[
https://issues.apache.org/jira/browse/CASSANDRA-6253?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13809082#comment-13809082
]
Aleksey Yeschenko commented on CASSANDRA-6253:
----------------------------------------------
bq. Also note that while it would make sense to rename CLUSTERING_KEY to
CLUSTERING_COLUMN, we serialize the kind using the enum string in the system
tables so it's worth breaking clients.
I think we really should rename it to CLUSTERING_COLUMN. Just special-case the
(de)serialization for system.schema_columns, since breaking clients just for
this would really be a bad thing.
Also, not a fan of "cfm.comparator instanceof CompositeType" all over the
place. Maybe add CFMetaData.hasCompositeComparator() instead?
Nitty nit: IndexExpression doesn't need the added
'org.apache.cassandra.cql3.ColumnIdentifier' import.
Other than that, LGTM.
> Refactor: remove CFDefinition
> -----------------------------
>
> Key: CASSANDRA-6253
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6253
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Fix For: 2.1
>
> Attachments: 0001-Remove-CFDefinition.txt, 0002-Unit-tests-updates.txt
>
>
> While writing CQL3 initially, we added CFDefinition to be a CQL3-friendly
> version of CFMetaData. However, a number of changes have it redundant. The
> infos of CFDefinition are now in CFMetaData in pretty much the same form and
> in particular CFDefinition.Name is now basically a copy of ColumnDefinition.
> This is not only useless, but it is confusing too: sometimes the CQL3 code
> uses ColumnDefinition because it has too, but more often it uses
> CFDefinition.Name and both use diferent vocabulary even though they are the
> exact same thing.
> So this ticket propose to remove CFDefinition and CFDefinition.Name since
> they are not useful anymore and this clean things up.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
