Mornin'

Good morning all

Morning

Morning all

mogge 😼

månmån

Thx @maleghast

Morning All :-)

@otfrom - de nada ;-)

Morning everyone

HELLO EVERYONE! 📣

Wow, hi there @danieleneal 😉

good morning everyone and HELLO DANIELNEAL!

sup

⬆️

I'm so sorry @lady3janepl

It was good while it lasted

I should be banned from the interwebs

do not ever stop being yourself

I am a connoisseur of bad puns

I'm not though. I have no taste in puns whatsoever and I have a degree in English Lit.

also had like 3h of sleep so it’s all good bar 3rd world war

ah, don't read the news then

>_>

do not think of the pink elephant…

given I have a mental auto-ignore on Trump and Brexit news, today’s big story is… summer solstice?

😉

happy summer solstice everyone 😄

hooray! after today every day gets darker until 21st December!

there will be many goths gently glowering in anticipation

no more sunburn, amirite?

read that as goths gently glowing and it took a moment to parse

on second thought I need more coffee

o/

Limavady sunset time isn't until 2211, so it won't be dark until 11pm.

MORNING

Er, with less caps

@jasonbell yeah, but your dawn is later you southwesterner you

I tend not to notice that bit.

so... my SAD starts today?

@seancorfield thanks for your comments on the uberjar stuff the other day, it was exactly what I needed :thumbsup:

Hello All… I’ve put some feelers out in #react-native already, but do any of you lot have any experience with ReactNative under Clojurescript? We are starting to talk about device-based apps and I am tempted to push hard on ReactNative as a way of keeping our tech-stack Clojure-based and uncluttered…

I shouldn’t be surprised, but - the git graph for linux is crazy. My parser doesn’t like 3-way merges - which are pretty rare elsewhere - but not in the linux kernel:

18-06-21 10:43:41 Crobuzon.local WARN [git-cd-log.core:132] - Can't diff commit with more than 2 parents:  790844941ae2d0ccde5892140cc499982800dd46  -  Merge branches 'for-4.16/hid-quirks-cleanup/asus', 'for-4.16/hid-quirks-cleanup/elecom', 'for-4.16/hid-quirks-cleanup/ish', 'for-4.16/hid-quirks-cleanup/multitouch', 'for-4.16/hid-quirks-cleanup/pixart', 'for-4.16/hid-quirks-cleanup/rmi', 'for-4.16/hid-quirks-cleanup/sony' and 'for-4.16/hid-quirks-cleanup/toshiba' into for-linus

@maleghast #cljsrn

@danieleneal has been doing cljs rn stuff iirc @maleghast

yeah, but he's very shouty. 😉

😂

yeah @maleghast I've been doing a fair amount of react native stuff

oops i meant @danieleneal HAS BEEN DOING CLJS RN STUFF IIRC @maleghast

is that why you are so angry @danieleneal?

hahaha

my process is (cycle [:react-native-red-screen-of-death :problem-with-npm :argh-android])

@maleghast I'm using expo - which is imo a decent trade off. It's a lot easier to get started and integrate CI but you can't drop down into native

I'm actually starting to get the hang of it now but the learning curve was pretty steep

ah, regular clojure is much easier (constantly :classpath-problem)

I think @olivergeorge recently put a guide together getting up and running, covering lots of the questions that come up as you begin development

ah yes, this is the one: https://confluence.condense.com.au/pages/viewpage.action?pageId=28082685

Bore da

@danieleneal - Thanks for the feedback, I did rather have the fear that it might be a lot to take in…

Initial investigations had led me to re-natal..?

Not spotted expo yet, but I will have a rummage on the internet…

Going to have a look at that link you posted as well…

Morning all! 👋

@maleghast you can try expo by doing lein new expo see https://github.com/seantempesta/expo-cljs-template

expo is like a curated react native experience that bundles some common native libraries and means you only ever write js (or cljs) code, not native code.

Sounds very appealing!

Does that mean there is a “water’s edge” of can do / can’t do though? For example can you access the device camera and / or mapping capabilities and GPS, for example?

(Sorry was away from Slack for lunch)

Yes there is definitely a "water's edge" of can do/ can't do but the can do functionality is quite broad. You can access device camera and mapping I think. I think their aim is to provide sensible js apis around the native features that aren't necessarily part of core react native

That sounds very encouraging... Sorry I am "in and out" today...

that's ok . I'm in and out all the time

I am out all of the time...

of my mind that is.

💥

If I have passed a class into a function as a var (e.g. (defn foo [klass] ...), how can I instantiate an instance of that class?

Where I would expect to call it like (foo Exception)

(new klass) is of course looking for a class called 'klass'

Is there a way to return the "middle" items in a vector without them being wrapped in a lazy-sequence? I mean drop-last returns a lazy-seq, but if I had a vector that looked like this:

[:order :order-id "order-1" "Order"]

(klass ...) also doesn't work, because Java interop expects the dot and it's also looking for a class called klass

(in the above I would do (drop-last (rest coll)) to get the "middle" items as a lazy-seq))

@maleghast so you want the middle item, or the middle items?

both the middle items

I’m a little unsure what you want

(some-fn [:order :order-id "order-1" "Order"]) => [:order-id "order-1"]

like that?

yeah

🙂

but NOT inside a collection

:thinking_face:

how do you get back multiple things if its not in a collection

@carr0t I looked for the java implementation of what you’re talking about - https://stackoverflow.com/questions/4872978/how-do-i-pass-a-class-as-a-parameter-in-java Does (.newInstance klass) work?

Basically I am (for...)-ing over a vector of vectors. I want to use the contents of the vector as parameters to a bidi call:

(bd/path-for routes (first coll) (drop-last (rest coll)))

but the above ^ will eval to this:

(bd/path-for [[vec of routes]] :order `(:order-id "order-1"))

@carr0t I’ve replied to your question about new in #clojure

and that won't work as the bidi/path-for (bd/path-for) expects the params for the URL / route definition to be parameters, like this:

(bd/path-for [[vec of routes]] :order :order-id "order-1")

(long story short, I am trying to create a re-usable Reagent component for our breadcrumb trail and some links can only be "built" by Bidi if I pass in the extra params)

dumb idea but could you desctructure ur coll

i assume u get it passed in

so its something like

(defn x [[a b c d]] ... (bd/path-for routes a b c))

@maleghast that’s what apply is for

(apply bd/path-for routes (first coll) (drop-last (rest coll)))

That's not a dumb idea, but for URLs that don't have extra params that would not work too well and some have more params and some less

@bronsa - O-RLY?

so apply will turn any collections into their constituent parts and then eval the function call..?

(apply f a b .. '(x y ..)) = (f a b .. x y ..)

Wow! That's so cool - I'd never really had to think about this ever before so I've been wondering what apply is for, if I am honest...

@bronsa Ace, ta 🙂

@bronsa - I am assuming that if the result of (drop-last (rest coll)) was nil / an empty sequence that would still be ok?

(apply f nil) would just call (f)

yes

[] or nil just splice into nothingness

So would you do something like this then? (apply bd/path-for routes (butlast coll)) @maleghast

Yeah, that's what I am doing, and it's working! 🙂

Thanks so much guys - your assistance has allowed me to turn another part of the app (reagent spa) into a reusable component and reduce the line count on the app by another 320 lines.

I am in refactor MODE - have dropped the overall line count by nearly 1300 lines since I started.

(though some of that drop is not a true drop, just moving functions that didn’t need to be in the main app namespace into another namespace / file and requiring them in) Still the main Clojurescript file is now a LOT smaller and easier to navigate around as a result. I reckon I can make it even smaller over the next couple of days too.

i ❤️ deleting code

Morning

Hi!

Have I mentioned how much I <3 the Heathrow Express..?

pop quiz for the evening

explain why this works

Clojure 1.10.0-master-SNAPSHOT
user=> (deftype T [x])
user.T
user=> (defn f [^T t] (.-x t))
#'user/f
user=> (deftype T [x])
user.T
user=> (f (T. 1))
1

but this throws a ClassCastException

Clojure 1.10.0-master-SNAPSHOT
user=> (deftype T2 [x])
user.T2
user=> (defn f2 [^T2 t] (.-x t))
#'user/f2
user=> (f2 nil)
NullPointerException   user/f2 (NO_SOURCE_FILE:6)
user=> (deftype T2 [x])
user.T2
user=> (f2 (T2. 1))
ClassCastException user.T2 cannot be cast to user.T2  user/f2 (NO_SOURCE_FILE:6)

so if the function is called before the type is re-defined the type-hint borks when called with an instance of the re-defined type, but if the function isn't called before the type is re-defined there is no borkage ?

yep

well i don't have time to dig into it properly, but i'll guess that the type hint targets are resolved lazily via the var ?

not quite but you're on the right track

the first example didn't use to work either before 1.8.0

i get the feeling you just gave me a massive hint and i'm not seeing the elephant in front of me 😬

linkage happens at class load time, since 1.8.0 clojure doesn't force loading of a class upon import, but just on first use

so in the first example f is linked to the second T

in the second f2 is linked to the first T2 and so a ClassCastException

ah, i didn't know clj had changed class-load-time in 1.8.0 neat!

Does anyone know when on Friday the Clojure eXchange deadline for submissions is?