bpo-36520: Email header folded incorrectly (#13608)

* bpo-36520: reset the encoded word offset when starting a new
line during an email header folding operation

* 📜🤖 Added by blurb_it.

* bpo-36520: add an additional test case, and provide descriptive
comments for the test_folding_with_utf8_encoding_* tests

* bpo-36520: fix whitespace issue

* bpo-36520: changes per reviewer request -- remove extraneous
backslashes; add whitespace between terminating quotes and
line-continuation backslashes; use "bpo-" instead of
"issue #" in comments

Lib/email/_header_value_parser.py

@@ -2789,6 +2789,7 @@ def _refold_parse_tree(parse_tree, *, policy):
             newline = _steal_trailing_WSP_if_exists(lines)
             if newline or part.startswith_fws():
                 lines.append(newline + tstr)
+                last_ew = None
                 continue
         if not hasattr(part, 'encode'):
             # It's not a terminal, try folding the subparts.

Lib/test/test_email/test_message.py

@@ -784,6 +784,137 @@ class TestEmailMessage(TestEmailMessageBase, TestEmailBase):
         m['Subject'] = 'unicöde'
         self.assertEqual(str(m), 'Subject: unicöde\n\n')
 
+    def test_folding_with_utf8_encoding_1(self):
+        # bpo-36520
+        #
+        # Fold a line that contains UTF-8 words before
+        # and after the whitespace fold point, where the
+        # line length limit is reached within an ASCII
+        # word.
+
+        m = EmailMessage()
+        m['Subject'] = 'Hello Wörld! Hello Wörld! '            \
+                       'Hello Wörld! Hello Wörld!Hello Wörld!'
+        self.assertEqual(bytes(m),
+                         b'Subject: Hello =?utf-8?q?W=C3=B6rld!_Hello_W'
+                         b'=C3=B6rld!_Hello_W=C3=B6rld!?=\n'
+                         b' Hello =?utf-8?q?W=C3=B6rld!Hello_W=C3=B6rld!?=\n\n')
+
+
+    def test_folding_with_utf8_encoding_2(self):
+        # bpo-36520
+        #
+        # Fold a line that contains UTF-8 words before
+        # and after the whitespace fold point, where the
+        # line length limit is reached at the end of an
+        # encoded word.
+
+        m = EmailMessage()
+        m['Subject'] = 'Hello Wörld! Hello Wörld! '                \
+                       'Hello Wörlds123! Hello Wörld!Hello Wörld!'
+        self.assertEqual(bytes(m),
+                         b'Subject: Hello =?utf-8?q?W=C3=B6rld!_Hello_W'
+                         b'=C3=B6rld!_Hello_W=C3=B6rlds123!?=\n'
+                         b' Hello =?utf-8?q?W=C3=B6rld!Hello_W=C3=B6rld!?=\n\n')
+
+    def test_folding_with_utf8_encoding_3(self):
+        # bpo-36520
+        #
+        # Fold a line that contains UTF-8 words before
+        # and after the whitespace fold point, where the
+        # line length limit is reached at the end of the
+        # first word.
+
+        m = EmailMessage()
+        m['Subject'] = 'Hello-Wörld!-Hello-Wörld!-Hello-Wörlds123! ' \
+                       'Hello Wörld!Hello Wörld!'
+        self.assertEqual(bytes(m), \
+                         b'Subject: =?utf-8?q?Hello-W=C3=B6rld!-Hello-W'
+                         b'=C3=B6rld!-Hello-W=C3=B6rlds123!?=\n'
+                         b' Hello =?utf-8?q?W=C3=B6rld!Hello_W=C3=B6rld!?=\n\n')
+
+    def test_folding_with_utf8_encoding_4(self):
+        # bpo-36520
+        #
+        # Fold a line that contains UTF-8 words before
+        # and after the fold point, where the first
+        # word is UTF-8 and the fold point is within
+        # the word.
+
+        m = EmailMessage()
+        m['Subject'] = 'Hello-Wörld!-Hello-Wörld!-Hello-Wörlds123!-Hello' \
+                       ' Wörld!Hello Wörld!'
+        self.assertEqual(bytes(m),
+                         b'Subject: =?utf-8?q?Hello-W=C3=B6rld!-Hello-W'
+                         b'=C3=B6rld!-Hello-W=C3=B6rlds123!?=\n'
+                         b' =?utf-8?q?-Hello_W=C3=B6rld!Hello_W=C3=B6rld!?=\n\n')
+
+    def test_folding_with_utf8_encoding_5(self):
+        # bpo-36520
+        #
+        # Fold a line that contains a UTF-8 word after
+        # the fold point.
+
+        m = EmailMessage()
+        m['Subject'] = '123456789 123456789 123456789 123456789 123456789' \
+                       ' 123456789 123456789 Hello Wörld!'
+        self.assertEqual(bytes(m),
+                         b'Subject: 123456789 123456789 123456789 123456789'
+                         b' 123456789 123456789 123456789\n'
+                         b' Hello =?utf-8?q?W=C3=B6rld!?=\n\n')
+
+    def test_folding_with_utf8_encoding_6(self):
+        # bpo-36520
+        #
+        # Fold a line that contains a UTF-8 word before
+        # the fold point and ASCII words after
+
+        m = EmailMessage()
+        m['Subject'] = '123456789 123456789 123456789 123456789 Hello Wörld!' \
+                       ' 123456789 123456789 123456789 123456789 123456789'   \
+                       ' 123456789'
+        self.assertEqual(bytes(m),
+                         b'Subject: 123456789 123456789 123456789 123456789'
+                         b' Hello =?utf-8?q?W=C3=B6rld!?=\n 123456789 '
+                         b'123456789 123456789 123456789 123456789 '
+                         b'123456789\n\n')
+
+    def test_folding_with_utf8_encoding_7(self):
+        # bpo-36520
+        #
+        # Fold a line twice that contains UTF-8 words before
+        # and after the first fold point, and ASCII words
+        # after the second fold point.
+
+        m = EmailMessage()
+        m['Subject'] = '123456789 123456789 Hello Wörld! Hello Wörld! '       \
+                       '123456789-123456789 123456789 Hello Wörld! 123456789' \
+                       ' 123456789'
+        self.assertEqual(bytes(m),
+                         b'Subject: 123456789 123456789 Hello =?utf-8?q?'
+                         b'W=C3=B6rld!_Hello_W=C3=B6rld!?=\n'
+                         b' 123456789-123456789 123456789 Hello '
+                         b'=?utf-8?q?W=C3=B6rld!?= 123456789\n 123456789\n\n')
+
+    def test_folding_with_utf8_encoding_8(self):
+        # bpo-36520
+        #
+        # Fold a line twice that contains UTF-8 words before
+        # the first fold point, and ASCII words after the
+        # first fold point, and UTF-8 words after the second
+        # fold point.
+
+        m = EmailMessage()
+        m['Subject'] = '123456789 123456789 Hello Wörld! Hello Wörld! '       \
+                       '123456789 123456789 123456789 123456789 123456789 '   \
+                       '123456789-123456789 123456789 Hello Wörld! 123456789' \
+                       ' 123456789'
+        self.assertEqual(bytes(m),
+                         b'Subject: 123456789 123456789 Hello '
+                         b'=?utf-8?q?W=C3=B6rld!_Hello_W=C3=B6rld!?=\n 123456789 '
+                         b'123456789 123456789 123456789 123456789 '
+                         b'123456789-123456789\n 123456789 Hello '
+                         b'=?utf-8?q?W=C3=B6rld!?= 123456789 123456789\n\n')
 
 class TestMIMEPart(TestEmailMessageBase, TestEmailBase):
     # Doing the full test run here may seem a bit redundant, since the two

Misc/NEWS.d/next/Library/2019-05-28-02-37-00.bpo-36520.W4tday.rst

+Lengthy email headers with UTF-8 characters are now properly encoded when they are folded. Patch by Jeffrey Kintscher.