[
https://issues.apache.org/jira/browse/DBUTILS-61?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12773150#action_12773150
]
Dan Fabulich commented on DBUTILS-61:
-------------------------------------
It seems that KeyedHandler was only a map of maps by convention; it was really
a KeyedHandler<K,V> implements ResultSetHandler<Map<K,V>>.
But there's no way to implement that API concretely! So I think the best thing
is just to create an explicit AbstractKeyedHandler<K,V> implements
ResultSetHandler<Map<K,V>>, so KeyedHandler extends
AbstractKeyedHandler<Object, Map<String, Object>>. This will also fix binary
compatibility, as the compiler will append a #createRow method that returns
Object.
> Backwards binary compatibility broken in KeyedHandler
> -----------------------------------------------------
>
> Key: DBUTILS-61
> URL: https://issues.apache.org/jira/browse/DBUTILS-61
> Project: Commons DbUtils
> Issue Type: Bug
> Reporter: Dan Fabulich
> Fix For: 1.3
>
>
> In revision r832220 KeyedHandler's API was modified. KeyedHandler#createRow
> used to be able to return any object, but now can only return a Map (and a
> map of Strings to Objects, at that).
> clirr reports this as a binary backwards compatibility error; even "correct"
> subclasses of KeyedHandler that attempt to override createRow will find that
> we ignore the subclass implementation and simply invoke
> KeyedHandler#createRow instead.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
