luk-f-a

I could really use such a feature. In case you are interested in a concrete use for this type of arrays, I am working with (time inhomogeneous) Markov Chains. The...

thanks for the explanation, and warning for deprecation (I just built a lot of components using that approach). Does the widget approach allow for creating custom props? I was able...

are you saying that those use cases (custom props and events) will remain possible only via `components` and won't be deprecated? Is it then worth documenting them here? ie, changing...

@stuartarchibald @sklam I made some progress today on finding the location of overloads. It's a bit convoluted, which gives me pause. Please have a look and let me know if...

@stuartarchibald thanks for the pointer on how to get the file information. Do you think it's ok that all templates are "logged" as (potential) dependencies? That seems easily doable with...

@stuartarchibald @sklam This is the second example. ```python from numba import njit, guvectorize import numpy as np from numba.core import types from numba.np.ufunc.wrappers import build_gufunc_wrapper @guvectorize(["(intp, intp[:])", "(float64, float64[:])"], "()->()",...

I've looked into this issue and it might require #2823 to be able to interact with via cffi with the new code.

nothing beyond this: https://numba-how-to.readthedocs.io/en/latest/numpy.html#using-the-new-1-17-numpy-random-generator

if you managed to solve the problem with `npydecl.registry` that would make `test_parfor.test_var` fail (describe here https://github.com/numba/numba/pull/4480#issuecomment-629698098), maybe this PR unlocks #4480

@stuartarchibald thanks a lot for the code example. I'm testing a new approach in #8396 but I'll keep this PR for experimenting with this type inference without lowering. I couldn't...