bpo-21315: Fix parsing of encoded words with missing leading ws. (GH-13425) (#13846)

* bpo-21315: Fix parsing of encoded words with missing leading ws.

Because of missing leading whitespace, encoded word would get parsed as
unstructured token. This patch fixes that by looking for encoded words when
splitting tokens with whitespace.

Missing trailing whitespace around encoded word now register a defect
instead.

Original patch suggestion by David R. Murray on bpo-21315.
(cherry picked from commit 66c4f3f38b867d8329b28c032bb907fd1a2f22d2)

Co-authored-by: Abhilash Raj <maxking@users.noreply.github.com>

4 files changed, 49 insertions, 3 deletions

diff --git a/Lib/email/_header_value_parser.py b/Lib/email/_header_value_parser.py
index 958ef5018c..18aecbffa7 100644
--- a/Lib/email/_header_value_parser.py
+++ b/Lib/email/_header_value_parser.py
@@ -97,6 +97,18 @@ EXTENDED_ATTRIBUTE_ENDS = ATTRIBUTE_ENDS - set('%')
 def quote_string(value):
     return '"'+str(value).replace('\\', '\\\\').replace('"', r'\"')+'"'
 
+# Match a RFC 2047 word, looks like =?utf-8?q?someword?=
+rfc2047_matcher = re.compile(r'''
+   =\?            # literal =?
+   [^?]*          # charset
+   \?             # literal ?
+   [qQbB]         # literal 'q' or 'b', case insensitive
+   \?             # literal ?
+  .*?             # encoded word
+  \?=             # literal ?=
+''', re.VERBOSE | re.MULTILINE)
+
+
 #
 # TokenList and its subclasses
 #
@@ -1050,6 +1062,10 @@ def get_encoded_word(value):
         _validate_xtext(vtext)
         ew.append(vtext)
         text = ''.join(remainder)
+    # Encoded words should be followed by a WS
+    if value and value[0] not in WSP:
+        ew.defects.append(errors.InvalidHeaderDefect(
+            "missing trailing whitespace after encoded-word"))
     return ew, value
 
 def get_unstructured(value):
@@ -1102,6 +1118,11 @@ def get_unstructured(value):
                 unstructured.append(token)
                 continue
         tok, *remainder = _wsp_splitter(value, 1)
+        # Split in the middle of an atom if there is a rfc2047 encoded word
+        # which does not have WSP on both sides. The defect will be registered
+        # the next time through the loop.
+        if rfc2047_matcher.search(tok):
+            tok, *remainder = value.partition('=?')
         vtext = ValueTerminal(tok, 'vtext')
         _validate_xtext(vtext)
         unstructured.append(vtext)
diff --git a/Lib/test/test_email/test__header_value_parser.py b/Lib/test/test_email/test__header_value_parser.py
index 676732bb3d..693487bc96 100644
--- a/Lib/test/test_email/test__header_value_parser.py
+++ b/Lib/test/test_email/test__header_value_parser.py
@@ -118,7 +118,7 @@ class TestParser(TestParserMixin, TestEmailBase):
                          '=?us-ascii?q?first?==?utf-8?q?second?=',
                          'first',
                          'first',
-                         [],
+                         [errors.InvalidHeaderDefect],
                          '=?utf-8?q?second?=')
 
     def test_get_encoded_word_sets_extra_attributes(self):
@@ -361,6 +361,25 @@ class TestParser(TestParserMixin, TestEmailBase):
             '=?utf-8?q?foo?==?utf-8?q?bar?=',
             'foobar',
             'foobar',
+            [errors.InvalidHeaderDefect,
+            errors.InvalidHeaderDefect],
+            '')
+
+    def test_get_unstructured_ew_without_leading_whitespace(self):
+        self._test_get_x(
+            self._get_unst,
+            'nowhitespace=?utf-8?q?somevalue?=',
+            'nowhitespacesomevalue',
+            'nowhitespacesomevalue',
+            [errors.InvalidHeaderDefect],
+            '')
+
+    def test_get_unstructured_ew_without_trailing_whitespace(self):
+        self._test_get_x(
+            self._get_unst,
+            '=?utf-8?q?somevalue?=nowhitespace',
+            'somevaluenowhitespace',
+            'somevaluenowhitespace',
             [errors.InvalidHeaderDefect],
             '')
 
@@ -546,7 +565,8 @@ class TestParser(TestParserMixin, TestEmailBase):
             '"=?utf-8?Q?not_really_valid?="',
             '"not really valid"',
             'not really valid',
-            [errors.InvalidHeaderDefect],
+            [errors.InvalidHeaderDefect,
+             errors.InvalidHeaderDefect],
             '')
 
     # get_comment
diff --git a/Lib/test/test_email/test_headerregistry.py b/Lib/test/test_email/test_headerregistry.py
index d1007099f6..e6db3acedc 100644
--- a/Lib/test/test_email/test_headerregistry.py
+++ b/Lib/test/test_email/test_headerregistry.py
@@ -1180,7 +1180,8 @@ class TestAddressHeader(TestHeaderBase):
 
         'rfc2047_atom_in_quoted_string_is_decoded':
             ('"=?utf-8?q?=C3=89ric?=" <foo@example.com>',
-            [errors.InvalidHeaderDefect],
+            [errors.InvalidHeaderDefect,
+            errors.InvalidHeaderDefect],
             'Éric <foo@example.com>',
             'Éric',
             'foo@example.com',
diff --git a/Misc/NEWS.d/next/Library/2019-05-19-10-48-46.bpo-21315.PgXVqF.rst b/Misc/NEWS.d/next/Library/2019-05-19-10-48-46.bpo-21315.PgXVqF.rst
new file mode 100644
index 0000000000..dd0dd7f72c
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2019-05-19-10-48-46.bpo-21315.PgXVqF.rst
@@ -0,0 +1,4 @@
+Email headers containing RFC2047 encoded words are parsed despite the missing
+whitespace, and a defect registered. Also missing trailing whitespace after
+encoded words is now registered as a defect.
+