AI, in your work.

What changes

Microsoft merged request #13272 into the main BCApps branch on October 27. It describes an issue with backdated payment unapplication: unrealized exchange differences could shift between periods while the overall balance still reconciled. Original request.

The change replays subsequent revaluations using their recorded dates and currency factors. Its stated scope covers customers, vendors and employees in W1 and RU. The inspected diff includes adjustment logic and regression tests; Nexo has not run them. Recorded commit.

Why it matters at period end

Business Central allows users to preview entries before posting an exchange-rate adjustment. Microsoft also explains that unapplying and reversing a previously adjusted payment creates reversal entries and may require another adjustment to obtain the correct current balance. Currency documentation.

Nexo interpretation: an overall reconciliation can conceal incorrect allocation between periods. Acceptance criteria should check the balance at each reporting date and the entries explaining any differences, keeping settings and filters consistent throughout the comparison.

Availability and qualifications

The backports to releases/29.x and releases/29.0 were merged on November 5. Their milestones are Version 29.1 and Version 29.0, respectively. These references identify code destinations; the inspected pages do not establish a distributed build or its installation in a customer environment. 29.x backport, 29.0 backport.

An automated review published on both requests flags a possible edge case: an adjustment rounded to zero could update the factor without creating the detail needed to replay it. This is a review observation without test execution, not a result reproduced by Nexo. The subsequent merge alone does not establish that the concern was resolved. Review on 29.0.

The behavior of individual installations when posting dates are blocked is also not established. Before applying a conclusion to production, the responsible team should identify the build, localization, extensions and date restrictions in the environment being evaluated.

Fictional example: the total does not tell the whole story

Suppose a test company expects to retain the following historical balances after completing an application and unapplication cycle. These synthetic figures are in local currency; they illustrate a comparison and neither reproduce the defect nor represent Business Central results.

Turning this into a functional test

In this example, July matches and June differs by 100. These are balances at different dates and should not be added together. The matching final balance neither explains nor validates the June change.

Nexo's proposed approach for a test company:

• Save reference historical balances and their supporting entries before the application and unapplication cycle.
• Compare every reporting date using identical filters, and explain any difference through specific entries.
• Include partial payments, multiple subsequent adjustments, repeated cycles and adjustments that round to zero.
• Check posting-date restrictions separately, without assuming that the correction permits operations in blocked periods.
• Record the build, localization and extensions alongside the observed result; do not extrapolate a successful test to other environments.