-
-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
test_bdist_egg fails with Distribution has no attribute entry_points (Python 3.6+) #836
Comments
Checking the CPython logs, I don't see anything there that I would have expected to trigger this behavior. |
I've decided to go back to the last successful build, and I selected this job which was passing and restarted it. If it passes, then we know the issue is with the setuptools code, but if it fails, then we know something changed with the CPython code in Travis. |
And it fails, so the issue must be related to CPython code having changed on Travis. |
The Travis output doesn't give enough information about which version of Python was being used.
|
I'm unable to troubleshoot because of the issue with tox rooted in a long-standing issue with virtualenv. |
Bypassing tox and using rwt, I was able to replicate the failure:
Interestingly, the failure doesn't happen if the tests are downselected:
Therefore, the failure is almost certainly an interaction with other tests. |
Passing |
I can bypass the failures with just these ignores: |
Okay, this is interesting - disabling doctest as so,
I'm able to work around pytest-dev/pytest#985 and get a pdb prompt at the failure (even though py36compat is no longer being tested. |
It appears as if Distribution.parse_command_line is implicated.
|
This gets weirder the deeper I dig.
It looks like self.entry_points is getting cleared by accessing self.display_options or self.negative_opt. I don't see anything else there that could be affecting |
Actually, it seems like it's the setting of
|
I'm starting to think there's an inherent flaw in the data model of Python:
|
Yes, there's definitely something wrong with our dear dictionary:
|
Given that the issue appears to be a Python bug, I'm going to bisect the changes between here and there to identify the offending commit, looking primarily at commits between when the tests last passed and when the tests started failing. |
The issue begins with |
In the last two days, the test_bdist_egg test has started failing on Python 3.6 dev and Python nightly (3.7):
The text was updated successfully, but these errors were encountered: