@dm3 Thanks again for the report that led to https://dev.clojure.org/jira/browse/CLJS-2101 . It was pretty cut-n-dry (laziness in the analyzer). Looks like you were 24 forms away from seeing it occur in regular JVM ClojureScript as well (it just happens sooner in self-host owing to lack of chunked sequences).

Is there a way to pass props to reagent components in devcards?

I have something like

(defn scaled-img [props]
  (let [url  (:src props)]
   [:img {:src url
          :style {:width "400px"
                  :height "300px"}}]))

(defcard simple-image
  "### a simple scaled image with a source"
  (let [url ""]
    (dc/reagent scaled-img))
  {:inspect-data true})

but I’m not sure how to pass something like {:src url} to scaled-img

I got it.

(defcard simple-image
  (dc/reagent [scaled-image {:src ""}]))

That does the trick.

Anyone know of a CLJS framework that allows for custom tags? Hiccup/Sablono CLJS implementations in Re-frame and Om(.next) don’t handle non-standard HTML the last time I checked (which was last December, admittedly) — I’m trying to get A-Frame working in CLJS, which would require something like [:a-scene {custom-properties-too ""} // etc. ]

@acuervo what's a-frame?

http://aframe.io a webvr framework

@acuervo you should be able to use non-standard html tags with React (reagent, om, etc.). One approach would be to use https://facebook.github.io/react/docs/dom-elements.html#dangerouslysetinnerhtml

`'clojure.core 

`'~'clojure.core 

Is there something special I have to configure in order to use json-response-fomat in cljs-ajax? Right now I get Error: No *print-fn* fn set for evaluation environment.

@zilti are you printing the result somewhere?

that's probably the issue - try js/console.log instead

@pesterhazy: no, no println or something anywhere.

I only use timbre logging, and that works flawlessly.

timbre may be printing using println

you have to call (enable-console-print!) somwhere

and it needs to be before the call to print

No, Timbre prints into the console the way it is supposed to. That doesn't cause any errors.

But I'll try (enable-console-print!) anyway, maybe it fixes cljs-ajax

http://blog.leaningtech.com/2017/06/announcing-cheerpj-java-compiler-for-webapps.html ^ how funny would it be to compile the clojure jar to js

Hmm, apparently cljs-ajax always prints to console what it receives

@selfsame do it! 🙂

I'm trying the new :npm-deps in cljsbuild

Should I be concerned about this? "WARNING: JSC_INVALID_ES3_PROP_NAME. Keywords and reserved words are not allowed as unquoted property names in older versions of JavaScript. If you are targeting newer versions of JavaScript, set the appropriate language_in option."

@zilti you're welcome!

@pesterhazy: thanks 🙂

Hm, it looks like there's a difference between uncaught instances of ex-info and js/Error. An uncaught js/Error produces much better stacktrace in Chrome, at least when thrown from core.async context. And Firefox shows no stacktrace at all on uncaught ex-info 🙂

@metametadata ex-info isn’t about browser support

adding to my previous question (shameless bump), the inner quote isn't necessary to demonstrate the difference:

`clojure.core   -> clojure/core

`~'clojure.core -> clojure.core

@metametadata that said if there’s a way to get ex-info to provide better traces, submit a patch! Seems like an easy one to do if it’s possible.

@dnolen got it! I'll need to look into it when I have some free time. I suspect there could be some browser-specific properties in js/Error objects which are not added into ex-info instances or smt. like that. So maybe it would be possible to create js/Error and then augment it with ExceptionInfo protocol instead of creating an instance of ExceptionInfo and copying data from js/Error into it. I dunno.

right somebody needs to play around with it

though I was pretty convinced we already did a bunch of work here

ExceptionInfo subclasses Error etc.

alright, thank you for the reply

@moxaj Can’t see your previous question.

@mfikes @moxaj I thought somebody filed a ticket for this recently? My question was whether this problem was at the reader?

@bronsa ? or maybe we’re doing something wrong?

i'll take a look

I think this might be caused by cljs's provided impl for tools.reader/resolve-symbol

yeah

cljs.analyzer> (env/ensure (resolve-symbol 'clojure.core))
clojure/core

@bronsa ok! thanks

https://dev.clojure.org/jira/browse/CLJS-2109

How would I go about serving purecss from a webjar through an aleph server? My understanding is yada can support webjars? The only yada example I see is for serving swagger, not a resource.

@bronsa @dnolen CLJS-2109 might be also leading to solving https://dev.clojure.org/jira/browse/CLJS-1982

ah yes, that seems to be the same issue

@richiardiandrea closed as dupe, thanks for finding it

Awesome no problem!

@deadghost use :language-in :es5 in the compiler options

Does anyone know of any supporting specification that shows that

(identical? (goog.json.parse "\"a\"") (goog.json.parse "\"a\""))
(identical? (JSON.parse "\"a\"") (JSON.parse "\"a\""))
(identical? (js/eval "\"a\"") (js/eval "\"a\""))

I think it mostly comes down to the behavior of JavaScript eval specification, but I haven’t found any good references stating what “same object instance” guarantees there are around strings in javascript.

Basically, I worry about the possibility of a parser etc to have this behavior instead (identical? "a" (js/String. "a")) ;= false

This relates back to cljs keywords, since they seem to be reliant on the identity semantics of their underlying string

ie

(let [a1 :a
      a2 (keyword "a")
      a3 (keyword (js/String. "a"))]
  [(= a1 a2)
   (= a2 a3)
   (= a1 a3)])

;;= [true false false]

In clj on jvm, keywords are assured to be unique and interned as a runtime Keyword constructor time cost, so the same concern didn’t apply

Re: keywords and identical? in cljs: https://github.com/clojure/clojurescript/wiki/Differences-from-Clojure#data-structures and some background: https://groups.google.com/forum/#!msg/clojurescript/bSFK6CEE3PE/nIoD11yjK_YJ

^ @mikerod

I've read both. I don't think they answer the question

keyword-identical? relies on the identical behavior of the underlying string.

As does Keyword Deftype's -equiv used by =

So my question is targeted at what we can rely on from JS when it comes to string identity.

@mikerod == always works on equal strings

your concerns about new String(...) applies to any of the native type constructors

don’t use them if you want to avoid such semantic pitfalls

Good chance this behavior is covered in the ECMA-262 spec? it’s a pretty small document

Yeah. I guess my concern was relying on thing like JSON parsers. If they have me new String() behavior unexpectedly then calling keyword on them would produce non = keywords.

I tried finding In spec. I'll have to look closer I guess. Seems to be a hard search topic to. I appreciate your inputs though.

unless your targeting strange runtimes I don’t see a cause for concern

if you’re targeting browser post IE >= 8 (2008), likely a non-issue

point being I don’t think any of the browser native JSON parsers exhibit this behavior

and not sure why you would be using anything else

(in browser I mean)

yeah, that’s what I was thinking

I was just trying to find any sort of “guarantee” of that because I started to question it in my head.

obviously testing it out is one way to know 😛

hi! is there something similar to redux-dev-tools for figwheel? I'm looking for a way to track/debug/timetravel the state of a webapp...

well, just realizing, that this is rather about re-agent than figwheel...

@mikerod (new Number(1) === 1); // false in JS

if JSON parsers worked like this no one would use them

@dnolen that is a good point

On the JVM side, I actually have been burned by JVM deserialization bringing in boolean values with new Boolean(), which CLJ doesn’t interact well with (relies on singleton Boolean/TRUE | FALSE). So I guess I am just paranoid of those types of things happening in deserialization layers. JS is a different world though of course.

yeah I think this madness in JS was copied over from Java

yeah, it’s not a “feature” I like about either one