stuartarchibald

Closing this question as it seems to be resolved. Thanks!

@AndrewEckart thanks for creating a draft for this. The way to do: > The str overload should fallback to __repr__ when __str__ is not defined. Seems like this might require...

> @AndrewEckart thanks for creating a draft for this. The way to do: > > > The str overload should fallback to **repr** when **str** is not defined. Seems like...

CC @DrTodd13

> Would the fact that AOT generates platform-specific machine code be a hurdle here? At least, it would clash with an ideal of JITing being a means of avoiding shipping...

> @stuartarchibald could you give us a pointer to how to solve this issue? can `ctypes` be used here? @bairdlampard534 I think it would requires changes to the way `pycc`...

> Thanks for the insight. What would be the fundamental difference between the output of caching vs. AOT compilation? Caching generates `nbi` and `nbc` files. Can AOT compilation produce something...

Thanks for the report. How to run tests is extensively documented in the developer/contributing part of the docs: https://github.com/numba/numba/blob/60723cb5a31d0b6ef516af95602a463322991420/docs/source/developer/contributing.rst?plain=1#L153-L211 I think a small addition to the installation docs explaining "this...

This is a start on documenting the sorts of checks that could be performed when reviewing NumPy/CPython feature implementation/changes. Outstanding issues: 1. Need to set up the project board on...

@apmasell Thanks for the patch. Was wondering if it would be preferable to expose remarks information in the function's per-signature metadata dictionary? I think it'd be possible to collect information...