@noprompt no

If you want that you have to invoke tools.reader yourself

thanks @shaun-mahood !

@noprompt could it be that the latest clojurescript version breaks garden css?

@noprompt it breaks with [org.clojure/clojurescript "1.9.293”]

@jrheard core.logic is not compatible with self-hosted ClojureScript. (FWIW, work has been done recently for core.match: http://blog.klipse.tech/clojure/2016/10/25/core-match.html)

I was trying to integrate my cljs-project with quickbooks api. should any one done please give me a suggestions?

Are there any downsides on using cookies for auth for an API only service?

@darwin For the record, I think anything but using <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/> is really asking for trouble. May be worth emphasizing it in the docs.

@pupeno i don’t see why not, this is how every rails application does authentication by default. the session information is encrypted and stored as a cookie in the client. checkout https://funcool.github.io/buddy-auth/latest/#signed-jwt

and a few SMS endpoint services i’ve used just simply have some small token they put in the URL or headers that identify your app

@naomarik isn’t jwt an alternative to using the session and thus, cookies?

@deas indeed

@pupeno from my understanding sessions use cookies to track user state. i haven’t implemented anything in clojure quite yet but at least in rails this is how it works

all i need to login as someone else is their cookie for that domain that includes all the session info

@naomarik yes, that’s how it works. jwt uses a non-cookie token to track users.

in any case, cookies can be used as the token transport header

there are no problem with mix jwt for generate tokens and cookie as transport header

yes this example i sent seems to just include the token in the authorization headers.. i don’t see why you wouldn’t be able to put it in a cookie though and be able to sign/unsign the data

the advantage to setting it on cookie is that in browser you don't need to hadle that headers manually 😛

@darwin In fact, it seems even worse than that. OOTB tomcat (need it) serves HTML setting ISO-8859-1 encoding header. And that even supercedes the meta header in the document.

also you can set expiration stuff and expire tokens by changing your secret key

https://github.com/funcool/buddy-auth/blob/master/examples/session/src/authexample/web.clj this example here seems closest to what i’m saying but i feel like it’s storing everything as plaintext, not sure how that info is being saved on the client

okay https://github.com/ring-clojure/ring/wiki/Sessions#session-stores the session middleware allows you to put your key there

so the data you put on the client is not visible

just have to use the ring.middleware.session.cookie/cookie-store

@deas at least you know where the problem is… 🙂

@darwin I may have missed something, but I think it may really be worth mentioning so other people don't run into similar issues.

well, people don’t read docs, so we would have to fix that “problem” first 😉

I think better time investment would be to check popular project templates and make sure meta charset it present there

@darwin Exactly what I was about to say!

@darwin ... and ring serving HTML using utf-8 response headers as well. People may be crazy enough run uberwars on tomcat.

@denik I discovered that last night. I've been working hard at trying to fix a number problems. It's been difficult.

is there a way to get some code prepended to the start of a generated js file from clojurescript ? I want to add the importScripts() line for webworkers

i mean, i could post-process the generated js files, but i was wondering if there's a hack or feature in clojurescript

I think there used to be a :preamble option afair

@octo221 https://github.com/clojure/clojurescript/wiki/Compiler-Options#preamble

Hey everyone, quick question, with the new 1.9.293 are we still not able to do :refer :all would it be better to do use?

@flyboarder: :refer :all isn't supported and may not ever be. use must be used with :only so you have to list everything you want to refer to anyway

Hey guys I am pretty new to ClojureScript, and I am really amazed by figwheel!

Are there any React wrappers for CLJS that support the context feature? I’ve looked at a bunch, but I didn’t see anything super-obvious.

@teslanick not aware of anything but I have done it manually and it isn't fun 😛

has anyone out there tried to use the closure library for matrix math? I can't even figure out how to initialize a matrix. The API docs say that it's a class ( https://google.github.io/closure-library/api/goog.math.Matrix.html ), the actual code looks like a function ( https://github.com/google/closure-library/blob/master/closure/goog/math/matrix.js#L57 ) and regardless of how I call it, I get a type error. Viz:

(require '[goog.math :as math])

(.log js/console (math/Matrix (clj->js [[1 2 3][4 5 6]])))

;; type error: goog.math.Matrix is not a function

(.log js/console (math/Matrix. (clj->js [[1 2 3][4 5 6]])))

;; type error: goog.math.Matrix is not a constructor

(.log js/console (new math/Matrix (clj->js [[1 2 3][4 5 6]])))

;; ditto, not a constructor

@gowder goog.math is not goog.math.Matrix

(require 'goog.math.Matrix)

(js/goog.math.Matrix. (clj->js [[1 2 3][4 5 6]])) should do it

ooh. that simple, eh. hah. thanks

Uff. That’s annoying. It seems like the cljs wrappers [for React] out there a lot ‘thicker’ than I want. 😕

@thheller also, so the name of the library is also a constructor?! that's... wow.

well technically the goog.math.Matrix is used to find the FILE that contains the source

which is goog/math/matrix.js or something

that then defines the goog.math.Matrix object

only then do you get the constructor

if you require goog.math it loads goog/math.js which has no knowledge of Matrix

but then how do things like :as and :refer work?

I guess maybe they don't, because it's not real clojurescript?

analyzer magic 😉

hah

in your case you could do

(:import [goog.math Matrix]) and then (Matrix. ...) if you don't like the js bit

refer of classes is funky

hmm. definitely not as clean as the jvm. thanks!

Hi. I'am trying to solve this, rather tricky problem with Reagent. However, it looks like an architectural problem in React based apps. Correct me if I'am wrong. We have a map component. That would be a parent component in Reagent

var mymap = L.map('mapid').setView([51.505, -0.09], 13);

var marker = L.marker([51.5, -0.09]).addTo(mymap);

@mik hehe welcome to the club https://github.com/facebook/react/issues/2763

answer basically is to use componentWillMount

but the official answer is do not build stuff like that via components but in your state instead

@thheller ComponentWillMount will work before rendering. So we can't initialize map on '#mapid' (it doesn't exist yet). Having a

mymap

@thheller I am using state approach right now. It works, but, I am looking for better ways. Thanks

dunno do the children that try to add things to the map have to be react components?

otherwise a component that just create a div with a react ref

should be enough to get a div into the page and then generate the map instance on it

dunno but be careful with componentWillMount as componentWillUnmount is the wrong order again 😛

Well, if we need them to auto update/remove themselves based on the state I guess.

@anmonteiro: thanks!

Hi! I'm having some issues with compiling my project with :optimizations set to anything. Is this a suitable channel to ask in? 🙂

it is!

Great! My build is, and has been, working fine all day. But just now I decided to try compiling using :optimizations :advanced, and Google Closure throws the error: >Oct 27, 2016 11:36:12 PM com.google.javascript.jscomp.LoggerErrorManager println >SEVERE: $HOME/flask-cljs/flask_cljs/target/cljsbuild-compiler-1/flask_cljs/core.js:6: >ERROR - required "flask_cljs.core.components.sidebar" namespace never provided >goog.require('flask_cljs.core.components.sidebar'); And the same for the only other file in the "components" folder, and thus the ns flask-cljs.core.components. My only dependecy is reagent. Also, compiling with any :optimizations, even :whitespace, throws the same error, aaaand I'm quite new to Clojure and Clojurescript 😊

@wielderofmjolnir hrm so your project is named flask-cljs right?

Yeah

what's your folder called?

flask_cljs

Supposed to have underscores for the filenames and hyphens for ns, yeah?

that's right

can you paste the NS form of one of those files that are erroring out?

>(ns flask-cljs.core.components.sidebar > (:require [flask-cljs.state :refer [app-state]]))

hrm looks right to me as well

your project folders are also structured according to the NS structure right?

i.e. you have flask_cljs/core/components/sidebar.cljs

Yeah 🙂

>{:id "min" > :source-paths ["src"] > :compiler {:output-to "../resources/public/js/compiled/flask_cljs.js" > :main flask-cljs.core > :optimizations :advanced > :pretty-print false}}]}

That's my build

config

(it's supposed to output to parent-dir, which has been working fine without optimization)

right

you can probably remove the ../ and have the same effect

Yeah, just tried it and got the same error..

@wielderofmjolnir hrm one last hunch, do you have a flask-cljs.core.components namespace?

Hah, I actually don't!

Is that it?

no...

the problem would be if you had one which declared a sidebar var

Aww.. I got excited there for a second 😄

hrm, one thing though, is that your output-to seems to differ from the path of the error

I figured the target folder was a temporary folder or something.

For compilation

try specifying a concrete output-dir

Also, I was def wrong earlier. :whitespace of course works.

Trying, a sec

Same error 😒

Different folder though

>flask-cljs/flask_cljs/../resources/public/js/compiled/out2/flask_cljs/core.js

can't say more without having a look at the project

also heading to bed. Probably something about those namespaces, but could be a compiler edge case

I'm trying to merge them all into core.cljs and see what I get then

Thanks though! Really grateful for your help @anmonteiro 😊

That worked!

Though it's not a very good solution..