VerifyError on javassist generated constructor

[jpschewe]

I'm generating classes on the fly using javassist and it seems that javassist is creating the incorrect bytecodes for super calls in constructors. I use this these API calls to create a constructor:
final CtConstructor cons1 = new CtConstructor(new CtClass[] {
pool.get("dome.interpretation.InterpretationFactory"),
pool.get("dome.datamodel.FirstClassObject")
}, clazz);
cons1.setBody("super((dome.interpretation.InterpretationFactory)$0, (dome.datamodel.FirstClassObject)$1);");
clazz.addConstructor(cons1);

And when it runs get this error:
java.lang.VerifyError: (class: dome/model/guid212906758570114/PlaceImpl, method: signature: (Ldome/interpretation/InterpretationFactory;Ldome/datamodel/FirstClassObject;)V) Expecting to find object/array on stack
at java.lang.Class.forName0(Native Method)
at java.lang.Class.forName(Class.java:242)
at dome.interpretation.InterpretationFactory.createInstanceOf(InterpretationFactory.java:133)
at dome.Test.main(Test.java)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
at java.lang.reflect.Method.invoke(Method.java:585)
at dome.Bootstrap.main(Bootstrap.java:33)

It also seems odd that I need to cast $0 and $1 to their datatypes, however if I don't then the javassist compiler complains that it cannot find the appropriate method. This is also true in methods as well as constructors. Does anyone have any workarounds?

[chiba]

$0 means "this". So maybe a right program is:

cons1.setBody("super((dome.interpretation.InterpretationFactory)$1, (dome.datamodel.FirstClassObject)$2);");

Note that the arguments are $1 and $2.

however if I don't then the javassist compiler complains that it cannot find the appropriate method. This is also true in methods as well as constructors.

This is because the type of $0 is not what you expected.

[jpschewe]

That might help. I'll have to try that. The documentation states that $0...$n are arguments, so I figured that $0 is the first arg. Perhaps someone can update the docs to explain this.

[chiba]

I see you misinterpreted the summary table of
special variables. I will update it to avoid such
misinterpretation.

Thank you for your suggestion.