[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
[paho-dev] Paho News - including 1.1 Release Approval
|
1) Release 1.1
(https://projects.eclipse.org/projects/technology.paho/releases/1.1.0)
has been approved
(https://bugs.eclipse.org/bugs/show_bug.cgi?id=440385). Yay! Now to
get it out.
2) I got travel approval to go to EclipseCon in San Francisco in March
to present on MQTT-SN. And I registered.
3) Ian Skerrett and I have been organizing a 2nd MQTT Interops Test Day
at EclipseCon. The number of participants signed up is borderline as to
whether we should go ahead. It's almost there, but if anyone else wants
to participate, please sign up here:
http://www.eventbrite.com/e/mqtt-interop-testing-day-tickets-13854369793?aff=es2&rank=0
I would like it to go ahead, and we could include some other activities,
such as:
a) discuss future improvements to MQTT, now that the 3.1.1 standard has
been approved (to be fed to the OASIS committee)
b) discuss MQTT-SN, uses and relevance
c) get feedback on the test material
or whatever the participants would feel is useful.
4) Andy Banks submitted a proposal to the MQTT OASIS TC to get a test
suite adopted for MQTT conformance. We hope that the Paho test material
will fit the bill, eventually.
5) I stumbled across this article in Wired:
http://www.wired.co.uk/news/archive/2013-04/12/super-mario-solved. Take
a look at the video, it's really cool :-).
Reading the paper made me realize how much harder I was making the MQTT
test generation by not being able to find best paths using backtracking
(you need an "undo last operation"). This is usually impossible if you
are using a real product to generate tests against (note to self: add
this feature to any product I write). But I'm using a Python broker
which I control. Using real sockets causes the Python deepcopy to fail
so I can't save/restore the broker state, so I'm just looking into
replacing the real sockets with a simple buffer, just for the test
generation. This should make the generation of high coverage test
suites much more efficient. If only I could make a cool video too...
I also fixed some other things in the test material, a couple thanks to
Dejan Bosanac (RedHat) who raised some really useful bugs.
6) I've been thinking about how to get more/better ports of the embedded
clients to the most popular embedded operating systems
(http://m.eet.com/media/1179248/20130227pcembeddedsurvey841.jpg). I've
put some feelers out for VxWorks. Started looking at FreeRTOS again.
I'd like to get ports for Contiki and RIOT too, which aren't on that
list. Apart from having ready-made ports available, I want to see if
there any obstacles that prevent them from being ported to any OS, and
to fix them if so. But I haven't got much further than that right now.
--
Ian Craggs
icraggs@xxxxxxxxxx IBM United Kingdom
Paho Project Lead; Committer on Mosquitto
