Fork me on GitHub
#clojurescript
<
2017-07-02
>
bhauman00:07:17

@dnolen @wildermuthn I just tested out reverting the following commit: https://github.com/clojure/clojurescript/commit/cdaeff298e0f1d410aa5a7b6860232270d287084 in order to fix the Node reloading issue in figwheel

wildermuthn02:07:30

bhauman: Thanks for following up on this. Had hoped to get time to test it out, but … life. 🙂

wildermuthn02:07:00

I guess maybe this indicates that there aren’t a lot of people using figwheel with node.js, at least on the latest versions. Which is too bad, because it really is an amazing dev experience to have reloading both on the server and client.

wildermuthn02:07:15

I mean, that the issue queue in figwheel wasn’t larger.

bhauman00:07:38

and as expected it fixed the issue

dnolen00:07:24

I would file issue so I don't forget

cjhowe03:07:24

is there a way i can get an atom of a part of an existing atom? for example, if i have an atom that is a map with some key :user, i want to be able to create an atom that represents the value of :user, to control which parts of my app have access to a specific parts of an atom map

noisesmith03:07:12

atoms don't do this, but reagent atoms allow cursors that behave this way

cjhowe03:07:53

well, that sounds nice, but does reagent support server side rendering in ring? i am using rum

noisesmith03:07:10

in cljs since there's no threading, there's no harm in just using multiple atoms though

noisesmith03:07:49

I haven't used it, and don't know how it compares to rum at all, but it exists https://github.com/reagent-project/reagent-cookbook/tree/master/recipes/reagent-server-rendering

noisesmith03:07:32

@cjhowe looks like rum has cursors too https://github.com/tonsky/rum

cjhowe03:07:42

oh, sweet!

noisesmith03:07:57

I couldn't find a link directly to it, but if you search cursor on that page, looks like it does what you want

cjhowe03:07:05

yeah, i did that, thanks 😄

stuartrexking09:07:03

Can anyone tell me what the 😆 symbol refers to?

thheller09:07:18

@stuartrexking someone in #reagent can probably answer that

crankyadmin09:07:30

Hi, Given:

(some-func some-arg (fn [err metadata] 
						 (if err(println err) 
						  (println metadata))
How would one go about dealing with this in a asynchronous way? Basically, I'm trying to build a map to ship over to my 'render thread' in electron.

crankyadmin09:07:38

I've got a funny feeling about about to learn how core.async works though 🙂

yury.solovyov09:07:08

how it works or how to use it?

crankyadmin09:07:40

How to use it... I basically need a way of calling a async function and waiting to built a map from it and ship it in one go... feels a little like an anti-pattern...

stuartrexking09:07:19

@thheller I can’t seem to see anything in the reagent docs. Any hints?

thheller09:07:58

no idea, never used reagent. looks like it has to do with native interop? ie. using a native JS component

thheller09:07:35

as far as CLJS is concerned it is just a normal keyword

crankyadmin09:07:56

@stuartrexking :> is for using react components.

crankyadmin09:07:11

[:> TitleBar {:controls true :inset true :style {:z-index "1000"
                                                      :position "fixed"
                                                      :top 0}}
      [:> Toolbar {:height              "43"
                   :horizontalAlignment "center"}]]

stuartrexking09:07:29

Yes, but what does that mean? What does it do? What does it refer to?

stuartrexking09:07:38

You can use react components without 😆

crankyadmin09:07:55

Its just syntactic sugar for react-create-class

crankyadmin09:07:51

Cheers man... thats exactly what I'm after.

crankyadmin10:07:00

@stuartrexking > You can use react components without 😆 Not quite... See: https://reagent-project.github.io/news/news060-alpha.html "Better interop with native React"

stuartrexking10:07:52

Thanks. I was using adapt-react-class.

metametadata11:07:59

@cjhowe in addition to Rum's cursors and derived-atoms you may find interesting this list of libs: https://metametadata.github.io/carry/faq/#is-there-a-way-to-code-a-view-model-if-i-dont-use-reagent For atom "entanglement" there're also these solutions: https://github.com/Frozenlock/entanglement, https://github.com/rymndhng/entangle And here's my solution to "read-only" entanglement: https://github.com/metametadata/carry/blob/143877b3e07f2d530ed0923f2f81cc06a94c5413/src/carry/core.cljc#L54

yury.solovyov12:07:03

is there a guide one can follow to publish a cljs package?

sekao16:07:02

Anyone running into a Use of undeclared Var cljs.js/lib warning with 1.9.671? Doesn't happen to me with 1.9.660

dnolen16:07:29

@sekao fixed in master

sekao16:07:35

gotcha thanks

dnolen16:07:58

:modules enhancements coming very soon. If you feel like giving this a run through https://gist.github.com/swannodette/6150d4213aeb9eba31e03ae522af4425, that would be helpful

dnolen16:07:20

the surface is boring, but hides a pretty massive change set - so if you try this also try the build on your own projects

dnolen16:07:36

I did a pretty crazy rounding of testing for this - but more is always better

dnolen16:07:56

also this guide should make it clear why having a resolve macro now is actually useful 🙂

yury.solovyov16:07:40

so, say, electron app might have 1 build that produces 2 files - for main and renderer, right?

dnolen16:07:34

@yury.solovyov I don’t know Electron dev that well - but is advanced compilation a thing there?

yury.solovyov16:07:55

yes, at least for browser-side

yury.solovyov16:07:33

main uses CJS from npm, so might be not as safe to use it there

yury.solovyov16:07:00

well, one might not use npm at all, and then yes, should be safe on both

dnolen16:07:01

hrm I dunno - there’s nothing really new here in this change

dnolen16:07:12

this is just a massive cleanup + ease of use thing

yury.solovyov16:07:53

ok, cool anyway

dnolen16:07:16

@bhauman thanks!

bhauman16:07:43

also noted a new relative path problem, which keeps rearing its ugly head

bhauman16:07:22

trying to see if its been reported

dnolen16:07:03

@bhauman what’s the cause? (or which combo of features)

dnolen16:07:17

people have mentioned :libs a couple of times (not :foreign-libs)

bhauman16:07:28

yep its in libs

bhauman16:07:07

for a single-segment GCL namespace in root of the specified directory

dnolen16:07:19

@favila mentioned that but I don’t think he opened an issue

dnolen16:07:34

(or something similar)

bhauman16:07:05

I will type it up then

dnolen16:07:47

@bhauman I’m assuming this case just blows up or something because of bad path computation logic

bhauman17:07:00

yep thats it

bhauman17:07:23

I've seen some churn around that computation

bhauman17:07:52

so a case has been probably missed

bhauman17:07:51

making a minimal case to be sure of the specific regression

dnolen17:07:39

cool, will make sure some tests come along with this one

dnolen17:07:39

@bhauman just made cljs.loader Figwheel friendly

bhauman17:07:40

uh ... pretend I've been under a rock, cljs.loader?

bhauman17:07:45

ok module manager

dnolen17:07:03

@bhauman yeah so like shadow-cljs, people can dev projects w/ code splitting the same under :none and :advanced

bhauman17:07:38

very very very cool

dnolen17:07:13

@bhauman thanks!

bhauman17:07:08

your are very welcome!!