Fix Openssl3 + py37 on windows failure

[akrherz]

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• Reset the build number to 0 (if the version changed)
• Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• Ensured the license file is being packaged.

Adds a test showing the openssl3 + python3.7 failure. Will iterate on a patch to then fix this.

Fixes #516

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[akrherz]

Great, the failure reproduced only on windows + openssl3

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "D:\bld\python_1634123290171\_test_env\lib\ssl.py", line 589, in create_default_context
    context.load_default_certs(purpose)
  File "D:\bld\python_1634123290171\_test_env\lib\ssl.py", line 490, in load_default_certs
    self._load_windows_store_certs(storename, purpose)
  File "D:\bld\python_1634123290171\_test_env\lib\ssl.py", line 482, in _load_windows_store_certs
    self.load_verify_locations(cadata=certs)
ssl.SSLError: unknown error (_ssl.c:4034)

Patch iteration coming in a bit

[akrherz]

My blissfully ignorant plan here is to iterate over cypthon 3.8 branch patches in this space and see when the failing test hopefully starts to pass. There are some larger patches that I am hoping to avoid. This first one may improve the logged error message to provide a better clue to the issue.

[akrherz]

Whoa, win_64_openssl3 is green! Will bump build number and un-draft this once the other builds are green.

[akrherz]

OK, this should be ready to go. My understanding of why this patch fixes the issue is that the ASN1 reader was incorrectly erroring when it found an unexpected EOF on the bytearray, but it should have been OK. This code path is unique to windows.

[dopplershift]

Wow, nice work @akrherz !

[akrherz]

Greeeeeeeeeen!!!

fixes #516

[jakirkham]

Woot! 🎉 Thanks @akrherz! 😄

Would it make sense to add the urlopen example from issue ( #516 ) as a test as well?

[akrherz]

Would it make sense to add the urlopen example from issue ( #516 ) as a test as well?

I don't believe so as that would introduce a network availability requirement for the test to succeed. I added a test that exercises the same failing codepath.

[jakirkham]

@conda-forge/python, could someone please take a second look at this?

[jakirkham]

cc @chenghlee @h-vetinari (for awareness)

[jakirkham]

Thanks Isuru! 😄

[github-actions]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• drone: passed
• travis: passed
• azure: passed

Thus the PR was passing and merged! Have a great day!

[xylar]

@akrherz, thanks very much for fixing this!

[h-vetinari]

Thanks a lot for tackling this (so far I only responded with emojis...)! :)

Unfortunately, this now leads to a hang in the cryptography test suite. I'm in the process of opening a new issue, just thought I'd let people in the thread know.

Edit: #518