Java source code generation frameworks

Question

I have a set of Java 5 source files with old-style Doclet tags, comments and annotations. And based on that I would like to write a generator for another set of Java classes.

What is the best way to do that? And are there any good standalone libraries for code analysis/generation in Java? Any shared exprience in this field is appreciated.

So, far I have found these:

• JaxME's Java Source Reflection - seems good, but it does not seem to support annotations. Also it had no release since 2006.

• Annogen - uses JDK's Doclet generator, which has some bugs under 1.5 JDK. Also it had no releases for a long time.

• Javaparser - seems good as well and pretty recent, but only supports Visitor pattern for a single class i.e. no query mechanism like in the 2 above packages.

Answer 1

Both the NetBeans IDE and Eclipse JDT projects have considerable Java code analysis/generation logic. I don't know what their dependencies are (i.e., can you use them as standalone libs), but other than that, I would take a good look at those two: it's unlikely there's a java code analysis library under more intensive development and more up to date.

Update:

PMD might be of interest as well:

PMD scans Java source code and looks for potential problems like:

* Possible bugs - empty try/catch/finally/switch statements
* Dead code - unused local variables, parameters and private methods
* Suboptimal code - wasteful String/StringBuffer usage
* Overcomplicated expressions - unnecessary if statements, for loops that could be while loops
* Duplicate code - copied/pasted code means copied/pasted bugs

Additionally, this blog entry discusses various static code analysis tools.

Answer 2

If you only need to generate syntactically correct Java code, check the Codemodel.

Answer 3

I ended up using PMD. Code example can be seen below:

    final Java15Parser parser = new Java15Parser();
    final FileInputStream stream = new FileInputStream("VehicleServiceType.java");

    final Object c = parser.parse(new InputStreamReader(stream));

    final XPath xpath = new BaseXPath("//TypeDeclaration/Annotation/NormalAnnotation[Name/@Image = 'WebService']",
        new DocumentNavigator());

    for (final Iterator iter = xpath.selectNodes(c).iterator(); iter.hasNext();) {
      final Object obj = iter.next();
      // Do code generation based on annotations...
    }