Configuring a Unicode server

The AR System Administration Console enables you to specify the minimum API that the server supports. AR System 7.x or later clients use API version 12, so enter the number 12 to prevent access to older clients.

Running a client from the command line on Windows

Use the --unicode flag when you invoke any of the following programs:

• arcache
• archgid
• archgsel
• ardisabled
• arhelp
• arl10nmenu
• arlabel
• artext
• arworkflow
• fbupdate

For example, see arlabel.txt in the Unsupported folder where AR System is installed. For more information about using ARLABEL, go to BMC Support.

Running a Unicode client from the command line on UNIX

To run a Unicode client from the command line, you must perform the following actions:

• Set the locale by using either of the following methods:
  • Set the LC_ALL and LANG environment variables to a value produced by the locale -a command.
  • Set the locale in your .profile file (sh/bash users) or .cshrc file (csh/tcsh users).
    The client uses the values you set in the Shell script. Of the locales listed, AR System supports those ending in UTF-8 (for Solaris and AIX) and utf8 (for HPUX and Linux).
    See your UNIX documentation for information about how to set your locale settings.
• Make sure that the dynamic libraries for AR System 7. x are available by adding the ARSystemServerInstallDir/server/bin directory to your dynamic library path with the appropriate environment variable:
  • LIBPATH on AIX
  • SHLIB_PATH on HP-UX
  • LD_LIBRARY_PATH on Solaris and Linux
    For example, on Solaris, enter:
    $ LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/bmc/ARSystem/server/bin; export LD_LIBRARY_PATH

To simplify the process, if you are running on the UNIX system where the AR System server is installed, you can use the arsystem env command:
$ <ARSystemInstallationDirectory>/server/bin/arsystem env <commandName> <arguments>

This automatically sets the dynamic library path and locale variables to the same values that the AR System server uses.

Filter and escalation workflow considerations for Unicode

Some server-side workflow actions call byte-oriented filter functions, such as LENGTH, SUBSTR, and STRSTR. There are also Character-oriented equivalents, such as LENGTHC, SUBSTRC, and STRSTRC.

Because UTF-8, Chinese, Japanese, and Korean characters can occupy one or more bytes, the byte-oriented, and character-oriented lengths and the offsets for these orientations are different. Make sure you understand the difference when creating workflow actions. See the following examples:

• LEFT($8$, 3) extracts a prefix of no more than 3 bytes from field 8.
• LEFTC($8$, 3) extracts a prefix of no more than 3 characters from field 8.