I am using a GetDictionary.jsp script to parse an XML document containing simple Japanese phrases encoded with Shift_JIS. What I observe is that after reading a phrase into a java.util.HashMap bean the phrase becomes longer and modified (really, corrupted). This (corrupted) phrase sent in an HTTP reply, is modified again inside the packet.
Example: 1. the original phrase (in binary), from the XML doc: be dd c0 b8 bc c3 b8 c0 de bb b2 2e 2e 2e 2. the phrase (1) logged from a HashMap object: efbf bd efbf bd 38 efbf bd c3 b8 efbf bd de bb efbf bd 2e 2e 2e 3. the phrase (2) in an HTTP reply: 21 29 21 29 38 21 29 21 29 21 29 21 29 21 29 2e 2e 2e It appears as if two transformations were going on, beyond the GetDictionary.jsp control. Interestingly the last three characters 2e 2e 2e (='...') are preserved all the way. Another script Send.jsp that sends a reply has the following page declarations: <%@ page pageEncoding="Shift_JIS"%> <%@ page contentType="text/xml; charset=Shift_JIS"%> The situation is the same bad if only one declaration is in the Send.jsp: <%@ page contentType="text/xml; charset=Shift_JIS"%> Does anyone have a suggestion what am I doing wrong? Especially - while is there a 'transformation' from the XML doc to the HashMap? And then, why is there a second 'transformation' into a packet? Thanks, Jerry. [By the way, the same GetDictionary.jsp works fine if the parsed/read XML document is utf-8 encoded, or ISO-8859-1 encoded, or ascii. The phrase in the HashMap object is binary identical with the original phrase from the XML doc. Then it is the same in the reply packet (sent with a script having declared a corresponding charset).] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]