cli: add --workspace flag to set base directory for Constellation workspace

[daniel-weisse]

Context

The goal of this PR is to let users specify the Constellation workspace using flags.

For the longest time, our CLI allowed setting a path for the config and master secret files. Other files and folders however, were always assumed to be in the current working directory.
This PR aims to address this issue by adding the --workspace flag, which allows a user to set the workspace for a CLI command.
All actions on files will then be run from the given workspace instead from the directory the command is executed from.

Proposed change(s)

• Remove --config flag
• Remove --master-secret flag
• Add --workspace flag
  • Replaces --config and --master-secret flags by allowing a user to specify the base directory context from where CLI commands are executed
  • All files are created and accessed relative to the path given by the flag
  • If omitted, uses the current working directory

Additional info

Not sure if we can properly test the flag using the e2e tests, so I request the reviewers to try out their workflows with the new flag and report any issues that may occur.

I already did some manual testing on GCP. Currently missing testing for IAM migration, general updates and manual tests on Azure/AWS.

e2e test Azure
e2e test AWS

Checklist

• Update docs
• Add labels (e.g., for changelog category)
• Is PR title adequate for changelog?
• Link to Milestone

[netlify]

✅ Deploy Preview for constellation-docs canceled.

Name	Link
🔨 Latest commit	4fb92ed
🔍 Latest deploy log	https://app.netlify.com/sites/constellation-docs/deploys/64cce188226ed800087ae88d

[elchead]

nit: change the title to "cli: add --workspace flag in favor of --config and --master-secret". seems to make it easier for the user what the change means in the changelog. you agree?

In the description I'd add:
The new flag makes the CLI consistent with the workspace concept (https://docs.edgeless.systems/constellation/architecture/orchestration#workspaces)

[malt3]

If I understand this PR correctly, you update every use of file paths and prepend the workspace path to it.
While this works and is stable, I wonder why we do not follow the convention of other tools and perform os.Chdir(workspace) and keep all paths relative to the CWD?
I'm just trying to understand the tradeoffs and why this implementation was chosen.

EDIT: Example implementation

[elchead]

Had a rough look over the changes and lgtm. Maybe we can test the flag explicitly in our e2e_test (simply cd .. and then use the workspace flag?)

[elchead]

If I understand this PR correctly, you update every use of file paths and prepend the workspace path to it. While this works and is stable, I wonder why we do not follow the convention of other tools and perform os.Chdir(workspace) and keep all paths relative to the CWD? I'm just trying to understand the tradeoffs and why this implementation was chosen.

I think this would simplify the changes a lot and remove unneeded parameters. I think it's better to choose this if there is no other reason for the current way. Even thought it would be some additional work :/

[daniel-weisse]

While this works and is stable, I wonder why we do not follow the convention of other tools and perform os.Chdir(workspace) and keep all paths relative to the CWD?

Tried this, and found a good solution that works.
Downside being that this won't correctly print filepaths relative to the execution directory. But it shouldn't be too much effort to fix that.

[malt3]

Thanks! Works great and LGTM.
Feel free to add or dismiss the short flag -C.

[3u13r]

The one file under my ownership looks fine. Unassigned review.

[derpsteb]

Did some basic manual testing and looked at codeowned files.
👍