Re: [rvm-research] Regarding the function call mechanism

On 3/25/2013 11:54 AM, Hari Krishnan wrote:

> I have a question. Is there any way I can say that the method arguments are mutated inMMTK or in
> Jikes RVM itself? how can it be done?

Dear Hari --

Do you mean: Is there a way to tell whether a method's arguments
are mutated?

Supposing that is your question, it could have two meanings:

1) For an argument x (say), there is an assignment x = ... that
occurs in the method.  This is easily checked statically by looking
at the bytecode of the method. If you mean "does it occur on a
particular invocation", then you require a dynamic check, but a
bytecode rewriting tool would allow you to instrument the places
where an argument is mutated.

By the way, in code generated by javac, I don't think argument
0 in an instance method (called 'this' in Java) is ever changed.

2) For an argument x that refers to some object, that *object*
is mutated by the method.  That is rather a harder question to
answer, since the object could be passed in a call, and mutated
that way.  If you mean a direct mutation (setting a field,
changing an array element) bytecode analysis could work fairly
well to determine that statically, or to instrument to tell you
dynamically.


If you are concerned about debugging a particular method, then
Eclipse watchpoints might be helpful.


As with many questions asked here, giving more background and
explaining what your overall objective is will likely obtain a
more useful answer.

Regards -- Eliot Moss

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_mar