<< 04-11-2022 >>
00:03:06FromDiscord<guttural666> this cannot possibly be a compile time value right? Nim will fall back to var I guess? https://media.discordapp.net/attachments/371759389889003532/1037879622731317258/image.png
00:03:33FromDiscord<guttural666> similar to constexpr in cpp
00:03:37FromDiscord<Elegantbeef> It can be `getHomeDir()` is overloaded for the VM iirc
00:03:48FromDiscord<Elegantbeef> So it returns your home dir
00:04:00FromDiscord<Elegantbeef> It's useful for macros, but a bug in disguise
00:04:08FromDiscord<Elegantbeef> That should be `let`
00:04:19FromDiscord<guttural666> yeah
00:05:05FromDiscord<guttural666> works just as well thanks 🙂
00:39:57*derpydoo joined #nim
00:48:42*pyrex left #nim (Leaving)
02:18:05*alystair joined #nim
02:18:37alystairI just saw a YT video about NIM, they mentioned it can compile to javascript? where do I read more about this
02:20:21FromDiscord<Elegantbeef> What do you want to know exactly?
02:29:38alystairah nevermind I need to just RTFM :)
02:52:02*derpydoo quit (Quit: derpydoo)
02:55:52*disso_peach joined #nim
03:10:49*arkurious quit (Quit: Leaving)
03:28:28FromDiscord<tangonov> sent a code paste, see https://play.nim-lang.org/#ix=4eTh
03:30:36FromDiscord<Bung> maybe try devel build version ?
03:32:01FromDiscord<tangonov> I am comfortable with my stable install from nixpkgs. If you're saying something upstream may fix this trivial issue, I can just accept it
03:32:54FromDiscord<tangonov> Stable on nix 1.6.6 if that's of any use in this conversation
03:37:44FromDiscord<Bung> 1.6.6 is fine, something changes to nimsuggest recent months, but maybe it's not right time try devel
03:57:34FromDiscord<proton> `/mnt/c/Users/low/nimbus-eth1/nimbus/p2p/validate.nim(247, 21) Error: getBalance(roDB, sender) can raise an unlisted exception: RlpError`↵Nim Compiler Version 1.6.8 [Linux: amd64]↵Compiled at 2022-11-04
04:04:40*dtomato4 quit (Ping timeout: 272 seconds)
04:05:02FromDiscord<Bung> that's probably lib side, add RlpError effect to that proc
04:07:49*rlr joined #nim
04:22:04*dtomato4 joined #nim
04:53:54FromDiscord<Bung> the effect system dont tell you the root cause? `can raise an unlisted exception Exception`, it doesn't tell me what proc call can raise that
05:12:20*disso_peach quit (Ping timeout: 248 seconds)
05:22:43*disso_peach joined #nim
05:47:08*rockcavera quit (Remote host closed the connection)
06:59:03*kenran joined #nim
07:02:29*lucerne joined #nim
07:17:23*arkanoid quit (Ping timeout: 246 seconds)
07:27:32*lucerne quit (Remote host closed the connection)
07:38:46*lucerne joined #nim
07:45:17*lucerne quit (Read error: Connection reset by peer)
07:50:07*lucerne joined #nim
07:59:19*jjido joined #nim
08:02:20*m5zs7k quit (Ping timeout: 252 seconds)
08:06:49*m5zs7k joined #nim
08:07:18FromDiscord<Girvo> sent a code paste, see https://play.nim-lang.org/#ix=4eTR
08:09:14FromDiscord<Girvo> (Ignore that the initial struct is called led_strip_s, thats just some C silliness with forward declarations so it can refer to itself)
08:12:27FromDiscord<Girvo> I guess what I'm asking is how the heck do I manage to generate `strip->clear()` instead of `(strip).clear()`?
08:13:40FromDiscord<Rika> Are they not equivalent?
08:14:56FromDiscord<Elegantbeef> They should be
08:15:19FromDiscord<Elegantbeef> `ptr->field` is dereference as is `(T)(ptr).field`
08:15:20FromDiscord<Girvo> ^ you are both correct. I wonder why it's choking then. I think the `importc: clear` that c2nim chucked on there is wrong
08:15:54FromDiscord<Girvo> Ah yep that's exactly what it was. c2nim was adding `importc` to the internal function pointers on the struct/procs on the object type. Remove them and it compiles!
08:15:56FromDiscord<Girvo> Now to see if it runs
08:16:02FromDiscord<Elegantbeef> It's unneeded in this case but you can importc fields iirc
08:16:05FromDiscord<Rika> Sounds strange
08:16:14FromDiscord<Rika> I thought it would still work with the import
08:16:22FromDiscord<Girvo> It was adding this `.ClE_0` stuff for some reason
08:16:35FromDiscord<Girvo> And removing the import on the `clear` removed it
08:16:38FromDiscord<Elegantbeef> Closures
08:16:47FromDiscord<Elegantbeef> `CLE` is closure environment
08:16:52FromDiscord<Elegantbeef> you needed the proc annotated `{.cdecl.}` or w/e
08:17:21FromDiscord<Girvo> Ah yeah, I just did that too lol
08:17:35FromDiscord<Elegantbeef> Remember that Nim assumes all procs in type defs are `closure`, fucks you around with C interop
08:17:41FromDiscord<Girvo> That's good to know, theres a few other libraries in ESP-IDF that are doing function pointers in structs everywhere
08:18:19FromDiscord<Girvo> sent a code paste, see https://play.nim-lang.org/#ix=4eTW
08:18:36FromDiscord<Elegantbeef> No problem
08:20:18*blackbeard420 quit (Ping timeout: 268 seconds)
08:22:40FromDiscord<Girvo> Now to wrap all this up in a way nicer interface lol
08:23:01FromDiscord<Girvo> But it works! I have a LED strip on our daughterboard lighting up through Nim 🙂
08:24:53FromDiscord<Elegantbeef> Damn the future is now
08:26:51FromDiscord<Girvo> It generates the PWM signal by hijacking the Remote Control (infrared) Transceiver hardware, because its the only thing fast enough that can generate a 24-bit 800khz signal for them lol
08:27:07FromDiscord<Girvo> The RMT driver is way too complicated for me to bother binding fully, but I only needed a couple types and two functions, and then `led_strip` itself. And they're currently blinding me on my desk, hooray
08:27:07FromDiscord<Elegantbeef> Look at this guy not thinking I know how remote controls used to work
08:27:07FromDiscord<Elegantbeef> Yes i said used to
08:27:30FromDiscord<Elegantbeef> Get your bluetooth wifi connected remote control for your forced smart IOT TV and be happy!
08:27:45*NimBot joined #nim
08:27:46FromDiscord<Elegantbeef> Do you get danger pay for the blinding
08:27:57FromDiscord<Elegantbeef> You really should high intensity light is no joke
08:28:30FromDiscord<Elegantbeef> I knew a guy whose retina had an image burned for a whole 5 minutes, couldnt see properly for 5 minutes, supremely dangerous!
08:28:54FromDiscord<Girvo> Bare LEDs driven at full power are crazy bright nowadays lol
08:28:59FromDiscord<Girvo> I have them turned down now
08:29:19FromDiscord<Girvo> sent a code paste, see https://play.nim-lang.org/#ix=4eTZ
08:29:38FromDiscord<Elegantbeef> The first step everyone takes when wrapping C, the ugliest stuff ever
08:30:05FromDiscord<Girvo> Those `_DEFAULT_CONFIG` things are `#define`-s in C land that give you a struct, I've done them as templates for now that just output the object with the default values set rather than a proc. I figure that's closest for now?
08:30:06FromDiscord<Elegantbeef> I might be wrong but atleast there isnt any resource management here
08:30:12FromDiscord<Girvo> Oh there absolutely isn't 😄
08:30:21FromDiscord<Girvo> Not yet anyway, that's the next step
08:30:36FromDiscord<Elegantbeef> Consts?
08:30:45FromDiscord<Elegantbeef> Or do they depend on runtime values?
08:31:21FromDiscord<Girvo> I know some have -- but these ones don't. But can you then edit the values after? One sec I'll show you what they do, and what I've done
08:33:57FromDiscord<Girvo> sent a code paste, see https://play.nim-lang.org/#ix=4eU5
08:34:13FromDiscord<Elegantbeef> Ah that's just a template constructor yea
08:34:14FromDiscord<Girvo> Can I assign a `const` object definition to a variable and modify its properties after?
08:35:13FromDiscord<Elegantbeef> I mean you can of course do `var a = myConst` but that's not what you want here, you want an expression
08:35:30FromDiscord<Girvo> Yeah sweet
08:36:00FromDiscord<Girvo> It's what c2nim was outputting and what I'd seen elsewhere and made sense to me as the direct translation of those defines
08:36:32FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4eU7
08:36:34FromDiscord<Girvo> Yeah that's what I also did haha
08:36:57FromDiscord<Girvo> It seemed to compile the same too yeah, with `-d:release`
08:37:27FromDiscord<Girvo> But all of this is going to be internal once I put some proper structure around all of it anyway, so I'm not too worried about it
08:37:27FromDiscord<Elegantbeef> I'd wager even without inline annotation in C it'd inline
08:37:45FromDiscord<Elegantbeef> This just seems like someone making a CMacro cause they are way way overly concerned
08:37:47FromDiscord<Girvo> Depends, the compiler they use for the LX7 processer here is... not great, sometimes
08:38:04FromDiscord<Elegantbeef> Right, i'm too used to gcc/clang
08:38:07FromDiscord<Girvo> And yep welcome to ESP-IDF lol. They use macros everywhere for this stuff
08:38:33FromDiscord<Elegantbeef> I guess if you want to not emit a procedure for basic operations C macros are the best you can do there
08:39:00FromDiscord<Girvo> Yeah. When all you have is text substitution and everything looks like a nail, or something
08:39:55FromDiscord<Elegantbeef> Metaprogramming was a mistake 😛
08:40:10FromDiscord<Girvo> Lispers recoil in horror
08:42:34FromDiscord<Girvo> Handy to know the `{.cdecl.}` requirement on the functions-in-structs. Will make binding some more of this stuff a lot easier!
08:42:43FromDiscord<Girvo> It'd tripped me up before and I couldn't work it out lol
08:45:06FromDiscord<srozb> hey, what is `nim secret` and why it's not in `--fullhelp` ?
08:45:49FromDiscord<Vindaar> if it were in `--fullhelp` it wouldn't be a secret, would it?
08:46:05FromDiscord<srozb> absolutely
08:46:41FromDiscord<Vindaar> it's a REPL for the nim VM. It's pretty limited and hence not really advertised
08:47:25FromDiscord<srozb> I see, I'm using inim and thought I could stick to the one that's already there (within the compiler)
08:48:13FromDiscord<ringabout> Does it screw up on Linux? It doesn't work properly on wsl2?
08:48:31FromDiscord<ringabout> (edit) "wsl2?" => "my wsl2."
08:49:45FromDiscord<ringabout> https://media.discordapp.net/attachments/371759389889003532/1038012164893573140/image.png
08:49:47FromDiscord<srozb> ok, thanks, I guess we'll have to wait for this secret to be less secret and more useful
08:51:33FromDiscord<ringabout> In reply to @flywind "Does it screw up": The prompt shows at the wrong place. It should have been before input not result.
08:56:55FromDiscord<Elegantbeef> Really i'd argue that inside a `importc`'d type it should infer as `cdecl` but that is likely more confusing
08:59:17*dnh joined #nim
09:04:29*alystair quit (Ping timeout: 246 seconds)
09:19:59*dnh quit (Quit: Textual IRC Client: www.textualapp.com)
09:22:12FromDiscord<4zv4l> I'm on linux and that does the same thing↵(@ringabout)
09:24:44FromDiscord<4zv4l> they should also probably add the `readLineFromStdin()` function to allow using the arrow keys to modify the input
09:28:29FromDiscord<ringabout> It is a regression
09:28:44*jjido quit (Quit: My laptop has gone to sleep. ZZZzzz…)
09:39:00FromDiscord<Phil> sent a long message, see http://ix.io/4eUn
09:39:47FromDiscord<Phil> (edit) "http://ix.io/4eUn" => "http://ix.io/4eUo"
09:40:11FromDiscord<srozb> sure, thanks. That explains a lot
10:12:28*pro joined #nim
10:23:22*pro left #nim (#nim)
10:42:32*wallabra quit (Ping timeout: 276 seconds)
10:44:20*wallabra joined #nim
11:51:05*jmdaemon quit (Ping timeout: 246 seconds)
11:57:13FromDiscord<Cheesy Brik> Where does the name nim come from?
12:01:33madpropsoriginally it was called Nimrod
12:01:42madpropsbut people complained i guess
12:02:17madpropshttps://en.wikipedia.org/wiki/Nimrod
12:05:41*disso_peach quit (Remote host closed the connection)
12:12:03FromDiscord<Generic> the story I heard was that in the Looney Tunes (idk whether it's a reoccuring thing or only happened ones, I don't know it that well) someone is called Nimrod in ridicule
12:12:47FromDiscord<Generic> and apparently that had such a lasting impact on American culture that some people to this day think Nimrod means idiot
12:13:47FromDiscord<Generic> not that I blame them, old cartoons had a lasting effect on me
12:13:56FromDiscord<Generic> (edit) "not that I blame them, old cartoons had a lasting effect on me ... " added "as well haha"
12:45:25*derpydoo joined #nim
12:47:14Amun-Rathere was similar story with Zig language (that reminded someone of "zig heil"…) :|
12:51:25FromDiscord<Prestige> that's so dumb lol
12:51:29FromDiscord<Generic> blah now I can't unhear it anymore
12:51:42FromDiscord<Rika> That is dumber than the Nim one ngl
12:52:56FromDiscord<Generic> it is, vowel length is afaik contrastive in German and English
12:53:53FromDiscord<Generic> though I currently can't come up with a minimal pair
13:06:37FromDiscord<tangonov> Zig as in "zigzag" is how I saw that one. Second to "why would anyone call a great programming language 'Nimrod'?" It took some growing up to realize that Nimrod was re-appropriated. I grew up understanding that Nimrod meant "moron" or baffoon, too.
13:07:07FromDiscord<tangonov> (edit) "baffoon," => "buffoon,"
13:07:42FromDiscord<tangonov> (edit) "growing up" => "reading"
13:11:25FromDiscord<jmgomez> converters doesnt work with procs, right?
13:24:19*jjido joined #nim
13:38:05FromDiscord<Require Support> anyone have an example of creating a `$` proc for a custom `ref object` ? is there a difference compared to a normal object?
13:41:10*blackbeard420 joined #nim
13:43:55FromDiscord<Duskhorn The Eclipsed> Hello, has anyone tried to load a custom shape window using the SDL2 bindings? ↵I had to manually import some functions but now I don't know why my code is not working, even after translating a running C example
13:44:07FromDiscord<Duskhorn The Eclipsed> And I'm stuck at the window being not opened
13:52:04FromDiscord<auxym> In reply to @Require Support "anyone have an example": no, there shouldn't be a difference. just define a `$` proc that takes your ref object type and returns a string
13:52:10*arkurious joined #nim
13:55:50*jjido quit (Quit: My laptop has gone to sleep. ZZZzzz…)
14:04:47*tinystoat joined #nim
14:07:16*tinytoast quit (Ping timeout: 248 seconds)
14:07:58*derpydoo quit (Quit: derpydoo)
14:10:31FromDiscord<nocturn9x> guys
14:10:38FromDiscord<nocturn9x> I have an idea that could be the cause of my problem
14:10:55FromDiscord<nocturn9x> does alloc() return nil when there's not enough memory in the system?
14:17:31FromDiscord<Rika> alloc() is pretty much equivalent to `malloc` so it should
14:17:45FromDiscord<Rika> unless you're using zephyrf
14:17:46FromDiscord<Rika> (edit) "zephyrf" => "zephyr"
14:17:48*xet7 quit (Ping timeout: 252 seconds)
14:20:26*derpydoo joined #nim
14:25:33*xet7 joined #nim
14:26:30FromDiscord<nocturn9x> I see
14:33:12*piertoni joined #nim
14:36:26FromDiscord<ChocolettePalette> What's that?↵(@Rika)
14:36:38FromDiscord<Rika> embedded rtos
14:37:46FromDiscord<ChocolettePalette> Uhh I'll google
14:39:13FromDiscord<ChocolettePalette> Zephyr is kinda cool ngl but there's a programming language called zephyr too
14:40:11FromDiscord<spoon> theres also fridges and shoes
14:43:16*kenran quit (Remote host closed the connection)
14:46:58*rockcavera joined #nim
14:46:59*rockcavera quit (Changing host)
14:46:59*rockcavera joined #nim
14:48:21*nisstyre joined #nim
15:42:03*piertoni quit (Ping timeout: 260 seconds)
15:54:05*alystair joined #nim
16:22:41*derpydoo quit (Ping timeout: 246 seconds)
16:23:34FromDiscord<jumping_sloth> Hey y'all, just found out about nim and I'm pretty impressed by its capabilities. I wanted to try the cross compilation, but when I try to run this command on windows\: `nim compile --cpu:amd64 --os:linux main.nim` the error `fatal error: sys/mman.h: No such file or directory 10 | #include <sys/mman.h> | ^~~~~~~~~~~~ compilation terminated.` appears. The program itself doesn't do anything special. It just prints some text.
16:45:58*derpydoo joined #nim
16:49:10Amun-Rayou have to have some "posixy" env installed
16:49:49FromDiscord<jumping_sloth> so I have to run the compile command in e.g. mingw?
16:51:07FromDiscord<auxym> In reply to @jumping_sloth "Hey y'all, just found": Nim uses a C compiler "under the hood". So, if you want to cross compile on windows targetting linux, you'll need a C compiler toolchain that can do that, first. A quick google seems to show that it's not straightforward, but might be possible using cygwin
16:52:04NimEventerNew question by The Modern Gamer: Getting hardware info in nim-lang, see https://stackoverflow.com/questions/74320430/getting-hardware-info-in-nim-lang
16:53:15FromDiscord<auxym> eg: https://stackoverflow.com/a/4770417/4092334
16:54:18FromDiscord<auxym> probably way easier to just use WSL
16:54:28FromDiscord<ShalokShalom> In reply to @auxym "Nim uses a C": Isnt this somehow possible with the Zig compiler?
16:54:56FromDiscord<auxym> no idea about zig
16:55:02FromDiscord<ShalokShalom> And always think about that↵↵https://github.com/jart/cosmopolitan
16:55:30Amun-Razig has all that built-in, nim generate C code for external compiler to compiler
16:55:37Amun-Ragenerates*
16:55:57Amun-Ratyping is not my strongest perk today
16:56:09FromDiscord<untoreh> can I iterate over a named tuple?
16:57:51FromDiscord<ShalokShalom> In reply to @Amun-Ra "zig has all that": Doesnt Zig offer a C compiler, that can do that?
16:58:19Amun-RaShalokShalom: zig can act as standalone C compiler
17:00:58NimEventerNew thread by freeflow: Importing from a parallel directory., see https://forum.nim-lang.org/t/9577
17:01:44FromDiscord<UnvaxxedPureBlood> In reply to @auxym "Nim uses a C": Zig can do that
17:01:47FromDiscord<jmgomez> sent a code paste, see https://play.nim-lang.org/#ix=4eX3
17:02:22FromDiscord<jmgomez> (edit) "https://play.nim-lang.org/#ix=4eX3" => "https://play.nim-lang.org/#ix=4eX4"
17:02:33FromDiscord<auxym> zig cc indeed seems pretty cool, just had a quick look at the blog post. I wonder if anyone has tried using nim with it. Should be possible with `--cc:clang` and `x64.linux.clang.exe = "zig cc"`, or something like that
17:02:36Amun-Rayou can generate C code with nim and compile for many platforms, you can compile zig code only for platforms defined in zig
17:03:02FromDiscord<jmgomez> I guess all that they have is a build script?
17:06:50FromDiscord<ShalokShalom> In reply to @auxym "zig cc indeed seems": Yes, I know that some people are doing that
17:06:58FromDiscord<auxym> zig cc is basically a CLI frontend to clang, but it also ships with libc for a ton of platform, and their compiler-rt thing. Or so I understood from a quick skim of the blog post.
17:07:07FromDiscord<ShalokShalom> But I am more excited about Cosmopolitian.
17:07:16FromDiscord<ShalokShalom> While I guess you can combine both.
17:07:40FromDiscord<auxym> In reply to @ShalokShalom "Yes, I know that": cool, would be a neat blog post to https://nim-lang.org/blog.html
17:08:12Amun-Raauxym: exactly
17:08:22FromDiscord<auxym> In reply to @ShalokShalom "But I am more": cosmopolitan is a (very) neat hack, but not something I'd consider using for "real" software
17:11:42FromDiscord<ShalokShalom> In reply to @auxym "cool, would be a": True
17:11:47FromDiscord<ShalokShalom> In reply to @auxym "cosmopolitan is a (very)": Why?/
17:12:05FromDiscord<ShalokShalom> Does Nim even get used for 'real' software 😋
17:18:20FromDiscord<auxym> good point 😉
17:19:28FromDiscord<auxym> for one thing, it does not have 100% syscall support, so anything you code which might make a syscall, you have to refer to this table: https://justine.lol/cosmopolitan/functions.html
17:20:09FromDiscord<auxym> also its dependent on OSes not breaking ABI compability, which is fine for linux apparently, but windows breaks the ABI regularly, and no idea for BSDs and whatnot
17:20:35FromDiscord<hovsater> How hard would it be to translate the extended BNF grammar (https://nim-lang.org/docs/manual.html#syntax-grammar) to regular BNF? I'm not too familiar with BNF or the extended BNF, but I might want to pursue translating it.
17:36:46FromDiscord<dlesnoff> In reply to @jmgomez "converters doesnt work with": Do you have a MWE of what you want?↵Do you want to implicitly change a procedure signature? Automatically add effects? Implicitly apply pragmas?↵You probably know but converters are really not appreciated (by Araq I guess?), since it breaks the rule "Explicit is better than implicit".
17:41:35FromDiscord<jmgomez> In reply to @Jiezron "Do you have a": I just ended up doing it explicitly. Basically I wanted to specify that two signatures were compatible. In particular to go from a regular proc into the same proc but returning a future instead..
17:48:40alystairI was going to explore nim as a single language to target both back and frontend effectively but once I saw the generated snake.js code from nim's home page I decided against it, are there no newer build-for-js examples?
17:49:05alystairI come from a deep vanilla js background so I'm hoping there has been a lot of improvement since that demo was made
17:49:27alystairas it's between 4-6 years old
17:51:21alystairI'd need to see more recent generated JS output to get a better understanding if nim is worth my time to pursue for my personal use cases
18:01:55FromDiscord<dlesnoff> In reply to @alystair "as it's between 4-6": What do you want for your JS output ?↵Do you want it to be easily readable ? small ? Fast ?↵Either way I have seen many compdaints about the JS output in the forum.↵There is yet a dedicated alternative std library for JS or nodeJS https://github.com/juancarlospaco/nodejs
18:05:30FromDiscord<dlesnoff> https://forum.nim-lang.org/t/8074
18:16:34FromDiscord<dlesnoff> https://forum.nim-lang.org/t/9525
18:16:45FromDiscord<ShalokShalom> In reply to @alystair "I'd need to see": Nim produces less readable results as Fable, Reason/Rescript and some others. ↵↵Is this important to you?
18:17:00FromDiscord<ShalokShalom> Or are you unhappy with the performance? ↵↵Security?
18:17:10FromDiscord<ShalokShalom> What is your pain point?
18:19:17FromDiscord<hector> What next, people asking for readable assembly code?
18:19:40FromDiscord<dlesnoff> In reply to @Jiezron "https://forum.nim-lang.org/t/9525": About the readability of the JS output
18:22:13FromDiscord<dlesnoff> In reply to @hector "What next, people asking": Don't be harsh to Nim newcomers please.↵Assembly and Javascript are complete different things. One is machine-dependent, the other is portable.
18:22:14FromDiscord<ShalokShalom> In reply to @hector "What next, people asking": JS can be done readable.
18:22:53FromDiscord<ShalokShalom> Nim doesnt do it. If you drop Nim later on, you depend on unreadable output code
18:23:20FromDiscord<ShalokShalom> If you do the same thing with Fable, you can drop it later and just succeed working with the produced output.
18:23:51FromDiscord<ShalokShalom> Particularly for companies, who might not find a Nim or FSharp developer that quick, is this a significant difference
18:26:05FromDiscord<hector> If it conflicts with performance I'd take performant JS over "readable" JS every time. WASM makes this moot anyway
18:31:27FromDiscord<dlesnoff> In reply to @hector "If it conflicts with": But readability improves tweaking possibilities and potential performance gains this way.↵It makes also the code reusable in non-Nim contexts. It allows you to extract specific parts of the output code. It helps in analyzing and comparing two code.↵This question is like any black-box vs white-box question: there is no definitive answer. It depends on what you want in the end.
18:35:05FromDiscord<dlesnoff> In reply to @hovsater "How hard would it": I am not sure what you mean by "regular grammar"?↵If it requires the language to be regular, i.e. specifiable by a regular expression, or equivalently by a finite automaton, I guess it is probably not doable.↵I am no expert in formal languages though.
18:35:43FromDiscord<hovsater> In reply to @Jiezron "I am not sure": I think you misunderstood me. I want to translate the current extended BNF into BNF.
18:36:41FromDiscord<hovsater> (edit) "In reply to @Jiezron "I am not sure": I think you misunderstood me. I want to translate the current extended BNF into BNF. ... " added "Since I'm not experienced with BNF I had a hard time understanding if I could express the extended BNF as BNF."
18:36:57FromDiscord<hovsater> (edit) "me." => "me or perhaps I'm misunderstand you. 😅"
18:37:24*FromDiscord quit (Remote host closed the connection)
18:37:37*FromDiscord joined #nim
18:42:24FromDiscord<dlesnoff> Oh yeah, as you see, I am no expert in formal languages, grammar specification and the likes 😦
18:45:51FromDiscord<dlesnoff> Any grammar defined in EBNF can also be represented in BNF, though representations in the latter are generally lengthier
18:46:04*jjido joined #nim
18:46:12FromDiscord<dlesnoff> Wikipedia quote
18:56:59*derpydoo quit (Quit: derpydoo)
18:58:41FromDiscord<fancy plants> I want to integrate Nim into an existing embedded C++ codebase. Can anyone help me reason with the memory model of the ARC gc? I want to use Nim for generating some JSON string, and so far I'm just passing the generated JSON string back to C as a `cstring` using an exported function. However, how is this cstring ref counted, and how do I safely deallocate the string after the C++ side uses the cstring (and likely copies it to another
18:59:08FromDiscord<fancy plants> Also hi! I am super new to Nim and I love everything about the language so far, and I decided the best way to learn was to dive right in
19:01:17*jjido quit (Quit: My laptop has gone to sleep. ZZZzzz…)
19:08:46FromDiscord<demotomohiro> I don't think creating cstring increase ref count.↵Easiest way is create a C function that read cstring but doesn't keep pointer to that string after return.↵Then, in Nim code, create a string and passed it to the C function.↵That string will be freed when the C function returned and the program get out the scope that string is defined.
19:11:30*wallabra quit (Ping timeout: 268 seconds)
19:12:47FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#ix=4eXy
19:13:10FromDiscord<Elegantbeef> Correct `cstring` is just the same as `myString[0].addr` so it does nothing for the GC
19:14:04FromDiscord<Elegantbeef> With refc you can do `GcRef`/`GcUnref` but with arc you need to use `ref string` to keep a reference alive, but it doesnt overly work well with C since you need a string implementation that matches that
19:15:31FromDiscord<fancy plants> @demotomohiro How would I add the ability for the C side to pass parameters to the JSON-string-making function? I'm essentially using Nim as a C JSON library and I can't just generate it purely from the Nim side
19:16:33FromDiscord<Elegantbeef> Expose the functions in Nim, and make sure any references are `GCRef`'d
19:17:45*derpydoo joined #nim
19:18:28FromDiscord<fancy plants> sent a code paste, see https://play.nim-lang.org/#ix=4eXz
19:19:02*wallabra joined #nim
19:20:54*phytolizer joined #nim
19:21:31FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4eXA
19:21:47FromDiscord<Elegantbeef> Whoops that should be `allocatedStrings[cstring(myStr[])] = myStr`
19:22:23FromDiscord<Elegantbeef> Due to the fact C strings are not pascal strings you need to do something like this to keep the string alive
19:22:59FromDiscord<Elegantbeef> Cause you cannot pass a pointer to your string to C and have it work with Cstrings, you either need to have a struct that matches the string in C, or do something like this
19:23:56FromDiscord<Elegantbeef> The Nim type def for a string in arc/orc is https://github.com/nim-lang/Nim/blob/devel/lib/system/strs_v2.nim#L12-L22
19:25:13FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4eXB
19:30:12FromDiscord<fancy plants> Would I not also need to `exportc` the dealloc function?
19:33:22*phytolizer quit (Quit: Client closed)
19:33:35FromDiscord<ShalokShalom> In reply to @hector "If it conflicts with": Performance and readability are usually both possible in JS. ↵↵A lot gets optimized by V8 and that is true for all JS code.
19:36:26FromDiscord<ShalokShalom> sent a long message, see http://ix.io/4eXI
19:57:52FromDiscord<haywireSSC> is there a way I can do something similar to making a sequence of concepts? ↵I could change all my code to use oop and inheritance insted of composition, allowing me to make sequences of the type they inheret from
19:59:22FromDiscord<Elegantbeef> Yes↵(@fancy plants)
20:00:08FromDiscord<Elegantbeef> No concepts are only compile time interfaces, to use them for dynamic dispatch you need to either manually or with a macro do that↵(@haywireSSC)
20:00:37FromDiscord<Elegantbeef> https://github.com/yglukhov/iface↵https://github.com/beef331/traitor↵Are some examples of making runtime interfaces
20:00:56FromDiscord<Elegantbeef> The latter of which is relatively untested and as it says mostly made cause an idea popped in my head
20:01:09FromDiscord<haywireSSC> thanks
20:02:03FromDiscord<jmgomez> you can also use inheritance
20:03:20FromDiscord<haywireSSC> yeah, I said
20:06:19FromDiscord<jmgomez> oh I missed it. I did use it for something like IRepository[T] that has multiple implementations but I prefer to stick to static dispatch when possible
20:06:35FromDiscord<jmgomez> does traitor support generics beef?
20:07:36FromDiscord<Elegantbeef> Dont recall
20:07:57FromDiscord<Elegantbeef> It needs a rewrite or two 😄
20:22:47*wallabra quit (Ping timeout: 246 seconds)
20:23:07FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#ix=4eY5
20:24:25*wallabra joined #nim
20:24:51FromDiscord<Elegantbeef> You dont `GcRef` the string
20:25:05FromDiscord<Elegantbeef> Also `GcRef` doesnt work on `seq`/`string` with orc/arc
20:26:46FromDiscord<Elegantbeef> Ah i misread you make an object and a ref for the destructor
20:33:22FromDiscord<ShalokShalom> Very nice talk↵↵https://youtu.be/HB5TrK7A4pI
20:33:39*cyraxjoe quit (Quit: I'm out!)
20:40:05FromDiscord<fancy plants> In reply to @demotomohiro "Another example to export": This is awesome. Again new to Nim, what is the use of creating the `NimString` and `NimStringObj` in the Nim code?
20:40:37FromDiscord<Elegantbeef> For the destructor since Nim only allows distincts and objects to have destructors
20:41:23FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4eY9
20:43:45FromDiscord<demotomohiro> In reply to @fancy plants "This is awesome. Again": ref object is allocated on heap and managed by Nim and object is allocated on stack.↵I defined object type because destructor is defined to object type but cannot be directly defined to ref object type.
20:56:28*cyraxjoe joined #nim
20:57:05FromDiscord<fancy plants> sent a code paste, see https://play.nim-lang.org/#ix=4eYe
20:58:24FromDiscord<fancy plants> Also thank you both for helping me with this! I'm really enjoying programming in Nim
20:58:45FromDiscord<fancy plants> Now for the harder challenge: convincing my coworkers this would be a good idea
21:03:26FromDiscord<Elegantbeef> you're getting the type of the dereferenced heap allocated type
21:03:26FromDiscord<Elegantbeef> It's how you can get the internal type without doing what demo did
21:03:46FromDiscord<fancy plants> Also, if there's no need to explicitly run anything or print out the destroyed string, do we need the `=destroy` proc implemented at all?
21:03:47FromDiscord<Elegantbeef> `[]` is how you dereference in Nim, `Type()` is how you construct an object, type of the dereferenced ref object is the heap allocate struct
21:03:58FromDiscord<fancy plants> Gotcha
21:04:29FromDiscord<Elegantbeef> You wouldnt need destroy since the memory management will go "Hey this ref has 0 on it's ref count, destroy on sight"
21:08:59FromDiscord<fancy plants> sent a code paste, see https://play.nim-lang.org/#ix=4eYn
21:09:09FromDiscord<fancy plants> Oh wait it'd need to be `ref` right cuz heap
21:09:52FromDiscord<Elegantbeef> Nim's new string isnt just a pointer so yea demo's is basically best you can do
21:11:18FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4eYo
21:11:19FromDiscord<Elegantbeef> Feel free to remove the destroy of course
21:23:07FromDiscord<fancy plants> Fascinating
21:26:29FromDiscord<jmgomez> sent a code paste, see https://play.nim-lang.org/#ix=4eYt
21:26:59FromDiscord<Elegantbeef> orc/arc use a string that stores the length on the stack
21:27:08FromDiscord<Elegantbeef> you can see this when you do `sizeof(string)` on refc vs. arc
21:27:47*jmdaemon joined #nim
21:28:24FromDiscord<jmgomez> ahh didnt know that! Does it applies to seq too?
21:28:29FromDiscord<Elegantbeef> Yes
21:29:13FromDiscord<Elegantbeef> In fact string literals in arc/orc are COW
21:29:20FromDiscord<jmgomez> Im curious about the reason behind it?
21:29:51FromDiscord<jmgomez> what's COW?
21:29:59FromDiscord<Elegantbeef> I'd assume it's cause `len` is often wanted and by storing it on the stack component you can speed up operations
21:29:59FromDiscord<Elegantbeef> Copy on write
21:30:37FromDiscord<jmgomez> I see, thanks!
21:31:11FromDiscord<Elegantbeef> I linked to the definition earlier but here it is again https://github.com/nim-lang/Nim/blob/devel/lib/system/strs_v2.nim#L12-L22
21:31:51FromDiscord<demotomohiro> Elegantbeef, you don't need to call `GcRef(result)` in `createStrInNim` proc in your code.↵Otherwise, you need to call `deallocNimString` twice to run destructor.
21:32:22FromDiscord<Elegantbeef> Right returning the type increments the counter
21:32:44FromDiscord<fancy plants> In reply to @demotomohiro "Another example to export": Another question. Because Nim would be using `arc` gc here, do I not need `NimMain` at the start of my C code?
21:33:06FromDiscord<Elegantbeef> Arc/Orc do not require `NimMain` but it's still good style to call it incase there is any top level code you rely on
21:33:37FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4eYv
21:34:24FromDiscord<fancy plants> sent a code paste, see https://play.nim-lang.org/#ix=4eYw
21:34:40FromDiscord<Elegantbeef> It doesnt have any top level code technicall
21:39:05FromDiscord<fancy plants> Good to know, thx
22:01:26FromDiscord<luteva> how can i set the id in norm to be a string/varchar and set the value explicitly? i want to use uuid here.
22:04:18*derpydoo quit (Ping timeout: 255 seconds)
22:45:50*jjido joined #nim
23:42:11FromDiscord<sOkam!> How would you get this type of value in nim? 🤔 https://media.discordapp.net/attachments/371759389889003532/1038236750402834473/unknown.png
23:43:36FromDiscord<Elegantbeef> `float.high`
23:46:23FromDiscord<Elegantbeef> Actually sorry that's positive infinity
23:47:37FromDiscord<Elegantbeef> https://nim-lang.org/docs/fenv.html#maximumPositiveValue.t%2Ctypedesc%5Bfloat32%5D
23:53:04FromDiscord<sOkam!> really handy, ty ✍️
23:54:08FromDiscord<sOkam!> what's different, in practical use terms, from `bycopy` to `byref`, and what's the default for Nim types?
23:54:40FromDiscord<sOkam!> seems like c2nim applies it to almost every type, but trying to understand why its doing that
23:54:41FromDiscord<Elegantbeef> It depends on the type in Nim
23:54:53FromDiscord<Elegantbeef> The reason it does that is cause C does that
23:54:59FromDiscord<Elegantbeef> It doesnt want to use optimisations that Nim applies
23:55:25FromDiscord<Elegantbeef> C passes by copy everywhere, Nim will pass by ref if the type is larger than 24bytes
23:55:56FromDiscord<Elegantbeef> Cause it's faster
23:56:18FromDiscord<sOkam!> Could I just remove the bycopy in the types, or will that create issues?
23:56:21FromDiscord<Elegantbeef> In C passing by reference is an explicit decision of `T`
23:56:30FromDiscord<Elegantbeef> That will create issues
23:56:44FromDiscord<Elegantbeef> If you're calling a procedure that expects a copy and you pass a pointer that's a C gen issue
23:57:21FromDiscord<sOkam!> so i will need to check each procedure one by one to search for possible problems, right?
23:57:47FromDiscord<Elegantbeef> Just use bycopy and dont worry
23:58:11FromDiscord<Elegantbeef> By copy is proper in C cause in C you explicitly do byref
23:58:43FromDiscord<sOkam!> yeah, this file does a lot of explicit ref passing. but i could adapt it, since im trying to convert it to pure nim
23:59:02FromDiscord<Elegantbeef> Well if you're making it pure nim you dont need to worry about that annotation
23:59:08FromDiscord<Elegantbeef> You just write Nim
23:59:27FromDiscord<sOkam!> yeah, but im starting with a c2nim translation as the base
23:59:57FromDiscord<sOkam!> so trying to convert it to standard nim, and trying to understand how it could be different