[
https://issues.apache.org/activemq/browse/CAMEL-1922?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=53706#action_53706
]
Adam Brewster commented on CAMEL-1922:
--------------------------------------
I wrote a unit test (the gist was FromBeanTest.java + the description field of
this issue), but apparently I suck at maven because can't I the test suite to
run.
Even after I checked out a fresh copy of the project, it still didn't work.
Occasionally when `mvn test` decides not to complain that camel-core is missing
(umm, didn't I just tell you to build that?) it takes ~ hours to finish (does
it start a new JVM for every test?) and finally complains about DataSet being
broken. I'm considering `rm -rf ~/.m2` but that seems pretty brutal.
I heard a rumor that with -Dtest=BridgeMethodTest, maven would only run my new
test, but given that option it reverts to complaining about camel-core.
I'll try again at some point.
> BeanEndpoint fails in the presence of bridge methods
> ----------------------------------------------------
>
> Key: CAMEL-1922
> URL: https://issues.apache.org/activemq/browse/CAMEL-1922
> Project: Apache Camel
> Issue Type: Bug
> Components: camel-core
> Affects Versions: 1.6.1
> Environment: Java SE 5
> Reporter: Adam Brewster
> Assignee: Hadrian Zbarcea
> Attachments:
> 0001-Reject-bridge-methods-while-introspecting-POJO-consu.patch
>
> Original Estimate: 2 days
> Remaining Estimate: 2 days
>
> When a message is sent to a bean: endpoint, camel searches for a method with
> the right signature. If multiple methods are found, an exception is thrown.
> If a have an interface like
> public interface <T> I {
> T function(T arg);
> }
> and a class like
> public class C implements I<Number> {
> Number function(Number arg) {
> return new Double(arg.doubleValue() * 2);
> }
> }
> then I can't use an instance of C as an endpoint.
> The problem is that under the hood java has created a bridge method, and the
> "real" class C looks like
> public class C implements I<Number> {
> Number function(Number arg) {
> return new Double(arg.doubleValue() * 2);
> }
> Object function(Object arg) {
> return function((Number) arg);
> }
> }
> Camel then discovers that there are two methods it could call (f(Object) and
> f(Number)) and throws an exception even though there's no real ambiguity
> because both methods do the same thing.
> I recommend rejecting bridge methods in BeanInfo.isValidMethod with somehting
> like
> if (method.isBridgeMethod()) return false;
> Does this sound reasonable?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
