Create arm-create-package.yml

[anthonydahanne]

This PR updates the pipeline builder create-package workflow automation to support multi-arch builds of this buildpack (ARM64 and AMD64). The ARM64 image should be considered experimental for now, since the official Paketo RFC and CNB RFCs related to multi-arch have not been merged in yet.
It uses an experimental version of the pack CLI (buildpacks/pack#2086), and is based off of the upstream CNB RFC that is currently IN PROGRESS. This workflow will definitely be subject to change when an official pack release comes out, and if there are any changes to the upstream RFC.
Once we have merged this and seen it work on main, we will contribute it into the upstream pipeline-builder repository.
(@sophiewigmore )

Checklist

• I have viewed, signed, and submitted the Contributor License Agreement.
• I have linked issue(s) that this PR should close using keywords or the Github UI (See docs)
• I have added an integration test, if necessary.
• I have reviewed the styleguide for guidance on my code quality.
• I'm happy with the commit history on this PR (I have rebased/squashed as needed).

[anthonydahanne]

this should be moved under pipeline builder, right @dmikusa ?

[dmikusa]

Yes, this is going to be a pipeline-builder change.

[dmikusa]

We've implemented some things in pipeline-builder, is this PR still needed?

[anthonydahanne]

@dmikusa well, we don't "need" it, it's more like a playground to test arm64 package publication.

AFAIK, pipeline-builder does not support generating such a workflow, so this workflow should go in pipeline-builder and be generated as pb-create-package-arm64.yml when a

package:
  arm64: true

property is set.

Also, it won't be the end of it, we'll also need another pb-merge-manifest.yml that will push the dual arch buildpack

[dmikusa]

Gotcha, OK.

[sophiewigmore]

Per convo with Anthony - we'll prob need to update some stuff in the buildpack.toml around the arch field for the dependencies. We will need libpak changes to work with that too, most likely.