[ http://issues.apache.org/jira/browse/HADOOP-414?page=comments#action_12432680 ] Doug Cutting commented on HADOOP-414: -------------------------------------
We seem to be talking at cross purposes. Are we still talking about the patch attached to this bug? Of course, if one changes a Writable from using UTF8 to Text, and makes no other changes, then one will have problems reading old data and/or with protocol compatibility. Folks who wish to be able to do things like this in user data should use VersionedWritable. But the question I was raising was about migrating internal uses of UTF8 to Text. For most of those, incrementing protocol version numbers and container format version numbers will suffice. The only exception I know of is the one I discussed above, namely the names of parameter classes used in RPC. That's the issue I was trying to raise here, not the general issue of version compatiblity for Writable. Users who have data files that use UTF8 may wish to upgrade their data to use Text. That's outside the scope of this bug, I think. Here we should be concerned with replacing internal uses. > replace class UTF8 with class Text > ---------------------------------- > > Key: HADOOP-414 > URL: http://issues.apache.org/jira/browse/HADOOP-414 > Project: Hadoop > Issue Type: Improvement > Affects Versions: 0.5.0 > Reporter: Hairong Kuang > Assigned To: Hairong Kuang > Fix For: 0.6.0 > > Attachments: utf2text.patch > > > Since class UTF8 is deprecated, all references of UTF8 in hadoop should be > replaced with class Text if the change does not break the system. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira
