> > That way the image data gets corrupted. Isn't this information supposed
> > to be in the HTTP Headers, and not in my output?
>
> This corruption is produced by an HTTP/1.1 transfer encoding called
> chunking. Its support is mandatory in HTTP/1.1. Which HTTP client are you
> using ?
> If you're directly talking on the socket or using a client without real
> support for HTTP/1.1, you may want to make a request using HTTP/1.0 (which
> doesn't have chunking), or set the content length of the response (in which
> case chunking will not be used).
I'm using Netscape 4.76 and Konqueror 2.1, and I tried Mozilla Seamonkey
on Windows. They all don't display the images correctly, and when I save
them to disk the files contain these "chunking" lines at start.
I also tried to set the content length, as you said, but that didn't
help: This is what's in the file before the jpeg data when I save it to
disk:
--- snip ---
HTTP/1.1 200 OK
Content-Type: image/jpeg
Content-Length: 108381
Date: Fri, 23 Mar 2001 17:44:30 GMT
Server: Apache Tomcat/4.0-b1 (HTTP/1.1 Connector)
--- snap ---
My code looks like this now (it worked with Resin, so I think it is all
right):
--- snip ---
// now, did we get the data? if yes, copy it
// to the response:
if ((blob != null) && (contentType != null)) {
response.setContentType (contentType);
response.setContentLength ((int) blob.length ());
ServletOutputStream out =
response.getOutputStream ();
byte b[];
long toRead = blob.length ();
long read = 0;
while (toRead > 0) {
b = blob.getBytes (read,
(toRead > 1024) ? 1024 : (int)toRead);
toRead -= b.length;
read += b.length;
out.write (b);
}
out.flush ();
out.close ();
--- snap ---
Please have a look at this problem.
with kind regards,
Stefan Schmidt
