-
* Remove the strict channel priority from condarc When using miniconda we had two channels named default and conda-forge. Now when using miniforge, the only channel is conda-forge. We had observed that cloning a new environment in JupyterHub failed due to this strict channel priority being in place. We therefore wish to remove this strict policy in order to address this failure to clone. Bug: T372417 * Stop using the classic solver in conda-analytics-clone We require the conda and conda-libmamba-solver packages to be present in environments that are created by the conda-analytics-clone script. However, the default behaviour of the `conda clone` command is to exclude these packages. We have a step that manually installs the packages from the local cache when clong the environment. Until today, we have had to specify the use of the 'classic' conda solver when installing these two packages, but it seems that the newer version of miniconda (or miniforge) no longer requires this workaround and fails if we try to use it. Bug: T372417 * Switch from miniconda to miniforge Signed-off-by: Balthazar Rouberol <brouberol@wikimedia.org> Bug: T372417
7d6b59ad