testReservedExchangeRedeclaredSameType

[Gordon Sim <gs...@redhat.com>]

A couple of new tests were added in http://svn.apache.org/r1686284. One 
of these, testReservedExchangeRedeclaredSameType, verifies that 
amq.direct can be redeclared using the correct type with passive set to 
false as well as to true.

I believe testing for declare of any standard exchange with 
passive=false is incorrect. The spec states:

     Exchange names starting with "amq." are reserved for pre-declared
     and standardized exchanges. The client MUST NOT attempt to create
     an exchange starting with "amq.".

A declare with passive=false is a request to create as far as the client 
is concerned. The broker should in my view raise an error in that case 
even if the exchange exists, since the client is doing something 
forbidden by the spec. Certainly testing that the broker allows it seems 
wrong.

If there is no disagreement, I will remove the line from the test. It 
will still test that you can passively declare the exchange, as that is 
completely legitimate.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org

Re: testReservedExchangeRedeclaredSameType

[Keith W <ke...@gmail.com>]

On 22 June 2015 at 19:16, Gordon Sim <gs...@redhat.com> wrote:

> A couple of new tests were added in http://svn.apache.org/r1686284. One
> of these, testReservedExchangeRedeclaredSameType, verifies that amq.direct
> can be redeclared using the correct type with passive set to false as well
> as to true.
>
> I believe testing for declare of any standard exchange with passive=false
> is incorrect. The spec states:
>
>     Exchange names starting with "amq." are reserved for pre-declared
>     and standardized exchanges. The client MUST NOT attempt to create
>     an exchange starting with "amq.".
>
> A declare with passive=false is a request to create as far as the client
> is concerned. The broker should in my view raise an error in that case even
> if the exchange exists, since the client is doing something forbidden by
> the spec. Certainly testing that the broker allows it seems wrong.
>
> If there is no disagreement, I will remove the line from the test. It will
> still test that you can passively declare the exchange, as that is
> completely legitimate.
>
>
Hi Gordon

Commit r1686284 was mine. I'll amend

I agree the test the passive=false portion of the test is erroneous,  I had
overlooked rule "reserved name" in the 0-10 spec, and was being influenced
by looser words "This command creates an exchange if it does not already
exist, and if the exchange exists, verifies that it is of the correct and
expected class. " at the start of exchange.declare chapter.

Kind regards, Keith.




> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
> For additional commands, e-mail: dev-help@qpid.apache.org
>
>