1 Reply Latest reply on Aug 22, 2018 8:22 AM by tomjenkinson

    Semantics of reading and writing config property values via arjPropertyManager


      I have created a JIRA ([JBTM-3049] Setting properties via arjPropertyManager should affect all related config bean instances - JBoss Issue Trac… ) to change the semantics of setting config properties via  com.arjuna.ats.arjuna.common.arjPropertyManager. This property manger is described in the Narayana Product Documentation.


      The JIRA explains that the mechanism by which the programmer changes config is via config beans which are returned by calls to arjPropertyManager or by calls to the BeanPopulator. Setting a property via arjPropertyManager will only update the default bean leaving other named instances unaffected. The JIRA proposes the following change in behaviour: setting a property via arjPropertyManager should update all named instances of the associated config bean. The getter methods on arjPropertyManager will continue to return the default bean instance for reading property values.


      The question is, are users okay with the change?