Phillip Schanely

Thank you for the report!! I think it's more than reasonable to disable side-effect detection during module import - I will do this. (it has been surprising to me to...

Yeah, while there are plenty of sensible limits to put on code under analysis, I'd certainly expect anything and everything to be import-able. Now released in 0.0.27 - please reopen...

Thank you for following up! Indeed, with your extra info I was able to reproduce (looks like importing torch is enough to make this crash happen for me). This problem...

Ok - v0.0.28 is released now and rips out the logic that causes trouble in your example. (I was able to reproduce approximately the same crash). But I'm still not...

The remaining ones may require even more capabilities. I'm not immediately prioritizing them. For instance, I'd like to simply disable `lru_cache` and friends, but it seems like we'd need to...

I figured out a workaround for disabling the caching behaviors of `lru_cache` and `cache`: we dynamically detect the wrapper functions at runtime and skip the caching bits. Implemented in 2ab5105aba83e8b26a40e9a1362a946fc8b13644...

Early experiments are look promising. Will wait for a few more betas/RCs before tying up all the loose ends.

What a neat idea! I'm also musing about a hybrid approach: if a good repr exists, it's nice to have something that's trivially copy+pasteable into an repl / debugger /...

Some exploration (and interesting CrossHair fixes related to `rich`'s dynamic imports) has happened. But there's a bit of an onion to unravel here: I'd like to make a few improvements...

Thanks! This is a really interesting idea. As a first step, perhaps experiment with a few examples; you're certainly on the right track with `StateSpace`. It should be easy to...