W3C home > Mailing lists > Public > html-tidy@w3.org > April to June 2000

Re: Java Tidy Backlog

From: <smarks@digisolutions.com>
Date: 19 Apr 2000 09:00:46 -0400
To: "Andy Quick" <ac.quick@sympatico.ca>, html-tidy@w3.org
Message-ID: <m31z42s0sh.fsf@hurin.innoverity.com>

I'd be willing to try and fix some bugs. However, right now I might be
adding one. 

Here goes:

When I parse this HTML 

<HTML>
<HEAD>
  <META NAME="GENERATOR" CONTENT="Adobe PageMill 3.0 Win">
  <TITLE>Spencer Museum of Art Printroom Home Page</TITLE>

<STYLE type="text/css">
  <!--  EM { font-family: garamond, serif; }  A { text-decoration: none; }   -->
</STYLE>

</HEAD>

<BODY BGCOLOR="#efefef" VLINK="#004040" ALINK="#0000a0" BACKGROUND="laid.gif">

You should see this.

</BODY>
</HTML>

and walk the DOM it get this output:

Element node: null
Element node: null
Element node: null
Element node: null
Element node: null
Text node: Spencer Museum of Art Printroom Home Page
Element node: null
Text node:   <!--  EM { font-family: garamond, serif; }  A { text-decoration: none; }   -->
Element node: null
Text node: You should see this.
Page parsed perfectly

If I remove the <STYLE type="text/css"> </STYLE> tags, but keep the
comment, so my input looks like this:

<HTML>
<HEAD>

  <META NAME="GENERATOR" CONTENT="Adobe PageMill 3.0 Win">
  <TITLE>Spencer Museum of Art Printroom Home Page</TITLE>


  <!--  EM { font-family: garamond, serif; }  A { text-decoration: none; }   -->


</HEAD>

<BODY BGCOLOR="#efefef" VLINK="#004040" ALINK="#0000a0" BACKGROUND="laid.gif">

You should see this.


</BODY>
</HTML>

and walk the DOM, I get this:

Element node: null
Element node: null
Element node: null
Element node: null
Element node: null
Text node: Spencer Museum of Art Printroom Home Page
Element node: null
Text node: You should see this.
Page parsed perfectly

Note: the big difference is the comment element is treated properly in
the second case. That is 
 <!--  EM { font-family: garamond, serif; }  A { text-decoration:
none; }   -->
does show up as  text node. 

So my question, perhaps bug, is why does the presence of the the Style
induce what appears to be incorrect behavior. That is why does the
presence of the STYLE tag cause the comment tag to be treated as a
text node and not an element node. 

Thanks in advance, and again, I am happy to try my hand at resolving some other
issues with JTiddy. Is there a bug database? If not, for starter, I could put
Bugzilla or Jitterbug (or something else) on my server. 

Spencer


Oh, here is the java that generated the above output:

/**
 * @author Spencer Marks <smarks@digisolutions.com>
 * @version $Id: HTMLTidyTagRemover.java,v 1.5 2000/04/18 22:19:49 sam Exp $
 */

package com.innoverity.util.html;

import java.io.InputStream;
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.io.OutputStream;
import java.io.IOException;
import java.io.PrintWriter;

import java.net.URL;
import java.net.URLConnection;

import com.innoverity.data.DynamicDictionary;
import com.innoverity.util.logging.MsgLog;

import org.w3c.dom.Document;
import org.w3c.dom.Node;
import org.w3c.dom.NodeList;
import org.w3c.tidy.Tidy;

import com.innoverity.data.ProcessedDocument;

/**
 * This implementation of the HTMLTagRemove interface
 * uses the HTML Tidy program to create a parse tree,
 * walk it and extract just the text from the element node.
 */
public class HTMLTidyTagRemover implements HTMLTagRemover  {

  HTMLTidyTagRemover() {

    /** pacakge level constructor so that only factory
     * can crate us.
     */
  }

  /**
   * Implement the HTMLTagRemoverInterface.
   *
   * @param in is a 'InputStream'
   * the HTML document to remove tags from.
   *
   * @param out is a 'PrintWriter - the new document that will contain
   * just the text of the supplied HTML document.
   *
   * @param errors is a 'OutputStream'
   * If the HTML document has parse errors they are recorded here.
   *
   * @return a 'HTMLPageInfo'
   * Info about the HTML document.
   */
  public HTMLPageInfo removeTags(InputStream in, PrintWriter out, PrintWriter errors) throws IOException  {

    HTMLPageInfo notReallyImplementedYet = new HTMLPageInfo();

    Tidy tidy = new Tidy();
    tidy.setMakeClean(true);
    tidy.setErrout(errors);

    // parse document
    Document doc = tidy.parseDOM(in, null);
    int errorCount =  tidy.getParseErrors();
    int warningCount = tidy.getParseWarnings();
    // udpate HTMLPageInfo
    notReallyImplementedYet.setNumberOfWarnings(warningCount);
    if (warningCount != 0 || errorCount != 0 ) {
      notReallyImplementedYet.setPageWasMalformed(true);
    } else {
      notReallyImplementedYet.setPageWasMalformed(false);
    }
    notReallyImplementedYet.setNumberOfErrors(errorCount);

    // generate output
    print(doc, out);

    errors.flush();
    out.flush();

    return notReallyImplementedYet;
  }

  /**
   * Prints the specified node, recursively.
   *
   * @param node a value of type 'Node'
   */
  private void print(Node node, PrintWriter out) {
    if ( node == null ) {
      return;
    }

    int type = node.getNodeType();
    switch ( type ) {

    case Node.DOCUMENT_NODE:
      print(((Document)node).getDocumentElement(),out);
      break;

    case Node.ELEMENT_NODE:
      String nodeValue = node.getNodeValue();
      System.out.println("Element node: " + nodeValue);
      NodeList children = node.getChildNodes();
      if ( children != null ) {
        int len = children.getLength();
        for ( int i = 0; i < len; i++ ) {
          print(children.item(i),out);
        }
      }
      break;

    case Node.TEXT_NODE:
       nodeValue = node.getNodeValue();
        System.out.println("Text node: " + nodeValue);
      out.write(nodeValue);
      out.write(" ");
      break;

    }

  }

  /**
   * Unit Test
   *
   * @param args[] is a 'String'
   */
  public static void main(String args[]) {

    if ( args.length != 3 ) {
      System.out.println("Usage: HTMLTidyProcessor <input file> <output file> <error log file>");
      System.out.println("You typed: " + args.length + "arguments");

      System.exit(1);
    }

    FileInputStream in = null;
    FileOutputStream out = null;
    FileOutputStream errors = null;
    HTMLTidyTagRemover  app = new HTMLTidyTagRemover();
    HTMLPageInfo pageInfo = null;
    try {

      in = new FileInputStream(args[0]);
      out = new FileOutputStream(args[1]);
      errors = new FileOutputStream(args[2]);
      PrintWriter textFile = new PrintWriter(out);
      PrintWriter errorOut = new PrintWriter(errors);
      pageInfo =  app.removeTags(in, textFile, errorOut);

      textFile.flush();
      errorOut.flush();

      textFile.close();
      errorOut.close();

      if (pageInfo.pageWasMalformed() == true) {
        System.out.println("The parser generated " + pageInfo.getNumberOfWarnings() + " warnings while removing tags");
        System.out.println("The parser generated " + pageInfo.getNumberOfErrors() + " errors while removing tags");
      } else {
        System.out.println("Page parsed perfectly");
      }

    }
    catch ( IOException e ) {
      System.err.println( e.toString() );
    }

  }

}

/**
 * @author Spencer Marks <smarks@digisolutions.com>
 * @version $Id: HTMLPageInfo.java,v 1.6 2000/04/18 22:19:49 sam Exp $
 */
package com.innoverity.util.html;

import java.net.URL;

/**
 * This class is used to hold statistics about
 * an HTML page.
 *
 * NOTE this class is more of a place holder than
 * anything else right now; It is very much a work in progress.
 * The idea being that at some point the search engine might
 * want to provide functionality along the lines of letting
 * the user search for pages with lots of images, etc.
 *
 * This data structure can be filled in when the html 2 text conversion
 * takes place.
 *
 * This object is included as member data in FeatureVector.
 * @see com.innoverity.clustering.FeatureVector
 *
 */

public class HTMLPageInfo {

  // true if the page required tidying or
  // could no tbe processed.
  private boolean wasMalFormed;

  private int numberOfLinks;
  private int numberOfImages;
  private int numberOfGIFs;
  private int numberOfMPEGs;
  private int numberOfJPEGs;
  private int numberOfWarnings;
  private int numberOfErrors;

  /**
   * Default constructor
   *
   */
  public HTMLPageInfo() {

    this.wasMalFormed = false;
    this.numberOfLinks = 0;
    this.numberOfImages = 0;
    this.numberOfGIFs = 0;
    this.numberOfMPEGs = 0;
    this.numberOfJPEGs = 0;
    this.numberOfWarnings = 0;
    this.numberOfErrors = 0;
  }



  /**
   * Returns true if the tag removing processing
   * detected an error parsing the page.
   *
   * @return a value of type 'boolean'
   */
  public boolean pageWasMalformed() {
    return wasMalFormed;
  }
  /**
   * Return the number of warnings the HTML parser
   * generated while processing the document. 
   * Warnings should be considered non fatal. 
   * That is the output should be OK. 
   * The parser is just reporting something it didn't like,
   * but it is something it can deal with. 
   * 
   * @return a value of type 'int'
   */
  public int getNumberOfWarnings() {
    return numberOfWarnings;
  }


  /**
   * Return the number of errors the HTML parser
   * generated while processing the document. 
   * Errros should be considered fatal. 
   * That is the output <b><i>might not</b></i> be OK. 
   * The parser is just reporting something it couln't really handle.
   *
   * @return a value of type 'boolean'
   */
  public int getNumberOfErrors() {
    return numberOfErrors;
  }

  /**
   * Returns the number of http:// hrefs in this document.
   *
   * @return a value of type 'int'
   */
  public int getNumberOfLinks() {
    return numberOfLinks;
  }

  /**
   * Returns the number of <IMG> tags in the document.
   *
   * @return a value of type 'int'
   */
  public int getNumberOfImages() {
    return numberOfImages;
  }

  /**
   * Returns the number of JPEGs in the document.
   *
   * @return a value of type 'int'
   */
  public int getNumberOfJPEGs() {
    return numberOfJPEGs;
  }

  /**
   * Returns the number of GIFs in the document.
   *
   * @return a value of type 'int'
   */
  public int getNumberOfGIFs() {
    return numberOfGIFs;
  }

  /**
   * Returns the number of MPEGs in the document.
   *
   * @return a value of type 'int'
   */
  public int getNumberOfMPEGs() {
    return numberOfMPEGs;
  }

  /**
   * Set whether or not the HTML page this object
   * contains info about was/is malformed.
   *
   * @return a value of type 'boolean'
   */
  public void setPageWasMalformed(boolean setting) {
    this.wasMalFormed = setting;
  }
  
  /**
   *
   * Set the number of errors the HTML parser
   * encountered. 
   * 
   * @return a value of type 'int'
   */
  public void setNumberOfErrors(int errors) {
    this.numberOfErrors = errors;
  }

  /**
   *
   * Set the number of warnings the HTML parser
   * encountered. 
   * 
   * @return a value of type 'int'
   */
  public void setNumberOfWarnings(int errors) {
    this.numberOfWarnings = errors;
  }
  /**
   * Sets the number of http:// hrefs in this document.
   *
   * @return a value of type 'int'
   */
  public void setNumberOfLinks(int num) {
    this.numberOfLinks = num;
  }

  /**
   * Sets the number of <IMG> tags in the document.
   *
   * @return a value of type 'int'
   */
  public void setNumberOfImages(int num) {
    this.numberOfImages = num;
  }

  /**
   * Sets the number of MPEGs in the document.
   *
   * @return a value of type 'int'
   */
  public void setNumberOfMPEGS(int num) {
    this.numberOfMPEGs = num;
  }

}










































"Andy Quick" <ac.quick@sympatico.ca> writes:

> Dave,
> 
> The backlog of Java tidy issues is growing a bit faster than I can
> handle them.  Generally speaking, Java tidy is quite stable and
> matches C tidy on a very high percentage of HTML documents.
> However, there are more bug-finders than bug-fixers right now :-)
> 
> I was wondering if I could ask people to post them to this mailing list
> so that I could increase the potential number of bug-fixers.  I would
> be willing to approve and integrate bug fixes into the source code.
> Maybe people could clearly indicate in the subject whether the issue
> applies to Tidy, Java Tidy, or both.  Another idea would be to set up
> another mailing list, but I don't think it's worth that at this point.
> 
> Regards,
> 
> Andy Quick
Received on Wednesday, 19 April 2000 09:01:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 April 2012 06:13:43 GMT