Streaming PDF files

[Dale DeMott]

I'm streaming a PDF files from my servlet to the browser. This seems to work great for my machine. Issue is that it doesn't seem to work for everyone. I dont' get an error. I just get a blank screen. It just stops loading and leaves the screen blank. Almost as though it was done but didn't quite complete. Not sure why this happens. Servlet code is below

public void dispatchPDF(String path, HttpServletResponse response)
    throws FileNotFoundException, IOException {
    File file = null;
    try {
        response.setContentType("application/pdf");
        file = new File(path);
        FileInputStream fileInputStream = new FileInputStream(file);
        OutputStream outputStream = response.getOutputStream();
        int currentByte = 0;
        int charsWritten = 0;
        while ((currentByte = fileInputStream.read()) != -1) {
            outputStream.write(currentByte);
            charsWritten++;
        }
        outputStream.flush();
    }
    catch (FileNotFoundException fnfe) {
        System.out.println(
            "FileNotFoundException thrown.  System was expecting to see the file at "
                + file.getAbsolutePath());
        throw fnfe;
    }
}

[Mathew Sam]

Check wheteher acrobat reader is installed in other systems.

Sam

[Dale DeMott]

Actually it is installed.

[Tim Holloway]

There's a fatal flaw in Netscape 4. I haven't checked to see if it's in Netscape 6 and/or Mozilla, and it may be in other browsers as well.
The problem is that when you're streaming, you presumably don't yet know the content-length, and without a valid (and accurate!) content-length, Netscape 4 comes up blank. The plugin's rejecting the data.
The usual work-arounds are to either 1) buffer the document until you can get its length, then output the content-length header followd by the PDF; or 2) inform the user to save the document as a file (Content-Disposition header helps) and open the file.
Bummer, nyet?

[Dale DeMott]

Actually I found the issue. It was that I had some users using Adobe Acrobat 4. Adobe Acrobat 5 works great though. Now the new issue is how can I speed up this dispatching. Right now it takes a good amount of time to dispatch a pdf using this code. Any hints on how to speed this up?

public void dispatchPDF(String path, HttpServletResponse response)
    throws FileNotFoundException, IOException {
    response.setContentType("application/pdf");
    File file = new File(path);
    FileInputStream fileInputStream = new FileInputStream(file);
    OutputStream outputStream = response.getOutputStream();
    int currentByte = 0;
    int charsWritten = 0;
    while ((currentByte = fileInputStream.read()) != -1) {
        outputStream.write(currentByte);
        charsWritten++;
    }
    outputStream.flush();
}

[Tim Holloway]

Interesting about Acrobat 5. Although that only helps if you have an audience you KNOW has upgraded!
You should see a significant improvement in performance if you read and write in larger chunks:

package com.listopt.servlet;
import javax.servlet.http.*;
import java.io.*;
public class PDF1Servlet extends HttpServlet
{
    final static int BUFLEN = 1024;
    public void doGet(HttpServletRequest request, HttpServletResponse response)
    {
byte[] buffer = new byte[BUFLEN];
        try {
        response.setContentType("application/pdf");
        response.setHeader("Pragma" , "no-cache");
        response.setHeader("Cache-Control" , "no-cache");
        response.setDateHeader("Expires" , 0);    
        response.setHeader("Content-Disposition" , "inline; filename=\"MailingLists.pdf\"");
        response.setHeader("Content-Description" , "Display database as PDF");
File tmpFile = File.createTempFile("WPD", "pdf");
java.io.FileOutputStream outf = new java.io.FileOutputStream( tmpFile );

java.io.OutputStream ostream = response.getOutputStream();
PrintSummary ps = new PrintSummary();
ps.process(outf);
outf.close();
response.setIntHeader("Content-length", (int)tmpFile.length() );
java.io.FileInputStream istream = new java.io.FileInputStream( tmpFile );
boolean keepGoing = true;
while ( keepGoing ) {
    int len = istream.read( buffer, 0, BUFLEN );
    if ( len > 0 ) {
ostream.write( buffer, 0, len );
    } else {
keepGoing = false;
    }
}
istream.close();
tmpFile.delete();
        } catch ( java.io.IOException ioex ) {
            response.setStatus(HttpServletResponse.SC_INTERNAL_SERVER_ERROR);
        }
    }
    public PDF1Servlet()
    {
    }
}

[Jim Yingst]

Also addressed here. Please don't duplicate questions in different forums - it just causes us to duplicate our answers.

[Dale DeMott]

Thanks Jim. You're right. Next time I'll only post once.

[Dale DeMott]

Tim Holloway
Can you explain a bit about this PrintSummary class. Is this going to be dispatching the temp file after its been created? Is this a class you created or did I miss this in the API?
[ August 09, 2002: Message edited by: Dale DeMott ]

[Dale DeMott]

Also, [Looking at Tim's code], do I need to worry about these temp files being created on the server. Will other people run into them if we get a lot of people on the server??
-Dale