chore: move from bingo to go tool

[bernot-dev]

This change moves tool dependencies to go tool instead of bingo.

Bingo will no longer be required.

[bwplotka]

Happy to remove bingo fully too.

Looks like some tests to fix still?

[bernot-dev]

I think it would be nice to not depend on Bingo, if go tool can meet our needs.

However, the latest version of Helm seems to have newer Kubernetes dependencies that forced a bunch of stuff to bump up, and broke things. I'll probably park this for now, since it's not a top priority.

[bwplotka]

This change moves the Helm dependency to a go tool instead of bingo.
As a result, bingo will no longer be required on local dev environment to run non-docker manifest generation.

Can we remove the use of bingo in hack/Dockerfile to test if we accomplished that with helm removal?

I also wonder why we needed to do this get if we use ${HELM} that should not need bingo get but instead use variables.env which will auto-build if needed.

Still helm build was taking long, so maybe it's faster with go tool plus we could remove bingo all together, so good for consistency.

Second question is around go module per tool. I'd prefer a single tool go module so we can minimize vendor and complexity (might be relevant for vendoring step in Louhi). With multiple go modules per tool we kind of build bingo again from scratch, manually 🙈

So to suggestions:

• Let's try one tool go mod
• Let's actually remove bingo get to see if things works (perhaps it worked without helm move 🤔 )

[bwplotka]

Let's go, thank you for helping!

[bwplotka]

Unrelated change, but low risk, so fine 👍🏽