hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-3642) add a HadoopIOException that can be thrown in any method that has IOException on its signature
Date Thu, 26 Jun 2008 14:03:45 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-3642?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12608438#action_12608438

Steve Loughran commented on HADOOP-3642:

no, we dont need this for java6+. Before then it makes it easier to create extensions of IOE
that have good constructors without having to replicate the same code to call initCause and
get messages from (non-null) causes if message==null in the constructor.

> add a HadoopIOException that can be thrown in any method that has IOException on its
> ----------------------------------------------------------------------------------------------
>                 Key: HADOOP-3642
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3642
>             Project: Hadoop Core
>          Issue Type: New Feature
>          Components: util
>    Affects Versions: 0.19.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>            Priority: Minor
>         Attachments: hadoop-io-exception.patch, hadoop-io-exception.patch, hadoop-io-exception.patch
> I find myself having to throw IOExceptions a lot, and create new ones -but the classes
signature varies from java1.5 to 1.6, and the base IOException is fairly meaningless. If Hadoop
added a HadoopIOException, it could be thrown whenever hadoop's own code needed to create
new IOExceptions, and possibly be differentiated in the catch() logic. 
> The biggest disadvantage of doing this is that as IOException is built into the JVM,
you can be sure that the far end will be able to deserialize it under RMI, without having
the rest of hadoop on the classpath. This is not a feature of hadoop, so should not be an
issue. For those of us who do use RMI, well, we'd better get our classpaths right.

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

View raw message