Utils: add parse_rfc2822date

(Useful for interpreting date header in inbound messages)

anymail/utils.py

@@ -12,7 +12,7 @@ from django.conf import settings
 from django.core.mail.message import sanitize_address, DEFAULT_ATTACHMENT_MIME_TYPE
 from django.utils.encoding import force_text
 from django.utils.functional import Promise
-from django.utils.timezone import utc
+from django.utils.timezone import utc, get_fixed_timezone
 # noinspection PyUnresolvedReferences
 from six.moves.urllib.parse import urlsplit, urlunsplit
 
@@ -466,3 +466,40 @@ def get_request_uri(request):
         url = urlunsplit((parts.scheme, basic_auth + '@' + parts.netloc,
                           parts.path, parts.query, parts.fragment))
     return url
+
+
+try:
+    from email.utils import parsedate_to_datetime  # Python 3.3+
+except ImportError:
+    from email.utils import parsedate_tz
+
+    # Backport Python 3.3+ email.utils.parsedate_to_datetime
+    def parsedate_to_datetime(s):
+        # *dtuple, tz = _parsedate_tz(data)
+        dtuple = parsedate_tz(s)
+        tz = dtuple[-1]
+        # if tz is None:  # parsedate_tz returns 0 for "-0000"
+        if tz is None or (tz == 0 and "-0000" in s):
+            # "... indicates that the date-time contains no information
+            # about the local time zone" (RFC 2822 #3.3)
+            return datetime(*dtuple[:6])
+        else:
+            # tzinfo = datetime.timezone(datetime.timedelta(seconds=tz))  # Python 3.2+ only
+            tzinfo = get_fixed_timezone(tz // 60)  # don't use timedelta (avoid Django bug #28739)
+            return datetime(*dtuple[:6], tzinfo=tzinfo)
+
+
+def parse_rfc2822date(s):
+    """Parses an RFC-2822 formatted date string into a datetime.datetime
+
+    Returns None if string isn't parseable. Returned datetime will be naive
+    if string doesn't include known timezone offset; aware if it does.
+
+    (Same as Python 3 email.utils.parsedate_to_datetime, with improved
+    handling for unparseable date strings.)
+    """
+    try:
+        return parsedate_to_datetime(s)
+    except (IndexError, TypeError, ValueError):
+        # despite the docs, parsedate_to_datetime often dies on unparseable input
+        return None

tests/test_utils.py

@@ -18,10 +18,11 @@ except ImportError:
     string_concat = None
 
 from anymail.exceptions import AnymailInvalidAddress
-from anymail.utils import (parse_address_list, ParsedEmail,
+from anymail.utils import (
+    parse_address_list, ParsedEmail,
     is_lazy, force_non_lazy, force_non_lazy_dict, force_non_lazy_list,
     update_deep,
-                           get_request_uri, get_request_basic_auth)
+    get_request_uri, get_request_basic_auth, parse_rfc2822date)
 
 
 class ParseAddressListTests(SimpleTestCase):
@@ -296,3 +297,35 @@ class RequestUtilsTests(SimpleTestCase):
                                             HTTP_AUTHORIZATION=self.basic_auth('user', 'pass'))
         self.assertEqual(get_request_uri(request),
                          "https://user:pass@secret.example.com:8989/path/to/?query")
+
+
+class ParseRFC2822DateTests(SimpleTestCase):
+    def test_with_timezones(self):
+        dt = parse_rfc2822date("Tue, 24 Oct 2017 10:11:35 -0700")
+        self.assertEqual(dt.isoformat(), "2017-10-24T10:11:35-07:00")
+        self.assertIsNotNone(dt.utcoffset())  # aware
+
+        dt = parse_rfc2822date("Tue, 24 Oct 2017 10:11:35 +0700")
+        self.assertEqual(dt.isoformat(), "2017-10-24T10:11:35+07:00")
+        self.assertIsNotNone(dt.utcoffset())  # aware
+
+        dt = parse_rfc2822date("Tue, 24 Oct 2017 10:11:35 +0000")
+        self.assertEqual(dt.isoformat(), "2017-10-24T10:11:35+00:00")
+        self.assertIsNotNone(dt.tzinfo)  # aware
+
+    def test_without_timezones(self):
+        dt = parse_rfc2822date("Tue, 24 Oct 2017 10:11:35 -0000")  # "no timezone information"
+        self.assertEqual(dt.isoformat(), "2017-10-24T10:11:35")
+        self.assertIsNone(dt.tzinfo)  # naive (compare with +0000 version in previous test)
+
+        dt = parse_rfc2822date("Tue, 24 Oct 2017 10:11:35")
+        self.assertEqual(dt.isoformat(), "2017-10-24T10:11:35")
+        self.assertIsNone(dt.tzinfo)  # naive
+
+    def test_unparseable_dates(self):
+        self.assertIsNone(parse_rfc2822date(""))
+        self.assertIsNone(parse_rfc2822date("  "))
+        self.assertIsNone(parse_rfc2822date("garbage"))
+        self.assertIsNone(parse_rfc2822date("Tue, 24 Oct"))
+        self.assertIsNone(parse_rfc2822date("Lug, 24 Nod 2017 10:11:35 +0000"))
+        self.assertIsNone(parse_rfc2822date("Tue, 99 Oct 9999 99:99:99 +9999"))