Shubhamkar Ayare

If a function has free (lexical) variables, then it is a closure, right? Am I missing something? Any other way you know of this could be done?

Hmm, true. Thanks for the pointers!

TIL: There is `(iterate::free-variables form)` that could be "light-weight" (and may be ~~somewhat~~ pretty restrictive) if someone is already using iterate elsewhere.

Yup, been using hu.dwim.walker, but it accounts for about half the size of the files generated in the ~/.cache folder. No proper alternative until implementations could emit a special condition...

Let me think if there's a better way to do this. `raw-py` anyways exists only for backward compatibility; for anything non-trivial, the rest of the interface through pycall, pyvalue, etc...

> other methods silently discard Python side issues such as `pymethod` I can confirm that there definitely seems to be some bug. I will try to get it fixed soon....

9c65f0eea1389796bc94ef126ce023ddf578caec should fix the pymethod not raising exceptions issue.

ed6f6f8dccb224220f0145fa5495acb235f5c50d should add support for capturing the python error even when raised in raw-py. I have incorporated your commit about importing sys and traceback before numpy, but I felt it...

> I do not believe it matters as long as it is consistent for the respective block. As long as we can detect it correctly, it should indeed be okay;...

I should rewrite that TODO as just "optimization" (EDIT: Done in https://github.com/digikar99/py4cl2-cffi/commit/163e52b9d63a9d4aa2d8f4d40ce7e31a4841eaf7). Apparantly, PyCall.jl is just 1.5 times slower* than native cpython; even though py4cl2-cffi might be 10 times faster...