username preceded by word character" do
+        def original_text; "meet@the beach"; end
+
+        it "should not be linked" do
+          Hpricot(@autolinked_text).search('a').should be_blank
+        end
+      end
+
+      context "username preceded by non-word character" do
+        def original_text; "great.@jacob"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('jacob')
+        end
+      end
+
+      context "username containing non-word characters" do
+        def original_text; "@zach&^$%^"; end
+
+        it "should not be linked" do
+          @autolinked_text.should link_to_screen_name('zach')
+        end
+      end
+
+      context "username over twenty characters" do
+        def original_text
+          @twenty_character_username = "zach" * 5
+          "@" + @twenty_character_username + "1"
+        end
+
+        it "should not be linked" do
+          @autolinked_text.should link_to_screen_name(@twenty_character_username)
+        end
+      end
+
+      context "username followed by japanese" do
+        def original_text; "@jacobの"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('jacob')
+        end
+      end
+
+      context "username preceded by japanese" do
+        def original_text; "あ@matz"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('matz')
+        end
+      end
+
+      context "username surrounded by japanese" do
+        def original_text; "あ@yoshimiの"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('yoshimi')
+        end
+      end
+
+      context "username using full-width at-sign" do
+        def original_text
+          "#{[0xFF20].pack('U')}jacob"
+        end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_screen_name('jacob')
+        end
+      end
+    end
+
+    describe "list path autolinking" do
+
+      context "when List is not available" do
+        it "should not be linked" do
+          @autolinked_text = TestAutolink.new.auto_link_usernames_or_lists("hello @jacob/my-list", :suppress_lists => true)
+          @autolinked_text.should_not link_to_list_path('jacob/my-list')
+        end
+      end
+
+      context "slug preceded by a space" do
+        def original_text; "hello @jacob/my-list"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_list_path('jacob/my-list')
+        end
+      end
+
+      context "username followed by a slash but no list" do
+        def original_text; "hello @jacob/ my-list"; end
+
+        it "should NOT be linked" do
+          @autolinked_text.should_not link_to_list_path('jacob/my-list')
+          @autolinked_text.should link_to_screen_name('jacob')
+        end
+      end
+
+      context "empty username followed by a list" do
+        def original_text; "hello @/my-list"; end
+
+        it "should NOT be linked" do
+          Hpricot(@autolinked_text).search('a').should be_blank
+        end
+      end
+
+      context "list slug at beginning of line" do
+        def original_text; "@jacob/my-list"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_list_path('jacob/my-list')
+        end
+      end
+
+      context "username preceded by alpha-numeric character" do
+        def original_text; "meet@the/beach"; end
+
+        it "should not be linked" do
+          Hpricot(@autolinked_text).search('a').should be_blank
+        end
+      end
+
+      context "username preceded by non-word character" do
+        def original_text; "great.@jacob/my-list"; end
+
+        it "should be linked" do
+          @autolinked_text = TestAutolink.new.auto_link("great.@jacob/my-list")
+          @autolinked_text.should link_to_list_path('jacob/my-list')
+        end
+      end
+
+      context "username containing non-word characters" do
+        def original_text; "@zach/test&^$%^"; end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_list_path('zach/test')
+        end
+      end
+
+      context "username over twenty characters" do
+        def original_text
+          @eighty_character_list = "jack/" + ("a" * 80)
+          "@#{@eighty_character_list}12345"
+        end
+
+        it "should be linked" do
+          @autolinked_text.should link_to_list_path(@eighty_character_list)
+        end
+      end
+    end
+
+    describe "hashtag autolinking" do
+      context "with an all numeric hashtag" do
+        def original_text; "#123"; end
+
+        it "should not be linked" do
+          @autolinked_text.should_not have_autolinked_hashtag('#123')
+        end
+      end
+
+      context "with a hashtag with alphanumeric characters" do
+        def original_text; "#ab1d"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag('#ab1d')
+        end
+      end
+
+      context "with a hashtag with underscores" do
+        def original_text; "#a_b_c_d"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag(original_text)
+        end
+      end
+
+      context "with a hashtag that is preceded by a word character" do
+        def original_text; "ab#cd"; end
+
+        it "should not be linked" do
+          @autolinked_text.should_not have_autolinked_hashtag(original_text)
+        end
+      end
+
+      context "with a page anchor in a url" do
+        def original_text; "Here's my url: http://foobar.com/#home"; end
+
+        it "should not link the hashtag" do
+          @autolinked_text.should_not have_autolinked_hashtag('#home')
+        end
+
+        it "should link the url" do
+          @autolinked_text.should have_autolinked_url('http://foobar.com/#home')
+        end
+      end
+
+      context "with a hashtag that starts with a number but has word characters" do
+        def original_text; "#2ab"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag(original_text)
+        end
+      end
+
+      context "with multiple valid hashtags" do
+        def original_text; "I'm frickin' awesome #ab #cd #ef"; end
+
+        it "links each hashtag" do
+          @autolinked_text.should have_autolinked_hashtag('#ab')
+          @autolinked_text.should have_autolinked_hashtag('#cd')
+          @autolinked_text.should have_autolinked_hashtag('#ef')
+        end
+      end
+
+      context "with a hashtag preceded by a ." do
+        def original_text; "ok, great.#abc"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag('#abc')
+        end
+      end
+
+      context "with a hashtag preceded by a &" do
+        def original_text; "&#nbsp;"; end
+
+        it "should not be linked" do
+          @autolinked_text.should_not have_autolinked_hashtag('#nbsp;')
+        end
+      end
+
+      context "with a hashtag that ends in an !" do
+        def original_text; "#great!"; end
+
+        it "should be linked, but should not include the !" do
+          @autolinked_text.should have_autolinked_hashtag('#great')
+        end
+      end
+
+      context "with a hashtag preceded by Japanese" do
+        def original_text; "の#twj_dev"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag('#twj_dev')
+        end
+      end
+
+      context "with a hashtag followed by Japanese" do
+         def original_text; "#twj_devの"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag('#twj_dev')
+        end
+      end
+
+      context "with a hashtag preceded by a full-width space" do
+        def original_text; "#{[0x3000].pack('U')}#twj_dev"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag('#twj_dev')
+        end
+      end
+
+      context "with a hashtag followed by a full-width space" do
+        def original_text; "#twj_dev#{[0x3000].pack('U')}"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_hashtag('#twj_dev')
+        end
+      end
+
+      context "with a hashtag using full-width hash" do
+        def original_text; "#{[0xFF03].pack('U')}twj_dev"; end
+
+        it "should be linked" do
+          link = Hpricot(@autolinked_text).at('a')
+          link.inner_text.should == "#{[0xFF03].pack('U')}twj_dev"
+          link['href'].should == 'http://twitter.com/search?q=%23twj_dev'
+        end
+      end
+
+    end
+
+    describe "URL autolinking" do
+      def url; "http://www.google.com"; end
+
+      context "when embedded in plain text" do
+        def original_text; "On my search engine #{url} I found good links."; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_url(url)
+        end
+      end
+
+      context "when surrounded by Japanese;" do
+        def original_text; "いまなにしてる#{url}いまなにしてる"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_url(url)
+        end
+      end
+
+      context "when surrounded by parentheses;" do
+        def original_text; "I found a neatness (#{url})"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_url(url)
+        end
+
+        context "when the URL ends with a slash;" do
+          def url; "http://www.google.com/"; end
+
+          it "should be linked" do
+            pending # our support for Wikipedia URLS containing parens breaks this corner case
+            @autolinked_text.should have_autolinked_url(url)
+          end
+        end
+      end
+
+      context "with a URL ending in allowed punctuation" do
+        it "does not consume ending punctuation" do
+          matcher = TestAutolink.new
+          %w| ? ! , . : ; ] ) } = \ ' |.each do |char|
+            matcher.auto_link("#{url}#{char}").should have_autolinked_url(url)
+          end
+        end
+      end
+
+      context "with a URL preceded in forbidden characters" do
+        it "should not be linked" do
+          matcher = TestAutolink.new
+          %w| \ ' / : ! = |.each do |char|
+            matcher.auto_link("#{char}#{url}").should_not have_autolinked_url(url)
+          end
+        end
+      end
+
+      context "when embedded in a link tag" do
+        def original_text; "<link rel='true'>#{url}</link>"; end
+
+        it "should be linked" do
+          @autolinked_text.should have_autolinked_url(url)
+        end
+      end
+
+      context "with multiple URLs" do
+        def original_text; "http://www.links.org link at start of page, link at end http://www.foo.org"; end
+
+        it "should autolink each one" do
+          @autolinked_text.should have_autolinked_url('http://www.links.org')
+          @autolinked_text.should have_autolinked_url('http://www.foo.org')
+        end
+      end
+
+      context "with multiple URLs in different formats" do
+        def original_text; "http://foo.com https://bar.com http://mail.foobar.org"; end
+
+        it "should autolink each one, in the proper order" do
+          @autolinked_text.should have_autolinked_url('http://foo.com')
+          @autolinked_text.should have_autolinked_url('https://bar.com')
+          @autolinked_text.should have_autolinked_url('http://mail.foobar.org')
+        end
+      end
+
+      context "with a URL having a long TLD" do
+        def original_text; "Yahoo integriert Facebook http://golem.mobi/0912/71607.html"; end
+
+        it "should autolink it" do
+          @autolinked_text.should have_autolinked_url('http://golem.mobi/0912/71607.html')
+        end
+      end
+
+      context "with a url lacking the protocol" do
+        def original_text; "I like www.foobar.com dudes"; end
+
+        it "links to the original text with the full href" do
+          link = Hpricot(@autolinked_text).at('a')
+          link.inner_text.should == 'www.foobar.com'
+          link['href'].should == 'http://www.foobar.com'
+        end
+      end
+
+    end
+
+    describe "Autolink all" do
+      before do
+        @linker = TestAutolink.new
+      end
+
+      it "should allow url/hashtag overlap" do
+        auto_linked = @linker.auto_link("http://twitter.com/#search")
+        auto_linked.should have_autolinked_url('http://twitter.com/#search')
+      end
+
+    end
+  end
+
+end