I’ve just noticed there is defmacro in cljs. Is there a document how to use it properly?

nvm, I found this https://groups.google.com/forum/#!topic/clojurescript/HINltkZ_NuM

@danielcompton I have wanted to potentially make use of defrecords but I dont know of a particular where it would shine over say using functions and maps of data. For what scenarios do you find defrecords a good fit?

@sbmitchell: this is pretty good explanation https://github.com/plumatic/eng-practices/blob/master/clojure/20130926-data-representation.md

nice

thanks

Is there a way to make closure compiler break up methods so none of them are over 64k compiled? I'm trying to target rhino and currently only :optimizations :advanced will work, but the rebuild time on that prevents a good development experience

alternately, if i could precompile the entire clojurescript library with :optimizations :advanced in advance and then have the rest of my code use :whitespace that would work. i don't need the dead code elimination features at all

In ClojureScript, is it possible to use the #my.thing[1 2 3] or #my.record{:a 1, :b 2} reader forms that are documented in http://clojure.org/reference/datatypes? I’m having difficulty using them.

Ah, looks like no, unfortunately: http://dev.clojure.org/jira/browse/CLJS-1328

I’m having trouble understanding the relationship between lein-figwheel and Figwheel Sidecar - can anyone explain?

@jonoflayham: lein-figwheel is lein plugin, and figwheel-sidecar(and also figwheel) is figwheel system.

Hi

I have some questions, need community's help.

I passed an atom: (defonce tasklist-atom (atom {:entries []})) to om/root function: (om/root tasklist tasklist-atom {:target tasklistel}) Now I can retrieve the atom content inside component constructor like this: (seq (:entries @tasklist-atom)) or (seq (:entries tasklist-atom)) . It should be an issue that I can retrieve the atom content via: (seq (:entries tasklist-atom)) right? Component Constructor Code Example: (defn render-tasklist [state owner {:keys [counter] :as local}] (println (seq (:entries state))) ;; print ({:subject "a", :id 2, :completed false}) (println (seq (:entries @state))) ;; print ({:subject "a", :id 2, :completed false})

Thanks in advance

@xfcjscn: from Om's docs: "During the render phase, you treat a cursor as a value, as a regular map or vector. " https://github.com/omcljs/om/wiki/Cursors

@anmonteiro: Thanks a lot. I didn't go through the doc carefully.

Hey, where can I read about cljsc files?

https://clojure.org/guides/reader_conditionals

Thanks turbopape

you’re welcome

Hey, how would I go about catching errors involving XmlHttpRequest ? can’t transfer this knowledge to cljs http://stackoverflow.com/questions/26756752/how-to-catch-errors-involving-xmlhttprequest

sorry I asked this before, but how does one cljs.js/eval forms that refer to functions in other namespaces ?

@ajchemist thanks!

@urbanslug: the second answer there by Francois Dermu is how I'd go about it: I'm not 100% sure about this, but I wouldn't expect, e.g., goog.net.XhrIo to throw exceptions, but rather just return a status code, which your cljs code responds to. Are you actually seeing exceptions being thrown by XmlHttpRequest or anything else in JS?

Has anyone tried making chrome extensions with cljs yet?

(I'm sure someone has but I'm mildly surprised I haven't bumped into any discussion of it)

@bwstearns: https://github.com/binaryage/chromex-sample

Dirac would be a real-world example: https://github.com/binaryage/dirac

@darwin Wicked cool. Thanks. I had seen the devtools repo but I must have misunderstood something and took it as not a chrome extension.

the chromex-sample is just what I was looking for though

(not= 'cljs-devtools 'dirac)

dirac is a DevTools fork wrapped in a Chrome Extension

ahhhh lol.

I'm still on my first coffee of the day.

cljs-devtools is just a library with some hooks into DevTools console (it does not need Dirac)

definitely look into https://github.com/binaryage/chromex, cannot imagine starting a new chrome extension project without it

awesome. I have historically avoided a lot of JS stuff just because I am not a fan of the language, but since finding cljs I have been looking around at a lot of areas that I had previously not dove too deep into, and the other day at work I found myself wishing for a chrome extension.

you should be familiar with core.async, chrome apis are mostly async, you will be writing code spiced with a lot of go blocks, I guess (an example https://github.com/binaryage/dirac/blob/master/src/background/dirac/background/tools.cljs)

Hello guys and gals, I’m a bit confused why this is returning true

(number? (js/parseInt (last "123x")))

js/parseInt returns NaN which is Not a Number, but clojurescript says it is….

All I want to do really is to see if the last character in a string is a number, maybe there is an easier way

I think JS treats NaN as a Number type

(number? (js/parseInt (last "sdfgsdfg")))

NaN is always a value representable by floating-point

also true

so how could I figure out if the last character is a number or not?

I think (js/isNaN n) is what you need

Since the only numeric available in JS is double-precision floating point, NaN (as a valid double-precision floating-point) will make number? true

ah, ok… I’ll try that

@victorbjelkholm: regexp?

to see if a number is a NaN

parseInt would return an Integer or NaN so sounds reasonable

cljs.user=> (number? NaN)
true

thanks @spinningtopsofdoom

@fasiha: I’ll try everything I can to not do regex 😉

isNan isn’t supported in IE from what I see

might have to polyfill

oh nvm, that’s just Number.isNan

blissdev: didn’t know, not a requirement from me though

ah

planck seems to think NaN is a number when tested with number?

I'd prefer (re-matches "[0-9]" (last "123x")). Makes it obvious what you're doing instead of complexting language specifics. Parsing to number and checking for NaN is much less obvious. Plus it'll work in Clojure when you want to share code.

base don the standard it should be a number

4.3.24 NaN number value that is an IEEE 754-2008 “Not-a-Number” value

fasiha: maybe, if you’re coming from clojure and are gonna write serverside and want to share. In this case is just the frontend and coming from a javascript background

http://www.ecma-international.org/ecma-262/6.0/#sec-terms-and-definitions-nan

@bwstearns: Right, and also (== NaN NaN) is false

No matter how much sanity cljs brings to the fight, js will always bring more insanity.

I think JS engines implement isNaN by checking if x==x because NaN is the only thing (so far) for which x!=x

@bwstearns: again, it's a floating point, IEEE 754 thing. All languages that use floating point will have those semantics.

@bwstearns: But NaN is a number, even though its name would lead you to believe otherwise. Clojure is the same: (number? Double/NaN) yields true

@fasiha @mfikes But JS seems to be the only language that I run into NaNs frequently.

I think this all strictly following IEEE 754

Don’t get me wrong, JS is a funky language, but IEEE 754 is what this is all about

Thats why we there are standards 😛 to make the counter intuitive "clear"

Yup: in Clojure (= Double/NaN Double/NaN) ; => false too welcome to the world of floating-point! FYI, John Gustafson has described a unum alternative to floating-point which will solve most of these problems: http://ubiquity.acm.org/article.cfm?id=2913029 I believe there's a Java port too (and we'll hopefully get hardware support in a few years)

What @bwstearns says makes sense tho: I feel that I do run into NaN in cljs more than Clojure. Is that just because of parseInt?

I think it's just more javascript-y to actually return NaN as a result so they end up getting passed around and pop up several layers later where numbers (real numbers, that you can do math with, you know what I mean) were expected. btw I didn't mean to wholly sidetrack the discussion from the initial topic but it does seem like a regex might be the easiest?

do you need interop or not?

rather clj + cljs solution I mean

(!= nil (re-matches #"[0-9]" (last "123abc")))

seems fine I suppose. I try to avoid regex in most instances but to each his own

some? is more idiomatic than (!= nil …) I think? @bwstearns

yeah i was just looking for a way to get rid of that lol

boolean maybe?

(boolean (re-matches #"[0-9]" (last "asdf5")))

hmm

re-find returns a boolean?

some? is what you want.

ah

nvm no it doesnt

seems like that'd be a helpful re-function

maybe re-present

or re-exists

re-findis a thing.

yea but I thought it returned a boolean 😕 heh still have to do something like (nil? (re-find ..))

(contains? "0123456789" (last "asdf4")) No regex 😛

lol

I would probably just stick with the regex 😛

true. although other than feeling silly about typing [0-9] (a substantial barrier for recommending using the literal) I'm not sure how bad occasional literals are like that. If you were to change your mind about which characters should return true for this whole contains, you'd still need to modify the regex.

@victorbjelkholm: checkout goog.string/isNumeric

@jr: ah, that’s even more helpful, thanks a lot

(it appears to be implemented via regex btw)

@bhauman: Tested my Reagent project with the latest Devcards commit and it works, probably the problem was fixed by: https://github.com/bhauman/devcards/pull/100

@juhoteperi: cool, I'll get a release out pronto