Enhance isEmail validation (#269)

This enhances the current `isEmail` validation by using a consistent
regex that will be used across protovalidate implementations.

Author: smaye81

protovalidate/internal/extra_func.py

@@ -13,8 +13,8 @@
 # limitations under the License.
 
 import math
+import re
 import typing
-from email.utils import parseaddr
 from ipaddress import IPv4Address, IPv4Network, IPv6Address, IPv6Network, ip_address, ip_network
 from urllib import parse as urlparse
 
@@ -23,6 +23,11 @@
 
 from protovalidate.internal import string_format
 
+# See https://html.spec.whatwg.org/multipage/input.html#valid-e-mail-address
+_email_regex = re.compile(
+    r"^[a-zA-Z0-9.!#$%&'*+/=?^_`{|}~-]+@[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*$"
+)
+
 
 def _validate_hostname(host):
     if not host:
@@ -49,23 +54,6 @@ def _validate_hostname(host):
     return not all_digits
 
 
-def validate_email(addr):
-    parts = parseaddr(addr)
-    if addr != parts[1]:
-        return False
-
-    addr = parts[1]
-    if len(addr) > 254:
-        return False
-
-    parts = addr.split("@")
-    if len(parts) != 2:
-        return False
-    if len(parts[0]) > 64:
-        return False
-    return _validate_hostname(parts[1])
-
-
 def validate_host_and_port(string: str, *, port_required: bool) -> bool:
     if not string:
         return False
@@ -157,10 +145,19 @@ def is_ip_prefix(val: celtypes.Value, *args) -> celpy.Result:
 
 
 def is_email(string: celtypes.Value) -> celpy.Result:
+    """Returns true if the string is an email address, for example "[email protected]".
+
+    Conforms to the definition for a valid email address from the HTML standard.
+    Note that this standard willfully deviates from RFC 5322, which allows many
+    unexpected forms of email addresses and will easily match a typographical
+    error.
+    """
+
     if not isinstance(string, celtypes.StringType):
         msg = "invalid argument, expected string"
         raise celpy.CELEvalError(msg)
-    return celtypes.BoolType(validate_email(string))
+    m = _email_regex.match(string) is not None
+    return celtypes.BoolType(m)
 
 
 def is_uri(string: celtypes.Value) -> celpy.Result:

tests/conformance/nonconforming.yaml

@@ -8,46 +8,6 @@ standard_constraints/well_known_types/timestamp:
  - gte_lte/invalid/above
  - lte/invalid
 
-library/is_email:
-  - invalid/left_side_empty
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - invalid/non_ascii
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"µ@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - invalid/quoted-string/a
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"\"foo bar\"@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - invalid/quoted-string/b
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"\"foo..bar\"@example.com"}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - invalid/trailing_dot
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"[email protected]."}
-  #        want: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #         got: valid
-  - valid/exhaust_atext
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ!#$%&'*+-/=?^_`{|}[email protected]"}
-  #        want: valid
-  #         got: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #             message: ""
-  #             for_key: false
-  - valid/label_all_digits
-  #       input: [type.googleapis.com/buf.validate.conformance.cases.IsEmail]:{val:"[email protected]"}
-  #        want: valid
-  #         got: validation error (1 violation)
-  #          1. constraint_id: "library.is_email"
-  #             message: ""
-  #             for_key: false
 library/is_host_and_port:
   - port_required/false/invalid/port_number_sign
   #       input: [type.googleapis.com/buf.validate.conformance.cases.IsHostAndPort]:{val:"example.com:+0"}