[ https://issues.apache.org/jira/browse/HADOOP-2663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12560637#action_12560637 ]
Runping Qi commented on HADOOP-2663: ------------------------------------ On the flip side, the Buffer should provide a constructor taking a utf-8 String object as a parameter. > Buffer class' toString method should accept a code name for "true" utf-8 > codeName > ---------------------------------------------------------------------------------- > > Key: HADOOP-2663 > URL: https://issues.apache.org/jira/browse/HADOOP-2663 > Project: Hadoop > Issue Type: Improvement > Components: record > Reporter: Runping Qi > Assignee: Milind Bhandarkar > > Currently, if one call toString("UTF-8"), a String object is created using > Java's converion code. > That does not work properly for some rare but still true utf-8 bytes. Hadoop > has its own utf-8 conversion > code for string serialization/deserialization. The same code should be used > here. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.