kodonnell

OK, MRE: ```python from persistqueue import Queue class Point: def __init__(self, x, y): self.x = x self.y = y def __repr__(self): return f"{self.x}|{self.y}" @staticmethod def dump(value, f): print("dumping:", value) f.write(f"{value.x}|{value.y}\n".encode('ascii'))...

Nice @nickdrozd - can you give the exact programs/commands used? I think I tried to replicate your graphs for some of the stuff I was working on, but couldn't figure...

So, I keep coming back to performance, so here are some ideas/summaries since that's what this issue is for. - #529 - I filed [this](https://github.com/PyCQA/pylint/issues/2912) in pylint, but may be...

(Also, it might pay to cross-post the above into pylint as well, as that's generally where users care most about performance of astroid.)

Another question - has anyone considered parallelism for performance, etc.? Since we're largely CPU bound (?), I don't think we'd get much benefit from standard threading (GIL etc.) so we'd...

Aside from custom colors/maps, it was brought up in #58 that supporting these *per feature* might be a requirement. Easy to do - just nailing down the API is important.

OK, figured it out - [here's](https://github.com/microsoft/MLOpsPython/blob/cbeadae5b53409f71ff69a34c27ed4aa1e6606da/ml_service/pipelines/diabetes_regression_build_train_pipeline.py#L110) the culprit. If the training step is re-used, then the `run.parent.log` will never get called, and hence the issue I'm having will always occur....

Ah, right. I'm certainly no expert (and couldn't find an exact answer) but you look to be right. Also, to clarify, I can run cythonised code from a pure-python service...

Nice! Let's keep this open as a way to cross-pollinate between them. Can you update us if you implement any cool new features?

Nice - I've always thought it'd be good to create a specification file which lists all of these assumptions/findings. Maybe you could start one with the above? Encoding was a...