IDE developing JSP And Hebrew Encoding

Question

Hi.
I�m using Sun One Studio 4 CE.
I�m new Java Programmer.
I create JSP that use Taglib and read XML file written in Hebrew (Encoding iso-8859-1);
When I run the file on a Tomcat I have no problems.
But when I run the file within the ide of Sun One Studio I get gibberish.
So. Do you know where I cane change the character settings so It work the same?
Thanks
Shimon

Answer 1

The default encodings specified by the JSP and Servlet spec are unfortunately no good for most of the world!
There are two issues you need to understand: page encoding, vs response encoding.
Page encoding defines what encoding was used to create a particular page (JSP, XML, HTML). Unless the page encoding is set correctly, the entity that reads the page (e.g. the JSP parser, the web container, or the web server) cannot understand the characters before they do anything with them.
Response encoding defines what encoding is used when the server sends the HTTP response to the browser. This is actually a separate issue from page encoding. For example, a JSP could have the page encoding set correctly (so that the JSP parser can read it), but if the response encoding is not set correctly, then what goes to the browser is still gibberish.
To set the page encoding of a JSP, use a page directive:
<%@page pageEncoding="UTF-8"%>
(UTF-8 will work in Studio, there is no need to pick another more specialized charset).
In a HTML file or XML file, use an XML directive to set the page encoding. The XML directive will not work in JSP though.
So now the parser can translate the JSP into a servlet. But that's not good enough, because we also have to produce a response that the browser can understand.
To set the response encoding from a servlet, you can use
response.setContentType("text/html;UTF-8")
In J2EE 1.4, you can also use response.setCharacterEncoding("UTF-8");
To set it from a JSP, use
<%@page contentType="text/html;UTF-8"%>
All modern web browsers understand UTF-8, so that is a safe encoding to pick for the response. (It is a good encoding at the page level too if you use Studio. If you use another tool for your JSPs and it prefers a different encoding, use that at the page level but stick to UTF-8 for the response unless you have wireless clients and know that you need something else).
Note that this can only be set before any of the response is written. So if this JSP is included into another one, and the response buffer size has been exceeded, it won't work. So the response encoding must be set as early as possible.
What if you have a servlet that forwards to a JSP? Well, if you consider how the response object works, you should only have to set it once and earlier is better, so the servlet controller would be the obvious place. However, in J2EE 1.3, the encoding model wasn't well spec'd out, so on some servers, even if you set the response in the servlet, the JSP will overwrite it with the default if you didn't specify the content type. So experiment and set it in both the servlet and the JSP if you have to.
From J2EE 1.4, it's OK to set the response encoding once, it the servlet - the JSP container doesn't overwrite it anymore.
Check out the following link for more on i18n in web apps:
http://java.sun.com/developer/technicalArticles/Intl/MultilingualJSP/