MarkBlair · 7 Dec 2025, 17:54 UTC
What identity survives with the completion, apart from the coupon name?
9 pointsRecipe number and a local counter. The counter resets when the job is reopened, so its value appears in two runs. No shared run identifier in the export.
22 pointsDoes the duplicate exist in the saved report too, or only on the screen?
20 pointsReplay the retained state offline. Keep the original history present, not an empty test database.
2 pointsOur shift once corrected a display total while leaving the exported report unchanged. The next planner reasonably trusted the file. I'd check both paths before describing this as a colour or counter problem.
15 pointsBoth paths, Anika. Aisha's point applies: the extra row is saved, not just drawn. Developer has a replay with the existing history; the original retained state remains unassigned to either run.
14 pointsKeep that uncertainty after restart too. Otherwise the next launch forgets why it was held.
20 pointsAnd repeat the same delivery. One successful reconnect test may still hide duplicate counting.
7 pointsWho reviews the physical coupon while the software history stays uncertain?
20 pointsMark, if quality can decide the physical coupon, link that reviewed decision without inventing a missing software sequence. We needed both the practical disposition and an honest explanation of why the old report changed.
24 pointsQuality owns that review. The proposed correction links its decision to the coupon and incident, leaving the unknown old attempt unassigned. I'm not treating the eventual physical decision as a recovered log.
23 pointsHas the replay stopped another row appearing in the exported report?
19 pointsIn the developer's replay, yes. Repeated delivery stays one recorded event. Restart with the held reason retained is the next test, so not a completed fix yet.
19 pointsInclude a genuinely different completion too. Rejecting everything would look impressively duplicate-free.
14 pointsAnd show the held reason to whoever covers the next shift.
12 pointsLiam's new-event case matters to the operator explanation as well. People need to distinguish waiting work, completed work and an old uncertain event without assuming every warning means they should repeat the inspection.
15 pointsNew completion and retained-hold restart cases are added. Quality has released the coupon under the review process. Thanks Aisha, that kept the physical decision moving without pretending we found the missing run identity.
19 pointsWho approves the final report adjustment when those software checks are finished?
21 pointsThis discussion is closed to new replies after six months without activity. Last activity: .