Re: [aspectj-users] array pointcuts?

[Matthew Webster <matthew_webster@xxxxxxxxxx>]

Mark,

>but I wonder if 'call' type joint points would be feasible.
If the join point doesn't exist then neither execution nor call pointcuts
will match. This isn't a question of the target class not being exposed to
the weaver e.g. you are tying to advise the construction of a library class
so you must use call rather than execution. I think what you are thinking
of is a "call to object allocation" join point which exists at the bytecode
level (new, newarray, ...) but not in the Java language: Java doesn't give
you uninitialized objects (array elements are initialized to their default
values).

Matthew Webster
AOSD Project
Java Technology Centre, MP146
IBM Hursley Park, Winchester,  SO21 2JN, England
Telephone: +44 196 2816139 (external) 246139 (internal)
Email: Matthew Webster/UK/IBM @ IBMGB, matthew_webster@xxxxxxxxxx
http://w3.hursley.ibm.com/~websterm/


"Mark van der Kraan" <mark@xxxxxxxxxxxxxxxx>@eclipse.org on 22/09/2004
11:07:59

Please respond to aspectj-users@xxxxxxxxxxx

Sent by:    aspectj-users-admin@xxxxxxxxxxx


To:    <aspectj-users@xxxxxxxxxxx>
cc:
Subject:    Re: [aspectj-users] array pointcuts?


Hi Matthew,

Thank you for your extensive answer. You guys must sometimes get a little
bit tired of these "why can't I do X" questions :-)

I can understand that you can't have 'execute' array construction join
points
(since the type is created at runtime), but I wonder if 'call' type joint
points
would be feasible. I'll take my questions to the dev list as you suggested.

Cheers,
Mark


> Mark,
>
> There is no initialization (or preinitialization) joint point for array
> types as they do not have constructors. They do have a
staticinitialization
> join point but this is not accessible as array classes are generated at
> runtime by the JVM.
>
>       String s = new String();            // Call constructor
>       String sa = new String[1];    // No constructor
>
> See this thread:
> http://dev.eclipse.org/mhonarc/lists/aspectj-users/msg02494.html
>
> In addition there is no array element get/set pointcut, these are both
> treated as get join points on the array object itself, so there is no way
> of telling which array element is being accessed or what it's value is.
The
> initial array allocation is a set join point. There has been some
> discussion on this subject
> http://dev.eclipse.org/mhonarc/lists/aspectj-users/msg00932.html before
> although none on aspectj-dev so you might like to request an enhancement
to
> the pointcut language there.
>
> Matthew Webster
> AOSD Project
> Java Technology Centre, MP146
> IBM Hursley Park, Winchester,  SO21 2JN, England
> Telephone: +44 196 2816139 (external) 246139 (internal)
> Email: Matthew Webster/UK/IBM @ IBMGB, matthew_webster@xxxxxxxxxx
> http://w3.hursley.ibm.com/~websterm/
>
>
> "Mark van der Kraan" <mark@xxxxxxxxxxxxxxxx>@eclipse.org on 21/09/2004
> 14:53:53
> To:    <aspectj-users@xxxxxxxxxxx>
> cc:
> Subject:    [aspectj-users] array pointcuts?
>
>
> Hi,
>
> I am somewhat surprised that there seems to be no support for pointcut
> expressions
> for array creation and array read/write access, similar to pointcuts for
> object creation and
> get/set field access. Am I wrong? And if I'm right, are there any plans
for
> adding
> these in the future?
>
> Cheers,
> Mark

_______________________________________________
aspectj-users mailing list
aspectj-users@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/aspectj-users