Re: [aspectj-users] @DeclareMixin with inner class - Syntax ?

2013-04-02 Thread Andy Clement
Or you could use com.nested.my.concurrent.*.* (depending on your needs - Alexanders will choose everything below the prefix, the .*.* is more selective). If you want to you can also combine it with the type category type pattern: com.nested.my.concurrent.*.* is(InnerType)

Re: [aspectj-users] AbstractMethodError calling method defined on class receiving ITD of same method

2013-04-02 Thread Matthew Adams
Hey, Andy, what did you decide to do with this scenario? Will this just remain a StupidUserError, or will a future version of the compiler emit an error? FWIW, I definitely favor the latter. I readily espouse the virtue of laziness http://c2.com/cgi/wiki?LazinessImpatienceHubris , and the

Re: [aspectj-users] @DeclareMixin with inner class - Syntax ?

2013-04-02 Thread Alexander Kriegisch
True, but com.nested.my.concurrent.*.* will not just match inner classes, but also regular classes in direct subpackages, which is why I said that if you need to be more specific you can concatenate subconditions with an or operator. ;-) Am 02.04.2013 um 17:01 schrieb Andy Clement

Re: [aspectj-users] @DeclareMixin with inner class - Syntax ?

2013-04-02 Thread Alexander Kriegisch
Sorry, it has gotten late here... I totally overlooked the is(InnerType) part. Mea culpa. Reminder to myself: read more carefully, then answer. And maybe get some sleep. Alexander Kriegisch Am 02.04.2013 um 19:35 schrieb Alexander Kriegisch alexan...@kriegisch.name: True, but

Re: [aspectj-users] @DeclareMixin with inner class - Syntax ?

2013-04-02 Thread Jean Andre
Thank you very-very much for theses answers and we updated our code according to the new knowledge. Did you test with || is(InnerType) (instead of ) ? With is(InnerType), we got only the inner classes. Using OR triggers some compiler errors. Well, from your side, the proposed solutions are

Re: [aspectj-users] @DeclareMixin with inner class - Syntax ?

2013-04-02 Thread Alexander Kriegisch
@Jean: The || was not for is() (which I had not even noticed, as I said), but for combinations of specific package and class name specifiers. Please do not mix up my answer with Andy's (which is superior because more precise anyway). BTW, I did not test anything, because I am nowhere near a PC