Explicitly check for non-ASCII characters with better error messages when SMTPUTF8 is off

Author: JoshData

email_validator/syntax.py

@@ -75,9 +75,25 @@ def validate_email_local_part(local, allow_smtputf8=True, allow_empty_local=Fals
     # for internationalized addresses. It's the same pattern but with additional
     # characters permitted.
     m = DOT_ATOM_TEXT_INTL.match(local)
-    if m and allow_smtputf8:
+    if m:
         # It's valid.
 
+        # But international characters in the local part may not be permitted.
+        if not allow_smtputf8:
+            # Check for invalid characters against the non-internationalized
+            # permitted character set.
+            # (RFC 2822 Section 3.2.4 / RFC 5322 Section 3.2.3)
+            bad_chars = set(
+                safe_character_display(c)
+                for c in local
+                if not ATEXT_RE.match(c)
+            )
+            if bad_chars:
+                raise EmailSyntaxError("Internationalized characters before the @-sign are not supported: " + ", ".join(sorted(bad_chars)) + ".")
+
+            # Although the check above should always find something, fall back to this just in case.
+            raise EmailSyntaxError("Internationalized characters before the @-sign are not supported.")
+
         # RFC 6532 section 3.1 also says that Unicode NFC normalization should be applied,
         # so we'll return the normalized local part in the return value.
         local = unicodedata.normalize("NFC", local)
@@ -111,7 +127,7 @@ def validate_email_local_part(local, allow_smtputf8=True, allow_empty_local=Fals
     bad_chars = set(
         safe_character_display(c)
         for c in local
-        if not (ATEXT_INTL_RE if allow_smtputf8 else ATEXT_RE).match(c)
+        if not ATEXT_INTL_RE.match(c)
     )
     if bad_chars:
         raise EmailSyntaxError("The email address contains invalid characters before the @-sign: " + ", ".join(sorted(bad_chars)) + ".")

tests/test_syntax.py

@@ -56,6 +56,31 @@
                 ascii_email="!#$%&'*+-/=?^_`.{|}~@example.tld",
             ),
         ),
+        (
+            'jeff@臺網中心.tw',
+            ValidatedEmail(
+                local_part='jeff',
+                ascii_local_part='jeff',
+                smtputf8=False,
+                ascii_domain='xn--fiqq24b10vi0d.tw',
+                domain='臺網中心.tw',
+                email='jeff@臺網中心.tw',
+                ascii_email='jeff@xn--fiqq24b10vi0d.tw',
+            ),
+        ),
+    ],
+)
+def test_email_valid(email_input, output):
+    # These addresses do not require SMTPUTF8. See test_email_valid_intl_local_part
+    # for addresses that are valid but require SMTPUTF8. Check that it passes with
+    # allow_smtput8 both on and off.
+    assert validate_email(email_input, check_deliverability=False, allow_smtputf8=False) == output
+    assert validate_email(email_input, check_deliverability=False, allow_smtputf8=True) == output
+
+
+@pytest.mark.parametrize(
+    'email_input,output',
+    [
         (
             '伊昭傑@郵件.商務',
             ValidatedEmail(
@@ -106,18 +131,6 @@
                 email='葉士豪@臺網中心.tw',
             ),
         ),
-        (
-            'jeff@臺網中心.tw',
-            ValidatedEmail(
-                local_part='jeff',
-                ascii_local_part='jeff',
-                smtputf8=False,
-                ascii_domain='xn--fiqq24b10vi0d.tw',
-                domain='臺網中心.tw',
-                email='jeff@臺網中心.tw',
-                ascii_email='jeff@xn--fiqq24b10vi0d.tw',
-            ),
-        ),
         (
             '葉士豪@臺網中心.台灣',
             ValidatedEmail(
@@ -200,10 +213,15 @@
         ),
     ],
 )
-def test_email_valid(email_input, output):
-    # print(f'({email_input!r}, {validate_email(email_input, check_deliverability=False)!r}),')
+def test_email_valid_intl_local_part(email_input, output):
+    # Check that it passes when allow_smtputf8 is True.
     assert validate_email(email_input, check_deliverability=False) == output
 
+    # Check that it fails when allow_smtputf8 is False.
+    with pytest.raises(EmailSyntaxError) as exc_info:
+        validate_email(email_input, allow_smtputf8=False, check_deliverability=False)
+    assert "Internationalized characters before the @-sign are not supported: " in str(exc_info.value)
+
 
 @pytest.mark.parametrize(
     'email_input,error_msg',
@@ -263,7 +281,6 @@ def test_email_invalid_syntax(email_input, error_msg):
     # checks do not arise.
     with pytest.raises(EmailSyntaxError) as exc_info:
         validate_email(email_input)
-    # print(f'({email_input!r}, {str(exc_info.value)!r}),')
     assert str(exc_info.value) == error_msg
 
 
@@ -283,7 +300,6 @@ def test_email_invalid_reserved_domain(email_input):
     # DNS deliverability checks do not arise.
     with pytest.raises(EmailSyntaxError) as exc_info:
         validate_email(email_input)
-    # print(f'({email_input!r}, {str(exc_info.value)!r}),')
     assert "is a special-use or reserved name" in str(exc_info.value)
 
 
@@ -317,16 +333,31 @@ def test_email_unsafe_character(s, expected_error):
     ('email_input', 'expected_error'),
     [
         ('white space@test', 'The email address contains invalid characters before the @-sign: SPACE.'),
+        ('test@white space', 'The part after the @-sign contains invalid characters: SPACE.'),
         ('\n@test', 'The email address contains invalid characters before the @-sign: U+000A.'),
+        ('test@\n', 'The part after the @-sign contains invalid characters: U+000A.'),
     ],
 )
 def test_email_invalid_character(email_input, expected_error):
-    # Check for various unsafe characters:
+    # Check for various unsafe test_email_invalid_character_smtputf8:
     with pytest.raises(EmailSyntaxError) as exc_info:
         validate_email(email_input, test_environment=True)
     assert str(exc_info.value) == expected_error
 
 
+@pytest.mark.parametrize(
+    ('email_input', 'expected_error'),
+    [
+        ('λambdaツ@test', 'Internationalized characters before the @-sign are not supported: \'λ\', \'ツ\'.'),
+    ],
+)
+def test_email_invalid_character_smtputf8(email_input, expected_error):
+    # Check for various unsafe characters:
+    with pytest.raises(EmailSyntaxError) as exc_info:
+        validate_email(email_input, allow_smtputf8=False, test_environment=True)
+    assert str(exc_info.value) == expected_error
+
+
 def test_email_test_domain_name_in_test_environment():
     validate_email("anything@test", test_environment=True)
     validate_email("anything@mycompany.test", test_environment=True)