[
https://issues.apache.org/jira/browse/HADOOP-16300?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16838006#comment-16838006
]
Ayush Saxena commented on HADOOP-16300:
---------------------------------------
Thanx [~ebugs-in-cloud-systems] for the report.
* May be all the reported issues can be done as part of Single Jira only.
* HADOOP-16295 - makes quite sense to me, If the exception is thrown because of
an interrupt, I guess we can avoid the disk check.
* For the MiniKdc ones, There is no functionality wise advantage, But I agree
Technically the exceptions could be turned better to make more sense as FNF
instead of RuntimeException if Workdir isn't there, Provided it doesn't have
any adverse affects in the present stage.
[[email protected]] Any thoughts on this?
> ebugs automated bug checker is reporting exception issues
> ---------------------------------------------------------
>
> Key: HADOOP-16300
> URL: https://issues.apache.org/jira/browse/HADOOP-16300
> Project: Hadoop Common
> Issue Type: Bug
> Affects Versions: 3.1.2
> Reporter: eBugs
> Priority: Major
>
> "ebugs-in-cloud-systems" is reporting issues related to exception handling
> Some of them may be be correct, some of them appear to have misunderstood
> issues.
> For reference
> * we aren't going to change the signature of public APIs, so new exception
> classes will not be added. anything which suggests that will be WONTFIX.
> * we probably aren't going to change code which throws unchecked exceptions
> to checked ones unless there is a good reason.
> * we can and do tighten the exceptions thrown in failures (e.g. replace an
> IOException with an InterruptedIOException. Patches welcome there, with tests.
> * making sure we don't lose the stack traces of inner causes would be nice
> too.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
