Line number counter using java AST parser

Question

I have some interesting exercise - to read .java file with Java, to find every code part in this .java file(like methods and etc.) and to say how many line numbers we have in every method without whitespaces and comments and also to find the count of items in every method.

Here is some example that I use - so could Java Parser. With the code I can say how many line numbers I have in method, but cannot remove the count of whitespaces and comments. Can someone help me?

The code:

public static void main(String[] args) throws ParseException, IOException {
    File f = new File(".").getAbsoluteFile();
    File srcRoot = new File(f, "src");
    String srcFilename = TestMethodLineNumber.class.getName().replaceAll("\\.", "/") + ".java";
    File src = new File(srcRoot, srcFilename);
    System.out.println(f);
    System.out.println(srcRoot);
    System.out.println(src);
    getMethodLineNumbers(src);
}

private static void getMethodLineNumbers(File src) throws ParseException, IOException {
    CompilationUnit cu = JavaParser.parse(src);
    new MethodVisitor().visit(cu, null);
}

/**
 * Simple visitor implementation for visiting MethodDeclaration nodes.
 */
private static class MethodVisitor extends VoidVisitorAdapter {
    @Override
    public void visit(MethodDeclaration m, Object arg) {
        System.out.println("From [" + m.getBeginLine() + "," + m.getBeginColumn() + "] to [" + m.getEndLine() + ","
                + m.getEndColumn() + "] is method:");
        System.out.println(m);
    }
}

}

Answer 1

It looks to me like JavaParser already removes blank lines, I would use a technique from this thread Removing all types of comments in java file to remove all the comments and then use JavaParser to get your results.

In general questions like these aren't always interesting unless you're actually going after executable code. Consider:

if (a)
{
    runA();
}
else
{
    runB();
}

You're count will be 8, what exactly does that tell you? Is it more interesting to know there is one if statement and/or two method calls? You'll really want better constraints to know what you're going after:

a ? runA() : b ? runB(): runError("nothing true");

Is only one line, but you have 3 potential execution branches. You can also run into some pretty crazy statements that span multiple lines.

MyFancyObject myFancyObject = new MyFancyObject("This is my really long name",
                                                123113212,
                                                "2014-11-20 18:43:12",
                                                anotherArg, yetAnother,
                                                moar);

Should this be 5 lines or 1?