Mailing-List: contact hdfs-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hdfs-issues@hadoop.apache.org
Date: Thu, 18 Sep 2014 22:20:34 +0000 (UTC)
From: "Eric Payne (JIRA)" <jira@apache.org>
To: hdfs-issues@hadoop.apache.org
Message-ID: <JIRA.12740833.1410459123000.70415.1411078834220@Atlassian.JIRA>
In-Reply-To: <JIRA.12740833.1410459123000@Atlassian.JIRA>
References: <JIRA.12740833.1410459123000@Atlassian.JIRA>
 <JIRA.12740833.1410459123856@arcas>
Subject: [jira] [Commented] (HDFS-7049)
 TestByteRangeInputStream.testPropagatedClose fails and throw NPE on
 branch-2
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HDFS-7049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14139613#comment-14139613 ] 

Eric Payne commented on HDFS-7049:
----------------------------------

Hi [~jyu@cloudera.com]. Thanks for merging this fix and creating the patch.

The patch doesn't apply because it has the 'a/' and 'b/' at the beginning of the filepaths. I downloaded the patch, removed those strings from the filepaths, and I was able to apply the patch cleanly to branch-2. The test also passes cleanly with no NPE.

Once you make that change to the patch, it looks good to me.

> TestByteRangeInputStream.testPropagatedClose fails and throw NPE on branch-2
> ----------------------------------------------------------------------------
>
>                 Key: HDFS-7049
>                 URL: https://issues.apache.org/jira/browse/HDFS-7049
>             Project: Hadoop HDFS
>          Issue Type: Bug
>            Reporter: Juan Yu
>            Assignee: Juan Yu
>            Priority: Minor
>         Attachments: HDFS-7049-branch-2.patch
>
>
> On branch-2, TestByteRangeInputStream.testPropagatedClose throw NPE when HftpFileSystem$RangeHeaderUrlOpener.connect
> This is due to fix of HDFS-6143 "WebHdfsFileSystem open should throw FileNotFoundException for non-existing paths"
> public ByteRangeInputStream(URLOpener o, URLOpener r) throws IOException {
>     this.originalURL = o;
>     this.resolvedURL = r;
>     getInputStream();
>   }
> the getInputStream() will be called in constructor now to verify if file exists.
> Since we just try to test if ByteRangeInputStream#close is called at proper time, we could mock(ByteRangeInputStream.class, CALLS_REAL_METHODS) for testing to avoid the NPE issue.
> I believe the trunk version already does this, we just need to merge the test from trunk.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)