[
https://issues.apache.org/jira/browse/HADOOP-1821?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jim Kellerman updated HADOOP-1821:
----------------------------------
Status: Patch Available (was: Open)
Apparently there was a job left from a previous build still running which could
explain the sluggishness of lucene.zones. Killed job, resubmitting.
> [hbase] Replace all String.getBytes() with String.getBytes("UTF-8")
> -------------------------------------------------------------------
>
> Key: HADOOP-1821
> URL: https://issues.apache.org/jira/browse/HADOOP-1821
> Project: Hadoop
> Issue Type: Bug
> Components: contrib/hbase
> Affects Versions: 0.15.0
> Reporter: Jim Kellerman
> Assignee: Jim Kellerman
> Fix For: 0.15.0
>
> Attachments: patch.txt
>
>
> We cannot rely on the default encoding being UTF-8 so a naked
> String.getBytes() will return the bytes in whatever the default encoding is
> for the platform on which code is running. If it is subsequently read on
> another machine with a different default encoding, converting the bytes back
> to a string will result in garbage.
> Consequently, we should always specify an encoding for getBytes() and new
> String. UTF-8 is the preferred encoding.
> The places where we use unqualified getBytes are:
> HConstants.DELETE_BYTES, HConstants.COMPLETE_CACHEFLUSH
> hbase.io.MapWritable.main (but this will not be an issue once HADOOP-1760 is
> completed)
> TestHMemcache.addRows
> PerformanceEvaluation.generateValue
> TestGet
> TestHRegion
> TestHBaseCluster
> TestTableMapReduce
> TestScanner2
> TestRegExpRowFilter
> TestRowFilterSet
> org.onelab.test.StringKey
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
