CoreRasurae
cc24d8c8ac
Completely removed all legacy plan commands from the CLI:
- Removed tell, build, new, current, cd, continue CLI commands
- Removed programmatic wrapper functions (tell_command, build_command, etc.)
- Removed legacy deprecation message
- Updated help text to indicate V3 Plan Lifecycle exclusively
- Removed stale references to tell/build in help output and command validation
Removes the legacy 'tell' and 'build' CLI shortcuts from main.py:
- Removed echo lines advertising tell/build commands
- Removed tell/build from valid_cmds list
- Removed tell/build from _LIGHTWEIGHT_COMMANDS frozenset
- These dead entries were preventing helpful error messages
Test infrastructure improvements:
- Event bus exception test: Patch the module-level logger during emit() so that
structlog.testing.capture_logs() can capture the logs. Without patching, the
module-level logger created at import time is not captured by the context manager.
- Session create/list commands: Suppress cleveragents.mcp logger to CRITICAL level
during JSON/YAML output formatting to prevent health check warnings with ANSI codes
from being written to stdout before JSON output, which breaks JSON parsing.
- Extended plan_cli_coverage_boost with scenarios for estimation_result,
invariants, execution_environment, validation_summary, and checkpoint
coverage in _plan_spec_dict
Documentation:
- Created docs/Legacy_to_V3_Guide.md with comprehensive migration instructions
- Updated CONTRIBUTING.md to document removal of legacy workflow
- Updated CHANGELOG.md to reference issue #4181 instead of PR #10800
ISSUES CLOSED: #4181
232 lines
10 KiB
Gherkin
232 lines
10 KiB
Gherkin
Feature: Plan CLI supports only V3 commands (legacy commands removed)
|
|
|
|
Background:
|
|
Given CLI environment is ready for plan command tests
|
|
|
|
Scenario: Help output shows only V3 commands
|
|
When I invoke "--help" on plan app
|
|
Then the help contains "use"
|
|
And the help contains "execute"
|
|
And the help contains "apply"
|
|
And the help contains "status"
|
|
And the help contains "list"
|
|
And the help contains "errors"
|
|
|
|
Scenario: V3 command group is clearly labeled
|
|
When I invoke "--help" on plan app
|
|
Then the help contains "V3"
|
|
And the help contains "Plan Lifecycle"
|
|
|
|
Scenario: Legacy command tell is not available
|
|
When I invoke "tell --help" on plan app
|
|
Then the command fails with exit code non-zero
|
|
|
|
Scenario: Legacy command build is not available
|
|
When I invoke "build --help" on plan app
|
|
Then the command fails with exit code non-zero
|
|
|
|
Scenario: Legacy command new is not available
|
|
When I invoke "new --help" on plan app
|
|
Then the command fails with exit code non-zero
|
|
|
|
Scenario: Legacy command current is not available
|
|
When I invoke "current --help" on plan app
|
|
Then the command fails with exit code non-zero
|
|
|
|
Scenario: Legacy command cd is not available
|
|
When I invoke "cd --help" on plan app
|
|
Then the command fails with exit code non-zero
|
|
|
|
Scenario: Legacy command continue is not available
|
|
When I invoke "continue --help" on plan app
|
|
Then the command fails with exit code non-zero
|
|
|
|
Scenario: CLI runner initialization preserves state across commands
|
|
When I invoke "--help" on plan app
|
|
And I invoke "--help" on plan app again
|
|
Then both invocations should have consistent help output
|
|
And the context should not leak state between invocations
|
|
|
|
Scenario: CLI environment handles missing subcommand gracefully
|
|
When I invoke "" on plan app
|
|
Then the command fails with exit code non-zero
|
|
And the output contains help information
|
|
|
|
Scenario: V3 commands are correctly registered in app
|
|
When I invoke "--help" on plan app
|
|
Then the help output should contain complete command list
|
|
And the help output should not contain deprecated command references
|
|
|
|
Scenario: CLI resource cleanup after failed commands
|
|
Given CLI environment is ready for plan command tests
|
|
When I invoke "nonexistent-command" on plan app
|
|
Then the command fails with exit code non-zero
|
|
And the CLI runner should be in a valid state for next command
|
|
When I invoke "--help" on plan app
|
|
Then the help output should be available
|
|
|
|
Scenario: Repeated command invocations maintain consistency
|
|
Given CLI environment is ready for plan command tests
|
|
When I invoke "--help" on plan app
|
|
And I invoke "--help" on plan app again
|
|
When I invoke "--help" on plan app again
|
|
Then all invocations should have identical output
|
|
And the context state should remain clean
|
|
|
|
Scenario: Failed command does not corrupt context state
|
|
Given CLI environment is ready for plan command tests
|
|
When I invoke "--help" on plan app
|
|
And I save the help output for comparison
|
|
And I invoke "invalid-plan-id" on plan app
|
|
Then the command fails with exit code non-zero
|
|
When I invoke "--help" on plan app
|
|
Then the help output should match the saved output
|
|
|
|
Scenario: ULID validation rejects legacy plan names
|
|
When I validate plan ID "my-legacy-plan" as ULID
|
|
Then validation should fail with legacy workflow message
|
|
|
|
Scenario: ULID validation accepts valid ULID identifiers
|
|
When I validate plan ID "01HXM8C2ZK4Q7C2B3F2R4VYV6J" as ULID
|
|
Then validation should succeed with original plan ID returned
|
|
|
|
Scenario: Actor name validation requires namespace/name format
|
|
When I validate actor name "invalid-actor" as namespaced
|
|
Then validation should fail with actor format message
|
|
|
|
Scenario: Actor name validation accepts openai/gpt-4 format
|
|
When I validate actor name "openai/gpt-4" as namespaced
|
|
Then validation should succeed with original actor name returned
|
|
|
|
Scenario: Actor name validation accepts local/custom-actor format
|
|
When I validate actor name "local/custom-actor" as namespaced
|
|
Then validation should succeed with original actor name returned
|
|
|
|
Scenario: Relative time formatting shows current timestamp as "just now"
|
|
When I format current timestamp as relative time
|
|
Then formatted time should be "just now"
|
|
|
|
Scenario: Relative time formatting shows hours ago correctly
|
|
When I format timestamp from 3 hours ago as relative time
|
|
Then formatted time should contain "hour"
|
|
And formatted time should contain "ago"
|
|
|
|
Scenario: Relative time formatting shows days ago correctly
|
|
When I format timestamp from 5 days ago as relative time
|
|
Then formatted time should contain "day"
|
|
And formatted time should contain "ago"
|
|
|
|
Scenario: Multiple validation errors maintain CLI runner state
|
|
When I invoke an invalid plan command "execute not-a-ulid"
|
|
Then the command should fail
|
|
And I invoke another invalid command "apply another-invalid"
|
|
Then that command should also fail
|
|
And the CLI runner should remain functional
|
|
|
|
Scenario: Decision label generation for root decisions
|
|
When I generate decision label for type "prompt_definition" with ordinal 0
|
|
Then the label should be "Root"
|
|
|
|
Scenario: Decision label generation for strategy choices
|
|
When I generate decision label for type "strategy_choice" with ordinal 0
|
|
Then the label should be "Strategy"
|
|
|
|
Scenario: Decision label generation for invariant decisions
|
|
When I generate decision label for type "invariant_enforced" with ordinal 1
|
|
Then the label should be "Invariant 1"
|
|
|
|
Scenario: Decision label generation for multiple invariants
|
|
When I generate decision label for type "invariant_enforced" with ordinal 3
|
|
Then the label should be "Invariant 3"
|
|
|
|
Scenario: Decision label generation for subplan spawning
|
|
When I generate decision label for type "subplan_spawn" with ordinal 1
|
|
Then the label should be "Spawn 1"
|
|
|
|
Scenario: Decision label generation for parallel subplan spawning
|
|
When I generate decision label for type "subplan_parallel_spawn" with ordinal 2
|
|
Then the label should be "Parallel 2"
|
|
|
|
Scenario: Decision label generation for implementation choices
|
|
When I generate decision label for type "implementation_choice" with ordinal 0
|
|
Then the label should be "Implementation"
|
|
|
|
Scenario: Decision label generation for unknown decision types
|
|
When I generate decision label for type "unknown_type" with ordinal 0
|
|
Then the label should be "unknown_type"
|
|
|
|
Scenario: ULID validation rejects empty string
|
|
When I validate plan ID "" as ULID
|
|
Then validation should fail with legacy workflow message
|
|
|
|
Scenario: ULID validation rejects whitespace-only string
|
|
When I validate plan ID " " as ULID
|
|
Then validation should fail with legacy workflow message
|
|
|
|
Scenario: ULID validation rejects too-short identifiers
|
|
When I validate plan ID "01HXM8C" as ULID
|
|
Then validation should fail with legacy workflow message
|
|
|
|
Scenario: ULID validation rejects identifiers with invalid characters
|
|
When I validate plan ID "01HXM8C2ZK4Q7C2B3F2R4VYV6@" as ULID
|
|
Then validation should fail with legacy workflow message
|
|
|
|
Scenario: Actor name validation rejects empty string
|
|
When I validate actor name "" as namespaced
|
|
Then validation should fail with actor format message
|
|
|
|
Scenario: Actor name validation rejects name with only slash
|
|
When I validate actor name "/" as namespaced
|
|
Then validation should fail with actor format message
|
|
|
|
Scenario: Actor name validation rejects uppercase in namespace
|
|
When I validate actor name "OpenAI/gpt-4" as namespaced
|
|
Then validation should fail with actor format message
|
|
|
|
Scenario: Actor name validation rejects spaces in actor name
|
|
When I validate actor name "openai/gpt 4" as namespaced
|
|
Then validation should fail with actor format message
|
|
|
|
Scenario: Relative time formatting handles 1 minute ago
|
|
When I format timestamp from 1 minute ago as relative time
|
|
Then formatted time should be "1 minute ago"
|
|
|
|
Scenario: Relative time formatting handles 2 minutes ago
|
|
When I format timestamp from 2 minutes ago as relative time
|
|
Then formatted time should be "2 minutes ago"
|
|
|
|
Scenario: Relative time formatting handles 1 hour ago
|
|
When I format timestamp from 1 hour ago as relative time
|
|
Then formatted time should be "1 hour ago"
|
|
|
|
Scenario: Relative time formatting handles 2 hours ago
|
|
When I format timestamp from 2 hours ago as relative time
|
|
Then formatted time should be "2 hours ago"
|
|
|
|
Scenario: Relative time formatting handles 1 day ago
|
|
When I format timestamp from 1 day ago as relative time
|
|
Then formatted time should be "1 day ago"
|
|
|
|
Scenario: Relative time formatting handles 2 days ago
|
|
When I format timestamp from 2 days ago as relative time
|
|
Then formatted time should be "2 days ago"
|
|
|
|
Scenario: Relative time formatting handles very old timestamps
|
|
When I format timestamp from 365 days ago as relative time
|
|
Then formatted time should contain "365"
|
|
And formatted time should contain "day"
|
|
|
|
Scenario: ULID validation provides helpful error message for human-readable names
|
|
When I validate plan ID "my-plan" as ULID
|
|
Then validation error should mention "v3 plan lifecycle"
|
|
And validation error should mention "ULID"
|
|
|
|
Scenario: Actor validation provides helpful error message format requirements
|
|
When I validate actor name "missing-slash" as namespaced
|
|
Then validation error should mention "namespace/name"
|
|
|
|
Scenario: CLI runner state after consecutive failed validations
|
|
When I attempt multiple actor validations
|
|
Then all validations should fail appropriately
|
|
And the validation context should remain clean
|