[smeagull]

The real issue is that you can't have multiple versions installed at the same time.

if you could import numpy==3.2 it'd solve many problems.

[thinker5555]

Granted, I'm more of a hobbiest than a dev, but I think this is part of the problem that virtualenvs are supposed to help solve. One project (virtualenv) can have numpy==3.2, and another can have numpy==3.1. Maybe I'm naive, but it seems like having a one project with multiple versions of numpy being imported/called at various times would be asking for trouble.

[Spivak]

The thing you can’t do is solve for this situation.

    A depends on B, C
    B depends on D==1.24
    C depends on D==2.02

There should in an super ideal world be no problem with this. You would just have a “linker” that inserts itself in the module loader that presents different module objects when deps ask for “D” but it hasn’t happened yet.

[goodpoint]

That's a feature, not a bug.

[smeagull]

It's really not. I can't install tensorflow more than 10 times before I'm done. Virtualenvs are not fit for purpose.