<< 11-10-2024 >>
01:01:19FromDiscord<majortrips1763> @eightbit_dboy you can try to do something like `CC=sh4-elf-gcc nim c --cc:env` and see if that targets SuperH4 correctly
01:01:45FromDiscord<majortrips1763> Well, set `CC` to whatever your sh4 CC is named
01:02:43FromDiscord<majortrips1763> > To use the CC environment variable, use nim c --cc:env myfile.nim. To use the CXX environment variable, use nim cpp --cc:env myfile.nim. --cc:env is available since Nim version 1.4.
01:50:13*xet7 joined #nim
02:02:57FromDiscord<nervecenter> sent a code paste, see https://play.nim-lang.org/#pasty=rKUxvzpB
02:38:02*rockcavera quit (Remote host closed the connection)
03:05:49*redj joined #nim
03:33:02FromDiscord<eightbit_dboy> In reply to @majortrips1763 "<@327881349924913152> you can try": Well, this is my config.nims https://media.discordapp.net/attachments/371759389889003532/1294140680473542736/config.nims?ex=6709edee&is=67089c6e&hm=077f21ef17b1ed0ae26f4377f56d641f8c09f33b5cff10cd8caf9527ad64af0d&
03:34:04FromDiscord<eightbit_dboy> Not sure how, exactly, that translates to nimscript
03:38:43*PMunch_ joined #nim
03:38:56*pbsds3 quit (Quit: Ping timeout (120 seconds))
03:39:23*pbsds3 joined #nim
03:39:57*Artea quit (Quit: ZNC 1.8.2 - https://znc.in)
03:41:06*ttkap quit (Ping timeout: 252 seconds)
03:41:50*PMunch quit (Ping timeout: 252 seconds)
03:46:06*ttkap joined #nim
03:51:36*Artea joined #nim
05:08:22FromDiscord<pmunch> In reply to @strogon14_ "<@392962235737047041>: something like https://www.n": Ooh, good idea
05:26:55*ntat joined #nim
06:24:33*ntat quit (Quit: Leaving)
06:51:49FromDiscord<mrgaturus> i found `{.emit.}` quoting won't capture symbols inside a template or it's arguments
06:51:56FromDiscord<mrgaturus> also may crash compiler into sigsegv
06:52:36FromDiscord<mrgaturus> (edit) "i found `{.emit.}` quoting won't capture symbols ... inside" added "or arguments" | removed "or it's arguments"
06:52:56FromDiscord<mrgaturus> (edit) "also may crash compiler into ... sigsegv" added "a"
06:58:14FromDiscord<classid> In reply to @classid "Could y'all give me": Hey guys
06:58:26FromDiscord<classid> Can I get some
07:05:37FromDiscord<nnsee> In reply to @classid "Can I get some": i mean we have no idea what your skill level or interests are, it's very hard to just "give you a project idea". what's something you wished existed but doesn't, at least not in the way you'd like? is there a tool or program or something you could use to improve your workflow? etc
07:05:41FromDiscord<nnsee> start there
07:54:01FromDiscord<eightbit_dboy> sent a code paste, see https://play.nim-lang.org/#pasty=lZPPdEfE
07:54:17FromDiscord<eightbit_dboy> This is a demo of a wrapper I'm trying to generate using futhark.
07:59:06FromDiscord<eightbit_dboy> Here's the repo: https://github.com/8bitprodigy/naylib64
08:00:39FromDiscord<eightbit_dboy> Thing is, it has a `config.nims` and `panicoverride.nim`.
08:08:48*PMunch_ is now known as PMunch
08:10:38Amun-Rahttps://chrismorgan.info/blog/unlicense/
08:26:30FromDiscord<eightbit_dboy> BSD0 it is, then.
08:26:47FromDiscord<eightbit_dboy> At any rate, the license is the least of my worries right now.
08:36:02PMunchYou still need the `os` module for certain things
08:36:21PMunchAnd Nim is quite dumb when deciding whether you have it or not
08:36:41PMunchYou can't use the os module on compile-time if your target doesn't have the os module
08:37:30FromDiscord<Elegantbeef> It's a hard limitation to overcome
08:37:33PMunchWhen working with FreeRTOS on a microcontroller for example I have to perform the wrap first and then import the modules because Futhark can't be run directly as it depends on the OS module
08:38:04PMunchEven though it does all its work during compile-time and the host machine has it
09:07:37FromDiscord<eightbit_dboy> Well, I'm doing the wrap first and then installing it with `nimble`.
09:44:43*beholders_eye joined #nim
10:01:53*beholders_eye quit (Ping timeout: 248 seconds)
10:03:18*mahlon quit (Ping timeout: 246 seconds)
10:07:04*beholders_eye joined #nim
10:14:35*mahlon joined #nim
10:32:38*Artea quit (Quit: ZNC 1.8.2 - https://znc.in)
10:34:20*beholders_eye quit (Ping timeout: 252 seconds)
10:55:36*Artea joined #nim
11:01:32*coldfeet joined #nim
11:47:47*beholders_eye joined #nim
11:54:28FromDiscord<demotomohiro> I wrote this library for path handling at compile time: https://github.com/demotomohiro/pathX
11:57:00FromDiscord<PATRYK SENS> alias chipset install
12:50:58*nisstyre quit (Quit: WeeChat 4.2.1)
12:58:47FromDiscord<kiloneie> Repost:↵What happened to Araq's object variants proposal from a while back ?
12:59:18FromDiscord<kiloneie> One of other people's proposals to his was, being able to reuse fields for each object variant
13:02:08*ntat joined #nim
14:00:16FromDiscord<majortrips1763> In reply to @eightbit_dboy "Well, this is my": @Nerve's examples can be translated to this file pretty easily.
14:13:55*PMunch quit (Quit: Leaving)
14:29:12FromDiscord<majortrips1763> sent a long message, see https://pasty.ee/JIAlmubU
14:29:23FromDiscord<majortrips1763> (edit) "https://pasty.ee/cmnYCZKz" => "https://pasty.ee/VjMEIyHe"
14:42:49FromDiscord<majortrips1763> Also, if I recall correctly, sh4 doesn't have normal atomic instructions (does it actually have an atomic xchg?) and it may need some special work there to setup the appropriate primitives.
14:45:17FromDiscord<majortrips1763> I thought gcc had a stack of compiler options for selecting the various atomic implementations.
14:56:01*Batzy joined #nim
14:58:21*Batzy_ quit (Ping timeout: 246 seconds)
15:02:45Amun-Rash4 is quite special with its branch slots
15:44:53FromDiscord<majortrips1763> Is there a addr32, or pointer, type or some such for memory addresses that are not on the heap or the stack? 😛
15:45:57Amun-Raconstants :)
15:46:15FromDiscord<majortrips1763> Yah, no .. memory structure created by the bootloader
15:46:34FromDiscord<majortrips1763> Up there with writing to devices that are attached via direct memory addressing
15:47:56FromDiscord<majortrips1763> Right now I am just using uint32 types and casting objects to the address, but I feel like there is likely a better way.
15:48:27Amun-Rajust like in C
16:08:55FromDiscord<majortrips1763> In reply to @Amun-Ra "just like in C": Yup, just thinking that there could be some syntax that better informs Nim's of the expected constraints. W/ C it is pretty easy to just do random things to memory and cast it anyway you want, which is fine, but it also makes it pretty easy to shoot yourself; and others; in the foot.
16:38:32FromDiscord<demotomohiro> How about to use `distinct int` or `SomePtr[T] = distinct ptr T`?
16:43:34*Jhonny2x4 quit (Quit: Jhonny2x4)
16:43:43*Jhonny2x4 joined #nim
16:44:21*Jhonny2x4 quit (Client Quit)
16:44:29*Jhonny2x4 joined #nim
16:45:39*Jhonny2x4 quit (Client Quit)
16:45:47*Jhonny2x4 joined #nim
17:06:35FromDiscord<drunkenalcoholic> sent a code paste, see https://play.nim-lang.org/#pasty=lRVYIcSR
17:22:38FromDiscord<spotlightkid> @drunkenalcoholic\: you want file open mode `fmReadWriteExisting` instead of `fmReadWrite`.↵https://nim-lang.org/docs/syncio.html#FileMode
17:25:53FromDiscord<kryp7on> In reply to @majortrips1763 "I am honestly not": what I meant is: I have first part of software where I work with some files using GC, then it comes part to manually map memory bcs Im using WinAPI and some low level stuff, so I just switch while writing in source, to manually map memory in stuff instead of using GC. Then when I finish working with that low lvl stuff, I proceed to use GC again
17:26:00FromDiscord<kryp7on> is something like this possible?
17:26:48FromDiscord<kryp7on> ALSO: how low-level is Nim? Im pentester and I need to write some security tooling. Comparing to Zig/Rust/C, is it capable of manipulating with kernel and stuff?
17:36:07FromDiscord<drunkenalcoholic> In reply to @spotlightkid "<@325338333314154509>\: you want file": Thanks mate, that was the problem, now its working, very much appreciated, thank you
17:40:19FromDiscord<majortrips1763> In reply to @kryp7on "ALSO: how low-level is": Well .. Nim seems to be curiously popular for malware and the like .. there are entire repos and articles out there regarding weaponizing it...
17:42:08FromDiscord<majortrips1763> Well .. snap ...↵> set i too large; use `std/sets` for orginal types with more than 2^16 elements↵↵Bleh, the damn AOUT_KLUDGE flag in MB1 🤮
17:42:15FromDiscord<majortrips1763> (edit) "i" => "is"
17:42:19FromDiscord<majortrips1763> (edit) "orginal" => "ordinal"
17:43:54FromDiscord<majortrips1763> sent a code paste, see https://play.nim-lang.org/#pasty=MbbUUMQb
17:52:02FromDiscord<majortrips1763> oh .. that was my bad on that thing
17:53:36FromDiscord<majortrips1763> `AoutKludge = 20`
17:55:52FromDiscord<demotomohiro> sent a long message, see https://pasty.ee/HCyebPdV
17:57:02FromDiscord<majortrips1763> In reply to @demotomohiro "It is not kernel": The stack pointer situation is something I am dealing with as well.
17:58:56FromDiscord<demotomohiro> In reply to @majortrips1763 "The stack pointer situation": https://github.com/demotomohiro/MinimumNimPico/blob/main/src/startup.nim#L138↵I just used inline assembler to set the stack pointer.
18:06:45*beholders_eye quit (Ping timeout: 248 seconds)
18:12:36FromDiscord<majortrips1763> In reply to @demotomohiro "https://github.com/demotomohiro/MinimumNimPico/blob": Yah, same.. the problem for me is that I am setting up a temporary stack pointer so that I do the work of parsing the multiboot information data, since total memory and valid memory information is in the MBI .. then I get to init all of memory and figure out my "final" stack pointer.. hehe.
18:13:13FromDiscord<majortrips1763> And then I get to throw the old stack away.
18:36:59*PMunch joined #nim
19:58:22FromDiscord<kryp7on> In reply to @demotomohiro "It is not kernel": what about thing I asked? can I switch in source code from GC to manually mapping then again to GC?
20:07:05*coldfeet quit (Remote host closed the connection)
20:22:01*ntat quit (Quit: Leaving)
20:47:02FromDiscord<eightbit_dboy> In reply to @majortrips1763 "I went through and": That example is from naylib64, which is for n64, though.
20:47:39FromDiscord<eightbit_dboy> (edit) "In reply to @majortrips1763 "I went through and": That example is from naylib64, which is for n64, though. ... " added "That, I can already compile for."
20:48:24FromDiscord<eightbit_dboy> That said, this info may be helpful, if I decide to create an sh4 target.
20:50:25FromDiscord<demotomohiro> I think, Nim manages only memory allocated for ref, seq or string types. If C library or API allocates heaps, Nim's memory memory management do nothing to them. C functions should work correctly as long as you use them correctly.
20:56:52FromDiscord<demotomohiro> In reply to @kryp7on "what about thing I": I don't know what you are trying to do, but I don't think you need to stop Nim's GC to use memory provided by C APIs as long as they don't overlap memory that Nim's memory managent uses.
21:22:19*PMunch quit (Quit: leaving)
21:25:03FromDiscord<kryp7on> sent a code paste, see https://play.nim-lang.org/#pasty=TNBMoymf
21:25:18FromDiscord<kryp7on> (edit) "https://play.nim-lang.org/#pasty=JXAPujQF" => "https://play.nim-lang.org/#pasty=nMVjXndX"
21:25:39FromDiscord<Elegantbeef> If you use Arc there is no real GC it's just RAII
21:25:56FromDiscord<Elegantbeef> So... it's just an abstraction over manual management
21:26:31FromDiscord<kryp7on> huh, sry I didnt understand what you said. Idk whats Arc not RAII
21:26:53FromDiscord<Elegantbeef> Well then you shouldn't be worried about GC
21:34:22FromDiscord<kryp7on> In reply to @Elegantbeef "Well then you shouldn't": so its actually possible switching from GC to manual manipulating memory?
21:34:32FromDiscord<Elegantbeef> No cause that makes no sense
21:34:51FromDiscord<Elegantbeef> You use automatic managed memory where you want it, and manually manage where you want it
21:37:36FromDiscord<kryp7on> In reply to @Elegantbeef "You use automatic managed": isnt GC there by default?
21:38:00FromDiscord<majortrips1763> Hmm, are there any docs/tutorials/examples of casting objects to memory addresses?
21:38:24FromDiscord<Elegantbeef> `.addr` gets the address
21:39:11FromDiscord<Elegantbeef> `cast[ptr T](myData)` is how you cast to a pointer
21:40:54Amun-Rakryp7on: yes, but GC manages only certain types mentioned above
21:41:41FromDiscord<majortrips1763> Well, I have a uint32 that holds an address, and I am trying to cast an object to point at it
21:42:20FromDiscord<Elegantbeef> So then `cast[ptr MyObjecy](myuint32)`
21:42:30FromDiscord<Elegantbeef> Assuming 32bit CPU
21:42:56FromDiscord<kryp7on> In reply to @Amun-Ra "<@1124352185916981328>: yes, but GC": ah, and for the rest, you need to do it yourself?
21:43:00FromDiscord<Elegantbeef> If it's 64bit you likely should raise to 64bit then cast
21:47:45FromDiscord<majortrips1763> Well, it might be a 64bit CPU .. but the address is guaranteed to be 32bit.
21:48:14Amun-Rakryp7on: yes
21:53:52FromDiscord<majortrips1763> sent a code paste, see https://play.nim-lang.org/#pasty=CagDZUXW
21:55:08FromDiscord<majortrips1763> > type mismatch
21:57:00FromDiscord<majortrips1763> Oh .. I thihk I figured that out
21:57:05*krux02 joined #nim
22:13:35FromDiscord<majortrips1763> I needed to cast the addres of the object to an int before doing maths ... https://forum.nim-lang.org/t/2569↵↵Fun
22:14:25FromDiscord<Elegantbeef> Or you can `cast[ptr UncheckedArray[...]](...)[1]`
22:14:46FromDiscord<majortrips1763> Oh, I saw a reference to that, but I wasn't really grok'ing it
22:15:15FromDiscord<Elegantbeef> If you are just offsetting by `sizeof(T)` it's the same as an array of `T`
22:15:28FromDiscord<Elegantbeef> As such if you treat it as a pointer to an array of T you can offset it
22:15:29FromDiscord<majortrips1763> oh, each element is of a different size 😦
22:16:00FromDiscord<Elegantbeef> Oh I see `Multiboot2HeaderTag` and `MultiBootHeader`
22:16:09FromDiscord<Elegantbeef> So yea time to cast to an int
22:16:12FromDiscord<Elegantbeef> Or use `ptr_math`
22:16:37FromDiscord<majortrips1763> Yah, initially it starts w/ the header+sizeOfHeader to find the 1st Tag, and then tag.next is basically tag.addr+tag.size
22:17:25FromDiscord<majortrips1763> In reply to @Elegantbeef "Or use `ptr_math`": What is this strange thing of which you speak?
22:17:37FromDiscord<Elegantbeef> It's a nimble package
22:17:39FromDiscord<majortrips1763> Oh, a module!
22:17:41FromDiscord<majortrips1763> Nice
22:17:41FromDiscord<Elegantbeef> It implements operators on `ptr T`
22:18:04FromDiscord<Elegantbeef> It's a single file so you can always just steal it and drop it in your project with the license added
22:18:34FromDiscord<majortrips1763> I might do that later as I refactor stuff .. right now I am really enjoying doing everything directly and basically drinking from the firehose.
22:18:47FromDiscord<majortrips1763> Been a great learning experience
22:18:49FromDiscord<Elegantbeef> I believe that's how you deface yourself
22:19:03FromDiscord<majortrips1763> Yah, especially on the 10" line...
22:25:16FromDiscord<majortrips1763> So .. are there are common set of patterns for `.gitignore` when working with Nim?
22:26:14FromDiscord<Elegantbeef> On linux I just git ignore `.exe` `.dll` `.so` and extensionless f iles
22:27:02FromDiscord<majortrips1763> Kk, I was about to add .out and was curious if there were others I should add.
22:30:44Amun-Ramajortrips1763: I have also these: https://dpaste.com/ADLR4J66Y
22:40:42FromDiscord<.bobbbob> In reply to @Elegantbeef "If you use Arc": seems like a lot of new players hear gc and think java in the 90s gc and think it will take a lot of resources (I did that too a bit)
22:48:38FromDiscord<Elegantbeef> Right, if you do not know what RAII is you likely should go learn about deterministic deallocations before questioning about GCs 😄
22:52:13FromDiscord<.bobbbob> basically it's arc is deterministic and zero overhead and orc is just arc with reference counting for recursive types right?
22:52:37FromDiscord<Elegantbeef> Well Orc has a mark and sweep pass for cycles
22:52:45FromDiscord<Elegantbeef> `ref` has a reference count on Arc
22:52:56FromDiscord<.bobbbob> ah
22:53:49*krux02 quit (Remote host closed the connection)
22:59:15FromDiscord<Nlits (Ping on reply)> I'm trying to make a sort-of ui engine in nim but am stuck on how to do sizing and positioning. I know you have to do multiple passes, with new restrictions each time but still can't wrap my head around how to do this properly. A link to an article or some code written for another library would be nice. I appreciate any help. (Not quite sure this is the right place to ask but it is at least a start)
23:00:47FromDiscord<Elegantbeef> For my UI stuff I just have procedure to calculate the used size for an element then use that to position/size the parent element
23:01:20FromDiscord<Elegantbeef> So like a vertical layout group iterates its children and calculates the size of each incrementing the start of the next by the size of the former
23:01:49FromDiscord<Elegantbeef> It's not very digestable due to using a lot of silliness https://github.com/beef331/gooey/blob/master/src/gooey/layouts.nim
23:02:19FromDiscord<Elegantbeef> This is for game stuff so It's all anchor based and not meant for real programs
23:02:38FromDiscord<Elegantbeef> So like I anchor a UI element to one of 9 positions on a screen, then the position is an offset from that point
23:22:04FromDiscord<__nycto__> Quick question: Are [PFrame](https://nim-lang.org/docs/system.html#PFrame) pointers stable? That is to say: if I want to store a stack trace, do I need to copy the content of a `PFrame` to my own object, or can I just store the `PFrame` itself and refer back to it later?
23:23:09FromDiscord<Elegantbeef> Should be fine as long as the object does not outlive the stack frame
23:24:46FromDiscord<__nycto__> It does. I'm working on a memory leak analyzer for a custom device. So I'm trying to record the stack frame anytime memory is allocated, then print a stack trace for anything that isn't released. Or for memory that gets double released
23:32:22FromDiscord<Nlits (Ping on reply)> In reply to @Elegantbeef "So like a vertical": The issue is things like text-wrap require you to give a specified width, and depend on that. That is what the second iteration is for. I also dk how to make flexbox and such, but i will look at your code
23:33:03FromDiscord<Elegantbeef> My method of text wrap is just scale the text down 😄
23:34:10FromDiscord<Elegantbeef> I believe my Layout groups are technically flexboxes
23:34:39FromDiscord<Nlits (Ping on reply)> In reply to @Elegantbeef "My method of text": That isn't possible in my particular situation, but i can probably just expand upon what i see in your code with the code i already wrote
23:35:04FromDiscord<Elegantbeef> Well you do newline insertion if you do not scale down