linkedin
Regression in support for Python 3.6 as well as 3.7 in 1.0.7 release
As of the latest release on PyPi, I have observed that this package indicates support for Python >=3.6 in its package metadata, however is presently only compatible with Python versions 3.8 or later after the introduction of commit https://github.com/linkedin/shiv/commit/377c70db139a7e1fc5070a2a294d1fcd731dfc55
Therefore, would it be possible to either:
- Increment the minimum Python to 3.7 (or later) per https://github.com/linkedin/shiv/pull/256
- Revert the incompatible changes to the type annotations introduced in the above commit
Maybe I'm missing someting in my setup but it doesn't work in my 3.8 setup either.
The type hints in the environment.py file from the commit indicated in the previous comment makes it impossible for me to run shiv at all.
A possible work-around (it works in my, so-far limited, testing) is to add the line:
from __future__ import annotations
as the first code line in environment.py.
(as mentioned here: https://stackoverflow.com/a/75317316)
Thank you for reporting this. I have a PR for fixing the types which changes dict to Dict.
In subsequent PR's I am also planning to address the incorrect >= 3.6 and change it to >= 3.8 to be consistent with the tests that are provided.
Finally, I am planning to add the CI pipeline to PRs so that the validation is done earlier in the process and issues can be found.
Version https://github.com/linkedin/shiv/releases/tag/1.0.8 has been released which should fix the regression for Python 3.8.