httpServletRequest.getParameter decoding problem

[ybenigot]

Using jboss 3.2.3 with tomcat on Linux/Intel

I send an HTTP GET with header Content-Type: text/xml ; charset=utf-8 to a servlet using apache HttpClient.

The parameters values are encoded with URLencoder.urlencode(value,"UTF-8").

-> In the servlet, request.getParameter(parameterName) returns a String
with malformed text : accented characters are replaced by pairs of special characters.

If I use URLencoder.urlencode(value,"ISO-8859-1") on the client side it
works just fine.

So I guess jboss is using the defaut 8859 encoding on the server side, instead of the HTTP header information.

Is there a way to fix this ?

[mbeyer73]

I am having the same problem with JSP pages using JBoss 3.2.5 on Win XP. The parameters are always decoded in ISO-8859-1. I even tried an Encoding-Filter with no success.

Demo page:

<%@ page pageEncoding="UTF-8" contentType="text/html;charset=UTF-8" %>
<%@ page import="java.net.URLEncoder" %>

<% request.setCharacterEncoding("UTF-8"); %>

<title>Test</title>

<h1>Umlaut-Test! öäüß</h1>

<form method="get" action="test.jsp">
<input type="hidden" name="p" value="Müller" />
<input type="submit" value="submit! (UTF-8, see contentType)" />
</form>

<p><a href="test.jsp?p=<%= URLEncoder.encode ("Müller", "UTF-8") %>">LINK encoded in UTF-8</a></p>
<p><a href="test.jsp?p=<%= URLEncoder.encode ("Müller", "iso-8859-1") %>">LINK encoded in iso-8859-1</a></p>

<h1>Result</h1>
<p>request.getCharacterEncoding ( ) &ndash; <%= request.getCharacterEncoding ( ) %></p>
<p>request.getParameter ("p") &ndash; <%= request.getParameter ("p") %></p>

[mbeyer73]

It seems that Tomcat ignores request.setCharacterEncoding. The request char encoding however can be set globally in server.xml, tag "Connector", attribute URIEncoding, e.g.

 <Connector port="8080" address="${jboss.bind.address}"
 maxThreads="150" minSpareThreads="25" maxSpareThreads="75"
 enableLookups="false" redirectPort="8443" acceptCount="100"
 connectionTimeout="20000" disableUploadTimeout="true"
 compression="on" compressionMinSize="2048" compressableMimeType="text/html,text/xml"
 URIEncoding="UTF-8"
 />

My goodness, I spent two days two find that out :-( Why don't do the Tomcat developers follow the spec? Or at least write out a warning or throw an exception if I call request.setCharacterEncoding ("pleaseIgnoreMe")?

angry,
Marcus