On 15/07/2011 19:14, Clebert Suconic wrote:
>> Can you log this in JIRA, preferably with some suggested javadoc text?
>>
>> Do you mean change the spec so that createTopic/createQueue *won't* throw an
>> exception if the destination doesn't exist? What would the method return in
>> that case?
>>
>
> Change the spec so the provider can create the destination on the
> broker side case the destination doesn't exist.
The javadoc says "Throws: JMSException - if the session fails to create a queue due to some internal error."
I interpret this as meaning that the method should throw a JMSException if the createQueue/createTopic can't return a
valid Queue/Topic. I don't think this defines the circumstances when this happens. So I think a provider is entitled to
decide for itself whether it requires the destination to have been created beforehand, or whether it creates it dynamically.
> There is a test on the TCK validating this, If I change the behaviour
> to create the destination that test will fail.
Which test is it? (I can ask the TCK team about it).
GlassFish MQ (my product) auto-creates destinations on demand by default.
Nigel
>
>
> I will log a JIRA.