The display says 166 of 171 and nearly finished. Quality has released 133 brackets. I checked the event list: repeats after failed checks and restart are inflating the display.
I want accepted brackets counted once, with the failed attempts still there. Deleting the failed rows would make a prettier number and a rotten inspection history.
Count each identified bracket from its current authorised disposition, with attempts linked underneath. How many are held and how many have not been inspected yet? The remainder needs a meaning too.
Seventeen held, twenty-one untested. That makes 171 with the 133 accepted. There are 150 distinct inspected brackets; sixteen additional attempts explain the 166 checks.
Drop nearly finished now. Held and untested are different work, and neither disappears because the attempt total approaches the order quantity. Your planner needs those three groups while the report is being fixed.
Who owns changing a held bracket to accepted after a repeat? A passing attempt should not silently become the final decision if quality still has something unresolved.
Quality owns that change. The developer has separated the attempt result from disposition in the report. The retained test set now shows 133 accepted, 17 held and 21 untested without deleting any attempts.
Both tests pass in the replay. Repeated import adds no accepted bracket; withdrawal and restoration follow quality's decision. The stored item record, screen and export agree on the original test totals after restart.
Good. Has the normal workstation reproduced that comparison? I would want the covering person to find a held bracket and see why it is held, not just admire the corrected total.
Normal workstation comparison passed, and the held reasons are accessible from the group list. Quality has since released five held brackets: now 138 accepted, 12 held, 21 untested. Report and export match that list. Counting fault closed; those remaining brackets still have work outstanding.