You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@activemq.apache.org by "Gary Tully (JIRA)" <ji...@apache.org> on 2018/01/03 16:16:01 UTC

[jira] [Updated] (AMQ-6880) Incorrect comparison behaviour in TransportConnector.isMatchesCluster with Tokenised Filter

     [ https://issues.apache.org/jira/browse/AMQ-6880?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Gary Tully updated AMQ-6880:
----------------------------
    Description: 
the clusterFilter should support a comma-delimited list of filter expressions.
Consider a use case where we have brokers with prefixes, like east and west and we have specified a broker filter of {code}"east.*,west.*"{code}
We add a new broker to the network, called "west-broker2." Looking at the code, this expression will fail out on the first test (against the {code}"east.*{code} filter token), setting result to false and failing to add the broker.
Similarly, if we add a new broker called "east-broker2," the name will successfully test against the first filter token of {code}"east.*"{code}, then will subsequently fail against the following token of {code}"west.*"{code} instead of breaking out.

When a filter is set, we need to check for a match against all of the tokens in the filter.

  was:
the clusterFilter should support a comma-delimited list of filter expressions.
Consider a use case where we have brokers with prefixes, like east and west and we have specified a broker filter of "east.*,west.*"
We add a new broker to the network, called "west-broker2." Looking at the code, this expression will fail out on the first test (against the "east.* filter token), setting result to false and failing to add the broker.
Similarly, if we add a new broker called "east-broker2," the name will successfully test against the first filter token of "east.*", then will subsequently fail against the following token of "west.*" instead of breaking out.

When a filter is set, we need to check for a match against all of the tokens in the filter.


> Incorrect comparison behaviour in TransportConnector.isMatchesCluster with Tokenised Filter
> -------------------------------------------------------------------------------------------
>
>                 Key: AMQ-6880
>                 URL: https://issues.apache.org/jira/browse/AMQ-6880
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: Broker
>    Affects Versions: 5.15.0
>            Reporter: Gary Tully
>            Assignee: Gary Tully
>             Fix For: 5.16.0
>
>
> the clusterFilter should support a comma-delimited list of filter expressions.
> Consider a use case where we have brokers with prefixes, like east and west and we have specified a broker filter of {code}"east.*,west.*"{code}
> We add a new broker to the network, called "west-broker2." Looking at the code, this expression will fail out on the first test (against the {code}"east.*{code} filter token), setting result to false and failing to add the broker.
> Similarly, if we add a new broker called "east-broker2," the name will successfully test against the first filter token of {code}"east.*"{code}, then will subsequently fail against the following token of {code}"west.*"{code} instead of breaking out.
> When a filter is set, we need to check for a match against all of the tokens in the filter.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)