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
{{ message }}
Repository navigation
assert_never raises different exception than documented #1857
raiseValueError(f"An unhandled Literal ({inp}) in an if/else chain was found")
else:
raiseexc
For comparison. there is an assert_never function in Python's typing module since 3.11 (and in typing_extensions). This function is similar but only has one parameter--which the first parameter of git.types.assert_never corresponds to--and it documents that it raises an exception if called at runtime, but does not document the type of exception raised. At least currently in Python 3.12, it raises AssertionError.
The GitPython function is a public member of git.types, and it is used internally in GitPython. I don't know if the docstring should be fixed to describe the existing behavior, or if that behavior should itself be considered the bug here and changed to match the docstring.
It does seem a bit strange to me that functions related to testing are part of the git public API in the first place, but it's probably nothing that can truly be fixed anymore.
With that said, even tough there mere existence could be considered a bug, I tend to leave it and its behaviour alone while fixing the docstring.
Activity
Byron commentedon Mar 5, 2024
Thanks for letting me know.
It does seem a bit strange to me that functions related to testing are part of the
gitpublic API in the first place, but it's probably nothing that can truly be fixed anymore.With that said, even tough there mere existence could be considered a bug, I tend to leave it and its behaviour alone while fixing the docstring.