This fixes purchase price calculation on requisition lines when planning dates would place the order date before the work date. The code now keeps price calculation on the work date while preserving the calculated starting date unless the Order Date field is validated directly. The main changed paths use Order Date as the price document date, so the old-price case is addressed without changing posting code or public APIs. The added regression test covers two vendor price periods and checks that the current price is selected.

No suggestions.

Risk: This touches requisition worksheet planning dates and purchase direct unit cost. A mistake here could choose the wrong vendor cost on planned supply, but the change is narrow and does not change posting code, event signatures, or public procedure signatures.

Necessity: The linked bug is valid because a rush order can otherwise use an obsolete purchase price when lead time pushes the calculated order date into the past. The scope is right for the issue, and the tests cover the price regression plus related planning date expectations.

Same fix as round 1: on requisition lines, keep purchase price calculation on the work date instead of an order date in the past, and only update Starting Date when Order Date is validated directly. The single new commit ran Invoke-Miapp and propagated the RequisitionLine.Table.al change to the IT layer and added the matching country-layer test copies. The IT BaseApp change is identical to the already-reviewed W1 change, so the guard against a past order date is applied the same way in both layers.

Round 1 raised no suggestions, so there is nothing to track here.

None. The new commit only propagates the accepted W1 logic to the IT layer and adds the parallel test copies; the production logic is unchanged and consistent across layers.

Risk: Still limited to requisition worksheet planning dates and purchase direct unit cost. No posting code, event signatures, or public procedure signatures change. The country-layer propagation matches W1 byte-for-byte, so it adds no new behavior.

Necessity: Unchanged from round 1. The bug is valid and the scope is right; propagating to the country layer is the correct follow-up.