gh-66889: Report which character made an address invalid

Lib/email/_header_value_parser.py

@@ -152,7 +152,10 @@ def value(self):
 
     @property
     def all_defects(self):
-        return sum((x.all_defects for x in self), self.defects)
+        defects = list(self.defects)
+        for x in self:
+            defects.extend(x.all_defects)
+        return defects
 
     def startswith_fws(self):
         return self[0].startswith_fws()
@@ -1859,8 +1862,10 @@ def get_invalid_mailbox(value, endchars):
     invalid_mailbox = InvalidMailbox()
     while value and value[0] not in endchars:
         if value[0] in PHRASE_ENDS:
-            invalid_mailbox.append(ValueTerminal(value[0],
-                                                 'misplaced-special'))
+            special = ValueTerminal(value[0], 'misplaced-special')
+            special.defects.append(errors.InvalidHeaderDefect(
+                "misplaced special character {!r}".format(value[0])))
+            invalid_mailbox.append(special)
             value = value[1:]
         else:
             token, value = get_phrase(value)

Lib/test/test_email/test__header_value_parser.py

@@ -16,6 +16,25 @@ def test_EWWhiteSpaceTerminal(self):
         self.assertEqual(x.value, '')
         self.assertEqual(x.token_type, 'fws')
 
+    def test_all_defects_collects_subtree_in_order(self):
+        leaf = parser.ValueTerminal('x', 'atext')
+        leaf.defects.append(errors.InvalidHeaderDefect('leaf'))
+        inner = parser.TokenList([leaf])
+        inner.defects.append(errors.InvalidHeaderDefect('inner'))
+        outer = parser.TokenList([inner])
+        outer.defects.append(errors.InvalidHeaderDefect('outer'))
+        self.assertEqual([str(d) for d in outer.all_defects],
+                         ['outer', 'inner', 'leaf'])
+
+    def test_all_defects_is_a_copy(self):
+        tl = parser.TokenList()
+        tl.defects.append(errors.InvalidHeaderDefect('only'))
+        defects = tl.all_defects
+        self.assertEqual([str(d) for d in defects], ['only'])
+        self.assertIsNot(defects, tl.defects)
+        defects.append(errors.InvalidHeaderDefect('added'))
+        self.assertEqual(len(tl.defects), 1)
+
 
 class TestParserMixin:
 
@@ -2158,6 +2177,9 @@ def test_get_mailbox_list_unparseable_mailbox_null(self):
                 ' "Fred Flintstone" <dinsdale@test. example.com>'),
             [errors.InvalidHeaderDefect,   # the 'extra' text after the local part
              errors.InvalidHeaderDefect,   # the local part with no angle-addr
+             errors.InvalidHeaderDefect,   # misplaced '[' in the extra text
+             errors.InvalidHeaderDefect,   # misplaced ']' in the extra text
+             errors.InvalidHeaderDefect,   # misplaced '@' in the extra text
              errors.ObsoleteHeaderDefect,  # period in extra text (example.com)
              errors.ObsoleteHeaderDefect], # (bird) in valid address.
             '')
@@ -2186,7 +2208,9 @@ def test_get_mailbox_list_junk_after_valid_address(self):
                 ' "Fred Flintstone" <dinsdale@test.example.com>'),
             ('"Roy A. Bear" <dinsdale@example.com>@@,'
                 ' "Fred Flintstone" <dinsdale@test.example.com>'),
-            [errors.InvalidHeaderDefect],
+            [errors.InvalidHeaderDefect,   # the junk after the valid address
+             errors.InvalidHeaderDefect,   # first misplaced '@' in the junk
+             errors.InvalidHeaderDefect],  # second misplaced '@' in the junk
             '')
         self.assertEqual(len(mailbox_list.mailboxes), 1)
         self.assertEqual(len(mailbox_list.all_mailboxes), 2)
@@ -2688,6 +2712,8 @@ def test_get_address_list_mailboxes_invalid_addresses(self):
              [errors.InvalidHeaderDefect,   # invalid address in list
               errors.InvalidHeaderDefect,   # 'Foo x' local part invalid.
               errors.InvalidHeaderDefect,   # Missing . in 'Foo x' local part
+              errors.InvalidHeaderDefect,   # misplaced '[' after the addr-spec
+              errors.InvalidHeaderDefect,   # misplaced ']' after the addr-spec
               errors.ObsoleteHeaderDefect,  # period in 'Is.' disp-name phrase
               errors.InvalidHeaderDefect,   # no domain part in addr-spec
               errors.ObsoleteHeaderDefect], # addr-spec has comment in it
@@ -2712,6 +2738,26 @@ def test_get_address_list_mailboxes_invalid_addresses(self):
             address_list.addresses[3].all_mailboxes[0].display_name,
                 "Nobody Is. Special")
 
+    def test_get_address_list_misplaced_special_has_defect(self):
+        address_list = self._test_get_x(parser.get_address_list,
+            'abc@xyz.c:om',
+            'abc@xyz.c:om',
+            'abc@xyz.c:om',
+            [errors.InvalidHeaderDefect,   # invalid address in list
+             errors.InvalidHeaderDefect],  # the misplaced ':'
+            '')
+        invalid_mailbox = address_list.addresses[1].all_mailboxes[0]
+        self.assertEqual(invalid_mailbox.token_type, 'invalid-mailbox')
+        misplaced = invalid_mailbox[0]
+        self.assertEqual(misplaced.token_type, 'misplaced-special')
+        self.assertEqual(misplaced, ':')
+        # The defect is attached to the misplaced-special token itself, so the
+        # offending character can be located in the parse tree.
+        self.assertDefectsEqual(misplaced.defects,
+                                [errors.InvalidHeaderDefect])
+        self.assertEqual(str(misplaced.defects[0]),
+                         "misplaced special character ':'")
+
     def test_get_address_list_group_empty(self):
         address_list = self._test_get_x(parser.get_address_list,
             'Monty Python: ;',

Misc/NEWS.d/next/Library/2026-09-25-10-46-12.gh-issue-66889.EtcdVD.rst

@@ -0,0 +1,7 @@
+The :mod:`email` header parser now attaches an ``InvalidHeaderDefect`` to the
+``misplaced-special`` token it creates for a stray special character in an
+invalid mailbox, instead of reporting the problem only on the enclosing
+address list or mailbox list.  As a result,
+:attr:`~email.headerregistry.BaseHeader.defects` for a malformed address
+header such as ``To: abc@xyz.c:om`` gains one entry per misplaced character.
+Patch by Dmitry Voropaev.

Misc/NEWS.d/next/Library/2026-09-25-11-02-30.gh-issue-66889.Kq7nXb.rst

@@ -0,0 +1,4 @@
+Parsing a malformed :mod:`email` address header no longer takes time
+quadratic in the number of defects the parser reports.  Collecting the
+defects of a parse tree copied the accumulated list once per node; it now
+appends to a single list.  Patch by Dmitry Voropaev.