[ 
https://issues.apache.org/jira/browse/FELIX-4565?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Clement Escoffier closed FELIX-4565.
------------------------------------

> Occasional ArrayIndexOutOfBoundException in iPOJO's ProvidedServiceHandler
> --------------------------------------------------------------------------
>
>                 Key: FELIX-4565
>                 URL: https://issues.apache.org/jira/browse/FELIX-4565
>             Project: Felix
>          Issue Type: Bug
>          Components: iPOJO
>    Affects Versions: ipojo-runtime-1.11.2
>         Environment: Karaf 2.3.3 / Windows 7 64 bit
>            Reporter: Benjamin Debeerst
>            Assignee: Clement Escoffier
>             Fix For: ipojo-runtime-1.12.1
>
>
> I have an iPOJO (1.11.2) component using the @ServiceController annotation to 
> control the service publishing. Most of the time this works perfectly fine, 
> but occasionally I get an ArrayIndexOutOfBoundsException.
> {code}
> java.lang.ArrayIndexOutOfBoundsException: 4
>                at 
> org.apache.felix.ipojo.handlers.providedservice.ProvidedServiceHandler.__M_onSet(ProvidedServiceHandler.java:416)[83:org.apache.felix.ipojo:1.11.2]
>                at 
> org.apache.felix.ipojo.handlers.providedservice.ProvidedServiceHandler.onSet(ProvidedServiceHandler.java)[83:org.apache.felix.ipojo:1.11.2]
>                at 
> org.apache.felix.ipojo.InstanceManager.onSet(InstanceManager.java:1401)[83:org.apache.felix.ipojo:1.11.2]
>                [..] (My own code following, where the service controller 
> boolean is manipulated)
> {code}
> Unfortunately I cannot share the consumer code nor reproduce the problem. I 
> looked into [iPOJOs 
> code|http://grepcode.com/file/repo1.maven.org/maven2/org.apache.felix/org.apache.felix.ipojo/1.11.2/org/apache/felix/ipojo/handlers/providedservice/ProvidedServiceHandler.java#416]
>  to see what's going on, and there we have:
> {code}
> for (int j = 0; j < svc.getProperties().length; j++) { Property prop = 
> svc.getProperties()[j];
>                [...]
> }
> {code}
> Which looks like a proper race condition to me, if svc.getProperties() 
> changes its value in the meantime because of missing synchronizations. I 
> don't see any synchronizations there, but I also don't know if they would be 
> appropriate in the first place.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to