oops.. now I did ctrl-c in REPL and now even rlwrap java -cp cljs.jar:src clojure.main repl.clj doesn’t work

Exception in thread "main" java.io.FileNotFoundException: Could not locate clojure/browser/repl__init.class or clojure/browser/repl.clj on classpath., compiling:(/Users/Chico/Projects/om-test/repl.clj:1:1)

chicoxyzzy: double check that your cljs.jar is an ~18mb file, common mistake to download the wrong thing

@chicoxyzzy: ^

@chicoxyzzy: otherwise make sure you don’t have a global classpath setup with some version of Clojure on it

I didn’t change it anyhow when completing previous steps of tutorial and it looks fine 19842263 Jul 1 22:28 cljs.jar

@chicoxyzzy: seems very weird, but the error says different

yeah, one of the main features of clojure/clojurescript that has kept me happy is the community. There is a lot of mutual respect between library authors and no feuding, compared to other communities.

@chicoxyzzy: you will only get that error if the classpath is wrong, or somehow the wrong version of Clojure is getting loaded

the only two possibilities really.

@chicoxyzzy: and reading back - namespaces map to Java packages more or less and those must match the classpath

no picking names that don’t match the directory structure

ok thank you. I’ll try to fix my issue and change my namespace to om-test.core

directory structure needs to be src/om_test/core.cljs, dashes become underscores

CLJS compiler question: When it analyzes an ns declaration, does it begin emitting JS immediately or does that occur after the entire file has been analyzed?

cool. just created new project from scratch and all works fine now. than you @dnolen

https://github.com/clojure/clojurescript/wiki/Rationale#javascript-engines-gain-power I think we should add some info about forthcoming Web Assembly there

yeah

@domkm: See very end of this post as a guess at an answer: http://blog.fikesfarm.com/posts/2015-06-15-see-js-in-cljs-repl.html

(I think the answer is “right away,” depending on what you are really after.)

@mfikes: Ah, interesting. Thank you.

@mfikes: I’d like to write a macro that can be used from cljs to alter the current namespace. It seems like that might not be possible or might be too prone to breakage by compiler changes.

@domkm: ns is just a special form. It can be used in the middle of a file, if that's what you are thinking.

@mfikes: Hmm, I thought ns was a macro. Is it a special form in cljs?

@mfikes: I’d like to be able to alter a cljs namespace in a similar way to how potemkin can import-vars. I think I have it mostly worked out except for namespace alternation because cljs doesn’t have require.

@domkm: If you write a macro that involves ns, I suppose you'd have to be sure it comes back into ClojureScript after expansion.

@domkm are you looking just to import vars like potemkin? Or something more?

@danielcompton: Yes, import vars

@domkm: ns is a macro in Clojure, but a special form in ClojureScript.

@domkm https://github.com/Day8/re-frame/issues/62

@mfikes: Interesting! I didn’t know.

http://side-effects-bang.blogspot.com.au/2015/06/importing-vars-in-clojurescript.html

Thanks. Unfortunately that just defines new vars which means that they have the wrong metadata. It also doesn’t work for macros.

Ah rats, I hadn’t tried it yet

@domkm: (doc ns) will work in the next release. A preview of the output (showing it is a special form) is in http://dev.clojure.org/jira/browse/CLJS-1307

question on using cljs-ajax if someone has time ..

http://screencast.com/t/6T4ViLZN

i have a liberator delivering a json api (above) and I am trying to make a ajax get call to load it into an atom (for looking at in figwheel repl) and printing to log …

i I load the page in the browser.. in figwheel repl I run (get-data) and this is the output

Figwheel: trying to open cljs reload socket
utils.cljs:40 Figwheel: socket connection established
ioc_helpers.cljs:41 Uncaught Error: No protocol method ReadPort.take! defined for type null: 
xhrio.js:561 XHR finished loading: GET "”.

i see in the network requests it accesses the url and response is data

are you using lighttable

i am using emacs

sorry missed the screencast link ( I usually do) heh

i look at the network tab and see my json data is Content-Type:text/plain;charset=UTF-8 .. maybe that has something to do with it ..

@clojuregeek: looks like something to do with core.async

ioc_helpers.cljs is a core.async file

i’m not using core.async for loading the data

cljs-ajax is though

ok

do you know where exactly the error is thrown? Inside get-data, or inside handle-response?

let try again, i had typo in my handle function.. i’ve been trying different versions

hm got something back..

http://screencast.com/t/9ohzC75g6

ok i think i’m getting back a string I need to parse to json.

@clojuregeek: cljs-ajax has some helpers to do this for you

oh so dont need cheshire ?

@clojuregeek: shouldn’t need to, https://github.com/JulianBirch/cljs-ajax#json-specific-settings

ah that works

ok thanks so much.. i better get to bed now before i get in trouble with hubby

np

Last night I converted bidi from cljx to cljc, following the 1.7 announcement. However, I'm struggling to persuade lein to pick up the .cljc files in testing - is there an approach to testing that people are using that is working for them? Could really use some help here

@malcolmsparks: I think there's some open issues around that, though if you pass particular files/namespaces (can't remember which) they'll be tested

To lein test that is

http://facebook.github.io/graphql/

I have a cljs project which was set up with chestnut

When I deploy to heroku with :optimizations advanced, it fails with cannot call .. on undefined,

but it does work with :optimizations :whitespace

https://github.com/slipset/moon/blob/master/project.clj#L78

it fails at this line https://github.com/slipset/moon/blob/master/src/cljs/moon/components.cljs#L134

which again calls (why, I don't know) goog.string.format.demuxes_["f"]

@slipset sounds like you're using a library with a bug. You need to find the source that makes that invocation.

From looking at the code, it seems as if it is the js reader thingy which is causing me problems, since I'm passing js-opts to an om-component,

The value of the key :now will be a float, which I guess has to be stringified so it can be passed to javascript.

What js reader thingy?

I'm pushing a commit with :optimization :advanced right now, so I can give more exact info

Om doesn't use goog.string.format nor does ClojureScript

in om, when you have a component you pass a first argument which becomes the attributes of the html-element, right?

sorry, not component

the thing that is nil here: (dom/p nil (:text data))

What you are talking about now seems unrelated to your bug

ok.

You said the error shows a call to format. You need find what code calls format. Nothing else is relevant really.

Ok, Seems like the error is slightly different today than it was last night.

I'll have to continue debugging, thanks!

https://twitter.com/ustunozgur/status/616519185108054016

🍷

@dnolen: Nice!

Using dependencies from http://cljsjs.github.io/ … Is there anything else you have to do after adding the dependency to project.clj and :require 'ing the namespace in a .cljs file?

@stuartsierra: nope, that should be all

perhaps an issue with the packaging on the cljsjs artifact? externs issue?

@martinklepsch: looks like your boilerplate metadata is showing 😉

Specifically, loading cljsjs.d3 causes "Uncaught SyntaxError: Unexpected token ILLEGAL" in the JavaScript Console. I know this is an issue with D3 in general — the JavaScript sources use UTF-8 characters.

The recommended solution is to make sure d3.js is sent to the browser with UTF-8 encoding.

Figwheel is sending the JavaScript to the browser.

stuartsierra: Just a stab in the dark, Are you using the <meta charset="UTF-8"> tag in your header?

@bhauman: No. Also not my HTML.

maybe try placing it in the head of your html. and see if that helps.

not sure it will help

@bhauman: Looks like that might have fixed it. Thanks!

At this point I think that should be on every default template. Adding it to the figwheel template now.

My standard index.html file, fwiw:

<!DOCTYPE html>
<html lang="en">
    <head>
        <meta charset="utf-8">
        <meta name="viewport" content="width=device-width, initial-scale=1">
    </head>
    <body>
        <div id="app"></div>
        <script src="app.js"></script>
    </body>
</html>

Is <meta charset="utf-8"> a real thing? I thought it was <meta http-equiv="Content-type" content="text/html; charset=UTF-8">

html standards have gone through a lot of changes

I guess I do old-fashioned HTML.

@stuartsierra: yep 😉

At the rate its moving I think we all do old fashioned html

the html I posted is the minimal amount to be compliant with good "responsive html" practices

pkobrien: thanks I'll take it

html5, language set so translations work, charset, viewport for mobile devices, etc.

except the body elements are for reactive apps

I still don't use any of the new html5 "semantic" tags, it just never occurs to me to use them.

header, footer, main, aside, etc

(html
   [:body
    [:header {:role "banner"}
     [:nav {:role "navigation"}]
     ]
    [:aside {:role "complementary"}
     ]
    [:main
     ]
    [:footer {:role "contentinfo"}
     ]
    ]))

@meow: How do you paste code like that in Slack?

my advanced compilation ClojureScript app.js file is quite large. How can I reduce the size? Is it core.async that's eating that much space?

(defn test [a b c])

Aha

@cfleming: you got it 😉

@dnolen: 0.1.56 just landed with a bunch of CLJC fixes. Still more to do but is hopefully more usable now. If you’re still having problems I’d like to know about them.

@petrus only the gzipped size matters

Size of uncompressed source - irrelevant

@cfleming: will give it a shot soon!

Hi. Is lein new mies still the recommended template? I saw no mention of it in the quickstart.

@shriphani: I’m actually thinking about deprecating them, don’t have any time to maintain them anymore

and Quick Start is a better starting point

if I stick a transit-encoded map in a goog.net.XhrIo, it appears the xhr escapes my transit string again. Is there a way to avoid that?

@arohner: really? cljs-ajax and cljs-http must have sorted that out, I would look there.

yeah, I’m not seeing anything special yet. It seems like they just do the expected (t/write (t/writer :json) x)

amazingly, the double escape thing only happens on some websites and not others

is there a way to make an xhr call block? I’m getting an error when I :refer <!!

@roberto: no. JS is single threaded, so you can’t block

okay, a better question: has anyone used clj-http with re-frame?

err, cljs-http

@roberto: I’m sure plenty of people have. You probably want (go (<! chan))

in cljs-http callbacks, put results on channels. Elsewhere, take data from channels, and update UI

yeah, that is what I’m using

but not sure why updating the state isn’t triggering the subscriptions

@dnolen: my double escape thing, which I know sounds super weird and sketchy 1) happens on some websites and not others 2) doesn’t appear to happen when using (t/writer :json-verbose)

I suspect the site is monkeypatching js/String or something

@arohner: this not something we’re going to address at all

@arohner: you’re the first person to report this issue so I don’t think it’s an actual problem

not expecting you to. It’s just weird and I’m trying to understand it

`cognitect.transit.write(cognitect.transit.writer("json"), 3) ""[\"~#'\", 3]”"`

different site:

cognitect.transit.write(cognitect.transit.writer("json"), 3) "["~#'",3]"

anyone able to point me in the right direction

if I am trying to convert an array of JS maps

to a clojure coll of maps

@samueldev: I think js->clj may work for you

http://www.spacjer.com/blog/2014/09/12/clojurescript-javascript-interop/

thank you @johnmendonca

so when it comes to working with Om

as a means to learn clojurescript

should I ask what essentially amount to "beginner Om/CLJS questions" here, or in #C03S1L9DN, or in #C053AK3F9

?

or in #C06DT2YSY*

@samueldev: whatever gets the job done 😉

I’d say actually #C03S1L9DN and #C06DT2YSY.

currently im using cljs-ajax

for my ajax requests

and its response objects are huge and seem to have the data nested inside somewhere

trying to figure out how to access those 😞

I use that as well but I get a map back.

And they are bloody huge.

(def app-state
  (GET ""
    {:handler (fn [response] (.log js.console response))
     :response-format :json
     :keywords? true}))

response logs out to this object: http://puu.sh/iLbcC/15819d9c9c.png

Are you trying to use JavaScript to process it?

and i need to access (in JS syntax cuz im a noob) response.tail.arr[0]

@samueldev: you don’t want to debug that, you should print it to a ClojureScript thing

im trying to get the contents of response.tail.arr[0] in the form of a collection of clojure maps

@dnolen: how?

this is my biggest issue right now i think, figuring out the "clojurescript" way to debug this stuff

ha

from the JavaScript Console cljs.core.pr_str(foo)

whatever local you are trying to print

I don't think I have access to it like that

or rather I do, because its in my browser, but I don't know where

just type that in the JavaScript Console

I understand that haha

I mean foo in your case

how do I assign a local to the window so I can know what to log out?

you need to set a break point in your ClojureScript source

Hah, I keep typing and then I see @dnolen typing and I’m all, ‘I defer.'

😛 every piece of your advice leads me to more questions... gonna use my googlefu before asking lol

@samueldev: Are you using Figwheel yet?

aye

lein new figwheel project -- --om

Okay, good, good. The latest Figwheel allows you to set breakpoints as @dnolen just described. At the same time, you should through the Figwheel REPL be able to query defs and so forth.

Uncaught SyntaxError: Unexpected token debugger

-_-

(def app-state
  (GET ""
    {:handler (fn [response]
                (.log js/console (type response))
                (js* "debugger;"))
     :response-format :json
     :keywords? true}))

nvm I wasn't thinking about how this would translate to javascript

first thing after params in a cljs fn is a return, so that debugger would never get hit

sweet! progress

so I get this when I run cljs.core.pr_str(response) after my breakpoint

http://puu.sh/iLcCX/1f62d8b453.png

which appears to be a valid vector?

or am I crazy

given that that's the case, should I not be able to iterate over it and build some om components with om/build-all? all of my questions derive from the first issue I ran into:

just used eval'ed some bootstrapped ClojureScript from a browser REPL

getting the following when I run om/build-all...: Uncaught Error: [object Object] is not ISeqable

complete context:

(defonce app-state
  (GET ""
    {:handler (fn [response]
                (js* "debugger;")
                response)
     :response-format :json
     :keywords? true}))

(defn project-view [project owner]
  (reify
    om/IRender
    (render [this]
      (dom/li nil (str (:name project) ", ")
        (dom/a #js {:href (:github-url project)} (:github-url project))))))

(defn projects-view [data owner]
  (reify
    om/IRender
    (render [this]
      (dom/div nil
        (dom/h2 nil "Project list")
        (apply dom/ul nil
          (om/build-all project-view (seq data)))))))

(om/root projects-view app-state {:target (. js/document (getElementById "app"))})

in other words om/build-all can't iterate over the PersistentVector? is my understanding?

samueldev: your app-state is borked. GET is async it’s not going to return anything meaningful right away.

gotcha

@samueldev: ^

do I need async channels to hold offon app initialization of the component then

you need something

word

figured as much

welp

if I throw my entire om component definition and rendering code into the callback of the cljs-ajax call

it works fine

so you were right and I was just making assumptions about my error

(the assumption being that "oh, object object isnt iseqable? that must mean that the http call IS finished, and its the response format thats the issue")

oh hey you wrote the damn thing @dnolen

thanks for helping a noob out

no problem