[trevor-scale]

Many secure environments disable the hashlib.md5 function because it is insecure, this causes the line below to fail making dateparser unusable in these environments:

dateparser/dateparser/conf.py

Line 52 in 02bd2e5

return hashlib.md5("".join(keys).encode("utf-8")).hexdigest()

This is remedied by replacing that line with return hashlib.new("md5", "".join(keys).encode("utf-8"), usedforsecurity=False).hexdigest() which appropriately bypasses the fips security check because dateparser is not using the md5 hash for security applications.

Corresponding issue: #1258

[ghmj2417]

I posted this in #1258 as well. Change usedforsecurity=True to usedforsecurity=False. The suggestion/example provided in issue 1258 was to use hashlib.new but it also included the full command which had usedforsecurity=True to test a failure when using FIPS compliant encryptions/hashes/etc. My apologies for not being clear.

[trevor-scale]

I posted this in #1258 as well. Change usedforsecurity=True to usedforsecurity=False. The suggestion/example provided in issue 1258 was to use hashlib.new but it also included the full command which had usedforsecurity=True to test a failure when using FIPS compliant encryptions/hashes/etc. My apologies for not being clear.

Thanks for catching this wasn't looking closely when I copy-pasted haha. Will update now!

[trevor-scale]

I posted this in #1258 as well. Change usedforsecurity=True to usedforsecurity=False. The suggestion/example provided in issue 1258 was to use hashlib.new but it also included the full command which had usedforsecurity=True to test a failure when using FIPS compliant encryptions/hashes/etc. My apologies for not being clear.

Thanks for catching this wasn't looking closely when I copy-pasted haha. Will update now!

@ghmj2417 are you able to approve and merge this, would love to get this added in!

[trevor-scale]

@Gallaecio can you give this a review and lmk if any changes are needed when you have the chance?

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.62%. Comparing base (c2ff31c) to head (f0b2675).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1267   +/-   ##
=======================================
  Coverage   96.62%   96.62%           
=======================================
  Files         235      235           
  Lines        2905     2905           
=======================================
  Hits         2807     2807           
  Misses         98       98           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Gallaecio]

Please, install pre-commit, run pre-commit install, then pre-commit run --all-files, keep running that command until everything is green (I think in this case a single run should suffice, but seeing everything green is nice 🙂), and commit those changes.

[ghmj2417]

hashlib.new was to accommodate Python 3.8. If setup.py is going to require >= 3.9 then hashlib.md5 is fine.

[trevor-scale]

Finally got the linting done! @Gallaecio are you ok to merge now?

[trevor-scale]

ghmj2417

/ @ghmj2417

[yannsartori]

I would be interested in this change as well, right now my workaround is to use unittest.mock.patch

import hashlib
from unittest.mock import patch

import dateparser

settings = ...

@classmethod
def patched_get_key(cls, settings=None):
    if not settings:
        return "default"
    keys = sorted([f"{key}-{settings[key]}" for key in settings])
    return hashlib.md5("".join(keys).encode("utf-8"), usedforsecurity=False).hexdigest()

with patch("dateparser.conf.Settings.get_key", patched_get_key):
    dt = dateparser.parse(s, settings=settings)