You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When you download the current master as zip archive from github.com, assigning a version number fails (empty string). As a consequence, packages depending on pandas (e.g. statsmodels) will fail to build as well.
I assume, the reason is the following: The version number is determined using git, but since the downloaded archive is not a git repository, this action fails and the version number remains an empty string.
I dont't think, this should be changed, but it would be helpful, if you added a node in the installation instruction, e.g. for people intending to install pandas on a machine without internet connection (like me ;)).
The text was updated successfully, but these errors were encountered:
@Funkensieper sorry for the long delay. You're completely right, that's bad behaviour.
Will merge PR starting the next release cycle in a few days, which will give you
a version string of the form "0.13.0.dev" if git (or .git) aren't available.
When you download the current master as zip archive from github.com, assigning a version number fails (empty string). As a consequence, packages depending on pandas (e.g. statsmodels) will fail to build as well.
I assume, the reason is the following: The version number is determined using git, but since the downloaded archive is not a git repository, this action fails and the version number remains an empty string.
I dont't think, this should be changed, but it would be helpful, if you added a node in the installation instruction, e.g. for people intending to install pandas on a machine without internet connection (like me ;)).
The text was updated successfully, but these errors were encountered: