[
https://issues.apache.org/jira/browse/DERBY-2493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12484857
]
Knut Anders Hatlen commented on DERBY-2493:
-------------------------------------------
Replacing the Hashtable with a HashMap was trivial, but because of differences
between the implementations of Hashtable and HashMap, the rows come out in
different order for some queries (in particular SELECT DISTINCT). And since the
implementation of HashMap changed between JDK5 and JDK6, the order varies
between JVMs as well. Three tests will be affected by this change:
JUnit:
LangScripts.aggregate
ResultSetsFromPreparedStatementTest.testDistinctScanResultSet
derbyall:
lang/distinct.sql (currently being converted to JUnit, see DERBY-2491)
I don't think the LangScripts tests can handle different results in different
JVMs, so that one will probably have to be converted to pure JUnit in order to
work with this change. In ResultSetsFromPreparedStatementTest it should be
possible to use an assert method which didn't require a specific order.
> Use unsynchronized collections in BackingStoreHashtable
> -------------------------------------------------------
>
> Key: DERBY-2493
> URL: https://issues.apache.org/jira/browse/DERBY-2493
> Project: Derby
> Issue Type: Improvement
> Components: Performance, Store
> Affects Versions: 10.3.0.0
> Reporter: Knut Anders Hatlen
> Assigned To: Knut Anders Hatlen
> Priority: Minor
>
> BackingStoreHashtable uses a Vector and a Hashtable, but doesn't need the
> synchronization provided by these classes (I think). Replacing them with
> ArrayList and HashMap could improve performance for some kinds of operations.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
