[ https://jira.jboss.org/jira/browse/WELD-309?page=com.atlassian.jira.plugi... ] Stuart Douglas commented on WELD-309: ------------------------------------- After looking into this it appears the problem is the way the equality is defined for RIBean. Because the equals method only uses the RIBeans id, which is just bean type and class name, only one ManagedBean of a given type can be installed. This problem will only be evident when adding beans in the BeforeBeanDiscovery phase, as in normal operation you will not have the same java class with different annotations. As far as I can tell the correct way to fix this is to encode annotation information into a beans id, but I am not familiar enough with the code to know if this will cause problems elsewhere. Is this a reasonable approach? -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/WELD-309?page=com.atlassian.jira.plugi... ] Stuart Douglas commented on WELD-309: ------------------------------------- True, really what we need is some kind of unique id string for RIBean, that contains all the methods, fields etc and associated annotations, which would result in a rather long bean id (not sure if this would cause a problem?). I see what you mean about the equality on AnnotatedType, it does not even check the type of item so a WeldMethod can be equal to a WeldClass if they have the same annotations. I think that this can be solved using the same approach, I will experiment and try an get a patch ready for review in the next two days. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/WELD-309?page=com.atlassian.jira.plugi... ] Pete Muir commented on WELD-309: -------------------------------- The patch is good enough to use as a workaround, but I would prefer to investigate if it is possible to switch to using object equality is possible. Off the top of my head, I believe it is (it will require a couple of assumptions to be fixed I guess), but as we cache loaded WeldClass's by underlying class and by AnnotatedType, I think we will always use the correct object and hence this should be sufficient. I will apply this patch in a bit and run the full suite. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira