please do share with us.

I have a sample here I played with a bit: https://github.com/campspot/time-tracker

Dropwizard, kotlin and JDBI

ignore the client side stuff, doesn’t work. But the server is a decent template

thank you.

back to what you were talking about before, my startup is using spring 👍

+ kotlin

@Pere Casafont cool, what kotlin stack do you use for your startup?

I am still moving to kotlin

core logic migration is complete, but most rest controllers are still java

if you ask about spring stack I use webmvc and hibernate+jpa

I haven't decided what DB layer I'll use, I'm leaning toward JOOQ

FWIW, bigger companies do tend to use Java because there are better monitoring tools...was at an AI meetup at Spotify and they mentioned the backend is all Java, so they had to do JNI to Tensorflow because Python monitoring was not well supported in their infrastructure...and JVM code will be faster. Lots of Scala too...very little Kotlin.

hibernate jpa supports hql methods

in the long term, blindly using hibernate generates technical debt, but it's really handy to prototype your project

@cedric that seems insane - how many jars is it downloading?

As many as DropWizard requires, I’d say medium size

I tried to limit that memory to 512m but it’s not picked up so far

not even sure Gradle can do that

@cedric try packaging the ZIP on travis and deploying that way?

yes was thinking about that, but what a hassle…

or maybe just deploy the jar file but build locally and not on heroku

if you use http://start.http4k.org you can get an entire CD pipeline to heroku in one line of bash. then switch out the dependencies and you're done

you just need a couple of api tokens

I always use the "application" plugin and create the zip/tar locally

@fred.deschenes Yeah going to explore this

+1 for the app plugin. it'd also be interesting to know what the size of the end zip is.

The fat jar is 17megs with a minimal app

running with Jetty backend?