[zeromq-dev] ZRE spec versus Zyre implementation on protocol header and message type

Pieter Hintjens ph at imatix.com
Wed Dec 18 13:07:39 CET 2013


Hi Ken,

Thanks for being the first person to spot this error. When I added the
sequence number I did it wrongly. I've looked at changing the code but
that's quite ugly. So I've changed the spec now.

-Pieter

On Tue, Dec 17, 2013 at 11:10 PM, Ken Nelson <ken.nelson at us.ibm.com> wrote:
>
> While reviewing Zyre, the open source implementation of ZRE, and the ZRE
> specification, I found something puzzling.
>
> Using the latest Zyre source -  https://github.com/zeromq/zyre and download
> the latest source ("Download ZIP" button on right) - view the file
> src/zre_msg.c and look at the function zre_msg_recv(). This is the function
> that receives and parses the ZRE protocol frames.  The order of reading
> fields from the header is: 1) signature - line 246,  2) id - line 258, and
> 3) sequence on the first line of all the valid case alternatives.  Note that
> "id" is what I would call the ZRE message type, e.g. Hello, Whisper, Shout,
> et al, although the ZRE spec is silent about what this number represents.
>
> In contrast, review the ZRE spec page http://rfc.zeromq.org/spec:20 and look
> at the order of fields there.  It specifies: 1) signature, 2) sequence, and
> 3) id.
>
> A search of the Open and Closed issues on the GitHub page revealed nothing
> relevant, and neither did a search of the ZeroMQ discussion archives.
>
> Am I interpreting something wrong?  If so, what did I miss?  If not, which
> one is correct?
>
>
> Ken
>
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
>



More information about the zeromq-dev mailing list