← Prev in month ← Prev in thread
Next in thread → Next in month →

Re: [mqtt-comment] Several comments/questions

From
Richard J Coppen <>
Date
2014-04-10T11:08:30+00:00
ID
Thread
Re: [mqtt-comment] Several comments/questions
Hi Martin, 

Many thanks for your comments and feedback
on CSD1. 

The MQTT TC has now addressed these
under Jira issue 193 > https://tools.oasis-open.org/issues/browse/MQTT-193
<

Best regards

Richard

From:      
 Martin Hurton <>

To:      
 ,

Date:      
 11/02/2014 20:16

Subject:    
   [mqtt-comment]
Several comments/questions

Sent by:    
   <>

1 - [MQTT-2.0.0-1] "Unless stated otherwise, if
either the Server or Client receives a Control Packet which does not meet
this specification, it MUST close the Network Connection"

Should the server publish Will message in this case?

Also,what confuses me is the fact that for some cases
the spec explicitly says that the server should close the connection (invalid
flags in packet header, non-zero Reserved bits in payload of SUBSCRIBE
Control packet) but not for others (zero packet identifier, invalid Topic
Filter in SUBSCRIBE or UNSUBSCRIBE Control Packet, invalid Topic Name in
PUBLISH Control Packet, non-zero length payload for DISCONNECT Control
Packet, etc.).

2 - [MQTT-3.8.3-1] The Payload of a SUBSCRIBE packet
MUST contain at least one Topic Filter / QoS pair. A SUBSCRIBE packet with
no payload is a protocol violation.

There is no similar requirement for UNSUBSCRIBE Control
packet. Is this by design? What is the rationale for this requirement/assymetry?

3 - [MQTT-3.1.3-4] The ClientId MUST comprise only
Unicode [Unicode63] characters, and the length of the UTF-8 encoding MUST
be at least zero bytes and no more than 65535 bytes.

It sound a bit strange to require that the string must
be at least zero bytes. Every UTF-8 encoding meets this requirement.

4 - What is meant by transient error and what is
the difference among transient error (4.8) protocol error
(section 3.1.2.5) and protocol violation (3.1)?

5 - In section 3.8.3.1, Description in the table should
read LSB for byte 2 and byte 8, rather than MSB.

6 - In section 3.10.2.2, Description in the table should
read LSB for byte 2 and byte 7, rather than MSB.

With Regards,

Martin Hurton

Unless stated otherwise above:

IBM United Kingdom Limited - Registered in England and Wales with number
741598. 

Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU
← Prev in month ← Prev in thread
Next in thread → Next in month →