-
-
Notifications
You must be signed in to change notification settings - Fork 1.9k
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
Cannot freeze pip version within the virtualenv #4111
Comments
@techalchemy I'm not sure where this behaviour is coming from, but I can confirm that I got bitten by this last week, as I had an old version of |
While attempting to work around this, I found that
|
Recategorising as an enhancement request to offer a way for users to decide for themselves which set of risks they want to tolerate ( |
What about pinning them if they are explicitly stated in the Pipfile? Assuming users want to pin them in that case. |
@gsemet pipenv vendors in specific version of pip that is part of the project -- so depending on the version of pipenv that you have installed you will have a specific version of pip. setuptools and wheel are its own beasts because pip and requirementslib and maybe other things choose to install the latest versions during resolution and locking cycles and that is very uncontrollable--I've tried recently when setuptools released a new version that broke a test of ours to pin it, but it wasn't really possible and was easier to realize what the actual bug was and fix that to work with the new version of setuptools. I live in fear of new setuptools major version updates after it ruined that weekend of mine. |
At this point in time |
Hi.
I do not find a way to freeze the pip version installed within the virtualenv (like
pipenv install pip==19.3.1
). The latest version as of today (20.0.1) has a bug that is cannot install one package (azure-iothub-device-client==1.4.6
), while the previous version works.pipenv seems to force installing to the latest version WITHIN the virtualenv, no matter what we put in the Pipfile. Same I see some of these packages (pip, setuptools) are NOT locked, so when they evolves and introduce a regression, the build fails for no reason.
I strongly recommend to treat these packages like every other ones, ie, allowing user to freeze them to ensure the build is fully reproductible, no matter what happens on pypi.org.
Upstream bug opened: pypa/pip#7644
The text was updated successfully, but these errors were encountered: