Forum:

Java in General

Creating a Parser

Greenhorn

Posts: 13

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

Sorry, I couldn't find an answer to this. I've been trying to create my own parser to convert a .java file into .html format and inserting syntax highlighting. I thought I would give it a try, but it seems harder than I thought. My code is very sorry, for that I'm sorry, but it actually highlights all the keywords defined. The only exception is when a keyword doesn't have a delimiter character before it. This is a big problem for me. Here is what I have:

import java.io.*;
import java.util.*;
 
public class JavaToHTML
{
    static String[] keywords = {"do", "while", "static", "void", "import", "class", "public", "new"};
 
    public static void parseJava() throws IOException
    {
        StringBuffer line = new StringBuffer();
        String separatorOfLines = System.getProperty( "line.separator" );
 
        line.append( "<HTML><BODY BGCOLOR=\"DDDEEFF\"><PRE>" );
 
        try
        {
            BufferedReader br = new BufferedReader( new FileReader( "JavaToHTML.java" ) );
            String temp = null;
            while( ( temp = br.readLine() ) != null )
            {
                line.append( temp );
                line.append( separatorOfLines );
            }
        }
 
        catch(FileNotFoundException e) {}
        catch(IOException e) {}
 
        line.append("</PRE></BODY></HTML>");
 
        String s = line.toString();
        String finalString = "";
 
        StringTokenizer st = new StringTokenizer(s, " \n\t", true);
 
        while(st.hasMoreTokens())
        {
            String s2 = st.nextToken();
            System.out.println( s2 );
 
            for( int i = 0; i < keywords.length; i++ )
            {
                if( s2.equals( keywords[i] ) )
                {
                    finalString += "<FONT COLOR = \"FF0000\">" + keywords[i] + "</FONT>";
                    break;
                }
 
                else
                {
                    if ( i == keywords.length - 1 )
                    {
                        finalString += s2;
                    }
                }
            }
        }
 
        try
        {
            BufferedWriter bw = new BufferedWriter( new FileWriter( "Source.html" ) );
            bw.write( finalString );
            bw.newLine();
            bw.close();
        }
        catch(Exception e) {}
    }
 
    public static void main(String[] args) throws IOException
    {
        parseJava();
    }
}

This code is reading it's own .java file and creates an HTML file and should color all the keywords red. The first import statement isn't highlighted red because it has no character before it, and there are other times when this error can occur (ex: if(true) {break;} or any other keyword in a way like this). I believe I have bad design or I'm overlooking something. Can anyone help me out? Thanks
PS: Also if anyone can find a better algorithm or can do something more efficiently, I'd appreciate help on it or a solution to that particular problem. I know this code is pretty inefficient, but this is the best I could do. Thanks again.
[ April 26, 2002: Message edited by: Peter Allen ]
[ April 26, 2002: Message edited by: Dirk Schreckmann ]

Jim Yingst

Wanderer

Posts: 18671

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

Urg.
I'd recommend editing your post to put in some line breaks, so we can read it.
Also, in general "parsing" is rarely a beginner-level topic. This question might be better asked in intermediate. I'd transfer it, but I'll let you edit the question first.

"I'm not back." - Bill Harding, Twister

Peter Allen

Greenhorn

Posts: 13

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

Oops, sorry I should have seen that.

Peter Allen

Greenhorn

Posts: 13

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

I'm not trying to get too fancy with this. It ALMOST works, I just need help recognizing keywords when they're at different positions like the one I mentioned (like if its not surrounded by a space, tab, etc.) I just wanted help finding a way to fix this, otherwise it works the way I wanted to, even if it is very very messy. Thanks

Dirk Schreckmann

Sheriff

Posts: 7023

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

I'd guess Jim would accept the code formatting, now. So, I'm moving this to Java in General (intermediate)...

[How To Ask Good Questions] [JavaRanch FAQ Wiki] [JavaRanch Radio]

Mapraputa Is

Leverager of our synergies

Posts: 10065

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

What you are writing is called "Lexical analyzer" or "Lexer". Are you doing it because you need this functionality, for the fun of it, or as an educational experience? If the first is true, there is a generic tool that takes a grammar specification and generates Java classes to turn "raw" input into tokens: http://www.cs.princeton.edu/%7Eappel/modern/java/JLex/current/manual.html
Or you could use Java Compiler Compiler (JavaCC). This has one Java grammar specification ready, and even "JavaCC grammar to convert Java or JavaCC code to HTML" http://www.cobase.cs.ucla.edu/pub/javacc/#Jsection

Dirk Schreckmann

Sheriff

Posts: 7023

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

Originally posted by Peter Allen:
The only exception is when a keyword doesn't have a delimiter character before it.

One idea that occurs to me then, look for the keyword in the entire String (or StringBuffer) rather than trying to use a delimiter.
Also, you have a couple of places where using a StringBuffer (which is mutable while a String is not) would improve efficiency overall.

[How To Ask Good Questions] [JavaRanch FAQ Wiki] [JavaRanch Radio]

Jim Yingst

Wanderer

Posts: 18671

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

Or, skip saving things in Strings or StringBuffers entirely, and write your output directly to the Writer bw. OK, you'll have to save somethings along the way, but many things are not necessary. I'd just read one line from the file, process it, and write the results to the Writer before moving on.
To be fair, this sort of performance optimization is less important than getting the logic right. Here are some perfectly legal test cases to consider:

Can you highlight this correctly?
Incidentally, a related discussion may be found here.
[ April 27, 2002: Message edited by: Jim Yingst ]

"I'm not back." - Bill Harding, Twister

Peter Allen

Greenhorn

Posts: 13

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

Thanks for the help. My friend found what was wrong, the HTML tags I was inserting was affecting the final String and how the StringTokenizer worked. I knew I had it working, but at the same time I feel stupid for not catching it. Thank you for your help though.

Steve Deadsea

Ranch Hand

Posts: 125

posted 20 years ago

Number of slices to send:

Optional 'thank-you' note:

Send

There are already libraries out there that will do what you want.
http://ostermiller.org/syntax/

Well THAT's new! Comfort me, reliable tiny ad:

free, earth-friendly heat - a kickstarter for putting coin in your pocket while saving the earth

https://coderanch.com/t/751654/free-earth-friendly-heat-kickstarter