Args pointcut with at least 1 matching param in undefined order wanted

[unknown12]

Hi,

i want to define a pointcut which matches all public methods of a certain packacke with at least one certain run-time parameter passed as argument to such a method. Example:

execution (public * my.simple.packacke..*(..)) and (args(authorizable))

This works for all methods that have exactly one parameter of type "my.simple.interfaces.Authorizable". But of course later on a developer could introduce a new method into the above package with a signature like this
my.simple.packacke.MyClass#foo(int i, Authorizable authorizable). Now the aspect handling the authorization mechanism would not be called.

So i need a way to define all methods in a desired package with at least one (or more) parameters passed which are of type X. Something like (args(*, authorizable, *)). But this and other combinations doesn't seem to work. I tried:
- args(*, authorizable, *) - no error but also no match *grr*
- args(authorizable, ..) - matches only if the first param is of type Authorizable
- args(authorizable) or args(*, authorizable) or (authorizable, *) - matches only if the first or last param is of type Authorizable or there is only one param of type Authorizable.

Of course i could implement the aspect in a way that it determines if a param is of type Authorizable via the JoinPoints#getArgs() method. This already works, but i would be a bit more satisfied if i could constrain the pointcut expression.

TIA for any ideas,
Dirk

[cgdavies]

Try:

execution (public * my.simple.package..*(..,authorizable,..))

[unknown12]

The special wildcard '..' may be used only once in such an expression.

[cgdavies]

Well - I'm pretty new to AOP (completely new to AspectJ style pointcut expressions), so I wrote some test cases before replying above. In my test case, I have:

<aop:config>
	<aop:aspect ref="testAdvice">
		<aop:before method="foo" pointcut="execution(* test..*(..,test.Foo,..))" />
	</aop:aspect>
</aop:config>

This works perfectly in my tests.

However, if you say it doesn't work for you, then it doesn't work - I'll leave it for someone more experienced with AOP. I'd be interested to see the suggested solution.