Has someone ever tried to "bend" code mirror and have a repl in the style of a terminal (like http://clojurescript.io) ? @jaredly

@richiardiandrea: http://jaredforsyth.com/reepl/ uses codemirror if that’s similar to what you’re imagining

yes, that's exactly why I am asking you 😄 basically I would like CodeMirror but with the familiar top-to-bottom terminal

it wouldn’t be too hard to change reepl to have the top-to-bottom style. just change the element that scrolls, really

because I am having problems with jqconsole and I was wondering if I should switch...still...i don't know if CodeMirror is reagent friendly enough

but you tell me that it is not difficult to change the style, well I might dig into it

I made a codemirror wrapper that sync w/ a reagent atom

oh

cool

I'll have a look

https://github.com/jaredly/reepl/blob/master/src/reepl/code_mirror.cljs#L129

I will either go with your solution or implement a custom re-console (I always propose it but everybody feels like reinventing the wheel)

have you seen the various codemirror-console libraries? http://kybernetikos.github.io/SoleMirror/ (and some others that I can’t find immediately)

Nice! I like jq-console because it has the classic feeling of being in a terminal

I guess that with some css-fu I can adapt code mirror to that style

I’m sure there’s a theme that’s green text on a black screen 😄

😄

Hey everyone, I was wondering is it necessary to learn Javascript prior to learning Clojure Script? It’s not my first language , I am a Java developer, but I rarely use Javascript.

@joshua.d.horwitz: it’s not necessary. you might run into some javascript things (just like clojure gives you java stacktraces) but the people here are super helpful 😄

It's good to know how JavaScript works though, like how its single threaded

I'd say it's good enough to Google your way out when your run into problems but that's still not ideal

Hmm, I guess it wouldn’t hurt to do some javascript learning quickly before

Hmm a "just enough JavaScript for clojurescript" tutorial would be nice

One thing I underestimated was how powerful the Google closure library was

So I've heard, but it's a very Java-ish, hardly documented style of Javascript unfortunately.

Most "typical" Javascript programmers would denounce it pretty fast.

I for one would appreciate a "just enough GClosure library for a Clojurescript developer" P ;

That is a great Idea for a tutorial!

I’ll help in anyway I can

After reading the state of clojure results I feel like there's a huge influx of people who want to try it out, and that we should do a better job to ease in adoption

one thing i'm curious about is what makes them stay

I think that the interactive development story is pretty strong

many people seem to struggle to get their environment up and running

Hopefully boot with tutorials like Modern Clojurescript helps with that

IMO boot's Clojurescript story is far less painful than lein-cljsbuild

I really liked mark mandel's idea of dockerizing everything

perhaps a container with working emacs + CIDER, or vim + fireplace

Or Cursive <3

Though I don't know how reliable docker is with X forwarding

My test project seemed to work well with docker though (with Cursive outside)

But I had problems running it on windows for example

And at least setting up Clojure projects is still simpler than Ruby or such

^ docker should precisely solve that

all the boot windows problems would dissolve

Should, but I couldn't even boot the image there

because they're essentially running it in unix

Though that might just be me not using windows for 4 years and not knowing what to do

That's probably not a Clojurescript topic to be honest

I started with cljs a couple months ago, and I’m using it for a client project now.

I stayed for the macros, and as you said, @jaen, the interactive development story

I ought to try boot; lein is definitely a pain point

@amonks: if you want to try boot, catch - https://github.com/magomimmo/modern-cljs/tree/master/doc/second-edition

@jaen: awesome! thanks!

Just noticed that (inc 3 12) returns 4 rather than an arity error. Is that a ClojureScript bug, or a reflection of some JS idiosyncrasy?

JS seems to handle extra params in the same way:

function myInc(n) { return n + 1; }
myInc (3, 17)
4

In the Speaking JavaScript book, I'm seeing that "JavaScript does not enforce a function’s arity: you can call it with any number of actual parameters, independent of what formal parameters have been defined...[given] more actual parameters than formal parameters, the extra parameters are ignored..." But I guess what I don't understand is whether there's a clear convention for when ClojureScript will throw an arity error and when it simply passes off to JS (in which case apparently there will never be an arity error).

I could be wrong about this, but I think CLJS just shows a compiler WARNING when it detects an arity error. As you referenced, the underlying JS will never complain when that happens at runtime.

@chrisoakman: I guess you're right. eg:

cljs.user=> (reduce + 0 '(0 1 2) :extra)
WARNING: Wrong number of args (4) passed to cljs.core/reduce at line 1
Invalid arity: 4
	 cljs$core$reduce (cljs/core.cljs:2147:1)

Still, no warning on (inc 3 12). So I'm still not sure whether there's a definitive cljs convention or whether it’s just a matter of how each function happens to be implemented.

Interestingly, if I define my own inc function, it throws a warning:

cljs.user=> (defn my-inc [x] (cljs.core/+ x 1))
#'cljs.user/my-inc
cljs.user=> (my-inc 3 12)
WARNING: Wrong number of args (2) passed to cljs.user/my-inc at line 1
4

(defn inc
  "Returns a number one greater than num."
  [x] (cljs.core/+ x 1))

@eggsyntax: There’s something odd with bootstrapped macros like inc as compared to if-not, probably worth tracking down. (What you are describing doesn’t occur in JVM ClojureScript.)

Is there a good way to see if the JavaScript environment you're in has a DOM?

In particular in the context of working with Google Closure on iOS react-native; in boot-react-native we're shimming window.document so that won't work.

@mfikes: oh, glad you noticed this. I should have checked to see if it was Planck-specific. Thanks!

@eggsyntax: Yeah, you can also see it in, say, http://clojurescript.io, which is based on bootstrap as well:

cljs.user=> (if-not true)
Could not eval (if-not true) - Invalid arity: 3 at line 1  - Invalid arity: 3
cljs.user=> (inc 2 3)
3

Heh, yeah, bootstrap-specific — didn’t mean to blame Planck 😉

Incidentally, Planck will give you

cljs.user=> (if-not true)
Invalid arity: 1

Dude, Planck is flawless. 💯

That last difference, (arity for macros), is the difference between 1.1.170 and 1.1.228. It solved the age-old issue of doing (max) an getting an invalid arity diagnostic indicating arity 2

Gotcha.

Cool, thanks! It’s been a good opportunity to go look at more of the cljs source code.

I may dig into the (inc 2 3) problem at some point—I recorded it in Planck’s issues. Maybe I can create a minimal repro, or even figure out the root cause.

I wonder if Planck signals the start of a totally independent self-hosted clojurescript.

as far as sentiment. I feel like we're a clojurescript version of lein away from it.

@gerred: you are thinking of the idea of using self-hosted ClojureScript to transpile for general targets, like the web?

yep, and Node.

remove the JVM dependency from the equation.

Can you extend existing JS types with ClojureScript protocols? I'm getting Extending an existing JavaScript type - use a different symbol name instead of js/String e.g string.

yeah, if you don't use :advanced, and don't rely heavily on macros, such a transpiler is probably becoming feasible. Macros in libs makes it challenging.

@jindrichm: yes but not native types directly because that’s problematic

@jindrichm: yes, use string

@jindrichm: follow the advice given by the warning

you lose a lot sadly not having the closure compiler.

uglifier and friends doesn't quite do it

so maybe it's not the right path.

@gerred: it’s certainly possible but as you say the tradeoffs are significant

yeah.

I know a lot of people, including myself at one point, have a pretty irrational response to anything that has anything to do with the JVM.

@jindrichm: Feedback on usefulness / accuracy of this proposed docstring change would be welcome: http://dev.clojure.org/jira/browse/CLJS-1550

@gerred: sure but no one involved in Clojure(Script) development is concerned about this problem at all

@gerred: We are still finding bugs in corner cases with bootstrap, FWIW

yup.

I'm thinking more along the lines of "how do we get more developers into this?!", and self-hosted probably isn't the correct answer, since they'll come JVM or not.

@gerred: For regular dev, JVM ClojureScript has the test of time on its side, definitely

@gerred: The question of JVM-less ClojureScript won’t go away, even if it is an irrational question. Hrm.

@dnolen: So I need to define a custom type?

@jindrichm: You can use string to essentially extend js/String to a protocol

@gerred: as far as I can tell Clojure(Script) usage grows at a fairly healthy rate every year - not quite enough that people don’t have staffing worries - but teams and companies that adopt Clojure consistently seem quite happy about it and stick with it. Not sure any other metric matters much.

very true.

Besides, a lot of people, including myself, have a partially rational response to anything that has anything to do with JS. If that hasn’t kept devs away, I don’t know if anything can 😉

tbh companies have staffing worries no matter the language

^

@jindrichm: just use string as @mfikes said

@mfikes: I see, thanks!

(defprotocol Foo (bar [_]))
(extend-type string Foo (bar [_] "foobar")
(bar "") ; => "foobar"

It is my opinion that someone in the community that is interested in a Node-based transpiler should give it a shot. We might learn the pros and cons. Can’t hurt. Whoever does it is likely to learn a lot about how the language ticks internally, which is a lot of what drives me to mess with Planck.

I actually came over to this side when I ceased being able to figure out what the hell was going on with ES6, Babel, Webpack, and everything else people are cobbling together into what they think a React app should be composed of/built with.

hell, ES6/ES7 are just huge languages on their own.

yes we’re also not anti-JS, Closure gives us the ability to integrate ES6 etc.

feels like a year ago I was happy with Gulp and a little React. 😉

oh yeah.

so we’ve been following along and will continue to do so

I think more what I mean is that the footprint of JS is growing at a far, far faster rate than ClojureScript's.

this like comparing Clojure usage to Java usage - this is non-information

everything we do is going to be drop in the bucket compared to #1 and #2 languages

was enough to make me interested enough / convinced enough to finally sit down and learn it though

but yeah, not useful as a basis for comparison

it made me actually enjoy writing JS again though.

so the personal value there is pretty massive.

any good references to learn clojurescript? video tutorials would be best.

@jcomplex: there was an O’Reilly series fairly recently (but I haven’t seen it myself so I can’t vouch for it). One downside is that it isn’t free.

@jcomplex: https://www.safaribooksonline.com/library/view/learning-clojurescript/9781771373890/

I think you can get access to it with the 30 day trial, and Safari books also has pretty much all of the Clojure books as well.

You can also go with http://www.purelyfunctional.tv/single-page-applications , but it's not free either - $25 / month for all his content though

I've also heard really good things about https://www.youtube.com/watch?v=aWJPlRVZX90

assuming you want to take on the burden of learning Emacs at the very same time

but yeah I’ve heard a lot of good things about that one too

would be cool to see a super beginner friendly video tutorial series using Atom & Parinfer

& Figwheel

& devcards

@jcomplex: If you have any specific things you really want to focus on or get stuck, the #C053AK3F9 channel is really helpful and is a great place to ask questions too (even if you think they're too basic)

@dnolen I use Atom, Parinfer, & Figwheel so I guess the Safari books are my options

@jcomplex: Unfortunately, they are not books, but videos.

oh videos my bad

A lot of the concepts should apply to all editors, so once you get a little more comfortable with the basics the others might be useful as well.

@jcomplex: yeah I looked around not too much out there specifically for ClojureScript

I've heard good things about Modern ClojureScript: https://github.com/magomimmo/modern-cljs

thanks for all the resources guys

@jcomplex: another approach might be to just use Clojure videos, there’s a lot more out there and lot of the concepts transfer

unless of course you just know you want to focus on ClojureScript

just focusing on cljs

@jcomplex: got it, yeah definitely a lot fewer resources - hopefully that will change in the near future with all the current interest

I am trying to get server-side rendering working and I get "navigator is not defined". Looking at the server-side.js file, it looks like react boots up like:

var Browser = {
  isDomPresent: true,
  navigator: navigator,
  window: window,
  document: document,

How would this ever work?

navigator is undefined (the default value for non bound variables) so {navigator: navigator} ==` {navigator: undefined}`

@chrisn you can defined nodejs globals, w/ global.navigator = null etc

then those variable references won’t die

@chrisn where are you finding that code? react is very server friendly — is that an old version?

It took me a long time to burrow down but basically we have something including quil that is including processor.js.

I mis-attributed the problem; I just got something rendering with react and now I am trying to pare down our system so I can pinpoint which namespaces aren't going to work server-side.

If you use (defroutes (ANY "*" [] (route-fn)) is there any way to get the request url in route-fn?

@chrisn I assume you are asking about compojure? If so, there is a :uri key on the requests that you can check:

(defroutes app
  (ANY "*" [] (fn [req] (ok (:uri req)))))