[
https://issues.apache.org/jira/browse/METAMODEL-1093?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15314752#comment-15314752
]
ASF GitHub Bot commented on METAMODEL-1093:
-------------------------------------------
Github user LosD commented on a diff in the pull request:
https://github.com/apache/metamodel/pull/107#discussion_r65771496
--- Diff: jdbc/src/main/java/org/apache/metamodel/jdbc/JdbcDataSet.java ---
@@ -197,8 +197,12 @@ public void close() {
if (_closed) {
return;
}
+
+ FileHelper.safeClose(_resultSet);
+
if (_jdbcDataContext != null) {
- _jdbcDataContext.close(_connection, _resultSet, _statement);
+ FileHelper.safeClose(_statement);
--- End diff --
No, we can't close the statement unconditionally, it makes almost every
single test fail. I'm not entirely sure why (didn't really pursue it, honestly).
> JDBC CompiledQuery ResultSets not closed properly
> -------------------------------------------------
>
> Key: METAMODEL-1093
> URL: https://issues.apache.org/jira/browse/METAMODEL-1093
> Project: Apache MetaModel
> Issue Type: Bug
> Reporter: Dennis Du Krøger
> Assignee: Dennis Du Krøger
>
> When running compiled queries, the JdbcResultSet only returns the lease on
> the query, it doesn't close the result set.
> For most JDBC drivers, that isn't really an issue, as they can easily handle
> connections with multiple open result sets. However, the MySQL JDBC sometimes
> uses streaming result sets (to save memory), of which only one can be open at
> a time.
> I almost have a PR ready, but I'm having a hard time making any kind of test
> that makes sense: When I make a simple test, the driver seems to close the
> resultset automatically right after execution, but that is not what we see in
> DataCleaner, and I'm not really sure why.
> I think I'll post it anyway, as it is rather large (I also tried to avoid
> repeats by making sure that the JdbcDataContext only handles the connection,
> and the rest is handled by try-with-resources as far as possible), so an
> early review would be nice.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
