AR-150

Error during export phase If PKPurgeQuery is empty

Bug

Fixed the processing of the repl_ExportTasks.PKPurgeQuery value: the query is considered as an empty string when the value contains nothing except spaces or tabs:

AR-167

Wrong exit code when an updater times out

Bug

Replicator returned a wrong exit code (-1000) when waiting for an update process which exceeded the specified timeout. The issue is fixed.

AR-202

Pick up only active sites to update parameters group

Change Request

Added the Show active sites only checkbox to the update parameters page. Instead of picking up active sites only, users can display all the sites they might need.

Active sites are listed by default (the checkbox is prechecked).

AR-211

Plugin Log grid - To add the sorting by column

Change Request

Added the possibility to sort rows in the grid by column. To sort grid rows in ascending order, simply click the column name. Clicking the same column name again will sort the rows in descending order.

Clicking a column name makes it displayed in bold. It means that the sorting mode is enabled.

To disable the sorting mode and return to the original view of the grid, click on the leftmost (empty) cell at the top of the cell.

By default, the grid rows are ordered first by the Event Time column, then by the Pin column.

Sorting only determines how the data that the replicator manager received from the service is displayed. At the same time, the data itself does not change its order. Therefore, sorting does not affect the set of data that the replicator service passes to the manager when the navigation button is clicked.

AR-265

Replication Monitoring Service: to display the list only of Active sites in the Android application

Change Request

The Hide inactive checkbox is added to the Replicator Monitor Android application.

If the checkbox is selected, then the application shows only active sites.

AR-268

Invalid pointer operation

Bug

Fixed the Invalid pointer operation error. The following enhancements have been made:

• Removed repeated calls of TWebSocketConnection.StartRead method from the message event handler (WSClientMessage method) for better matching between the TWebSocketConnection class and the WebSocket protocol. From now, StartRead is called only once in the lifetime of a websocket connection object. After switching to the new design of reading income websocket messages (one call of StartRead instead of many), the ‘invalid pointer operation’ error is no longer available since the _Release method is no longer called in the income message handler, which means that the connection object cannot be destroyed inside this handler.

• Fixed the problem of missing HelloFromServer message in an open websocket connection to the monitoring server and added parsing data of the receive buffer to the StartRead method to handle messages that could come from the monitoring server before this method is called. In particular, if there is a HelloFromServer message in the read buffer, the replicator parses it and sends a ReplLoginRequest message from the StartRead method in order to authenticate with the server and complete the handshake procedure. Parsing of subsequent data from the server is performed as before – from the incoming message handler when new data is received.

AR-270

Adjusting the warning message for the case when export is performed for site with missing channel direction types

Change Request

Added a clear error message in case the tasks for a channel are misconfigured on the Export stage:

AR-297

Updates the city name or zip code in the office, creates duplicates with replication

Bug

Implemented the replication solution which allows handling the insert/update of the record into the tables with unique indexes.

If unique key violations occur during import, the replicator puts such rows into the import storage and continues importing until it processes all task files from the CAB (RPL) file. If a CAB file contains a task for repl_Deleted table, the replicator searches the import storage for records with error 2627 “cannot insert duplicate key” at the end of the CAB file import. If the Replicator finds such records in the import storage, it will try to import them.

AR-289

Update status windows aren't showing the correct current status for other sites

Change Request

Added the upgrade enhancements:

• If Site was upgraded before via Repplicator, then the ReplManager shows the latest update date/time of the application on the Site

• If Site has never been updated before via Replicator, then ReplManager shows the following message in Status: “Version XXX is available on the site“

AR-290

Add the "Hide inactive sites" checkbox to the Android app

Change Request

The Hide inactive checkbox is added to the Replicator Monitor Android application.

If the checkbox is selected, then the application shows only active sites.

AR-292

Plugin Log: the error when clicking on the empty grid

Bug

Fixed the error that occurred when a user clicked an empty grid in the PluginLog view.

AR-293

Handling of unique indexes by the Replicator

Change Request

Implemented the replication solution which allows handling the insert/update of the record into the tables with unique indexes.

If unique key violations occur during import, the replicator puts such rows to the import storage and continues importing until it processes all task files from the CAB (RPL) file. If a CAB file contains a task for repl_Deleted table, the replicator searches the import storage for records with error 2627 “cannot insert duplicate key” at the end of the CAB file import. If the Replicator finds such records in the import storage, it will try to import them.

AR-296

Handling of calculated columns by the Replicator

Change Request

The replicator can handle persisted computed columns properly: they are included into export queries and they are correcly handled by the import queries.

AR-300

Update documentation on the auto-update functionality

Change Request

Updated the manual with the auto-update functionality.

AR-303

Change the static loading of sqlite3.dll to dynamic loading

Change Request

Added the dynamic load of sqlite3.dll. Replicator can now run successfully even if this file is not available. Without this DLL, the Replicator can run replication and perform other functions except for saving replication events and sending them to the monitoring server (that is the monitoring of the current site gets disabled). The absence of sqlite3.dll will not prevent Replicator from connecting to the monitoring server and transmitting to it monitoring events received from other sites as heartbeat files.