Hey Roger,
BTW, if you peek under the hood of the C++ wrapper, please let me
know what you think. After being totally swamped for the better part
of a year, I've finally freed up my schedule to get this into a
release worthy-state. I will be concentrating on some MQTT/Paho
stuff the next few weeks.
I have a few updates already. A few months back the C library added
better tracking of the message token/ID from the async library, so
that makes the job of a wrapper much easier. So the next update will
be a little simpler with not so much locking.
I also started working on a wrapper for the "D" language. An
interesting aspect of that one is the ability of the language to
promise immutability of the messages - or at least individually for
the payloads and topic strings. That significantly reduces the
amount of locking and copying that needs to be done, and the number
of race conditions that you need to fret over. If everyone promises
not to change the message, then you don't need to copy it or keep
protecting it for exclusive access,
I'm wondering if it wouldn't be nice to try something like that in
the C++ library. It could be as "simple" as making the shared pointers
const. That might have a ripple effect, but might make the library
simpler and more efficient.
Frank
On 05/19/2015 03:57 AM, Roger Light
wrote:
Hi Frank,
Thanks, that does the job.
Cheers,
Roger
_______________________________________________
paho-dev mailing list
paho-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/paho-dev
|