Scrambled images when downloading across HttpURLConnection

[blanoel]

When attempting a .gif download from an URL address with an http protocol, the file is the right size, however the image is not loadable when viewing in a windows appliaction. Below is a bit of the sample code being used:

HttpURLConnection imgConnection = (HttpURLConnection fileName.openConnection(); // filename being the http URL
imgConnection.addRequestProperty("User-Agent", "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; .NET CLR 1.1.4322)");
InputStreamReader fileIn = new InputStreamReader(imgConnection.getInputStream());
FileWriter fileOut = new FileWriter(newName); // newName being client side filename
int c = 0;
while ((c = fileIn.read()) != -1) {
fileOut.write(c);
}
fileOut.flush();
fileOut.close();
fileIn.close();

Is there a better way? Maybe some kind of URL Encoding issue? Please help.

[ray326]

It looks to me like you're not parsing the GIF image data out of the response. Open the file up in a text editor and I think you'll see the HTTP headers as the first few lines.

[Khalid Ali]

you will need to parse the image from the http stream.
1. make sure that you set the
enctype="multipart/form-data"
then parse it http stream and strip the image from it.
Then I think you can create an image and save that image to the file...

[blanoel]

Thanks all.

Actually I took a look at the binary image in text format and I did not see any html headers in the file.

I also added a the line:

imgConnection.addRequestProperty("Enctype", "multipart/form-data");

Is this proper way to add the request property?

It stll seems to be scrambled. Any ideas?

[ray326]

I don't know about that encoding for a request but I'm a bit leery of it. Are the first 3 bytes of the downloaded image "GIF"? If not then does "GIF" appear anywhere in the image?

[blanoel]

After changing instances of InputStreamReader to BufferedInputStream, as well as changing FileWriter to BufferedOutputStream, all is working properly.

It turns out that Reader and Writer classes cannot handle binary data very well.

Thanks to all those that helped with this problem.

[ray326]

Ahh. The Readers and Writers may do end-of-line conversions ala FTP in text mode.