you cant use normal aop on ejb beans since they are changed when deployed on
jboss. but if you use ejb3 you can use the interceptor util thats included in
the spec to do what you ask for.
so since you call entitymanager.find(..) im guessing you are using ejb3. if you
are take a look at
the code i provided above will catch any exception thrown in a method that is
intercepted. just add the above interceptor to the methods that can/will throw
an exception and it will catch it.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4147521#4147521
Reply
hi, i just created a small example you can download here:
http://wiki.jboss.org/wiki/attach?page=JBossAOPMavenPlugin%2Fmaven-jbossaop-example.tar.gz
its very basic, but it has two modules and the second module use a interceptor
defined in the first one. the poms are a bit different, but see the
the xml file was a bit hard to read, please include it in the code tags :)
- and yes, you can define which packages you want to transform and which ones
you want to skip. look at
jboss/server/all/deploy/jboss-aop-jdk50.deployer/META-INF/jboss-service.xml.
change the Include/Exclude tags as
in your example you might set:
| property name=includecom.test.core.aop/property
| property name=excludeorg, com, etc/property
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4113466#4113466
Reply to the post :
after some input i changed the groupid to org.jboss.maven.plugins as im told
thats the default groupid for maven plugins.
the goal was also changed from aopc to compile (now its jbossaop:compile)
there is also support for compiling tests with: jbossaop:compile-test
View the original post :
thanks, my pleasure :)
i can only agree with you that the maven respositories could easily be much
better, but i believe its getting there. jboss have its own repository, a
snapshot repository and think that modules will be updated faster in the future.
remember that many projects (jboss aop
currently the plugin depends on jboss aop 2.0.0alpha2 since thats the latest
maven release atm. the plan is that it will depend on 2.0 when its released.
mainly since thats what we want people to use and it also has less depencies on
other modules. - but there is nothing in the maven plugin
i added maven-jbossaop-plugin to the aop svn respository. i guess we'll move it
to the jboss-plugins repository when the time is right, but i just added it
here so the hardcore aop users could test it out :)
small howto:
1. get the source and run (in the maven-jbossaop-plugin folder): mvn
i got a compile plugin working on my local repository, it needs a bit more work
before ill push it out though. sorry that this has taken so long, but ive been
swamped at work for a while. ill try to get around and atleast push it out on
svn in a few days...
View the original post :
hi, i believe youre in the wrong forum. that class is provided by jbosscache.
- but from what i can see in the jbosscache repository this file has been moved
to org.jboss.cache.pojo.memory and has fixed the log error.
View the original post :
not atm. - but im looking for one too, so there might be one in
the-not-too-distant-future :)
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4030617#4030617
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=4030617
hi, im afraid its impossible to use introduction with dynamic aop. the reason
is that for introduction, the class is altered (which is impossible to do when
the class is already loaded by the jvm). for pointcuts its not needed to alter
the class after its been prepared.
so im afraid you have to
hi, ive been working on a ejb3 application that need to cache as much as
possible. queries are cached manually by storing the result (only the primary
keys) in jbosscache (the same cache as the entity beans).
when that query is called again we fetch the id's from cache and for each id we
just
hello, im working on a ejb3 application and i noticed that all the entities are
stored in the cache root node (as far as i can see). im wondering how optimal
this is...? in our application we will try to have as much in cache as possible
(100k-1m), but if all entities are added to the root i
ykrishnaprasad wrote :
| a) Can get the type of an invocation(whether its a METHOD)?
|
i guess you mean in the interceptor, so yes. just either try to cast it to a
MethodInvocation or check if its an instanceof MethodInvocation.
ykrishnaprasad wrote :
| b) Can i use
hm, thats strange, could you tell what files you copied over?
if jbossaop isnt working at all anymore many of the jboss applications will not
work either. eg (ejb3, cache, etc), so i guess you should have gotten a lot of
other errors too (even though i havent tested it).
View the original post
hi, look at the injboss example provided by the jbossaop package. you can also
find it online here:
http://docs.jboss.org/aop/1.3/aspect-framework/examples/injboss/aopInJbossPackaging.html.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3991954#3991954
Reply
hi, i havent used aspectj in many years so im outdated and probably a bit
biased, but...
aspectj is a more popular implementation, its older, it got a larger userbase
and it got a few features jboss doesnt have (but we're catching up :). the
reason i started to use jboss aop was because i just
i would say that we have a deadlock issue when one thread is using jbossaop
classloader and another is using the default one ( eg not the jbossaop
classloader). if all threads are using the jbossaop classloader we shouldnt see
this error (i hope).
i havent had the time to look more deeply into
hm, i might be far off, but if i remember correctly tomcat doesnt work well
with annotations (dont sue me if im wrong, its early in the morning).
could you try to deploy the code in jboss and not tomcat and see how it works?
since it works when you test is outside tomcat i believe that its no
hm, its seems like we got a general concurrent bug here. i was hoping it was
only generated advisors (SuperClassesFirstWeavingStrategy) that caused the
problem...
i looks like our systemclassloader is creating a deadlock, if you got the time.
could you try to deploy the app in jboss and see if
hm, its quite possible that we have several concurrent issues with the
generated advisors instrumentation that uses SuperClassesFirstWeavingStrategy.
you can easily tell jboss aop to use a different instrumentor (its only from
jboss aop 2.0 that generated advisors are default instrumentor).
add
hi, i guess this question should be answered by kabir. atm he is on vacation,
but he will answer you when he's back.
in the meantime i guess you have to add @Aspect on all subclasses (i dont know
how many classes you got, but hopefully its not too big of a job :)
View the original post :
hm, even though i havent tested it myself but i thought it should work. the
code you are calling is CtMethod/CtBehavior which is javassist, but thats not
your problem :)
have you tested .getAvailableParameterAnnotations() too?
if that doesnt work either, could i bother you to make a small test
hi, thanks for the feedback. SuperClassesFirstWeavingStrategy is used with the
generated advisors instrumentation. it is very different from the other weaving
types and as you can see, we havent fully tested it yet.
i have commited a quickfix to the problem (i hope) in svn. if you could try it
no, only for eclipse afaik.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3988093#3988093
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=3988093
___
jboss-user mailing list
hi, when you introduce an interface to a class, that class will always also
implement Advised. i do not know why it fails to compile, but it might be
because the rmic do not have Advised in its classpath?
- will it work to aopc after rmic?
View the original post :
if you are creating an interceptor within a ejb3 application i really reccomend
that you use the ejb3 interceptors. they are managed by the appserver and are
really easy to set up. - and they are part of the ejb3 spec too.
View the original post :
hi, ive tested your example and i get the same error. i can see that it matches
the pointcut and the generated code looks correct. so if im not mistaken ill
call this a bug.
i tried to change the pointcut to:
@Aspect(scope=Scope.PER_INSTANCE)
| public class MyAspect {
| @Bind
hi, you have an error in your jboss-aop.xml file, you need to wrap the
introduction defenitions in a mixin tag, eg:
| aop
|introduction class=Calculator
| mixin
| interfaces
| Main
|/interfaces
|class
|
compile time may be the easiest to set up and use if you're not very familiar
with jbossaop. compile and load time do not differ in functionality. with both
you can add a new pointcut/interceptor to a method/field/etc in runtime if the
class has been prepared first.
View the original post :
This is now fixed in svn.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3979905#3979905
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=3979905
___
jboss-user mailing list
hi, it looks like there is something wrong with the instrumentation. what
excactly is it that you are trying to do?
if you could post a small example that reproduce you problem, that would be
great.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3978289#3978289
this problem is now fixed, it was a bug in javassist. the fix is atm only
available from the javassist cvs repository, but the upcoming jboss-aop 2.0
will include this fix.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=3977088#3977088
Reply to the post :
try to add the property;
| property name=jboss.aop.verbose value=true /
| and see what the output, [debug] jboss.aop.class.path: says.
i would guess that the errormessage: [error] Unable to find
jboss.aop.class.path: will show up in the console log of javaws (with javaws
you can enable a
hi, if you could provide a small example ill take a look at it, but as far as i
can remember (long time since i made a webstart app) webstart is a web-deployed
application, and hence the classes used by the app must be packaged in a jar
and referenced in the .jnlp file.
- and i think that that
hm, how did you install jbossaop 1.5.1 into jboss as?
jbossaop 1.5.1 do come with org.jboss.aop.MethodInfo.getInterceptors(), so it
seems like there are some problems with old/new jbossaop jars.
earlier versions of jbossaop did not have this method (and in the repository
MethodInfo is moved away
as far as i can tell there is nothing wrong with your code, but the errormsg
you get is typical that the pointcut you defined with your @Bind doesnt exist.
eg: that you mistyped the classname/package/etc.
i tried your example where i just changed packagename to com, and pointcuted
another
hi, i misunderstood your question a bit, but if you test your subclasses on the
inherited method you will see that they will be intercepted even though the
subclass hasnt changed. For some transformation the subclass will also change,
but not always. eg:
aop
| interceptor name=MethodInt
hi, a simple workaround for the fieldproblem is to either to rename the field
variable so that it doesnt match the field of the subclass or make the field
variable private accessible. the only workaround for the overflow/loop problem
is to rename one of the methods calling. i'll make a post
hi, im not certain i understood you question, but ill try to clarify a bit.
first off your jboss-aop.xml is completly valid and will work, but its a bit
wierd. for readability i would recommend:
| ?xml version=1.0 encoding=UTF-8 standalone=yes?
| aop
| interceptor
42 matches
Mail list logo
