UTF8 Support

Question

My platform: Ora 8.1.7 UTF8 Encoded, Portal 3.0, Win 2000 Prof. Enviromental Variable NLS_LANG=AMERICAN_AMERICA.UTF8
owa_util.get_cgi_env('REQUEST_CHARSET') returns UTF8
If I enter for example text "omega" in greek alphabet into the database using portlet with standard html form, then all portlets accesing this data return error message: "Error: The portlet could not be contacted.". The same happens if I log out and log back and querry the greek record.
What is the problem?
I have written my own servlet connecting to my UTF8 database. In my servlet I have had to specialy handle the java streams. It has worked.
In release notes you write: "The character set setting that you use on your middle tier should be identical to the character set setting of your database."
How should I set this character set on the midle tier? I have 9iAS and 8i on the same box.
Do you have some recomedations?

Answer 1

Hi,
We have a portlet that gives similar behavior - anytime you enter an 'accuting vowel', the portlet breaks. But entering the same thing in the PL/SQL application that the Portlet is based on, doesn't break the application.
There are bugs in WebIV about the XML parser that causes similar problems in
the Apps (our portlet uses the XML parser) but those problems are supposed to be resolved since XML parser 2.0.2.8 (and 2.0.2.9 is the version that comes with 8.1.7 and 9ias also). I am not sure what to think, or who we should talk to about this - i.e., whether it's a Portal problem w/rendering the portlets that contain the 'strange' characters, or whether it's a database thing....