Same bracket, new attempt. Delayed old completion finishes it. Random sleeps make the test unreliable.
Old FR10 completion marks the next offline attempt complete
KaiBennett0715 · 2026年3月25日 06:23 UTC
主帖中文翻译暂不可用，以下显示英文原文；回复保留原文。 查看英文原文
6 条回复
Control the event order directly. Start the old attempt, hold its completion, start the next attempt, then deliver the held event. What identity does your update check use?
9分Bracket identity only. Explicit event delivery reproduces it every time. Attempt reference is present but ignored.
11分Then don't just throw away every delayed event. It still belongs in the old attempt's history. Check that the new one stays pending until its own completion arrives.
5分Repeat the old completion too. Same event shouldn't add history twice or finish new work on its second arrival.
-1分Fixed the ownership check. Old and repeated old completions leave the new attempt pending; its genuine completion finishes it.
23分Keep that exact order as the regression case, Kai. Much better than hoping a sleep lands in the useful bit. Offline bookkeeping fixed, not a test of the physical inspection.
22分讨论已关闭
此讨论已连续六个月没有活动，现已关闭新回复。 最后活动： .