Sorry the tests are TCK.
It looks like the specifics of the tests can only be discussed on a private
channel btw members who have signed the NDA.

It's kind of frustrating as only I have access to the TCK, and I am sure the
situation will improve  when martin gets access to the TCK.

Regards,

Rajith

On 12/4/06, Martin Ritchie <[EMAIL PROTECTED]> wrote:

On 02/12/06, Rajith Attapattu <[EMAIL PROTECTED]> wrote:
> Hi Folks,
>
> Some tests are failing as it throws MessageNotWriteableException when
trying
> to add properties.

I'd love to know more about what tests are failing as we have quite a
lot of testing around this as I just recently added the ability to
throw the MNWExceptions for properties.

> This exception is throw when _readableProperties is set to true in the
> AbstractJMSMessage.
>
> After some close inspection I see the following logic in the
> AbstractJMSMessage constructors
>
>        _readableProperties = (_contentHeaderProperties != null);
>        _readableMessage = data != null;
>
> Why is this done this way ?

We need to determine when the message has arrived so that we can throw
the correct exception. the _readableProperties setting above is wrong.
I realised that later but haven't committed the change it should be
false. However I believe that this should only be the case for the
constructor that takes the ByteBuffer, as this is used to create a new
message. The other constructors are used for creating a message for
delivery.


The _readableMessage though should be as above as this method is only
called with a value for data when the message is being delivered.
Setting it to false should cause our existing Qpid test cases to fail.


> As a temorary messure I just set the above to properties to false and
now it
> results in more tests being passed.

Does it not also result in some tests failing?

> However I would like to understand the reason behind the above logic to
> provide a proper fix.
>
> Comments are appreciated.
>
> Regards,
>
> Rajith
>
>


--
Martin Ritchie

Reply via email to