fix: 区分 provider responses 失败分类

src/backend/codex-stdio.ts

@@ -1160,7 +1160,9 @@ function classifyMessageFailureKind(message: string, fallback: FailureKind): Fai
   if (/invalid[_ -]?prompt/u.test(text) && /invalid function arguments json string|tool_call_id/u.test(text)) return "provider-invalid-tool-call";
   if (/invalid function arguments json string/u.test(text)) return "provider-invalid-tool-call";
   if (/rate.?limit|too many requests|\b429\b/u.test(text)) return "provider-rate-limited";
-  if (/\b401\b|\b403\b|unauthori[sz]ed|forbidden|invalid api key|authentication|auth failed|oauth|access token/u.test(text)) return "provider-auth-failed";
+  if (/\b401\b|\b403\b|unauthori[sz]ed|forbidden|invalid api key|api key (?:is )?(?:required|missing)|authentication|auth failed|oauth|access token/u.test(text)) return "provider-auth-failed";
+  if (isProviderStreamDisconnectedMessage(text)) return "provider-stream-disconnected";
+  if (isProviderHttpErrorMessage(text)) return "provider-http-error";
   if (isProviderUnavailableMessage(text)) return "provider-unavailable";
   if (/timed out|timeout|idle timeout/u.test(text)) return "backend-timeout";
   if (/invalid json|json parse/u.test(text)) return "backend-json-parse-error";
@@ -1171,6 +1173,16 @@ function isProviderCompactUnsupportedMessage(text: string): boolean {
   return /responses\/compact|\/compact\b/u.test(text) && /\b404\b|not found|unsupported|no route|not implemented/u.test(text);
 }
 
+function isProviderStreamDisconnectedMessage(text: string): boolean {
+  return /responsestreamdisconnected|response stream disconnected|stream disconnected before completion|disconnected before completion/u.test(text);
+}
+
+function isProviderHttpErrorMessage(text: string): boolean {
+  if (/\b(?:http(?:\s+status)?|status(?:\s+code)?|unexpected status|status|code)\s*[:=]?\s*[45]\d\d\b/u.test(text)) return true;
+  if (/\b[45]\d\d\b/u.test(text) && /http|status|service unavailable|bad gateway|gateway timeout|internal server error|not found|provider|upstream/u.test(text)) return true;
+  return false;
+}
+
 function isProviderUnavailableMessage(text: string): boolean {
   if (/\b(?:http(?:\s+status)?|status(?:\s+code)?|code)\s*[:=]?\s*5\d\d\b/u.test(text)) return true;
   if (/\b5\d\d\b/u.test(text) && /service unavailable|bad gateway|gateway timeout|internal server error|provider|upstream|response\s*stream\s*disconnected|responsestreamdisconnected/u.test(text)) return true;

src/common/types.ts

@@ -19,6 +19,8 @@ export type FailureKind =
   | "backend-timeout"
   | "provider-auth-failed"
   | "provider-rate-limited"
+  | "provider-stream-disconnected"
+  | "provider-http-error"
   | "provider-invalid-tool-call"
   | "provider-compact-unsupported"
   | "provider-unavailable"

src/selftest/cases/30-codex-stdio.ts

@@ -225,9 +225,11 @@ const selfTest: SelfTestCase = async (context) => {
     await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-429-terminal", expectedStatus: "failed", expectedFailureKind: "provider-rate-limited" });
     await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-invalid-tool-call", expectedStatus: "failed", expectedFailureKind: "provider-invalid-tool-call" });
     await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-compact-404-terminal", expectedStatus: "failed", expectedFailureKind: "provider-compact-unsupported" });
-    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-503-rpc-error", expectedStatus: "failed", expectedFailureKind: "provider-unavailable" });
-    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-503-terminal", expectedStatus: "failed", expectedFailureKind: "provider-unavailable" });
-    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-503-retry-event", expectedStatus: "failed", expectedFailureKind: "provider-unavailable", expectRetryError: true });
+    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-stream-disconnected-rpc-error", expectedStatus: "failed", expectedFailureKind: "provider-stream-disconnected" });
+    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-503-rpc-error", expectedStatus: "failed", expectedFailureKind: "provider-stream-disconnected" });
+    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-503-terminal", expectedStatus: "failed", expectedFailureKind: "provider-http-error" });
+    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-unavailable-terminal", expectedStatus: "failed", expectedFailureKind: "provider-unavailable" });
+    await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "provider-503-retry-event", expectedStatus: "failed", expectedFailureKind: "provider-stream-disconnected", expectRetryError: true });
     await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "invalid-json", expectedStatus: "failed", expectedFailureKind: "backend-json-parse-error" });
     await runFailureCase({ client, managerUrl: server.baseUrl, context, mode: "missing-terminal", expectedStatus: "failed", expectedFailureKind: "backend-timeout", timeoutMs: 500 });
     await runSlowProgressIdleCase({ client, managerUrl: server.baseUrl, context });
@@ -239,7 +241,7 @@ const selfTest: SelfTestCase = async (context) => {
     await runSessionStorageSubdirCase({ client, managerUrl: server.baseUrl, context });
     await runSessionStorageNoSecretLeakCase({ client, managerUrl: server.baseUrl, context });
 
-    return { name: "codex-stdio", tests: ["runner-lease-heartbeat", "runner-lease-conflict-recovery", "codex-stdio-fake-turn", "codex-stdio-k8s-sandbox-override", "codex-stdio-projected-writable-home", "codex-stdio-deepseek-profile-fake-turn", "codex-stdio-dsflash-go-profile-fake-turn", "codex-stdio-dsflash-go-config-metadata", "codex-stdio-minimax-m3-profile-fake-turn", "codex-stdio-deepseek-missing-secret-no-fallback", "codex-stdio-minimax-m3-missing-secret-no-fallback", "codex-stdio-config-model-authoritative", "codex-stdio-explicit-model-forwarded", "codex-stdio-final-agent-message-only", "codex-stdio-web-search-progress", "codex-stdio-stale-thread-resume-failed", "codex-stdio-live-tool-events", "codex-stdio-noisy-reasoning-suppression", "codex-stdio-missing-turn-result", "codex-stdio-provider-auth-failed", "codex-stdio-provider-rate-limited", "codex-stdio-provider-invalid-tool-call", "codex-stdio-provider-compact-unsupported", "codex-stdio-provider-503-rpc-error", "codex-stdio-provider-503-terminal", "codex-stdio-provider-503-retry-event", "codex-stdio-invalid-json", "codex-stdio-timeout", "codex-stdio-idle-timeout-progress-refresh", "codex-stdio-command-failure-keeps-run-open", "codex-stdio-secret-unavailable", "codex-stdio-spawn-failure"] };
+    return { name: "codex-stdio", tests: ["runner-lease-heartbeat", "runner-lease-conflict-recovery", "codex-stdio-fake-turn", "codex-stdio-k8s-sandbox-override", "codex-stdio-projected-writable-home", "codex-stdio-deepseek-profile-fake-turn", "codex-stdio-dsflash-go-profile-fake-turn", "codex-stdio-dsflash-go-config-metadata", "codex-stdio-minimax-m3-profile-fake-turn", "codex-stdio-deepseek-missing-secret-no-fallback", "codex-stdio-minimax-m3-missing-secret-no-fallback", "codex-stdio-config-model-authoritative", "codex-stdio-explicit-model-forwarded", "codex-stdio-final-agent-message-only", "codex-stdio-web-search-progress", "codex-stdio-stale-thread-resume-failed", "codex-stdio-live-tool-events", "codex-stdio-noisy-reasoning-suppression", "codex-stdio-missing-turn-result", "codex-stdio-provider-auth-failed", "codex-stdio-provider-rate-limited", "codex-stdio-provider-invalid-tool-call", "codex-stdio-provider-compact-unsupported", "codex-stdio-provider-stream-disconnected", "codex-stdio-provider-503-rpc-error", "codex-stdio-provider-503-terminal", "codex-stdio-provider-unavailable", "codex-stdio-provider-503-retry-event", "codex-stdio-invalid-json", "codex-stdio-timeout", "codex-stdio-idle-timeout-progress-refresh", "codex-stdio-command-failure-keeps-run-open", "codex-stdio-secret-unavailable", "codex-stdio-spawn-failure"] };
   } finally {
     await new Promise<void>((resolve) => server.server.close(() => resolve()));
   }
@@ -310,7 +312,7 @@ async function runFailureDoesNotTerminalRunCase(options: { client: ManagerClient
   }) as JsonRecord;
   assert.equal(result.stopped, "idle-timeout", "non one-shot runner should remain alive after a failed command until idle timeout");
   assert.equal(result.terminalStatus, "failed");
-  assert.equal(result.failureKind, "provider-unavailable");
+  assert.equal(result.failureKind, "provider-http-error");
   const command = await options.client.get(`/api/v1/runs/${item.runId}/commands/${item.commandId}`) as { state?: string };
   assert.equal(command.state, "failed");
   const run = await options.client.get(`/api/v1/runs/${item.runId}`) as { status?: string; terminalStatus?: string | null; failureKind?: string | null };

src/selftest/fake-codex-app-server.ts

@@ -82,6 +82,10 @@ for await (const line of rl) {
       respond(message.id, null, { code: -32000, message: "responseStreamDisconnected: HTTP 503 Service Unavailable from provider" });
       continue;
     }
+    if (mode === "provider-stream-disconnected-rpc-error") {
+      respond(message.id, null, { code: -32000, message: "stream disconnected before completion: error sending request for url (http://138.2.51.180:8083/responses)" });
+      continue;
+    }
     if (mode === "provider-401-rpc-error") {
       respond(message.id, null, { code: -32000, message: "HTTP 401 Unauthorized: invalid api key" });
       continue;
@@ -122,6 +126,14 @@ for await (const line of rl) {
       respond(message.id, { turn });
       continue;
     }
+    if (mode === "provider-unavailable-terminal") {
+      turnCounter += 1;
+      const turn = { id: `turn_selftest_${turnCounter}`, status: "failed", error: { message: "provider is temporarily unavailable" } };
+      notify("turn/started", { turn: { id: turn.id, status: "running" } });
+      notify("turn/completed", { turn });
+      respond(message.id, { turn });
+      continue;
+    }
     if (mode === "provider-429-terminal") {
       turnCounter += 1;
       const turn = { id: `turn_selftest_${turnCounter}`, status: "failed", error: { message: "HTTP 429 Too Many Requests: rate limit exceeded" } };