[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
RE: [aspectj-users] set() capture oldvalue
|
> the interface to
> provide access to the "before" value at a set join point is a little
> awkward: it really has to go on the JoinPoint interface
Not a binding form of set?:
before(int oldValue, int newValue) : set(oldValue) && set(int Foo.i)
&& args(newValue) {
System.out.println("setting Foo.i from " + oldValue + " to " + newValue);
}
The binding form is a tad bit more explicit in terms of what's required
to set up a join point. The reflective form is easier to make lazy.
Either way, is this as a field-get join point? If not, would that defeat
some forms of access control? If so, where/when is it? At the top
of the advice (which advice, if there are multiple advice at the join point?)
Is the value bound before the join point or before each advice? In cases
where some advice writes the field, should less-precedent advice see the
field as it was before any advice ran or as it is before that advice ran?
So a field-get before any advice? before any actual access (i.e., back
to thisJoinPoint.getFieldValue()?)
Assuming for the moment that reflection isn't too costly, I imagine that
a user could write a nice set of libraries for the standard reflective
access AspectJ users need:
public class AJReflect {
// usable in if() PCD
public static boolean isExactly(Class c, Object o) { ... }
public static boolean targetIsExactly(Class c, JoinPoint jp) { ... }
public static boolean thisIsExactly(Class c, JoinPoint jp) { ... }
// used mostly in advice
public static Object getFieldValue(JoinPoint jp) { ... }
public static Class getDeclaringClass(JoinPoint jp) { ... }
... others?
}
As a publicly-available library, it might reduce boilerplate code...
Wes
> ------------Original Message------------
> From: Adrian Colyer <adrian_colyer@xxxxxxxxxx>
> To: aspectj-users@xxxxxxxxxxx
> Date: Wed, May-18-2005 1:41 AM
> Subject: RE: [aspectj-users] set() capture oldvalue
>
> This is a reasonable requirement. There are some potential
> complications
> in the implementation to support after advice, and the interface to
> provide access to the "before" value at a set join point is a little
> awkward: it really has to go on the JoinPoint interface alongside
> getThis,
> getTarget, and getArgs, even though it only applies to set join points.
>
>
> If we did something like that you could write:
>
> after(Object newValue) returning : set( * *) && args(newValue) { //
> the
> pointcut would obviously be refined for a 'real' application...
>
> Object oldValue =
> thisJoinPoint.getOriginalFieldValueAtAFieldSetJoinPoint() ; // would
> need
> a better name...
> if (!oldValue.equals(newValue) {
>
> firePropertyChangeEvent(oldValue,newValue,thisJoinPoint.getSignature());
> }
> }
>
> Please raise an enhancement request on bugzilla if you think something
> like this would be useful, and then I encourage others who have
> opinions
> (for or against) to comment there.
>
> -- Adrian
> Adrian_Colyer@xxxxxxxxxx
>
>
>
> "Nitzan Volman" <nitzan@xxxxxxxxxxx>
> Sent by: aspectj-users-bounces@xxxxxxxxxxx
> 17/05/2005 15:34
> Please respond to
> aspectj-users@xxxxxxxxxxx
>
>
> To
> <aspectj-users@xxxxxxxxxxx>
> cc
>
> Subject
> RE: [aspectj-users] set() capture oldvalue
>
>
>
>
>
>
> Hi,
>
> I encountered the same problem, and I think what Tim suggested is
> needed
> in the language.
> My use case is simple: I am trying to audit changes to an object by
> monitoring field modifications, on each field modification (set
> operation)
> I need to compare the previous and value of the field with the new
> value
> to make sure the value really changed, no point in auditing fields that
>
> are reset to the same value.
> A typical scenario where this happens is when you submit web forms,
> most
> of the values are the same but the "set" is called anyway.
>
> I can think of several solutions for this use case that would not
> require
> adding a feature to AspectJ... But doing this using AspectJ would be a
> great enhancement
>
>
> /Nitzan Volman
>
>
>
>
>
>
> -----Original Message-----
> From: aspectj-users-bounces@xxxxxxxxxxx
> [mailto:aspectj-users-bounces@xxxxxxxxxxx] On Behalf Of Tim Schafer
> Sent: Monday, May 16, 2005 5:39 PM
> To: aspectj-users@xxxxxxxxxxx
> Subject: Re: [aspectj-users] set() capture oldvalue
>
>
> It's true in this particular application, the performance of reflection
>
> won't
> kill the application.
> But it's just something that I feel is missing from the language. I
> wanted
> to see what other use cases might exist. Performance could be a problem
> in
> some cases.
>
>
> Tim Schafer
> tschafer@xxxxxxxxxxx
>
>
>
> ----Original Message Follows----
> From: Matthew Webster <matthew_webster@xxxxxxxxxx>
> Reply-To: aspectj-users@xxxxxxxxxxx
> To: aspectj-users@xxxxxxxxxxx
> Subject: Re: [aspectj-users] set() capture oldvalue
> Date: Mon, 16 May 2005 07:50:11 +0100
> MIME-Version: 1.0
> Received: from mail.eclipse.org ([206.191.52.53]) by
> mc4-f24.hotmail.com
> with Microsoft SMTPSVC(6.0.3790.211); Sun, 15 May 2005 23:50:49 -0700
> Received: from node3.eclipse.org (localhost [127.0.0.1])by
> mail.eclipse.org
> (Postfix) with ESMTP id E1D40CB4D;Mon, 16 May 2005 02:52:44 -0400 (EDT)
> Received: from mtagate2.uk.ibm.com (mtagate2.uk.ibm.com
> [195.212.29.135])by
> mail.eclipse.org (Postfix) with SMTP id E5C276AC0Ffor
> <aspectj-users@xxxxxxxxxxx>; Mon, 16 May 2005 02:51:44 -0400 (EDT)
> Received: from
> d06nrmr1407.portsmouth.uk.ibm.com(d06nrmr1407.portsmouth.uk.ibm.com
> [9.149.38.185])by mtagate2.uk.ibm.com (8.12.10/8.12.10) with ESMTP id
> j4G6oD4G320918for <aspectj-users@xxxxxxxxxxx>; Mon, 16 May 2005
> 06:50:13
> GMT
> Received: from d06av04.portsmouth.uk.ibm.com
> (d06av04.portsmouth.uk.ibm.com[9.149.37.216])by
> d06nrmr1407.portsmouth.uk.ibm.com (8.12.10/NCO/VER6.6) with ESMTP
> idj4G6oDhG282280for <aspectj-users@xxxxxxxxxxx>; Mon, 16 May 2005
> 07:50:13
>
> +0100
> Received: from d06av04.portsmouth.uk.ibm.com (loopback [127.0.0.1])by
> d06av04.portsmouth.uk.ibm.com (8.12.11/8.13.3) with ESMTP
> idj4G6oDPY003034for <aspectj-users@xxxxxxxxxxx>; Mon, 16 May 2005
> 07:50:13
>
> +0100
> Received: from d06ml067.portsmouth.uk.ibm.com
> (d06ml067.portsmouth.uk.ibm.com[9.149.38.140])by
> d06av04.portsmouth.uk.ibm.com (8.12.11/8.12.11) with ESMTP
> idj4G6oCZe003029for <aspectj-users@xxxxxxxxxxx>; Mon, 16 May 2005
> 07:50:12
>
> +0100
> X-Message-Info: GQXpnklFM/ejFGBEcP5COAwEw6lHHLtKPmpYmPhOgiU=
> X-Original-To: aspectj-users@xxxxxxxxxxx
> Delivered-To: aspectj-users@xxxxxxxxxxx
> Sensitivity:
> X-Mailer: Lotus Notes Release 6.5.3 September 14, 2004
> X-MIMETrack: Serialize by Router on D06ML067/06/M/IBM(Release 6.53HF247
>
> |January 6, 2005) at 16/05/2005 07:50:12
> X-BeenThere: aspectj-users@xxxxxxxxxxx
> X-Mailman-Version: 2.1.5
> Precedence: list
> List-Id: aspectj-users.eclipse.org
> List-Unsubscribe:
> <https://dev.eclipse.org/mailman/listinfo/aspectj-users>,<
> mailto:aspectj-users-request@xxxxxxxxxxx?subject=unsubscribe>
> List-Archive: <http://eclipse.org/pipermail/aspectj-users>
> List-Post: <mailto:aspectj-users@xxxxxxxxxxx>
> List-Help: <mailto:aspectj-users-request@xxxxxxxxxxx?subject=help>
> List-Subscribe:
> <https://dev.eclipse.org/mailman/listinfo/aspectj-users>,<
> mailto:aspectj-users-request@xxxxxxxxxxx?subject=subscribe>
> Errors-To: aspectj-users-bounces@xxxxxxxxxxx
> Return-Path: aspectj-users-bounces@xxxxxxxxxxx
> X-OriginalArrivalTime: 16 May 2005 06:50:50.0047 (UTC)
> FILETIME=[98796CF0:01C559E3]
>
>
>
>
>
> Tim,
>
> If you have a concern about the performance of something measure it.
> Many
> frameworks use performance. I don't believe it's impact on overall
> application performance will be as bad as you think.
>
> 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/
>
> "Tim Schafer" <tschafer@xxxxxxxxxxx>@eclipse.org on 13/05/2005 19:11:36
>
> Please respond to aspectj-users@xxxxxxxxxxx
>
> Sent by: aspectj-users-bounces@xxxxxxxxxxx
>
>
> To: aspectj-users@xxxxxxxxxxx
> cc:
> Subject: Re: [aspectj-users] set() capture oldvalue
>
>
> That's exactly what I'm doing.
> Using reflection.
> I'm concerned about performance.
> If I were using aspects I wouldn't use refelection in each setter to
> get
> the old value. The aspect compiler should be able to do this work for
> me.
> I believe it's a major defeciency that this isn't avaiblable. Are there
>
> already plans to fix this? Should I file a bug?
>
> Tim Schafer
> tschafer@xxxxxxxxxxx
>
>
>
> ----Original Message Follows----
> From: Matthew Webster <matthew_webster@xxxxxxxxxx>
> Reply-To: aspectj-users@xxxxxxxxxxx
> To: aspectj-users@xxxxxxxxxxx
> Subject: Re: [aspectj-users] set() capture oldvalue
> Date: Fri, 13 May 2005 16:14:30 +0100
> MIME-Version: 1.0
> Received: from mail.eclipse.org ([206.191.52.53]) by
> mc7-f35.hotmail.com
> with Microsoft SMTPSVC(6.0.3790.211); Fri, 13 May 2005 08:15:02 -0700
> Received: from node3.eclipse.org (localhost [127.0.0.1])by
> mail.eclipse.org
>
> (Postfix) with ESMTP id 1DAB9E70F;Fri, 13 May 2005 11:17:14 -0400 (EDT)
> Received: from mtagate4.uk.ibm.com (mtagate4.uk.ibm.com
> [195.212.29.137])by
>
> mail.eclipse.org (Postfix) with SMTP id 6D0276ABD5for
> <aspectj-users@xxxxxxxxxxx>; Fri, 13 May 2005 11:16:05 -0400 (EDT)
> Received: from
> d06nrmr1407.portsmouth.uk.ibm.com(d06nrmr1407.portsmouth.uk.ibm.com
> [9.149.38.185])by mtagate4.uk.ibm.com (8.12.10/8.12.10) with ESMTP id
> j4DFEc78328870for <aspectj-users@xxxxxxxxxxx>; Fri, 13 May 2005
> 15:14:38
> GMT
> Received: from d06av02.portsmouth.uk.ibm.com
> (d06av02.portsmouth.uk.ibm.com[9.149.37.228])by
> d06nrmr1407.portsmouth.uk.ibm.com (8.12.10/NCO/VER6.6) with ESMTP
> idj4DFEbN9284252for <aspectj-users@xxxxxxxxxxx>; Fri, 13 May 2005
> 16:14:37
> +0100
> Received: from d06av02.portsmouth.uk.ibm.com (loopback [127.0.0.1])by
> d06av02.portsmouth.uk.ibm.com (8.12.11/8.13.3) with ESMTP
> idj4DFEb7M026517for <aspectj-users@xxxxxxxxxxx>; Fri, 13 May 2005
> 16:14:37
> +0100
> Received: from d06ml067.portsmouth.uk.ibm.com
> (d06ml067.portsmouth.uk.ibm.com[9.149.38.140])by
> d06av02.portsmouth.uk.ibm.com (8.12.11/8.12.11) with ESMTP
> idj4DFEbFm026512for <aspectj-users@xxxxxxxxxxx>; Fri, 13 May 2005
> 16:14:37
> +0100
> X-Message-Info: GQXpnklFM/deOXuyugSvCR2mxkvbZGrT7xa62XqVDbE=
> X-Original-To: aspectj-users@xxxxxxxxxxx
> Delivered-To: aspectj-users@xxxxxxxxxxx
> Sensitivity:
> X-Mailer: Lotus Notes Release 6.5.3 September 14, 2004
> X-MIMETrack: Serialize by Router on D06ML067/06/M/IBM(Release 6.53HF247
> |January 6, 2005) at 13/05/2005 16:14:37
> X-BeenThere: aspectj-users@xxxxxxxxxxx
> X-Mailman-Version: 2.1.5
> Precedence: list
> List-Id: aspectj-users.eclipse.org
> List-Unsubscribe:
> <https://dev.eclipse.org/mailman/listinfo/aspectj-users
> >,<
> mailto:aspectj-users-request@xxxxxxxxxxx?subject=unsubscribe>
> List-Archive: <http://eclipse.org/pipermail/aspectj-users>
> List-Post: <mailto:aspectj-users@xxxxxxxxxxx>
> List-Help: <mailto:aspectj-users-request@xxxxxxxxxxx?subject=help>
> List-Subscribe:
> <https://dev.eclipse.org/mailman/listinfo/aspectj-users>,<
> mailto:aspectj-users-request@xxxxxxxxxxx?subject=subscribe>
> Errors-To: aspectj-users-bounces@xxxxxxxxxxx
> Return-Path: aspectj-users-bounces@xxxxxxxxxxx
> X-OriginalArrivalTime: 13 May 2005 15:15:03.0565 (UTC)
> FILETIME=[89BCABD0:01C557CE]
>
>
>
>
>
> Tim,
>
> I'm sure there was a thread on either aspectj-users or aspectj-dev but
> I
> can't find it. Bottom line is the existing field value is not a
> property
> of the join point so you must get it directly. The simple examples
> below
> shows you how to use reflection or direct access:
>
> public class SomeClass {
>
> private int intField;
>
> public void setInt (int i) {
> intField = i;
> }
> }
>
> public aspect Aspect {
>
> pointcut fieldSet(int newValue) :
> set(* SomeClass+.*) && args(newValue)
> && withincode(* SomeClass+.set*(..))
> ;
>
> before (int newValue) : fieldSet (newValue) {
>
> Object oldValue = null;
> try {
> Class clazz =
> thisJoinPoint.getSignature().getDeclaringType();
> String fieldName =
> thisJoinPoint.getSignature().getName();
> Field field = clazz.getDeclaredField(fieldName);
> field.setAccessible(true);
> oldValue = field.get(thisJoinPoint.getThis());
> }
> catch (Exception ex) {
> ex.printStackTrace();
> }
> System.out.println("? beforeFieldSet() oldVaue=" +
> oldValue
> + ", newValue=" + newValue);
> }
> }
>
> public privileged aspect PrivilegedAspect {
>
> pointcut intFieldSet(SomeClass obj, int newValue) :
> set(* intField) && target(obj) && args(newValue)
> && withincode(* SomeClass+.set*(..))
> ;
>
> before (SomeClass obj, int newValue) : intFieldSet
> (obj,newValue)
> {
> int oldValue = obj.intField;
> System.out.println("? beforeIntFieldSet() oldVaue=" +
> oldValue
> + ", newValue=" + newValue);
> }
>
> }
>
> 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/
>
> "Tim Schafer" <tschafer@xxxxxxxxxxx>@eclipse.org on 11/05/2005 21:47:44
>
> Please respond to aspectj-users@xxxxxxxxxxx
>
> Sent by: aspectj-users-bounces@xxxxxxxxxxx
>
>
> To: aspectj-users@xxxxxxxxxxx
> cc:
> Subject: [aspectj-users] set() capture oldvalue
>
>
> I've used the following pointcut to capture all field modifications
> that
> occur within a setter method.
>
> pointcut fieldSet(Object newValue) : set(* SomeClass+.*) &&
> args(newValue)
> && withincode(* SomeClass+.set*(..));
>
> I can capture the new value, but I can't capture the old value without
> resorting to reflection. In theory aspectJ could provide access to the
> old
> value more statically. But aspectJ doesn't seem to support this. Or am
> I
> missing something?
>
>
> Tim Schafer
> tschafer@xxxxxxxxxxx
>
>
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
>
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
>
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
>
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
>
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
>
> _______________________________________________
> aspectj-users mailing list
> aspectj-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
