--workspace-id ignored on deployment commands: coalesceWorkspace() checks current-workspace context before reading the flag

[jremigio-circle]

Summary

With an organization-scoped API token (no workspaceId claim) used via ASTRO_API_TOKEN in a non-interactive/CI context, astro deployment variable update (and other commands routing through coalesceWorkspace) fail with:

Error: failed to find a valid workspace: failed to get current Workspace: current workspace context not set

EVEN WHEN --workspace-id <id> is passed. The explicit flag is effectively ignored.

Affected versions

v1.40.1, v1.42.1 (latest stable), and v1.43.0-nightly — the relevant functions are byte-identical across these, so upgrading does not help.

Root cause (code)

In cmd/cloud/workspace.go, coalesceWorkspace() calls workspace.GetCurrentWorkspace() and returns its error BEFORE it checks the --workspace-id flag:

func coalesceWorkspace() (string, error) {
    wsFlag := workspaceID
    wsCfg, err := workspace.GetCurrentWorkspace()   // called first
    if err != nil {
        return "", errors.Wrap(err, "failed to get current Workspace")  // returns here — flag never read
    }
    if wsFlag != "" { return wsFlag, nil }           // --workspace-id only honored after
    if wsCfg != "" { return wsCfg, nil }
    return "", errors.New("no valid Workspace source found")
}

GetCurrentWorkspace() (cloud/workspace/workspace.go) errors whenever c.Workspace == "". In the API-token path, checkAPIToken() (cmd/cloud/setup.go) derives the workspace from token claims, falling back to c.Workspace (if wsID == "" { wsID = c.Workspace }); an org-scoped token has no workspaceId claim, so in a fresh config home c.Workspace stays empty and the flag cannot rescue it.

Reproduction

1. ASTRO_HOME=$(mktemp -d), set ASTRO_API_TOKEN to an organization-scoped token, no astro workspace switch performed.
2. Run astro deployment variable update --workspace-id <valid-ws-id> --deployment-id <valid-dep-id> --key FOO --value bar.
3. Observe the current workspace context not set error despite --workspace-id.
4. Running astro workspace switch <ws-id> first makes it succeed.

Expected

An explicit --workspace-id should be honored before requiring a current-workspace context.

Suggested fix

In coalesceWorkspace(), return wsFlag when non-empty BEFORE calling GetCurrentWorkspace() (or have checkAPIToken seed the context workspace from --workspace-id for org-scoped tokens).

Related

• Improve message for the missed "workspace" option while using Org level tokens #1635 (symptom: improve the error message for org-level tokens)
• astro org switch still prompts for workspace selection when --workspace-id flag is provided #1962 (analogous --workspace-id ignored in org switch)

This issue is specifically about the coalesceWorkspace ordering for deployment/variable commands.

Impact

Breaks non-interactive CI flows that use org-scoped tokens and pass --workspace-id explicitly.

[github-actions]

This issue is stale because it has been open for 60 days with no activity.