@mfikes: can't seem to call that ns-interns macro with cljs-bootstrap: https://gist.github.com/tolitius/c17b2d4a98eabb591926 (neither via cljs.js nor via replumb/read-eval-call), but I maybe missing something. I see that *load-macros* is true by default

@tolitius: Need to keep state around. Commented: https://gist.github.com/tolitius/c17b2d4a98eabb591926#gistcomment-1639027

@tolitius: In other words, your (def foo 42) is being done in “regular” ClojureScript, but then you are passing an empty compiler state to cljs.js.

hi, i started working on the om tutorial https://github.com/omcljs/om/wiki/Basic-Tutorial, its up and running, but the figwheel repl doesn't come up after i've refreshed the browser, thanks

@mfikes: so that is why it's an atom thanks a lot, makes sense!

how ready is om next to be used in production?

Figwheel: trying to open cljs reload socket socket.js:62 WebSocket connection to '<ws://localhost:5060/figwheel-ws/dev>' failed: Error in connection establishment: net::ERR_CONNECTION_REFUSED

@tolitius: Yeah. I suppose you can use cljs.js like that in a regular ClojureScript REPL, but in a deployment scenario you are dealing only with the state atom, essentially. If you need to load analysis metadata for stuff that is compiled and ships with your app, I suspect you can (it has been a little while since I’ve looked at that code).

@mfikes: did not occur to me how the pieces are used first time I looked at it, but it is coming together

@mfikes: yea, I believe meta/vary/alter! are supported, but will find out

@tolitius: It appears they all work in Planck (a pure bootstrapped env), FWIW.

@mfikes: nice, thanks for checking

@tesseract: there is an #C06DT2YSY specific channel. Answer not quite, but we’re working steadily towards beta

people are building out production systems with the knowledge that beta is near

@brw314: you also might want to direct your questions to #C06DT2YSY

is there a way to interact with clojure/jvm from the figwheel repl? or for figwheel watchers to produce output to that repl?

the only way i've found to produce output is by throwing an exception

Could you please let me know what is this error trying to say?

it may be raise from below go block. (go (<! loaded-channel)...)

Why do I get “Uncaught TypeError: Cannot read property 'appendChild' of null” when trying (defonce conn (repl/connect "http://localhost:9000/repl”)) ? I’m running ring/compojure to serve HTML and I have started ClojureScript REPL with "lein trampoline cljsbuild repl-listen"

are macros impossible to implement in javascript?

why do they have to get compiled from the jvm?

@tmtwd: because of google-closure compiler. It removes everything that is not used

is there a way to iterate through my project's namespaces in clojurescript?

it doesn't have to be at runtime, i just want to build a tree of my namespace and a given value that each of them provides

currently i'm requiring them individually and building the tree by hand, but this is going to get pretty tedious as the project grows

to be a bit more concrete, say I have three namespaces 'foo.one', 'foo.two', 'foo.bar.one' and each of these have an 'info' symbol. I want a map {:foo {:one foo.one/info :two foo.two/info} :bar {:one foo.bar.one/info}}

I can't know for sure since I never tried, but you could see if the analyzer API is available during macroexpansion - https://github.com/clojure/clojurescript/blob/master/src/main/clojure/cljs/analyzer/api.clj#L152-L158

I don't know if it is though.

thanks, I will look into that

meh, no such namespace 😕

You tried it in a clj file, right? Then I guess tough luck. Unless someone more experienced with that comes along and has a brighter idea.

yep, a clj file that I load with :require-macros

grepping around on github, there seem to be people doing something similar with all-ns in clojurescript, e.g. https://github.com/darwin/plastic/blob/7c81a4fcc0bb53c25e1c1bbe5d41e23913ddbfcd/src/test/plastic/test/karma.clj

i'm afraid that's a bit too advanced for day #3 of using clojure/cljs .)

Yeah, I imagined it would work like that, not sure why it doesn't. But then again I never tried it in practice, so there might be something else you need to do I'm unaware of.

hah, had a mistake in my macro... i syntax-quoted the outer form and thought I'd deleted the backtick. turns out there was some dust on my screen over that exact position :3

works

wow, that is really neat ❤️ that's got to be the easiest way of metaprogramming i've ever done. doing that with python or java would have taken dozens of lines of code

Haha, glad it's working then

Btw, is there a way to see the expanded macros in clojurescript?

I can only see the cljs with macros and the generated js in the out directory

Hey, @jaen, do you remember that gist you sent me

@trancehime: I dunno, what was it about?

tables

containing data from an atom which has vector

@jbaiter: I guess you could just use the usual Clojure macroexpand since it's Clojure code.

@trancehime: ah, you mean this? - https://gist.github.com/jaen/dd22b3432581fc7ab8f5

yep

what is .. in [new-value (.. e -target -value)]

i glossed over that last time

it's probably a reference to something but not sure what

You know what (.-property object) does?

interop with js to get a property of an object? iirc?

Yep

So .. macro is another layer of syntax sugar over it

oh .. is a macro

It just expands to (.-value (.-target e))

So you don't have to nest too much

Ahhh.

I think you can do even fancier things, like (.. e (function arg) -property-on-result) which should be eqivalent to e.function(arg).property-on-result

But Id' have to check it in repl to be sure

Every element in a seq should have a unique :key wat

but.. pretty sure there is

oh wait the keys are like "this" not like :this

pls

Hmm, no

What it means is

The thing described here - https://facebook.github.io/react/docs/multiple-components.html#dynamic-children

If you have a dynamic list of children react has an optimisation to reuse existing DOM to make the updates faster

You can tag each component with a unique (in the scope of a list) key to make use of that

In the example from gist it would look like this

(fn [idx {:keys [name min pos points]}]
  ^{:key "your-key-here"}
  [:tr ...])

Where ^{} thing means adding Clojure metadata on the vector.

Reagent uses the metadata to set the React keys on components.

yeah, i did ^{ } for some other stuff

In this case something like (str "score-row-" idx)would be a good key, I guess.

Unless you have some server-based id, then use that instead, I guess.

my problem is my data is keyed {"like" "this"} so doing it your way makes the thing behave weird even if I have server-based id

it makes it believe all the children have "id" which makes flattenChildren throw a warning

So far I haven't run into any problems until now...

Hmm, can you show the code?

That should not happen I think

derefing the atom gives me a vector of data which each element looks like this: {"end_level" 18, "schedule_game_detail_id" 157356, "id" 760501, "deaths" 1, "win" nil, "kills" 1, "player_name" "Huni", "schedule_game_id" 78749, "assists" 1}

> Huni Hehe.

That looks legit, let me think

In the mean time you might appreciate this:

(defn transform-keys
  "Recursively transforms all map keys from strings to with given function."
  [m f]

  (walk/postwalk (fn [x] (if (map? x) (into {} (map f x)) x)) m))

(defn api->keyword-key [[key value]]
    [(keyword (str/dasherize (name key))) value])

str is this lib - https://github.com/funcool/cuerdas

But yeah, I think I misunderstood you when you said "oh wait the keys are like "this" not like :this", I thought you were referring to the metadata map.

In that case you probably were right - you're using :keys instead of :strs in the destructuring

So your id is nil each time, because you indeg by :id not "id" and all rows end up being mapped to the same single row.

I suggest you try the code I've pasted above so you can get nice, dasherized keywords keys instead

(`walk` is clojure.walk)

@jaen thanks, I will try that when I can. Right now I have to take care of some chores now that I'm back home. And my sister is still sick.

Sure, feel free to ping me how it worked out.

If namespace foo.core depends upon bar.core (via a :require and a fn call), and I change bar.core’s on-disk implementation, should I see evidence of those changes if I do (require ‘foo.core :reload-all)? (Say by calling a fn in foo.core that then calls a revised fn in bar.core?)

yes

:reload-all effectively forgets anything that has been loaded

@dnolen: Cool. I think I may write a minimal ticket then. (Not working for me for a very simple case with script/noderepljs.)

@mfikes: whether it works or doesn’t work is REPL dependent

and all of them manage this by tracking the loaded set in the JS environment via Google Closure

goog.dependencies_ specfically

sometimes this breaks because GCL changes the details

gmorning, Clojurists! Can anyone pointout resources for debugging go loop bugs via sourcemap in browser, or if there’s an issue already in progress on it?

@dnolen: Ahh… I think I see stuff you did specifically for the browser REPL for :reload-all. I’ll start there then. (I want to make it work in Ambly.)

@wildermuthn: nobody has ever picked up that problem

great thing for somebody in the community to work on

I’d be happy to dig in with a little pointing in the right direction

Should I start with sourcemaps? Learning how that works?

@wildermuthn: nothing to do with source maps at all

cljs.core.async needs to be fixed to use tools.analyzer a la clojure.core.async

this work has never been done

Currently, it uses something else, or just uses tools.analyzer differently?

cljs.core.async still uses a buggy hand rolled analyzer

@wildermuthn: I don’t have any more pointers to give you

Ok, thanks for pointing the right direction

you’ll have to poke around, the person to talk to is Timothy Baldridge

:thumbsup:

Hi, I'm found some trouble when trying compile some example code with :advanced/:simple compilation modes but I'm always getting this error: No implementation of method: :-find-sources of protocol: #'cljs.closure/Compilable found for class: nil, the same configuration but changing the :optimizations to :none everything works correctly.

The concrete build script is here: https://github.com/funcool/catacumba/blob/master/scripts/build-postal-chat.clj

http://blog.cognitect.com/blog/2015-clojure-community-survey

^^ would love to see your responses! thanks...

@alexmiller: just FYI, when I press next on page 2 on an iPad I get back to page 1.

@martinklepsch: that happened to me too, I had to go back and click next again \cc @alexmiller

@alexmiller: Page 2 answers didn’t save for me.

@alexmiller: I filled it out!

@dnolen: It did that for me multiple times, I'll try again later with proper computer

I got sent back to page 2 as well, from my laptop

well I stopped changing it so if that's the problem it will be fixed!

Wait, so changing values in place can lead to problems? Who would have figured! 😉

well my changes were additive, so there was no reason they needed to force a re-do :)

Worked on first try with laptop

@martinklepsch: I restored my old build.boot file for some of my projects and I can confirm that now the speed is much much better in almost all tasks.

@dnolen @mfikes Would it make sense to open a cljs-bootstrap Slack channel?

@richiardiandrea: I don’t think so

unless it’s about people with various bootstrap projects that want to chat

actual bootstrap questions should be here or in #C07UQ678E

TIL there's such a thing as #C07UQ678E.

@dnolen yes I was more thinking about projects

Maybe with a different name

@richiardiandrea: yes that makes sense to me

Ok then, I will try to come out with a better name, and wait for @mfikes opinion as well

Sometimes I’m a little bit concerned if we get off into the weeds about bootstrapped stuff. But, it is a part of ClojureScript proper now.

We could end up with too many channels if we aren’t careful

Well on the official wiki there are I think ten bootstrapped project in total, I just was thinking that it might be useful as a point of reference

But again, the community opinion counts more and that's why I was asking :)

I was starting to think we might need another channel, but I hadn’t reached that level of pain.

@mfikes I fine with people talking about it, it’s a very useful thing for people who are experienced with ClojureScript

some people seem to want to jump into the deep end, but there’s not really much that can be done about that

but a room for people who maintain bootstrap projects might make some sense

since that discussion is more focused on sharing ideas, approaches to building generic stuff

Yes that's true, and ease the pain for us as well, for instance today I wanted to check how it ended up with ns-resolve and I had to scroll up quite a bit. Let's keep it under the radar, if questions increase a bit, we will create it I propose...

Hm, this reminds me of a random idea I've had. I've re-played Ruby Warrior recently and thought it's a pretty cool way to introduce people to a programming language. I figured doing something like Champion of Clojure would be awesome and with bootstrapped Clojurescript it would actually also be possible.

Yes analyzing code and comparing it to (optimal) solutions was something I was thinking of for http://clojurescript.io, then I went for something simpler...

Heh, playing Ruby Warrior on intermediate made me appreciate how Clojurescript is a first-class implementation. Ruby Warrior is (I assume) using Opal and some common idioms don't Just Work™ (nothing against Opal guys though, I can imagine what a pain implementing Ruby must be with it's schizophrenic non-spec).

Hey everyone im trying to check a url and having problems getting a window function to work. Any ideas? (re-seq #"/#/login$" (.href js/window.location))

you need (.-href js/window.location)

(.href …) would be trying to call that function with no args

@noonian: ahh thanks!

np!

@cfleming: any reason when I run a ClojureScript task or REPL from Cursive that I get a target directory with a bunch of stuff in it?

this creates lots of annoying problems since target/classes is often on the classpath implicitly