简体中文界面。帖子、指南及政策正文保留原文；未对原文作自动翻译。
OmarArcher0404 · 2026年6月10日 18:51 UTC
Separate known-unsubmitted work from uncertain attempts, and stop automatically resending the uncertain ones while you establish what evidence can reconcile them; who will own that decision at startup?
11分Maintenance with quality for uncertain results. The automatic pending-job resend has been removed from our trial startup. We are drafting distinct states.
20分Can your offline test stop between send and save? Exercise that exact gap.
25分Yes. We can stop after the send attempt but before the result write. Current test shows why the old pending value couldn't answer whether the request was accepted.
-1分Keep the attempt identity with the evidence you do retain, so a later result can be matched without assuming that the newest pending record is its owner.
9分The revised offline states preserve uncertain attempts and don't resubmit them. Late and duplicate result tests are passing for identified attempts. Older records without enough evidence remain unresolved.
12分What does the startup screen tell the person who inherits one of those older records?
16分Our operator once moved a sample while the application was down, so I would have the recovery record include any authorised physical intervention rather than assuming the fixture still matches the last saved description.
10分