Thoughts: how to design utility similar to the unix/linux

Question

I want to take the "ls -la" output, in order to create binary tree nodes, which will contain the name of the file, its location, its size, and the permission.
I did "ls -la > tmp" and i am trying to use the output of the tmp file. Here's a snapshot:
-rw-r--r--   1 root   root   160920 Jun 11 21:10 BOOKOSI.TTF
-rw-r--r--   1 root   root   160940 Jun 11 21:10 BOOKOS.TTF
-rw-r--r--   1 root   root     7216 Jun 11 21:10 cga40737.fon
-rw-r--r--   1 root   root     6352 Jun 11 21:10 cga40850.fon
-rw-r--r--   1 root   root     6672 Jun 11 21:10 cga40852.fonAs you can see there are 9 collumns, which are THEORETICALLY separated with spaces(or tabs) and there is also a EOL at the end. Unfortunately both assumptions are wrong.
Hers's a part of the program, i have written:
    private static void analyzeInputCreateNodes() throws IOException
        File f= new File("/home/kostas/tmp3");
        FileInputStream fin= new FileInputStream(f);
        DataInputStream in= new DataInputStream(fin);               
        StreamTokenizer tokenizer=new StreamTokenizer(in); 
        tokenizer.resetSyntax();
        //tokenizer.wordChars(0,100);
        String s="";
        String s2="";
        String[] entryTable= new String[10];
       // String[] entryTable={"0","0","0","0","0","0","0","0","0","0"};
        StringBuffer sb= new StringBuffer();
        int lineCounter=0;
        int entryCounter=0;
        while(tokenizer.nextToken()!= StreamTokenizer.TT_EOF)
                s=(new Character((char)tokenizer.ttype)).toString();
                if (tokenizer.ttype==StreamTokenizer.TT_EOL)
                    for (int i=0; i<=9; i++)
                        System.out.print(entryTable[i]+" ");
                        //System.out.println();
                    entryCounter=-1;                                                          
                    entryTable[++entryCounter]=new Integer(lineCounter++).toString();
                if (!(s.equals(" ")))
                    s2=sb.append(s).toString();
                else
                    if (s.equals(" "))
                    entryTable[++entryCounter]=s2;
                lineCounter++;
         }And here's what i found out when i changed the code and added the following part:
                if (tokenizer.ttype==StreamTokenizer.TT_EOL)
                  System.out.println("EOL");
                s=(new Character((char)tokenizer.ttype)).toString();
               if (s.equals(" ")) System.out.println("space");This output proved my assumptions wrong:
space
space
space
space
space
space
space
space
EOL
space
space
space
space
space
space
space
space
EOL
space
space
space
space
space
space
space
space
EOL
space
......Do you have any ideas about how i can improve this?
Is there a way to create string patterns? (I don't know if it's called that way).
Post some keywords, links about it.
Thanks

Answer 1

Not clear what the problem is, that is, what you feel the problem is. Is it that you're getting one whitespace character between tokens on the input lines? Are you sure that this isn't an artifact of StreamTokenizer? (maybe it's collapsing whitespace?)
The problem with parsing the output of "ls -la" is that its output can be more complex than you may be anticipating. For example, I believe the date format will change if the file is sufficiently old. This is going to be system-dependent, of course, and thus not so portable. This is the kind of problem that explodes in complexity when you least expect it.
Is there any particular reason that you're not just using java.io.File?