Why Huckleberry?
This is a very, very interesting project. May I ask the reason you're building it, and planned usage?
Edit: Ah, I just found your other project calvin, so I'm guessing it's for that. Cool! I found your project while searching and thinking about a pure cljs build tool... something BADLY needed and awesome to see someone working on one.
Yeap Calvin is one of the uses, but it goes beyond that, the clasapaptj in the jvm is pretty Static and hard to work with, once you start a jvm instance with is painful to work with it, but that's not the case with node and I think that opens opportunities
Given that we are kind of talking here, first of all good job, this and calvin could be the replacement of lein for node!
I haven't checked the code so I was wondering if what hackleberry does is to use the Nexus Rest API. This could be the best thing to do for consuming maven/lein/boot artifacts.
I also found this project which can be relevant.
interesting I didn't know about the nexus rest api. No HB doesn't, it basically searches for maven artifacts in two places: clojars and maven-central, and resolves dependencies manually from there. I tried to follow the work done in https://github.com/cemerick/pomegranate and https://github.com/sonatype/sonatype-aether
Well yes that's even better 😀 I wonder whether there is a way to be sure that what you have is the correct algorithm.
In particular we could create tests that cross check with pomegranade, embedding the logic in cljc files, data in data out.
Not that I don't trust you 😀😀😀
I can tell you I haven't ported everything, I haven't though about crosscheck, but that's an interesting idea.