Handle parsing of LF-only body with separate parts

.github/workflows/adhoc.yml

@@ -0,0 +1,80 @@
+name: Adhoc RSpec tests
+
+on:
+  workflow_dispatch:
+
+  pull_request:
+    branches: [ feature/parse_lf ]
+  push:
+    branches: [ feature/parse_lf ]
+
+jobs:
+  ruby:
+    name: ${{ matrix.ruby }} (timeout ${{ matrix.timeout }})
+    runs-on: ubuntu-latest
+    timeout-minutes: ${{ matrix.timeout }}
+    strategy:
+      fail-fast: false
+      matrix:
+       include:
+         - ruby: 2.5
+           timeout: 5
+         - ruby: 2.6
+           timeout: 5
+         - ruby: 2.7
+           timeout: 5
+         - ruby: '3.0'
+           timeout: 5
+         - ruby: 3.1
+           timeout: 5
+         - ruby: jruby
+           timeout: 10
+         - ruby: jruby-head
+           timeout: 10
+         - ruby: jruby-9.4
+           timeout: 10
+         - ruby: jruby-9.3
+           timeout: 10
+         - ruby: jruby-9.2
+           timeout: 10
+    steps:
+    - name: Psych 5.0.0 needs libyaml-dev
+      run: |
+        sudo apt-get update
+        sudo apt-get install libyaml-dev
+    - uses: actions/checkout@v2
+      with:
+        ref: feature/parse_lf
+    - uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: ${{ matrix.ruby }}
+        rubygems: 3.2.3
+        bundler-cache: true
+        cache-version: 4
+    - name: Show psych
+      run: |
+        gem list psych --details
+      continue-on-error: true
+    - name: Run tests1
+      run: bundle exec rake spec || echo "action_state=failed" >> $GITHUB_ENV
+      continue-on-error: true
+    - name: Run tests2
+      run: bundle exec rake spec || echo "action_state=failed" >> $GITHUB_ENV
+      continue-on-error: true
+    - name: Run tests3
+      run: bundle exec rake spec || echo "action_state=failed" >> $GITHUB_ENV
+      continue-on-error: true
+    - name: Run tests4
+      run: bundle exec rake spec || echo "action_state=failed" >> $GITHUB_ENV
+      continue-on-error: true
+    - name: Run tests5
+      run: bundle exec rake spec || echo "action_state=failed" >> $GITHUB_ENV
+      continue-on-error: true
+    - name: Run tests6
+      run: bundle exec rake spec || echo "action_state=failed" >> $GITHUB_ENV
+      continue-on-error: true
+    - name: Run Summary
+      run: |
+        echo "${{ env.action_state }}"
+        # This will be be true for a successful run
+        test "${{ env.action_state }}" != 'failed'

lib/mail/body.rb

@@ -39,9 +39,9 @@ def initialize(string = '')
       else
         # Do join first incase we have been given an Array in Ruby 1.9
         if string.respond_to?(:join)
-          @raw_source = ::Mail::Utilities.to_crlf(string.join(''))
+          @raw_source = string.join('')
         elsif string.respond_to?(:to_s)
-          @raw_source = ::Mail::Utilities.to_crlf(string.to_s)
+          @raw_source = string.to_s
         else
           raise "You can only assign a string or an object that responds_to? :join or :to_s to a body."
         end
@@ -272,7 +272,7 @@ def extract_parts
       parts_regex = /
         (?:                    # non-capturing group
           \A                |  # start of string OR
-          \r\n                 # line break
+          \r?\n                # line break with optional CR
          )
         (
           --#{Regexp.escape(boundary || "")}  # boundary delimiter

spec/fixtures/emails/attachment_emails/attachment_pdf_non_ascii.eml

@@ -0,0 +1,70 @@
+From [email protected] Tue May 10 11:28:07 2005
+Return-Path: <[email protected]>
+X-Original-To: [email protected]
+Delivered-To: [email protected]
+Received: from localhost (localhost [127.0.0.1])
+	by xxx.xxxxx.com (Postfix) with ESMTP id 50FD3A96F
+	for <[email protected]>; Tue, 10 May 2005 17:26:50 +0000 (GMT)
+Received: from xxx.xxxxx.com ([127.0.0.1])
+ by localhost (xxx.xxxxx.com [127.0.0.1]) (amavisd-new, port 10024)
+ with LMTP id 70060-03 for <[email protected]>;
+ Tue, 10 May 2005 17:26:49 +0000 (GMT)
+Received: from xxx.xxxxx.com (xxx.xxxxx.com [69.36.39.150])
+	by xxx.xxxxx.com (Postfix) with ESMTP id 8B957A94B
+	for <[email protected]>; Tue, 10 May 2005 17:26:48 +0000 (GMT)
+Received: from xxx.xxxxx.com (xxx.xxxxx.com [64.233.184.203])
+	by xxx.xxxxx.com (Postfix) with ESMTP id 9972514824C
+	for <[email protected]>; Tue, 10 May 2005 12:26:40 -0500 (CDT)
+Received: by xxx.xxxxx.com with SMTP id 68so1694448wri
+        for <[email protected]>; Tue, 10 May 2005 10:26:40 -0700 (PDT)
+DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
+        s=beta; d=xxxxx.com;
+        h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type;
+        b=g8ZO5ttS6GPEMAz9WxrRk9+9IXBUfQIYsZLL6T88+ECbsXqGIgfGtzJJFn6o9CE3/HMrrIGkN5AisxVFTGXWxWci5YA/7PTVWwPOhJff5BRYQDVNgRKqMl/SMttNrrRElsGJjnD1UyQ/5kQmcBxq2PuZI5Zc47u6CILcuoBcM+A=
+Received: by 10.54.96.19 with SMTP id t19mr621017wrb;
+        Tue, 10 May 2005 10:26:39 -0700 (PDT)
+Received: by 10.54.110.5 with HTTP; Tue, 10 May 2005 10:26:39 -0700 (PDT)
+Message-ID: <[email protected]>
+Date: Tue, 10 May 2005 11:26:39 -0600
+From: Test Tester <[email protected]>
+Reply-To: Test Tester <[email protected]>
+To: [email protected], [email protected]
+Subject: Another PDF with 🎉 Unicode chars in it 🍿
+Mime-Version: 1.0
+Content-Type: multipart/mixed; 
+	boundary="----=_Part_2192_32400445.1115745999735"
+X-Virus-Scanned: amavisd-new at textdrive.com
+
+------=_Part_2192_32400445.1115745999735
+Content-Type: text/plain; charset=ISO-8859-1
+Content-Transfer-Encoding: quoted-printable
+Content-Disposition: inline
+
+Just attaching another PDF, here, to see what the message looks like,
+and to see if I can figure out what is going wrong here.    Â
+
+------=_Part_2192_32400445.1115745999735
+Content-Type: application/pdf; name="broken.pdf"
+Content-Transfer-Encoding: base64
+Content-Disposition: attachment; filename="broken.pdf"
+
+JVBERi0xLjQNCiXk9tzfDQoxIDAgb2JqDQo8PCAvTGVuZ3RoIDIgMCBSDQogICAvRmlsdGVyIC9G
+bGF0ZURlY29kZQ0KPj4NCnN0cmVhbQ0KeJy9Wt2KJbkNvm/od6jrhZxYln9hWEh2p+8HBvICySaE
+ycLuTV4/1ifJ9qnq09NpSBimu76yLUuy/qzqcPz7+em3Ixx/CDc6CsXxs3b5+fvfjr/8cPz6/BRu
+rbfAx/n3739/fuJylJ5u5fjX81OuDr4deK4Bz3z/aDP+8fz0yw8g0Ofq7ktr1Mn+u28rvhy/jVeD
+QSa+9YNKHP/pxjvDNfVAx/m3MFz54FhvTbaseaxiDoN2LeMVMw+yA7RbHSCDzxZuaYB2E1Yay7QU
+x89vz0+tyFDKMlAHK5yqLmnjF+c4RjEiQIUeKwblXMe+AsZjN1J5yGQL5DHpDHksurM81rF6PKab
+gK6zAarIDzIiUY23rJsN9iorAE816aIu6lsgAdQFsuhhkHOUFgVjp2GjMqSewITXNQ27jrMeamkg
+1rPI3iLWG2CIaSBB+V1245YVRICGbbpYKHc2USFDl6M09acQVQYhlwIrkBNLISvXhGlF1wi5FHCw
+wxZkoGNJlVeJCEsqKA+3YAV5AMb6KkeaqEJQmFKKQU8T1pRi2ihE1Y4CDrqoYFFXYjJJOatsyzuI
+8SIlykuxKTMibWK8H1PgEvqYgs4GmQSrEjJAalgGirIhik+p4ZQN9E3ETFPAHE1b8pp1l/0Rc1gl
+fQs0ABWvyoZZzU8VnPXwVVcO9BEsyjEJaO6eBoZRyKGlrKoYoOygA8BGIzgwN3RQ15ouigG5idZQ
+fx2U4Db2CqiLO0WHAZoylGiCAqhniNQjFjQPSkmjwfNTgQ6M1Ih+eWo36wFmjIxDJZiGUBiWsAyR
+xX3EekGOizkGI96Ol9zVZTAivikURhRsHh2E3JhWMpSTZCnnonrLhMCodgrNcgo4uyJUJc6qnVss
+nrGd1Ptr0YwisCOYyIbUwVjV4xBUNLbguSO2YHujonAMJkMdSI7bIw91Akq2AUlMUWGFTMAOamjU
+OvZQCxIkY2pCpMFo/IwLdVLHs6nddwTRrgoVbvLU9eB0G4EMndV0TNoxHbt3JBWwK6hhv3iHfDtF
+yokB302IpEBTnWICde4uYc/1khDbSIkQopO6lcqamGBu1OSE3N5IPSsZX00CkSHRiiyx6HQIShsS
+HSVNswdVsaOUSAWq9aYhDtGDaoG5a3lBGkYt/lFlBFt1UqrYnzVtUpUQnLiZeouKgf1KhRBViRRk
+ExepJCzTwEmFDalIRbLEGtw0gfpESOpIAF/NnpPzcVCG86s0g2DuSyd41uhNGbEgaSrWEXORErbw
+------=_Part_2192_32400445.1115745999735--
+

spec/fixtures/emails/attachment_emails/attachment_pdf_non_ascii_lf.eml

@@ -0,0 +1,70 @@
+From [email protected] Tue May 10 11:28:07 2005
+Return-Path: <[email protected]>
+X-Original-To: [email protected]
+Delivered-To: [email protected]
+Received: from localhost (localhost [127.0.0.1])
+	by xxx.xxxxx.com (Postfix) with ESMTP id 50FD3A96F
+	for <[email protected]>; Tue, 10 May 2005 17:26:50 +0000 (GMT)
+Received: from xxx.xxxxx.com ([127.0.0.1])
+ by localhost (xxx.xxxxx.com [127.0.0.1]) (amavisd-new, port 10024)
+ with LMTP id 70060-03 for <[email protected]>;
+ Tue, 10 May 2005 17:26:49 +0000 (GMT)
+Received: from xxx.xxxxx.com (xxx.xxxxx.com [69.36.39.150])
+	by xxx.xxxxx.com (Postfix) with ESMTP id 8B957A94B
+	for <[email protected]>; Tue, 10 May 2005 17:26:48 +0000 (GMT)
+Received: from xxx.xxxxx.com (xxx.xxxxx.com [64.233.184.203])
+	by xxx.xxxxx.com (Postfix) with ESMTP id 9972514824C
+	for <[email protected]>; Tue, 10 May 2005 12:26:40 -0500 (CDT)
+Received: by xxx.xxxxx.com with SMTP id 68so1694448wri
+        for <[email protected]>; Tue, 10 May 2005 10:26:40 -0700 (PDT)
+DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
+        s=beta; d=xxxxx.com;
+        h=received:message-id:date:from:reply-to:to:subject:mime-version:content-type;
+        b=g8ZO5ttS6GPEMAz9WxrRk9+9IXBUfQIYsZLL6T88+ECbsXqGIgfGtzJJFn6o9CE3/HMrrIGkN5AisxVFTGXWxWci5YA/7PTVWwPOhJff5BRYQDVNgRKqMl/SMttNrrRElsGJjnD1UyQ/5kQmcBxq2PuZI5Zc47u6CILcuoBcM+A=
+Received: by 10.54.96.19 with SMTP id t19mr621017wrb;
+        Tue, 10 May 2005 10:26:39 -0700 (PDT)
+Received: by 10.54.110.5 with HTTP; Tue, 10 May 2005 10:26:39 -0700 (PDT)
+Message-ID: <[email protected]>
+Date: Tue, 10 May 2005 11:26:39 -0600
+From: Test Tester <[email protected]>
+Reply-To: Test Tester <[email protected]>
+To: [email protected], [email protected]
+Subject: Another PDF with 🎉 Unicode chars in it 🍿
+Mime-Version: 1.0
+Content-Type: multipart/mixed; 
+	boundary="----=_Part_2192_32400445.1115745999735"
+X-Virus-Scanned: amavisd-new at textdrive.com
+
+------=_Part_2192_32400445.1115745999735
+Content-Type: text/plain; charset=ISO-8859-1
+Content-Transfer-Encoding: quoted-printable
+Content-Disposition: inline
+
+Just attaching another PDF, here, to see what the message looks like,
+and to see if I can figure out what is going wrong here.    Â
+
+------=_Part_2192_32400445.1115745999735
+Content-Type: application/pdf; name="broken.pdf"
+Content-Transfer-Encoding: base64
+Content-Disposition: attachment; filename="broken.pdf"
+
+JVBERi0xLjQNCiXk9tzfDQoxIDAgb2JqDQo8PCAvTGVuZ3RoIDIgMCBSDQogICAvRmlsdGVyIC9G
+bGF0ZURlY29kZQ0KPj4NCnN0cmVhbQ0KeJy9Wt2KJbkNvm/od6jrhZxYln9hWEh2p+8HBvICySaE
+ycLuTV4/1ifJ9qnq09NpSBimu76yLUuy/qzqcPz7+em3Ixx/CDc6CsXxs3b5+fvfjr/8cPz6/BRu
+rbfAx/n3739/fuJylJ5u5fjX81OuDr4deK4Bz3z/aDP+8fz0yw8g0Ofq7ktr1Mn+u28rvhy/jVeD
+QSa+9YNKHP/pxjvDNfVAx/m3MFz54FhvTbaseaxiDoN2LeMVMw+yA7RbHSCDzxZuaYB2E1Yay7QU
+x89vz0+tyFDKMlAHK5yqLmnjF+c4RjEiQIUeKwblXMe+AsZjN1J5yGQL5DHpDHksurM81rF6PKab
+gK6zAarIDzIiUY23rJsN9iorAE816aIu6lsgAdQFsuhhkHOUFgVjp2GjMqSewITXNQ27jrMeamkg
+1rPI3iLWG2CIaSBB+V1245YVRICGbbpYKHc2USFDl6M09acQVQYhlwIrkBNLISvXhGlF1wi5FHCw
+wxZkoGNJlVeJCEsqKA+3YAV5AMb6KkeaqEJQmFKKQU8T1pRi2ihE1Y4CDrqoYFFXYjJJOatsyzuI
+8SIlykuxKTMibWK8H1PgEvqYgs4GmQSrEjJAalgGirIhik+p4ZQN9E3ETFPAHE1b8pp1l/0Rc1gl
+fQs0ABWvyoZZzU8VnPXwVVcO9BEsyjEJaO6eBoZRyKGlrKoYoOygA8BGIzgwN3RQ15ouigG5idZQ
+fx2U4Db2CqiLO0WHAZoylGiCAqhniNQjFjQPSkmjwfNTgQ6M1Ih+eWo36wFmjIxDJZiGUBiWsAyR
+xX3EekGOizkGI96Ol9zVZTAivikURhRsHh2E3JhWMpSTZCnnonrLhMCodgrNcgo4uyJUJc6qnVss
+nrGd1Ptr0YwisCOYyIbUwVjV4xBUNLbguSO2YHujonAMJkMdSI7bIw91Akq2AUlMUWGFTMAOamjU
+OvZQCxIkY2pCpMFo/IwLdVLHs6nddwTRrgoVbvLU9eB0G4EMndV0TNoxHbt3JBWwK6hhv3iHfDtF
+yokB302IpEBTnWICde4uYc/1khDbSIkQopO6lcqamGBu1OSE3N5IPSsZX00CkSHRiiyx6HQIShsS
+HSVNswdVsaOUSAWq9aYhDtGDaoG5a3lBGkYt/lFlBFt1UqrYnzVtUpUQnLiZeouKgf1KhRBViRRk
+ExepJCzTwEmFDalIRbLEGtw0gfpESOpIAF/NnpPzcVCG86s0g2DuSyd41uhNGbEgaSrWEXORErbw
+------=_Part_2192_32400445.1115745999735--
+

spec/mail/attachments_list_spec.rb

@@ -234,12 +234,23 @@ def check_decoded(actual, expected)
       expect(mail.attachments[0].decoded.length).to eq 1026
     end
 
+    it "should decode an attachment and non-ascii text" do
+      mail = read_fixture('emails/attachment_emails/attachment_pdf_non_ascii.eml')
+      expect(mail.attachments[0].decoded.length).to eq 1026
+    end
+
     it "should decode an attachment with linefeeds" do
       mail = read_fixture('emails/attachment_emails/attachment_pdf_lf.eml')
       expect(mail.attachments.size).to eq(1)
       expect(mail.attachments[0].decoded.length).to eq 1026
     end
 
+    it "should decode an attachment with linefeeds and non-ascii text" do
+      mail = read_fixture('emails/attachment_emails/attachment_pdf_non_ascii_lf.eml')
+      expect(mail.attachments.size).to eq(1)
+      expect(mail.attachments[0].decoded.length).to eq 1026
+    end
+
     it "should find an attachment that has an encoded name value" do
       mail = read_fixture('emails/attachment_emails/attachment_with_encoded_name.eml')
       expect(mail.attachments.length).to eq 1