[
https://issues.apache.org/jira/browse/HBASE-10255?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13860532#comment-13860532
]
stack commented on HBASE-10255:
-------------------------------
-1
The justification is irrational so not to be trusted. The context in hbase is
pb (pb uses pb LIS). In hadoop the quoted use is web and log aggregation.
Irrelevant.
Test numbers look good but I don't trust them given all that has proceeded here
and they are provided out of thin air: nothing on whether with or without hdfs,
on cluster or not.
> Remove dependency on LimitInputStream
> -------------------------------------
>
> Key: HBASE-10255
> URL: https://issues.apache.org/jira/browse/HBASE-10255
> Project: HBase
> Issue Type: Task
> Reporter: Ted Yu
> Assignee: Ted Yu
> Priority: Minor
> Fix For: 0.98.0, 0.99.0
>
> Attachments: 10255-v1.txt, 10255-v2.txt, alternate_lis.txt
>
>
> LimitInputStream has always been a @Beta API and beta apis aren't guaranteed
> to remain stable over such a long period (v12 to v15).
> LimitInputStream was copied from Guava v12
> The recommended replacement is to use ByteStreams#limit(java.io.InputStream,
> long) instead.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
