Files
llmx/codex-cli/tests/agent-server-retry.test.ts
Michael Bolin ae5b1b5cb5 add support for -w,--writable-root to add more writable roots for sandbox (#263)
This adds support for a new flag, `-w,--writable-root`, that can be
specified multiple times to _amend_ the list of folders that should be
configured as "writable roots" by the sandbox used in `full-auto` mode.
Values that are passed as relative paths will be resolved to absolute
paths.

Incidentally, this required updating a number of the `agent*.test.ts`
files: it feels like some of the setup logic across those tests could be
consolidated.

In my testing, it seems that this might be slightly out of distribution
for the model, as I had to explicitly tell it to run `apply_patch` and
that it had the permissions to write those files (initially, it just
showed me a diff and told me to apply it myself). Nevertheless, I think
this is a good starting point.
2025-04-17 15:39:26 -07:00

169 lines
4.2 KiB
TypeScript
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import { describe, it, expect, vi } from "vitest";
// Utility: fake OpenAI SDK that can be instructed to fail with 5xx a set
// number of times before succeeding.
function createStream(events: Array<any>) {
return new (class {
public controller = { abort: vi.fn() };
async *[Symbol.asyncIterator]() {
for (const ev of events) {
yield ev;
}
}
})();
}
const openAiState: { createSpy?: ReturnType<typeof vi.fn> } = {};
vi.mock("openai", () => {
class FakeOpenAI {
public responses = {
create: (...args: Array<any>) => openAiState.createSpy!(...args),
};
}
class APIConnectionTimeoutError extends Error {}
return {
__esModule: true,
default: FakeOpenAI,
APIConnectionTimeoutError,
};
});
vi.mock("../src/approvals.js", () => ({
__esModule: true,
alwaysApprovedCommands: new Set<string>(),
canAutoApprove: () => ({ type: "auto-approve", runInSandbox: false } as any),
isSafeCommand: () => null,
}));
vi.mock("../src/format-command.js", () => ({
__esModule: true,
formatCommandForDisplay: (c: Array<string>) => c.join(" "),
}));
vi.mock("../src/utils/agent/log.js", () => ({
__esModule: true,
log: () => {},
isLoggingEnabled: () => false,
}));
import { AgentLoop } from "../src/utils/agent/agent-loop.js";
describe("AgentLoop automatic retry on 5xx errors", () => {
it("retries up to 3 times then succeeds", async () => {
// Fail twice with 500 then succeed.
let call = 0;
openAiState.createSpy = vi.fn(async () => {
call += 1;
if (call <= 2) {
const err: any = new Error("Internal Server Error");
err.status = 500;
throw err;
}
return createStream([
{
type: "response.output_item.done",
item: {
type: "message",
role: "assistant",
id: "m1",
content: [{ type: "text", text: "ok" }],
},
},
{
type: "response.completed",
response: {
id: "r1",
status: "completed",
output: [
{
type: "message",
role: "assistant",
id: "m1",
content: [{ type: "text", text: "ok" }],
},
],
},
},
]);
});
const received: Array<any> = [];
const agent = new AgentLoop({
model: "any",
instructions: "",
approvalPolicy: { mode: "auto" } as any,
additionalWritableRoots: [],
onItem: (i) => received.push(i),
onLoading: () => {},
getCommandConfirmation: async () => ({ review: "yes" } as any),
onLastResponseId: () => {},
});
const userMsg = [
{
type: "message",
role: "user",
content: [{ type: "input_text", text: "hi" }],
},
];
await agent.run(userMsg as any);
await new Promise((r) => setTimeout(r, 20));
expect(openAiState.createSpy).toHaveBeenCalledTimes(3);
const assistant = received.find((i) => i.role === "assistant");
expect(assistant?.content?.[0]?.text).toBe("ok");
});
it("fails after 3 attempts and surfaces system message", async () => {
openAiState.createSpy = vi.fn(async () => {
const err: any = new Error("Internal Server Error");
err.status = 502; // any 5xx
throw err;
});
const received: Array<any> = [];
const agent = new AgentLoop({
model: "any",
instructions: "",
approvalPolicy: { mode: "auto" } as any,
additionalWritableRoots: [],
onItem: (i) => received.push(i),
onLoading: () => {},
getCommandConfirmation: async () => ({ review: "yes" } as any),
onLastResponseId: () => {},
});
const userMsg = [
{
type: "message",
role: "user",
content: [{ type: "input_text", text: "hello" }],
},
];
await expect(agent.run(userMsg as any)).resolves.not.toThrow();
await new Promise((r) => setTimeout(r, 20));
expect(openAiState.createSpy).toHaveBeenCalledTimes(5);
const sysMsg = received.find(
(i) =>
i.role === "system" &&
typeof i.content?.[0]?.text === "string" &&
i.content[0].text.includes("Network error"),
);
expect(sysMsg).toBeTruthy();
});
});