Open jordens opened 4 years ago
I have added a conda package to hydra, and checked that it installs fine with its lasersdk
dependency on a fresh miniconda install, with the modifications below to install-with-conda.py
.
https://git.m-labs.hk/M-Labs/nix-scripts/commit/961dcac63d587fdb8a389ccef5585acb53c0fa79
Do you insist on the gitlab pipeline stuff?
CONDA_PACKAGES = [
"artiq",
"artiq-comtools",
"toptica-lasersdk-artiq",
]
...
if ADD_CHANNELS:
run("conda config --prepend channels quartiq")
run("conda config --prepend channels m-labs")
run("conda config --prepend channels https://conda.m-labs.hk/artiq")
run("conda config --append channels conda-forge")
...
I need a CI that runs at leas the current set of tests. And the conda packages need to work. Gitlab pipeline or nix, I don't care. It just needs to work.
Is there any problem with the conda package from hydra on conda.m-labs.hk?
The test are run (https://nixbld.m-labs.hk/build/20804/nixlog/1) except that the LASER
environment variable isn't set. Is there a sensible value for it considering that I don't have TOPTICA lasers?
I haven't tested the conda packages from hydra.
That's fine then. Importing the code already covers some very basic things like https://github.com/quartiq/highfinesse-net/issues/1. But the README needs updating. And the conda recipe also doesn't work anymore I'd think. And the gitlab CI definition should go as well if we move this over to nix. And the conda packages in the quartiq channel need some plan to phase them out. Also opticlock needs updating.
Importing the code already covers some very basic things like quartiq/highfinesse-net#1.
It actually does not cover it; Python only reports that error when it actually tries to create the Server.
Yes
https://gitlab.com/quartiq/lasersdk-artiq/commit/f96d0e81aa47beb468fa4d73d92dad0872d8a960/pipelines?ref=master