Thales Maciel
18b8fd2a7d
Implement the 2.8.0 workspace milestone with named snapshots and full-sandbox reset across the CLI, Python SDK, and MCP server. Persist the immutable baseline plus named snapshot archives under each workspace, add workspace reset metadata, and make reset recreate the sandbox while clearing command history, shells, and services without changing the workspace identity or diff baseline. Refresh the 2.8.0 docs, roadmap, and Python example around reset-over-repair, then validate with uv lock, UV_CACHE_DIR=.uv-cache make check, UV_CACHE_DIR=.uv-cache make dist-check, and a real guest-backed create/snapshot/reset/diff smoke test outside the sandbox.
380 lines
13 KiB
Python
380 lines
13 KiB
Python
from __future__ import annotations
|
|
|
|
import asyncio
|
|
from pathlib import Path
|
|
from typing import Any, cast
|
|
|
|
import pytest
|
|
|
|
import pyro_mcp.server as server_module
|
|
from pyro_mcp.server import create_server
|
|
from pyro_mcp.vm_manager import VmManager
|
|
from pyro_mcp.vm_network import TapNetworkManager
|
|
|
|
|
|
def test_create_server_registers_vm_tools(tmp_path: Path) -> None:
|
|
manager = VmManager(
|
|
backend_name="mock",
|
|
base_dir=tmp_path / "vms",
|
|
network_manager=TapNetworkManager(enabled=False),
|
|
)
|
|
|
|
async def _run() -> list[str]:
|
|
server = create_server(manager=manager)
|
|
tools = await server.list_tools()
|
|
return sorted(tool.name for tool in tools)
|
|
|
|
tool_names = asyncio.run(_run())
|
|
assert "vm_create" in tool_names
|
|
assert "vm_exec" in tool_names
|
|
assert "vm_list_environments" in tool_names
|
|
assert "vm_network_info" in tool_names
|
|
assert "vm_run" in tool_names
|
|
assert "vm_status" in tool_names
|
|
assert "workspace_create" in tool_names
|
|
assert "workspace_diff" in tool_names
|
|
assert "workspace_export" in tool_names
|
|
assert "workspace_logs" in tool_names
|
|
assert "workspace_sync_push" in tool_names
|
|
assert "shell_open" in tool_names
|
|
assert "shell_read" in tool_names
|
|
assert "shell_write" in tool_names
|
|
assert "shell_signal" in tool_names
|
|
assert "shell_close" in tool_names
|
|
assert "service_start" in tool_names
|
|
assert "service_list" in tool_names
|
|
assert "service_status" in tool_names
|
|
assert "service_logs" in tool_names
|
|
assert "service_stop" in tool_names
|
|
assert "snapshot_create" in tool_names
|
|
assert "snapshot_delete" in tool_names
|
|
assert "snapshot_list" in tool_names
|
|
assert "workspace_reset" in tool_names
|
|
|
|
|
|
def test_vm_run_round_trip(tmp_path: Path) -> None:
|
|
manager = VmManager(
|
|
backend_name="mock",
|
|
base_dir=tmp_path / "vms",
|
|
network_manager=TapNetworkManager(enabled=False),
|
|
)
|
|
|
|
def _extract_structured(raw_result: object) -> dict[str, Any]:
|
|
if not isinstance(raw_result, tuple) or len(raw_result) != 2:
|
|
raise TypeError("unexpected call_tool result shape")
|
|
_, structured = raw_result
|
|
if not isinstance(structured, dict):
|
|
raise TypeError("expected structured dictionary result")
|
|
return cast(dict[str, Any], structured)
|
|
|
|
async def _run() -> dict[str, Any]:
|
|
server = create_server(manager=manager)
|
|
executed = _extract_structured(
|
|
await server.call_tool(
|
|
"vm_run",
|
|
{
|
|
"environment": "debian:12",
|
|
"command": "printf 'git version 2.0\\n'",
|
|
"ttl_seconds": 600,
|
|
"network": False,
|
|
"allow_host_compat": True,
|
|
},
|
|
)
|
|
)
|
|
return executed
|
|
|
|
executed = asyncio.run(_run())
|
|
assert int(executed["exit_code"]) == 0
|
|
assert "git version" in str(executed["stdout"])
|
|
|
|
|
|
def test_vm_tools_status_stop_delete_and_reap(tmp_path: Path) -> None:
|
|
manager = VmManager(
|
|
backend_name="mock",
|
|
base_dir=tmp_path / "vms",
|
|
network_manager=TapNetworkManager(enabled=False),
|
|
)
|
|
manager.MIN_TTL_SECONDS = 1
|
|
|
|
def _extract_structured(raw_result: object) -> dict[str, Any]:
|
|
if not isinstance(raw_result, tuple) or len(raw_result) != 2:
|
|
raise TypeError("unexpected call_tool result shape")
|
|
_, structured = raw_result
|
|
if not isinstance(structured, dict):
|
|
raise TypeError("expected structured dictionary result")
|
|
return cast(dict[str, Any], structured)
|
|
|
|
async def _run() -> tuple[
|
|
dict[str, Any],
|
|
dict[str, Any],
|
|
dict[str, Any],
|
|
dict[str, Any],
|
|
list[dict[str, object]],
|
|
dict[str, Any],
|
|
]:
|
|
server = create_server(manager=manager)
|
|
environments_raw = await server.call_tool("vm_list_environments", {})
|
|
if not isinstance(environments_raw, tuple) or len(environments_raw) != 2:
|
|
raise TypeError("unexpected environments result")
|
|
_, environments_structured = environments_raw
|
|
if not isinstance(environments_structured, dict):
|
|
raise TypeError("environments tool should return a dictionary")
|
|
raw_environments = environments_structured.get("result")
|
|
if not isinstance(raw_environments, list):
|
|
raise TypeError("environments tool did not contain a result list")
|
|
created = _extract_structured(
|
|
await server.call_tool(
|
|
"vm_create",
|
|
{
|
|
"environment": "debian:12-base",
|
|
"ttl_seconds": 600,
|
|
"allow_host_compat": True,
|
|
},
|
|
)
|
|
)
|
|
vm_id = str(created["vm_id"])
|
|
await server.call_tool("vm_start", {"vm_id": vm_id})
|
|
status = _extract_structured(await server.call_tool("vm_status", {"vm_id": vm_id}))
|
|
network = _extract_structured(await server.call_tool("vm_network_info", {"vm_id": vm_id}))
|
|
stopped = _extract_structured(await server.call_tool("vm_stop", {"vm_id": vm_id}))
|
|
deleted = _extract_structured(await server.call_tool("vm_delete", {"vm_id": vm_id}))
|
|
|
|
expiring = _extract_structured(
|
|
await server.call_tool(
|
|
"vm_create",
|
|
{
|
|
"environment": "debian:12-base",
|
|
"ttl_seconds": 1,
|
|
"allow_host_compat": True,
|
|
},
|
|
)
|
|
)
|
|
expiring_id = str(expiring["vm_id"])
|
|
manager._instances[expiring_id].expires_at = 0.0 # noqa: SLF001
|
|
reaped = _extract_structured(await server.call_tool("vm_reap_expired", {}))
|
|
return (
|
|
status,
|
|
network,
|
|
stopped,
|
|
deleted,
|
|
cast(list[dict[str, object]], raw_environments),
|
|
reaped,
|
|
)
|
|
|
|
status, network, stopped, deleted, environments, reaped = asyncio.run(_run())
|
|
assert status["state"] == "started"
|
|
assert network["network_enabled"] is False
|
|
assert stopped["state"] == "stopped"
|
|
assert bool(deleted["deleted"]) is True
|
|
assert environments[0]["name"] == "debian:12"
|
|
assert int(reaped["count"]) == 1
|
|
|
|
|
|
def test_server_main_runs_stdio_transport(monkeypatch: pytest.MonkeyPatch) -> None:
|
|
called: dict[str, str] = {}
|
|
|
|
class StubServer:
|
|
def run(self, transport: str) -> None:
|
|
called["transport"] = transport
|
|
|
|
monkeypatch.setattr(server_module, "create_server", lambda: StubServer())
|
|
server_module.main()
|
|
assert called == {"transport": "stdio"}
|
|
|
|
|
|
def test_workspace_tools_round_trip(tmp_path: Path) -> None:
|
|
manager = VmManager(
|
|
backend_name="mock",
|
|
base_dir=tmp_path / "vms",
|
|
network_manager=TapNetworkManager(enabled=False),
|
|
)
|
|
source_dir = tmp_path / "seed"
|
|
source_dir.mkdir()
|
|
(source_dir / "note.txt").write_text("ok\n", encoding="utf-8")
|
|
|
|
def _extract_structured(raw_result: object) -> dict[str, Any]:
|
|
if not isinstance(raw_result, tuple) or len(raw_result) != 2:
|
|
raise TypeError("unexpected call_tool result shape")
|
|
_, structured = raw_result
|
|
if not isinstance(structured, dict):
|
|
raise TypeError("expected structured dictionary result")
|
|
return cast(dict[str, Any], structured)
|
|
|
|
async def _run() -> tuple[dict[str, Any], ...]:
|
|
server = create_server(manager=manager)
|
|
created = _extract_structured(
|
|
await server.call_tool(
|
|
"workspace_create",
|
|
{
|
|
"environment": "debian:12-base",
|
|
"allow_host_compat": True,
|
|
"seed_path": str(source_dir),
|
|
},
|
|
)
|
|
)
|
|
workspace_id = str(created["workspace_id"])
|
|
update_dir = tmp_path / "update"
|
|
update_dir.mkdir()
|
|
(update_dir / "more.txt").write_text("more\n", encoding="utf-8")
|
|
synced = _extract_structured(
|
|
await server.call_tool(
|
|
"workspace_sync_push",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"source_path": str(update_dir),
|
|
"dest": "subdir",
|
|
},
|
|
)
|
|
)
|
|
executed = _extract_structured(
|
|
await server.call_tool(
|
|
"workspace_exec",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"command": "cat subdir/more.txt",
|
|
},
|
|
)
|
|
)
|
|
diffed = _extract_structured(
|
|
await server.call_tool("workspace_diff", {"workspace_id": workspace_id})
|
|
)
|
|
snapshot = _extract_structured(
|
|
await server.call_tool(
|
|
"snapshot_create",
|
|
{"workspace_id": workspace_id, "snapshot_name": "checkpoint"},
|
|
)
|
|
)
|
|
snapshots = _extract_structured(
|
|
await server.call_tool("snapshot_list", {"workspace_id": workspace_id})
|
|
)
|
|
export_path = tmp_path / "exported-more.txt"
|
|
exported = _extract_structured(
|
|
await server.call_tool(
|
|
"workspace_export",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"path": "subdir/more.txt",
|
|
"output_path": str(export_path),
|
|
},
|
|
)
|
|
)
|
|
service = _extract_structured(
|
|
await server.call_tool(
|
|
"service_start",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"service_name": "app",
|
|
"command": "sh -lc 'touch .ready; while true; do sleep 60; done'",
|
|
"ready_file": ".ready",
|
|
},
|
|
)
|
|
)
|
|
services = _extract_structured(
|
|
await server.call_tool("service_list", {"workspace_id": workspace_id})
|
|
)
|
|
service_status = _extract_structured(
|
|
await server.call_tool(
|
|
"service_status",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"service_name": "app",
|
|
},
|
|
)
|
|
)
|
|
service_logs = _extract_structured(
|
|
await server.call_tool(
|
|
"service_logs",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"service_name": "app",
|
|
"all": True,
|
|
},
|
|
)
|
|
)
|
|
service_stopped = _extract_structured(
|
|
await server.call_tool(
|
|
"service_stop",
|
|
{
|
|
"workspace_id": workspace_id,
|
|
"service_name": "app",
|
|
},
|
|
)
|
|
)
|
|
reset = _extract_structured(
|
|
await server.call_tool(
|
|
"workspace_reset",
|
|
{"workspace_id": workspace_id, "snapshot": "checkpoint"},
|
|
)
|
|
)
|
|
deleted_snapshot = _extract_structured(
|
|
await server.call_tool(
|
|
"snapshot_delete",
|
|
{"workspace_id": workspace_id, "snapshot_name": "checkpoint"},
|
|
)
|
|
)
|
|
logs = _extract_structured(
|
|
await server.call_tool("workspace_logs", {"workspace_id": workspace_id})
|
|
)
|
|
deleted = _extract_structured(
|
|
await server.call_tool("workspace_delete", {"workspace_id": workspace_id})
|
|
)
|
|
return (
|
|
created,
|
|
synced,
|
|
executed,
|
|
diffed,
|
|
snapshot,
|
|
snapshots,
|
|
exported,
|
|
service,
|
|
services,
|
|
service_status,
|
|
service_logs,
|
|
service_stopped,
|
|
reset,
|
|
deleted_snapshot,
|
|
logs,
|
|
deleted,
|
|
)
|
|
|
|
(
|
|
created,
|
|
synced,
|
|
executed,
|
|
diffed,
|
|
snapshot,
|
|
snapshots,
|
|
exported,
|
|
service,
|
|
services,
|
|
service_status,
|
|
service_logs,
|
|
service_stopped,
|
|
reset,
|
|
deleted_snapshot,
|
|
logs,
|
|
deleted,
|
|
) = asyncio.run(_run())
|
|
assert created["state"] == "started"
|
|
assert created["workspace_seed"]["mode"] == "directory"
|
|
assert synced["workspace_sync"]["destination"] == "/workspace/subdir"
|
|
assert executed["stdout"] == "more\n"
|
|
assert diffed["changed"] is True
|
|
assert snapshot["snapshot"]["snapshot_name"] == "checkpoint"
|
|
assert [entry["snapshot_name"] for entry in snapshots["snapshots"]] == [
|
|
"baseline",
|
|
"checkpoint",
|
|
]
|
|
assert exported["artifact_type"] == "file"
|
|
assert Path(str(exported["output_path"])).read_text(encoding="utf-8") == "more\n"
|
|
assert service["state"] == "running"
|
|
assert services["count"] == 1
|
|
assert service_status["state"] == "running"
|
|
assert service_logs["tail_lines"] is None
|
|
assert service_stopped["state"] == "stopped"
|
|
assert reset["workspace_reset"]["snapshot_name"] == "checkpoint"
|
|
assert reset["command_count"] == 0
|
|
assert reset["service_count"] == 0
|
|
assert deleted_snapshot["deleted"] is True
|
|
assert logs["count"] == 0
|
|
assert deleted["deleted"] is True
|