[ 
https://issues.apache.org/jira/browse/FLINK-4053?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15351104#comment-15351104
 ] 

ASF GitHub Bot commented on FLINK-4053:
---------------------------------------

Github user zentol commented on the issue:

    https://github.com/apache/flink/pull/2128
  
    I tihnk this looks good.
    
    One remark for future contributions though: I think some of your tests are 
too specific. 
    These include `openCallDeclaresQueue`,  `throwExceptionIfChannelIsNull` and 
`invokePublishBytesToQueue`.
    
    Effectively you are not testing functionality but putting a lock on every 
single line of the implementation. 
    
    We will now get a test failure (which _should_ mean that the functionality 
is broken) if we change the exception message (or even handle it completely), 
use a different `basicPublish` method (or pass different arguments) or pass 
different `queueDeclare` arguments.


> Return value from Connection should be checked against null
> -----------------------------------------------------------
>
>                 Key: FLINK-4053
>                 URL: https://issues.apache.org/jira/browse/FLINK-4053
>             Project: Flink
>          Issue Type: Bug
>            Reporter: Ted Yu
>            Assignee: Ivan Mushketyk
>            Priority: Minor
>
> In RMQSource.java and RMQSink.java, there is code in the following pattern:
> {code}
>       connection = factory.newConnection();
>       channel = connection.createChannel();
> {code}
> According to 
> https://www.rabbitmq.com/releases/rabbitmq-java-client/current-javadoc/com/rabbitmq/client/Connection.html#createChannel()
>  :
> {code}
> Returns:
> a new channel descriptor, or null if none is available
> {code}
> The return value should be checked against null.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to