Uploading Files Using Non-Latin Characters
When uploading an attachment file with a filename that uses non-Latin characters, such as Hebrew or Japanese, the system uses the exact filename or UTF-8 % sequenced numbers. In previous releases, the system renamed the file and converted to strings of numbers.
When uploading non-Latin characters, the following error can occur: "Attachment Filename is too long (maximum is 64 characters)."
To resolve this issue on UNIX servers:
Make sure that the corresponding language locale, such as the UTF-8 character set, is installed at the operating system level.
Use "locale -a" to get the list of installed locales.
Use the LANG environment variable to set for the session that runs the application server to a valid locale for the character set.
For example, en_US.utf8, iw_IL.utf8, ja_JP.utf8, and so on. See your "locale -a" output.
Use PSADMIN to set the Character Set parameter to the appropriate value, as in Character Set=utf8.
The character set values specified in the shell locale (LANG) and PSAPPSRV.CFG (Character Set) must agree, and, in this case, PSAPPSRV.CFG requires utf8.
Note: If the shell locale is set to something other than the default of C, Tuxedo will issue warnings when the application server boots. In such a case, add a symlink from $TUXDIR/prod/locales/<your_shell_locale> to $TUXDIR/prod/locales/C to resolve that issue.
On a Windows-based application server, make sure the server operating system has the corresponding language installed.
To resolve this issue on Windows servers:
Make sure that the corresponding language locale, such as CP932 (for Japanese) character set, is installed at the operating system level.
Use the Regional and Language Options utility in the Control Panel to get the list of installed locales, and use “chcp” in the command prompt window to see the current default code page.
Use Control Panel/Regional and Language Options to set the system locale for the session that runs the application server with a valid language.
For example, on the Advanced tab for the Language for non-Unicode programs option, select Japanese for the Japanese system locale of CP932.
Use PSADMIN to set the Character Set parameter so that it reflects the appropriate code set.
Using the Japanese example: the default code page CP932 (equivalent to SJIS) and the character set value in PSAPPSRV.CFG must agree. So in this case, PSAPPSRV.CFG requires “sjis” to be set as the character set, as in Character Set=sjis.