[jboss-jira] [JBoss JIRA] (WFLY-1656) JCA adaptor model name should be unique

[Bartosz Baranowski (JIRA)]

    [ https://issues.jboss.org/browse/WFLY-1656?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12789086#comment-12789086 ] 

Bartosz Baranowski commented on WFLY-1656:
------------------------------------------

[~dmlloyd] I may have made a bad description. There are two things here:
1. incremental suffix added to model when un *id* 'ed def of resource-adapter is present in:
 - multiple times in one profile
 - in each profile( leak between profiles via static def )
2. seems like incremental suffix in service name - if un *id* 'ed defs are activated, it creates services with incremental suffix

Now trick is that any addition to profile or shift of profile defs will cause different IDs to be generated. 
                
> JCA adaptor model name should be  unique
> ----------------------------------------
>
>                 Key: WFLY-1656
>                 URL: https://issues.jboss.org/browse/WFLY-1656
>             Project: WildFly
>          Issue Type: Bug
>          Components: JCA
>    Affects Versions: 8.0.0.Alpha2
>            Reporter: Bartosz Baranowski
>            Assignee: Bartosz Baranowski
>             Fix For: 8.0.0.Alpha3
>
>
> This is an offspring of: https://issues.jboss.org/browse/WFLY-1492 .
> Example RA definition:
> {noformat}
> <resource-adapters>
>     <resource-adapter id="XXX">
>        <archive>test</archive>
>        <transaction-support>NoTransaction</transaction-support>
>     </resource-adapter>
> </resource-adapters>
> {noformat}
> The *id* attribute is optional. This leads to a bit undeterministic model expansion. There are two bits that dont add up or make this a bad apple.
> 1. RA service names are not unique, despite their model paths, the service name boils down to *archive name* ( and possibly *id* if its present). This seems odd, since in model we can point exact, unique path(profile, subsystem,.. and so on).
> 2. Second thing is the *id* piece. Its optional. If its not present, AS injects counter suffix( either due to #1 or because there is doubled definition of RA in the same profile). In the end RA paths look like:
> {noformat}
> /xxx/xxx/xxx/${archive-name}/${archive-name}
> /xxx/xxx/xxx/${archive-name}/${archive-name}->1
> /xxx/xxx/xxx/${archive-name}/${archive-name}->2
> /xxx/xxx/xxx/${archive-name}/${archive-name}->3
> {noformat}
> With *id* present:
> /xxx/xxx/xxx/${archive-name}/${id}
> Im not sure ATM, but its even possible to double with *id*
> {noformat}
>  /xxx/xxx/xxx/${archive-name}/${id}->1  
> {noformat}

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira