02:55 <FromGitter> <Dan-Do> But browser can only run javascript, doesn't it? Others (Typescript / CoffeeScript/ Mint-lang...) are complied to jvs right?

02:56 <FromGitter> <Dan-Do> We need isomorphic for both client/server side.

02:57 <Andriamanitra> web assembly is a thing these days, it's possible to almost completely avoid javascript

02:57 <FromGitter> <Dan-Do> Can you give me link of document how to build web asm using crystal?

02:58 <Andriamanitra> i have no clue how tbh, all i know is it should be possible since crystal is built on llvm

02:59 <FromGitter> <Dan-Do> https://forum.crystal-lang.org/t/crystal-in-possible-wasm-everywhere-future/1383

03:02 <FromGitter> <Dan-Do> It's more difficult than jvs :)

03:06 <Andriamanitra> for now yes, but i think that's a much better goal to work towards than embedding jsx in crystal

08:15 <raz> yup there has been rumors of porting hyperstack to crystal. imho that's the way forward (isomorphic with wasm); https://forum.crystal-lang.org/t/crystal-js-transpiler/903/29

08:15 <raz> unfortunately it's a major undertaking and doesn't have much manpower behind it

08:17 <raz> (woops google gave me a deeplink - i just meant to link to that thread as a whole, not that specific comment)

08:17 <yxhuvud> I'm not certain I see the point. people seems to do well without it.. especially Lucky seems to go around the issue by not being magic.

08:18 <raz> yup, hyperstack is essentially lucky, just with react on the client side

08:20 <raz> so you for example have as a page template: `class MyPage; def render; div { "Data is: ${model.data }; end; end". and whenever you update the model using the ORM it automatically updates on the screen of all users who are looking at it

08:20 <raz> without having to write any JS code

08:21 <raz> (simplified example of course, but that's how it works in a nutshell)

08:22 <raz> i've dabbled with the ruby verison, it still has rough edges, but works and could be a wonderful way out of JS hell if someone (ideally with deep pockets) was to create a fully polished impl :)

08:28 <raz> i think flutter by google comes with a similar promise (apart from being hybrid). but well, it's dart and it's google. will probably remain a niche monstrosity like GWT.

09:59 <repo> Blacksmoke16: hey again o/ I'm trying to understand how exactly to use dependency injection for athena controllers. So i need a service that holds an mqtt client connection and allows me to publish messages from controller actions

10:00 <repo> i guess i'm asking how exactly can i inject a service into a controller that i've annotated with `@[ADI::Register..]`

10:02 <repo> hm i guess i'll try just adding an initializer with that service as type restriction

10:13 <repo> hm nope, that didn't work

10:15 <repo> ah i had to register the controller as well

12:59 <FromGitter> <Blacksmoke16> and be sure to make the controller public, via `public: true`

12:59 <FromGitter> <Blacksmoke16> ideally i think it would register controllers behind the scenes for you, but that depends on that custom macro method PR

13:20 <repo> yup i got a compile time error telling me i should be using public: true :)

13:20 <FromGitter> <Blacksmoke16> ah nice

14:23 <FromGitter> <Blacksmoke16> really want that bug to get fixed so it actually points to the correct thing

14:23 <FromGitter> <Blacksmoke16> error message is good but the location is not super helpful atm :/

19:50 <oprypin> I'm binding a C function that has `...` in its args. and the function ends up receiving just zeros from me

19:50 <oprypin> this is the binding https://github.com/oprypin/crystal-imgui/search?q=igText

19:50 <oprypin> and this is the function https://github.com/cimgui/cimgui/blob/18af68901e7b8608f3c69a5b2993610c03317083/cimgui.cpp#L527

19:52 <oprypin> and this is the call site

19:52 <oprypin> i can add printf("!!! %f\n", va_arg(args, int)); inside the function - that indeed prints a zero like i'm observing

19:57 <oprypin> mm i actually have a nice local repro https://carc.in/#/r/9oor

19:58 <oprypin> no it's not a repro, i justmade a mistake

19:59 <oprypin> this might be specific to floats?

20:01 <oprypin> https://carc.in/#/r/9op5

20:10 <FromGitter> <franciscoadasme> hey everyone, can anyone point me to some guide about profiling a crystal program?, I found this old article (https://www.mikeperham.com/2016/06/24/profiling-crystal-on-osx/) but it doesn't seem to work any longer as Instruments only showed GC calls for the example

20:26 <oprypin> yea i think Crystal fails to do the promotion of va_args

20:28 <oprypin> https://carc.in/#/r/9oq6

20:37 <oprypin> reported the bug https://github.com/crystal-lang/crystal/issues/9742 😬

21:04 <oprypin> mm this is also a big oof https://carc.in/#/r/9oqg

21:06 <oprypin> https://carc.in/#/r/9oqi ? ? !? ? !

21:06 <FromGitter> <Blacksmoke16> idt map keeps the typing like that

21:07 <oprypin> i know that my types are a subset of incoming types always

21:07 <oprypin> https://github.com/crystal-lang/crystal/blob/5999ae29beacf4cfd54e232ca83c1a46b79f26a5/src/tuple.cr#L113 i hate when Array is used as the restriction. should be Indexable yet again

21:08 <FromGitter> <Blacksmoke16> prob yea

21:10 <oprypin> any way to avoid this to_a

21:11 <oprypin> because other than that blemish i'm quite happy with https://carc.in/#/r/9oqk

21:29 <oprypin> found my solution but only by reopening tuple. https://carc.in/#/r/9or2 any way to avoid that?

21:34 <oprypin> pushed my fix https://github.com/oprypin/crystal-imgui/commit/5c305064bfd2b8333ae00c07eb2572f9dbad575b