What changed, between which versions

attachments are copied to the companion archive database (<Database>_Archive) before delete, to keep the documents for the 7-year e-invoice retention requirement. Optional "Reclaim space" step (LOB compaction / shrink) after cleanup.

Non-critical start-up work (version check, Shopee and SQL Express checks, DynaMod Update, Tip of the Day, GST/SST/E-Invoice tips, fiscal and tax validators, alerts, unread-message badge) now runs about half a second after the menu is usable, so no pop-ups interrupt the loading screen. Measured login-to-ready on a warm start: about 3 to 4 seconds instead of about 3 to 9.5 seconds.

Best for low-activity months – about 23% faster on the Regenerate Stock Cost step in testing, same figures.

block a stock item from being used on Sales, Purchase or Inventory documents, by unit of measure and by location, so a discontinued item can keep selling down its remaining balance while reordering stops.

supply location reviews incoming requests, adds them into a Picking List, enters Accepted Qty, then rejects with reason or generates DO / Stock Transit Request (cross-database supported).

Stock Month End Closing Process and Regenerate Stock Cost now run much faster on large databases, and the month-end close runs in the background on the server with checkpoint, resume and cancel.

The working table is now indexed and rows are processed in a fixed order. Measured on a large production database (372 GB, 9,163 stocks, Weighted Average): the full Regenerate Stock Cost step takes about 55 minutes instead of about 27 hours, and the deepest single stock about 17 seconds instead of about 39 minutes.

It is now calculated once per period, with supporting indexes: about 4 minutes instead of about 60 minutes. The whole month-end close now completes in about 1 hour instead of about 27 hours.

Databases on those methods still gain from the faster Closing Stock Value steps and the background flow below.

the run continues on the server (SQL Server Agent; DynaMod Process Manager on SQL Express; otherwise in-process). Re-open the screen at any time to re-attach and follow the live progress log, or request a cancel. Every step, period and location is checkpointed, so an interrupted run resumes where it stopped instead of starting over. Month-end and the Process Manager deferred cost recalculation no longer block each other on the same stock.

the new process was compared row by row against the original method on production data: identical stock cost for every stock, and identical Closing Stock Value across all 134,272 detail, location and company rows.

Each page now fetches exactly one page of rows however deep into the list you are, instead of re-fetching everything from the first page. Scrolling to the bottom loads the next page; scrolling to the top loads the previous page and keeps the current record in view; the page selector only fetches pages that are not already loaded. Ordering now includes the document key, so pages can never overlap or skip records. Stock Maintenance gains scroll paging (it was not connected before).

Promotion Studio is a single console for every promotion in the company, and Promotion Builder is one editor that replaces all eight of the old promotion wizards.

Stock, Mix & Match, Range Quantity Discount, Buy & Free, Purchase With Purchase, Cash Rebate, Add On and Member Group — through the same four steps: Plan → Schedules → Batches & rules → Items.

Each one can be applied, acknowledged, deferred or removed, and a nightly sweep catches anything changed outside the usual screens.

request stock from another location/subsidiary by Stock ID / Description / Barcode; requests are auto-numbered and tracked in Saved Item / Processing / Rejected / Completed tabs.

1. Last Purchase 2. Last Purchase Include Quantity & Discount 3. Supplier SKU Price 4. Reference Cost 5. System Cost

Measured on a customer database: the Cash Sales listing sort takes about 0.14 seconds instead of about 89 seconds; Customer Invoice and Supplier Invoice listing sorts are 2 to 3 times faster. The indexes are created by the 2026.2 upgrade script.

user can continue working and open the report from My Reports once it is ready.

User can open a Ready report, cancel a running preparation or delete old runs; the screen auto-refreshes.

new one-pass engine closes all locations for a period in a single pass: about 7 minutes per month instead of about 1.9 hours (tested on 294 locations). A 42-month backlog now takes about 5 hours instead of about 81 hours.

This also speeds up General Ledger Listing (Show B/F), Trial Balance and Balance Sheet.

the new process was verified against the original method on production data (74 of 74 affected tables matched by checksum; row-level report comparisons showed zero differences).

Financial reports now run much faster on large databases: Profit and Loss, Trial Balance, Balance Sheet (including the Multi-Level versions) and General Ledger Listing with Show B/F.

The reports now build only the combinations that actually carry figures. The gain is largest on multi-location / multi-department / multi-job databases with Suppress Zero on; small databases are unaffected.

Preview now takes about 31 seconds instead of about 50 seconds without filter, and about 18 seconds instead of about 48 seconds with a location filter (3 completed fiscal years; the saving grows with the number of fiscal years).

the retained earnings and current-year profit adjustment previously calculated in the DynaMod client is now done in the database. Together with Profit and Loss and Profit and Loss Multi-Level, they can be run in the background (see Reports Enhancement – Run in Background).

every changed report was compared row by row against the original method on production data with zero differences. Other grouping options and customised report queries continue to use the original method.

(Ctrl+Alt+C & Ctrl+Alt+ V)

indexes are now maintained based on their actual fragmentation level (rebuild heavily fragmented indexes, reorganize lightly fragmented ones, skip healthy ones) followed by update statistics, instead of rebuilding every index blindly. Much faster on large database.

Options include Online rebuild (Enterprise edition only), Sort in TempDB, MaxDOP, statistics sample %, time limit and resume prior run.

user can close the screen and continue working; re-open the screen to attach to the running job, view the live progress log or cancel it. This also resolves the middleware out-of-memory error when maintaining very large database.

On SQL Server Express or when Agent is not running, the screen automatically falls back to foreground mode — the active mode is shown in the log header.

it now queues the maintenance as a background request and the upgrade completes immediately.

user can close the screen and continue working, and check the progress anytime from the Requests tab. Runs can be cancelled and resumed.

the run is refused with a clear message stating the space needed vs available if there isn't enough space (checks the source data/log, archive database and snapshot volumes).

Message shown:

The gain is largest over branch or VPN links – about 2.5 seconds saved per login on a 40 ms link and about 7.5 seconds on a 120 ms link. The full saving needs the 2026.2 Server Socket; an older Server Socket still works, the client simply falls back to the previous calls. Deploy the Server Socket before the clients. The splash message now reads "Reading licence from server socket" – the old "Verifying license with activation server" wording was misleading, as no activation server is contacted.

Verifying licence, Checking version updates, Loading company settings, Validating fiscal/tax, Checking integrations, Opening menus, Ready – with a moving progress bar on longer phases so the application never looks frozen. A bridge screen covers the gap between the login dialog closing and the main window appearing.

Cash Sales Advance Search on very large databases is otherwise unchanged in this release.

Trial Balance, Profit & Loss, Profit & Loss Multi-Level, Balance Sheet and Balance Sheet Multi-Level now generate up to 8–37× faster (procedures pfims71900, pfims71910, pfims72460, pfims71920, pfims72470), with identical report results.

the live log now shows "[INFO] EDITION [ONLINE not supported — rebuilding OFFLINE]" and the run continues offline.