Independent review of full PR diff at 3afac00 vs main. Blocking P1 remains: the new title-term stripping preserves topical fallbacks for the two examples from #1701, and excludes…
Независимое ревью актуального diff 9832c0c..cf510f0: обнаружен blocking finding (P1) в backend/search/query.py:44-49,139. is_named_act_query отключ…
Проверил diff PR #51 на коммите 1ce4349: блокирующих замечаний нет. Смена email сериализуется с выдачей reset/resend токенов через блокировку user row; регрессионный тест покрывает этот сценарий. Проверил mail logging, mobile CSS, APP_BASE_URL, версии и changelog. Backend integration tests не запускались: в среде нет подключения к PostgreSQL.
Final independent review of the complete PR diff at cf510f0 vs main, including CHANGELOG.md and prior query fixes. Code findings: none; no blocking code defects found. The minimum-two…
Independent review of the full PR diff at a3d82b4 vs main. Code findings: none remain. The fallback now requires at least two analyzed matches even for short clauses; the regression test…
Independent review of the full PR diff at 6f8b668 vs main. One code finding remains: [query.py:143-156] counts whitespace words before analysis, then uses 75% for 3+ words. The mapped RU…
Independent follow-up review of 3b56445 and the full PR diff vs main: blocking findings remain.
- P1 (residual single-token fallback): the 75% threshold still rounds down for the…
Independent review findings (formal REQUEST_CHANGES review is unavailable because Gitea rejects review by the PR author):
- P1:
backend/search/query.py:138-151lets each relaxed `multi_match…