RE: Doug's interface for traversal-related concerns


Subject: RE: Doug's interface for traversal-related concerns
From: Karl Lieberherr (lieber@ccs.neu.edu)
Date: Sat Mar 02 2002 - 17:21:47 EST


Hi John:

>From john_j_sung@yahoo.com Sat Mar 2 16:46:41 2002
>
>Hi Karl,
>
>I really like the declare visitor part, it's a very nice way of adding

I am glad you like it too.

>visitors to the whole system. I think that this is the way I should
>implement the syntax for the visitor as well. It shouldn't be too hard to
>find the definition for the particular visitor in the parse tree or loaded
>classes.

loaded classes please (reflection).
We want to add this again without touching the
AspectJ compiler.
Some of my COM 3362 students, e.g. Laura and Paul, already do this.

>
>I'm just wondering about syntax for adding advice to edges. Since, AspectJ
>is advising on methods and there's a wrapper method call for traversing an
>edge. We should be able to do this. So, what would be the syntax for adding
>advice to those?

We don't need new syntax. We use the Java syntax that Doug has
for DJ. That is our official new aspect-specific language for
traversal related concerns.

>
>John
>

-- Karl

>
>-----Original Message-----
>From: Pengcheng Wu [mailto:wupc@ccs.neu.edu]
>Sent: Friday, March 01, 2002 10:07 AM
>To: Karl Lieberherr
>Cc: dougo@ccs.neu.edu; johan@ccs.neu.edu; john_j_sung@yahoo.com;
>neerajsangal@mediaone.net; skotthe@ccs.neu.edu; com3362@ccs.neu.edu
>Subject: Re: Doug's interface for traversal-related concerns
>
>
>
>
>On Thu, 28 Feb 2002, Karl Lieberherr wrote:
>
>> Hi Doug, Pengcheng and John:
>>
>> I like the new design by Doug for edge patterns very much.
>> This will be our new interface for expressing
>> traversal related-concerns in DJ and in AspectDJ.
>>
>> John: please can you check whether you want to put this
>> interface into AspectDJ?
>> See: /home/lieber/.www/com3362/w02/hw/7/assign7-com3362.txt
>>
>> Pengcheng: do you agree with Doug re. calling all matching methods?
>
>
>I agree with Doug about calling all matching methods in that it more
>conforms to the adaptive programming implementation style. But I have some
>conservations about that in terms of standard object-oriemted programming
>style, i.e., in O-O, only most specific method get executed, unless the
>users
>explicitly call 'super.m()'.
>
>--Pengcheng
>
>
>
>
>
>_________________________________________________________
>Do You Yahoo!?
>Get your free @yahoo.com address at http://mail.yahoo.com
>
>



This archive was generated by hypermail 2b28 : Sat Mar 02 2002 - 17:21:50 EST