Investigation: ACCESS-OM2-BGC Versioning

[CodeGat]

Some questions that came up on slack regarding the spack.yaml of this repository:

• Should the name of the spec be access-om2 like the existing https://github.com/ACCESS-NRI/access-om2 repo? Could lead to clashes in the vk83 spack environment. See

  ACCESS-OM2-BGC/spack.yaml

  Lines 10 to 11 in af942a8

  	specs:
  	- access-om2@git.2023.11.23

  . I'm not too sure about remediation actions for this one (@harshula ?) but it might require the creation of a new package in spack-packages? Also note that a change to the package name will mean the removal of the -bgc bit here:

  ACCESS-OM2-BGC/spack.yaml

  Line 59 in af942a8

  access-om2: '{name}-bgc/2023.11.23'

• Should the versions of access-om2 and access-om2-bgc be in lockstep? Currently, both have the version @git.2023.11.23. Would an update to one of the models require an update to the other, or will they drift out of sync? If it's the latter, we should start access-om2-bgc at @git.2024.03.0.

And just for my own understanding:

• What is the type syntax here?

  ACCESS-OM2-BGC/spack.yaml

  Line 16 in af942a8

  require: '@git.2023.11.09 type=ACCESS-OM-BGC'

Tagging @aidanheerdegen as well

[aidanheerdegen]

Soz. I've pre-empted this with a forum topic, in case this repo "goes away"/"gets taken"

https://forum.access-hive.org.au/t/access-om2-biogeochemistry-builds-and-deployment/1922

[CodeGat]

This has been kinda solved:

• The first point is solved because we are using the new access-om2-bgc package here: Add ACCESS-OM2-BGC Package spack-packages#81.
• The second point is kinda solved because using a CalVer-like format means that updates will probably be mostly in lockstep.