Fork me on GitHub
#onyx
<
2017-01-03
>
len14:01:55

Hi all, new to onyx, looking very good so far

len14:01:45

Do all workflows need an output ?

Travis14:01:31

Yes I believe so, if you don't need to write to anything you can basically make a noop leaf function

len14:01:23

ok thanks will try that - side effect only 🙂

jasonbell15:01:11

@len if I’m not looking to store (to S3 for example) or pass through to a Kafka queue then I channel all outputs to a core.async :out channel.

len15:01:01

can I use core,aync in production ?

len15:01:13

i understand for an embedded job, I am deploying to a cluster feeding off the datomic txn log

jasonbell15:01:01

@len most of my jobs end up going to Kafka or storing the output so I don’t feel to be the best to answer, part of me doesn’t see any huge reason not to

len15:01:40

agreed @jasonbell thanks - am rethinking what I am doing

Travis15:01:15

can just use clojure identity function

Travis15:01:23

if you don’t need to do anything in the out

jasonbell15:01:37

that’s the way I see it too

jasonbell15:01:01

I always assume to be an :in and :out channel of some form.

Travis15:01:33

thats what we do since we have a batch function writer that writes to a DB and use the identity function as a leaf to satisfy the workflow

len15:01:18

oh thats perfect thanks @colinhicks

len15:01:10

its considered and good

michaeldrogalis15:01:54

@jasonbell Send a PR in with your name appended. I usually update it once every few weeks but sometimes I forget.

len15:01:14

leaft-tasks is the word I was missing thanks @michaeldrogalis

jasonbell15:01:58

@michaeldrogalis okay, thanks, most kind.