[ https://issues.apache.org/jira/browse/DERBY-4443?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13014032#comment-13014032 ]
Houx Zhang commented on DERBY-4443: ----------------------------------- Thanks for your checking, Knut. This problem has perplexed me for some days. After checking again, I accept that private static void rollBackAndThrowSQLException(Connection conn, SQLException se) throws SQLException { try { conn.rollback(); } catch (SQLException e) { se.setNextException(e); } throw se; } is better, we will get the full SQLException chain. However, it's very strange that when debugging in this situation, the chain will be broken down before thrown to EmbedPreparedStatement.executeUpdate() in Line 309 of EmbedPreparedStatement, and the rollback exception will be lost. Is this situation normal, or is there something special for SQLException launched by rollback(), please? > Wrap rollback in exception handlers in try-catch > ------------------------------------------------ > > Key: DERBY-4443 > URL: https://issues.apache.org/jira/browse/DERBY-4443 > Project: Derby > Issue Type: Bug > Components: Demos/Scripts, Documentation, Eclipse Plug-in, JDBC, > Network Client, Network Server, Replication, Services, SQL, Test, Tools > Affects Versions: 10.5.3.0 > Reporter: Aaron Digulla > Assignee: Houx Zhang > Labels: derby_triage10_8 > Attachments: DERBY-4443-1.patch, DERBY-4443-2.patch, > DERBY-4443-3.patch, DERBY-4443-4.patch, DERBY-4443-4.png, DERBY-4443-5.patch, > DERBY-4443.patch > > > Avoid this pattern everywhere: > }catch(SQLException se){ > //issue a rollback on any errors > conn.rollback(); > throw se; > } > because an error in rollback will shadow the original exception. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira