ci(autofix): number the status comment like every other round message - #7748
Merged
Conversation
`effective_round` counts rounds already finished, and "Push and report" posts ROUND + 1 as the round it just performed. The status comment used the raw value, so the same round showed two numbers in one thread — observed on #7724 at 10:10 UTC: the report said "round 6/100" while the status said "AutoFix round 5 finished". Display ROUND + 1 in both status messages, guarded so a missing or non-numeric round degrades to the raw value instead of failing the step.
Collaborator
Author
|
@qwen-code /takeover |
qwen-code-dev-bot
added
the
autofix/takeover
Collaborator
|
✅ Qwen Triage finished — CI landed green on ✅ Qwen Triage 已完成 —— |
Collaborator
|
🤝 Takeover engaged: the autofix loop now manages this PR — it will address new review feedback and resolve base conflicts until the label is removed or the round cap is reached. Remove the 中文说明🤝 已接管：autofix 循环现在管理此 PR —— 将持续处理新的评审反馈与 base 冲突，直到移除标签或达到轮次上限。移除 |
Collaborator
|
Thanks for the PR! Template looks good ✓ Problem: observed bug with concrete evidence — on #7724 the status comment said "round 5 finished" while the round report in the same thread said "round 6/100". Same round, two different numbers. That's exactly the confusion the status comment exists to prevent. Direction: aligned. This is a display-only consistency fix in the autofix CI loop — no product direction concerns. Size: not applicable (no core module paths touched). Approach: the scope is exactly right — two identical 6-line additions at the two status comment sites, incrementing Moving on to code review. 🔍 中文说明感谢贡献！ 模板完整 ✓ 问题：已观测到的 bug，有具体证据——在 #7724 上，状态评论写 "round 5 finished"，而同一线程中的轮次报告写 "round 6/100"。同一轮出现两个编号，恰恰是状态评论本该消除的困惑。 方向：对齐。这是 autofix CI 循环中的纯显示一致性修复，无产品方向顾虑。 规模：不适用（未触及核心模块路径）。 方案：范围恰好——在两个状态评论位置各加 6 行相同代码，将 进入代码审查 🔍 — Qwen Code · qwen3.8-max-preview Reviewed at |
Collaborator
Code ReviewIndependent proposal: the round report numbers the round being performed ( Comparison with the diff: the PR does exactly this, and nothing more. Both sites get the same 6-line block — set One thing I verified: No findings. TestingThis is an unattended CI run — PR code is never executed here. Evidence comes from the PR's own CI checks: Final CI results for
One row per check name (latest run); skipped checks omitted; failures sort first. / 每个检查名一行（取最新一次运行），省略 skipped，失败项排在最前。 Ubuntu unit test is still running. macOS/Windows/Integration are skipped (expected for this change — CI-only, no runtime code). The finalize workflow will update this table when CI settles. Not verified: live autofix round behavior (requires a real autofix run; the author reports exercising the display logic directly and mutation-testing the pinning assertions). 中文说明代码审查独立方案： 轮次报告按正在执行的轮次编号（ 与 diff 对比： PR 完全按此方案实现，且无多余改动。两处各加相同的 6 行代码块——设置 已验证： 无发现。 测试这是无人值守 CI 运行——此处不执行 PR 代码。证据来自 PR 自身的 CI 检查。Ubuntu 单元测试仍在运行中。macOS/Windows/集成测试已跳过（符合预期——仅 CI 改动，无运行时代码）。finalize 工作流会在 CI 结束后更新此表。 未验证：实际 autofix 轮次行为（需要真实的 autofix 运行；作者报告已直接验证显示逻辑并对钉住断言做了变异测试）。 — Qwen Code · qwen3.8-max-preview Reviewed at |
Collaborator
|
Confidence: 5/5 — clean, minimal fix for an observed inconsistency; would merge without hesitation. This is exactly the kind of PR the gate should wave through quickly. The bug is real and concretely evidenced (two messages in #7724, same round, different numbers). The fix is the smallest possible change — two identical 6-line blocks that increment the display value, guarded so a missing round number degrades instead of crashing a best-effort step. No control flow touched, no round accounting changed, nothing beyond what the stated goal needs. The test pins both sites and the author mutation-verified the assertions. My independent proposal matched the PR's approach exactly — I didn't find a simpler path it missed. If I had to maintain this in six months, the comment at each site tells me precisely why the increment is there. Approval deferred until CI lands green on 中文说明置信度：5/5 —— 干净、最小化的修复，针对已观测到的不一致；毫不犹豫地合并。 这正是 gate 应该快速放行的 PR。bug 真实且有具体证据（#7724 中两条消息，同一轮，不同编号）。修复是最小可能的改动——两个相同的 6 行代码块递增显示值，加守卫使缺失的轮次号降级而非让尽力而为的步骤崩溃。未触及控制流，未改轮次计数，没有超出目标所需的任何改动。测试钉住了两处，作者还做了变异验证。 我的独立方案与 PR 方案完全一致——没找到更简路径。如果半年后维护这段代码，每处的注释会告诉我为什么需要这个递增。 审批延迟至 CI 在 — Qwen Code · qwen3.8-max-preview Reviewed at |
Contributor
Code Coverage Summary
CLI Package - Full Text ReportCore Package - Full Text ReportFor detailed HTML reports, please see the 'coverage-reports-22.x-ubuntu-latest' artifact from the main CI run. |
qwen-code-ci-bot
approved these changes
Jul 26, 2026
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
LGTM, looks ready to ship — CI landed green after the review. ✅
qwen-code-ci-bot
approved these changes
Jul 26, 2026
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
No issues found. LGTM! ✅
中文说明
未发现问题。LGTM！✅
— qwen3.7-max via Qwen Code /review
Collaborator
Independent local verification report (Linux)Verdict: verified — the off-by-one premise is real in the workflow source, the fix's shell behaves exactly as claimed on all three inputs, the pinning test kills the mutation, and the test-failure set is byte-identical to main. 1. Premise check (the report really does print ROUND + 1)The round report step computes 2. Shell behavior, executed directly (tmux)Extracted the exact patched snippet and ran it:
3. Tests — attribution and mutation
ConclusionThesis: one thread, one numbering convention — and the convention to converge on is the one the report (and the markers) already use. Evidence: the premise located in the workflow source (§1), the three claimed input/output pairs reproduced by executing the real snippet (§2), and an identical failure set plus a mutation-killed pin (§3). Display-only, guarded, matches its description exactly. LGTM. 中文摘要Linux 本地独立验证：前提属实——报告步骤用 — independent review loop, executed-snippet verification on Linux |
Collaborator
Collaborator
|
🤖 Reviewed the latest feedback — no changes needed. Why, point by point: · 已审阅最新反馈——无需改动。逐点说明原因如下： No changes this roundThere is no actionable feedback to address on PR #7748.
The PR stays as-is: a display-only, guarded change that numbers the status 中文说明本轮无需修改PR #7748 没有任何需要处理的可执行反馈。
PR 保持现状：一处纯显示层、带保护的修改，让状态评论采用与轮次报告及标记相同的「正在执行的轮次」编号约定，外加一个钉桩测试。 Base-conflict check · 基分支冲突检查: no conflict with main. · 与 main 无冲突。 🧠 Handled by Qwen Code · model/模型 |
yiliang114
approved these changes
Jul 26, 2026
yiliang114
left a comment
yiliang114
left a comment
Collaborator
There was a problem hiding this comment.
Approve ✅
Clean off-by-one fix for #7738's status comment. ROUND counts completed rounds, but the report posts ROUND + 1 — using the raw value made the same round carry two different numbers in one thread. The ^[0-9]+$ regex guard before arithmetic is correct defensive coding. Tests pin both steps (post + finalize). No issues.
github-actions
Bot
added
the
skip-changelog-auto
Collaborator
|
Released in v0.21.1. |
Collaborator
|
Acknowledged — PR #7748 ( ✅ completed |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this PR does
Fixes an off-by-one in the AutoFix status comment (#7738): it now numbers the round being performed, like every other message the loop posts.
Why
effective_roundcounts rounds already finished, andPush and reportpostsROUND + 1as the round it just performed. The status comment used the raw value, so the same round carried two different numbers in one thread.Observed live on #7724 at
10:10 UTC— both written by the bot, minutes apart, about the same round:🤖 Addressed the latest review feedback (round **6**/100)✅ AutoFix round **5** finishedThat is exactly the confusion the status comment exists to remove, so it is worth a follow-up rather than leaving it.
How
Display
ROUND + 1in both status steps, guarded by a numeric check so a missing or non-numeric round degrades to the raw value instead of failing the step (it is a best-effort step — it must never cost a round).Reviewer Test Plan
effective_round=5→ shows 6 (matches the report'sround 6/100); emptyeffective_roundfalls back toROUND=3→ shows 4; a non-numeric value stays as-is and does not crash.npx vitest run scripts/tests/qwen-autofix-workflow.test.js— 103/103 (two consecutive clean runs; the suite's known load-flakes appeared in one earlier run and are unrelated).bash -nclean on both run blocks, actionlint 16 vs main 16 (no new findings).Risk & Scope
ROUNDitself is untouched — only what the status comment prints.Linked Issues
Follow-up to #7738. Defect observed on #7724.
中文说明
本 PR 做了什么
修复 AutoFix 状态评论(#7738)的 off-by-one:现在按正在执行的轮次编号,与循环发布的其它消息一致。
为什么
effective_round统计的是已完成的轮数,而Push and report发布的是ROUND + 1(它刚执行的那一轮)。状态评论直接用了原始值,导致同一轮在同一线程里出现两个编号。在 #7724 于
10:10 UTC实测到 —— 两条都由 bot 发出、相隔几分钟、指同一轮:轮次报告写（第 **6**/100 轮）,状态评论写✅ AutoFix round **5** finished。这恰恰是状态评论本该消除的困惑,因此值得一个跟进修复。
怎么做
两个状态步骤都显示
ROUND + 1,并加数字校验:若轮次值缺失或非数字,则退回原始值而不是让步骤失败(它是尽力而为的步骤,绝不能拖垮一轮)。评审验证
effective_round=5→ 显示 6(与报告的round 6/100一致);effective_round为空 → 回退ROUND=3→ 显示 4;非数字值原样保留且不崩溃。npx vitest run scripts/tests/qwen-autofix-workflow.test.js—— 103/103(连续两次干净;该 suite 的已知负载 flake 在更早一次出现,与本 PR 无关)。bash -n干净、actionlint 16 vs main 16(无新增)。风险与范围
ROUND本身未动,只改状态评论打印的内容。关联 Issue
#7738 的跟进;缺陷在 #7724 上观察到。