Echo Web Framework

Is did trace the request sent from FF3 to the server.

It has the following header:

content-type: text/xml; charset=ISO-8859-1

The XML parser on the serverside then trys to interprete this as UTF-8
and of course fails.

I assume that the clientside part of echo2/3 does not specify what encoding the xml requests are sent, and we are then on mercy of the browser to use UTF-8 (or in the case of FF3 NOT).

No idea in which .js file this is handled...

Without having to patch the Echo source I implemented a servlet filter which reencodes requests which have the "wrong" encoding

The HeaderControlFilter:

public class HeaderControlFilter implements Filter {
protected String reencoding = null;
public void doFilter(ServletRequest request, ServletResponse response,
FilterChain chain) throws IOException, ServletException {
HttpServletRequest httpRequest = (HttpServletRequest)request;
HttpServletResponse httpResponse = (HttpServletResponse)response;
if(reencoding!=null && request.getCharacterEncoding()!=null && !reencoding.equals(request.getCharacterEncoding())) {
request = new HeaderControlRequest(httpRequest,reencoding);
}
chain.doFilter(request, response);
}
public void init(FilterConfig filterConfig) throws ServletException {
this.reencoding = filterConfig.getInitParameter("request.reencoding");
}
}


The HeaderControlRequest wrapper:

public class HeaderControlRequest extends HttpServletRequestWrapper {
String encoding;
ServletInputStream is;
public HeaderControlRequest(HttpServletRequest request, String encoding) {
super(request);
this.encoding = encoding;
}
public ServletInputStream getInputStream() throws IOException {
if(is==null) {
is = new InputStreamReencoder(super.getInputStream(), super.getCharacterEncoding(), encoding);
}
return is;
}
}


The InputStreamReencoder:

public class InputStreamReencoder extends ServletInputStream {
ByteArrayInputStream is;
public InputStreamReencoder(ServletInputStream in, String inEncoding, String outEncoding) {
ByteArrayOutputStream byteOut = new ByteArrayOutputStream();
byte[] buffer = new byte[4096];
int bytesRead = 0;
try {
do {
bytesRead = in.read(buffer);
if (bytesRead > 0) {
byteOut.write(buffer, 0, bytesRead);
}
} while (bytesRead > 0);
} catch (IOException e) {
e.printStackTrace();
} finally {
if (in != null) {
try {
in.close();
} catch (IOException ex) {
// Ignore
}
}
}
byte[] data = byteOut.toByteArray();
try {
data = new String(data, inEncoding).trim().getBytes(outEncoding);
is = new ByteArrayInputStream(data);
} catch (UnsupportedEncodingException e) {
// Ignore
}
}
public int read() throws IOException {
return is.read();
}
}


Add the following to web.xml:

<filter>
<filter-name>headercontrol</filter-name>
<filter-class>HeaderControlFilter</filter-class>
<init-param>
<param-name>request.reencoding</param-name>
<param-value>UTF-8</param-value>
</init-param>
</filter>
<filter-mapping>
<filter-name>headercontrol</filter-name>
<url-pattern>/echoservlet</url-pattern>
</filter-mapping>


Hi,

This is indeed still an issue in the official Firefox release. I have produced a workaround based on issues https://bugzilla.mozilla.org/show_bug.cgi?id=431701 and https://bugzilla.mozilla.org/show_bug.cgi?id=407213#c8.
This workaround will only be applied for Firefox 3.0 browsers. This issue will probably be fixed by Mozilla in one of the next updates. This workaround forces Firefox to deal with UTF-8 and saves resources on the server.

Technically speaking, Echo does not honor a browser's request for a specific encoding. This could be solved by implementing some kind of re-encoding as suggested earlier in this thread.

Niels

P.S. diff is based on SVN 1207