[
https://issues.apache.org/jira/browse/HBASE-8430?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13641301#comment-13641301
]
Sergey Shelukhin commented on HBASE-8430:
-----------------------------------------
[~stack] [~devaraj] what do you guys think? In a test of new WAL with
CellScanner, when calling advance(), Hadoop stream inside HBase threw
IOException (EOF, for a EOF test). None of the handling logic was executed
because it is wrapped in BaseDecoder (and cannot be not wrapped as advance()
doesn't throw), so the exception happily went all the way to the top level.
> Cell decoder/scanner/etc. should not hide exceptions
> ----------------------------------------------------
>
> Key: HBASE-8430
> URL: https://issues.apache.org/jira/browse/HBASE-8430
> Project: HBase
> Issue Type: Bug
> Components: IPC/RPC, Protobufs
> Reporter: Sergey Shelukhin
>
> Cell scanner, base decoder, etc., hide IOException inside runtime exception.
> This can lead to unexpected behavior because a lot of code only expects
> IOException. There's no logical justification behind this hiding so it should
> be removed before it's too late (the sooner we do it the less throws
> declarations need to be added)
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
