fix(compose): preserve container state across daemon restarts

[bilby91]

Closes #71.

Summary

• Shellout path (the one DAP runs on today): thread a single existingContainer bool from up.go through createFreshCompose → upComposeShellout → ComposeUpSpec.NoRecreate. When set, buildUpArgs appends --no-recreate to docker compose up -d. Mirrors upstream devcontainers/cli's gate (container || expectExistingContainer).
• Native orchestrator: when an existing container's LabelConfigHash and LabelImageDigest match the desired spec but its state is not Running, StartContainer it instead of stop+remove+recreate. The previous gate (c.State == runtime.StateRunning as a third reuse condition) destroyed perfectly-good containers any time dockerd had been restarted, because daemon-restored containers come back Exited.

Both branches share one root cause: drift in compose-side config-hash (or a temporarily-stopped container) was being treated as a recreate signal, which destroys the container's writable layer. For workspaces where ~/.claude, ~/.bash_history, install caches, or anything else lives in $HOME, that's data loss.

Test plan

• go test ./... clean.
• New TestBuildUpArgs_NoRecreate / TestBuildUpArgs_NoRecreateOmittedByDefault cover the argv edges. Verified the positive case fails against the pre-fix buildUpArgs.
• New TestUp_StartsStoppedContainerOnConfigMatch simulates a daemon-restart by stopping a previously-Up'd container and Up-ing again; asserts the container ID is preserved, StartContainer is called once, and RunContainer/RemoveContainer are not. Verified it fails against the pre-fix orchestrator.
• Live reproduction in DAP: pre-fix path destroys ~/.claude/projects/<encoded>/<sid>.jsonl across pod restart and Claude SDK errors with "No conversation found with session ID: <sid>". With this patch the JSONL survives.

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features

  • Stopped containers with matching configuration are now automatically restarted instead of being recreated.
  • Added --no-recreate flag support to preserve existing containers when configuration changes.

• Tests

  • Added tests to verify container reuse and --no-recreate flag behavior.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 95257fe3-4ab6-442b-a8e1-1a10acf0624f

📥 Commits

Reviewing files that changed from the base of the PR and between 4dd3c64 and 571db62.

📒 Files selected for processing (6)

• compose/orchestrator.go
• compose/orchestrator_test.go
• runtime/docker/compose.go
• runtime/docker/compose_test.go
• runtime/runtime.go
• up.go

---

📝 Walkthrough

Walkthrough

This PR fixes container state destruction across daemon restarts by preventing recreation when compose config hashes match. It adds a NoRecreate flag to ComposeUpSpec, threads existing-container detection through the engine's compose refresh path, implements --no-recreate flag passing to docker compose, and updates the native orchestrator backend to start stopped containers instead of recreating them on config match.

Changes

Container preservation on compose refresh

Layer / File(s)	Summary
ComposeUpSpec NoRecreate field runtime/runtime.go	Adds NoRecreate bool field to ComposeUpSpec with documentation explaining it gates --no-recreate for docker compose to preserve existing containers during config-matched refresh.
Docker compose --no-recreate flag implementation runtime/docker/compose.go, runtime/docker/compose_test.go	buildUpArgs conditionally appends --no-recreate when NoRecreate is set; two test cases verify the flag is included or omitted based on the boolean value.
Engine.Up threading and upComposeShellout wiring up.go	Engine.Up detects existing containers and threads existing != nil through createFreshCompose → upComposeShellout, where it sets ComposeUpSpec.NoRecreate to preserve containers across compose refresh.
Orchestrator stopped container reuse compose/orchestrator.go, compose/orchestrator_test.go	ensureService now starts stopped containers when config hash and image digest match, instead of falling through to stop-and-recreate; regression test verifies container ID is preserved and StartContainer is called on reuse after daemon-like restart.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Poem

A rabbit hops through compose with glee,
Preserving state across restarts free,
No recreate flags make containers stay,
Through daemon sleep, they live another day! 🐰✨

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: preserving container state across daemon restarts through the compose backend fix.
Linked Issues check	✅ Passed	The PR fully implements all coding requirements from issue #71: NoRecreate field added, buildUpArgs appends --no-recreate flag, existing container signal threaded through up.go/upComposeShellout, native orchestrator updated to start stopped config-matched containers, and comprehensive tests added.
Out of Scope Changes check	✅ Passed	All changes are directly scoped to issue #71: runtime specs, compose arg building, shellout path threading, orchestrator reuse logic, and corresponding unit/regression tests—no unrelated modifications present.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/compose-shellout-no-recreate

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}