Fork me on GitHub
#yada
<
2017-01-02
>
malcolmsparks12:01:10

please RT to get the word out

dominicm14:01:04

RE: yada/lean not having SSE support. Is that just because core.async is being kept out & the current SSE api revolves around it?

dominicm14:01:04

So I could still return an async response of some kind, using something manifold-compatible and include the data: prefix (or use yada/async)

malcolmsparks15:01:09

@dominicm - yes, it's to avoid the clojure.core.async dependency. The idea is that lean shows you how to start from a bare-bones yada and build it up (lean actually brings in bidi and aleph).

malcolmsparks15:01:24

yada/core only depends on a small set of dependencies (manifold, schema, byte-streams, clj-time, ring-core, hiccup)

malcolmsparks15:01:25

It's really for advanced users, and specifically requested by @jeroenvandijk but others have also complained about the number of dependencies involved in the full 'batteries-included' yada - this provides another option

malcolmsparks15:01:54

yada beginners and existing users can just continue as before and use [yada "1.2.0"] - there should be no breakages for those who have only depended on the public API (in yada.yada).

malcolmsparks15:01:28

I'll be more deliberate in future in steering people to use yada.yada rather than requiring functions from other yada namespaces - that said, the upgrade path for the majority of existing users should be seamless

malcolmsparks15:01:20

I'm trying really hard nowadays not to break compatibility, there's a lot of features planned for 2017 and I'll try to bring all these in without introducing upgrade hurdles for existing users

malcolmsparks15:01:41

The tricky one will be clojure.spec - this recent refactoring has one eye on that transition...

dominicm15:01:30

I was just contemplating how a clojure.spec integration could happen for query params more easily now.

lmergen16:01:53

I would really like spec integration

lmergen16:01:25

then again, it's not necessarily that complex to do right now

lmergen16:01:22

what I would really like is if yada detected when a handler's fn is a core.spec/fdef, and return 400 when the http request parameters do not match what the spec'ed function expects

lmergen16:01:51

I don't think that would be all to difficult to achieve

dominicm21:01:17

@lmergen would that mean you have to destructure all params, cookies, etc. in the function signature?