Our offline FR5 result tracker lets an old attempt's completion finish the newer attempt when it arrives late; how should I make that sequence repeatable so the developer can test the fix, rather than hope a random delay reproduces it?
How can I replay a delayed completion without random sleeps?
RaviBarnes0563 · 2025年6月5日 10:03 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
2 条回复
Use a controlled event queue and test clock so you choose the order explicitly: start the first attempt, start the second, then deliver the first completion. Do both request and completion carry an attempt identity the tracker can compare?
10分They only carried the reused job label; the revised fake interface now carries a separate attempt identity, and the queued late completion leaves the newer attempt waiting. The real interface change still needs agreement with its owner.
10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： .