How does it work for reproducing scientific experiments using legacy Python 2 code?
All of the major scientific libraries and core packages are dropping python 2 support. Are there legacy versions of these packages being maintained or at least preserved?
Maybe if you're reproducing a scientific experiment you should use the original package versions? New releases of pip or scientific libraries won't support Python 2, but the old versions should still work. And you won't be running different code from using up-to-date versions.
> How does it work for reproducing scientific experiments using legacy Python 2 code?
Install Python 2.7, which is by does with pip. Upgrade pip if you want, which will get you the last version of pip that supports Python 2 (the old versions haven't gone away, new versions just don't support py2 and won't show up when upgrading on py2). Install the needed scientific libraries. Run code.
> Are there legacy versions of these packages being maintained or at least preserved?
AFAIK, there is no plan to remove py2 packages from pyPI.
I assumed that the old pip versions would keep working, as long as the python2 packages kept being hosted, and this is just freeing new pip development from being python2 compatible. Does anyone know if this is the case?
If not, itβs a good argument for vendoring dependencies when publishing scientific code.
All of the major scientific libraries and core packages are dropping python 2 support. Are there legacy versions of these packages being maintained or at least preserved?
Is this even a valuable use case?