| 00:03:35 | rockcavera | Is there any way to transform a `ptr T` into an `openArray[T]`? The size of `ptr T` is known |
| 01:10:11 | FromDiscord | <Elegantbeef> Rockavera `cast[ptr UncheckedArray[T]](yourPtr).toOpenArray(0, len -1 )` |
| 01:10:55 | * | xet7 joined #nim |
| 01:12:38 | * | krux02_ quit (Remote host closed the connection) |
| 01:20:49 | rockcavera | Elegantbeef, I didn't know there was a `toOpenArray()` for `UncheckedArray[T]`. Thanks |
| 01:59:02 | FromDiscord | <sOkam! 🫐> In reply to @PMunch "<@259277943275126785>, well you can,": wait, please explain 👀 |
| 02:02:31 | FromDiscord | <sOkam! 🫐> In reply to @strogon14 "I'm using nimscript, i.e.": I made `confy` for this reason. might be worth taking a look if you are already pushing nimscript to its limit with such "complex" tasks as making a symlink↵_sarcastic, btw. its a simple task, and easily doable in non-nimscript. but it's not exposed to the vm, because... nimscript 🤷♂️_↵https://github.com/heysokam/confy |
| 02:03:50 | strogon14 | sOkam!, thanks, I'll definitely have a look. |
| 02:05:59 | FromDiscord | <sOkam! 🫐> @ElegantBeef do you know anything about that comment from @pmunch about not importing the implictely imported `system`? |
| 02:15:32 | * | SchweinDeBurg quit (Quit: WeeChat 4.3.0-dev) |
| 02:23:22 | * | def- quit (Quit: -) |
| 02:24:30 | * | def- joined #nim |
| 02:27:42 | * | def- quit (Client Quit) |
| 02:27:53 | * | def- joined #nim |
| 02:57:55 | FromDiscord | <aryzen> So, maybe I'm approaching this from the wrong angle, but I'm trying to use "cinterop" which supposedly works with complex libraries (Creator said it works for Diligent, which I know is more complex than Filament), but it has little documentation; I haven't even begun trying to link the binaries I'm already having trouble getting it to find the headers |
| 02:59:21 | FromDiscord | <aryzen> It recognises when I ask it to open a header file myself in nim, but if that header wants to include another header, it can't find it; even though I'm asking it to open that exact header file too |
| 03:13:33 | FromDiscord | <aryzen> I mean, I know I'm doing something wrong because I can't get the compiler arguments right either, supposedly I should use a cpp switch but I can't get it to recognise it |
| 03:21:05 | FromDiscord | <aryzen> I managed to get cpp to work but my error messages are now kinda insane https://media.discordapp.net/attachments/371759389889003532/1233256513124958319/image.png?ex=662c6f21&is=662b1da1&hm=da58f34b9149524cc1f834f50bf6eef8c6044e2fd70adc6a41c8fb638a0fae82& |
| 03:40:53 | FromDiscord | <aryzen> alright... a lot further along than I was before now. I'm keenly in the "I don't know how cpp works" realm now, since I don't know wtf a protected constructor and destructor is |
| 03:51:18 | * | SchweinDeBurg joined #nim |
| 04:56:18 | FromDiscord | <pmunch> In reply to @heysokam "<@145405730571288577> do you know": Sorry, I went to bed 😅 https://nim-lang.org/docs/nimscript.html#patchFile%2Cstring%2Cstring%2Cstring I believe you can replace system by putting this in your config.nims. alternatively you can use `--lib` but that changes the entire standard library path |
| 06:04:43 | * | rockcavera quit (Remote host closed the connection) |
| 06:34:19 | * | PMunch joined #nim |
| 06:39:37 | FromDiscord | <pmunch> Hmm, it doesn't seem like `patchFile` works for system.. That's a shame |
| 06:54:44 | PMunch | Had a poke around the compiler and system.nim is in fact treated different than other modules and thus doesn't pass through the patchFile check |
| 06:54:59 | PMunch | A bit unfortunate, but should be easily fixable if anyone cares to do it |
| 07:50:03 | FromDiscord | <sp8061> hi, im new to nim |
| 07:51:13 | FromDiscord | <sp8061> i was looking to use this webgpu style library https://github.com/heysokam/ngpu?tab=readme-ov-file |
| 07:52:14 | FromDiscord | <sp8061> so I got the src directory and the first example, the e00_hellongpu.nim↵cd to it and "nim c -r e00_hellongpu.nim" |
| 07:52:32 | FromDiscord | <sp8061> e00_hellongpu.nim(9, 8) Error: cannot open file: nsys |
| 07:52:51 | FromDiscord | <Elegantbeef> You did not do `nimble install -d` |
| 07:57:16 | FromDiscord | <sp8061> that- that makes sense |
| 07:57:19 | FromDiscord | <sp8061> thanks |
| 08:56:29 | FromDiscord | <sOkam! 🫐> In reply to @sp8061 "so I got the": `nimble hello` |
| 08:57:08 | FromDiscord | <sOkam! 🫐> https://github.com/heysokam/ngpu/blob/cbf8ec66edcb6cac7729e3c0b7bb71b52a5954cc/ngpu.nimble#L39-L52 |
| 08:58:12 | FromDiscord | <sOkam! 🫐> In any case, its likely that none of them will work because I deprecated wgpu and I haven't updated it in a while↵the readme in both ngpu and wgpu is a must-read, it explains everything |
| 08:59:59 | FromDiscord | <sp8061> sent a code paste, see https://play.nim-lang.org/#pasty=UJPxipxPWuSZ |
| 09:00:35 | FromDiscord | <sp8061> for some reason this happens, im trying to edit the nim files in your wgpu library but to no avail |
| 09:00:48 | FromDiscord | <sOkam! 🫐> yeah because I figured nobody was using ngpu anyway, and Its deprecating wgpu, so I started working on a futhark wgpu wrapper but didn't fully implement it |
| 09:01:24 | FromDiscord | <sp8061> oh 😦 is there an alternative' |
| 09:01:25 | FromDiscord | <sp8061> ? |
| 09:02:00 | FromDiscord | <sOkam! 🫐> in short, I recommend cloning an older commit of the wgpu wrapper and going that route↵or I can help you fix whatever issues come up in the wgpu futhark wrapper, but it would be tedious |
| 09:02:08 | FromDiscord | <sOkam! 🫐> ngpu will definitely not work for a long time |
| 09:02:50 | FromDiscord | <sOkam! 🫐> wgpu-native is a C library, so it can be wrapped with futhark "fairly easily"↵although if you are new to nim, that easy won't be too easy |
| 09:03:26 | FromDiscord | <sp8061> indeed haha. I will try the old clone for now |
| 09:03:42 | FromDiscord | <sOkam! 🫐> let me try to find you that commit |
| 09:03:59 | FromDiscord | <sp8061> From what I understood, ngpu wasn't only a wrapper right? it attempted to also make life easier |
| 09:05:01 | FromDiscord | <sOkam! 🫐> @sp8061 b88fb2dcf80eda0f94b77c407bf6b3ba6afb95a5 |
| 09:05:44 | FromDiscord | <sOkam! 🫐> In reply to @sp8061 "From what I understood,": yes, but its not finished by any means, and i'm refactoring it to be built in C with Vulkan. so it will be a long while until its usable again (think a couple of years, given the current pace) |
| 09:15:32 | FromDiscord | <sp8061> I understand 100% |
| 09:16:20 | FromDiscord | <sp8061> I started using WebGPU with JS and although the setup took a while I kinda learnt it and did my own examples |
| 09:17:36 | FromDiscord | <sp8061> Issue is, I want to develop a native app, on a decently fast language, and the little maturity of Dawn and webgpu-native are cornering me into c/c++ which I'm not experienced with |
| 09:19:05 | FromDiscord | <sp8061> i think it's time for me to man up and do it in c++ 😢 😄 |
| 09:19:34 | * | def- quit (Quit: -) |
| 09:20:00 | * | def- joined #nim |
| 09:21:35 | * | ntat joined #nim |
| 09:25:40 | * | def- quit (Quit: -) |
| 09:27:13 | * | def- joined #nim |
| 09:47:46 | FromDiscord | <sOkam! 🫐> In reply to @sp8061 "i think it's time": I would be willing to help setup the futhark wrapper for wgpu-native |
| 09:48:31 | FromDiscord | <sOkam! 🫐> I haven't done it because I haven't had a personal reason to revive it yet. But if you want to give it a go, and help debug it, I can give you a hand and collab so we have a reliable wgpu wrapper again |
| 10:05:43 | FromDiscord | <sp8061> realistically there is little time for me to invest on anything for the next half a year so I wouldn't be able to contribute |
| 10:06:00 | FromDiscord | <sp8061> but thanks a lot for the offer, really |
| 10:22:52 | * | coldfeet joined #nim |
| 10:30:07 | * | coldfeet quit (Remote host closed the connection) |
| 11:39:25 | * | krux02 joined #nim |
| 12:00:30 | * | def- quit (Quit: -) |
| 12:00:47 | * | def- joined #nim |
| 12:12:22 | PMunch | Playing around some more with libfuse. Incorporated libvlc and metatag (will probably replace metatag with functions from libvlc in the future) and now I have a small music player as a file system :P |
| 12:14:01 | PMunch | You give it a folder and it creates by-artist, by-album, and by-year folders, along with a control folder. In the by-X folders there is a bunch of song names as files, the content of a file is the path to the original file. To play a song you cat a song into control/current and write "play" to control/command. You can also write a list of files to control/up-next and they will be played in order. |
| 12:14:49 | PMunch | So if you want to shuffle all songs by an artist you do something like `cat by-artist/<artist>/* | sort --random > control/up-next` |
| 12:15:28 | FromDiscord | <nervecenter> In reply to @patitotective "Hi, I'm using datamancer": Try deep copying the frame over to the thread before doing anything with it? |
| 12:44:29 | FromDiscord | <patitotective> Well, the dataframe is created inside the thread, so I don't think that would be the problem |
| 13:13:05 | * | def- quit (Quit: -) |
| 13:13:54 | * | def- joined #nim |
| 13:25:12 | * | rockcavera joined #nim |
| 13:55:07 | FromDiscord | <vindaar> In reply to @patitotective "Well, the dataframe is": as long as the DF is created on the separate thread, it should work fine. if you can share more code, I can try to help more. Sounds more like an issue with the compiler being too paranoid |
| 13:58:38 | FromDiscord | <patitotective> In reply to @vindaar "as long as the": Well, I didn't want to create the DF on another thread since it's created from a 18MB CSV file and I don't want the GUI to freeze while loading the CSV |
| 14:11:04 | NimEventer | New thread by ninovanhooff: Please review my types: ref object or plain object, see https://forum.nim-lang.org/t/11499 |
| 14:11:42 | * | pbsds quit (Ping timeout: 252 seconds) |
| 14:11:57 | * | pbsds3 joined #nim |
| 14:24:32 | * | ntat quit (Quit: Leaving) |
| 14:34:20 | FromDiscord | <hyperdbg> hey, how many TCP socket can I create from one device in parallel and without interruption use them?↵does this depend on networking hardware? if so, which exactly? I don |
| 14:34:37 | FromDiscord | <hyperdbg> (edit) "don" => "don't control my ethernet cable, but I can buy, for example, better router\switch" |
| 14:46:32 | * | ntat joined #nim |
| 15:31:47 | * | Mister_Magister quit (Excess Flood) |
| 15:32:13 | * | Mister_Magister joined #nim |
| 16:42:15 | FromDiscord | <wileycoleman> sent a long message, see https://pasty.ee/oDukrVQscWdR |
| 16:45:45 | * | krux02 quit (Remote host closed the connection) |
| 16:50:17 | FromDiscord | <threefour> In reply to @hyperdbg "hey, how many TCP": What do you mean without interruption? I believe sockets are essentially just a software abstraction for multiplexing your network interface(s), which are all serial, including Ethernet, afaik. |
| 16:51:14 | FromDiscord | <threefour> The socket limit depends on what your file descriptor limit is set to, which on Linux, you'd do via `ulimit`. |
| 16:53:03 | FromDiscord | <hyperdbg> In reply to @threefour "What do you mean": imagine me opening N connection to Z hosts. Intuitevely, aside from software limits there are hardware limits as well. That's the my questions, in case of N connection to Z hosts, what is the first bottleneck cause that I should look for (in networking hardware)? |
| 16:53:09 | FromDiscord | <hyperdbg> (edit) "connection" => "connections" |
| 17:00:49 | FromDiscord | <aintea> sent a long message, see https://pasty.ee/sSoqTtFxtNQs |
| 17:01:54 | strogon14 | you can always generalize things too far. |
| 17:02:31 | * | SchweinDeBurg quit (Quit: WeeChat 4.3.0-dev) |
| 17:08:10 | FromDiscord | <polylokh_39446> a loop is less lines of assembly than the unrolled loop. The CPU might execute the unrolled loop faster because there are no branches to predict and no branch-predictions to fail. |
| 17:09:10 | FromDiscord | <.lisuwu_> In reply to @aintea "Nim compiler has two": by speed it means nim compiler will focus on speed, then nim compiler might add some things to make it faster↵↵i cant really think of any examples where nim compiler might do things, but i heard of this example that might happen in C side:↵if you do `x 15`, the C compiler might convert that into `(x << 4) - x`, overall ending up faster because CPU can do bitshifts faster than it can d |
| 17:09:25 | FromDiscord | <.lisuwu_> (edit) "In reply to @aintea "Nim compiler has two": by speed it means nim compiler will focus on speed, then nim compiler might add some things to make it faster↵↵i cant really think of any examples where nim compiler might do things, but i heard of this example that might happen in C side:↵if you do `x 15`, the C compiler might convert that into `(x << 4) - x`, overall ending up faster because CPU can do bitshifts faster than |
| 17:11:43 | FromDiscord | <polylokh_39446> in general, you're probably viewing the CPU as much simpler and mechanical than it is. With something like intel, proprietary microcode, deep pipelines and layers of cache and branch prediction, a better mental model is probably that your assembly is high level program code for a virtual machine with a JIT compiler, than like 'assembly'. |
| 17:14:48 | FromDiscord | <polylokh_39446> anyway the conventional explanation for the flags that there are space/speed tradeoffs that Nim will consider, like inlining and loop unrolling, that wil lead to larger binaries but maybe faster execution. And an example that still works with a primitive-CPU model is inlining: there is "more machine code" in the binary, but on any given execution path there is actually less machine code because you're skipping calls/returns and doi |
| 17:15:00 | FromDiscord | <polylokh_39446> (edit) "wil" => "will" |
| 17:19:16 | FromDiscord | <demotomohiro> In reply to @aintea "Nim compiler has two": Optimizing speed might not inline functions or expand loop in order to generate smaller code while optimizing speed might do both. |
| 17:30:25 | Amun-Ra | "Optimizing speed" → "Optimizing size" |
| 17:43:42 | FromDiscord | <Robyn [She/Her]> The urge to make a custom package manager for Nim :p |
| 17:44:02 | FromDiscord | <aintea> It's pretty easy to make one |
| 17:44:17 | FromDiscord | <Robyn [She/Her]> Is it? |
| 17:44:33 | FromDiscord | <aintea> yes |
| 17:44:34 | FromDiscord | <aintea> here |
| 17:44:48 | Amun-Ra | but… why? |
| 17:45:11 | FromDiscord | <aintea> sent a code paste, see https://play.nim-lang.org/#pasty=EwIMBYZnwHjA |
| 17:45:19 | FromDiscord | <aintea> run this in your shell and you're good to go |
| 17:45:42 | FromDiscord | <Robyn [She/Her]> In reply to @Amun-Ra "but… why?": Boredom and because the lang server has ruined my motivation to work on my parser for now lol |
| 17:45:45 | Amun-Ra | won't work in my shell |
| 17:45:51 | FromDiscord | <aintea> nushell |
| 17:45:54 | FromDiscord | <aintea> or fish |
| 17:45:58 | Amun-Ra | fish |
| 17:46:05 | FromDiscord | <Robyn [She/Her]> In reply to @aintea "run this in your": >.> |
| 17:46:12 | Amun-Ra | but won't work in dash, csh, ksh, zsh either ;) |
| 17:46:17 | FromDiscord | <Robyn [She/Her]> In reply to @Amun-Ra "fish": Eyyyy, same |
| 17:46:21 | FromDiscord | <aintea> works in zsh |
| 17:46:26 | Amun-Ra | hmm |
| 17:46:35 | FromDiscord | <aintea> because zshrc is in ~/.zshrc |
| 17:47:04 | Amun-Ra | I thought ${SHELL##/} is bashism |
| 17:47:29 | FromDiscord | <aintea> nope, it's just taking `$SHELL` and only getting what's after the last `/` |
| 17:47:44 | FromDiscord | <aintea> so `/bin/zsh` would return zsh |
| 17:47:46 | Amun-Ra | yes, the problematic part is ##/ |
| 17:47:48 | FromDiscord | <aintea> (edit) "zsh" => "`zsh`" |
| 17:48:17 | FromDiscord | <Robyn [She/Her]> Anyone know what the point of a JSON library that uses 'pointers' is? Is it quicker and less memory intensive to access keys or? |
| 18:01:26 | FromDiscord | <odexine> Pointers usually make things slower |
| 18:01:33 | FromDiscord | <Robyn [She/Her]> Make sense |
| 18:01:39 | * | lucasta joined #nim |
| 18:01:43 | Amun-Ra | depends vs what |
| 18:02:04 | FromDiscord | <Robyn [She/Her]> Something like Jsony when it's kept in memory |
| 18:02:37 | Amun-Ra | for example, I wouldn't say pointers are slower than copying whole objects |
| 18:12:19 | FromDiscord | <Robyn [She/Her]> That also checks out :p |
| 18:12:35 | FromDiscord | <Robyn [She/Her]> I'd assume pointers would also be much more compact? |
| 18:15:33 | Amun-Ra | but with great power comes great responsiblity |
| 18:22:55 | FromDiscord | <nocturn9x> question |
| 18:23:14 | FromDiscord | <nocturn9x> can I use nim's views to tell `std/algorithm`'s `sort` to only sort a subset of an array |
| 18:23:46 | FromDiscord | <nocturn9x> I use an array of static size with a lenght field because it's cheaper than allocating and sizing a dynamic list every time, but that means not all elements of the array are valid |
| 18:24:08 | FromDiscord | <nocturn9x> and I want these invalid elements (which are all at the end) to be sliced off the sorting |
| 18:24:23 | FromDiscord | <nocturn9x> do I have to reimplement like quicksort or is it possible to do in some other way? |
| 18:36:21 | FromDiscord | <demotomohiro> In reply to @nocturn9x "question": How about to use this?↵https://nim-lang.org/docs/system.html#toOpenArray%2Carray%5BI%2CT%5D%2CI%2CI |
| 18:37:58 | FromDiscord | <nocturn9x> does that copy |
| 18:40:13 | FromDiscord | <demotomohiro> I don't think it creates copy. It should just returns sub-range of array as openArray. |
| 18:41:00 | FromDiscord | <demotomohiro> I also wrote array + lenght field module: https://github.com/demotomohiro/littlesugar/blob/main/src/littlesugar/staticSeq.nim |
| 18:42:35 | FromDiscord | <demotomohiro> And it has `toOpenArray` with bound checks: https://github.com/demotomohiro/littlesugar/blob/main/src/littlesugar/staticSeq.nim#L117 |
| 18:43:32 | FromDiscord | <Phil> In reply to @wileycoleman "I get an error": I'm not aware of active gintro usage so I can't claim to have a huge understanding there.↵I think Stefan lurks in the discord so he might answer, but that's asking for luck |
| 18:44:13 | FromDiscord | <Phil> He's using matrix though and goes by the user StefanSalewski there |
| 19:30:47 | * | coldfeet joined #nim |
| 19:38:10 | * | lucas_ta joined #nim |
| 19:38:40 | * | lucasta quit (Ping timeout: 245 seconds) |
| 19:53:34 | * | lucas_ta quit (Quit: Leaving) |
| 19:54:37 | * | ntat quit (Quit: Leaving) |
| 20:50:15 | * | def- quit (Quit: -) |
| 20:51:02 | * | def- joined #nim |
| 21:07:24 | FromDiscord | <norbert3873> sent a code paste, see https://play.nim-lang.org/#pasty=qHEzyuVTdxyZ |
| 21:07:29 | * | krux02 joined #nim |
| 21:08:20 | FromDiscord | <nnsee> In reply to @norbert3873 "Since upgrading to 2.0.x,": i've personally never seen that. what does your nim.cfg look like? |
| 21:10:25 | FromDiscord | <norbert3873> In reply to @nnsee "i've personally never seen": Not sure I have one... I'm using choosenim |
| 21:11:32 | FromDiscord | <norbert3873> derp found it let me peruse it for anything abnormal |
| 21:14:16 | FromDiscord | <Elegantbeef> @norbert3873 `--deepcopy` must be defined in the project config file on any projects that depend on code that uses it |
| 21:14:26 | FromDiscord | <Elegantbeef> This means that if you import code that depends on it you need to make a `config.nims` with it |
| 21:15:38 | FromDiscord | <norbert3873> In reply to @Elegantbeef "This means that if": is the naming `config.nims` important here? I would thing a `mypkgname.config` would work, same with pasing it to `nimble build` |
| 21:15:51 | FromDiscord | <norbert3873> (edit) "thing" => "think" |
| 21:16:11 | FromDiscord | <Elegantbeef> you can do `config.nims` or put a `mypkgname.nims` next to the entry file |
| 21:16:38 | FromDiscord | <Elegantbeef> But config.nims are only searched for in the present directory or the entry file |
| 21:23:07 | FromDiscord | <norbert3873> Thanks beef. It was an issue with the upstream package not building, installing with nimble let me fallback to #head and that built |
| 21:23:25 | FromDiscord | <Elegantbeef> You need to do `nimble build --deepcopy:on` |
| 21:23:28 | FromDiscord | <Elegantbeef> Or make a config file |
| 21:23:31 | FromDiscord | <zidsal> sent a code paste, see https://play.nim-lang.org/#pasty=VueqXJmltWZN |
| 21:23:43 | FromDiscord | <Elegantbeef> deepcopy is explicitly not enabled by default as it's expected to be removed |
| 21:23:51 | FromDiscord | <Elegantbeef> Enums and unicode operators are default now |
| 21:23:54 | FromDiscord | <Elegantbeef> strictfuncs has changed |
| 21:24:01 | FromDiscord | <Elegantbeef> Orc is also default |
| 21:24:07 | FromDiscord | <zidsal> thanks beef |
| 21:24:16 | FromDiscord | <norbert3873> the error was confusing me because of the `arc/orc` mention in the error. Thanks for the help. |
| 21:38:23 | * | coldfeet quit (Remote host closed the connection) |
| 21:52:28 | * | SEP quit (Ping timeout: 268 seconds) |
| 21:54:13 | * | SEP joined #nim |
| 22:20:21 | * | def- quit (Quit: -) |
| 22:39:01 | * | def- joined #nim |
| 23:23:36 | * | krux02 quit (Remote host closed the connection) |
| 23:24:20 | * | SchweinDeBurg joined #nim |
