@alex340 could you show an example usage of your macro? I’m struggling:

(defstyled header :h1 "color: blue;")

(defn hello-world []
  [:div {:style {:font-family "sans-serif"}}
   [header (:text @app-state)]
   [:h3 "Edit this and watch it change!"]])

core.cljs:305 Uncaught Error: No protocol method ICollection.-conj defined for type string: Hello world!

hmm, I see. by changing it to [header {} (:text @app-state)] it now works as expected

OK, I’ve gotten up and running with shadow-cljs and it’s easier than I expected

according to this: https://clojurescript.org/about/differences

Currently ClojureScript numbers are just JavaScript numbers

@ajs CLJS doesn't have int types...

(type 1) in cljs returns number

(defn int
  "Coerce to int by stripping decimal places."
  [x]
  (bit-or x 0))

i see, the functions int? and integer? were leading me to believe it did, but i see now they are not what I think they are

Hi. I've been struggling with converting the following javascript into clojurescript, and hoped someone here can help me:

let Inline = Quill.import('blots/inline');
class BoldBlot extends Inline { }

http://teavm.org

Is it interesting for clojurescript?

with this,

(defn clear-button []
  [:a {:style {:visibility (if (str/blank? (@app-state :url))
                             "hidden"
                             "visible")}}
   [:span.icon.is-right.is-small {:on-click #(.log js/console "hello")}
    [:i.fas.fa-times]]])

(defn clear-button []
  [:a 
   [:span.icon.is-right.is-small {:on-click #(.log js/console "hello")}
    [:i.fas.fa-times]]])

hold on - chrome's inspector doesn't show <span onClick="blah"> as i was expecting it to

so i'm guessing that the span just got rendered below another element and can't receive the "click"

@ackerleytng React never places the event handlers on DOM. It handles the events internally.

@ackerleytng He means that it registers the event programmatically, that's why it does not show up on the span element in the dom displayed by the inspector.

Your span may not receive the click if it occupies no space on the web page. Double-check that, add some text inside your span tag.

You may try to set pointer-events: none; on the style of :i.fas.fa-times and see if it helps.

ah i see

thanks so much!

I'm pretty sure it's something to do with layering

i'm using bulma and the cross was actually an icon, something like https://bulma.io/documentation/form/input/#with-font-awesome-icons

when i right click on the cross and click inspect, chrome highlights the input and not the span, so i guess bulma might have hidden the cross below

i switched to using a button, as in https://bulma.io/documentation/form/general/#form-addons so that fixed it

The discussion of (.method js/object) vs (js/object.method) comes up a bunch and people seem to favor the former. But just to be clear: the latter form is passed through to javascript without being subject to advanced compilation, correct? In dealing with external libraries (without extern declarations), it seems like this is a big advantage.

wait i’m confused. what does that demonstrate?

oh. nevermind i got it

only object["method"] would be safe

right

back to using cljs-oops 🙂

btw. (.toString #(.method js/object 1 2 3)) produces exactly the same js code, so (I assume) there is no difference when the code gets passed to google closure compiler

i forgot that the evil happens after clojurescript is done

@lee.justin.m Just enable externs inference and this will be covered automaticallly

@rauh i guess. the documentation on externs inference is not exactly reassuring. particularly the part where it says (Alpha) in the title

Well you can always inspect the generated externs in your source folder and see if what's in there. Unless you have thousands you can quicky verify it.

It's working pretty well for most people.

my problem is that it’s just me working on this project. i can’t be going off into the deep end of compiler internals, or trying to figure out some subtle bug with externs, or trying to translate from es7 code. using cljs-oops is dead simple and it will definitely work, so that’s kind of what I’m leaning towards

cljs-oops also has dev-time diagnostics, which could help catch some class of mistakes

all of the documentation on inference and externs and so forth tend to assume that I understand what on earth google closure compiler is doing, and how its type system works. i really don’t want to know about that if i don’t have to

on the other hand, cljs-oops could slow down your compilation times (it uses macros heavily)

thankfully for me that is the least of my worries. i care only about correctness and developer power

personally I’m in the same boat, also cljs compilation is pretty slow anyways, so during development I have to rely on incremental compilation setup anyways

I’m trying to understand how to use shadow-cljs with the macchiato framework that yogthos wrote

I’m not quite understanding how figwheel knows how to start/stop the http server in the template

@lilactown happy to help in the #shadow-cljs chan

ah OK, i’ll switch there 🙂

With a clj + cljs project developed in cider where we want to share code, should it be 2 Lein projects or one? If one, how do I specify client vs server dependencies?

@alexisvincent you could do either one

client dependencies can be mixed with server dependencies - when you compile the projects, the server will only import the dependencies it needs. same with the client