Add a lexer for emails

rouge-ruby · pyrmont · Oct 13, 2020 · Jul 22, 2020 · Jul 22, 2020 · Jul 22, 2020
commit bd456920637bcaa518fe7c20673ce51137daaeb5
diff --git a/lib/rouge/demos/email b/lib/rouge/demos/email
@@ -0,0 +1,11 @@
+From: Me <me@example.com>
+To: You <you@example.com>
+Date: Tue, 21 Jul 2020 15:14:03 +0000
+Subject: A very important message
+
+> Please investigate.  Thank you.
+
+I have investigated.
+
+-- 
+This message is highly confidential and will self-destruct.
diff --git a/lib/rouge/lexers/email.rb b/lib/rouge/lexers/email.rb
@@ -0,0 +1,46 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+module Rouge
+  module Lexers
+    class Email < RegexLexer
+      tag 'email'
+      aliases 'eml', 'mail'
+      filenames '*.eml', '*.mail'
+      mimetypes 'message/rfc822'
+
+      title "Email"
+      desc "An email message"
+
+      def self.detect?(text)
+        return true if text.start_with?('From: ')
+      end
+
+      state :root do
+        rule %r/^(From|To|Cc|Bcc):\s/, Keyword, :address
+        rule %r/^Date:\s/, Keyword, :date
+        rule %r/^Subject:\s/, Keyword, :subject
+        rule %r/\n/m, Text::Whitespace
+        rule %r/^>.*/, Comment
+        rule %r/^--\s\n/m, Comment::Doc, :signature
+        rule %r/.*/, Text
+      end
+
+      state :address do
+        rule %r/[^\n]+\n/m, Name, :pop!
+      end
+
+      state :date do
+        rule %r/[^\n]+\n/m, Literal::Date, :pop!
+      end
+
+      state :subject do
+        rule %r/[^\n]+\n/m, Name::Label, :pop!
+      end
+
+      state :signature do
+        rule %r/.*/m, Comment::Doc
+      end
+    end
+  end
+end
diff --git a/spec/lexers/email_spec.rb b/spec/lexers/email_spec.rb
@@ -0,0 +1,23 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+describe Rouge::Lexers::Email do
+  let(:subject) { Rouge::Lexers::Email.new }
+
+  describe 'guessing' do
+    include Support::Guessing
+
+    it 'guesses by filename' do
+      assert_guess :filename => 'foo.eml'
+      assert_guess :filename => 'foo.mail'
+    end
+
+    it 'guesses by mimetype' do
+      assert_guess :mimetype => 'message/rfc822'
+    end
+
+    it 'guesses by source' do
+      assert_guess :source => 'From: Me <me@example.com>'
+    end
+  end
+end
diff --git a/spec/visual/samples/email b/spec/visual/samples/email
@@ -0,0 +1,11 @@
+From: Me <me@example.com>
+To: You <you@example.com>
+Date: Tue, 21 Jul 2020 15:14:03 +0000
+Subject: A very important message
+
+> Please investigate.  Thank you.
+
+I have investigated.
+
+-- 
+This message is highly confidential and will self-destruct.