[rabbitmq-discuss] consuming problems

Matthias Radestock matthias at lshift.net
Fri Jun 5 18:26:00 BST 2009


Jack Moffitt wrote:
>> The class_id refers to the AMQP content class, of which there is only one
>> defined in the current spec. It's called 'basic' and the id of it is indeed
>> 60.
> May I suggest that you put that as the default in the erlang record
> then?  Or at mention it in the comments in rabbit_framing.hrl?  I
> couldn't even find information about this in the specification.  It
> seems like unless I was an active participant in the AMQP standards
> process, I'd have no way to know what this was.

If you use lib_amqp then you are shielded from that particular aspect.

>> You'll probably want to use amqp_channel:subscribe/3 instead.
> So basically the only documentation that exists:
> http://hopper.squarespace.com/blog/2008/1/12/introducing-the-erlang-amqp-client.html
> is wrong.

That page is nearly 18 months old. A few things have changed in the API
since that post.

>> Also, have you looked at lib_amqp.erl? It contains a number of convenient
>> wrappers around the core API.
> How would I have found out about this?  I'll go investigate it now
> that you've mentioned it.
> I think this would have gone a lot smoother had there been one or two
> examples in the README that had some comments and that got maintained
> along with the code.  It's not very fun to not have documentation, but
> its even worse when the documentation you can find is incorrect.

There are a whole bunch of tests that ship with the code which can serve
as a starting point for further exploration.

Having said that, the Erlang client is experimental, and clearly marked
as such. One reason for its experimental status is the sparseness of

We welcome patches to improve the code, comments, examples, docs, etc.
Also, perhaps you could write a blog post that substitutes the
out-of-date one above.



More information about the rabbitmq-discuss mailing list