[jira] [Comment Edited] (TEZ-2407) Drop references to the old DataInputBuffer impl in PipelinedSorter

["Gopal V (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TEZ-2407?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14527144#comment-14527144 ] 

Gopal V edited comment on TEZ-2407 at 5/4/15 7:49 PM:
------------------------------------------------------

No, the issue is that {{DataInputBuffer::getLength()}} has bad semantics - it returns capacity instead of length of data.

We are always forced to do {{DataInputBuffer::getLength() - DataInputBuffer::getPosition()}} to get the accurate value & that's an easy thing to forget.

Since {{DataInputBuffer}} comes from hadoop, we can't change the original - however, we can make our code more readable as it is a simple class to replace & make getLength() meaningful.


was (Author: gopalv):
No, the issue is that {{DataInputBuffer::getLength()}} has bad semantics - it returns capacity instead of length of data.

Since that comes from hadoop, we can't change the original - however, we can make our code more readable as it is a simple class to replace & make getLength() meaningful.

> Drop references to the old DataInputBuffer impl in PipelinedSorter
> ------------------------------------------------------------------
>
>                 Key: TEZ-2407
>                 URL: https://issues.apache.org/jira/browse/TEZ-2407
>             Project: Apache Tez
>          Issue Type: Bug
>            Reporter: Rajesh Balamohan
>




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