So the node js route won't be ideal for browser

Basically what happens is that your calls to require execute correctly on node, and when you compile with basic optimizations the times you call require stay in place

From there using browserify on your code would work

But browserify/webpack, as dnolen would be quick to point out, don't do a great job at optimizing and dead code eliminating generated clojurescript code

*the JavaScript generated from clojurescript

In browser, where code size is a big deal, you basically have to use advanced optimizations, which doesn't work with the require function, but instead with js externs and Google closure namespaces

Work is being done to make using code from node doable (both on the clojurescript team and the Google closure team)

But it's all really alpha right now

For now just look on cljsjs for the js libraries you want to use and if you can't find what you need then come back here and someone will help you out

*probably

@msuess require gets compiled away so probably something went wrong. The guide is clear you can’t use ES6 yet w/ Node modules support

but it’s a cutting edge feature so expect a bumpy ride unless you have a lot of patience and are OK with digging into how the feature actually works

Anyone have a line on how to do sha256 hashing in cljs? All the links I'm finding seem to be out of date. TIA

Hello, if I'm looking to use third-party react component from npm, is http://blob.tomerweller.com/reagent-import-react-components-from-npm a good guide to follow?

Specifically, the approach is to compile the react components with webpack, attach it to window, and exclude react and react-dom from reagent.

hey all I have been playing witht the cljs.js/compile-str and I noticed the library resolution function passes a string for the library name. I was wondering what would happen if there are two files name the same coming from different libraries? Anybody knows how collisions get resolved?

@yanglin That guide works iff you have the externs for the libraries you are using or you don't use advanced compilation

Also be sure to add the generated bundle.js to your unit tests

Anyone knows how to get around importing foreign libs that uses flow types ? I can't find anything about this and I am not familiar with flow type...

what library? usually projects that code with type annotations should be running their code through a compiler step to remove them; there might be another .js file in the package you should refer to instead

gl-react

I see a scripts/flow.sh

Oh, you're getting the source directly from github?

you should get the release builds instead, they'll have the compiled output as normal javascript in the lib/ folder

Thanks, that makes sense 🙂 I'll try that!

does cljs have access to gen ?

does cljs have access to spec.gen? I'm trying to (:require ,clojure.spec.gen :as gen] in a cljs file, and it can't find it

err, when I :module-type :es6 in my :foreign-libs it fails to find the :provides ["gl-react"] namespace (No such namespace: gl-react, could not locate gl_react.cljs...)

but when I remove the :module-type :es6 then it builds and later fails executing js cause it's a es6 module 😞

@qqq the ns is cljs.spec.impl.gen due to a conflict with cljs.spec/gen otherwise

I believe I found out my mistake is I had not browserify'ed my foreign libs.

my foreign lib expects window.react to be defined when it's :required... Any idea how I could (set! js/React react) when I :require cljsjs.react ?

I could fix it by importing react in my index.html but there has to be a more idiomatic way...

I found this googling around for a better solution: http://blob.tomerweller.com/reagent-import-react-components-from-npm

I find it's an interesting option that I believe would fix my problem (foreign lib needs window.react define prior to :require) but I'd loose cljsjs... maybe there is an even better solution?

@sineer the :module-type for :foreign-libs is very much a work in progress, webpack should be more reliable for a while

unless you really feel like trying cutting edge stuff

clsjs looks a lot like cljsjs when you're tired!

how to write this in clojurescript: messages.indexOf(link.source) > -1 ? link.target : link.source ?

messages is an array, link.source & .target are integers

by this point i know that in the array it's only one of them that is present and one of them is always present

and i would like to get back the one that isn't

well, i wrote (if (> (.indexOf messages link.source) -1) link.target link.source) hopefully it will work

now i need to figure out how to modify js Objects 😄

your code appears correct, to modify js object use (set! js/obj val)

js/obj ?

replace that by your javascript object ref.. js/ is to access javascript global ns

i am using variables from local scope

yeah, ok, then i understood correctly, thanks

problem is i want to set! a property, not the whole object

and (set! objname "prop" val) is incorrect

ashnur read this http://squirrel.pl/blog/2013/03/28/two-ways-to-access-properties-in-clojurescript/

its important to understand it properly for when you'll use advanced optimization...

thanks, this is great

@ashnur native interop is usually done via the the . (dot) operator

(if (> (.indexOf messages link.source) -1) link.target link.source) this shouldn't work

(if (> (.indexOf messages link.source) -1) (.-target link) (.-source link)) use (.-property obj) to access properties of objects

same for set (set! (.-property obj) new-value)

oh nvm, the article covers that

yeah, but thanks anyway, better twice than not even once 🙂

what happens when i doseq over a javascript array?

try it!

in lumo:

(doseq [x (into-array [:a :b :c])] (prn x))

@pesterhazy i don't understand

how would trying it tell me what it does actually?

i am already using it, but my question is not about the result, it is about it's action

ah, I didn't understand then

here's the source of seq: https://github.com/clojure/clojurescript/blob/47fa30dd95c850aa0643aae9c274c09bc202065c/src/main/cljs/cljs/core.cljs#L1113

as you can see it hands over to -seq of ISeqable

can't find where ISeqable is defined for arrays though

yeah, that's why i asked :)]

ah wait it checks if it's an array in seq

can't decipher, is it converting into something or there is some default

then it creates an IndexedSeq of the array

here's the implementation: https://github.com/clojure/clojurescript/blob/47fa30dd95c850aa0643aae9c274c09bc202065c/src/main/cljs/cljs/core.cljs#L1510

it uses aget and keeps track of the index

@ashnur finding the anwer you were looking for?

not yet, no, but i figured first i write it and see if it's slow

so for now i jsut use it

(defn somefn [] (let [localvalue (.jsfunc js/somejsobj (fn [] ( ... how to access localvalue here? ... )))]))

localvalue

says undeclared

and i think it's correct

you're not showing the actual code

i have not yet written the actual code

you are saying i should be able to access it?

in that example no

i can show more

https://gist.github.com/ashnur/910a4fdc3556a32f435f248c04108c8f#file-index-js-L72-L91

ah I see

see how i use timer variable from inside the anonymous function

the lack of identation threw me off

you can use an atom for that if you want

i am afraid of using atoms, because i don't know what they are doing internally

i am not even sure how would help me here

this type of assignment won't work in CLJS

that's what i suspected

(let [!localvalue (atom nil)] (reset! !localvalue (do-something (fn [] (prn @!localvalue)))))

kind of thing. I'm not sure I balanced parentheses correctly there 🙂

oh, i know then, this @/deref thing does a next-tick

come again?

a skip in the event loop

not sure how to say it

blocks

@ashnur There is no async in cljs.core

rauh how is that relevant here, i don't see

The deref doesn't do next-tick like you guessed

All of cljs.core is sync

oh, ok, thanks for the correction then

@ashnur deref is just "give me the value of the thing now", basically a property access

@ is just syntax sugar. nothing more.

@thheller so, i should do what @pesterhazy suggested to solve this situation? use an atom?

yes

i am not grokking this yet

ok

thanks to everyone

(let [timer-ref
      (atom nil)
      
      timer-fn
      (fn [elapsed]
        ...
        (.stop @timer-ref))
      
      timer
      (js/d3.timer timer-fn)]
  (reset! timer-ref timer))

maybe that is clearer, you need to name for the timer before it actually exists, so timer-ref is a placeholder basically. ie. there will be a thing in here when timer-fn is called

yeah, i realized what was "wrong" with the original situation 🙂

hmm. i am having a brain freeze... how to have multiple expressions in a conditional? (do) ?

yes

yay

i am sure when i finish this i will see some ways to make it more idiomatic to clojurescript, but for now, it's not very nice ;D

const notCurrentEdge = (fe) => currentEdges.find((ce) => ce.index != fe.index) how to have a predicate like this? currentEdges is an array of objects with index and i am trying to make sure that the object i am checking has an index that's not present. kinda obvious .. not sure what's the best way to do this

basically because i don't have a .find for cljs

ashnur: you can emulate .find with (first (filter #(not= (.-index %) (.-index fe)) currentEdges))

i did (every? (fn [edge] (!= (aget fe "index") (aget edge "index"))) current-edges)

but maybe it's wrong

i have another error to fix before i can test this

ah yeah, that works if you’re just trying to make sure find finds nothing

?

ignore me, you are doing the right thing 🙂

oh, 😄

one thing is sure, this learning project is actually doing its job, i am running into a lot of weird stuff

https://owensd.io/2017/02/06/clojurescript-electron/

aaand i have found where i need mutation of a list. how can i do forEach? 😄

https://gist.github.com/ashnur/910a4fdc3556a32f435f248c04108c8f#file-index-js-L75 this mutates the array currentEdges

http://blog.cognitect.com/blog/2017/1/31/state-of-clojure-2016-results

very illuminating results

what I see: cljs needs better JS libs support (which is a wip) and better marketing (wonder how this will end up)

^ 2nd may be a consequence of the 1st too

how do i write this in cljs? const inTheDark = (edge) => (messages.indexOf(edge.source) > -1 ) ^ (messages.indexOf(edge.target) > -1 ) 😄

also, this is weird Use of undeclared Var combo-network.core/!=

i checked, this works https://clojuredocs.org/clojure.core.logic/!=

so why it says undeclared?

@ashnur that is core.logic, you probably want not=

huh

what's the difference

i used = and changed every to not-any

if you want != as it is in JS use not=. core.logic is a logic programming library, an entirely different thing

https://clojuredocs.org/clojure.core/not=

i understand that i have to use not=

what i don't understand is core.logic, but guess that can wait

i mean i thought that core.logic is just the part where logic stuff is separated but that doesn't mean i can't use it

so, how can i have XOR?

bit-xor

i am fairly sure i can't do the hack from javascript to just use the bitwise oeprator instead, right?

yeah, it needs number arguments

(or one-thing other-thing)

that's not xor

i really hope it isn't

🙂

@ashnur bit-xor

ah ok, you want to call it without numbers but still have the a number as the result

@rauh i am not sure why would you say that as i just mentioned that i tried and it wants number inputs

yeah not in clojure

yeah, i want (xor true false) -> true

not really a new idea

(or true false) => true?

you keep using bools 😛

...

you realize that or is not xor, right?

this is the second time you suggest or

(or true true) gives true while (xor true true) gives false

yes, but xor with bools doesn't compute in my head

what do you mean?

it's just exclusive or

x stands for exclusive

it does the same thing, but not on bits but on "a single bit" which is just a boolean

that's why it works in js where type coercion casts true/false to 1/0

yes, I get that ... just in my history whenever I used binary ops I used it on numbers. just ignore me .. I'm old.

ashnur: bit-xor compiles into x ^ x

thheller like i am young 😄

https://github.com/clojure/clojurescript/blob/master/src/main/clojure/cljs/core.cljc#L1141

@jr that's cool, but if i try to use it with booleans it complains, guess i could cast the boolean into numbers

anyway, i just wrote xor function

@ashnur it seems to be more common in the clojure world to use all, some, and, or or when doing boolean logic. Not that you can't use xor if you want, just not overly common in clojure

yeah the analyzer tags the input as a number so it will complain

@roman01la better marketing you mean for the staffing concerns?

@anmonteiro I think better integration with JS will have some impact - but I do think the marketing problem is a broader issue

as I’ve said before nearly 60% of users come from Java/Python/Ruby so I think there’s a lot we can do here

I remain skeptical of attacking the JS behemoth directly 🙂

i know that if it wasn't for the java tight coupling, i would've tried cljs 2 years earlier

although now that i said this, you probably want to tie it to java even more just to keep people like me away 😛

@ashnur you are in a persistent minority that we are well aware of 😉

😛

but survey’s prove it doesn’t matter that much

i am not sure how can that survey tell anything about this, but sure, as you said, doesn't matter much

we have questions about pain points

Java is not high on the list

but it would be crazy for you to have java high on that list because it would go against the tide where you perceive money comes

so to me its kinda obvious why even if i were right you would deny it 🙂

it never appears as a problem

people can add freeform feedback

but again, i don't know if i am right, and it doesn't really matter

yeah, never, only for me :))

see, you deny my existence :))

someone wrote a library for CLJS/CLJ macros to make it easier to generate platform specific code, anyone remember the name?

other than the (:ns &env) thingy

@ashnur no, just not a priority and unlikely to be - there’s just more important things to work on

@thheller I believe this is what you're looking for https://github.com/cgrand/macrovich

@anmonteiro thats the one, thanks

@ashnur if you have bools your xor is not=

FYI, I discovered dirac (cljs repl for devtools!) two days ago and it rocks! Coolest repl I ever seen 🙂 Check it out if you haven't already...

Also, I'm extremely happy to have a working gl-react canvas in devcards with om next 🙂

@rauh ?

@ashnur Well (not= true true), true false, false false, false true are all equal to (xor a b)

wow

it's so obvious

thanks rauh

@rauh, clever

(defn xor [a b] (not= (boolean a) (boolean b)))

@dnolen: I meant better marking for broader adoption

Has any seen the error

[Violation] Parser was blocked due to document.write(<script>)

@rayk I’m assuming this is in Firefox or something?

Chrome Canary

we only do that under dev - so not a big deal

Thanks... will not waste the time chasing it down...

I found cljs-http to do http get but I'm wondering if there's a simpler or more common way to load resource files? a simple blocking load would do...

@sineer blocking load isn’t desirable since JavaScript is single threaded

browsers don't do sync 😞

@pesterhazy you can do a sync get but I don’t recommend it

yeah I was just thinking about that trying to think of a way if at all possible.. it's for devcards stuff and to defonce some shaders data..

@dnolen, really? you can send synchronous network requests from javascript in a browser?

node.js has blocking fs read also if I remember correctly

I guess I'll just have to use cljs-http... Out of curiosity is there anything simpler (no core.async dep maybe) ?

cljs-ajax, https://developers.google.com/closure/library/docs/xhrio

yeah that's what I was looking for 🙂 Thanks! I remember using something similar with react-native

From a cljs repl, is there any way to block until some async event happens?

I can do this:

cljs.user=> (def r (atom nil))
#'cljs.user/r
cljs.user=> (goog.net.XhrIo/send "" #(reset! r (-> % .-target .getResponseJson js->clj)))
#object[Object [object Object]]
cljs.user=> @r

but it would be neat if I could programmatically wait for the result (instead of manually checking the atom periodically)

@pesterhazy can you not use channels for that

@stbgz how would that work?

@sineer do you use any wrapper for gl-react or does it "just work" with om-next ? I’m a reagent user so not very familiar with om-next.

I'm working on a cljs lib/wrapper for gl-react it's just prototype yet but I'm making great progress 🙂

I havent seen gl-react used with cljs anywhere but there isn't much to it... the hard work is getting the foreign libs to work

@pesterhazy I swear I’ve seen that being done with channels, but I can’t find the doc anymore, I’ll ping you if I find it

gl-react is just a react component so you can use it with plain react or which ever cljs wrapper you prefer 🙂

@stbgz cool thanks

@stbgz maybe you're thinking of using <!! i.e. blocking take, which is intentionally missing from the cljs implementation of core.async

does anyone know how to get goog.DEBUG to end up false in CLJS 1.9.229 ? I've tried :closure-defines {:goog.DEBUG false} and :closure-defines {"goog.DEBUG" false} to no avail.

[both in my profile with :optimizations :advanced]

@johanatan either 'goog.DEBUG false or "goog.DEBUG" false should work

what's the symptom you get?

hmm, ok. I think i've found the problem

do i need lein cljsbuild once prod to actually build the prod profile?

i think i was still actually building the debug one with lein cljsbuild once

I don't use cljsbuild, so I couldn't tell you

yep, lein cljsbuild once prod is invoking the right build now. will see how that turns out. thx!

don't forget to add the ^boolean type hint too

helps with DCE

use goog/DEBUG, the js/goog.DEBUG variant won’t work in .456: https://github.com/binaryage/cljs-devtools/commit/4e368408986bf739b56aad64e064b854fd7989d2

@darwin 👍

in general I stay away from :closure-defines, too many sharp edges

macros will do the same job predictably

do you have an example of using macros for prod vs dev builds?

give me a sec

hmm, my prod build seems to be broken with this cryptic stack trace:

java.lang.ClassCastException: clojure.lang.PersistentList cannot be cast to clojure.lang.Named
	at clojure.core$name.invokeStatic(core.clj:1546)
	at clojure.core$name.invoke(core.clj:1540)
	at cljs.closure$make_options.invokeStatic(closure.clj:263)
	at cljs.closure$make_options.invoke(closure.clj:253)
	at cljs.closure$optimize.invokeStatic(closure.clj:1266)
	at cljs.closure$optimize.doInvoke(closure.clj:1259)
	at clojure.lang.RestFn.applyTo(RestFn.java:139)
	at clojure.core$apply.invokeStatic(core.clj:648)
	at clojure.core$apply.invoke(core.clj:641)
	at cljs.closure$build.invokeStatic(closure.clj:2020)
	at cljs.closure$build.invoke(closure.clj:1927)
	at cljs.build.api$build.invokeStatic(api.clj:199)
	at cljs.build.api$build.invoke(api.clj:187)
	at cljs.build.api$build.invokeStatic(api.clj:190)
	at cljs.build.api$build.invoke(api.clj:187)
	at cljsbuild.compiler$compile_cljs$fn__5973.invoke(compiler.clj:60)
	at cljsbuild.compiler$compile_cljs.invokeStatic(compiler.clj:59)
	at cljsbuild.compiler$compile_cljs.invoke(compiler.clj:48)
	at cljsbuild.compiler$run_compiler.invokeStatic(compiler.clj:168)
	at cljsbuild.compiler$run_compiler.invoke(compiler.clj:122)
	at user$eval6110$iter__6146__6150$fn__6151$fn__6169.invoke(form-init7676291134357241625.clj:1)
	at user$eval6110$iter__6146__6150$fn__6151.invoke(form-init7676291134357241625.clj:1)
	at clojure.lang.LazySeq.sval(LazySeq.java:40)
	at clojure.lang.LazySeq.seq(LazySeq.java:49)
	at clojure.lang.RT.seq(RT.java:521)
	at clojure.core$seq__4357.invokeStatic(core.clj:137)
	at clojure.core$dorun.invokeStatic(core.clj:3024)
	at clojure.core$doall.invokeStatic(core.clj:3039)
	at clojure.core$doall.invoke(core.clj:3039)
	at user$eval6110.invokeStatic(form-init7676291134357241625.clj:1)
	at user$eval6110.invoke(form-init7676291134357241625.clj:1)
	at clojure.lang.Compiler.eval(Compiler.java:6927)
	at clojure.lang.Compiler.eval(Compiler.java:6917)
	at clojure.lang.Compiler.load(Compiler.java:7379)
	at clojure.lang.Compiler.loadFile(Compiler.java:7317)
	at clojure.main$load_script.invokeStatic(main.clj:275)
	at clojure.main$init_opt.invokeStatic(main.clj:277)
	at clojure.main$init_opt.invoke(main.clj:277)
	at clojure.main$initialize.invokeStatic(main.clj:308)
	at clojure.main$null_opt.invokeStatic(main.clj:342)
	at clojure.main$null_opt.invoke(main.clj:339)
	at clojure.main$main.invokeStatic(main.clj:421)
	at clojure.main$main.doInvoke(main.clj:384)
	at clojure.lang.RestFn.invoke(RestFn.java:421)
	at clojure.lang.Var.invoke(Var.java:383)
	at clojure.lang.AFn.applyToHelper(AFn.java:156)
	at clojure.lang.Var.applyTo(Var.java:700)
	at clojure.main.main(main.java:37)
Subprocess failed

Any idea on that one?

is that from the 'goog.DEBUG as recommended above?

I may have missed the correct syntax 🙂

try 'goog/DEBUG

yep, changing it to "goog.DEBUG" has fixed that

slash instead of dot

oh, hmm. ahh ok

@johanatan in a macro you can have access to compiler options, and you can decide to emit clojurescript code or not, for example depending on the :optimizations key https://github.com/binaryage/cljs-oops/blob/276de0b73fbcef6a7ad43b31470ecaf2139da097/src/lib/oops/config.clj#L21-L23

@darwin cool, thx!

this technique can be combined with preloads and leiningen profiles

and what are the "sharp edges" involved with the closure approach?

for a simple "debug" flag, would it be fine?

the sharp edges were related to DCE, I expected it to happen and it silently didn’t, once because of missing ^boolean type hint, second time because of using js/goog.DEBUG form

but code worked, it just prevented DCE, so it wasn’t that bad 🙂

anyways :closure-defines is not widely used I would say, so there might be some quirks not only with DCE

what is DCE?

Dead Code Elimination

ahh

btw, i had to use goog.DEBUG here. goog/DEBUG did not work

perhaps you're on a later version than I ?