Re: [rules-users] Re: Odd static inner class behavior
Hi Eric, after some tests I noticed that in java language you can't
import the same class name twice even if the class is defined in
different packages, ie try creating a java file and import both
java.awt.List and java.util.List, it won't work, you will get a
similar error from javac compiler
So, if it is really necessary you can try using two package builders
as I did in the following example
public void testImportColision () throws Exception {
final PackageBuilder builder = new PackageBuilder();
final PackageBuilder builder2 = new PackageBuilder();
builder.addPackageFromDrl( new InputStreamReader( getClass
().getResourceAsStream( nested1.drl ) ) );
builder2.addPackageFromDrl( new InputStreamReader( getClass
().getResourceAsStream( nested2.drl ) ) );
final Package pkg = builder.getPackage();
final Package pkg2 = builder2.getPackage();
final RuleBase ruleBase = getRuleBase();
ruleBase.addPackage( pkg );
ruleBase.addPackage( pkg2 );
final WorkingMemory workingMemory =
ruleBase.newStatefulSession();
workingMemory.insert( new FirstClass() );
workingMemory.insert( new SecondClass() );
workingMemory.insert( new FirstClass.AlternativeKey() );
workingMemory.insert( new SecondClass.AlternativeKey() );
workingMemory.fireAllRules();
}
Regards
Fernando
On Aug 1, 2007, at 1:15 PM, Eric Miles wrote:
Edson,
I was finally able to get around to making a test case to determine
if I did in fact see this behavior. I can reproduce it and have
opened up a Jira for it. I have also attached my test case for
easy reproduction. For your (and other's) knowledge, here is the link:
http://jira.jboss.com/jira/browse/JBRULES-1053
Truly odd behavior for sure!
Thanks,
Eric
Edson Tirelli wrote:
Eric,
Please do! Thanks,
Edson
2007/7/26, Eric Miles [EMAIL PROTECTED]
mailto:[EMAIL PROTECTED]:
Edson,
That certainly makes sense. However I'm fairly certain that in
referencing the inner class in rule definition, I always
qualified it
with the outer class name, ie:
DataClass.AlternativeKey()
or
AnotherClass.AlternativeKey()
I appreciate your explaination of the merge process. Rather
than have
you spend any more time on this, I'll try to put together a
test case to
ensure I was seeing the behavior I thought I was seeing. I
probably
won't get around to this until tonight or the weekend.
If I was mistaken, I'll let you (and the mailing list) know.
If I was
not, would you like me to open a JIRA with the attached test
case? I
would assume that if the inner classes contain the qualified
name that
the engine should be able to handle that?
Thanks,
Eric
Edson Tirelli wrote:
Eric,
Thanks, I understand now.
What happens is that if both DRL files declare the
same package
name, all their contents will be merged. It means that you
would
end up
with both imports in the same namespace:
import com.company.DataClass.AlternativeKey;
import com.company.AnotherClass.AlternativeKey;
And so the engine will raise an error saying that it does
not know
which one you are refering to when you write simply:
AlternativeKey
I think the engine behavior is correct, since the idea
of loading
two different files with the same name space into the same
package
builder is to merge them, or even replace (update) that
eventually have
the same name.
What do you think?
Edson
2007/7/26, Eric Miles [EMAIL PROTECTED]
mailto:[EMAIL PROTECTED]
mailto: [EMAIL PROTECTED]
mailto:[EMAIL PROTECTED]:
Edson,
I have since changed my schema but here was my issue:
rule1.drl:
import com.company.DataClass.AlternativeKey;
import com.company.DataClass;
rule Some rule
when
DataClass.AlternativeKey(someParm == true)
then
...
end
Different drlf file:
rule2.drl
import com.company.AnotherClass.AlternativeKey;
import com.company.AnotherClass;
rule Another rule
when
AnotherClass.AlternativeKey(diffParm == 1)
then
...
end
This was the gist of what I was doing. The outer
classes' names
were different, it was the INNER class of each of these
classes that
had the same name. I was actually getting compile
errors on the
import statements. Like I said, these rules worked
fine if
loaded
separately, but once I tried to put them all int he
same
[rules-users] Re: Odd static inner class behavior
Edson, That certainly makes sense. However I'm fairly certain that in referencing the inner class in rule definition, I always qualified it with the outer class name, ie: DataClass.AlternativeKey() or AnotherClass.AlternativeKey() I appreciate your explaination of the merge process. Rather than have you spend any more time on this, I'll try to put together a test case to ensure I was seeing the behavior I thought I was seeing. I probably won't get around to this until tonight or the weekend. If I was mistaken, I'll let you (and the mailing list) know. If I was not, would you like me to open a JIRA with the attached test case? I would assume that if the inner classes contain the qualified name that the engine should be able to handle that? Thanks, Eric Edson Tirelli wrote: Eric, Thanks, I understand now. What happens is that if both DRL files declare the same package name, all their contents will be merged. It means that you would end up with both imports in the same namespace: import com.company.DataClass.AlternativeKey; import com.company.AnotherClass.AlternativeKey; And so the engine will raise an error saying that it does not know which one you are refering to when you write simply: AlternativeKey I think the engine behavior is correct, since the idea of loading two different files with the same name space into the same package builder is to merge them, or even replace (update) that eventually have the same name. What do you think? Edson 2007/7/26, Eric Miles [EMAIL PROTECTED] mailto:[EMAIL PROTECTED]: Edson, I have since changed my schema but here was my issue: rule1.drl: import com.company.DataClass.AlternativeKey; import com.company.DataClass; rule Some rule when DataClass.AlternativeKey(someParm == true) then ... end Different drlf file: rule2.drl import com.company.AnotherClass.AlternativeKey; import com.company.AnotherClass; rule Another rule when AnotherClass.AlternativeKey(diffParm == 1) then ... end This was the gist of what I was doing. The outer classes' names were different, it was the INNER class of each of these classes that had the same name. I was actually getting compile errors on the import statements. Like I said, these rules worked fine if loaded separately, but once I tried to put them all int he same rule base, I was getting the import collision error. Later on this evening (when I'm not at work), I'll try to put together a small test case and upload it. In the meantime, you can look skim over this and let me know if something jumps out at you. Thanks, Eric On Thu, 2007-07-26 at 10:32 -0300, Edson Tirelli wrote: Eric, Not sure if I understood your problem, but if you have multiple classes with the same name, and the only difference is that they are inner classes of different classes, I guess what you need to do is to fully qualify your class names in your rules... rule xxx when my.package.MyClass.MyInnerClass( ... ) ... end If this is not your problem, can you please show us an example so we understand it better? Edson 2007/7/25, Eric Miles [EMAIL PROTECTED] mailto:[EMAIL PROTECTED]: Due to how JAXB treats anonymous inner complex types, I ended up with a public static inner classes named AlternativeKey in several of my data classes I have several rules written to deal with each data class individually that all work ok. However, when I attempt to put them all in the same rule base (all belong to the same package), I get an import collision exception on the AlternativeKey inner class. Depending on where in the builder I add the resource depends on which AlternativeKey the compiler bitches about (validity). I'm not familiar with the source at all, so I'm unsure as to where to look for this. However, this sounds like a bug to me? There is an easy workaround for this as I I just don't use anonymous types and define them in my schema explicitly. Just thought I'd identify this as a possi ble issue. Thanks, Eric ___ rules-users mailing list rules-users@lists.jboss.org mailto:rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users -- Edson Tirelli Software Engineer - JBoss Rules Core Developer Office: +55 11 3529-6000 Mobile: +55 11 9287-5646 JBoss, a division of Red Hat @ www.jboss.com http://www.jboss.com -- Edson Tirelli Software Engineer - JBoss Rules Core Developer Office: +55 11 3529-6000 Mobile: +55 11 9287-5646 JBoss, a division of Red
