generate missing text parts on incoming messages

fixes CNVS-19005

test plan
- reply to an email from canvas with an email that does not have a
 text part
- ensure that the posted discussion/conversation/whatever does not
 include html tags in the posted text

Change-Id: I3223f10600e84aa0fb44f13435165557aaf59603
Reviewed-on: https://gerrit.instructure.com/53071
Tested-by: Jenkins
Reviewed-by: Andrew Butterfield <abutterfield@instructure.com>
QA-Review: Steven Shepherd <sshepherd@instructure.com>
Product-Review: Joel Hough <joel@instructure.com>

gems/incoming_mail_processor/lib/incoming_mail_processor/incoming_message_processor.rb

@@ -87,16 +87,27 @@ module IncomingMailProcessor
         text_part =  incoming_message.text_part
 
         html_body = self.class.utf8ify(html_part.body.decoded, html_part.charset) if html_part
-        body = self.class.utf8ify(text_part.body.decoded, text_part.charset)
+        text_body = self.class.utf8ify(text_part.body.decoded, text_part.charset) if text_part
       else
-        body = self.class.utf8ify(incoming_message.body.decoded, incoming_message.charset)
+        case incoming_message.mime_type
+        when 'text/plain'
+          text_body = self.class.utf8ify(incoming_message.body.decoded, incoming_message.charset)
+        when 'text/html'
+          html_body = self.class.utf8ify(incoming_message.body.decoded, incoming_message.charset)
+        else
+          raise "Unrecognized Content-Type: #{incoming_message.mime_type.inspect}"
+        end
       end
 
-      if !html_body
-        html_body = self.class.format_message(body).first
+      if html_body && !text_body
+        text_body = self.class.html_to_text(html_body)
       end
 
-      return body, html_body
+      if text_body && !html_body
+        html_body = self.class.format_message(text_body).first
+      end
+
+      return text_body, html_body
     end
 
     def self.mailbox_keys

gems/incoming_mail_processor/spec/incoming_mail_processor/incoming_message_processor_spec.rb

@@ -63,11 +63,11 @@ describe IncomingMailProcessor::IncomingMessageProcessor do
   def test_message (filename)
     message = get_processed_message(filename)
 
-    text_body =  message.body.strip!
-    text_body.should == get_expected_text(filename)
+    text_body =  message.body.strip
+    text_body.should == get_expected_text(filename).strip
 
-    html_body =  message.html_body.strip!
-    html_body.should == get_expected_html(filename)
+    html_body =  message.html_body.strip
+    html_body.should == get_expected_html(filename).strip
   end
 
   def get_processed_message(name)
@@ -128,7 +128,9 @@ describe IncomingMailProcessor::IncomingMessageProcessor do
 
   describe "#process_single" do
     it "should not choke on invalid UTF-8" do
-      IncomingMessageProcessor.new(message_handler, error_reporter).process_single(Mail.new { body "he\xffllo".force_encoding(Encoding::BINARY) }, '')
+      IncomingMessageProcessor.new(message_handler, error_reporter).process_single(Mail.new {
+          content_type 'text/plain; charset=UTF-8'
+          body "he\xffllo".force_encoding(Encoding::BINARY) }, '')
 
       message_handler.body.should == "hello"
       message_handler.html_body.should == "hello"
@@ -169,6 +171,26 @@ describe IncomingMailProcessor::IncomingMessageProcessor do
       IncomingMessageProcessor.new(message_handler, error_reporter).process_single(incoming_bounce_message, '')
     end
 
+    it "creates text body from html only messages" do
+      IncomingMessageProcessor.new(message_handler, error_reporter).process_single(Mail.new {
+          content_type 'text/html; charset=UTF-8'
+          body '<h1>This is HTML</h1>'
+        }, '')
+      message_handler.body.should == "************\nThis is HTML\n************"
+      message_handler.html_body.should == '<h1>This is HTML</h1>'
+    end
+
+    it "creates missing text part from html part" do
+      IncomingMessageProcessor.new(message_handler, error_reporter).process_single(Mail.new {
+          html_part do
+            content_type 'text/html; charset=UTF-8'
+            body '<h1>This is HTML</h1>'
+          end
+        }, '')
+      message_handler.body.should == "************\nThis is HTML\n************"
+      message_handler.html_body.should == '<h1>This is HTML</h1>'
+    end
+
     it "works with multipart emails with no html part" do
       test_message('multipart_mixed_no_html_part.eml')
     end