[jira] Resolved: (DERBY-4732) Release system resources in CanonTestCase thoroughly

["Kristian Waagan (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/DERBY-4732?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Kristian Waagan resolved DERBY-4732.
------------------------------------

    Issue & fix info:   (was: [Patch Available])
       Fix Version/s: 10.7.0.0
          Resolution: Fixed

Seems like an improvement.
I committed the patch to trunk with revision 963716.

> Release system resources in CanonTestCase thoroughly
> ----------------------------------------------------
>
>                 Key: DERBY-4732
>                 URL: https://issues.apache.org/jira/browse/DERBY-4732
>             Project: Derby
>          Issue Type: Improvement
>          Components: Test
>            Reporter: Yun Lee
>            Assignee: Yun Lee
>            Priority: Minor
>             Fix For: 10.7.0.0
>
>         Attachments: derby-4732.patch, derby-4732.stat
>
>
> Now, in the method of compareCanon(String canon) in CanonTestcase, two BufferedReaders----cannonReader and testOutput are closed in try catch block, however if some exception occurs above, the two BufferReaders won't be closed, and related system resources won't be released. 
> Besides, a releative problem is about the ByteArrayOutputStream rawBytes in CanonTestcase, it's created in getOutputStream() and closed in compareCanon(String canon). It's pointed out by Knut in Derby-2708, ByteArrayOutputStream's javadoc says that closing it has no effect, so tearDown() method has been added  to CanonTestCase and sets rawBytes to null. It certainly can work. However, is it also OK to null out rawBytes in compareCanon(String canon)? If so, it seems cleaner than to null it out in tearDown() method.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.