@nowcrew/daemon 0.6.63 → 0.6.64

package/README.md

@@ -380,6 +380,7 @@ CREW_EXECUTION_MAX_STARTING_TOTAL=10
 CREW_EXECUTION_MAX_STARTING_PER_RUNTIME=10
 CREW_EXECUTION_START_GAP_MS=500
 CREW_EXECUTION_STARTUP_TIMEOUT_MS=120000
+# Retained for compatibility; accepted executions wait indefinitely for a slot.
 CREW_EXECUTION_MAX_QUEUE_WAIT_MS=300000
 CREW_EXECUTION_MAX_FIRST_OUTPUT_WAIT_MS=120000
 CREW_EXECUTION_FIRST_OUTPUT_GRACE_MS=180000
@@ -391,14 +392,14 @@ protocol-v1 and legacy work.
 Runtime startup is a separate FIFO gate: by default up to ten Claude, Codex, or Kimi processes may be
 initializing at a time, and launches are spaced by 500ms. A process leaves the startup gate
 only after its runtime-specific ready event; startup timeout cancels the owned process tree.
-An accepted execution may wait up to five minutes for a local and host slot. After runtime readiness,
-the daemon waits up to two minutes for the first model event, with a three-minute grace window for
-slow providers. When one of these local liveness limits expires, the supervisor is stopped and the
-terminal result is written to the execution journal before it is reported, so the execution remains
-replayable instead of being silently dropped. Queue and first-output timeouts are retried locally at
-most twice (three total attempts). Each retry releases its current machine/host slot and reserves a
-new one, which puts it behind work already waiting in the queue. A third timeout is terminal and is
-not requeued; other failures are not automatically retried.
+An accepted execution waits indefinitely for a local and host slot unless it is explicitly cancelled.
+After runtime readiness, the daemon waits up to two minutes for the first model event, with a
+three-minute grace window for slow providers. When the first-output liveness limit expires, the
+supervisor is stopped and the terminal result is written to the execution journal before it is
+reported, so the execution remains replayable instead of being silently dropped. First-output
+timeouts are retried locally at most twice (three total attempts). Each retry releases its current
+machine/host slot and reserves a new one, which puts it behind work already waiting in the queue. A
+third timeout is terminal and is not requeued; other failures are not automatically retried.
 
 Local policy can reduce server-requested access and limits; it cannot grant more access than requested.
 Provider credentials and configured environment are prepared locally and never carried in execution

package/dist/execution-runner.js

@@ -299,12 +299,6 @@ class ExecutionCancelledError extends Error {
         this.name = "ExecutionCancelledError";
     }
 }
-class ExecutionQueueTimeoutError extends Error {
-    constructor(timeoutMs) {
-        super(`Execution queue wait exceeded ${timeoutMs}ms`);
-        this.name = "ExecutionQueueTimeoutError";
-    }
-}
 async function cancellable(promise, cancellation) {
     if (cancellation === undefined)
         return promise;
@@ -315,24 +309,9 @@ async function cancellable(promise, cancellation) {
         cancellation.requested.then(() => { throw new ExecutionCancelledError(); }),
     ]);
 }
-async function cancellableWithTimeout(promise, timeoutMs, cancellation, timeoutError) {
-    let timer;
-    try {
-        return await Promise.race([
-            cancellable(promise, cancellation),
-            new Promise((_resolve, reject) => {
-                timer = setTimeout(() => reject(timeoutError), timeoutMs);
-            }),
-        ]);
-    }
-    finally {
-        if (timer !== undefined)
-            clearTimeout(timer);
-    }
-}
 function failedCompletion(spec, error, startedAt, finishedAt) {
     const message = error instanceof Error ? error.message : String(error);
-    const errorCode = error instanceof ExecutionQueueTimeoutError || error instanceof HostReservationCancelledError
+    const errorCode = error instanceof HostReservationCancelledError
         ? "queue_timeout"
         : error instanceof ProjectContextUnavailableError
             ? error.code
@@ -519,7 +498,9 @@ export async function runExecution(config, input, dependencies) {
     });
     try {
         if (dependencies.slot !== undefined) {
-            await cancellableWithTimeout(dependencies.slot.ready, config.executionLimits.maxQueueWaitMs, dependencies.cancellation, new ExecutionQueueTimeoutError(config.executionLimits.maxQueueWaitMs));
+            // Queueing is not a liveness failure. Keep the execution durable until its
+            // machine/host slot is granted or the server explicitly cancels it.
+            await cancellable(dependencies.slot.ready, dependencies.cancellation);
         }
         let projectContext;
         let sessionContextFingerprint;
@@ -562,17 +543,15 @@ export async function runExecution(config, input, dependencies) {
                 if (cancel !== null && timeout === undefined && !dependencies.cancellation?.isRequested()) {
                     const firstOutputDeadline = config.executionLimits.maxFirstOutputWaitMs
                         + config.executionLimits.firstOutputGraceMs;
-                    if (effectiveTimeoutMs > firstOutputDeadline) {
-                        firstOutputTimer = setTimeout(() => {
-                            firstOutputTimedOut = true;
-                            try {
-                                void cancel().catch(rejectCancellationFailure);
-                            }
-                            catch (error) {
-                                rejectCancellationFailure(error);
-                            }
-                        }, firstOutputDeadline);
-                    }
+                    firstOutputTimer = setTimeout(() => {
+                        firstOutputTimedOut = true;
+                        try {
+                            void cancel().catch(rejectCancellationFailure);
+                        }
+                        catch (error) {
+                            rejectCancellationFailure(error);
+                        }
+                    }, firstOutputDeadline);
                     timeout = setTimeout(() => {
                         timedOut = true;
                         try {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@nowcrew/daemon",
-  "version": "0.6.63",
+  "version": "0.6.64",
   "type": "module",
   "description": "crew daemon — 运行在用户机器:拉起/管理 agent 进程,注入 crew CLI,归一化 runtime 事件",
   "license": "Apache-2.0",