Understand any codebase instantly. System intelligence for codebases, built for humans and AI.
TypeScript aiai-memoryclaude-codeclicode-analysiscode-mappingcode-memorycodebase-understandingcodexdeveloper-toolsllmopenclawpersistent-memoryprogram-analysissoftware-architecturesystem-intelligencesystem-mapping
test(mcp): cover the read-scope reset after a failed run
KageBinary4473b3cdd3e0
diff --git a/ix-cli/src/cli/__tests__/mcp-runner.test.ts b/ix-cli/src/cli/__tests__/mcp-runner.test.tsindex 33ffe6bb..f3c7937e 100644--- a/ix-cli/src/cli/__tests__/mcp-runner.test.ts+++ b/ix-cli/src/cli/__tests__/mcp-runner.test.ts@@ -405,6 +405,21 @@ describe("in-process ix runner", () => {});});+ it("invalidates the read-scope cache after a failed run", async () => {+ const run = testRunner();+ resetReadScope.mockClear();++ // A failed read is often "no workspace here"; the next call must look the+ // scope up again rather than answer from the one that just failed.+ expect(await run(["soft-fail"])).toMatchObject({ ok: false });+ expect(resetReadScope).toHaveBeenCalledTimes(1);+ expect(await run(["boom"])).toMatchObject({ ok: false });+ expect(resetReadScope).toHaveBeenCalledTimes(2);++ await run(["say", "read"]);+ expect(resetReadScope).toHaveBeenCalledTimes(2);+ });+it("fails the run when output is truncated to protect the server heap", async () => {const run = createInProcessRunner({createProgram: createTestProgram,