Accomodate Hatch use case

[flying-sheep]

Hatch currently monkeypatches propose_interpreters to apply its own filter logic: https://github.com/RRosio/hatch/blob/ecff6e6f3fff42efbd30866fb64e6273c93774d7/src/hatch/env/virtual.py#L378-L398

This broke because virtualenv moved the functionality to this package.

Would it be possible to add a filter callback or so to get_interpreter to make Hatch’s implementation fully based on public APIs, or is there another way to achieve the same?

[gaborbernat]

Can you detail how would you like this to look and behave? Thanks!

[flying-sheep]

Added a link to the code. It’s simply:

def _patched_propose_interpreters(*args, **kwargs):
    for interpreter, impl_must_match in propose_interpreters(*args, **kwargs):
        if not self._interpreter_is_compatible(interpreter):
            continue
        yield interpreter, impl_must_match

so a straightforward way would be to replace all this with

-[monkeypatching code]
-python_info = virtualenv_discovery.get_interpreter(python_version, (), env=self.get_interpreter_resolver_env())
+python_info = virtualenv_discovery.get_interpreter(python_version, (), env=self.get_interpreter_resolver_env(), filter=self._interpreter_is_compatible)

or maybe a variant where impl_must_match is also passed to the callback or so.

[flying-sheep]

Thank you!

[ofek]

Thanks a lot 🙇‍♂️