Fix body line iteration to properly decode Content-Transfer-Encoding

The previous fix (removing decode=True) stopped the crash but left
quoted-printable and base64 bodies undecoded, causing e.g. QP soft
line breaks (=20) to appear literally in addresses and break pattern
matching.

Add _body_line_iterator() to SimpleMatch which uses get_payload(decode=True)
to get CTE-decoded bytes, then decodes to str using the part's charset.
SimpleWarning imports and uses the same helper. Tagger likewise imports it
so topic matching works correctly on encoded message bodies.

Author: thegushi

Mailman/Bouncers/SimpleMatch.py

@@ -17,10 +17,30 @@
 
 """Recognizes simple heuristically delimited bounces."""
 
+import io
 import re
 import email.iterators
 
 
+def _body_line_iterator(msg):
+    """Iterate body lines with Content-Transfer-Encoding properly decoded.
+
+    Python 3's body_line_iterator only yields str payloads; get_payload(
+    decode=True) returns bytes (CTE-decoded). This helper bridges the gap so
+    quoted-printable and base64 bodies are decoded before pattern matching.
+    """
+    for subpart in email.iterators.typed_subpart_iterator(msg):
+        payload = subpart.get_payload(decode=True)
+        if not isinstance(payload, bytes):
+            continue
+        charset = subpart.get_content_charset('us-ascii') or 'us-ascii'
+        try:
+            text = payload.decode(charset, errors='replace')
+        except LookupError:
+            text = payload.decode('us-ascii', errors='replace')
+        yield from io.StringIO(text)
+
+
 
 def _c(pattern):
     return re.compile(pattern, re.IGNORECASE)
@@ -224,7 +244,7 @@ def process(msg, patterns=None):
     # we process the message multiple times anyway.
     for scre, ecre, acre in patterns:
         state = 0
-        for line in email.iterators.body_line_iterator(msg):
+        for line in _body_line_iterator(msg):
             if state == 0:
                 if scre.search(line):
                     state = 1

Mailman/Bouncers/SimpleWarning.py

@@ -21,7 +21,7 @@
 import email.iterators
 
 from Mailman.Bouncers.BouncerAPI import Stop
-from Mailman.Bouncers.SimpleMatch import _c
+from Mailman.Bouncers.SimpleMatch import _c, _body_line_iterator
 
 
 
@@ -75,7 +75,7 @@ def process(msg):
     addrs = {}
     for scre, ecre, acre in patterns:
         state = 0
-        for line in email.iterators.body_line_iterator(msg):
+        for line in _body_line_iterator(msg):
             if state == 0:
                 if scre.search(line):
                     state = 1

Mailman/Handlers/Tagger.py

@@ -26,6 +26,7 @@
 from email.header import decode_header
 
 from Mailman import Utils
+from Mailman.Bouncers.SimpleMatch import _body_line_iterator
 from Mailman.Logging.Syslog import syslog
 from Mailman.Handlers.CookHeaders import change_header
 
@@ -97,7 +98,7 @@ def scanbody(msg, numlines=None):
     # the first numlines of body text.
     lines = []
     lineno = 0
-    reader = list(email.iterators.body_line_iterator(msg))
+    reader = list(_body_line_iterator(msg))
     while numlines is None or lineno < numlines:
         try:
             line = reader.pop(0)