[jboss-jira] [JBoss JIRA] (JBJCA-1101) XAException::XA_RB∗ isn't fatal

[Jan Martiska (JIRA)]

    [ https://issues.jboss.org/browse/JBJCA-1101?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12999794#comment-12999794 ] 

Jan Martiska commented on JBJCA-1101:
-------------------------------------

I think the condition to determine if error code is fatal is not entirely correct:

{noformat}
return !(errorCode >= XAException.XA_RBBASE && errorCode < XAException.XA_RBEND);
{noformat}

The upper bound should include XA_RBEND too, this way it treats XA_RBEND and XA_RBTRANSIENT (both have the same value of 107) as fatal, which contradicts the name of this issue. So I think the correct line is:

{noformat}
return !(errorCode >= XAException.XA_RBBASE && errorCode <= XAException.XA_RBEND);
{noformat}

> XAException::XA_RB∗ isn't fatal
> -------------------------------
>
>                 Key: JBJCA-1101
>                 URL: https://issues.jboss.org/browse/JBJCA-1101
>             Project: IronJacamar
>          Issue Type: Bug
>          Components: JDBC
>    Affects Versions: 1.0.21.Final, 1.1.0.Final
>            Reporter: Tom Jenkinson
>            Assignee: Jesper Pedersen
>            Priority: Critical
>             Fix For: 1.0.22.Final, 1.1.1.Final
>
>
> In org.jboss.jca.adapters.jdbc.xa.XAManagedConnection:
> {code}
> try
> {
>      xaResource.end(xid, flags);
> }
> catch (XAException e)
> {
>      broadcastConnectionError(e);
>      throw e;
> }
> {code}
> should not consider XA_RB* error codes as fatal. Same goes for .start().



--
This message was sent by Atlassian JIRA
(v6.3.1#6329)