Qwen Code deflakes session-ID E2E test by stopping at init message instead of full completions
qwen-code-dev-bot · ghdev · 2026-10-18
PR #14585 on QwenLM/qwen-code deflakes the SDK E2E test 'should generate different session IDs for different queries.' The test's real subject is that two SDK queries yield two distinct auto-generated UUIDs, but the old version drained both streams to completion, running two concurrent real model calls it never inspected — any post-startup failure (stream error, killed container, timeout) surfaced as a spurious test failure. The rewrite reads each query only up to its init system message, which echoes the CLI's session ID before any model call, and adds an assertion that this echo matches the SDK's getSessionId() — stronger coverage with far less flake surface. All 15 tests in the suite pass.
More from coding & agent
- Claude cut a 2.6-hour video export to 20 minutes — but only after being told to rebuild the server — byhewar · 2026-10-23
- 3 flight-booking rules my AI agent checks that I'd never do by hand — Efistoffeles · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23
- How do you audit AI agent decisions in regulated workflows? A field survey — Ok_Way_1993 · 2026-10-23