{:compile (fn [... has nothing to do with clojure compiler. This is just "define foo as a map with key :compile"

Reitit middlewares? If so, maybe it's about this? https://cljdoc.org/d/metosin/reitit/0.7.0-alpha7/doc/ring/compiling-middleware

Ah, of course. It's been more than a year since I have done anything with Clojure. Thanks. 🙂

Damn. I'm really ashamed of myself. A really rookie misunderstanding. Shame on you, Kari! 🙂

But I must say that I miss Clojure. Sometimes I ask myself that did I make a good choice switching Clojure to a considerably bigger pay check. 🙂

It's easy to switch between both so you don't really have to choose :)

bb is more limited but more light-weight (faster to start), if you run into a limitation you can still switch to JVM Clojure.

Feel free to join #CLX41ASCS if you have any bb-specific questions

Thanks for the answers @U04V15CAJ! I’m really interested in the announcement of self contained binary for babashka. I’ll try to find the library’s or utils for start with generator and cli and ask in the bb channel if have doubts 😃

I work on this tool called https://github.com/djblue/portal and there is this https://github.com/djblue/portal/issues/68 that could use some work. The code base isn't particularly small, but the issues should mostly be isolated to the https://github.com/djblue/portal/blob/master/src/portal/runtime/browser.cljc#L23-L32 namespace. The underlying problem is allowing users to specify the name of a binary of a chromium based browser to use instead of the default chrome.

Very cool! I've seen several Clojure talks about Portal. I'll check it out!

try it, that will not work

macros receive unevaluated forms

https://clojurians.slack.com/archives/C053AK3F9/p1677785599197789?thread_ts=1677777234.151169&cid=C053AK3F9

Hmmm, yeah, I’m doing maintenance work on a legacy codebase and I do wish that first rule you mention had been followed more closely – that it depended on fewer custom macros.

I guess something like

(defmacro my-macro [v]
  `(let [v# ~v]
     (if-not (var? v#)
       :not-a-var
       ...)))

That’s an interesting approach, but the assert is executed immediately, not in a ns-quoted block I think the idea is to make ns loading fail the instant it tries to expand a macro with bad args

I guess then your macro must call eval on it's arguments, and check if the result is a var or not.

I can rant at you a bit

if that would help

but the basics is the the places in the code where you "make a choice" you don't do mutation or side effects

the places where you actually affect the world or need to gather information from the world, you go wild

since most programs are "talk to the world" -> "think" -> "talk to the world", that is a "shell" around "thinking"

(let [stuff (make-http-request client)]
  (when (should-do-thing? stuff)
    (save-to-database conn (shape stuff)))

so the goal is to put as much of your program in should-do-thing? and shape as you can

Do you have code examples on a forge somewhere that I could take a look at?