Message forwarded from Reza Rahman (who is having trouble with the java.net spam filter):
> I agree with you.
On 15/12/2011 18:27, Nigel Deakin wrote:
> On 29/09/2011 15:24, Nigel Deakin wrote:
>> I'd like to raise this as a discussion topic rather than a concrete proposal.
>>
>> I've also logged this in JIRA
>> http://java.net/jira/browse/JMS_SPEC-53
>>
>> My question is: should we change all the JMS interfaces that currently implement a close() method to implement the
>> java.lang.AutoCloseable interface?
>>
>> This would affect Connection, Session, MessageConsumer, MessageProducer, QueueBrowser.
>
> This proposal received support and no objections when I raised it some time ago.
>
> I've been looking at this in a bit more detail.
>
> Making Connection, MessageConsumer, MessageProducer, QueueBrowser extend AutoCloseable seems straightforward.
>
> However there's a problem with Session. This already extends Runnable, so I can't make it extend AutoCloseable. (I can't
> think of an obvious way around this, but if I've missed something obvious please let me know)
>
> This suggests two options:
>
> 1. Leave Session unchanged and make all the other interfaces extend AutoCloseable
>
> 2. Option (1) is confusing and inconsistent to users. It's better to simply make Connection extend AutoCloseable and
> leave the other interfaces unchanged. In practice the only object that actually needs closing in practice is COnnection
> anyway.
>
> Any preferences? I'm inclined towards (2)...
>
> Nigel