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。
BenAbbott0022 · 2026年6月25日 22:53 UTC
What is the counter reading from: accepted attempt rows, or a saved total incremented as results arrive? Those fail differently.
15分Accepted attempt rows. Restart rebuilds the total from them, including two accepted attempts belonging to the same coupon. The running display had hidden that duplication.
15分Same coupon, twice. Does each attempt already carry the coupon identity, or would somebody have to infer that now?
21分And who decides which disposition remains current when a later review changes it, because I couldn't hand maintenance a rule that just picks whichever row arrived last
17分Louis, agreed about arrival order. But choosing one accepted attempt also misses a later withdrawal of acceptance; the item disposition needs its own reviewed history.
7分Grace, both attempts have the same coupon ID. Louis and Leah, quality owns the disposition sequence; the developer is using its current reviewed state, not result arrival order.
3分Keep a coupon with no completed decision in the test as well. Missing isn't a quiet pass.
12分Leah, yes, I was thinking of choosing a current decision rather than an accepted attempt; thanks for making that distinction explicit before it became my misleading handover sentence
13分The rebuild now gives 183 accepted, nine rejected and five awaiting disposition, matching quality's 197-item ledger. Duplicate and reordered attempt deliveries leave those totals unchanged; history remains available.
7分Withdrawal and reinstatement tests also passed against the ledger, including restart between decisions. The dashboard and export now use that same disposition view. Reconciliation closed.
18分