(This is a pretty dry technical cleanup build with various fixes and small improvements, as well as a very important multi-level addition for Unicode names in databases which we want to get out as soon as possible. We are following up with the UI build soon.)
Dataports for remote databases support Unicode names for all database objects and all schema items. (Temporary exceptions: ODBC and Oracle.)
(Backstory: We have always supported Unicode in string values and string parameters. However, support for Unicode names of fields and tables has been very sporadic. The dataports relied on the locale of the connecting user to be similar to the database locale and even that was not very consistent and did not work in all places. The change extends all database connections with new logic that allows using Unicode names in queries according to the rules for the specific database, and adds that logic for all supported database types. Two temporary exceptions are ODBC and Oracle, we are working to allow Unicode names there as well. The change allows seamlessly working with tables, views, fields, indexes and other native database constructs with Unicode names both from the UI and programmatically.)
Connecting to a DB2 database uses ODBC instead of the previously used interface specific to DB2 (CLI). (CLI was similar to ODBC and over time the differences between the offered functionality were becoming ever smaller until they almost completely disappeared. We no longer see any need to support a specialized interface that requires separate configuration. A note to DB2 users: connecting to a 64-bit version of DB2 requires installing the latest fix pack dated December 7, 2017.)
Database.Export and Database.ExportFile script functions accept empty string for a component name to export the entire database.
(Fix) Exporting a database to a MAP or MXB file correctly handles autogenerated fields. (The fix allows exporting data sets like SQLITE or GDB.)
Exporting a database to a MAP file builds intermediate levels for images with an rtree index on x-y-tile fields.
Opening an MXB file builds intermediate levels for images with an rtree index on x-y-tile fields.
File - Save As sets the initial filename to the currently used one.
File - Export Project sets the initial filename to that of the currently opened MAP file.
Exporting an MXB file reports the number of exported records and bytes.
SELECT INTO reports the number of inserted records and bytes.
Linking an ADF or GDB data set sets the name of the created data source to that of the folder.
Linking an SDTS data set sets the name of the created data source to the filename prefix (SDTS data sets consist of multiple files named xxxxYYYY.DDF with the YYYY.DDF parts dictated by the standard).
Dropping an MXB file from Windows Explorer into the application window opens it.
Installation packages register the extension and the icon for MXB files. (This particular change cannot be tested using a portable install.)
(Fix) Reading a PNG file no longer sometimes misinterprets transparent pixels as black.
(Fix) Reading a malformed PNG file no longer has a potential to crash.
Reading a 3d geometry value in a SHP file forces invalid Z values such as NaN or Inf to 0.
Many tables in ADF and several other dataports report the number of records they contain.
(Fix) Connecting to a WFS server no longer fails to read tables with no identity data. (If the server exposes a table without identity data, it is better to turn off the cache for it. Without caching, zooming to a particular extent in such a table will request data only for that extent. With caching, zooming to a particular extent will try to avoid repeat requests and since the table has no identity data to determine duplicates, the system will be fetching an entire table.)
End of list.