GitLab

MNT: Revert to previous logic - create full environment on each deploy job

See merge request fsl/conda/fsl-ci-rules!136

job. Hopefully can use quicker approach soon

MNT: use py3

See merge request fsl/conda/fsl-ci-rules!135

Mnt/python

See merge request fsl/conda/fsl-ci-rules!134

Bf/clone

See merge request fsl/conda/fsl-ci-rules!133

MNT: Skip environment creation for deploy jobs, as it is really slow and unnecessary

See merge request fsl/conda/fsl-ci-rules!132

unnecessary

ENH,RF: Pin core/base C++ dependencies at build time, using a conda_build_config.yaml and base_deps.yaml file.

See merge request fsl/conda/fsl-ci-rules!131

want to re-list fsl-base=last-stable-version from fsl-release.yml

to prepare package specs for use as CLI args

dependency, and pinned to specific version. This could alternately be added to
the projects that use binary components of boost (currently only ciftilib and
mist I think?)

based on the recipe type. Only pin C++ requirements for compiled projects.
Only pin numpy/python if they are listed as host requirements.

conda_build_config.yaml and base_deps.yaml file.

MNT: Ensure that correct GCC is installed.

See merge request fsl/conda/fsl-ci-rules!130

latest available on conda-forge, so the catch-all clause in previous code
should take effect, and I shouldn't need to rebuild the images.

Mnt/cuda11.4 1.5

See merge request fsl/conda/fsl-ci-rules!129

Mnt/cuda11.4 1.5

See merge request fsl/conda/fsl-ci-rules!128

using this feature at the moment

used in a rules: if: statement, as gitlab says the YAML is invalid

Bf/update dependants

See merge request fsl/conda/fsl-ci-rules!127