michaelr: this is common if you have split your app into multiple files -- it only reloads what has changed so if you don't change the file with om/root it doesn't re-render

you can add a function which figwheel always runs and put om/root in there or i believe there is some metadata you can add to the root ns to always reload that file

You add ^:figwheel-always to the namespace statement.

Got @maria’s CommonJS ClojureScript compiler support working with Ambly on a device for the first time! One less hurdle. If you are interested, recorded a demo of it: https://youtu.be/CS0RU2oUq7s

@nullptr, @akiva thanks

Well, it seems I am completely stuck with "ERROR: JSC_DUPLICATE_EXTERN_INPUT. Duplicate extern input" from lein. I only have one file with externs, referenced once in project.clj (as :externs). No matter where I try to place it, I get this message. And it seems to be a lein-cljsbuild problem, because I get it from lein uberjar, but lein with-profile uberjar cljsbuild once works and builds correct JavaScript. Sigh.

@jrychter: what is the name of the extern?

@thheller: the file, you mean? Currently :externs ["semantic.ext.js"], although I tried various locations for it.

https://github.com/google/closure-compiler/tree/master/externs

these are the externs added by default

I meant the name of the object you are writing externs for, not the filename

not sure what causes the error in closure but that would be the first place I check

@thheller: the file contains a bunch of lines of the form jQuery.fn.checkbox=function(){};. But I don't think this is the problem — things do work if I just invoke lein with-profile uberjar cljsbuild once.

haven't used cljsbuild for over 2 years, can't comment on that.

Exposition of how Ambly is experimentally using CommonJS lib support (towards React Native): http://blog.fikesfarm.com/posts/2015-06-17-commonjs-libs-in-ambly.html

has anyone experimented with using bootcards with clojurescript?

https://brendaneich.com/2015/06/from-asm-js-to-webassembly/

ah cool @shaunlebron !

@dnolen: Interesting.

Yes, we are aiming to develop the Web’s polyglot-programming-language object-file format.

Bottom line: with co-evolution of JS and wasm, in a few years I believe all the top browsers will sport JS engines that have become truly polyglot virtual machines.

this blog post is amazing - I'm currently watching the videos

@escherize: definitely a cool development but we’re definitely looking at at likely 3-4 year timeline before there will be anything worth targeting for ClojureScript

https://github.com/WebAssembly/design/blob/master/FAQ.md#is-webassembly-only-for-cc-programmers

sounds remarkably like ClojureScript’s compilation strategy 😉

anybody that’s ever submitted a patch to ClojureScript or even think they may one day, #C06E3HYPR is a good place to hang

@alexmiller just mentioned a nice patch to unify clojure.string across Clojure/ClojureScript

what are the consequences/gains of having clojurescript in clojurescript?

no consequences since it will always be optional

independence of JVM?

gains are you can now target places the JVM can’t reach easily (Electron/Atom Shell), hopefully write scripts w/ zero startup time over Node.js, show off your library with a REPL on a public website w/o a separate compilation web service.

I'm just asking the obvious maybe, but doesn't Google Closure compiler use a JVM?

yes so Google Closure will supported but it will need to be a separate step if you want to use it

however for the use cases I stated above, it doesn’t really offer much value

yeah, I got it. cool

@dnolen: we could achieve macros in cljs, don't we? Do you think there is value on that?

I'm curious about your experience

@andrewhr: a bootstrapped thing will never support that and in fact there will be safe guards to make it very, very hard to mix macros and runtime code.

because A) it’s seriously not important B) it would fork the ClojureScript world, i.e. uncompilable stuff for ClojureScript JVM

people will keep asking for it and the answer will always be “NO"

get used to it

hahaha got it

Does that mean cljs-in-cljs REPLs will have no support for macros? Or can they use macros but not create them?

yeah, community split makes everyone lose on the long run

@aengelberg: you will be able load macros just fine and even properly expand them

@aengelberg: none of safe guards will be unsurmountable

now our tools enforce us to the data > functions > macros spectrum

So were you just saying NO for "macros in cljs source code"?

it will just make people trying to mix code & macros into one source have to do ugly hacks

and every one downstream will have to adopt them.

so libraries that break the rules will be poisoned

@aengelberg: no the only thing we will prevent is mixing macros and runtime code into the same .cljs or .cljc file

macros will actually be present at runtime

we’re just going to have some static analysis to discourage mixing stages

that is as long as the macro files do not interact with the runtime files you’re fine - even if they both actually exist at runtime

this means you have something that will work for ClojureScript JVM and ClojureScript JS

it’s actually really important that this work

ClojureScript macros are not like Clojure macros, they can have the same names

cljs.core/+ macro, and cjls.core/+ fn can coexist

not possible in Clojure

outside of :inline construct which isn’t really the same thing.

Never understood that… how do you use the cljs.core/+ macro? Is it automatic?

since there are 2 stages (unlike Clojure), these things just never interact

you basically have separate non-interacting namespaces for macros and runtime stuff

Cool. Captured the naming aspect: https://github.com/clojure/clojurescript/wiki/Differences-from-Clojure#macros

Ahh… I think I got it: The analyzer never sees, for example defn. Macroexpansion occurs first, and after that, only specials or fns exist to be analyzed and then compiled.

If I’m understanding correctly (+ 1 1) involves the macro (at least initially), while (reduce + [1 1]) involves the fn.

mfikes: yep that’s exactly right

@mfikes: ^

@dnolen: Thanks! I’ll add some stuff to the Wiki to help others avoid meta-circular meltdown.

Did my best to capture this. Feel free to correct my inaccuracies: https://github.com/clojure/clojurescript/wiki/Differences-from-Clojure#macros

So is someone writing a clojurescript package for atom now?

Sounds like it would be easier than what the LightTable people had to do?

@mfikes, thanks for asking that question, I never made that connection with the naming conflicts between the core macros and functions

@shaunlebron: Yeah. I suppose it caused a question or two when you are generating your doc tooling and you see both.

@mfikes: looks good

yeah, I just include both sources, but default to the function implementation: https://github.com/cljsinfo/api-refs/blob/catalog/refs/cljs.core_PLUS.md

@shaunlebron: I bet that’s safe. If you have a name collision then it is probably only sane if the macro and function have roughly the same semantics.

This could be beneficial. In ClojureScript, you could support something like (map or xs) (ignoring what that really means for now).

In other words, if you define a macro foo and feel compelled to be a good citizen and provide a foo* function for use in cases where a macro won’t do, perhaps a foo function could be defined.

is this slack channel being archived somewhere? I feel like there’s a lot of good information here that I would like to search through someday

ugh, there’s a 10,000 message limit, but that’s a sum limit across all clojurians channels

@shaunlebron: They all are. I’m looking for the links right now.

FWIW, I’ve been trying to take clarifications, etc., and incorporate them into the Wiki.

https://clojurians.slack.com/team/logbot

@mfikes: yeah that’s really the best idea, thanks for doing that

@akiva, nice thanks: http://clojurians-log.mantike.pro/clojurescript/index.html

@shaunlebron: I suppose only an extremely small subset of the info here really is suitable for the Wiki. (Things that the Wiki covered, but perhaps not clearly or comprehensively.) But, you are right that some odd side things are best just being searchable.

@shaunlebron: Not a problem.

love the integrated google search, someone gave this some thought

I’m pretty sure @ul’s behind the curtain.

that’s awesome @mfikes thank you!

@akiva, added a wiki link to the chat logs, could be useful: https://github.com/clojure/clojurescript/wiki#chat

I wonder if this has supplanted irc

I’ve heard that it has died down there somewhat.

that’s hard to say, there are as many lurkers on IRC as before Slack, ~200

so looking forward to porting this to ClojureScript http://www.amazon.com/gp/product/0262527952/ref=ox_sc_act_title_1?ie=UTF8&psc=1&smid=ATVPDKIKX0DER

@dnolen: How is the port of core.logic to ClojureScript going?

@markstang: it works, I haven’t touched it in a while and probably won’t have time to any time soon.

I’ll probably revisit a little bit when CLJS can compile itself as an online CLJS mK would be fun

@akiva: Do you know if Log Bot is a thing that can be leveraged elsewhere? I’m working with another group that’s considering using Slack, but people there really want logging and this looks like it might help a lot.

@dnolen: I am sure the “Dr’s” would love an online ClJS mK. I think they were asking about it a while back...

@nicmcphee: I’m pretty sure it could be. It just needs to be installed and configured.

https://github.com/kidach1/slack_bot

@akiva: That rocks – thanks for the pointer!

My pleasure.

Is there a way (or roadmap) to get the terminal to recognize clojurescript test suites failing as a nonzero exit code?

Currently CI systems will accept it even if they do lein cljsbuild test and get failures on the node output e.g.

I'm not even sure which component (clojurescript, cljs.test, lein-cljsbuild) can be improved to make that work properly...

Is there a good reagent example of a reusable dynamic list component, something where items get added/updated/removed, and the values have to bubble up to the parent container? Specifically, is there a clean way to update changes in a vector without contaminating the line components with parent state?

@aengelberg: this seems very specific to the runner you are using and very little to do with ClojureScript. When your tests are done invoke the proper exit (with code).

that's what I'm curious about... how to exit with a code?

maybe there's a way to do that specific to node.

http://stackoverflow.com/questions/5266152/how-to-exit-in-node-js

http://phantomjs.org/api/phantom/method/exit.html

stuff like this

Sweet. So that would be translated to (.exit js/process 1)?

yes

thanks

Just asked in #C0617A8PQ, perhaps more relevant here; do cider and the figwheel repl play nicely, or do people tend to drop back to inf-clojure?

still haven’t heard many confirmed reports of it working using all up-to-date dependencies

Righto, thanks.

inf-clojure definitely works as does anything that supports inferior-lisp like interaction, Cursive and I assume Sublime

Yea, inf-clojure does the trick.

Can't figure out why my defmethod isn't working here (per cljs.test wiki instructions).

cljs.user=> (require '[cljs.test :as t :refer-macros [deftest]])
nil
cljs.user=> (deftest my-test (t/is (= 1 2)))
#'cljs.user/my-test
cljs.user=> (defmethod cljs.test/report [:cljs.test/default :end-run-tests] [m] (println "Look what I found" m))
#<[object Object]>
cljs.user=> (t/test-ns 'cljs.user)

Testing cljs.user

FAIL in (my-test) (at /home/alex/git/instaparse-cljs/.cljs_node_repl/cljs/test.js:384:41)
expected: (= 1 2)
  actual: (not (= 1 2))
nil

@aengelberg: try t/run-tests instead of t/test-ns

cljs.user=> (t/run-tests 'cljs.user)

Testing cljs.user

FAIL in (my-test) (at /home/alex/git/instaparse-cljs/.cljs_node_repl/cljs/test.js:384:41)
expected: (= 1 2)
  actual: (not (= 1 2))

Ran 1 tests containing 1 assertions.
1 failures, 0 errors.
nil

https://twitter.com/id_aa_carmack/status/611279873852731392

Doesn't work on nashorn repl either

cljs.user=> (t/run-tests 'cljs.user)

Testing cljs.user

FAIL in (my-test) (at <anonymous> (.cljs_nashorn_repl/goog/../cljs/test.js:NaN:384)
expected: (= 1 2)
  actual: (not (= 1 2))

Ran 1 tests containing 1 assertions.
1 failures, 0 errors.
nil

@aengelberg: what version of ClojureScript ? 0.0-3308?

pretty sure this works as it was added by people who needed it

3165... ick

How are the version numbers determined? I'm just curious since they look rather arbitrary

Now I'm getting:

Exception in thread "main" java.io.FileNotFoundException: Could not locate cljs/util__init.class or cljs/util.clj on classpath: , compiling:(cljs/closure.clj:1:1)

or I need clojure 1.7.

yup, that's it.

https://github.com/clojure/clojurescript/commit/87f39511a54d7b91ae86f581fda90f280f985816

@aengelberg the version number is the number of the commit

aha. thanks.

AMD Module processing landed thanks again @maria

@maria is on a roll! 🎢

Wow, that was fast, I just uploaded the patch 20 minutes ago…thank you for merging @dnolen

np

Probably a question that has been answered multiple times by now: does AMD/CommonJS support mean we don’t need externs when we have JS in those module formats?

https://twitter.com/mayaneise/status/611214781878374400

What does “might” mean in that context?

you will need externs if the library is going to do something dynamic w/ strings & Object properties

fortunately that doesn’t seem quite so common as it once was at least for major libraries

Ah, I now remember reading that in the scrollback some day, thanks

@martinklepsch: Here is a simple example for when you would need externs:

greeting["bye"]  = function(name) {
        return "Bye, " + name;
    };

And also reading this might help to understand when you need externs: https://developers.google.com/closure/compiler/docs/limitations

> The Compiler renames properties in Advanced mode, but it never renames strings.

@martinklepsch: thanks to the vagaries of JSPerf abuse there’s now a useful myth that .foo is faster than [“foo”]