fix(doctor): make docker.daemon check work under Podman

[hectorvent]

Summary

floci doctor's docker.daemon check reported "Docker daemon not reachable" under Podman even when the daemon was clearly up — the same run showed container.running, image.present, and endpoint.reachable all passing, and docker.socket correctly resolving the DOCKER_HOST Podman socket.

✗ docker.daemon    Docker daemon not reachable
                   Fix: Start Docker: sudo systemctl start docker
✓ docker.socket    /run/user/100033/podman/podman.sock accessible
✓ container.running  Container 'floci' is running
✓ endpoint.reachable http://localhost:4566 is reachable

Root cause

isDaemonReachable() probed with docker info --format '{{.ServerVersion}}'. .ServerVersion is a Docker-only info template field; under the podman-docker shim, podman info has no such field, so the template errors and exits non-zero → false "not reachable". (The reported "Docker 5.8.2" is Podman's version, confirming the shim.)

Fix

Probe reachability by exit code only — docker info / podman info exit 0 when the daemon (or DOCKER_HOST socket) is reachable, non-zero when not — with no schema dependency. Output is discarded so the unread pipe can't fill and deadlock waitFor().

Process p = new ProcessBuilder("docker", "info")
        .redirectOutput(ProcessBuilder.Redirect.DISCARD)
        .redirectError(ProcessBuilder.Redirect.DISCARD)
        .start();
return p.waitFor() == 0;

Testing

mvn test — 33 passing. (Daemon reachability is integration-only; no unit test added.)

Follow-up (not in this PR)

Cosmetic Podman gaps remain: docker.installed/docker.version label Podman as "Docker" (and the <20.10 warning is meaningless for Podman), and fix hints say "Docker Desktop"/"systemctl start docker". Worth a separate Podman-awareness enhancement.