2eae0b2507
Removes the dashboard's stuck-task tagging per operator request: the Stuck card/status badges, stuck row styling, the footer Stuck segment and stuckTaskCount stat, utils/taskStuck.ts, the isStuck agent-activity gate, and the taskStuckTimeoutMs prop plumbing (App -> Board/Lane/Column/ WorktreeGroup/MainContent -> TaskCard/ListView/ExecutorStatusBar). Stuck-task tests are deleted or reconciled. The taskStuckTimeoutMs setting and the engine's recovery sweeps (including the stuck-killed status) are unchanged — the setting is engine-side only now. Also repoints the FN-6756 liveness-gate ratchet's facade scans at executor/task-executor-session-facades.ts, where the wave20 extraction moved hasLiveSessionSurface/clearPhantomExecutorBinding (the two pre-existing red tests on main). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
86 lines
4.2 KiB
TypeScript
86 lines
4.2 KiB
TypeScript
import type { Task } from "@fusion/core";
|
|
import { enrichRunningAgentTaskShapeFromFlags, isRunningAgentTask } from "../../../core/src/agents/live-agent-count";
|
|
import { isArchivedColumnRole, isCompleteColumnRole } from "./columnRoles";
|
|
|
|
/** The shared status vocabulary for active task phases and lock/model policy. */
|
|
export const ACTIVE_STATUSES = new Set([
|
|
"planning",
|
|
"researching",
|
|
"executing",
|
|
"finalizing",
|
|
"merging",
|
|
"merging-pr",
|
|
"merging-fix",
|
|
"reviewing",
|
|
"landing",
|
|
]);
|
|
|
|
export interface TaskAgentActivityOptions {
|
|
globalPaused?: boolean;
|
|
queued?: boolean;
|
|
// FNXC:StuckTagRemoval 2026-08-17-22:30: the isStuck gate was deleted with the dashboard's stuck-task tagging; the engine-written "stuck-killed" status below still suppresses the pulse.
|
|
/*
|
|
FNXC:WorkflowResolvedColumns 2026-07-29-00:00 (U12 — R8 drift conversion):
|
|
The task's own column traits, when the caller has them. This predicate drives the
|
|
pulsing status badge, the agent-active row border, and the column header's executing
|
|
count, so a card in a renamed lane must resolve its roles from traits rather than ids.
|
|
|
|
Optional, and the legacy ids remain the fallback: callers without resolved metadata
|
|
(pre-load, or a card stranded in a vanished lane) must keep their current behaviour
|
|
rather than lose activity detection entirely.
|
|
*/
|
|
/*
|
|
FNXC:TaskActivity 2026-08-01-17:53:
|
|
Widened with the review/merge roles because the positive arm now delegates to the shared
|
|
live-agent predicate, which resolves merge statuses through mergeOrchestration/mergeBlocker.
|
|
*/
|
|
columnFlags?: { intake?: boolean; hold?: boolean; complete?: boolean; archived?: boolean; countsTowardWip?: boolean; mergeOrchestration?: boolean; mergeBlocker?: boolean };
|
|
}
|
|
|
|
/*
|
|
FNXC:TaskActivity 2026-07-16-00:00:
|
|
FN-8055 makes the agent-active border and pulsing badges represent the same ground truth: an agent is working now. Reject render-context global pause, queue, and derived freshness-stuck gates before checking activity.
|
|
|
|
FNXC:TaskActivity 2026-08-01-17:53:
|
|
Operator requirement: activity chrome and lane counts must NEVER show more work than the engine's
|
|
actual live-agent population. Summing lane headers used to exceed the concurrency cap (e.g. 10 glowing
|
|
cards under a 9-slot limit) because this predicate unioned extra render-only signals — the FN-8494
|
|
`needs-replan` REVISING chrome, the FN-8300 fresh planner-log window, ACTIVE_STATUSES in any column,
|
|
and running unified-progress items. Those extras glowed on cards that hold no concurrency slot, which
|
|
operators read as a capacity breach.
|
|
The positive arm is now exactly the shared `isRunningAgentTask` predicate used by footer Running and
|
|
project admission, so card glow (and the header counts derived from it) is a strict subset of the
|
|
slot-holding population. The suppression gates above it only ever subtract (queued/stuck/paused/failed
|
|
cards can still hold a slot briefly but must not glow), preserving the "never more" direction.
|
|
|
|
Stuck-killed and both terminal columns are never active, even when stale execution status or workflow-step data remains on the task.
|
|
|
|
Model-resolution and routing locks intentionally import only ACTIVE_STATUSES and retain their status-or-in-progress policy; using this rendering predicate there would change lock behavior during status-null workflow steps.
|
|
*/
|
|
export function isTaskAgentActive(
|
|
task: Pick<Task, "column" | "status" | "paused" | "userPaused" | "steps" | "enabledWorkflowSteps" | "workflowStepResults" | "recentAgentActivityAt" | "sessionFile" | "checkedOutBy">,
|
|
options: TaskAgentActivityOptions = {},
|
|
): boolean {
|
|
const status = task.status;
|
|
|
|
if (
|
|
options.globalPaused === true ||
|
|
options.queued === true ||
|
|
status === "queued" ||
|
|
status === "stuck-killed" ||
|
|
task.paused === true ||
|
|
task.userPaused === true ||
|
|
status === "paused" ||
|
|
status === "failed" ||
|
|
status === "awaiting-approval" ||
|
|
status === "awaiting-user-input" ||
|
|
isCompleteColumnRole(options.columnFlags, task.column) ||
|
|
isArchivedColumnRole(options.columnFlags, task.column) ||
|
|
status === "done"
|
|
) {
|
|
return false;
|
|
}
|
|
|
|
return isRunningAgentTask(enrichRunningAgentTaskShapeFromFlags(task, options.columnFlags));
|
|
}
|