Email tweaks (#988)

* small tweaks to email validation

* more test examples

* change and better error messages

* remove duplicate tests

* linting

* tweak from @hramezani

changes/988-samuelcolvin.md

@@ -0,0 +1,3 @@
+Changes to email validation: whitespace is stripped from names, e.g. `' Fred Smith <fred@example.com>'`,
+unicode NFC normalization is applied to the local part, the domain part is cleaned to the internationalized form
+by round-tripping through IDNA ASCII, better error messages are provided when validation fails

pydantic/errors.py

@@ -186,7 +186,10 @@ class DictError(PydanticTypeError):
 
 
 class EmailError(PydanticValueError):
-    msg_template = 'value is not a valid email address'
+    msg_template = 'value is not a valid email address: {reason}'
+
+    def __init__(self, reason: str):
+        super().__init__(reason=reason)
 
 
 class UrlError(PydanticValueError):

pydantic/networks.py

@@ -700,19 +700,17 @@ def validate(cls, value: NetworkType) -> Union[IPv4Network, IPv6Network]:
             raise errors.IPvAnyNetworkError()
 
 
-pretty_email_regex = re.compile(r'([\w ]*?) *<(.*)> *')
+pretty_email_regex = re.compile(r' *([\w ]*?) *<(.+?)> *')
 
 
 def validate_email(value: Union[str]) -> Tuple[str, str]:
     """
-    Brutally simple email address validation. Note unlike most email address validation
+    Email address validation using https://pypi.org/project/email-validator/
+
+    Notes:
     * raw ip address (literal) domain parts are not allowed.
     * "John Doe <local_part@domain.com>" style "pretty" email addresses are processed
-    * the local part check is extremely basic. This raises the possibility of unicode spoofing, but no better
-        solution is really possible.
     * spaces are striped from the beginning and end of addresses but no error is raised
-
-    See RFC 5322 but treat it with suspicion, there seems to exist no universally acknowledged test for a valid email!
     """
     if email_validator is None:
         import_email_validator()
@@ -725,12 +723,8 @@ def validate_email(value: Union[str]) -> Tuple[str, str]:
     email = value.strip()
 
     try:
-        email_validator.validate_email(email, check_deliverability=False)
+        parts = email_validator.validate_email(email, check_deliverability=False)
     except email_validator.EmailNotValidError as e:
-        raise errors.EmailError() from e
-
-    at_index = email.index('@')
-    local_part = email[:at_index]  # RFC 5321, local part must be case-sensitive.
-    global_part = email[at_index:].lower()
+        raise errors.EmailError(*e.args) from e
 
-    return name or local_part, local_part + global_part
+    return name or parts['local'], parts['email']

tests/test_networks.py

@@ -6,6 +6,7 @@
     AnyUrl,
     BaseModel,
     CockroachDsn,
+    EmailError,
     EmailStr,
     FileUrl,
     HttpUrl,
@@ -746,6 +747,8 @@ class Model(BaseModel):
         ('foo.bar@example.com ', 'foo.bar', 'foo.bar@example.com'),
         ('foo BAR <foobar@example.com >', 'foo BAR', 'foobar@example.com'),
         ('FOO bar   <foobar@example.com> ', 'FOO bar', 'foobar@example.com'),
+        (' Whatever <foobar@example.com>', 'Whatever', 'foobar@example.com'),
+        ('Whatever < foobar@example.com>', 'Whatever', 'foobar@example.com'),
         ('<FOOBAR@example.com> ', 'FOOBAR', 'FOOBAR@example.com'),
         ('ñoñó@example.com', 'ñoñó', 'ñoñó@example.com'),
         ('我買@example.com', '我買', '我買@example.com'),
@@ -759,6 +762,11 @@ class Model(BaseModel):
         ('foo.bar@example.com', 'foo.bar', 'foo.bar@example.com'),
         ('foo.bar@exam-ple.com ', 'foo.bar', 'foo.bar@exam-ple.com'),
         ('ιωάννης@εεττ.gr', 'ιωάννης', 'ιωάννης@εεττ.gr'),
+        ('foobar@аррӏе.com', 'foobar', 'foobar@аррӏе.com'),
+        ('foobar@xn--80ak6aa92e.com', 'foobar', 'foobar@аррӏе.com'),
+        ('аррӏе@example.com', 'аррӏе', 'аррӏе@example.com'),
+        ('xn--80ak6aa92e@example.com', 'xn--80ak6aa92e', 'xn--80ak6aa92e@example.com'),
+        ('葉士豪@臺網中心.tw', '葉士豪', '葉士豪@臺網中心.tw'),
     ],
 )
 def test_address_valid(value, name, email):
@@ -767,31 +775,36 @@ def test_address_valid(value, name, email):
 
 @pytest.mark.skipif(not email_validator, reason='email_validator not installed')
 @pytest.mark.parametrize(
-    'value',
+    'value,reason',
     [
-        'f oo.bar@example.com ',
-        'foo.bar@exam\nple.com ',
-        'foobar',
-        'foobar <foobar@example.com',
-        '@example.com',
-        'foobar@.example.com',
-        'foobar@.com',
-        'foo bar@example.com',
-        'foo@bar@example.com',
-        '\n@example.com',
-        '\r@example.com',
-        '\f@example.com',
-        ' @example.com',
-        '\u0020@example.com',
-        '\u001f@example.com',
-        '"@example.com',
-        '\"@example.com',
-        ',@example.com',
-        'foobar <foobar<@example.com>',
+        ('@example.com', 'There must be something before the @-sign.'),
+        ('f oo.bar@example.com', 'The email address contains invalid characters before the @-sign'),
+        ('foobar', 'The email address is not valid. It must have exactly one @-sign.'),
+        ('foobar@localhost', 'The domain name localhost is not valid. It should have a period.'),
+        ('foobar@127.0.0.1', 'The domain name 127.0.0.1 is not valid. It is not within a valid top-level domain.'),
+        ('foo.bar@exam\nple.com ', None),
+        ('foobar <foobar@example.com', None),
+        ('foobar@.example.com', None),
+        ('foobar@.com', None),
+        ('foo bar@example.com', None),
+        ('foo@bar@example.com', None),
+        ('\n@example.com', None),
+        ('\r@example.com', None),
+        ('\f@example.com', None),
+        (' @example.com', None),
+        ('\u0020@example.com', None),
+        ('\u001f@example.com', None),
+        ('"@example.com', None),
+        ('\"@example.com', None),
+        (',@example.com', None),
+        ('foobar <foobar<@example.com>', None),
+        ('foobar <foobar@example.com>>', None),
+        ('foobar <<foobar<@example.com>', None),
+        ('foobar <>', None),
     ],
 )
-def test_address_invalid(value):
-    with pytest.raises(ValueError):
+def test_address_invalid(value, reason):
+    with pytest.raises(EmailError, match=f'value is not a valid email address: {reason or ""}'):
         validate_email(value)
 
 

tests/test_types.py

@@ -1137,8 +1137,24 @@ class MoreStringsModel(BaseModel):
             'type': 'value_error.any_str.min_length',
             'ctx': {'limit_value': 5},
         },
-        {'loc': ('str_email',), 'msg': 'value is not a valid email address', 'type': 'value_error.email'},
-        {'loc': ('name_email',), 'msg': 'value is not a valid email address', 'type': 'value_error.email'},
+        {
+            'loc': ('str_email',),
+            'msg': (
+                'value is not a valid email address: The email address contains invalid characters before the @-sign: '
+                'LESS-THAN SIGN.'
+            ),
+            'type': 'value_error.email',
+            'ctx': {'reason': 'The email address contains invalid characters before the @-sign: LESS-THAN SIGN.'},
+        },
+        {
+            'loc': ('name_email',),
+            'msg': (
+                'value is not a valid email address: '
+                'The email address contains invalid characters before the @-sign: SPACE.'
+            ),
+            'type': 'value_error.email',
+            'ctx': {'reason': 'The email address contains invalid characters before the @-sign: SPACE.'},
+        },
     ]