@dnolen: Yeah, I assume they want the rights to it. (Nothing on their website, but I’m assuming that’s the case.)

@dnolen: Anyway, yeah. Awesome talk!

@mfikes: was fun to demo code I got working yesterday 😛

@dnolen: Very brave. Hah!

@mfikes: the first time I demoed ClojureScript at a conf was JSConf 2012

the story between then and now - can hardly be more dramatically different

Wow. Jelly stains. Thoughts on JavaScript, LISP & Play https://youtu.be/AnzMEdC2FDk

@mfikes: haha, yeah

@mfikes: I think that was just after cljsbuild appeared, good REPL story, source maps - non-existent

no proper incremental building, etc.

Watching it now. John McCarthy was still alive at the time.

@mfikes: yeah Dan Ingalls (original Smalltalk team) was in the audience

that was frightening

Holy shite

👏

woah

just bootstrapped cljs

eval working in the browser!

I never thought it would happen

hmm

is it known to work with advanced optimization?

@ericnormand: David said bootstrapped would only work with :none and :simple

ah

thanks

I looked around in the docs, but couldn't find it

Docs

I think the :simple stuff was mentioned in here or IRC FWIW. Can’t recall whre.

can I help out by recording it somewhere?

That’s a good question. I don’t even think the new cljs.js stuff or things about bootstrapped even have a page for stuff like this.

Perhaps it could just go as notes to the existing compiler options page (where :none et. al. are described.)

was thinking here: https://github.com/clojure/clojurescript/wiki/Bootstrapped-ClojureScript-FAQ

In truth, that page was created to answer the frequently asked question about whether bootstrapped ClojureScript means we can develop without Java.

The other place is https://github.com/clojure/clojurescript/wiki/Compiler-Options#optimizations

you say "the new cljs.js"

where is that?

or what is it?

That is the new namespace that contains the support for making use of bootstrapped ClojureScript.

I'm trying to make an embeddable compiler widget for my blog

and my courses

but it's kind of big to put on the blog

Ahh… are you using cljs.js to do it?

no

https://github.com/clojure/clojurescript/blob/master/src/main/cljs/cljs/js.cljs

That has eval-str, etc.

I just took the browser/cljs_bootstrap/dev.cljs and modified it

oh, nice

I did not see that

interesting

The overall goal is, if you want to use bootstrapped ClojureScript, you can simply make use of that cljs.js and call, maybe eval-str

it can load libraries

Yes. If you want to get fancy. Planck is making use of that now.

Replete would also do so in the future.

I can't see needing that just yet

I really wanted a textbox that people could try to implement map in

Yeah, eval-str is all you’d need.

Planck is using it here, if you are looking for another example client https://github.com/mfikes/planck/blob/master/ClojureScript/planck/src/planck/core.cljs#L97

I think Joel Martin recently updated http://ClojureScript.net to use cljs.js

cool

is state to hold def's and things?

the environment, that is

I’m not sure…. I think source map info is dropped in there. I haven’t looked in the state atom yet to see what’s really in there

so you just need to create an empty one and keep using it?

wow

I just started the Jelly Stains video

Yeah

and I'm amazed by how small @dnolen's hair is

Hah!

He had hair back then. Wait….

well, it would be nice if advanced worked

would be much smaller

hmmm

is there a reason cljs.js/compile would always return ""?

or, rather, always return "" when compiling a literal value

Try telling it :context :expr

I wonder if that is available for cljs.js/compile or just eval-str

ok

thanks

what does context do?

It is telling it the context of what is being evaluated. Like expression, statement, return. AST concepts. So for a REPL you are giving it an expression to evaluate. Especially a literal like 1

@ericnormand: See http://dev.clojure.org/jira/browse/CLJS-1357

yeah

but now it's throwing an assertion error

not sure why

ah

my fault

wrong number of arguments 😞

FWIW, David has only recently (a day or so ago) finished coding up that stuff and Joel Martin and I have been trying it, finding defects, logging JIRAs.

IOW there could still be defects to be found.

Some JIRAs have been logged, but not yet fixed, etc.

thanks

hmmmmm

it's amazing this works

still some things I don't understand

but they'll have to wait till morning

thanks so much!

this is awsome!

I am having fun with eval as well. Fantastic work!

does anyone know how to bring other namespaces into the compiler environment? (e.g. clojure.core)

@mhuebert: An ns form with a :require

@mfikes what do you use for load-fn?

@mhuebert: In Planck: https://github.com/mfikes/planck/blob/master/ClojureScript/planck/src/planck/core.cljs#L101

I was looking at that… slurp, is that for in a browser or in node?

@mhuebert: Planck is still in the process of being updated for cljs.js

@mhuebert: Neither. See the demo at the bottom of http://blog.fikesfarm.com/posts/2015-07-16-fast-javascriptcore-desktop-clojurescript-repl.html

@mfikes looks great

@mfikes so for browser usage, the load-fn needs to find js or cljs source for whatever path is required

To make Planck be useful, my plan is to implement I/O facilities and other stuff as well as making it support this

Usage:  planck [init-opt*] [main-opt] [args]

  With no options or args, runs an interactive Read-Eval-Print Loop

  init options:
    -i, --init path     Load a file or resource
    -e, --eval string   Evaluate expressions in string; print non-nil values

  main options:
    -m, --main ns-name  Call the -main function from a namespace with args
    -r, --repl          Run a repl
    path                Run a script from a file or resource
    -                   Run a script from standard input
    -h, -?, --help      Print this help message and exit

  The init options may be repeated and mixed freely, but must appear before
  any main option.

  Paths may be absolute or relative in the filesystem

@mhuebert: For any JavaScript environment, in order to support requiring code, *load-fn* needs to be properly implemented. The requirements of that function are documented fairly well.

@mfikes: yeah I’ve been going over those docs. I’m a bit confused about what kind of string to return for, say, clojure.core.

@mhuebert: Are you considering a case like (ns cljs.user (:require clojure.set)) ?

@mfikes: as a basic example, (ns cljs.user (:require clojure.core :refer [shuffle]))

i mean I’d like to refer all of clojure.core

@mhuebert: Is clojure.core something that can be compiled as ClojureScript?

@mhuebert: There is a cljs.core/shuffle

@mfikes yeah, that would be it. I can type in (clojure.core/shuffle [1 2 3 4]) and that works too but maybe its mapped to cljs.core/shuffle

@mhuebert: Interesting. I’ve never gone down that path. There is this: http://dev.clojure.org/jira/browse/CLJS-1362

@mhuebert: In other words you should be able to do (shuffle [1 2 3 4])

if I eval-str (fn[{:keys [a b]}] nil ) I get warnings - WARNING: Use of undeclared Var cljs.core/not for not, apply, hash-map, get

@mhuebert: Maybe that is the same root cause as CLJS-1362

@mfikes: ah that would be it

@dnolen: How safe is cljs.analyzer for tooling to use? Boot-cljs currently uses *cljs-warning-handlers*, *cljs-warnings* and with-warning-handlers.

They are used to check if any warnings happened during the build.

@juhoteperi we recently discussed this in #C07UQ678E , cljs.analyzer changes often and tooling should use cljs.analyzer.api instead.

@juhoteperi: (that is in general, boot-cljs might have special needs)

Yeah using api namespaces would be preferred, but it doesn't currently provide anything for this use case.

Has anyone experienced [object Object] is not iSeqable error ?

i was trying to do this (for [i vector-of-list] [:custom-reagent-component i])

@juhoteperi: it’s not safe to use at all. if you use it you should expect your library to break all the time.

@juhoteperi: if there’s something people need start opening enhancement tickets + patches.

to the api namespaces

there’s already work underway to present better names and to hide the dynamic vars.

this approach should be replicated where possible.

@juhoteperi: for example changes likes this https://github.com/clojure/clojurescript/blob/master/src/main/clojure/cljs/analyzer/api.clj#L20-L29

now users don’t have to know about the cljs.env namespace, or have a confusing name that clashes with analysis environments, etc.

https://twitter.com/mark_l_watson/status/624980508452130817

Are there any spread-sheet library in Cljs like http://handsontable.com ?

@lazy-lambda: It probably means that vector-of-list is not a sequence, since for requires a sequence and (seq? []) => false

@bensu: the vector-of-list thing was actually an atom and I forgot to deref it. Silly Me. I should have updated here though. Thanks.

Does anyone know of a way to pass figwheel a port number at runtime?

Or, barring that, to assert a leiningen option at runtime?

@chancerussell I think that is configurable: https://github.com/bhauman/lein-figwheel#figwheel-server-side-configuration but maybe I didn't understand your question?

Yeah, I’m just looking for a way to set it at launch, rather than by modifying project.clj

I think it may look to the env variable PORT

I've not used figwheel directly; I've always used chestnut. There are two ports one for the HTTP server and one for websockets. If you are using the Chestnut template you can control the HTTP port via the PORT env variable.

Hmm, tried but no luck. May have to just use sed the project.clj

Cool, I’ll check on how Chestnut does it.

Thanks!