[
https://issues.apache.org/jira/browse/DERBY-3579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12604627#action_12604627
]
Rick Hillegas commented on DERBY-3579:
--------------------------------------
Hi Kim,
Thanks for the patch. It looks good. However, I think that there is still a
statement in cdevspecial847513 which could be misleading:
"Note: You cannot roll back this statement, because commits occur within the
procedure itself."
This suggests to me that a procedure always either commits or rolls back its
work. This is not the case. It all depends on how the procedure is coded. If
the procedure does not explicitly call commit() or rollback() and no errors
occur while the procedure executes, then the calling code can roll back the
work done by the procedure.
> The Developer's Guide incorrectly describes the behavior of transactions
> inside procedures and functions
> --------------------------------------------------------------------------------------------------------
>
> Key: DERBY-3579
> URL: https://issues.apache.org/jira/browse/DERBY-3579
> Project: Derby
> Issue Type: Bug
> Components: Documentation
> Affects Versions: 10.3.2.1
> Reporter: Rick Hillegas
> Assignee: Kim Haase
> Attachments: DERBY-3579.diff, DERBY-3579.zip, z.java
>
>
> The "Programming database-side JDBC procedures" section of the Developer's
> Guide misleads users about how transactions can be managed inside database
> procedures. For instance, the section titled "Invoking a procedure using the
> CALL command" says the following: "Procedures that use nested connections, on
> the other hand, are not permitted to commit or roll back and can therefore be
> rolled back after the calling statement." This is not true. User-coded
> procedures can issue both commit() and rollback() on the nested connection
> bound to the "jdbc:default:connection" URL and those methods have the desired
> behavior. The whole "Programming database-side JDBC procedures" section could
> use an overhaul since it appears to make other misleading statements about
> the behavior of transactions inside routines. Thanks to Dag for pointing out
> the following email thread:
> http://www.nabble.com/Transaction-Problems-of-a-Derby-Stored-Procedure-td15494178.html
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
