Repository navigation
[PR] release: arm merge-gated lane publish (lane-side release-bugfix.yml) - #4654
Conversation
…bugfix.yml v2) Same file as master's f009043 (#4653). GitHub executes the BASE (lane-side) copy for pull_request_target events, so this must land on the lane BEFORE the first merge-gated publish can fire. This merge itself cannot publish (at merge time the event still runs the older lane-side copy, which has no such trigger): armed, not fired.
(ai: vllm-qwen/qwen3.8-27b) 查重: 不重复——本 PR 是 owner 2026-10-25 决策(v2 合入门控发布流)的 lane 侧武装件,与 #4653(master 侧同款文件)、#4655(advisory bc-2026-001,首轮流载体)同属一个发布闭环,三者互补不重叠。 已按 [PR] 流程核对代码与 CI(未重做工作),八问如下:
验证证据: 单文件 commit(+156/−42,与正文声明一致)✓ · 与 #4653 逐字节相同 ✓ · YAML 解析通过(triggers: pull_request / pull_request_target / workflow_dispatch)✓ · head 建议序列确认(与正文一致): ① 给 判定: 审核通过,可合入——等你按序列执行(分支保护 → 合并本 PR)。 |
(ai: vllm-qwen/qwen3.8-27b) 标签: |
[PR] release: land the TRUE v3 gate-branch publish model (v2 arm self-fired at #4654)
…to 0.1.206) EMERGENCY MIGRATION COMMIT, bypasses no direct... actually via reviewed PR as usual. Context: #4654 (v2 lane-side arming) was merged BEFORE the model moved to v3 gate-branch publishing. Under v2 rules, the NEXT merge into this lane would publish 0.1.207 without the advisory note or gate-branch review. This PR replaces the lane-side workflow with the v3 file (whose resolve path no-ops for lane-targeted closes) AND temporarily reverts package.json to 0.1.206 so that if the OLD v2 copy still governs this very merge (it executes the pre-merge base-side file), the monotonic guard refuses (0.1.206 == npm latest) and nothing ships. EXPECTED: this run goes RED at the publish guard. That red is the disarm working. Follow-up PR restores the 0.1.207 pre-bump (under v3 it is then a silent no-op).
What
Lane-side copy of
release-bugfix.ymlv2 (identical file to #4653 on master) so that merging a reviewed PR INTO this lane becomes the release action for v0.1.207 (per the 2026-10-25 owner decision; thebackport-shiplabel auto-publish is decommissioned).Why this order is safe ("armed, not fired")
GitHub executes the BASE (lane-side) workflow copy for
pull_request_targetevents. At the moment THIS PR merges, the event still runs the old lane-side yml (no such trigger) → nothing publishes. From the next merge onward, the new copy governs: a merged same-repo PR into this lane publishes its POST-MERGE head (=merge_commit_sha) through the full guard chain, then auto-opens the alignment PR to master. Non-merged closes no-op within milliseconds; fork PRs are rejected.Diff
.github/workflows/release-bugfix.ymlonly (+156/−42 vs the lane's pre-#4596 copy: legacy closed-PR path kept dormant, all version guards / full-CI dispatch+wait / fast mode unchanged, label trigger removed, p_r_target + drift warning + audit comment + auto-alignment-PR added).Suggested sequence (owner)
bugfix-v0.1.206(same status as master; payload in [PR] release: v3 gate-branch publish model (ship PR lane→bugfix fires; auto-cut next lane) #4653 thread).