<< 19-09-2025 >>
00:00:26*SchweinDeBurg joined #nim
00:13:21*beholders_eye quit (Ping timeout: 244 seconds)
00:15:53*xutaxkamay_ joined #nim
00:16:13*xutaxkamay quit (Ping timeout: 260 seconds)
00:16:50*xutaxkamay_ is now known as xutaxkamay
01:29:48FromDiscord<nelsono6> I'll help the first 15 people interested on how to start earning $90k within 72hours, but you will reimburse me 10% of your profits when you receive it. Ask me how!!
01:29:52FromDiscord<nelsono6> If interested, send me a direct message. For more information…contact via.. TG↵@Nelson_owen↵https://t.me/Nelson_owen
01:37:34*rockcavera quit (Remote host closed the connection)
01:40:12FromDiscord<riku5543> sent a code paste, see https://play.nim-lang.org/#pasty=AExzbFXa
01:45:32FromDiscord<riku5543> (also <@&371760044473319454> there is a bit of spam in these channels)
01:58:19*SchweinDeBurg quit (Quit: WeeChat 4.8.0-dev)
02:19:35FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#pasty=GKMyZMUK
02:30:55FromDiscord<riku5543> Ah okay I see. Would it be bad if my wrapper kept track of the filename variable and just updated that when trying to assign to it? (and then I could destroy it later when, say, shutdown is called)
02:33:26FromDiscord<riku5543> I'm just not sure of wrappers having an internal state of sorts haha. But I guess stuff like sdl does that already
03:46:48*SchweinDeBurg joined #nim
05:52:02*alexdaguy joined #nim
06:10:48*alexdaguy quit (Quit: WeeChat 4.7.1)
07:57:50*skippy8 joined #nim
08:09:30*xtr00 quit (Read error: Connection reset by peer)
08:15:38*xtr00 joined #nim
08:56:15*xtr00 quit (Ping timeout: 258 seconds)
08:57:11*xtr00 joined #nim
10:40:45*mahlon quit (Ping timeout: 245 seconds)
11:14:55*mahlon joined #nim
11:18:49*xtr00 quit (Ping timeout: 256 seconds)
11:18:49*nils` quit (Ping timeout: 256 seconds)
11:19:06*nils` joined #nim
11:19:39*xtr00 joined #nim
11:28:36*xutaxkamay quit (Ping timeout: 244 seconds)
11:28:37*yeti quit (Ping timeout: 244 seconds)
11:28:37*FromDiscord quit (Ping timeout: 244 seconds)
11:28:53*xutaxkamay joined #nim
11:34:48*SchweinDeBurg quit (Ping timeout: 256 seconds)
11:35:53*dicot quit (*.net *.split)
11:35:53*Jjp137_ quit (*.net *.split)
11:35:53*computerquip quit (*.net *.split)
11:35:54*johuck quit (*.net *.split)
11:35:54*joast quit (*.net *.split)
11:35:54*GreaseMonkey quit (*.net *.split)
11:35:54*redj quit (*.net *.split)
11:36:48*SchweinDeBurg joined #nim
11:40:23*dicot joined #nim
11:40:23*feinfinger joined #nim
11:40:23*Jjp137_ joined #nim
11:40:23*computerquip joined #nim
11:40:23*johuck joined #nim
11:40:23*joast joined #nim
11:40:23*GreaseMonkey joined #nim
11:40:23*redj joined #nim
11:40:59*feinfinger is now known as Guest8695
11:41:15*FromDiscord joined #nim
11:44:29*Guest8695 quit (Changing host)
11:44:29*Guest8695 joined #nim
11:44:37*Guest8695 is now known as yeti
11:45:28*xtr00 quit (Quit: ZNC 1.9.1 - https://znc.in)
11:50:47*xtr00 joined #nim
11:58:22*dicot quit (Read error: Connection reset by peer)
12:04:05*dicot joined #nim
12:12:53FromDiscord<ayex> [pmunch](https://matrix.to/#/%40_discord_392962235737047041%3At2bot.io) I assume that is referring to the message visible from matrix from nelsono6 from very early this day?
12:16:18*beholders_eye joined #nim
13:09:21*dicot left #nim (#nim)
14:13:05*bcksl quit (Server closed connection)
14:19:48*bcksl joined #nim
15:10:45FromDiscord<mitranim> Seems strange that empty types like `tuple[]` or `object` have `sizeof == 1`. In Go it's 0, which makes perfect sense.
15:41:55FromDiscord<Robyn [She/Her]> In reply to @mitranim "Seems strange that empty": in Go they're probably compiled away
15:42:09FromDiscord<Robyn [She/Her]> You can't have an object that exists at runtime and have it take up nothing, after all
15:42:16FromDiscord<tauruuuuuus> In reply to @mitranim "Seems strange that empty": IIRC it's actually implementation-dependent in go, but the main implementation has indeed that property
15:42:37FromDiscord<tauruuuuuus> actually I was messing around and it seems like they are compiled away in nim too..
15:42:57FromDiscord<mitranim> It has some funky side effects. For example, a sequence of zero-sized values is also zero-sized. A zero-sized struct field shares its address with the next field.
15:43:18FromDiscord<mitranim> (edit) "sequence" => "backing array for a slice"
15:43:20FromDiscord<Robyn [She/Her]> huh weird
15:44:13FromDiscord<mitranim> This stuff sometimes saves space when you need to implement an interface but your type has no state at all
15:48:27FromDiscord<tauruuuuuus> in go it's useful in chans and for sets I guess, that's where it came in handy for me, bu anyways this is more for #offtopic lol
15:50:27FromDiscord<kiloneie> In reply to @Zoom "I don't know what": How is this ?
15:50:33FromDiscord<kiloneie> (edit) "In reply to @Zoom "I don't know what": How is this ? ... " added "https://pastecode.io/s/kmjza34m"
16:27:56*n|Phreak quit (Ping timeout: 256 seconds)
16:29:42*n|Phreak joined #nim
16:45:22*skippy8 quit (Quit: WeeChat 4.7.1)
17:18:56*SchweinDeBurg quit (Quit: WeeChat 4.8.0-dev)
17:47:35FromDiscord<tauruuuuuus> Requesting nimscript expertise
17:49:13FromDiscord<tauruuuuuus> sent a code paste, see https://play.nim-lang.org/#pasty=jgjEWqpy
17:52:43Amun-Rarun the command manually and test return code with echo $?
17:52:51Amun-Raor echo $rc in fish
17:54:50Amun-Ratauruuuuuus: wait… do you need to run that in static context?
17:55:40Amun-Raah, nimscript
17:55:42Amun-Ranvm
18:00:28FromDiscord<tauruuuuuus> already done and tested both there and in nimscript by printing `code` through echo, or I wouldn't be thinking I'm going crazy lol
18:06:44Amun-Rahmm, I added my own if before yours and mine works fine ;)
18:10:14Amun-RaI can confirm that that code doesn't work here neither
18:12:08Amun-Rahttps://play.nim-lang.org/#pasty=oGTqWLil
18:23:15*n|Phreak quit (Ping timeout: 256 seconds)
18:23:23*n|Phreak joined #nim
18:30:56*n|Phreak_ joined #nim
18:31:00*drawkula joined #nim
18:32:04*yeti quit (Killed (copper.libera.chat (Nickname regained by services)))
18:32:04*drawkula is now known as yeti
18:34:22*n|Phreak quit (*.net *.split)
18:34:23*Jjp137_ quit (*.net *.split)
18:34:23*computerquip quit (*.net *.split)
18:34:23*johuck quit (*.net *.split)
18:34:23*joast quit (*.net *.split)
18:34:23*GreaseMonkey quit (*.net *.split)
18:34:24*redj quit (*.net *.split)
18:34:50*Jjp137_ joined #nim
18:34:50*computerquip joined #nim
18:34:50*johuck joined #nim
18:34:50*joast joined #nim
18:34:50*GreaseMonkey joined #nim
18:34:50*redj joined #nim
18:37:25*yeti quit (Ping timeout: 256 seconds)
18:37:57*yeti joined #nim
18:37:59*nils` quit (Ping timeout: 256 seconds)
18:38:17*nils` joined #nim
18:42:15FromDiscord<Zoom> You're suppressing the error messages the OS gives you when the exception is thrown. Use `except IOError as e: echo e.msg`↵(@kiloneie)
18:43:21FromDiscord<Zoom> And don't catch defects, those are for logic errors in your code. Use asserts.
18:59:20FromDiscord<kiloneie> In reply to @Zoom "You're suppressing the error": Thank you 🙂
19:02:53FromDiscord<riku5543> In reply to @pmunch "Which ones?": There was a message about telegram in all of the primary channels but they got deleted
19:13:29*oisota9 joined #nim
19:15:21*oisota quit (Ping timeout: 258 seconds)
19:15:21*oisota9 is now known as oisota
19:17:49*SchweinDeBurg joined #nim
19:31:20*skippy8 joined #nim
20:05:13FromDiscord<nasuray> sent a code paste, see https://play.nim-lang.org/#pasty=gZRyqWUz
21:42:18*syl quit (Quit: C-x C-c)
21:44:32*syl joined #nim
21:45:24FromDiscord<cloudperry> https://github.com/nim-lang/Nim/issues/25181
21:45:51FromDiscord<cloudperry> Is this a bug or am I missing something obvious about how macros work?
21:46:41FromDiscord<Elegantbeef> You're taking in a typed section
21:46:53FromDiscord<cloudperry> Basically I'm trying to add align pragmas to object fields from a macro, but the align pragma does nothing
21:48:11FromDiscord<Elegantbeef> Change it to `untyped` and it'll work, you also can write a macro that you can use like `type MyType {.myMacro.} = object`
21:48:27FromDiscord<cloudperry> Ok thanks I will try that
21:49:16FromDiscord<cloudperry> So pragmas (or at least align) are applied already when taking in typed NimNodes?
21:49:59FromDiscord<Elegantbeef> Well typed is already semantically checked and compiled so adding things to `Type2` does nothing
21:50:12FromDiscord<Elegantbeef> You'd need to emit a new type with a different symbol
21:50:26FromDiscord<cloudperry> Ok makes sense
21:50:48FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=MORThLdG
21:50:58FromDiscord<Elegantbeef> You're going to have a lot of fun btw
21:51:49FromDiscord<cloudperry> Haha I already made the OpenGL UBO generator and it works other than the align pragma not applying
21:52:16FromDiscord<Elegantbeef> Yea I was going to do the same thing but was going to just going to detect it instead of annotate it
21:52:24FromDiscord<Elegantbeef> Easier that way 😄
21:53:30FromDiscord<cloudperry> Yep align does apply with untyped, but now I can't fix my macro to work with untyped
21:53:56FromDiscord<Elegantbeef> Right you need to remove the `nkSym` from the code and replace it with `nkIdent`
21:55:11FromDiscord<cloudperry> Wow Discord msg limits are extremely short https://media.discordapp.net/attachments/371759389889003532/1418717082622886019/message.txt?ex=68cf22be&is=68cdd13e&hm=ea99ec35f7a2c1066a449c282bb2f75c0558488558e3d12dacdbdc63b8ba9d16&
21:55:29FromDiscord<cloudperry> I shouldn't have made this as my literal first macro :D
21:55:50FromDiscord<Elegantbeef> You're making a dogs breakfast of it all indeed
21:55:56FromDiscord<cloudperry> Now its going to be a nightmare to understand how to fix this
21:56:30FromDiscord<Elegantbeef> Oh how happy I am to only ever use std430 😄
21:56:58FromDiscord<cloudperry> Wait do you not use UBOs? I thought UBOs can only be std140.
21:57:12FromDiscord<Elegantbeef> I only use SSBOs
21:57:22FromDiscord<cloudperry> Ok
21:58:00FromDiscord<cloudperry> Can I add new object fields from a typed macro though?
21:58:12FromDiscord<cloudperry> I could just add padding fields if thats possible
21:58:46FromDiscord<Elegantbeef> Like I said if you remove the symbols, yes
21:58:48*zgasma joined #nim
22:00:27*zgasma quit (Client Quit)
22:00:37*zgasma joined #nim
22:01:08FromDiscord<Elegantbeef> The fun part about your align table is that it does not work with aliases
22:01:46FromDiscord<cloudperry> Yep I know working with types as strings is messy
22:02:24FromDiscord<Elegantbeef> You really would want to use a `std/macrocache`'s `CacheSeq`
22:02:59FromDiscord<Elegantbeef> Then use `a == b or sameType(a, b) or sameType(b, a) # There is a bug in Nim that is still not fixed afaik`
22:03:16FromDiscord<cloudperry> I'm too tired to fix this mess so I will have to look at this another day
22:03:31FromDiscord<cloudperry> Or is there a simple fix for my existing macro?
22:04:01FromDiscord<Elegantbeef> Replace the `nkSym` with `nkIdent`
22:04:07FromDiscord<Elegantbeef> It's like 5 lines of code
22:04:42FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=NKcyFbXG
22:04:58FromDiscord<Elegantbeef> Run that after running your pragma macro
22:05:09FromDiscord<cloudperry> Yeah I'm just confused about where I even have an nkSym
22:05:26FromDiscord<Elegantbeef> `echo nimNode.treeRepr`
22:05:33FromDiscord<Elegantbeef> You have plenty of them cause you have `typed` code
22:05:39FromDiscord<Elegantbeef> Also sorry that's `nnkSym`
22:06:42FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=rZpTZiaX for example
22:07:41FromDiscord<cloudperry> Ok I see my type names/aliases are Sym and the object type is Sym
22:07:57FromDiscord<Elegantbeef> Yes identifiers are looked up inside a `typed` macro
22:10:15FromDiscord<cloudperry> In reply to @Elegantbeef "Run that after running": Uhh so I just need to do that to all nnkSym nodes at the end of my macro? Or some time after I have added the pragmas?
22:10:43FromDiscord<Elegantbeef> Well not before adding the pragmas cause you want the type information
22:11:35*beholders_eye quit (Ping timeout: 245 seconds)
22:19:30FromDiscord<cloudperry> Idk what I'm doing that proc tbh https://media.discordapp.net/attachments/371759389889003532/1418723198429696010/message.txt?ex=68cf2871&is=68cdd6f1&hm=77da2e8c8f2de64c114f3e2e27c65dc0ebdc770f9731822076b04b112176abba&
22:20:23FromDiscord<cloudperry> I tried desymming just the type of the fields, but that doesn't seem to do anything
22:20:54FromDiscord<cloudperry> The only other Sym I have in my AST is the object type itself
22:28:19FromDiscord<Elegantbeef> just do `result.desym()` a the end
22:29:49FromDiscord<cloudperry> Oh wow that actually worked! Thanks :D
22:33:03FromDiscord<cloudperry> My Nim OpenGL project has been going quite nicely before this macro mess. I made tiny abstractions for shaders, UBOs, SSBOs, VAOs, VBOs and EBOs. For now I don't plan to make any more abstractions, because other OpenGL things I need seem quite simple to use just through the GLAD wrapper.
22:33:45FromDiscord<Elegantbeef> My abstraction for SSBOs is just `type SSBO[T: array or seq] = distinct Gluint` 😄
22:34:54FromDiscord<Elegantbeef> https://codeberg.org/ElegantBeef/truss3D if you want a gander at any of my opengl work 😄
22:35:04FromDiscord<cloudperry> I'm going to do a CSG/SDF sphere tracing shader and also trying to speed it up by rasterizing some parts of the scene
22:35:52FromDiscord<cloudperry> Its for my bachelor's thesis about rendering constructive solid geometry in real-time, but we'll see if I ever get there
22:38:45FromDiscord<cloudperry> In reply to @Elegantbeef "https://codeberg.org/ElegantBeef/truss3D if you wan": Good to see that there is at least some slightly larger Nim OpenGL projects out there
22:39:15FromDiscord<cloudperry> I haven't seen much actual projects just hello triangle level stuff
22:39:25FromDiscord<cloudperry> (edit) "is" => "are"
22:40:15FromDiscord<cloudperry> [This GLFW wrapper](https://github.com/johnnovak/nim-glfw) + GLAD generated OpenGL has been really nice so far
22:41:06FromDiscord<cloudperry> Only thing I'm missing is some good Imgui wrapper that would be easy to use with GLFW, but I don't need UI yet
22:42:29FromDiscord<Elegantbeef> simplescreenrecorder-2025-08-26\_11.24.33.mp4 [simplescreenrecorder-2025-08-26_11.24.33.mp4](https://t2bot.io/_matrix/media/v3/download/matrix.org/IdLRrVqmcYxPgepeTZJpqPeu)
22:42:39FromDiscord<Elegantbeef> Heh I can say for certainty the opengl stuff works
22:43:08FromDiscord<Elegantbeef> I don't exactly like imgui so I don't use them
22:43:17FromDiscord<Elegantbeef> But I did take a gander at koi for a bit
22:45:02FromDiscord<cloudperry> The main reason I will probably use some immediate mode GUI for this project is that I just need some simple debug UI type things
22:45:19FromDiscord<cloudperry> And I don't want to spend too much time on that, because there are hard graphics things to solve
22:46:11FromDiscord<cloudperry> I didn't know about Koi and it seems pretty much perfect for this project since I'm already using nim-glfw
22:58:38FromDiscord<mitranim> sent a long message, see https://pasty.ee/lrCWsHYl
22:59:45FromDiscord<Elegantbeef> Yes you can mixin procedures inside a generic
23:00:12FromDiscord<mitranim> Where to look for examples?
23:01:26FromDiscord<Elegantbeef> https://github.com/beef331/nimtrest/wiki/Code-snippets#using-generic-interfaces
23:05:19FromDiscord<mitranim> Looking into this... what the heck is `concept type`? The language manual doesn't mention this, and the compiler internals are rather sparse on the topic 😅
23:06:42FromDiscord<mitranim> > In contrast to other programming languages, the vtable in Nim is stored externally to the object, allowing you to create multiple different vtable views for the same object. Thus, the polymorphism in Nim is unbounded - any type can implement an unlimited number of protocols or interfaces not originally envisioned by the type's author.↵Not in contrast with Go actually, the two languages completely align on this one point 😄
23:07:28FromDiscord<Elegantbeef> Experimental feature, don't worry about it
23:07:39FromDiscord<mitranim> Oh wait, they don't! In Go, you can't extend the definition of a type from another package, which is sometimes a pain point. This paragraph implies you can in Nim.
23:07:49FromDiscord<Elegantbeef> Nim does not have vtables so idk where you got that
23:07:56*skippy8 quit (Quit: WeeChat 4.7.1)
23:08:13FromDiscord<mitranim> `doc/manual_experimental.md`
23:08:16FromDiscord<Elegantbeef> You don't use interfaces in Nim you use generics which instantiate on invoked
23:09:20FromDiscord<Elegantbeef> Nim types are just types procedures are unattached
23:09:44FromDiscord<mitranim> That works for single values, and also results in static dispatch, and that's great. But Go interface values / Rust boxed traits / "fat pointers" from this paragraph are required when constructing collections of values with different underlying concrete types which implement the same interface
23:10:24FromDiscord<mitranim> Also to me this seems like the only sane approach; bundling vtables _into_ objects is madness
23:11:00FromDiscord<mitranim> (edit) "That works for single values, and also results in static dispatch, and that's great. But Go interface values / Rust boxed traits / "fat pointers" from this paragraph are required when constructing collections of values with different underlying concrete types which implement the same interface ... " added "(at the cost of more memory usage and dynamic dispatch)"
23:11:38FromDiscord<Elegantbeef> Nim has no mechanism for this
23:11:54FromDiscord<Elegantbeef> You have to manually make you objects and bring the table around
23:18:30*xtr00 quit (Read error: Connection reset by peer)
23:18:58FromDiscord<mitranim> sent a code paste, see https://play.nim-lang.org/#pasty=nnicEHeH
23:22:53FromDiscord<Elegantbeef> You're just using generics
23:22:53FromDiscord<Elegantbeef> The concept does no binding though
23:22:54FromDiscord<Elegantbeef> Which means static dispatch
23:23:30FromDiscord<mitranim> Exactly how it should be. My worry was whether the compiler is able to pass the interface implementation back from user code to library code
23:24:04FromDiscord<mitranim> As in, does the language depend _entirely_ on in-scope proc overloads for such behaviors or not
23:25:24FromDiscord<mitranim> (edit) "not" => "can interface implementations cross module boundaries"
23:26:08*xtr00 joined #nim
23:29:54FromDiscord<Elegantbeef> You need in scope procedures at the place of instantiation
23:29:59FromDiscord<Elegantbeef> Like I said
23:30:16FromDiscord<Elegantbeef> You do not even need the concept
23:30:42FromDiscord<mitranim> So this doesn't carry over to a _third_ module which uses library code from A and type from B?
23:30:53FromDiscord<mitranim> Unless it imports everything from B including its proc?
23:38:37FromDiscord<mitranim> sent a code paste, see https://play.nim-lang.org/#pasty=DkyqQZAH
23:39:22FromDiscord<mitranim> (edit) "https://play.nim-lang.org/#pasty=gxRMNpLH" => "https://play.nim-lang.org/#pasty=YoaEEDFZ"
23:39:36FromDiscord<mitranim> (edit) "https://play.nim-lang.org/#pasty=tmJqeEGq" => "https://play.nim-lang.org/#pasty=grZZVVTX"
23:48:16FromDiscord<Elegantbeef> So perhaps stop using that import method 😄