Executing exe using getRuntime.exec() through servlet

[Sanjaya Sugiarto]

Hello everyone,

I have a problem. I need to execute an exe. I have tried with a simple main method and it works like a charm. But as I try to integrate into Tomcat, it hangs. The program uses pdftohtml to convert a pdf into html.

Here is the code:

    
import java.io.BufferedReader;
 
import java.io.File;
 
import java.io.FileNotFoundException;
 
import java.io.IOException;
 
import java.io.InputStream;
 
import java.io.InputStreamReader;
 
import org.apache.log4j.Logger;
 
 
 
/**
 
 * Adding the ability to support PDF and converting it to HTML
 
 *
 
 * @author sugiarto
 
 */
 
public class PdfSupport extends ADocumentSupport
 
{
 
    static Logger m_logger = Logger.getLogger(PdfSupport.class.getName());
 
    
 
    public PdfSupport()
 
    {
 
    }
 
    
 
    public PdfSupport(File p_pathToSourcePdf, File p_pathToCreatedHtml, String p_LibPath)
 
    {
 
        convertToHtml(p_pathToSourcePdf, p_pathToCreatedHtml, p_LibPath);
 
    }
 
    
 
    public File convertToHtml(File p_pathToSourcePdf, File p_pathToCreatedHtml, String p_LibPath)
 
    {
 
        String l_wherePdfToHtmlInstalled = p_LibPath;
 
        convertPdfToHtml(l_wherePdfToHtmlInstalled, p_pathToSourcePdf, p_pathToCreatedHtml);
 
        return p_pathToCreatedHtml;
 
    }
 
    
 
    /**
 
     * Converts the PDF to HTML
 
     * @param p_pathWherePdfToHtmlInstalled Defines the path where the tool pdftohtml installed
 
     * ( eg. C:/pdftohtml/ )
 
     * @param p_pathToSourcePdf Defines the path where the will-be-converted source PDF saved
 
     * ( eg. C:/data/will_be_converted_file.pdf )
 
     * @param p_pathToCreatedHtml Defines the path where the resulting HTML file of the
 
     * converted PDF file will be saved ( eg. C:/temp/convertedFromPDF.htm )
 
     */
 
    public void convertPdfToHtml(String p_pathWherePdfToHtmlInstalled, File p_pathToSourcePdf, File p_pathToCreatedHtml)
 
    {
 
        //FileTypeDetection ftd = new FileTypeDetection();
 
        //if (ftd.detectByMimeType(p_pathToSourcePdf.getAbsolutePath()) == FileTypeDetection.TYPE_PDF) {
 
        try
 
        {
 
            if (System.getProperty("os.name").toUpperCase().indexOf("WINDOWS") == 0)
 
            {
 
                Process process = Runtime.getRuntime().exec(p_pathWherePdfToHtmlInstalled + "pdftohtml.exe -noframes -i -c " + p_pathToSourcePdf + " " + p_pathToCreatedHtml);
 
                process.waitFor();
 
                process.destroy();
 
            }
 
            else
 
            {
 
                Process process = Runtime.getRuntime().exec(p_pathWherePdfToHtmlInstalled + "pdftohtml -noframes -i -c " + p_pathToSourcePdf + " " + p_pathToCreatedHtml);
 
                process.waitFor();
 
                process.destroy();
 
            }
 
            // tidy up the converted HTML
 
            //tidyHtml(p_pathToCreatedHtml, p_pathToCreatedHtml.concat("_cleaned.html"));
 
            if (m_logger.isInfoEnabled())
 
            {
 
                m_logger.info("The HTML has been tidied up");
 
            }
 
        }
 
        catch (FileNotFoundException fnf)
 
        {
 
            m_logger.error("Failure during tidying up the html \n", (Throwable) fnf);
 
        }
 
        catch (IOException ioe)
 
        {
 
            m_logger.error("Failure during executing the executable pdftohtml \n", (Throwable) ioe);
 
        }
 
        catch (InterruptedException ie)
 
        {
 
            m_logger.error("PDFToHTML is not successfully executed \n", (Throwable) ie);
 
        }
 
    }
 
 
 
    public boolean checkFileProperty(String p_pathWherePdfToHtmlExecuteableIntalled)
 
    {
 
        boolean isFileOk = false;
 
        File file = new File(p_pathWherePdfToHtmlExecuteableIntalled);
 
        if(!file.isFile() || !file.exists())
 
        {
 
            if(!file.canRead())
 
            {
 
                m_logger.warn("The file is protected");
 
                isFileOk = false;
 
            }
 
            m_logger.warn("File not found");
 
            isFileOk = false;
 
        }
 
        else
 
        {
 
            isFileOk = true;
 
        }
 
        return isFileOk;
 
    }
 
}

It hangs just before the waitFor().
Have you guys any idea how to solve this?
Many thanks.

PS: I tried this with tomcat bundled with Netbeans 5.

[William Brogden]

Sounds like a very familiar problem - when you exec an external program, the Process takes over the standard input, output and error streams. If you do not provide for consuming these streams, the program probably hangs indefinately because it is trying to write to the output and/or error stream.

This is all covered in the java.lang.Process JavaDocs discussion.

Bill

[Sanjaya Sugiarto]

Sorry for my ignorant, but how can I provide to consume those streams? In Javadoc there is no detail how to do that.

Thanks.

[Sanjaya Sugiarto]

Hmmm....you meant:

process.getErrorStream();
process.getInputStream();
process.getOutputStream();

Is adding those 3 lines before the waitFor should be enough? Or you meant something different?

Thanks

[Paul Clapham]

Not enough. Read this.

[William Brogden]

Is adding those 3 lines before the waitFor should be enough? Or you meant something different?

Just obtaining the output streams is not enough, you have to provide for reading from them - just like it says in the JavaDocs for Process.

Note that if your exe is producing an error, it will likely be reported as messages to either stderr or stdout. You will need a separate Thread to read each stream and write the result somewhere so you can check it later.

You would only need to write to the stdin stream if your program requires keyboard input.

Bill