UAT: `PlanApplyService.apply_with_validation_gate()` silently fails to transition plan state — returns `outcome=applied` while plan remains in `Apply/QUEUED` state #1910

New issue

Open

opened 2026-04-03 00:11:53 +00:00 by freemo · 1 comment

freemo commented

2026-04-03 00:11:53 +00:00

Owner

Metadata

Branch: fix/uat-plan-apply-service-silent-state-transition-failure
Commit Message: fix(plan): call start_apply() before complete_apply() in apply_with_validation_gate() to prevent silent state transition failure
Milestone: v3.2.0
Parent Epic: #372

Background and Context

PlanApplyService.apply_with_validation_gate() in src/cleveragents/application/services/plan_apply_service.py silently fails to transition the plan's processing state to APPLIED after a successful apply. The method returns ApplyResult(outcome=APPLIED) even though the plan remains in Apply/QUEUED state.

Root Cause:
The method calls self._lifecycle.complete_apply() directly without first calling self._lifecycle.start_apply(). The complete_apply() method requires the plan to be in PROCESSING state, but the plan is in QUEUED state when apply_with_validation_gate() is called. The resulting PlanError is caught and silently swallowed (logged at DEBUG level only), and the method returns a false success result.

Code Location:

src/cleveragents/application/services/plan_apply_service.py, lines ~632–651 (the apply_with_validation_gate method)
The complete_apply() call at line ~634 fails because start_apply() was never called
The exception is caught at line ~640 and silently swallowed

Expected Behavior (from spec and CLI implementation):
The correct flow (as implemented in src/cleveragents/cli/commands/plan.py lines ~826–832) is:

Call start_apply(plan_id) — transitions plan from QUEUED to PROCESSING
Call complete_apply(plan_id) — transitions plan from PROCESSING to APPLIED

Actual Behavior:

apply_with_validation_gate() calls complete_apply() directly (skipping start_apply())
complete_apply() raises PlanError: Plan X is not processing (current: queued)
The exception is caught and logged at DEBUG level: "Could not transition to applied (plan may not be in Apply phase)"
The method returns ApplyResult(outcome=APPLIED) — a false success
The plan remains in Apply/QUEUED state forever (stuck)

Steps to Reproduce:

from cleveragents.application.services.plan_apply_service import PlanApplyService
from cleveragents.application.services.plan_executor import PlanExecutor
from cleveragents.application.services.plan_lifecycle_service import PlanLifecycleService
from cleveragents.config.settings import Settings
from cleveragents.domain.models.core.plan import ProjectLink, ProcessingState

settings = Settings()
lifecycle = PlanLifecycleService(settings=settings)
action = lifecycle.create_action(
    name='apply-test', description='Test', definition_of_done='- Step 1',
    strategy_actor='local/strategy-stub', execution_actor='local/execute-stub',
)
plan = lifecycle.use_action(action_name='apply-test', project_links=[ProjectLink(project_name='test-project')])
plan_id = plan.identity.plan_id

executor = PlanExecutor(lifecycle_service=lifecycle)
executor.run_strategize(plan_id)
lifecycle.execute_plan(plan_id)
executor.run_execute(plan_id)
lifecycle.apply_plan(plan_id)

apply_service = PlanApplyService(lifecycle_service=lifecycle)
result = apply_service.apply_with_validation_gate(plan_id, allow_empty=True)

print(f'Result outcome: {result.outcome}')  # Prints: applied (FALSE SUCCESS)
plan_final = lifecycle.get_plan(plan_id)
print(f'Plan state: {plan_final.processing_state}')  # Prints: queued (BUG: should be applied)

Impact:

Any code path that uses apply_with_validation_gate() (rather than the CLI's manual start_apply + complete_apply sequence) will silently fail to apply the plan
Plans get stuck in Apply/QUEUED state permanently
Callers receive a false success result (outcome=APPLIED) with no indication of failure

Why Tests Don't Catch This:
The existing BDD tests for apply_with_validation_gate use mocked lifecycle services where complete_apply is patched to succeed regardless of plan state. The bug only manifests with a real (non-mocked) PlanLifecycleService.

Subtasks

Reproduce the bug with a real (non-mocked) PlanLifecycleService and confirm the false-success return value
Add a call to self._lifecycle.start_apply(plan_id) before complete_apply() in apply_with_validation_gate() in src/cleveragents/application/services/plan_apply_service.py
Verify the fix transitions the plan correctly through QUEUED → PROCESSING → APPLIED
Add or update BDD scenario(s) for apply_with_validation_gate using a real (non-mocked) PlanLifecycleService to prevent regression
Ensure exception handling around the state transition is tightened so silent swallowing of PlanError is no longer possible (raise or surface the error appropriately)
Run full nox suite and confirm coverage ≥ 97%

Definition of Done

apply_with_validation_gate() correctly calls start_apply() before complete_apply()
The plan transitions to Apply/APPLIED (not Apply/QUEUED) after a successful apply via apply_with_validation_gate()
ApplyResult(outcome=APPLIED) is only returned when the plan state has actually been transitioned to APPLIED
A regression test using a real PlanLifecycleService (not a mock) covers this exact flow
No PlanError from state transition is silently swallowed in apply_with_validation_gate()
All nox stages pass
Coverage >= 97%

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## Metadata - **Branch**: `fix/uat-plan-apply-service-silent-state-transition-failure` - **Commit Message**: `fix(plan): call start_apply() before complete_apply() in apply_with_validation_gate() to prevent silent state transition failure` - **Milestone**: v3.2.0 - **Parent Epic**: #372 ## Background and Context `PlanApplyService.apply_with_validation_gate()` in `src/cleveragents/application/services/plan_apply_service.py` silently fails to transition the plan's processing state to `APPLIED` after a successful apply. The method returns `ApplyResult(outcome=APPLIED)` even though the plan remains in `Apply/QUEUED` state. **Root Cause:** The method calls `self._lifecycle.complete_apply()` directly without first calling `self._lifecycle.start_apply()`. The `complete_apply()` method requires the plan to be in `PROCESSING` state, but the plan is in `QUEUED` state when `apply_with_validation_gate()` is called. The resulting `PlanError` is caught and silently swallowed (logged at DEBUG level only), and the method returns a false success result. **Code Location:** - `src/cleveragents/application/services/plan_apply_service.py`, lines ~632–651 (the `apply_with_validation_gate` method) - The `complete_apply()` call at line ~634 fails because `start_apply()` was never called - The exception is caught at line ~640 and silently swallowed **Expected Behavior (from spec and CLI implementation):** The correct flow (as implemented in `src/cleveragents/cli/commands/plan.py` lines ~826–832) is: 1. Call `start_apply(plan_id)` — transitions plan from `QUEUED` to `PROCESSING` 2. Call `complete_apply(plan_id)` — transitions plan from `PROCESSING` to `APPLIED` **Actual Behavior:** 1. `apply_with_validation_gate()` calls `complete_apply()` directly (skipping `start_apply()`) 2. `complete_apply()` raises `PlanError: Plan X is not processing (current: queued)` 3. The exception is caught and logged at DEBUG level: "Could not transition to applied (plan may not be in Apply phase)" 4. The method returns `ApplyResult(outcome=APPLIED)` — a false success 5. The plan remains in `Apply/QUEUED` state forever (stuck) **Steps to Reproduce:** ```python from cleveragents.application.services.plan_apply_service import PlanApplyService from cleveragents.application.services.plan_executor import PlanExecutor from cleveragents.application.services.plan_lifecycle_service import PlanLifecycleService from cleveragents.config.settings import Settings from cleveragents.domain.models.core.plan import ProjectLink, ProcessingState settings = Settings() lifecycle = PlanLifecycleService(settings=settings) action = lifecycle.create_action( name='apply-test', description='Test', definition_of_done='- Step 1', strategy_actor='local/strategy-stub', execution_actor='local/execute-stub', ) plan = lifecycle.use_action(action_name='apply-test', project_links=[ProjectLink(project_name='test-project')]) plan_id = plan.identity.plan_id executor = PlanExecutor(lifecycle_service=lifecycle) executor.run_strategize(plan_id) lifecycle.execute_plan(plan_id) executor.run_execute(plan_id) lifecycle.apply_plan(plan_id) apply_service = PlanApplyService(lifecycle_service=lifecycle) result = apply_service.apply_with_validation_gate(plan_id, allow_empty=True) print(f'Result outcome: {result.outcome}') # Prints: applied (FALSE SUCCESS) plan_final = lifecycle.get_plan(plan_id) print(f'Plan state: {plan_final.processing_state}') # Prints: queued (BUG: should be applied) ``` **Impact:** - Any code path that uses `apply_with_validation_gate()` (rather than the CLI's manual `start_apply` + `complete_apply` sequence) will silently fail to apply the plan - Plans get stuck in `Apply/QUEUED` state permanently - Callers receive a false success result (`outcome=APPLIED`) with no indication of failure **Why Tests Don't Catch This:** The existing BDD tests for `apply_with_validation_gate` use mocked lifecycle services where `complete_apply` is patched to succeed regardless of plan state. The bug only manifests with a real (non-mocked) `PlanLifecycleService`. ## Subtasks - [ ] Reproduce the bug with a real (non-mocked) `PlanLifecycleService` and confirm the false-success return value - [ ] Add a call to `self._lifecycle.start_apply(plan_id)` before `complete_apply()` in `apply_with_validation_gate()` in `src/cleveragents/application/services/plan_apply_service.py` - [ ] Verify the fix transitions the plan correctly through `QUEUED → PROCESSING → APPLIED` - [ ] Add or update BDD scenario(s) for `apply_with_validation_gate` using a real (non-mocked) `PlanLifecycleService` to prevent regression - [ ] Ensure exception handling around the state transition is tightened so silent swallowing of `PlanError` is no longer possible (raise or surface the error appropriately) - [ ] Run full nox suite and confirm coverage ≥ 97% ## Definition of Done - [ ] `apply_with_validation_gate()` correctly calls `start_apply()` before `complete_apply()` - [ ] The plan transitions to `Apply/APPLIED` (not `Apply/QUEUED`) after a successful apply via `apply_with_validation_gate()` - [ ] `ApplyResult(outcome=APPLIED)` is only returned when the plan state has actually been transitioned to `APPLIED` - [ ] A regression test using a real `PlanLifecycleService` (not a mock) covers this exact flow - [ ] No `PlanError` from state transition is silently swallowed in `apply_with_validation_gate()` - [ ] All nox stages pass - [ ] Coverage >= 97% --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo added this to the v3.2.0 milestone

2026-04-03 00:12:17 +00:00

freemo added a new dependency

2026-04-03 00:14:41 +00:00

#372 Legendary: Plan Lifecycle (Workstream A)

freemo referenced this issue

2026-04-03 00:18:37 +00:00

[UAT-SUPERVISOR] Progress Report — Cycle 481 — v3.7.0 #1938

freemo commented

2026-04-03 00:19:50 +00:00

Author

Owner

Issue triaged by project owner:

State: Verified
MoSCoW: MoSCoW/Should Have — bug or error handling improvement.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: ca-project-owner

Issue triaged by project owner: - **State**: Verified - **MoSCoW**: MoSCoW/Should Have — bug or error handling improvement. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: ca-project-owner