Handle question mark in encoded text (Backports mikel#1452)

GChuf · Oct 20, 2021 · 873aa38 · 873aa38
1 parent 51c05e0
commit 873aa38
Show file tree

Hide file tree

Showing 3 changed files with 30 additions and 5 deletions.
diff --git a/CHANGELOG.rdoc b/CHANGELOG.rdoc
@@ -4,6 +4,7 @@ Bugs:
 * Fix duplicate dots at beginning of lines. (c960657)
 * Don't alter attachment filenames when Content-Type parsing fails. (nbianca)
 * Fix decoding fields ending in question marks. (jeremy)
+* Fix decoding quoted-printable fields containing question marks. (lewispb)
 
 
 == Version 2.7.1 (2018-10-13)

diff --git a/lib/mail/constants.rb b/lib/mail/constants.rb
@@ -34,8 +34,28 @@ module Constants
     ATOM_UNSAFE   = /[#{Regexp.quote aspecial}#{control}#{sp}]/n
     PHRASE_UNSAFE = /[#{Regexp.quote aspecial}#{control}]/n
     TOKEN_UNSAFE  = /[#{Regexp.quote tspecial}#{control}#{sp}]/n
-    ENCODED_VALUE = /\=\?([^?]+)\?([QB])\?[^?]*?\?+\=/mi
-    FULL_ENCODED_VALUE = /(\=\?[^?]+\?[QB]\?[^?]*?\?+\=)/mi
+
+    ENCODED_VALUE = %r{
+      \=\?     # literal =?
+      ([^?]+)  #
+      \?       # literal ?
+      ([QB])   # either a "Q" or a "B"
+      \?       # literal ?
+      .*?      # lazily match all characters
+      \?\=     # literal ?=
+    }mix # m is multi-line, i is case-insensitive, x is free-spacing
+
+    FULL_ENCODED_VALUE = %r{ # Identical to ENCODED_VALUE but captures the whole rather than components of
+      (
+        \=\?    # literal =?
+        [^?]+   #
+        \?      # literal ?
+        [QB]    # either a "Q" or a "B"
+        \?      # literal ?
+        .*?     # lazily match all characters
+        \?\=    # literal ?=
+      )
+    }mix # m is multi-line, i is case-insensitive, x is free-spacing
 
     EMPTY          = ''
     SPACE          = ' '

diff --git a/spec/mail/encodings_spec.rb b/spec/mail/encodings_spec.rb
@@ -736,6 +736,10 @@
       expect(Mail::Encodings.value_decode(encoded)).to eq expected
     end
 
+    it "should handle a question mark in the text" do
+      encoded = "=?iso-8859-1?Q?Hello World?_-_How are you??="
+      expect(Mail::Encodings.value_decode(encoded)).to eq "Hello World? - How are you?"
+    end
   end
 
   describe "pre encoding non usascii text" do
@@ -939,7 +943,7 @@ def convert(from, to)
     it "splits adjacent encodings into separate parts" do
       convert "A=?iso-2022-jp?B?X=?==?iso-2022-jp?B?Y=?=B", ["A", "=?iso-2022-jp?B?X=?=", "=?iso-2022-jp?B?Y=?=", "B"]
     end
-    
+
     it "splits adjacent encodings without unencoded into separate parts" do
       convert "=?iso-2022-jp?B?X=?==?iso-2022-jp?B?Y=?=", ["=?iso-2022-jp?B?X=?=", "=?iso-2022-jp?B?Y=?="]
     end
@@ -951,10 +955,10 @@ def convert(from, to)
     it "does not join different encodings" do
       convert "A=?iso-2022-jp?B?X=?==?utf-8?B?Y=?=B", ["A", "=?iso-2022-jp?B?X=?=", "=?utf-8?B?Y=?=", "B"]
     end
-    
+
     it "does not keep the separator character between two different encodings" do
       rfc_1342_newline_separators = ["\x0A", "\x20"]
-      
+
       rfc_1342_newline_separators.each do |rfc_1342_separator|
         convert "=?iso-2022-jp?B?X=?=#{rfc_1342_separator}=?utf-8?Q?Y=?=", ["=?iso-2022-jp?B?X=?=", "=?utf-8?Q?Y=?="]
       end