ngam

> Feel free to open an issue on the JuliaPkg repo to discuss the order to search for Julia. It seems off-topic for the present PR. Of course it is...

> > Nevertheless, JuliaPkg and JuliaCall are functioning Python packages which can deal with their own Julia dependencies, so distributing them as-is via Conda is fine, no? I appreciate they...

I am fine with merging this, but not the other yet. There is no reason to call this `pyjuliapkg` as `juliapkg` would suffice. In the future `julia-pkg` will be the...

> Also there is no JuliaCall.jl. That package is PythonCall.jl and would presumably be called julia-pythoncall in conda-forge Wouldn't it be awesome if cjdoris wrote a Julia package just to...

> Anaconda package conda-forge* > what to put there to disable PyPy builds Under build section, use `skip: true # [condition]` like below ```yaml build: skip: true # [python_impl ==...

xref https://github.com/conda-forge/nuitka-feedstock/pull/156

> As discussed in the mailing list central_diff_weights and derivative should be deprecated and completely removed in the future. Within SciPy, misc.derivative is used internally only in scipy.stats.rv_continuous [here](https://github.com/scipy/scipy/blob/e3cd846ef353b10cc66972a5c7718e80948362ac/scipy/stats/_distn_infrastructure.py#L2069). I'm...

I see, thanks for the background information and links! The plan (obviously) makes sense. I have heard about findiff and numdifftools, but I do believe a differentiate/derivatives scipy modulde/subpackage would...

Issue: [Many packages use the environment variable `CUDA_HOME` to point to the root of where the tools used to be](https://github.com/search?q=org%3Aconda-forge%20CUDA_HOME&type=code). Question: What's a good replacement for this now? I suppose...

Another question: > 2\. There are other components users would like access to ( [ptxas executable cudatoolkit-feedstock#72](https://github.com/conda-forge/cudatoolkit-feedstock/issues/72) ) Where is `ptxas` located now? We need it for tensorflow and jaxlib.