qwen-code review misreads comma-grouped counts: "2,239 passed" logged as 239
wenshao · ghdev · 2026-10-09
Issue #13735 in QwenLM/qwen-code reports that the /review Test Plan check misparses counts with thousands separators: "2,239 passed" is recorded as 239, triggering a false mismatch against the test suite output.
Root cause in packages/cli/src/commands/review/test-plan.ts:
- COUNTRES anchors on \b(\d+); a word boundary exists between the comma and the next digit, so matching starts mid-number and captures only "239".
- ruleCount() parses via Number(/(\d+)/.exec(text)?.[1]), which would read just "2" even on a full match.
The author provides three reproduction cases and cites a real instance on #12250 where the bot review carried "239 passed" for a "2,239 passed" description.
Suggested fix: accept \d{1,3}(?:,\d{3})+|\d+ with a (?<![\d,]) lookbehind, strip separators before Number(), and add comma-grouped test cases.
More from coding & agent
- 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
- Capy 0.5.0 lets Claude Pro/Max subscribers reuse their plan quota, adds OIDC cloud access — garrytan · 2026-10-23
- You.com CEO Richard Socher: legacy code is human slop, let AI clean it up — RichardSocher · 2026-10-23
- Let agents design a single-line glance component so you can track them from the sidebar — blixt · 2026-10-23