Hi Andy,
Thank you so much!
Unfortunately we have a little ltw vs. source weaving problem. Consider I
have the following setup:
Source Weaving:
##############
package ch.annotation;
import java.lang.annotation.ElementType;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target;
@Retention(RetentionPolicy.RUNTIME)
@Target(ElementType.FIELD)
public @interface Anno {
}
package ch.aspects;
import ch.annotation.Anno;
public aspect TriggerAll {
declare @field : * *.myInt : @Anno;
before(Anno anno) : @annotation(anno) && set(* *.myInt) {
System.out.println("Triggered");
}
}
package ch.tests;
public class MyObject {
public int myInt;
public int getMyInt() {
return myInt;
}
public void setMyInt(int myInt) {
this.myInt = myInt;
}
}
package ch.tests;
public class Main {
public static void main(String[] args) {
MyObject mo = new MyObject();
mo.myInt = 13;
mo.setMyInt(13);
}
}
This outputs:
Triggered
Triggered
// two times "Triggered". As expected.
Now LTW:
########
Infos:
- -javaagent:lib/aspectjweaver.jar is set
- the compiled aspect "TriggerAll.aj" is in the class searchpath
(classpath) packed within a jar file.
public class MyObject2 {
public int myInt;
public int getMyInt() {
return myInt;
}
public void setMyInt(int myInt) {
this.myInt = myInt;
}
}
public class Main {
public static void main(String[] args) {
MyObject2 mo = new MyObject2();
mo.myInt = 13;
mo.setMyInt(13);
}
}
This outputs:
Triggered
// just one "Triggered" message. And this comes from mo.setMyInt(13).
So mo.myInt = 13 does not trigger the pointcut!
Note: It has nothing to do with your recent changes. This happens
on AspectJ 1.6.12.
Output form the weaving "process":
[AppClassLoader@5acac268] info AspectJ Weaver Version 1.6.12 built on
Tuesday Oct 18, 2011 at 17:52:06 GMT
[AppClassLoader@5acac268] info register classloader
sun.misc.Launcher$AppClassLoader@5acac268
[AppClassLoader@5acac268] info using configuration /path/META-INF/aop.xml
[AppClassLoader@5acac268] info using configuration
/other-path/lib/TriggerAll.jar!/META-INF/aop-ajc.xml
[AppClassLoader@5acac268] info register aspect ch.aspects.TriggerAll
[AppClassLoader@5acac268] info weaver operating in reweavable mode. Need
to verify any required types exist.
[AppClassLoader@5acac268] debug weaving 'Main'
[AppClassLoader@5acac268] debug weaving 'MyObject2'
[AppClassLoader@5acac268] weaveinfo 'public int myInt' of type 'MyObject2'
(MyObject2.java) is annotated with @Anno field annotation from
'ch.aspects.TriggerAll' (TriggerAll.aj:6)
[AppClassLoader@5acac268] weaveinfo Join point 'field-set(int
MyObject2.myInt)' in Type 'MyObject2' (MyObject2.java:11) advised by before
advice from 'ch.aspects.TriggerAll' (TriggerAll.aj:8)
[AppClassLoader@5acac268] debug weaving 'ch.aspects.TriggerAll'
[AppClassLoader@5acac268] info processing reweavable type
ch.aspects.TriggerAll: ch\aspects\TriggerAll.aj
[AppClassLoader@5acac268] debug cannot weave
'org.aspectj.lang.NoAspectBoundException'
[AppClassLoader@5acac268] debug weaving 'ch.annotation.Anno'
[AppClassLoader@5acac268] debug weaving '$Proxy3'
Triggered
But again: Thanks very much for you effort,
domi
On Mon, Apr 2, 2012 at 11:28 PM, Andy Clement <[email protected]>wrote:
> Ok, implemented under https://bugs.eclipse.org/bugs/show_bug.cgi?id=375881
>
> Dev build is on the downloads page or here:
>
> http://www.eclipse.org/downloads/download.php?file=/tools/aspectj/dev/aspectj-DEVELOPMENT-20120402142200.jar
>
> It enables you to write declare-annotation segments in the
> concrete-aspect section of aop.xml. Here is one of my testcase XMLs:
>
> <aspectj>
> <aspects>
> <concrete-aspect name="ConjuredUp">
> <declare-annotation field="* field1"
> annotation="@Annot(someBoolean = false, 'abc' )"/>
> </concrete-aspect>
> </aspects>
>
> <weaver options="-Xreweavable -verbose -XlazyTjp -showWeaveInfo">
> <include within="Hello6"/>
> </weaver>
> </aspectj>
>
> You can use field or method (to indicate declare @field or declare
> @method), you cannot specify type yet. All annotation value types are
> supported apart from arrays and nested annotations. Notice to avoid
> the need to write double quotes as ", you can use single quotes
> for strings. If you are using a class reference it must either be in
> java.lang or fully qualified.
>
> There is still more error checking code to go in but hopefully you can
> use it as it is now.
>
> cheers,
> Andy
>
> On 1 April 2012 07:51, Dominik Mengelt <[email protected]> wrote:
> > Andy,
> >
> > Thanks for your efforts. It is really appreciated.
> >
> >> I suspect markers wont be enough for you? If so if you can tell me
> >> the types of value (string,int, etc) you need to specify in the
> >> annotation I'll do those first.
> >
> >
> > Thats correct. I need the following types:
> >
> > - string
> > - long
> > - boolean
> > - class
> >
> > It would be great to have a dev / nightly build with the ability to
> define
> > annotations within the aop.xml
> >
> > thanks again,
> > domi
> >
> >
> > On Fri, Mar 30, 2012 at 8:58 PM, Andy Clement <[email protected]>
> > wrote:
> >>
> >> > So I'm curious if there are currently any plans to implement this
> >> > feature or
> >> > can anybody provide an alternate solution to this problem?
> >>
> >> I didn't have plans but as you are asking for it I've done a bit of
> >> hacking this morning. I can currently add annotations to
> >> fields/methods but only simple marker annotations.
> >>
> >> I can do this
> >>
> >> <declare-annotation method="* foo(..))" annotation="@Marker"/>
> >>
> >> you cannot say
> >>
> >> annotation="@Marker(x=y,z=a)"
> >>
> >> I suspect markers wont be enough for you? If so if you can tell me
> >> the types of value (string,int, etc) you need to specify in the
> >> annotation I'll do those first. (Obviously supporting all of them
> >> would be nice, but not sure i have time to do that right now).
> >>
> >> Also declare @type won't work right now.
> >>
> >> cheers,
> >> Andy
> >>
> >> On 30 March 2012 04:30, Dominik Mengelt <[email protected]>
> wrote:
> >> > Hi all,
> >> >
> >> > I'm interested in the possibility to declare field and method
> >> > annotations
> >> > directly on concrete aspects in an aop.xml. Basically I have the same
> >> > "problem / feature request" as Thomas on the following post:
> >> >
> >> >
> http://aspectj.2085585.n4.nabble.com/abstract-quot-declare-method-quot-td3823887.html
> >> >
> >> > What I also found is the following enhancement request from Andrew:
> >> > https://bugs.eclipse.org/bugs/show_bug.cgi?id=359159
> >> >
> >> > On the end of comment 1 Andrew says:
> >> >
> >> >> "A similar model could be extended to
> >> >> support any declares too (anntation/parents)".
> >> >
> >> >
> >> > So I'm curious if there are currently any plans to implement this
> >> > feature or
> >> > can anybody provide an alternate solution to this problem?
> >> >
> >> > Thanks in advance.
> >> >
> >> > Domi
> >> >
> >> > _______________________________________________
> >> > aspectj-users mailing list
> >> > [email protected]
> >> > https://dev.eclipse.org/mailman/listinfo/aspectj-users
> >> >
> >> _______________________________________________
> >> aspectj-users mailing list
> >> [email protected]
> >> https://dev.eclipse.org/mailman/listinfo/aspectj-users
> >
> >
> >
> > _______________________________________________
> > aspectj-users mailing list
> > [email protected]
> > https://dev.eclipse.org/mailman/listinfo/aspectj-users
> >
> _______________________________________________
> aspectj-users mailing list
> [email protected]
> https://dev.eclipse.org/mailman/listinfo/aspectj-users
>
_______________________________________________
aspectj-users mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/aspectj-users
