Re: [paho-dev] Contributing and coding standards?

[Frank Pagliughi <fpagliughi@xxxxxxxxxxxxxx>]

Hey Daniel,

Sorry, I was just trying to possibly save you some development time.

I don't know the Embedded Library to be any more (or less) performant 
than the Linux library, The Embedded lib was not designed to be a more 
efficient version of the original, but rather to run on different 
systems like RTOS's and bare metal chips that lack an MMU.

It would be cool to do some performance tests of the different Paho 
libraries (and any others), but no one has gotten around to that yet.

Frank

On 12/6/18 8:55 PM, Daniel Santos wrote:
> Hello Frank,
>
> That's not a very acedemic argument for what type of software to use.
> I'm sure the full version will "work fine", but it is most certainly not
> optimial for my particular application.  Just because you *can* do
> something doesn't mean that you *should*.
>
>
> On 12/4/18 10:05 AM, Frank Pagliughi wrote:
> > Hey Daniel,
> >
> > Your board is an utter powerhouse compared to ones on which I have
> > deployed the "full" Paho C library!!!
> >
> > For comparison, just a few years back I successfully deployed the Paho
> > C library on an _ancient_ ARM9TDMI processor running:
> >
> > 200MHz with 32MB flash and 32MB RAM, with 8kB caches!!!
> > It used Linux 2.6 and gcc 4.2.
> >
> > Mosquitto was also running on the board to create an internal
> > message/event bus, and I used Redis for MQTT message persistence and
> > SQLite for longer-term storage. The board was also collecting,
> > aggregating, and transmitting a fair amount of data.
> >
> > The performance of the Paho library was never an issue.
> >
> > So you should give it a try before attempting to transform the
> > Embedded C library into something different than it was meant to be.
> I'll let you make your own decisions about what you should do in your
> projects and I'll likewise make my own.  What was it "meant to be"
> anyway?  And what does that even mean?
>
> Daniel