Terminal success/failure closes kept the acceptance-phase operation state,
emitting self-contradicting envelopes (outcome=success with state=processing).
WithOperationTerminalState now closes the envelope at the observed terminal
status and clears timed_out.
WaitSpec.Validate trimmed status values only for checking and never wrote them
back, so padded declarations passed validation, published a padded Schema, and
then fail-closed at runtime as unknown statuses. NormalizeWaitSpec is now the
single canonical form shared by declaration (corecmd.New / AttachContract),
ToolSpec, and validation: trimmed values, duplicate/conflict rejection,
defensive copy.
Also covers the waitTimeoutDuration secs<=0 branch flagged by the coverage
gate.
Completes the wait capability per review guidance ("add corresponding
execution hooks and fallback tests for the modes"):
- contract.WaitSpec restores event/auto modes with event_key,
match_field, and a new resource_query (dotted path into the accepted
result data yielding the identifier events correlate against);
per-mode validation of required fields
- internal/wait adds EventStream (leaf-owned transport) and RunEvent:
correlated-event filtering, the same terminal/pending/unknown mapping
as polling, timed-out pending on deadline during consumption, and an
ErrEventStreamEnded sentinel distinguishing stream termination from
fail-closed status errors
- Spec.WaitEvents hook; validateWaitDecl pairs mode with hooks
(poll<->WaitPoll, event<->WaitEvents, auto<->both; surplus hooks
rejected too)
- the wait phase runs event-first in auto mode and falls back to polling
when the stream ends or the subscription fails, under one deadline
spanning both phases; strict event mode surfaces stream errors
- output.CommandResult gains Data() (deep copy) so the framework can
resolve the resource identifier without exposing mutable state
Changed-code coverage re-verified at 100% (CI cross-package recipe).
Addresses the three P1 findings from review 4942891040:
1. event/auto modes were declared but always executed polls. WaitSpec now
accepts poll only (event/auto fail validation with a not-implemented
message); event_key/match_field dead fields removed. Event waiting will
land with its own execution path and mode constant.
2. a deadline reached during the between-poll sleep re-polled with a
cancelled context, so a context-aware poller surfaced its error as a poll
failure instead of the contracted timed-out pending. The wait between
polls now uses a timer + select on ctx.Done(), the deadline is checked
before each poll, and a poll error on a cancelled context closes as
timed-out pending with the last observed status.
3. legacy Invoke/Orchestrate/RunE commands declaring Wait observed a failure
terminal while still exiting 0. validateWaitDecl now requires the
ResultInvoke dispatcher (the only path whose unified envelope can be
closed); the wait phase no longer wraps legacy paths.
Also: dropped the unreachable nonPendingTerminal branch, simplified
waitTimeoutSecs to the flag value (registration always seeds the reviewed
default), and raised changed-code coverage to 100% (new wait-engine edge
tests, output With* unit tests, contractfinal deep-copy coverage,
AttachContract invalid-Wait panic path).
Honor KindBool aliases in BuildArgs/hasEffectiveValue/constraintProvided;
reject MarkRequired+Aliases combinations at registration; validate
env-sourced values against declared enums; drop dead
ValidationEffective/ProjectDeclaredParameters constants and the unused
AnnotateRuntimeFlag parameter; route confirmationBypass through BoolFlag;
refresh retired-flow comments.
Pin admission race shards to timeout_budget (12m/cli 15m), cover
runtimeannotate and schema_source_root success paths for platform/main
gates, and make Windows absolute catalog path checks platform-safe.
Co-authored-by: Cursor <cursoragent@cursor.com>
Runtime assemble was stamping Snapshot.SourceHash with the registry
surface hash, so schema --all catalog_hash diverged from the CI dump
content source_hash and failed Policy. Also remap ContractFinal
Interface.Ref onto pinned MCP metadata so interface_type stays aligned.
Co-authored-by: Cursor <cursoragent@cursor.com>
Move schema authority to declare-time ParamDecl/ContractFinal and
ResolveSchemaBuild so CI/runtime assemble instead of shipping JSON
exclusions/meta-index as delivery sources.
Co-authored-by: Cursor <cursoragent@cursor.com>
Break the remaining corecmd→cli reverse dependency by owning
runtimeannotate and contractfinal on the framework side, with cli
keeping thin re-exports. Document the three authoring tiers and that
Shortcut may use DeclareLeafMetadata.
Co-authored-by: Cursor <cursoragent@cursor.com>
Align ContractFinal godoc and AttachContract comments with the
contractfinal/runtimeannotate seams, clarify CHANGELOG that corecmd
still may import cli subpackages, and gofmt shortcut_test imports.
Co-authored-by: Cursor <cursoragent@cursor.com>
Move AnnotateRuntime* into cli/runtimeannotate and the Cobra-keyed
ContractFinal store into cli/contractfinal so corecmd depends on thin
subpackages instead of the cli delivery root. Keep contract as DTO-only,
document Description declare-vs-delivery, and house homology gates under
cli/homology.
Co-authored-by: Cursor <cursoragent@cursor.com>
SchemaDecl confused authoring with Catalog/ToolSpec delivery. Authors now
declare ContractDecl (nested contract.* types) on Spec/LeafSpec/Shortcut;
AttachContract registers only through cli.RegisterRuntimeContractFinal, and
description provenance stamps cobra_help when Long wins.
Co-authored-by: Cursor <cursoragent@cursor.com>
Drop the dual-entry thin alias layer so helpers/shortcut/framework author
contract.* types directly; keep only AnnotateRuntime* delivery helpers in cli
and document the corecmd→cli seam.
Co-authored-by: Cursor <cursoragent@cursor.com>
玉澜