Add e-mail lexer (rouge-ruby#1567)

This commit adds a lexer for e-mails. It follows RFC 2822 as well as supporting quoted lines in the form described in RFC 3676. Co-authored-by: Michael Camilleri <mike@inqk.net>
NSHipster · May 19, 2021 · 3a0e6e3 · 3a0e6e3
1 parent 2dd87f2
commit 3a0e6e3
Show file tree

Hide file tree

Showing 4 changed files with 91 additions and 0 deletions.
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 +1,39 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+module Rouge
+  module Lexers
+    class Email < RegexLexer
+      tag 'email'
+      aliases 'eml', 'e-mail'
+      filenames '*.eml'
+      mimetypes 'message/rfc822'
+
+      title "Email"
+      desc "An email message"
+
+      start do
+        push :fields
+      end
+
+      state :fields do
+        rule %r/[:]/, Operator, :field_body
+        rule %r/[^\n\r:]+/, Name::Tag
+        rule %r/[\n\r]/, Name::Tag
+      end
+
+      state :field_body do
+        rule(/(\r?\n){2}/) { token Text; pop!(2) }
+        rule %r/\r?\n(?![ \v\t\f])/, Text, :pop!
+        rule %r/[^\n\r]+/, Name::Attribute
+        rule %r/[\n\r]/, Name::Attribute
+      end
+
+      state :root do
+        rule %r/\n/, Text
+        rule %r/^>.*/, Comment
+        rule %r/.*/, Text
+      end
+    end
+  end
+end
diff --git a/spec/lexers/email_spec.rb b/spec/lexers/email_spec.rb
@@ -0,0 +1,18 @@
+# -*- 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'
+    end
+
+    it 'guesses by mimetype' do
+      assert_guess :mimetype => 'message/rfc822'
+    end
+  end
+end
diff --git a/spec/visual/samples/email b/spec/visual/samples/email
@@ -0,0 +1,23 @@
+From: Me <me@example.com>
+To: You <you@example.com>
+Cc: Somebody <somebody@example.com>,
+ And One More <andonemore@example.com>
+Bcc: Secret Person <secretperson@email.com>
+X-Spam-Status: Definitely not spam
+Date: Tue, 21 Jul 2020 15:14:03 +0000
+Subject: RE: A very important message
+ that continues onto the next line.
+
+Greetings and salutations.
+
+>> A second-level quotation.
+>
+> Please investigate.  Thank you.
+
+I have investigated.
+
+Note: A space-stuffed line starting with > is not a quote.
+ > like this.
+
+--
+This message is highly confidential and will self-destruct.