[Htmlparser-cvs] htmlparser/src/org/htmlparser/lexer Lexer.java,1.2,1.3 Page.java,1.6,1.7

[<der...@us...>]

Update of /cvsroot/htmlparser/htmlparser/src/org/htmlparser/lexer
In directory sc8-pr-cvs1:/tmp/cvs-serv23027/src/org/htmlparser/lexer

Modified Files:
	Lexer.java Page.java 
Log Message:
Fifth drop for new i/o subsystem.
There is now a mainline for the lexer.
Try:
    java -jar lexer.jar http://whatever
or the integration build has a new lexer execution script:
    bin/lexer http://whatever




Index: Lexer.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/lexer/Lexer.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** Lexer.java	21 Aug 2003 01:52:23 -0000	1.2
--- Lexer.java	23 Aug 2003 01:33:06 -0000	1.3
***************
*** 26,33 ****
--- 26,39 ----
  // CA 94708, USA
  // Website : http://www.industriallogic.com
+ // 
+ // This class was contributed by 
+ // Derrick Oswald
+ //
  
  package org.htmlparser.lexer;
  
+ import java.io.IOException;
  import java.io.UnsupportedEncodingException;
+ import java.net.URL;
  import java.net.URLConnection;
  import java.util.Vector;
***************
*** 546,548 ****
--- 552,580 ----
      }
  
+     /**
+      * Mainline for command line operation
+      */
+     public static void main (String[] args) throws IOException, ParserException
+     {
+         URL url;
+         Lexer lexer;
+         Node node;
+ 
+         if (0 >= args.length)
+             System.out.println ("usage: java -jar htmllexer.jar <url>");
+         else
+         {
+             url = new URL (args[0]);
+             try
+             {
+                 lexer = new Lexer (url.openConnection ());
+                 while (null != (node = lexer.nextNode ()))
+                     System.out.println (node.toString ());
+             }
+             catch (ParserException pe)
+             {
+                 System.out.println (pe.getMessage ());
+             }
+         }
+     }
  }

Index: Page.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/lexer/Page.java,v
retrieving revision 1.6
retrieving revision 1.7
diff -C2 -d -r1.6 -r1.7
*** Page.java	21 Aug 2003 01:52:23 -0000	1.6
--- Page.java	23 Aug 2003 01:33:06 -0000	1.7
***************
*** 26,29 ****
--- 26,33 ----
  // CA 94708, USA
  // Website : http://www.industriallogic.com
+ // 
+ // This class was contributed by 
+ // Derrick Oswald
+ //
  
  package org.htmlparser.lexer;
***************
*** 39,45 ****
  import java.net.URLConnection;
  import java.net.UnknownHostException;
! 
! import org.apache.commons.logging.Log;
! import org.apache.commons.logging.LogFactory;
  
  import org.htmlparser.util.ParserException;
--- 43,47 ----
  import java.net.URLConnection;
  import java.net.UnknownHostException;
! import java.util.Random;
  
  import org.htmlparser.util.ParserException;
***************
*** 61,69 ****
  
      /**
-      * The logging object.
-      */
-     protected static Log mLog = null;
- 
-     /**
       * The source of characters.
       */
--- 63,66 ----
***************
*** 113,117 ****
          catch (UnknownHostException uhe)
          {
!             throw new ParserException ("the host (" + connection.getURL ().getHost () + ") was not found", uhe);
          }
          catch (IOException ioe)
--- 110,116 ----
          catch (UnknownHostException uhe)
          {
!             Random number = new Random ();
!             int message = number.nextInt (mFourOhFour.length);
!             throw new ParserException (mFourOhFour[message], uhe);
          }
          catch (IOException ioe)
***************
*** 348,352 ****
                      if (!ret.equalsIgnoreCase (content))
  					{
!                         getLog ().info (
                              "detected charset \""
                              + content
--- 347,351 ----
                      if (!ret.equalsIgnoreCase (content))
  					{
!                         System.out.println (
                              "detected charset \""
                              + content
***************
*** 408,417 ****
  			// return the default
  			ret = _default;
!             getLog ().debug (
                  "unable to determine cannonical charset name for "
                  + name
                  + " - using "
!                 + _default,
!                 ita);
  		}
          
--- 407,415 ----
  			// return the default
  			ret = _default;
!             System.out.println (
                  "unable to determine cannonical charset name for "
                  + name
                  + " - using "
!                 + _default);
  		}
          
***************
*** 506,523 ****
          getText (buffer, 0, mSource.mOffset);
      }
- 
-     //
-     // Bean patterns
-     //
- 
-     public Log getLog ()
-     {
-         if (null == mLog)
-             mLog = LogFactory.getLog (this.getClass ());
- //        String name = this.getClass ().getName ();
- //        java.util.logging.Logger logger = java.util.logging.Logger.getLogger (name);
- //        logger.setLevel (java.util.logging.Level.FINEST);
-         return (mLog);
-     }
- 
  }
--- 504,506 ----