Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
67 changes: 36 additions & 31 deletions tests/control_plane/test_turn_envelope_settlement.py
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
"""Settlement remains executable and signed through the opt-in compact view."""
"""Verify historical settlement signatures and the current C1 CLI plan."""
from copy import deepcopy
import shlex

import pytest

Expand All @@ -11,7 +12,7 @@
from tests.test_turn_envelope import _full_decision
from tests.control_plane.test_quota_settlement_cli import (
AGENT_ID, GOAL_ID, TODO_ID, TURN_ID,
_configure_read_only_todo, _run_cli, _run_generated_cli,
_classification_count, _configure_read_only_todo, _run_cli, _run_generated_cli,
_spend_run_count, _write_fixture, _initialize_fixture_authority,
)

Expand Down Expand Up @@ -91,31 +92,37 @@ def test_absent_plan_is_not_invented_for_historical_or_unbound_packets():
("--runtime-profile", "generic_cli"),
("--runtime-profile", "codex_app_ssh_goal"),
])
def test_generated_json_commands_settle_original_turn_once(tmp_path, profile_args):
def test_generated_c1_commands_settle_original_turn_once(tmp_path, profile_args):
project, runtime, registry = _write_fixture(tmp_path / "quoted inputs")
_configure_read_only_todo(project)
_initialize_fixture_authority(project, runtime)
args = ("quota", "should-run", *profile_args, "--goal-id", GOAL_ID,
"--agent-id", AGENT_ID, "--todo-id", TODO_ID,
"--turn-instance-id", TURN_ID, )
rc, full = _run_cli(registry, runtime, *args)
assert rc == 0, full
rc, envelope = _run_cli(registry, runtime, *args, "--turn-envelope")
assert rc == 0, envelope
assert envelope["schema_version"] == "loopx_turn_envelope_v0"
plan = envelope["writeback"]["settlement_plan"]
assert plan == full["interaction_contract"]["cli_channel"]["settlement_plan"]
assert plan["identity"] == full["heartbeat_receipt"]["settlement_identity"]
binding = (
"--goal-id", GOAL_ID, "--agent-id", AGENT_ID,
"--todo-id", TODO_ID, "--turn-instance-id", TURN_ID,
)
args = ("quota", "should-run", *profile_args, *binding)
rc, default = _run_cli(registry, runtime, *args)
assert rc == 0 and default["schema_version"] == "loopx_turn_envelope_v1", default
rc, alias = _run_cli(registry, runtime, *args, "--turn-envelope")
assert rc == 0 and alias["schema_version"] == "loopx_turn_envelope_v1", alias
plan = alias["interaction_contract"]["cli_channel"]["settlement_plan"]
assert plan == default["interaction_contract"]["cli_channel"]["settlement_plan"]
identity = default["heartbeat_receipt"]["settlement_identity"]
assert plan["identity"] == alias["heartbeat_receipt"]["settlement_identity"] == identity
assert plan["schema_version"] == "quota_settlement_plan_v2"
assert [step["kind"] for step in plan["ordered_steps"]] == [
"validation", "durable_writeback", "quota_spend", "terminal_closeout",
]
assert plan["ordered_steps"][-1]["conditional"] is True
assert plan["host_handoff"]["inside_agent_settlement"] is False
spend_command = plan["ordered_steps"][2]["command_template"]
rc, premature = _run_generated_cli(spend_command, registry_path=registry)
spend = plan["ordered_steps"][2]
assert spend["owner"] == "runtime" and "command_template" not in spend
rc, premature = _run_cli(registry, runtime, "quota", "spend-slot", *binding,
"--slots", "1", "--source", "heartbeat", "--execute")
assert rc != 0, premature
assert _spend_run_count(runtime) == 0
writeback = plan["ordered_steps"][1]["command_template"]
assert {"--settle-turn", "--settlement-receipt"} <= set(shlex.split(writeback))
# A validated exclusion advances research without claiming the objective is met.
# The same negative result is not a blocker merely because a candidate failed.
guidance = plan["ordered_steps"][1]["precondition"]
Expand All @@ -140,18 +147,16 @@ def test_generated_json_commands_settle_original_turn_once(tmp_path, profile_arg
assert _spend_run_count(runtime) == 0
writeback = writeback.replace("<validated_progress>", "validated_exclusion").replace(
"<scale>", "single_surface").replace("<outcome>", "outcome_progress")
writeback += " --progress-result-class advanced --progress-evidence-id exclusion-proof"
rc, refreshed = _run_generated_cli(
writeback + " --delivery-boundary in_flight_continuation --no-global-sync --suppress-external-sinks",
registry_path=registry,
)
assert rc == 0, refreshed
# Recovery instructions must also execute verbatim, with no format repair.
owed_command = refreshed["settlement_owed"]["command"]
for replay, command in ((False, owed_command), (True, spend_command)):
rc, spent = _run_generated_cli(command, registry_path=registry)
assert rc == 0, spent
assert spent["settlement_result"]["ok"] is True
if replay:
assert spent["idempotent_replay"] is True
assert _spend_run_count(runtime) == 1
writeback += (" --progress-result-class advanced --progress-evidence-id exclusion-proof"
" --delivery-boundary in_flight_continuation --no-global-sync --suppress-external-sinks")
for replay in (False, True):
rc, settled = _run_generated_cli(writeback, registry_path=registry)
assert rc == 0 and settled["settlement_progress"]["state"] == "settled", settled
assert settled["settlement_identity"] == identity
receipts = settled["settlement_result"]["receipts"]
assert [item["step_kind"] for item in receipts] == ["validation", "durable_writeback", "quota_spend"]
assert all(item["status"] == "committed" and item["effect_id"] == identity["effect_id"]
for item in receipts)
assert settled.get("idempotent_replay", False) is replay
assert _classification_count(runtime, "validated_exclusion") == 1
assert _spend_run_count(runtime) == 1
Loading