On Sat, 16 Mar 2002, Mete Kural wrote:
> Date: Sat, 16 Mar 2002 11:05:08 -0800 (PST)
> From: Mete Kural <[EMAIL PROTECTED]>
> Reply-To: Tomcat Users List <[EMAIL PROTECTED]>
> To: [EMAIL PROTECTED]
> Subject: Tomcat and Unicode parameters in URLs ???
>
> Hello all,
>
> I tried a million ways of making Tomcat 4.0.3 work
> with Unicode URL parameters, but nothing seems to
> work. It always corrupts the parameters. Does anybody
> know a workaround to make Unicode request parameters
> work with Tomcat?
>
> For instance, I changed the SnoopServlet example given
> with Tomcat 4 to output the response in the unicode
> with setContentType("text/html;charset=utf-8"). But
> when I write unicode parameters in the URL text area
> of Internet Explorer as parameters to SnoopServlet, it
> always corrupts my parameters. Instead of printing
> them a two-bye unicode characters, it prints every
> unicode character as two one-byte garbage or otherwise
> ASCII characters. I also tried making a URL request
> using the URL class in JAVA SDK 1.4. That didn't work
> as well. The URLEncode and URLDecode classes in the
> SDK don't seem to do their job right also. Has anyone
> been able to make use of these classes? Any
> workarounds and bug reports will be greatly
> appreciated.
>
Setting the content type, as you did above, only affects the *output* of
that particular response -- it has nothing to do with how the next *input*
request from that browser will be handled.
In order to deal with request parameters in an incoming request, you must
tell Tomcat what encoding to use, *before* processing the parameters.
This is done by calling the request.setCharacterEncoding() method that was
added in Servlet 2.3. As long as you call this before calling methods
like request.getParameter(), the proper encoding will be applied.
One way to do this without modifying your application itself is to use a
Filter that looks at incoming requests and decides what encoding should be
used -- perhaps by looking at the <code>Accept-Language</code> header, or
based on attributes you have stored in the current session that indicate
what the user will be supplying. A very simple example of such a filter
is included in the "/examples" webapp shipped with Tomcat 4 -- the source
code for this filter is in file SetCharacterEncodingFilter.java in the
$CATALINA_HOME/webapps/examples/WEB-INF/classes/filters subdirectory.
This example is fairly simpleminded -- you just configure a filter
initialization parameter that is used to set the encoding for all requests
-- but you can use it as a starting point for more sophisticated
processing by subclassing it and overriding the selectEncoding() method.
This filter can be enabled by copying the appropriate class file to your
own WEB-INF/classes directory, and adding a filter definition to your
web.xml file:
<filter>
<filter-name>Character Encoding Filter</filter-name>
<filter-class>filters.SetCharacterEncodingFilter</filter-class>
<init-param>
<param-name>encoding</param-name>
<param-value>UTF-8</param-value>
</init-param>
</filter>
Then, you select which requests this filter applies to with a filter
mapping -- the "/*" pattern says apply it to *all* requests:
<filter-mapping>
<filter-name>Character Encoding Filter</filter-name>
<url-pattern>/*</url-pattern>
</filter-mapping>
With filter mappings like this, you can be more selective about which URLs
it applies to by using a more precise URL pattern, or apply different
filters to different URLs -- all without affecting your servlets or JSP
apges at all. The syntax for the <url-pattern> element in a filter
mapping is the same as that used for a servlet mapping.
Be sure to put these elements in the correct places in your web.xml file
to maintain the element order that is required by the DTD.
> Thanks,
> Mete
>
Craig
--
To unsubscribe: <mailto:[EMAIL PROTECTED]>
For additional commands: <mailto:[EMAIL PROTECTED]>
Troubles with the list: <mailto:[EMAIL PROTECTED]>
