define "compile constants" and why they have to be in a config.edn?

For example: • wether the app is compiled in debug mode or not • the links the app will point too (different across dev/prod environments) They need to be in a config.edn because both the backend and frontend need to access them, and different environment have different config files.

ok, but for the backend these are not compile constants

they are runtime configuration

arguably for the frontend they are also runtime configuration

whether its compiled in debug mode or not I'd say is a different problem

what backend to talk to I personally configure at runtime not compile time

and usually that is something my backend sends to my frontend, usually directly with the initial HTML

Yeah you’re right, for the backend it’s runtime configuration

I guess I could send it to the frontend, yes

I mean you can also all make these compile time constants if you really wanted

but it’d be more practical having it directly compiled into the js file imo.

for that you just need to build your own solution. since there is nothing provided out of the box for that

Okay

I mean if they have to be in config.edn.

https://shadow-cljs.github.io/docs/UsersGuide.html#closure-defines

https://shadow-cljs.github.io/docs/UsersGuide.html#shadow-env

I guess I could have shadow cljs to read the config.edn file and then generate closure defines according to it

https://shadow-cljs.github.io/docs/UsersGuide.html#_release_specific_vs_development_configuration

are your options otherwise

Thank you!

I'd say its easier and more flexible by other means. for example not having the recompile the entire CLJS code if your backend url changes. or using the same frontend build for prod and testing

IMHO runtime configuration should not be compile contstants

I see.

shadow-cljs is a CLJS development tool. ring is CLJ server side, which I always recommend doing without shadow-cljs

since you won't be using shadow-cljs in production either. it is best to just setup your own CLJ server. the CLJS setup doesn't change

if you really want to use shadow-cljs for this then you need to use the correct http config. https://shadow-cljs.github.io/docs/UsersGuide.html#dev-http

so just :dev-http {8600 {:root "public" :handler when.helpers/handler}} at the top level. not inside the build config, and removing the http related things from there

Oh, now I understand. Thank you very much, Thomas and thank you for your great tool.

I updated the .edn to

;; shadow-cljs configuration

{:source-paths
 ["src"]

 :dependencies [[binaryage/devtools "1.0.6"]
                [reagent "0.8.0-alpha2"]]

 :nrepl        {:port 8777}

 :dev-http {8600
            {:root "public"
             :handler when.helpers/handler}}

 :builds

 {:app {:target :browser
        :output-dir "public/js/compiled"
        :asset-path "/js/compiled"

        :modules
        {:main
         {:entries [when.core]}}

        :devtools
        {:before-load when.core/stop
         :after-load when.core/start
         :preloads     [devtools.preload]}}}}

:shadow.cljs.devtools.server.dev-http/handler-load-ex - {:http-handler when.helpers/handler}
FileNotFoundException Could not locate when/helpers__init.class, when/helpers.clj or when/helpers.cljc on classpath.

...

[:app] Build failure:
The required namespace "when.helpers" is not available, it was required by "when/core.cljs".
"when/helpers.clj" was found on the classpath. Should this be a .cljs file?

And here is my when.helpers file:

as I said. server side ring handlers are running in CLJ thus this needs to be a .clj file not .cljs