| 00:37:04 | FromDiscord | <System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#pasty=QCoQrooi |
| 00:40:41 | FromDiscord | <Elegantbeef> You can set the destructor for the object it points to `proc =destroy(obj: typeof(ModuleSynthComputer()[])) = ....` |
| 00:41:10 | FromDiscord | <Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=bsHwlgsx |
| 00:41:36 | FromDiscord | <Elegantbeef> Bit odd that you're namespacing your types if you ask me 😄 |
| 00:42:48 | FromDiscord | <System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#pasty=ELdnXxli |
| 00:43:03 | FromDiscord | <Elegantbeef> You're prefixing them with `Module` |
| 00:43:32 | FromDiscord | <System64 ~ Flandre Scarlet> Should I remove the prefix? |
| 00:43:55 | FromDiscord | <Elegantbeef> I would as it's just pointless verbiage if you want my view |
| 00:54:47 | FromDiscord | <.tokyovigilante> Hey, on my continuing journey to use Nim to import every C library there is, I've extended the `wayland-scanner` tool to generate a Nim FFI module to be able to use protocol extensions, along the same lines as Futhark, and am seeing a similar problem where I try to pass a C struct declared externally by pointer to another Nim function in another module, and getting an incompatible type error. |
| 00:55:23 | FromDiscord | <.tokyovigilante> sent a code paste, see https://play.nim-lang.org/#pasty=ajbUijeg |
| 00:56:51 | FromDiscord | <Elegantbeef> I do see `interface` is there which is a keyword |
| 00:58:20 | FromDiscord | <.tokyovigilante> Ah yeah that could be it... |
| 01:00:57 | FromDiscord | <Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=hbpgXSAM |
| 01:06:02 | FromDiscord | <.tokyovigilante> I thought I had escaped them all but maybe not |
| 01:06:41 | FromDiscord | <.tokyovigilante> `proc wl_registry_bind (wl_registry: ptr struct_wl_registry, name: uint32, `interface`: ptr struct_wl_interface, version: uint32): pointer =↵` |
| 01:07:04 | FromDiscord | <.tokyovigilante> (edit) "`proc wl_registry_bind (wl_registry: ptr struct_wl_registry, name: uint32, `interface`: ptr struct_wl_interface, version: uint32): pointer =↵`" => "sent a code paste, see https://paste.rs/lmH83" |
| 01:07:18 | FromDiscord | <.tokyovigilante> yup did escape it |
| 01:12:37 | FromDiscord | <.tokyovigilante> I found this by just casting to a `pointer` before passing worked when I saw the same issue with Vulkan |
| 01:22:10 | FromDiscord | <.tokyovigilante> is there any way to get any extra type information from the compiler? The error doesn't really make any sense as the types should be the same |
| 01:22:13 | FromDiscord | <eightbit_dboy> I added documentation for my latest project, `nerc`, the Nim web anti-framework: https://github.com/8bitprodigy/nerc |
| 01:49:41 | * | shrewd joined #nim |
| 03:08:25 | * | rockcavera quit (Remote host closed the connection) |
| 04:14:46 | * | shrewd quit (Ping timeout: 252 seconds) |
| 05:30:05 | * | ntat joined #nim |
| 05:48:46 | FromDiscord | <Elegantbeef> Well @.elcritch it's not overly informative or well written but I did at least sorta document the HCR path I went down https://www.jasonbeetham.ca/writeups/hcr.html 😄 |
| 05:49:19 | FromDiscord | <Elegantbeef> \Probably should expand it to include the full `HcrObj` I describe there, but it's a lot of code |
| 07:22:06 | * | coldfeet joined #nim |
| 07:34:38 | * | coldfeet quit (Quit: Lost terminal) |
| 08:43:17 | FromDiscord | <.elcritch> In reply to @Elegantbeef "Well <@703717429230174229> it's not": No hipster photo at the top?! 😛 |
| 08:50:00 | PMunch | "Fun" fact, there is no way to tell Nim to free all resources held by a dynamic library when you close it |
| 08:51:24 | PMunch | More relevant when loading dynamic libraries into non-Nim applications though |
| 08:55:46 | * | coldfeet joined #nim |
| 08:57:56 | FromDiscord | <odexine> In reply to @Elegantbeef "Well <@703717429230174229> it's not": Time to proofread |
| 09:02:56 | * | coldfeet quit (Quit: Lost terminal) |
| 09:36:53 | * | derpydoo joined #nim |
| 10:20:17 | * | ntat quit (Quit: Leaving) |
| 10:31:27 | * | derpydoo quit (Ping timeout: 276 seconds) |
| 10:45:39 | * | ntat joined #nim |
| 10:55:23 | * | FromDiscord quit (Remote host closed the connection) |
| 10:55:59 | * | FromDiscord joined #nim |
| 11:35:13 | * | ntat quit (Quit: leaving) |
| 11:35:30 | * | ntat joined #nim |
| 11:38:34 | * | ntat quit (Client Quit) |
| 11:41:11 | * | ntat joined #nim |
| 12:07:48 | * | ntat quit (Read error: Connection reset by peer) |
| 12:27:28 | FromDiscord | <System64 ~ Flandre Scarlet> In reply to @PMunch ""Fun" fact, there is": Using destructors maybe? |
| 12:32:43 | FromDiscord | <dawidek.2137> when compiled to wasm32, reading a field from a nil variable returns garbage, is there a way to make nim error out? |
| 13:10:26 | * | derpydoo joined #nim |
| 13:17:25 | FromDiscord | <Robyn [She/Her]> In reply to @dawidek.2137 "when compiled to wasm32,": could manually check if it's nil... do you get this behaviour in native x86? |
| 13:38:42 | * | derpydoo quit (Ping timeout: 252 seconds) |
| 13:41:29 | * | derpydoo joined #nim |
| 14:09:30 | * | ntat joined #nim |
| 14:22:41 | * | CypherCat joined #nim |
| 14:56:57 | * | derpydoo quit (Quit: derpydoo) |
| 15:09:04 | PMunch | @System64_~_Flandre_Scarlet, nope. |
| 15:09:06 | * | PMunch quit (Quit: Leaving) |
| 15:23:49 | * | ntat quit (Quit: Leaving) |
| 15:24:42 | * | ntat joined #nim |
| 15:54:53 | * | coldfeet joined #nim |
| 15:58:40 | FromDiscord | <alehander92> hey |
| 15:58:46 | FromDiscord | <alehander92> we're trying to cross compile |
| 15:58:53 | FromDiscord | <alehander92> codetracer's CLI to windows |
| 16:03:37 | FromDiscord | <alehander92> i have almost no xp with nim on windows |
| 16:03:46 | FromDiscord | <alehander92> is mingw still the best way to do that ( i assume so) |
| 16:04:00 | * | coldfeet quit (Quit: Lost terminal) |
| 16:39:11 | * | PotentialUser-25 joined #nim |
| 16:39:38 | * | PotentialUser-25 quit (Client Quit) |
| 17:14:00 | * | ntat quit (Remote host closed the connection) |
| 17:15:01 | * | ntat joined #nim |
| 17:54:14 | FromDiscord | <Elegantbeef> @alehander92 you can also use zig to cross compile, which uses clang |
| 18:00:27 | FromDiscord | <Elegantbeef> Pmunch you could use a dynamic library destructor that calls your own procedure to free resources, it's not pretty and means all global variables have to be manually destroyed. It should work though. Especially you use malloc instead of the Nim allocator(which let's be honest is wise so you can share resources) |
| 18:15:21 | FromDiscord | <alehander92> In reply to @Elegantbeef "Well <@703717429230174229> it's not": What happened to the built-in experimental hcr support in Nim |
| 18:15:40 | FromDiscord | <alehander92> In reply to @Elegantbeef "<@384062055080001545> you can also": Thank you, it seems mingw works, but we're aware of zig as well |
| 18:16:56 | FromDiscord | <Elegantbeef> it never left experimental 😄 |
| 18:18:03 | FromDiscord | <Elegantbeef> You also need the NimRTL(less so now with Arc/Orc) plus I don't think it does data migration |
| 18:18:17 | FromDiscord | <Elegantbeef> So if you add a new field to an object you have to restart |
| 18:20:02 | FromDiscord | <Elegantbeef> Potato on the otherhand allows new fields, has a watcher so you don't have to manually reload, and allows using the same module for entry for a real program and a hot code program(with a few `when`s) |
| 18:22:41 | FromDiscord | <alehander92> Sounds very interesting, nice ! |
| 18:23:26 | FromDiscord | <Elegantbeef> Oh just remembered don't know if you've checked with virustotal using the mingw build, but in my experience antiviruses hate mingw. A C hello world compiled with mingw was reported as malware by like 17 separate vendors |
| 21:24:18 | * | rockcavera joined #nim |
| 22:07:50 | * | rockcavera quit (Remote host closed the connection) |
| 22:10:38 | * | rockcavera joined #nim |
| 22:25:08 | * | Jhonny2x4 quit (Quit: Jhonny2x4) |
| 22:25:16 | * | Jhonny2x4 joined #nim |
| 22:25:42 | FromDiscord | <Robyn [She/Her]> I wonder if it would be stupid for when allowing WASM runtimes to spawn their own WASM runtime, the memory would have to be `malloc`ed by the 1st runtime, and then the host system would start a runtime with that section of the memory being where the data is written and read from |
| 22:59:05 | * | ntat quit (Quit: Leaving) |
| 23:10:23 | FromDiscord | <Elegantbeef> @Robyn [She/Her] It's runtimes all the way down |
| 23:10:23 | FromDiscord | <Elegantbeef> I feel like that's a bad idea |
| 23:10:24 | FromDiscord | <Elegantbeef> Just depend upon another module |
| 23:20:09 | FromDiscord | <Robyn [She/Her]> In reply to @Elegantbeef "<@524288464422830095> It's runtimes all": Runtime is still managed by the host, it's just that the WASM runtime sacrifices it's own memory to give to a child WASM runtime |
| 23:20:27 | FromDiscord | <Elegantbeef> But why would you want that? |
| 23:20:45 | FromDiscord | <Robyn [She/Her]> To ensure memory limits aren't bypassed and a semblance of safety is kept |
| 23:21:00 | FromDiscord | <Elegantbeef> I mean why would you want wasm inside wasm? |
| 23:21:12 | FromDiscord | <Elegantbeef> Why would you not just want to depend upon another module and load both into the same environment? |
| 23:21:43 | FromDiscord | <Robyn [She/Her]> In reply to @Elegantbeef "I mean why would": JIT but in WASM :> |
| 23:21:56 | FromDiscord | <Elegantbeef> There are JIT'd wasm runtimes though |
| 23:21:58 | FromDiscord | <Robyn [She/Her]> Tho ig it wouldn't be too performant |
| 23:22:24 | FromDiscord | <Robyn [She/Her]> In reply to @Elegantbeef "There are JIT'd wasm": Yeah ig nested JITs are stupid |
| 23:52:49 | FromDiscord | <leorize> you can't build a JIT within wasm anyways |
| 23:53:11 | FromDiscord | <leorize> iirc wasm data memory is not executable in any way |
| 23:53:35 | FromDiscord | <Elegantbeef> Hey compile those opcodes into native code then uh... ask the host to run the native code 😄 |
| 23:55:15 | FromDiscord | <leorize> lol |
| 23:55:38 | FromDiscord | <Robyn [She/Her]> In reply to @leorize "iirc wasm data memory": That was why I was wondering if this silly way of creating another WASM runtime with the generated WASM code and calling into it |
| 23:57:58 | FromDiscord | <leorize> that could work, yes |
| 23:58:27 | FromDiscord | <leorize> Google famously used a similar trick to obfuscate reCaptcha |
| 23:58:48 | FromDiscord | <leorize> they build a VM in JS then use that VM to run reCaptcha |