Makefile: make dev uses packer for install

[lbajolet-hashicorp]

When building the development version of the plugin, we now use
packer plugins install for that instead of moving it to the plugins
directory manually.

This target will require Packer 1.10.2 and above to function, as prior
versions don't support instaling pre-release versions of a plugin with
that command.

[wolfganghuse]

LGTM

[lbajolet-hashicorp]

I see an E2E test failing because the updated dev target needs packer installed for the process to succeed. Given this is a build test I would think invoking make build would be an acceptable replacement, alternatively we could also include the setup-packer action to install packer in the workflow's env.

I'll be happy to amend the PR however you think is more relevant, please let me know which direction's your preferred!

[tuxtof]

Hello @lbajolet-hashicorp
for now external action is blocked on the org
please amend with the first solution

[lbajolet-hashicorp]

Hi @tuxtof,

Should be good to go on the E2E front now! Had to change a few more things than just the target since there's some build tests later in the pipeline that do have Packer installed, so I updated the plugin installation to use the updated workflows.

Not sure if there's something to do for the black duck policy check to go green on my end, but please let me know if you need anything from me to fix that.

Thanks for the reviews!

[tuxtof]

@lbajolet-hashicorp i miss the fact that hashicorp/setup-packer was already authorized and used in some workflow
you can take this path , don't worry for the blackduck error , we have work in progress issue with

[lbajolet-hashicorp]

@tuxtof regarding make dev, since the workflows are split in two I'd think we don't need to add the setup-packer action to the build step, so long as the dependent step that does the tests is able to get the binary we're good to go.
I don't think installing Packer during the build step improves the workflow here, but please correct me if I'm wrong and I can amend the PR again if you prefer.

[tuxtof]

just check the entire PR more deeply :-D and that's OK for me