email_reply_trimmer 0.1.8 → 0.1.13

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: 402d5791ea53b316a7ea2d7a972621060fbb632c
-  data.tar.gz: c8c406d99a009bbf32b9ebef739cc46eaa361cfa
+SHA256:
+  metadata.gz: c094ba472eabd06ed02276c4ec9e42cb32ccb873871e46e55c46cd4640425470
+  data.tar.gz: 968cee2c9b3de46ef86814dd84ee8449c400e9c5fa8d61363cbb90064204fadf
 SHA512:
-  metadata.gz: 530b38bed9536b989a66146001f7514b54b5b905c6eedce1032d82267cc72f0072bac8483d5ac126d41848ce9963014a5330cfd89f9db4aea62138e7ab5bfc6c
-  data.tar.gz: 721fd78143d81634eefa49a6df73b6565f1a5ea9fae598fc57e67509f988efb7b47de654ab4d16f3b5513799d80b686bd50a5721eb8b383630b1024173b02961
+  metadata.gz: 6c71cad161a2aaa27e813ff63f91b1e10359d73ffc95d0d01df0374dc8cc9a2ba506985cd4e33d12e63b3f41cd9e6f66bbe42293feea3f6e968c7d90e79a88de
+  data.tar.gz: e7e3661094d8b1183f572b7e847cac62a71215a341d12b627e0b2d5cfedfdad161f858b8aba3d97463d5ffda5919211da2b268dbd75e7edc48a95811d1e5b481

data/email_reply_trimmer.gemspec

@@ -20,5 +20,5 @@ Gem::Specification.new do |s|
 
   s.add_development_dependency 'rake', '~> 12'
   s.add_development_dependency 'minitest', '~> 5'
-  s.add_development_dependency 'rubocop', '~> 0'
+  s.add_development_dependency 'rubocop', '~> 0.52.1'
 end

data/lib/email_reply_trimmer.rb

@@ -6,7 +6,7 @@ require_relative "email_reply_trimmer/email_header_matcher"
 require_relative "email_reply_trimmer/quote_matcher"
 
 class EmailReplyTrimmer
-  VERSION = "0.1.8"
+  VERSION = "0.1.13"
 
   DELIMITER    = "d"
   EMBEDDED     = "b"
@@ -54,7 +54,7 @@ class EmailReplyTrimmer
     end
 
     # when the reply is at the end of the email
-    if pattern =~ /^(b[^t]+)*b[bqeh]+t[et]*$/
+    if is_reply_at_end?(pattern)
       index = pattern =~ /t[et]*$/
       pattern = ""
       lines = lines[index..-1]
@@ -78,8 +78,8 @@ class EmailReplyTrimmer
 
     # if there is some text before a huge quote ending the email,
     # then remove the quote
-    if pattern =~ /te*[qbe]+$/
-      index = pattern =~ /te*[qbe]+$/
+    if pattern =~ /t?e*[qbe]+$/
+      index = pattern =~ /t?e*[qbe]+$/
       pattern = pattern[0..index]
       lines = lines[0..index]
     end
@@ -144,6 +144,12 @@ class EmailReplyTrimmer
 
     if index = pattern =~ /(?:h[eqd]*?){3,}[tq]/
       embedded = lines[index..-1].join("\n").strip
+    elsif index = pattern =~ /b(?:[eqd]*){3,}[tq]/
+      # Exception for email clients (macOS / iOS) which embed fwd emails in quotes.
+      embedded = lines[index + 1..-1].map { |l| l.gsub(/^>\s*/, '') }.join("\n").strip
+    end
+
+    if index
       before = lines[0...(pattern[0...index] =~ /e*(b[eqd]*|b*[ed]*)$/)].join("\n").strip
       return [embedded, before]
     end
@@ -151,59 +157,62 @@ class EmailReplyTrimmer
 
   private
 
-    def self.preprocess!(text)
-      # normalize line endings
-      text.gsub!("\r\n", "\n")
-
-      # remove PGP markers
-      text.gsub!(/\A-----BEGIN PGP SIGNED MESSAGE-----\n(?:Hash: \w+)?\s+/i, "")
-      text.gsub!(/^-----BEGIN PGP SIGNATURE-----$[\s\S]+^-----END PGP SIGNATURE-----/, "")
-
-      # remove unsubscribe links
-      text.gsub!(/^Unsubscribe: .+@.+(\n.+http:.+)?\s*\z/i, "")
-
-      # remove alias-style quotes marker
-      text.gsub!(/^.*>{5} "[^"\n]+" == .+ writes:/, "")
-
-      # change enclosed-style quotes format
-      text.gsub!(/^>>> ?(.+) ?>>>$\n([\s\S]+?)\n^<<< ?\1 ?<<<$/) { $2.gsub(/^/, "> ") }
-      text.gsub!(/^>{4,}[[:blank:]]*$\n([\s\S]+?)\n^<{4,}[[:blank:]]*$/) { $1.gsub(/^/, "> ") }
-
-      # fix all quotes formats
-      text.gsub!(/^((?:[[:blank:]]*[[:alpha:]]*[>|])+)/) { $1.gsub(/([[:alpha:]]+>|\|)/, ">") }
-
-      # fix embedded email markers that might span over multiple lines
-      (
-        EmbeddedEmailMatcher::ON_DATE_SOMEONE_WROTE_REGEXES +
-        EmbeddedEmailMatcher::SOMEONE_WROTE_ON_DATE_REGEXES +
-        EmbeddedEmailMatcher::DATE_SOMEONE_WROTE_REGEXES +
-        [EmbeddedEmailMatcher::DATE_SOMEONE_EMAIL_REGEX]
-      ).each do |r|
-        text.gsub!(r) do |m|
-          m.count("\n") > 4 ? m : m.gsub(/\n+[[:space:]]*/, " ")
-        end
-      end
+  def self.preprocess!(text)
+    # normalize line endings
+    text.gsub!("\r\n", "\n")
+
+    # remove PGP markers
+    text.gsub!(/\A-----BEGIN PGP SIGNED MESSAGE-----\n(?:Hash: \w+)?\s+/i, "")
+    text.gsub!(/^-----BEGIN PGP SIGNATURE-----$[\s\S]+^-----END PGP SIGNATURE-----/, "")
+
+    # remove unsubscribe links
+    text.gsub!(/^Unsubscribe: .+@.+(\n.+http:.+)?\s*\z/i, "")
+
+    # remove alias-style quotes marker
+    text.gsub!(/^.*>{5} "[^"\n]+" == .+ writes:/, "")
+
+    # change enclosed-style quotes format
+    text.gsub!(/^>>> ?(.+) ?>>>$\n([\s\S]+?)\n^<<< ?\1 ?<<<$/) { $2.gsub(/^/, "> ") }
+    text.gsub!(/^>{4,}[[:blank:]]*$\n([\s\S]+?)\n^<{4,}[[:blank:]]*$/) { $1.gsub(/^/, "> ") }
+
+    # fix all quotes formats
+    text.gsub!(/^((?:[[:blank:]]*[[:alpha:]]*[>|])+)/) { $1.gsub(/([[:alpha:]]+>|\|)/, ">") }
 
-      # remove leading/trailing whitespaces
-      text.strip!
+    # fix embedded email markers that might span over multiple lines
+    (
+      EmbeddedEmailMatcher::ON_DATE_SOMEONE_WROTE_REGEXES +
+      EmbeddedEmailMatcher::SOMEONE_WROTE_ON_DATE_REGEXES +
+      EmbeddedEmailMatcher::DATE_SOMEONE_WROTE_REGEXES +
+      [EmbeddedEmailMatcher::DATE_SOMEONE_EMAIL_REGEX]
+    ).each do |r|
+      text.gsub!(r) do |m|
+        m.count("\n") > 4 ? m : m.gsub(/\n+[[:space:]]*/, " ")
+      end
     end
 
-    def self.compute_elided(text, lines)
-      elided = []
+    # remove leading/trailing whitespaces
+    text.strip!
+  end
+
+  def self.compute_elided(text, lines)
+    elided = []
 
-      t = 0
-      l = 0
+    t = 0
+    l = 0
 
-      while t < text.size
-        while l < lines.size && text[t] == lines[l]
-          t += 1
-          l += 1
-        end
-        elided << text[t]
+    while t < text.size
+      while l < lines.size && text[t] == lines[l]
         t += 1
+        l += 1
       end
-
-      elided.join("\n").strip
+      elided << text[t]
+      t += 1
     end
 
+    elided.join("\n").strip
+  end
+
+  def self.is_reply_at_end?(pattern)
+    pattern =~ /^b[^t]+t[et]*$/
+  end
 end

data/lib/email_reply_trimmer/embedded_email_matcher.rb

@@ -54,6 +54,8 @@ class EmbeddedEmailMatcher
   end
 
   # суббота, 14 марта 2015 г. пользователь etewiah написал:
+  # 23 mar 2017 21:25 "Neil Lalonde" <meta@discoursemail.com> napisał(a):
+  # 30 серп. 2016 р. 20:45 "Arpit" no-reply@example.com пише:
   DATE_SOMEONE_WROTE_MARKERS = [
     # Russian
     ["пользователь", "написал"],
@@ -65,8 +67,8 @@ class EmbeddedEmailMatcher
 
   DATE_SOMEONE_WROTE_REGEXES = DATE_SOMEONE_WROTE_MARKERS.map do |user, wrote|
     user.size == 0 ?
-      /^.*\d{4}.*?(?:(?!#{wrote})[\s\S])*#{wrote}:/ :
-      /^.*\d{4}.*?#{user}.*?(?:(?!#{wrote})[\s\S])*#{wrote}:/
+      /\d{4}.{1,80}\n?.{0,80}?#{wrote}:/ :
+      /\d{4}.{1,80}#{user}.{0,80}\n?.{0,80}?#{wrote}:/
   end
 
   # Max Mustermann <try_discourse@discoursemail.com> schrieb am Fr., 28. Apr. 2017 um 11:53 Uhr:
@@ -84,7 +86,7 @@ class EmbeddedEmailMatcher
   # 2013/10/2 camilohollanda <info@discourse.org>
   # вт, 5 янв. 2016 г. в 23:39, Erlend Sogge Heggen <info@discourse.org>:
   # ср, 1 апр. 2015, 18:29, Denis Didkovsky <info@discourse.org>:
-  DATE_SOMEONE_EMAIL_REGEX = /^.*\d{4}.+\s?<[^@<>]+@[^@<>.]+\.[^@<>]+>:?$/
+  DATE_SOMEONE_EMAIL_REGEX = /\d{4}.{1,80}\s?<[^@<>]+@[^@<>.]+\.[^@<>]+>:?$/
 
   # codinghorror via Discourse Meta wrote:
   # codinghorror via Discourse Meta <info@discourse.org> schrieb:

data/lib/email_reply_trimmer/signature_matcher.rb

@@ -20,7 +20,7 @@ class SignatureMatcher
     /^[[:blank:]]*[[:word:]]+ from mobile/i,
     /^[[:blank:]]*[\(<]*Sent (from|via|with|by) .+[\)>]*/i,
     /^[[:blank:]]*From my .{1,20}/i,
-    /^[[:blank:]]*Get Outlook for iOS/i,
+    /^[[:blank:]]*Get Outlook for /i,
     # French
     /^[[:blank:]]*Envoyé depuis (mon|Yahoo Mail)/i,
     # German
@@ -34,10 +34,17 @@ class SignatureMatcher
     /^[[:blank:]]*Enviado do meu /i,
     # Spanish
     /^[[:blank:]]*Enviado desde mi /i,
+    # Dutch
+    /^[[:blank:]]*Verzonden met /i,
+    /^[[:blank:]]*Verstuurd vanaf mijn /i,
+    # Swedish
+    /^[[:blank:]]*från min /i,
   ]
 
   def self.match?(line)
-    SIGNATURE_REGEXES.any? { |r| line =~ r }
+    # remove any markdown links
+    stripped = line.gsub(/\[([^\]]+)\]\([^\)]+\)/) { $1 }
+    SIGNATURE_REGEXES.any? { |r| stripped =~ r }
   end
 
 end

data/test/before/forwarded_apple.txt

@@ -0,0 +1 @@
+This email has been forwarded from Apple Mail.

data/test/before/forwarded_gmail.txt

@@ -0,0 +1 @@
+This email has been forwarded from Gmail.

data/test/elided/embedded_email_polish_2.txt

@@ -0,0 +1,7 @@
+23 mar 2017 21:25 "Neil Lalonde" <meta@discoursemail.com> napisał(a):
+
+> neil <https://meta.discourse.org/users/neil> Neil Lalonde
+> <https://meta.discourse.org/users/neil> Team
+> March 23
+>
+> I have made required changes. Enjoy!

data/test/elided/embedded_email_ukrainian.txt

@@ -0,0 +1,17 @@
+30 серп. 2016 р. 20:45 "Arpit" <no-reply@example.com> пише:
+
+> meg <http://forum.example.com/users/foo> Foo
+> <http://forum.example.com/users/foo>
+> Август 30
+>
+> Когда будет точное название, напишите в личку и мы поменяем название.
+> ------------------------------
+>
+> Visit Message
+> <http://forum.example.com/t/slug/24>
+> or reply to this email to respond.
+>
+> To unsubscribe from these emails, click here
+> <http://forum.example.com/email/unsubscribe/0123456789>
+> .
+>

data/test/elided/forwarded_apple.txt

@@ -0,0 +1,15 @@
+> Begin forwarded message:
+>
+> From: Foo Discourse <foo@discourse.org>
+> Subject: This email is going to be forwarded
+> Date: 01 January 2019 at 12:00:00 EET
+> To: bar@discourse.org
+>
+> Lorem ipsum dolor sit amet, consectetur adipiscing elit. Morbi auctor
+> lacus in metus vehicula, non dictum mi consequat. Nunc eget quam sit
+> amet dolor scelerisque cursus. Etiam laoreet lectus a rutrum
+> sollicitudin. Nullam sit amet lorem id ante lobortis fringilla. Duis
+> dictum tempus sem et ultrices. Sed ac lectus ante. Vivamus interdum,
+> risus non tincidunt accumsan, purus mi accumsan dolor, vitae euismod
+> turpis orci ut est. Integer ut ante eu velit dignissim ullamcorper non
+> id nibh. Nulla auctor magna erat, et posuere sapien gravida convallis.

data/test/elided/forwarded_gmail.txt

@@ -0,0 +1,15 @@
+---------- Forwarded message ---------
+From: Foo Discourse <foo@discourse.org>
+Date: Tue, Jan 01, 2019 at 12:00 PM
+Subject: This email is going to be forwarded
+To: <bar@discourse.org>
+
+
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Morbi auctor
+lacus in metus vehicula, non dictum mi consequat. Nunc eget quam sit
+amet dolor scelerisque cursus. Etiam laoreet lectus a rutrum
+sollicitudin. Nullam sit amet lorem id ante lobortis fringilla. Duis
+dictum tempus sem et ultrices. Sed ac lectus ante. Vivamus interdum,
+risus non tincidunt accumsan, purus mi accumsan dolor, vitae euismod
+turpis orci ut est. Integer ut ante eu velit dignissim ullamcorper non
+id nibh. Nulla auctor magna erat, et posuere sapien gravida convallis.

data/test/elided/signatures.txt

@@ -24,5 +24,8 @@ Sent from my Windows 8 PC
 (sent from a phone)
 (Sent from mobile device)
 從我的 iPhone 傳送
+Sent from [mail](https://go.microsoft.com/fwlink/?LinkId=550986) for windows 10
+Verzonden met BlackBerry Work
 
 Get Outlook for iOShttps://aka.ms/o0ukef
+Get [Outlook for Android](https://aka.ms/ghei36)

data/test/elided/spam_1.txt

@@ -0,0 +1,75 @@
+> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>
+>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>
+>>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>>
+>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>
+>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>
+>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>
+>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>
+>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>
+>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>
+>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>
+>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>
+>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>
+>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>
+>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:

data/test/elided/spam_2.txt

@@ -0,0 +1,152 @@
+> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>
+>> Hello world!
+>>
+>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>
+>>> Hello world!
+>>>
+>>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>>
+>>>> Hello world!
+>>>>
+>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>
+>>>>> Hello world!
+>>>>>
+>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>
+>>>>>> Hello world!
+>>>>>>
+>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>
+>>>>>>> Hello world!
+>>>>>>>
+>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>
+>>>>>>>> Hello world!
+>>>>>>>>
+>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>
+>>>>>>>>> Hello world!
+>>>>>>>>>
+>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>
+>>>>>>>>>> Hello world!
+>>>>>>>>>>
+>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>
+>>>>>>>>>>> Hello world!
+>>>>>>>>>>>
+>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>
+>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>
+>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>
+>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>
+>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>
+>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>
+>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

data/test/emails/embedded_email_polish_2.txt

@@ -0,0 +1,11 @@
+Hi Neil,
+Thanks for prompt response!
+
+
+23 mar 2017 21:25 "Neil Lalonde" <meta@discoursemail.com> napisał(a):
+
+> neil <https://meta.discourse.org/users/neil> Neil Lalonde
+> <https://meta.discourse.org/users/neil> Team
+> March 23
+>
+> I have made required changes. Enjoy!

data/test/emails/embedded_email_ukrainian.txt

@@ -0,0 +1,19 @@
+Спасибо еще раз.
+
+30 серп. 2016 р. 20:45 "Arpit" <no-reply@example.com> пише:
+
+> meg <http://forum.example.com/users/foo> Foo
+> <http://forum.example.com/users/foo>
+> Август 30
+>
+> Когда будет точное название, напишите в личку и мы поменяем название.
+> ------------------------------
+>
+> Visit Message
+> <http://forum.example.com/t/slug/24>
+> or reply to this email to respond.
+>
+> To unsubscribe from these emails, click here
+> <http://forum.example.com/email/unsubscribe/0123456789>
+> .
+>

data/test/emails/forwarded_apple.txt

@@ -0,0 +1,17 @@
+This email has been forwarded from Apple Mail.
+
+> Begin forwarded message:
+>
+> From: Foo Discourse <foo@discourse.org>
+> Subject: This email is going to be forwarded
+> Date: 01 January 2019 at 12:00:00 EET
+> To: bar@discourse.org
+>
+> Lorem ipsum dolor sit amet, consectetur adipiscing elit. Morbi auctor
+> lacus in metus vehicula, non dictum mi consequat. Nunc eget quam sit
+> amet dolor scelerisque cursus. Etiam laoreet lectus a rutrum
+> sollicitudin. Nullam sit amet lorem id ante lobortis fringilla. Duis
+> dictum tempus sem et ultrices. Sed ac lectus ante. Vivamus interdum,
+> risus non tincidunt accumsan, purus mi accumsan dolor, vitae euismod
+> turpis orci ut est. Integer ut ante eu velit dignissim ullamcorper non
+> id nibh. Nulla auctor magna erat, et posuere sapien gravida convallis.

data/test/emails/forwarded_gmail.txt

@@ -0,0 +1,17 @@
+This email has been forwarded from Gmail.
+
+---------- Forwarded message ---------
+From: Foo Discourse <foo@discourse.org>
+Date: Tue, Jan 01, 2019 at 12:00 PM
+Subject: This email is going to be forwarded
+To: <bar@discourse.org>
+
+
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Morbi auctor
+lacus in metus vehicula, non dictum mi consequat. Nunc eget quam sit
+amet dolor scelerisque cursus. Etiam laoreet lectus a rutrum
+sollicitudin. Nullam sit amet lorem id ante lobortis fringilla. Duis
+dictum tempus sem et ultrices. Sed ac lectus ante. Vivamus interdum,
+risus non tincidunt accumsan, purus mi accumsan dolor, vitae euismod
+turpis orci ut est. Integer ut ante eu velit dignissim ullamcorper non
+id nibh. Nulla auctor magna erat, et posuere sapien gravida convallis.

data/test/emails/signatures.txt

@@ -26,5 +26,8 @@ Sent from my Windows 8 PC
 (sent from a phone)
 (Sent from mobile device)
 從我的 iPhone 傳送
+Sent from [mail](https://go.microsoft.com/fwlink/?LinkId=550986) for windows 10
+Verzonden met BlackBerry Work
 
 Get Outlook for iOShttps://aka.ms/o0ukef
+Get [Outlook for Android](https://aka.ms/ghei36)

data/test/emails/spam_1.txt

@@ -0,0 +1,75 @@
+> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>
+>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>
+>>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>>
+>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>
+>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>
+>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>
+>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>
+>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>
+>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>
+>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>
+>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>
+>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>
+>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>
+>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:

data/test/emails/spam_2.txt

@@ -0,0 +1,174 @@
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec sit amet dictum
+lectus, at efficitur neque. Class aptent taciti sociosqu ad litora torquent per
+conubia nostra, per inceptos himenaeos. Aenean dictum pulvinar tortor non
+ultricies. Maecenas cursus lobortis dolor quis feugiat. Integer efficitur,
+mauris sit amet lobortis consectetur, urna augue feugiat mi, et tempor dui
+sapien vehicula dolor. Aenean sit amet ultrices tortor. Praesent vel porttitor
+risus, eu fringilla nibh. Phasellus sollicitudin elit id tortor dapibus, at
+euismod odio sodales. Ut est ex, pharetra malesuada dui maximus, ullamcorper
+fringilla turpis. Nullam gravida, elit blandit commodo blandit, velit sem
+ornare odio, eu ultricies arcu ex sit amet libero. Vestibulum eu fringilla
+ante, nec blandit nunc.
+
+Nullam nec viverra ipsum. Nullam semper eu urna id facilisis. Nam blandit
+dictum eros, sit amet auctor dui fermentum a. Quisque finibus magna nisi, ac
+hendrerit purus lacinia ut. Donec maximus, enim id tincidunt vulputate, libero
+ex tempor elit, vitae sagittis libero leo sed leo. Quisque non gravida ex,
+efficitur interdum neque. Interdum et malesuada fames ac ante ipsum primis in
+faucibus. Phasellus rhoncus nisi vitae turpis bibendum luctus. Duis placerat
+aliquam lectus, pellentesque commodo dolor hendrerit quis. Suspendisse ipsum
+sapien, interdum ac felis ac, mattis varius neque. Maecenas nec risus sit amet
+ex aliquam vestibulum. Maecenas a turpis et turpis volutpat congue.
+
+> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>
+>> Hello world!
+>>
+>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>
+>>> Hello world!
+>>>
+>>> On Thu, Apr 30, 2020 at 10:07 AM, "someone" <someone@example.com> wrote:
+>>>
+>>>> Hello world!
+>>>>
+>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>
+>>>>> Hello world!
+>>>>>
+>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>
+>>>>>> Hello world!
+>>>>>>
+>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>
+>>>>>>> Hello world!
+>>>>>>>
+>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>
+>>>>>>>> Hello world!
+>>>>>>>>
+>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>
+>>>>>>>>> Hello world!
+>>>>>>>>>
+>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>
+>>>>>>>>>> Hello world!
+>>>>>>>>>>
+>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>
+>>>>>>>>>>> Hello world!
+>>>>>>>>>>>
+>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>
+>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>
+>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>
+>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>
+>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>
+>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>
+>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Apr 30, 2020 at 10:06 AM, "someone" <someone@example.com> wrote:
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> Hello world!
+>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

data/test/embedded/forwarded_apple.txt

@@ -0,0 +1,13 @@
+From: Foo Discourse <foo@discourse.org>
+Subject: This email is going to be forwarded
+Date: 01 January 2019 at 12:00:00 EET
+To: bar@discourse.org
+
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Morbi auctor
+lacus in metus vehicula, non dictum mi consequat. Nunc eget quam sit
+amet dolor scelerisque cursus. Etiam laoreet lectus a rutrum
+sollicitudin. Nullam sit amet lorem id ante lobortis fringilla. Duis
+dictum tempus sem et ultrices. Sed ac lectus ante. Vivamus interdum,
+risus non tincidunt accumsan, purus mi accumsan dolor, vitae euismod
+turpis orci ut est. Integer ut ante eu velit dignissim ullamcorper non
+id nibh. Nulla auctor magna erat, et posuere sapien gravida convallis.

data/test/embedded/forwarded_gmail.txt

@@ -0,0 +1,14 @@
+From: Foo Discourse <foo@discourse.org>
+Date: Tue, Jan 01, 2019 at 12:00 PM
+Subject: This email is going to be forwarded
+To: <bar@discourse.org>
+
+
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Morbi auctor
+lacus in metus vehicula, non dictum mi consequat. Nunc eget quam sit
+amet dolor scelerisque cursus. Etiam laoreet lectus a rutrum
+sollicitudin. Nullam sit amet lorem id ante lobortis fringilla. Duis
+dictum tempus sem et ultrices. Sed ac lectus ante. Vivamus interdum,
+risus non tincidunt accumsan, purus mi accumsan dolor, vitae euismod
+turpis orci ut est. Integer ut ante eu velit dignissim ullamcorper non
+id nibh. Nulla auctor magna erat, et posuere sapien gravida convallis.

data/test/test_email_reply_trimmer.rb

@@ -19,6 +19,10 @@ class TestEmailReplyTrimmer < Minitest::Test
     assert_match(/\r\n/, File.read("test/emails/normalize_line_endings.txt"))
   end
 
+  def test_is_reply_at_end?
+    Timeout::timeout(1) { assert_nil(EmailReplyTrimmer.send(:is_reply_at_end?, "bq" * 100)) }
+  end
+
   EMAILS.each do |filename|
     name = File.basename(filename, ".txt")
     define_method("test_#{name}") do
@@ -30,7 +34,7 @@ class TestEmailReplyTrimmer < Minitest::Test
   EMBEDDED_EMAILS = %w{
     email_headers_1 email_headers_2 email_headers_3 email_headers_4
     embedded_email_10 embedded_email_german_3 embedded_email_spanish_2
-    forwarded_message
+    forwarded_message forwarded_gmail forwarded_apple
   }
 
   EMBEDDED_EMAILS.each do |name|

data/test/trimmed/embedded_email_polish_2.txt

@@ -0,0 +1,2 @@
+Hi Neil,
+Thanks for prompt response!

data/test/trimmed/embedded_email_ukrainian.txt

@@ -0,0 +1 @@
+Спасибо еще раз.

data/test/trimmed/forwarded_apple.txt

@@ -0,0 +1 @@
+This email has been forwarded from Apple Mail.

data/test/trimmed/forwarded_gmail.txt

@@ -0,0 +1 @@
+This email has been forwarded from Gmail.

data/test/trimmed/spam_2.txt

@@ -0,0 +1,21 @@
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec sit amet dictum
+lectus, at efficitur neque. Class aptent taciti sociosqu ad litora torquent per
+conubia nostra, per inceptos himenaeos. Aenean dictum pulvinar tortor non
+ultricies. Maecenas cursus lobortis dolor quis feugiat. Integer efficitur,
+mauris sit amet lobortis consectetur, urna augue feugiat mi, et tempor dui
+sapien vehicula dolor. Aenean sit amet ultrices tortor. Praesent vel porttitor
+risus, eu fringilla nibh. Phasellus sollicitudin elit id tortor dapibus, at
+euismod odio sodales. Ut est ex, pharetra malesuada dui maximus, ullamcorper
+fringilla turpis. Nullam gravida, elit blandit commodo blandit, velit sem
+ornare odio, eu ultricies arcu ex sit amet libero. Vestibulum eu fringilla
+ante, nec blandit nunc.
+
+Nullam nec viverra ipsum. Nullam semper eu urna id facilisis. Nam blandit
+dictum eros, sit amet auctor dui fermentum a. Quisque finibus magna nisi, ac
+hendrerit purus lacinia ut. Donec maximus, enim id tincidunt vulputate, libero
+ex tempor elit, vitae sagittis libero leo sed leo. Quisque non gravida ex,
+efficitur interdum neque. Interdum et malesuada fames ac ante ipsum primis in
+faucibus. Phasellus rhoncus nisi vitae turpis bibendum luctus. Duis placerat
+aliquam lectus, pellentesque commodo dolor hendrerit quis. Suspendisse ipsum
+sapien, interdum ac felis ac, mattis varius neque. Maecenas nec risus sit amet
+ex aliquam vestibulum. Maecenas a turpis et turpis volutpat congue.

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: email_reply_trimmer
 version: !ruby/object:Gem::Version
-  version: 0.1.8
+  version: 0.1.13
 platform: ruby
 authors:
 - Régis Hanol
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-09-14 00:00:00.000000000 Z
+date: 2020-06-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -44,14 +44,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.52.1
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.52.1
 description: EmailReplyTrimmer is a small library to trim replies from plain text
   email.
 email:
@@ -61,7 +61,6 @@ extensions: []
 extra_rdoc_files: []
 files:
 - Gemfile
-- Gemfile.lock
 - LICENSE
 - README.md
 - Rakefile
@@ -80,6 +79,8 @@ files:
 - test/before/embedded_email_10.txt
 - test/before/embedded_email_german_3.txt
 - test/before/embedded_email_spanish_2.txt
+- test/before/forwarded_apple.txt
+- test/before/forwarded_gmail.txt
 - test/before/forwarded_message.txt
 - test/elided/delimiters.txt
 - test/elided/dual_embedded.txt
@@ -121,7 +122,8 @@ files:
 - test/elided/embedded_email_german_6.txt
 - test/elided/embedded_email_italian.txt
 - test/elided/embedded_email_norwegian.txt
-- test/elided/embedded_email_polish.txt
+- test/elided/embedded_email_polish_1.txt
+- test/elided/embedded_email_polish_2.txt
 - test/elided/embedded_email_portuguese.txt
 - test/elided/embedded_email_quote_text.txt
 - test/elided/embedded_email_russian_1.txt
@@ -129,12 +131,17 @@ files:
 - test/elided/embedded_email_spanish_1.txt
 - test/elided/embedded_email_spanish_2.txt
 - test/elided/embedded_email_swedish.txt
+- test/elided/embedded_email_ukrainian.txt
+- test/elided/forwarded_apple.txt
+- test/elided/forwarded_gmail.txt
 - test/elided/forwarded_message.txt
 - test/elided/normalize_line_endings.txt
 - test/elided/quote_and_text.txt
 - test/elided/quote_only.txt
 - test/elided/retains_spaces_and_formatting.txt
 - test/elided/signatures.txt
+- test/elided/spam_1.txt
+- test/elided/spam_2.txt
 - test/elided/strip.txt
 - test/elided/text_only.txt
 - test/elided/usenet.txt
@@ -178,7 +185,8 @@ files:
 - test/emails/embedded_email_german_6.txt
 - test/emails/embedded_email_italian.txt
 - test/emails/embedded_email_norwegian.txt
-- test/emails/embedded_email_polish.txt
+- test/emails/embedded_email_polish_1.txt
+- test/emails/embedded_email_polish_2.txt
 - test/emails/embedded_email_portuguese.txt
 - test/emails/embedded_email_quote_text.txt
 - test/emails/embedded_email_russian_1.txt
@@ -186,12 +194,17 @@ files:
 - test/emails/embedded_email_spanish_1.txt
 - test/emails/embedded_email_spanish_2.txt
 - test/emails/embedded_email_swedish.txt
+- test/emails/embedded_email_ukrainian.txt
+- test/emails/forwarded_apple.txt
+- test/emails/forwarded_gmail.txt
 - test/emails/forwarded_message.txt
 - test/emails/normalize_line_endings.txt
 - test/emails/quote_and_text.txt
 - test/emails/quote_only.txt
 - test/emails/retains_spaces_and_formatting.txt
 - test/emails/signatures.txt
+- test/emails/spam_1.txt
+- test/emails/spam_2.txt
 - test/emails/strip.txt
 - test/emails/text_only.txt
 - test/emails/usenet.txt
@@ -202,6 +215,8 @@ files:
 - test/embedded/embedded_email_10.txt
 - test/embedded/embedded_email_german_3.txt
 - test/embedded/embedded_email_spanish_2.txt
+- test/embedded/forwarded_apple.txt
+- test/embedded/forwarded_gmail.txt
 - test/embedded/forwarded_message.txt
 - test/matchers/does_not_contain_embedded_email.txt
 - test/test_email_matcher.rb
@@ -246,7 +261,8 @@ files:
 - test/trimmed/embedded_email_german_6.txt
 - test/trimmed/embedded_email_italian.txt
 - test/trimmed/embedded_email_norwegian.txt
-- test/trimmed/embedded_email_polish.txt
+- test/trimmed/embedded_email_polish_1.txt
+- test/trimmed/embedded_email_polish_2.txt
 - test/trimmed/embedded_email_portuguese.txt
 - test/trimmed/embedded_email_quote_text.txt
 - test/trimmed/embedded_email_russian_1.txt
@@ -254,12 +270,17 @@ files:
 - test/trimmed/embedded_email_spanish_1.txt
 - test/trimmed/embedded_email_spanish_2.txt
 - test/trimmed/embedded_email_swedish.txt
+- test/trimmed/embedded_email_ukrainian.txt
+- test/trimmed/forwarded_apple.txt
+- test/trimmed/forwarded_gmail.txt
 - test/trimmed/forwarded_message.txt
 - test/trimmed/normalize_line_endings.txt
 - test/trimmed/quote_and_text.txt
 - test/trimmed/quote_only.txt
 - test/trimmed/retains_spaces_and_formatting.txt
 - test/trimmed/signatures.txt
+- test/trimmed/spam_1.txt
+- test/trimmed/spam_2.txt
 - test/trimmed/strip.txt
 - test/trimmed/text_only.txt
 - test/trimmed/usenet.txt
@@ -282,8 +303,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.6.13
+rubygems_version: 3.0.3
 signing_key: 
 specification_version: 4
 summary: Library to trim replies from plain text email.

data/Gemfile.lock

@@ -1,38 +0,0 @@
-PATH
-  remote: .
-  specs:
-    email_reply_trimmer (0.1.7)
-
-GEM
-  remote: https://rubygems.org/
-  specs:
-    ast (2.3.0)
-    minitest (5.10.3)
-    parallel (1.11.2)
-    parser (2.4.0.0)
-      ast (~> 2.2)
-    powerpack (0.1.1)
-    rainbow (2.2.2)
-      rake
-    rake (12.0.0)
-    rubocop (0.49.1)
-      parallel (~> 1.10)
-      parser (>= 2.3.3.1, < 3.0)
-      powerpack (~> 0.1)
-      rainbow (>= 1.99.1, < 3.0)
-      ruby-progressbar (~> 1.7)
-      unicode-display_width (~> 1.0, >= 1.0.1)
-    ruby-progressbar (1.8.1)
-    unicode-display_width (1.3.0)
-
-PLATFORMS
-  ruby
-
-DEPENDENCIES
-  email_reply_trimmer!
-  minitest (~> 5)
-  rake (~> 12)
-  rubocop (~> 0)
-
-BUNDLED WITH
-   1.16.0.pre.2