fix(tooler): detect both legacy docker-compose and the docker cli compose plugin

[smaEti]

Fixes

• dockercomposetooler.Gauge now detects the Docker CLI Compose plugin in addition to the legacy docker-compose binary

Context

Gauge previously called AnyOneExecChecker(ctx, "docker-compose", "docker compose"), which resolves to exec.LookPath on each name. The "docker compose" arm could never succeed, the modern Compose ships as a Docker CLI plugin under Docker's plugin directory, not as a standalone binary on PATH. The check only passed incidentally on systems that also expose a docker-compose shim.

Changes

• internal/tooler/dockercomposetooler/tooler.go: replace the single AnyOneExecChecker call with explicit fallback:
  • ExecChecker("docker-compose"): legacy binary (unchanged behaviour)
  • else ExecChecker("docker") + docker compose version: actually invokes the CLI so the plugin dispatch can resolve (mirrors dockerswarmtooler)
  • else errors.Newf(errors.TypeNotFound, …) with a descriptive message

Test plan

• System with only legacy docker-compose — passes
• System with only the docker compose plugin — passes
• System with neither — returns TypeNotFound with a clear message

[Nageshbansal]

Hey @smaEti, can you push a small commit (or an empty git commit --allow-empty -m "ci: retrigger" && git push) to retrigger CI? The postci labeler job failed due to a stale event payload, and a fresh push should clear it. Thanks!

[Nageshbansal]

Hey @smaEti, thanks for this. It's approved, and the code checks are fine. The only thing blocking the merge is our branch protection requiring signed commits, and the commits on this PR are currently unsigned. Could you re-sign them and update the pr?

[smaEti]

Hey @smaEti, thanks for this. It's approved, and the code checks are fine. The only thing blocking the merge is our branch protection requiring signed commits, and the commits on this PR are currently unsigned. Could you re-sign them and update the pr?

Hey,
I ran into some issues while trying to re-sign the commits on this PR because there’s a merge commit in the history that makes the rebase process messy and complicated. (because i commited into the main branch of the forked repo, mybad :) )

Because of that, I’ll create a new PR with the same changes on a clean branch with properly signed commits.

Could you close this PR and review the new one instead?
Sorry for the inconvenience.

[Nageshbansal]

Hey @smaEti, no issues. Please create a new PR. Thanks :)

[therealpandey]

Works @smaEti :)