Isn't it clear? It basically says that while for other values with-meta has no effect on equality comparison, for functions it does:

(= (with-meta [10 20] {:some :meta}) [10 20]) => true
(defn foo [] "hi")
(= foo foo) => true
(= (with-meta foo {:some-more :meta}) foo) => false

(= foo foo) is to compare vars not functions, isn’t it?

I was confused by the sentence fragment in the test.

So a function is equal to itself, but a function with metadata attached is not equal to the original function.

Functions compare with identity and once you add metadata, they are no longer identical

are any other objects which are compared for identity also capable of attaching to meta data?

can't think of any, this is kind of a special case

I'm mostly interested in status and headers.

I'm talking specifically about [day8.re-frame/http-fx "0.2.1"] which uses cljs-ajax under the hood

did you try response-format :raw ?

Yeah, but I don't think it will change anything, sincie only response body is returned. So in effect, a string is returned instead of JSON

perhaps you could write your own response format then

but yeah, it seems somewhat tedious

Anyway, thanks for help. I'm curious how others tackle this issue in re-frame/clojurescript though.

No idea; but maybe someone in #clojurescript and/or #re-frame would know something

You actually have a directory named foo.bar ? I suspect that cljs, like clj, expects the namespace foo.bar to be found in a directory bar inside of a directory foo

Or rather, in a file named bar.cljs inside of a directory named foo

No sorry. I have a directory foo, with a file bar.cljs . These are both in the same folder as core.cljs. This was a reply to reply 1 btw XD

Understood.

Have you attempted to delete any/all caches of compiler output from previous cljs compiler runs you have made, and then run the compiler again?

I tried running shadow-cljs compile again. I'll try to delete all js-files in target/cljsbuild too.

Did not work 😕 I must be doing something stupid.

What does the complete ns form look like inside of your file foo/core.cljs ?

And the function baz isn't commented out in some way? Or perhaps it is defined after some point where there is a compiler error in the file foo/bar.cljs, so the compiler never gets that far to see baz ?

(ns ouija.core
  (:require
    [day8.re-frame.http-fx]
    [reagent.dom :as rdom]
    [reagent.core :as r]
    [re-frame.core :as rf]
    [goog.events :as events]
    [goog.history.EventType :as HistoryEventType]
    [markdown.core :refer [md->html]]
    [ouija.ajax :as ajax]
    [ouija.events]
    [ouija.highlight :refer [highlight] ; offending line
    [reitit.core :as reitit]
    [reitit.frontend.easy :as rfe]
    [clojure.string :as string]
    [com.rpl.specter :refer [MAP-VALS]])
  (:import goog.History))

https://github.com/endrebak/ouija/

If you are willing to try an experiment where highlight is the first function defined in ouija/highlight.cljs, that may help determine whether there is some other error in ouija/highlight.cljs that could be causing trouble.

It is just a toy project. I think you should just be able to git clone and compile

lein shadow watch app

I will experiment

I tried adding the line (defn highlight [])to the top of the file. Did not change anything.

It might well be due to you having highlight.clj and highlight.cljs in the clj tree as well as highlight.cljs in the cljs tree?

> find . -name 'highlight.*'
./src/clj/ouija/highlight.clj
./src/clj/ouija/highlight.cljs
./src/cljs/ouija/highlight.cljs

Oh, wow! Nice catch

Now it works 🙂 :thumbsup:

Thanks to the both of you.

If you have required specter, (eval (read-string "MAP-VALS")) will give com.rpm.specter/...

But there is no eval in clojurescript...

There is resolve: https://clojuredocs.org/clojure.core/resolve huzzah!

For #() the argument bindings are implicit, so if you ignore the (fn [args ...] it slides right in.

At least that's how I see it

@UCW9TUDNK Sure, but that's not the problem here. Its the implicit function application, unless its a let.

Oh, its because let is also surrounded by ()....I think.

Very odd, thanks btw. 🙂

There is no function application though, as #() is a noop (same as (fn []).

Yes. I attribute it to lisp macro "magic" [=

the expression following the # becomes the body of the fn, it is confusing because the parens in #(...) both act as delimiters of the anonymous function, and as part of the expression (...) which becomes the body of the anonymous function

another way to think of it is # starts an anonymous function, where the form that follows it is the body, but that form has to be a list (is wrapped in parens)

another way to understand it would be figuring out which parenthesis correspond to each other in the #(...) syntax and the (fn [...] (...)) syntax

#(...) => (fn [...] (...))

not sure how well that formatting emphasis comes through

#(let [...] ...) => (fn [...] (let [...] ...))

mechanically it is the same thing

That's probably one of the best explanations I've heard for it, to be honest!

#parenthesized-thing => (fn [...] parenthesized-thing) or #`(...)` => (fn [...] (...))

Ya, I get it. Thanks. 🙂

I'm just a little more used to Racket macros where that sort of thing is less common. 🙂

Anyway, thanks.