Python 3.12 raises deprecation warning related to `datetime.utcnow()`

[FelixSchwarz]

Python 3.12 raises deprecation warning related to `datetime.utcnow()`

[alex]

How fun. Is the suggested solution something that works on Python 3.7?

…

On Thu, Jul 6, 2023, 4:41 PM Felix Schwarz ***@***.***> wrote: Using datetime.utcnow() triggers a deprecation warning in Python 3.12. This leads to test failures in projects using PyOpenSSL (e.g. josepy <https://github.com/certbot/josepy>). _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ self = <OpenSSL.crypto.X509 object at 0x7f57096df2c0> def has_expired(self) -> bool: """ Check whether the certificate has expired. :return: ``True`` if the certificate has expired, ``False`` otherwise. :rtype: bool """ time_bytes = self.get_notAfter() if time_bytes is None: raise ValueError("Unable to determine notAfter") time_string = time_bytes.decode("utf-8") not_after = datetime.datetime.strptime(time_string, "%Y%m%d%H%M%SZ") > return not_after < datetime.datetime.utcnow() E DeprecationWarning: datetime.utcnow() is deprecated and scheduled for removal in a future version. Use timezone-aware objects to represent datetimes in UTC: datetime.now(datetime.UTC). /usr/lib/python3.12/site-packages/OpenSSL/crypto.py:1437: DeprecationWarning =========================== short test summary info ============================ FAILED tests/util_test.py::ComparableX509Test::test_getattr_proxy - Deprecati... ======================== 1 failed, 193 passed in 0.51s ========================= — Reply to this email directly, view it on GitHub <#1228>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAGBHWMLZRKWJ7FNZWW6DXO4PIHANCNFSM6AAAAAA2A6ATNM> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[FelixSchwarz]

How fun. Is the suggested solution something that works on Python 3.7?

No, I don't think so but this works:

$ python3.7
Python 3.7.16 (default, May 29 2023, 00:00:00) 
[GCC 13.1.1 20230511 (Red Hat 13.1.1-2)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from datetime import *
>>> datetime.now(timezone.utc)
datetime.datetime(2023, 7, 6, 20, 52, 13, 270136, tzinfo=datetime.timezone.utc)

[alex]

Great. I guess we should use that everywhere. Are you interested in submitting a PR?

…

On Thu, Jul 6, 2023, 4:53 PM Felix Schwarz ***@***.***> wrote: How fun. Is the suggested solution something that works on Python 3.7? No, I don't think so but this works: $ python3.7 Python 3.7.16 (default, May 29 2023, 00:00:00) [GCC 13.1.1 20230511 (Red Hat 13.1.1-2)] on linux Type "help", "copyright", "credits" or "license" for more information. >>> from datetime import * >>> datetime.now(timezone.utc) datetime.datetime(2023, 7, 6, 20, 52, 13, 270136, tzinfo=datetime.timezone.utc) — Reply to this email directly, view it on GitHub <#1228 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAGBHYWJ6RYFDWJOJIQUDXO4QTDANCNFSM6AAAAAA2A6ATNM> . You are receiving this because you commented.Message ID: ***@***.***>

[reaperhulk]

The proposed solution creates a tzaware datetime, which can't be compared to a naïve. So the strptime parsed object will need to have a utc tzinfo added to it as well.

[FelixSchwarz]

The proposed solution creates a tzaware datetime, which can't be compared to a naïve. So the strptime parsed object will need to have a utc tzinfo added to it as well.

Yeah, I thought that was obvious. :-) My reply was intended to show that we can build a tz-aware "utcnow" even with Python 3.7.

Actually we could approach the problem in the code posted at the top in two ways:

• compare tz-aware datetime instances (basically use datetime.strptime(time_string, "%Y%m%d%H%M%S%z"))
• strip tzinfo from a tz-aware "utcnow" (datetime.now(timezone.utc).replace(tzinfo=None).

I am in favor of the second approach because self.get_notAfter() returns the "ASN.1 TIME" format which I think is supposed to always have a "Z" at the end. If we use %z as a format string for .strptime() we would accept also other formats.

Great. I guess we should use that everywhere. Are you interested in submitting a PR?

#1229

[FelixSchwarz]

PR merged, closing.

[alex]

Thanks!

…

On Fri, Jul 7, 2023 at 8:31 AM Felix Schwarz ***@***.***> wrote: Closed #1228 <#1228> as completed. — Reply to this email directly, view it on GitHub <#1228 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAGBGNCZFRDNFYWNSUTIDXO76SXANCNFSM6AAAAAA2A6ATNM> . You are receiving this because you commented.Message ID: ***@***.***>

-- All that is necessary for evil to succeed is for good people to do nothing.