ABL constructs that support Unicode filenames

A subset of ABL constructs support the use of the full Unicode character set in filenames, including both ASCII and non-ASCII characters.

The following table lists those constructs and the associated elements that can contain any Unicode character. Windows natively stores filenames as Unicode (UTF-16), and for these constructs, the AVM converts filenames between Unicode and the code page set in the startup parameter -cpinternal. Most UNIX/Linux implementations support UTF-8 Unicode filenames. In both Windows and UNIX/Linux systems, the parameter -cpinternal should be set to UTF-8 to ensure that the AVM will be able to properly access files with Unicode names. For more information on -cpinternal, see Startup Command and Parameter Reference.

The ABL constructs listed in the following table do not support the full Unicode character set. Some constructs have strict limits, as noted in the table, but the rest are limited according to the operating system. In Windows, they are restricted to characters from the code page set in the Language for non-Unicode Programs (which is found under Region Settings in the Control Panel for Windows). A total of 256 characters can be used in Windows filenames, including non-ASCII characters, but the code page set in -cpinternal must match the code page selected in the Language for non-Unicode Programs. If your filenames contain Unicode characters outside of that code page, the AVM will not be able to access the file. In UNIX/Linux, the set of characters should be restricted to those available in the code page set in -cpinternal.

For both Windows and UNIX/Linux, -cpinternal must always be the same if a filename used with one of the constructs in the following table contains characters outside of ASCII. Restricting all filenames used with these constructs to ASCII characters only (unless a tighter restriction is noted) is the recommended practice.