Why has our UR5e nearly finished by checking the same brackets again?

Item labels exist. Screen totals successful checks, including repeats. Quality has current decisions on a separate list.

Reconcile that list to the 112 physical items first. Keep accepted, held, rejected and not yet inspected distinct. A held bracket is not spare capacity for the counter to fill with another successful retry.

Can the developer replay the existing checks against those item identities offline? That should expose the duplicate problem without putting more brackets through inspection just for the dashboard.

And let the operator open one bracket's history from the current disposition, so failed checks stay understandable instead of disappearing when the total is corrected.

Quality reconciled 103 accepted, four held, two rejected, three uninspected. All 112 accounted for. Replay arranged.

Who can change a disposition after that reconciliation? A later quality withdrawal needs to reach the item view and count, not remain on another list which the dashboard never sees.

Test that withdrawal explicitly. Start with an accepted test bracket, withdraw it, then deliver its earlier successful check again. It should not creep back into accepted because the event arrived late.

Quality owns changes. First replay fails AmyAbbott's case: an old successful check reinstates the withdrawn bracket.

Then the developer needs the quality decision ordering as well as duplicate detection. Preserve who made the withdrawal and why. The latest received machine result cannot silently cancel it.

Jo, would the revised screen show that bracket as held with the reason, rather than displaying a successful check beside an unexplained zero in the count?

Yes. Current disposition above the checks, with quality's reason. Tutor preferred that to redrawing every old result.

Revised replay passes duplicate, withdrawal and delayed-old-result cases. Restart also preserves the withdrawn test bracket's state.

Does a genuine later reviewed acceptance still work? The fix must not freeze every withdrawn bracket in review forever.

It does. Installed comparison reconciles to 103/4/2/3, with test withdrawal and reviewed acceptance behaving correctly. Shift lead checked item histories.

Did the three uninspected brackets stay plainly uninspected during that check? Those are the ones I'd expect somebody to mistake for missing data when they first use the new view.