refactor(cli): drop --format=json end-to-end
test / unit (pull_request) Successful in 18s
test / integration (pull_request) Successful in 1m2s

Companion to the compact preflight in #31 — the JSON format was
the structured alternative to the verbose text summary. With the
new compact text already on screen, no consumer was using the
JSON shape, and the abstract `BottlePlan.to_dict` was the
biggest piece of API surface no one is implementing against.

Removed:
  - `--format` CLI flag from `start` and `resume`.
  - `output_format` kwarg from `_launch_bottle`.
  - `BottlePlan.to_dict` abstract method.
  - `DockerBottlePlan.to_dict` (60-line dict builder).
  - The `_PlanView` dataclass — `print` was the only remaining
    caller, so the env-name computation is inlined.
  - `tests/integration/test_dry_run_plan.py` (JSON-shape
    integration test).
  - `tests/unit/test_cli_start_format.py` (flag-conflict unit).

Plan-introspection is still possible by reading the
`DockerBottlePlan` dataclass directly — fields like `image`,
`container_name`, `stage_dir`, `use_runsc` are all there. Tooling
that needs a stable wire shape can JSON-serialize the dataclass
themselves.

411 unit + integration tests pass.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
2026-05-25 20:54:51 -04:00
parent 5d5f118fb4
commit 572106d98f
6 changed files with 17 additions and 334 deletions
-149
View File
@@ -1,149 +0,0 @@
"""Integration: cli.py start --dry-run --format=json renders a stable
machine-readable plan and creates zero Docker resources. The shape of
the JSON document is part of the CLI's user-facing contract."""
import json
import os
import subprocess
import sys
import tempfile
import unittest
from pathlib import Path
from tests._docker import skip_unless_docker
REPO_ROOT = Path(__file__).resolve().parent.parent.parent
@skip_unless_docker()
class TestDryRunPlan(unittest.TestCase):
def test_dry_run_emits_structured_plan(self):
work_dir = Path(tempfile.mkdtemp())
try:
# PRD 0011 layout: per-file MD under .claude-bottle/.
# work_dir doubles as $HOME and as cwd for this test.
cb = work_dir / ".claude-bottle"
(cb / "bottles").mkdir(parents=True)
(cb / "agents").mkdir(parents=True)
(cb / "bottles" / "dev.md").write_text(
"---\n"
"egress:\n"
" allowlist:\n"
" - example.org\n"
"---\n"
)
(cb / "agents" / "demo.md").write_text(
"---\n"
"bottle: dev\n"
"---\n"
)
# Under act_runner with a host-mounted docker socket, the
# `docker network ls` / `docker ps -a` calls from inside the
# job container exit non-zero (see docs/ci.md for the same
# topology issue affecting other integration tests). Skip
# the side-effects guard there; locally the check still
# catches accidental docker resource creation by the dry
# run.
check_side_effects = os.environ.get("GITEA_ACTIONS") != "true"
nets_before = self._count_claude_bottle_networks() if check_side_effects else 0
ctrs_before = self._count_claude_bottle_containers() if check_side_effects else 0
env = os.environ.copy()
env["HOME"] = str(work_dir)
env.pop("CLAUDE_BOTTLE_DRY_RUN", None)
# The HOME override above isolates the manifest under test
# from the dev's real ~/claude-bottle.json. On Docker Desktop
# that same override breaks docker CLI endpoint resolution,
# since the active context lives in $HOME/.docker/config.json
# and the per-user socket sits under $HOME/.docker/run/.
# Pin DOCKER_HOST to the parent's resolved endpoint so the
# subprocess reaches the same daemon regardless of $HOME.
endpoint = subprocess.run(
["docker", "context", "inspect",
"--format", "{{.Endpoints.docker.Host}}"],
capture_output=True, text=True, check=True,
).stdout.strip()
if endpoint:
env["DOCKER_HOST"] = endpoint
result = subprocess.run(
[
sys.executable, str(REPO_ROOT / "cli.py"),
"start", "--dry-run", "--format", "json", "demo",
],
cwd=work_dir,
env=env,
capture_output=True,
text=True,
check=False,
)
self.assertEqual(
0, result.returncode,
f"start --dry-run failed: stderr={result.stderr}",
)
plan = json.loads(result.stdout)
self.assertEqual("demo", plan["agent"])
self.assertEqual("dev", plan["bottle"])
self.assertEqual("runc", plan["runtime"],
"runsc isn't available on the CI runner")
self.assertEqual([], plan["skills"])
self.assertEqual([], plan["git_remotes"])
self.assertEqual([], plan["git_gate"])
self.assertEqual(False, plan["remote_control"])
self.assertEqual(0, plan["prompt"]["length"])
# User-declared host + a baked default both present; the union
# is sorted and deduplicated.
hosts = plan["egress"]["hosts"]
self.assertIn("example.org", hosts)
self.assertIn("api.anthropic.com", hosts)
self.assertEqual(plan["egress"]["host_count"], len(hosts))
self.assertEqual(sorted(set(hosts)), hosts,
"hosts must be sorted and deduplicated")
# PRD 0006: TLS interception is on for every launched
# bottle. Fingerprint is null at dry-run (no CA exists
# yet); real launches log it from provision_ca.
self.assertEqual(
{"enabled": True, "ca_fingerprint": None},
plan["egress"]["tls_interception"],
)
# No Docker side effects (see the GITEA_ACTIONS skip note
# above — this guard runs locally only).
if check_side_effects:
self.assertEqual(nets_before, self._count_claude_bottle_networks(),
"no networks created")
self.assertEqual(ctrs_before, self._count_claude_bottle_containers(),
"no containers created")
finally:
import shutil
shutil.rmtree(work_dir, ignore_errors=True)
def _count_claude_bottle_networks(self) -> int:
return self._count_with_prefix(
["docker", "network", "ls", "--format", "{{.Name}}"], "claude-bottle"
)
def _count_claude_bottle_containers(self) -> int:
return self._count_with_prefix(
["docker", "ps", "-a", "--format", "{{.Names}}"], "claude-bottle"
)
def _count_with_prefix(self, cmd: list[str], prefix: str) -> int:
# capture_output + explicit returncode check so a docker
# failure surfaces its stderr in the test report instead of
# the bare CalledProcessError we used to get.
result = subprocess.run(cmd, capture_output=True, text=True, check=False)
if result.returncode != 0:
self.fail(
f"{' '.join(cmd)!r} failed (exit {result.returncode}): "
f"stderr={result.stderr.strip()!r}"
)
return sum(1 for n in result.stdout.splitlines() if n.startswith(prefix))
if __name__ == "__main__":
unittest.main()