[rabbitmq-discuss] Different Stomp Issue: rabbitmq-stomp and Codehaus PHP Stomp class

Tony Garnock-Jones tonyg at lshift.net
Fri Jun 5 15:41:43 BST 2009

Hi Gavin,

Yes, the ambiguity here is unfortunate.

Gavin M. Roy wrote:
> appends a linefeed after the frame's null terminator,
> which rabbitmq-stomp does not like.  Instead of sending frame endings
> as \0 they are sending them as \0\n.

The code currently (and at both the 1.5.3 and 1.5.5 tags) accepts any
number of leading \r and \n characters before a command -- what exactly
does the STOMP adapter do when it is sent a \0\n sequence as a frame

> In addition they send no-op \0's before senidng a frame, which
> rabbitmq-stomp does not like.

Yes, that's a current problem. I'll fix that in a similar way to the way
excess \n's are (or at least *should be* :-) ) being ignored now.

> However, I do believe that the rabbitmq-stomp connector should be more
> protocol variance tolerant and accept the \0\n frame ends and out of
> frame \0 characters without causing issues.

Yes. I agree. I think it should, however, stay strict in terms of what
it emits (i.e. continue to use plain \0 for its frame terminator when
it's speaking) -- do you think that makes sense?

 [][][] Tony Garnock-Jones     | Mob: +44 (0)7905 974 211
   [][] LShift Ltd             | Tel: +44 (0)20 7729 7060
 []  [] http://www.lshift.net/ | Email: tonyg at lshift.net

More information about the rabbitmq-discuss mailing list