Hi,
Applying compression filter to response with non-default character encoding will mess
the output.
The reason is CompressionServletResponseWrapper.getWriter() returns writer with
system-default encoding.
In my case response is UTF-8 and system-default encoding is ISO-something (Latin-1). I
get bunch of '?' instead of Cyrillic characters.
Attached patch fixes the problem.
Regards, Anton
Index: CompressionServletResponseWrapper.java
===================================================================
RCS file:
/home/cvspublic/jakarta-tomcat-4.0/webapps/examples/WEB-INF/classes/compressionFilters/CompressionServletResponseWrapper.java,v
retrieving revision 1.6
diff -u -r1.6 CompressionServletResponseWrapper.java
--- CompressionServletResponseWrapper.java 18 Jun 2002 19:53:26 -0000 1.6
+++ CompressionServletResponseWrapper.java 9 Jul 2002 08:55:53 -0000
@@ -259,7 +259,18 @@
if (debug > 1) {
System.out.println("stream is set to "+stream+" in getWriter");
}
- writer = new PrintWriter(stream);
+ final String charEnc = origResponse.getCharacterEncoding();
+ if (debug > 1) {
+ System.out.println("character encoding is " + charEnc);
+ }
+ // HttpServletResponse.getCharacterEncoding() shouldn't return null
+ // according the spec, so feel free to remove that "if"
+ if (charEnc != null) {
+ writer = new PrintWriter(new OutputStreamWriter(stream, charEnc));
+ } else {
+ writer = new PrintWriter(stream);
+ }
+
return (writer);
}
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
