python -m ipykernel install --user --name=myenv fails to create kernel

After creating an environment with something like "conda create --name myenv -y -c anaconda -c conda-forge python=3 ipykernel pandas matplotlib," then installing the kernal with "python -m ipykernel install --user --name=myenv," the "~/miniconda3/share/jupyter/kernels/myenv" folder doesn't get created. When I create a new notebook using myenv, I cannot import any libraries. I just figured out that if I copy the python3 folder to myenv, I could then import libraries in the notebook. This just started happening yesterday, ipykernel 5.3.4.

Reproduce the issue with the conda environment and the `python -m ipykernel install --user --name=myenv` command from the report. Inspect whether `~/miniconda3/share/jupyter/kernels/myenv` is created and verify that a notebook using `myenv` can import the installed libraries; done means the kernel is registered correctly without copying the `python3` folder.

Written by the indexing model from the issue text.