[zeromq-dev] czmqjni status?

[James Gatannah]

I think I got it.

Actually, I'm not sure why this approach didn't work the last time I
tried it. I'm pretty sure this is exactly the same thing I was doing
when I reached out for help. Maybe it was something weird with clojure
and boxing?

In case it might be useful, I slapped together a silly
request/response example at

https://github.com/jimrthy/cljeromq/tree/test/czmq/src/jython

I'm torn about whether this might be a valid candidate for a position
in the "HOWTO use" spot in the JNI bindings doc.

It's probably better than nothing, and python syntax really should be
easy enough for any programmer to understand. But I've run across a
lot of java people who just refuse to look at it.

If it seems like something worth tossing into the README, I'll be
happy to add a PR. If not, I'll try to figure out how to publish a
gist.

Thanks again!

On Wed, Apr 12, 2017 at 9:32 PM, James Gatannah
<james.gatannah at gmail.com> wrote:
>> Date: Tue, 11 Apr 2017 12:15:15 +0200
>> From: Kevin Sapper <kevinsapper88 at gmail.com>
>>
>> Hi James,
>>
>> the czmqjni wrapper is a very thin layer above the czmq API. So any
>> example in the
>> czmq README should work.
>
> That's what I expected, but I think I'm missing a translation step
> between the java object reference to a void* that the Zframe/message
> calls could use as a sockish.
>
>
>> Though some APIs might not be available in the
>> wrapper
>> due to technical reasons. If you're not able to convert the C examples
>> in the README,
>> let me know then I'll try to create one later this week.
>
> I haven't had any luck yet, so I appreciate the offer. I'll let you
> know if I manage to figure anything out before you put anything
> together.
>
>
>> While the czmq API is stable the jni wrapper's API is not. So we are
>> happy to receive
>> improvements.
>
> Cool. I've started to take a look at the GSL/Zproject docs. Time will
> tell whether I'll be able to find any bandwidth to actually
> contribute.
>
>
>> APIwhise jzmq and jeromq are the same. Also those APIs a more javalike
>> and both
>> libraries have been more tested than czmqjni. But they don't support
>> encryption,
>> if I'm not  mistaken,
>
> Last I heard/tried, it seemed like jzmq had that covered. And, with
> TweetNacl in java, I think it wouldn't be too hard to add to jeromq.
>
>> which would be your reason to use czmqjni instead.
>
> In this case, it's a lighter political/marketing lift. I already
> foisted the python czmq bindings as the most effective way for us to
> move our current tech stack (which is python up front and java at the
> bottom) from a slow brokered networking layer to using CLISRV.
>
> Since we're already building libzmq and czmq, adding this binding is
> less daunting than also adding something like jzmq. I know, it
> shouldn't be, but there you have it.
>
> Thank you,
> James
>
>
>> //Kevin