Default encoding switched to UTF-8

[julien1]

I have worked on that issue we had for a while. Basically when we change the response content type, at the same time the request content type must be changed to the same charset. Otherwise the textarea and textfield will not be decoded with the right content type.

I had to change the compress filter which was not using the declared charset to encode the chars as bytes and some other things.

[julien1]

BTW I tried the chinese examples given some time ago and it works well without using XML char entities. So for these guys which had issues with that, I advise you to get a snapshot from the CVS and it should be ok now.

[innovate]

Cool,

I will have a look into the code. Have you used the Filter from Struts?

I just installed nukes at home where I have MacOS X. No problems with the encoding.

At work, I have running Linux and there, I have problems with the encoding. Have you any idea what is different between those platforms???


Regards,

Cyrill

[julien1]

what kind ? usually the default encoding is different

[innovate]

Everythin OK under Linux :-)

[jae77]

this does not seem to work if i click the "view" button when viewing the details about a file stored in the database.

also, it seems that pulling images out of the database is now broken w/ these changes. from what i can tell, it is b/c the ByteArrayOutputStream in the "Blah" class is set to null.

if i initalize that variable when calling getOutputStream and i build the ServletOutputStreamWrapper w/ that, everything seems to work, but i'm not sure if what i'm doing is ultimately correct.

here's the snippet for getOutputStream from the Blah class.

 public ServletOutputStream getOutputStream() throws IOException
 {
 if (wrappedWriter != null)
 {
 throw new IllegalStateException();
 }
 if (wrappedStream == null)
 {
 OutputStream out = null;
 if (contentType == null || exclude.contains(contentType))
 {
 // That means we copy the stream directly to the output
 ServletResponse response = getResponse();
 response.setContentType(contentType);
 response.setBufferSize(Tools.DEFAULT_BUFFER_SIZE);
 out = response.getOutputStream();
 }
 else
 {
 // We bufferize for compression
 baos = new ByteArrayOutputStream();
 }

 if (out == null)
 {
 wrappedStream = new ServletOutputStreamWrapper(baos);
 }
 else
 {
 wrappedStream = new ServletOutputStreamWrapper(out);
 }
 }
 return wrappedStream;
 }

[cookman]

"cooper" wrote:
BTW I tried the chinese examples given some time ago and it works well without using XML char entities. So for these guys which had issues with that, I advise you to get a snapshot from the CVS and it should be ok now.

For chinese language, I just change the CoreModule's Charset attribute from "UTF-8" to "ISO-8859-1" via jmx-console, all things be ok.

[julien1]

yes I spent some time to understand how these things interracts.

[cookman]

"cookman" wrote:
For chinese language, I just change the CoreModule's Charset attribute from "UTF-8" to "ISO-8859-1" via jmx-console, all things be ok.

except for the mail received from the watched forums/topics, and chinese characters that nukes saved in database are all in unicode