Is there a way to get a multimethod to dispatch on the type of an object? Just dispatching on type of an object does not work because a list could be a IndexedSeq, LazySeq, or List.

This works in Clojure but not Clojurescript

(defmulti my-multi #?(:clj  class
                      :cljs type))

(defmethod my-multi #?(:clj  ISeq
                       :cljs cljs.core/List)
  [x]
  ;; handle Lists here
  )

Hi, I'm utilizing a button component that only supports href (and not on-click). What is the proper way to execute arbitrary ClojureScript function in the href attribute value?

Oops, nevermind. Undocumented [and uncoded actually] (and thus probably default) support for on-click apparently does exist.

FWIW, to solve my earlier multimethod problem, I just wrote a new type function:

(defn type->kw
  [x]
  (cond
    (seq? x) ::seq
    (vector? x) ::vec
    (map? x) ::map
    (or (string? x) (number? x)) ::string
    :else (type x)))

ha, nice

@kenny @jrheard here is a live demo of the code http://app.klipse.tech/?cljs_in=(defn%20type-%3Ekw%0A%20%20%5Bx%5D%0A%20%20(cond%0A%20%20%20%20(seq%3F%20x)%20%3A%3Aseq%0A%20%20%20%20(vector%3F%20x)%20%3A%3Avec%0A%20%20%20%20(map%3F%20x)%20%3A%3Amap%0A%20%20%20%20(or%20(string%3F%20x)%20(number%3F%20x))%20%3A%3Astring%0A%20%20%20%20%3Aelse%20(type%20x)))%0A%0A%0A(defmulti%20my-multi%20type-%3Ekw)%0A%0A(defmethod%20my-multi%20%3A%3Aseq%20%0A%20%20%5Bx%5D%0A%20%20(first%20x))%0A%0A(defmethod%20my-multi%20%3A%3Avec%20%0A%20%20%5Bx%5D%0A%20%20(second%20x))%0A%0A%0A%5B(my-multi%20%27(10%202%203))%0A%20(my-multi%20%5B10%202%203%5D)%5D

Cool! Haven't seen that app yet. Only slightly annoying thing is the URL 😉

hello. Trying to compile https://github.com/danielsz/system-websockets but getting an error:

@dimovich system-websockets require Clojure 1.8

added boot.properties to the project with clojure 1.8.0 requirement

compiles now 🙂

👍

@juhoteperi Do you have any idea why the Websocket REPL error :occupied error would occur in React Native only on Android but not on iOS when using boot-cljs-repl? Possibly related to Weasel?

@vikeri No idea

Apologies if this is the wrong channel. I’m trying to get a simple ajax call going from my webapp, and it’s been a character-developing couple of days. I do not feel like a smart man.

The call is for login, and it sends no data - just an authorization header.

(try
       (ajax/POST "/api/login”      ;; ajax is the cljs-ajax library
                {:headers        {"Authorization" (encode-auth (string/trim id) password)}
                 :handler        #(do
                                    (prn "Success" %)
 ;;                                    (session/put! :identity id)
 ;;                                    (js/setTimeout session-timer timeout-ms)
 ;;                                    (reset! fields nil)
 ;;                                    (set!  (.-href js/location) "#/dashboard")
                                    )
                 :params {}      ;; same symptoms whether this is present or absent
                 :error-handler  #(prn "Failed" %)
                 })
       (catch js/Object e
         (log "Caught " e "\n" (.-stack e))
         )

Apologies for the commented lines - I’ve stripped it down to the most basic success handler, just a prn

The :params key was added as an experiment, the symptom is the same with or without.

In any case, when this code is executed, the server seems to choke on request:

2016-09-28 05:08:06,111 [XNIO-1 I/O-2] DEBUG  - Error reading request
java.io.IOException: Connection reset by peer
	at sun.nio.ch.FileDispatcherImpl.read0(Native Method) ~[na:1.8.0_60]
	at sun.nio.ch.SocketDispatcher.read(SocketDispatcher.java:39) ~[na:1.8.0_60]
	at sun.nio.ch.IOUtil.readIntoNativeBuffer(IOUtil.java:223) ~[na:1.8.0_60]
	at sun.nio.ch.IOUtil.read(IOUtil.java:192) ~[na:1.8.0_60]
	at sun.nio.ch.SocketChannelImpl.read(SocketChannelImpl.java:380) ~[na:1.8.0_60]
	at org.xnio.nio.NioSocketConduit.read(NioSocketConduit.java:286) [xnio-nio-3.3.6.Final.jar:3.3.6.Final]
	at org.xnio.conduits.ConduitStreamSourceChannel.read(ConduitStreamSourceChannel.java:127) ~[xnio-api-3.3.6.Final.jar:3.3.6.Final]
	at io.undertow.server.protocol.http.HttpReadListener.handleEventWithNoRunningRequest(HttpReadListener.java:153) ~[undertow-core-1.3.23.Final.jar:1.3.23.Final]
	at io.undertow.server.protocol.http.HttpReadListener.handleEvent(HttpReadListener.java:131) ~[undertow-core-1.3.23.Final.jar:1.3.23.Final]
	at io.undertow.server.protocol.http.HttpReadListener.handleEvent(HttpReadListener.java:57) ~[undertow-core-1.3.23.Final.jar:1.3.23.Final]
	at org.xnio.ChannelListeners.invokeChannelListener(ChannelListeners.java:92) ~[xnio-api-3.3.6.Final.jar:3.3.6.Final]
	at org.xnio.conduits.ReadReadyHandler$ChannelListenerHandler.readReady(ReadReadyHandler.java:66) ~[xnio-api-3.3.6.Final.jar:3.3.6.Final]
	at org.xnio.nio.NioSocketConduit.handleReady(NioSocketConduit.java:88) [xnio-nio-3.3.6.Final.jar:3.3.6.Final]
	at org.xnio.nio.WorkerThread.run(WorkerThread.java:559) [xnio-nio-3.3.6.Final.jar:3.3.6.Final]

But when I issue the same call with curl instead of the web client, it works fine (the curl text was obtained by looking at the failed call in Chrome Dev Tools and “copying to curl”, so it should be the same as the XHR call from the browser)

~/dev/cavalry-admin # curl '' -H 'Accept: application/transit+json' -H 'Referer: ?' -H 'Origin: ' -H 'x-csrf-token: APvC0olHJjLRz61TKRRIdQlSE4sEloCdE3yCuBwuad9rt9F9Uxdmvh9cv5e5oLL4emN5IkwxyKTZJycW' -H 'User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/53.0.2785.116 Safari/537.36' -H 'Authorization: Basic cm9kZnJleTp0d2VlZGxl' -H 'Content-Type: application/transit+json; charset=UTF-8' --data-binary $'["~#\'",null]' --compressed

["^ ","~:result","~:ok”]

I really don’t know what the next step is - this is such a baby step (issue an XHR call from the browser) that I’m not able to accomplish, despite following a recipe.

@juhoteperi Alright I’m investigating to see if I find someting

Does anyone now a good way to inspect websockets?

The chrome inspector allows you to inspect the ws in the network tab

@credulous How'd you setup your server? Did you use a template or anything like that? Or are you experienced enough with clojure to know for sure that that part should be working 🙂

trying to use danielsz/system (ripped apart the example Sente project), and I keep getting this error: java.lang.Exception: #'namen.systems/dev-system is not a SystemMap

if anyone bumped into this before, let me know plz 🙂

@dimovich could you paste your namen.systems/dev-system var?

@dimovich right so the problem is that you’re using defn

try to use (def dev-system ...)

(obviously removing the argument vector etc)

hrm actually that might not be right

@dimovich hrm, you’re using Boot, right?

could you show me the call to the system task?

yeah, def doesn't fix it...

(system :sys #'dev-system :auto true :files ["handler.clj"])

@dimovich that’s just weird, I’ve got a very similar configuration which works perfectly

did you (require ‘[namen.systems :refer [dev-system]]) in your build.boot?

yes I did...

i took this project template https://github.com/danielsz/system-websockets, which compiles fine, and deleted everything I didn't need...

seems I also deleted something I shouldn't had to

@anmonteiro thx for your help... will try again from scratch 🙂

@kauko I used Luminus. I’m pretty sure it’s working because I have logging statements throughout the server code and I get positive logs right up until the return value of the last handler. Also, when I call the server using curl it works fine.

Huh.. maybe try asking in the #luminus channel? I'm sorry I can't help you

Yeah tried but the response was that it looked like a server problem. I dunno.

I guess you could try creating a fresh luminus project and try to see where you've done something different? IIRC the template includes some ajax calls

Hi. I'm working with the Google Maps API places autocomplete search box and I need to select the first entry on Enter or Tab. I saw some answers on StackOverflow that involve wrapping the listener and simulating a down arrow press on enter or tab. Here's the code I have so far http://pastebin.com/Ex6Zu48q can someone guide me on how I can wrap the event listener and simulate a down arrow press?

@viebel - one question i’ve had re: using klipse in blog posts - is there a way to include third-party libraries? like, say i want to have a blog post that plays around with core.logic

seems like a tall order, i’m not sure how one would go about implementing it, but i figured i’d check just in case 😄

@kenny I feel like you really want extend-protocol instead of multimethod on type.

@jrheard there is a way to use third-party libraries using data-external-libs. See https://github.com/viebel/klipse#clojure-only-1

siiiiick

nice dude

But there is a catch: the library must be self-hosted cljs compatible

aha, makes sense

Check this blog post for instance: http://blog.klipse.tech/lambda/2016/07/24/lambda-calculus-2.html

one other thing hasn’t been obvious to me - how is it that klipse windows later on in the document are able to access definitions from earlier on in the document?

Where I’m requiring a library of mine named gadjett

eg in one window you define a record Foo, and later on you use Foo, and the second window still works even if i change Foo’s definition in the first

@jrheard all the evaluations occur in the same context

makes sense?

not sure i follow, sorry

are you saying that on pageload, each window is evaluated one after the other, and is able to access definitions from earlier windows?

(i’m using the word “window” here because i don’t know what word to use for “kilpse plugin instance” or whatever) 🙂

the proper word is: “snippet"

gotcha, makes sense

You’re right about evaluation order of the snippets

cool, makes sense

but they’re not, like, permanently linked - so after pageload happens, if i modify or remove the definition of T in http://blog.klipse.tech/lambda/2016/07/24/lambda-calculus-2.html , later windows aren’t affected, and that’s expected+good, right?

100% correct

awesome, makes sense

thanks a bunch for answering my questions! 🙂

until you re-evalute the other snippets

with Ctrl-Enter

then the new definitions will affect the snippets

ah, yeah, for some reason i thought that that wasn’t the case, but i see that it works now

i think what i was doing was just removing definitions and ctrl-entering them, and of course that wouldn’t affect the rest of the document

BTW, are you planning to write a blog post with klipse snippets?

you’ll be the first to know when i do 🙂 i know you’re trying to get a lot of publicity for this tool and it makes a ton of sense

i’ve mainly just been curious, no immediate plans atm

but hopefully in the next month or two i’ll crank something out, and i’ll definitely let you know

fwiw i learned about it by watching one of your talks on youtube, probably from strange loop

i’ve been following the blog since then, good stuff

It was clojutre https://www.youtube.com/watch?v=33mTFjPIN10&feature=youtu.be

ah yeah that’s the one

I am so enjoying writing the blog posts

😄 nice

I want the community to engage

you can call it publicity

when using the bootstrap compiler, evaluating "(ns foo.bar) [1 2 3]" ;=> [1 2 3] but evaluating "(ns foo.bar) 42" ;=> nil. Is this correct?

trying to respond to an ajax GET with a clojure data structure, but client receives ""...

using cljs-ajax.

if responding with a string, the client gets everything ok

using (str {...}) doesn't seem right

ok.. Cheshire seems to work on server side, and on client {:response-format (cljs-ajax/json-response-format) seems to do the trick

Does anyone know how to simulate a key press event using closure?

Google Closure

@andres-alonso: don’t know about closure library, but this is what I use for my tests: https://github.com/eventualbuddha/keysim.js

@darwin thanks!

@pat the evaluation of 42 to nil is related to the evaluation context. See: https://github.com/viebel/klipse#clojure-only-1

and https://groups.google.com/forum/#!topic/clojurescript/MzDFXdDyiRg

@pat expectations about the top level are unrealistic given we’re compiling to JS source files

absolutely no guarantee that the value of the last form in a ns that looks like that is returnable

and it’s perfectly fine that the REPL and the compiler might not even correspond here

@dnolen what about without a ns form? (eval-str "42") ;=> nil. My use case is simple scripts

@pat it’s not well supported yet

there are a few tickets floating around to make it work but it’s non-trivial

also simply - just not a priority - unless someone wants to deliver the patch

im looking at emit* for constants, skips when not statement

not complaining just want to understand

it’s not supported 🙂 this is the explanation

and it’s not supported because no one has worked on it

if you’re interested in fixing that then the right place for this conversation is in #cljs-dev

right

i’ve found myself in a situation where i have ~7 lines i’d like to add to each of my :cljsbuild :builds in my project.clj: https://github.com/jrheard/voke/blob/commonjs/project.clj#L34-L40

right now i’ve just copy-pasted them into each build, but that seems gross and bad

is there a saner way of saying: always have these :foreign-libs in every build?

i read through https://github.com/technomancy/leiningen/blob/master/doc/PROFILES.md , but it doesn’t seem like i can use leiningen profiles to accomplish this

i feel like there must be some sane + standard way of doing this sort of thing, so i figured i’d check with the channel. thanks!

@jrheard you can put them in a deps.cljs file in your classpath

huh! neat! are the semantics of this deps.cljs file documented somewhere? now that i’m googling for it, i see it referred to in a few pages on how to package libraries for consumption, but i’m building an app and this feature looks useful in this context as well