cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pavel Yaskevich (JIRA)" <j...@apache.org>
Subject [jira] [Assigned] (CASSANDRA-3119) Cli syntax for creating keyspace is inconsistent in 1.0
Date Thu, 01 Sep 2011 14:23:10 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-3119?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Pavel Yaskevich reassigned CASSANDRA-3119:
------------------------------------------

    Assignee: T Jake Luciani  (was: Pavel Yaskevich)

it seems like the best move here will be to support both [{}] and {} in there because it's
imposible to say what caused that "null" because it fails somewhere deep in ANTLR (in most
cases when recognition was failed ANTLR throws standard exceptions which we handle correctly
as "Command not found" or "Syntax error").

> Cli syntax for creating keyspace is inconsistent in 1.0
> -------------------------------------------------------
>
>                 Key: CASSANDRA-3119
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3119
>             Project: Cassandra
>          Issue Type: Bug
>    Affects Versions: 1.0
>            Reporter: Sylvain Lebresne
>            Assignee: T Jake Luciani
>            Priority: Minor
>              Labels: cli
>             Fix For: 1.0
>
>
> In 0.8, to create a keyspace you could do:
> {noformat}
> create keyspace test with placement_strategy = 'org.apache.cassandra.locator.SimpleStrategy'
and strategy_options = [{replication_factor:3}]
> {noformat}
> In current trunk, if you try that, you get back "null". Turns out this is because the
syntax for strategy_options has changed and you should not use the brackets, i.e:
> {noformat}
> strategy_options = {replication_factor:3}
> {noformat}
> (and note that reversely, this syntax doesn't work in 0.8).
> I'm not sure what motivated that change but this is very user unfriendly. The help does
correctly mention the new syntax, but it is the kind of changes that takes you 5 minutes to
notice. It will also break people scripts for no good reason that I can see.
> We should either:
> # revert to the old syntax
> # support both the new and old syntax
> # at least print a meaningful error message when the old syntax is used
> Imho, the last solution is by far the worst solution.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message