<< 01-02-2025 >>
00:00:01*marcus quit (Remote host closed the connection)
00:00:44*marcus joined #nim
00:01:16FromDiscord<Elegantbeef> Not really without reaching to macros
00:01:50FromDiscord<Elegantbeef> https://github.com/beef331/micros/blob/master/src/micros/introspection.nim#L3-L7 if you so wish to do that
00:03:12FromDiscord<arkanoid> 90% of your answers are links to a lib you have already done. How do you do that
00:03:21FromDiscord<Elegantbeef> I am everything
00:04:16FromDiscord<luteva> In reply to @takemichihanagaki3129 "Is there a way": I mean nimsuggest is just a programm you can start if you want to. If you don't want it, just don't start it.
00:06:13FromDiscord<luteva> In reply to @takemichihanagaki3129 "Is there a way": what are you trying to do? where does there problem occur?
00:07:31FromDiscord<luteva> oh, maybe you are talking about some IDE configuration?
00:08:36FromDiscord<takemichihanagaki3129> In reply to @luteva "just do not start": This simply starts by the vs code extension. ↵But, btw, I just removed the binary.
00:08:43FromDiscord<takemichihanagaki3129> In reply to @takemichihanagaki3129 "This simply starts by": This solved my problem.
00:11:06FromDiscord<takemichihanagaki3129> sent a long message, see https://pasty.ee/VuUzZHqR
00:11:19FromDiscord<luteva> you can configure your IDE to start some programms to do some "magic". the "clean" solution would be to fiy the configuration. but if the ide can't find the programm anymore, this may work, too. 🙂
00:11:41FromDiscord<takemichihanagaki3129> In reply to @takemichihanagaki3129 "I didn't wanted. But": But this time vscode just spawned +10 instances at once, so I prefered to remove it completely.
00:12:45FromDiscord<takemichihanagaki3129> In reply to @luteva "oh, maybe you are": IDE settings would be the ideal to me.
00:13:26FromDiscord<luteva> i think there are more nim extensions in vscode. maybe a different one may work better for you,
00:13:31FromDiscord<takemichihanagaki3129> First time nimsuggest worked to me was using Helix, on a single file project.↵Again, running with 4gb and Windows consumes 3/4 of everything...
00:13:43FromDiscord<takemichihanagaki3129> In reply to @luteva "i think there are": I'll have a look.
00:14:03FromDiscord<luteva> (nit sure though)
00:38:04FromDiscord<System64 ~ Flandre Scarlet> In reply to @takemichihanagaki3129 "Mine is breaking my": I'm glad that I'm not alone to have this issue
00:38:23FromDiscord<System64 ~ Flandre Scarlet> It already crashed my computer
00:44:21FromDiscord<takemichihanagaki3129> In reply to @sys64 "It already crashed my": My computer got locked by 6min due to it...
00:47:05FromDiscord<takemichihanagaki3129> In reply to @sys64 "It already crashed my": My older one was able to run nimsuggest at some point.↵Not the fasted way, but worked.↵6Gb on win10.↵Now, I need to buy a new one, less GB and Win11...
00:48:50*GnuYawk67 quit (Quit: The Lounge - https://thelounge.chat)
00:49:15*GnuYawk67 joined #nim
00:56:16FromDiscord<lainlaylie> i dont want to sound like a broken record here but is this not https://github.com/nim-lang/choosenim/issues/13
01:35:43FromDiscord<Laylie> matrix bridge missed a whole hour of messages?
01:36:55FromDiscord<Elegantbeef> It tends to drop messages like a hot potato
01:44:03*alexdaguy joined #nim
01:44:25FromDiscord<arkanoid> why the two templates are different here? the first compiles, the second doesn't, but they should be the same https://play.nim-lang.org/#pasty=lCPeMABR
01:55:07*alexdaguy quit (Quit: restart)
01:55:25*alexdaguy joined #nim
02:10:46*arkanoid joined #nim
02:13:18FromDiscord<Elegantbeef> `ref Exception` vs `ref CatchableError`
02:13:32FromDiscord<Elegantbeef> Nim inherit types are no covariant so inside generic parameters a child does not match for the parent
02:14:16FromDiscord<arkanoid> gotcha
02:15:23FromDiscord<arkanoid> interesting point. Next would be why std/logging raises Exception instead of anything under CatchableError
02:20:26FromDiscord<Elegantbeef> Inheritance be damned 😄
02:21:05FromDiscord<Elegantbeef> Joke aside I think it's just cause a procedure it calls could not be effect analyzed
03:14:46FromDiscord<arkanoid> how would you keep a bunch of non constant configs (numbers, strings) readable by X-1 threads and writable by 1?↵There is no atomicity required here, I don't care about the order of write and reads. At the moment I'm using a stack allocated object where strings are StackString(s) and it works, but I quite find the approach less flexible than what it could be. Any hint?
03:16:50FromDiscord<Elegantbeef> What's your issue with the flexibillity?
03:19:13FromDiscord<arkanoid> for example one of those settings is an Uri, and other are Path(s). Saving all string as StackString(s) forces continuous conversion between higher level abstractions within each thread, when I really need to parse it them once every change
03:20:25FromDiscord<arkanoid> I could generate a ParsedConfig object that is the validated and converted mirror of the one limited by StackString, and run each every now and then from each thread, but it feels like polling from within the same process
03:23:34FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=wwOFfwvM
03:23:34FromDiscord<Elegantbeef> 😄
03:24:23FromDiscord<Elegantbeef> Whoops `{.guard:lock.}`
03:27:16FromDiscord<arkanoid> Hm, I tried something similar days ago, but I failed fight with the compiler that wanted me to not call global GCed objects from within thread code
03:28:25FromDiscord<Elegantbeef> So wrap it with some `{.cast(gcsafe).}` 😄
03:29:02FromDiscord<arkanoid> that's too much violence for me
03:29:40FromDiscord<Elegantbeef> Eh throw it inside guarded
03:29:46FromDiscord<Elegantbeef> Then it's no problemo
03:34:07FromDiscord<arkanoid> black wizard telling kids to play with weapons
03:34:32FromDiscord<arkanoid> but ok, I like locks
03:34:42FromDiscord<Elegantbeef> Well then you'll like guarded
03:34:53FromDiscord<Elegantbeef> Instead of `path: string` you do `path: Guarded[string]`
03:37:22FromDiscord<arkanoid> what is the pragma for? why aren't get and set sufficient for locking?
03:37:41FromDiscord<Elegantbeef> the program enforces you to always `{.lock:[guarded.lock].}: ....`
03:37:53FromDiscord<Elegantbeef> pragma\
03:38:10FromDiscord<Elegantbeef> So a smart alec cannot do `guarded.data` without the lock annotation
03:43:33FromDiscord<mjsdev> sent a code paste, see https://play.nim-lang.org/#pasty=gNPmkCxU
03:52:14FromDiscord<mjsdev> Nevermind, I may have found a solution using getImp on the symbol
03:52:24FromDiscord<arkanoid> it's a double guard against users peeping into private data. Cool
04:12:04*ensyde joined #nim
04:15:44FromDiscord<arkanoid> `Error: 'configurationJob' is not GC-safe as it calls 'innerProc'`, where `configurationJob` is a `{.thread.}`, and `innerProc` is a proc defined into `configurationJob` function body. How can it be no t GC-safe?! it's part of the thread body!
04:16:08FromDiscord<Elegantbeef> It's probably turned into a closure
04:16:31FromDiscord<arkanoid> do I need explicit nimcall then?
04:16:55FromDiscord<Elegantbeef> Try it and see what it complains about
04:17:03FromDiscord<Elegantbeef> I'm guessing "Cannot capture X in nimcall"
04:17:39FromDiscord<arkanoid> again ` Error: 'configurationJob' is not GC-safe as it calls 'inner'` where `inner` is `proc inner(): Result[void, string] {.nimcall.} =`
04:18:26FromDiscord<Elegantbeef> Does inner access any global memory?
04:20:47FromDiscord<arkanoid> it should not. I need to cleanup other errors probably, gimme a minute
04:36:08FromDiscord<arkanoid> sent a code paste, see https://play.nim-lang.org/#pasty=fCfmucKn
04:39:33FromDiscord<Elegantbeef> Where is it not gcsafe?
04:43:11FromDiscord<arkanoid> `Result[Ssh2Lib, string]` allocates a string
04:43:23FromDiscord<arkanoid> I've converted error type into enum
04:48:30FromDiscord<arkanoid> or better, avoiding checking initialization from thread and just assume it
05:14:59FromDiscord<arkanoid> sent a code paste, see https://play.nim-lang.org/#pasty=MFEFBolT
05:19:53FromDiscord<Elegantbeef> It makes a procedure inside of `sortedbyit` with `{.raises: [].}`
05:21:32FromDiscord<arkanoid> thanks again
05:31:27FromDiscord<arkanoid> argh, why cmd proc needs to be a closure! this won't work inside threads https://github.com/nim-lang/Nim/blob/6bf9265d24c6243cd5cf9e10f7fe766fa5372396/lib/pure/algorithm.nim#L368
05:31:57FromDiscord<Elegantbeef> Closures work on threads
05:32:03FromDiscord<Elegantbeef> Just don't send them across
05:45:38FromDiscord<arkanoid> sent a code paste, see https://play.nim-lang.org/#pasty=eVGqVsGl
05:47:06FromDiscord<Elegantbeef> `int` vs `BiggestInt`
05:47:14FromDiscord<arkanoid> here it is
05:47:22FromDiscord<arkanoid> thanks
06:08:11*xet7 joined #nim
06:35:08*SchweinDeBurg quit (Quit: WeeChat 4.6.0-dev)
06:35:36*SchweinDeBurg joined #nim
08:03:58*coldfeet joined #nim
08:04:35*ntat joined #nim
09:05:27*ensyde quit (Quit: WeeChat 4.5.1)
09:34:16FromDiscord<luteva> In reply to @takemichihanagaki3129 "My older one was": sorry, but I can't help with windows systems. But I am sure there are many other guys here, using windwos.
09:39:22*beholders_eye joined #nim
09:42:11*xet7 quit (Quit: Leaving)
10:49:33*coldfeet quit (Quit: Lost terminal)
11:13:37*beholders_eye quit (Ping timeout: 252 seconds)
13:28:18*rockcavera joined #nim
13:56:50*alexdaguy quit (Quit: WeeChat 4.5.1)
14:07:43*ntat quit (Quit: Leaving)
14:14:22FromDiscord<firasuke> Why are these not synced?↵- https://github.com/nim-lang/nightlies/releases/tag/latest-devel↵- https://github.com/nim-lang/nightlies/releases/tag/2025-02-01-devel-0861dabfa70f40f00c5c95d58c344ad5e0fd19ec
15:03:09*arkanoid quit (Ping timeout: 244 seconds)
16:46:40FromDiscord<nocturn9x> hello there
16:47:11FromDiscord<nocturn9x> sent a code paste, see https://play.nim-lang.org/#pasty=LCrrXtSC
16:50:54FromDiscord<nocturn9x> ah, seems like I'm missing the mingw libraries
16:56:46FromDiscord<nocturn9x> it seems like Nim is passing `-ldl` to the linker script
16:56:52FromDiscord<nocturn9x> but that doesn't work for windows cross compilation
16:56:56FromDiscord<nocturn9x> how do I tell it to like, not do that?
17:00:03*arkanoid joined #nim
17:00:41strogon14when defined(windows):
17:02:00FromDiscord<nocturn9x> ah I need `--os:windows` at the top of nim.cfg?
17:02:22FromDiscord<Robyn [She/Her]> In reply to @strogon14 "when defined(windows):": Isn't it `@defined windows` for the cfgm
17:02:42FromDiscord<nocturn9x> sent a code paste, see https://play.nim-lang.org/#pasty=jPzhCLie
17:02:47FromDiscord<Robyn [She/Her]> In reply to @nocturn9x "ah I need `--os:windows`": don't see why that'd be an issue
17:02:48FromDiscord<nocturn9x> I am not adding `-ldl` explicitly
17:02:52FromDiscord<nocturn9x> nim is doing it on its own
17:02:53FromDiscord<nocturn9x> lol
17:03:15FromDiscord<Robyn [She/Her]> In reply to @nocturn9x "I am not adding": can you check in the Nim config dir?
17:03:29FromDiscord<Robyn [She/Her]> Like, the Nim compiler dir
17:03:34FromDiscord<nocturn9x> uh
17:03:38FromDiscord<nocturn9x> where would I check lol
17:03:48FromDiscord<nocturn9x> there's nothing nim-related in `~/.config` except Inim stuff
17:04:34FromDiscord<Robyn [She/Her]> In reply to @nocturn9x "there's nothing nim-related in": It's the same dir as your Nim compiler dir, so if you used choosenim, check .choosenim
17:05:33FromDiscord<nocturn9x> sent a code paste, see https://play.nim-lang.org/#pasty=jpgylaIk
17:05:34FromDiscord<nocturn9x> haha
17:05:35FromDiscord<nocturn9x> found it
17:05:42FromDiscord<spotlightkid> Are you using the `dynlib` pragma or any library that does?
17:05:49FromDiscord<nocturn9x> do I just `--clang.options.linker = ""`
17:05:52FromDiscord<nocturn9x> In reply to @spotlightkid "Are you using the": nop
17:06:34FromDiscord<Robyn [She/Her]> In reply to @nocturn9x "this is my nim.cfg:": No clue why this is a problem, fairly recent post did this fine, but I do notice Nim 1.3.0 is defined?
17:06:51FromDiscord<nocturn9x> nono
17:06:57FromDiscord<nocturn9x> those version numbers are from my own engine
17:07:03FromDiscord<nocturn9x> it's a program specific version number
17:07:06FromDiscord<nocturn9x> nim version is 2.2.0
17:07:09FromDiscord<Robyn [She/Her]> aaah sorry
17:07:10FromDiscord<nocturn9x> anyway I figured it out
17:07:15FromDiscord<nocturn9x> In reply to @nocturn9x "do I just `--clang.options.linker": this fixes it
17:07:20FromDiscord<Robyn [She/Her]> Then yeah not too sure
17:07:25FromDiscord<nocturn9x> it builds the exe
17:07:26FromDiscord<nocturn9x> so all good
17:07:30FromDiscord<Robyn [She/Her]> In reply to @nocturn9x "do I just `--clang.options.linker": oh i didn't see that sorry
17:08:43FromDiscord<spotlightkid> Anyway, I would always put compiler options in a project specific config file, e.g. a `config.nims` in your `src` directory.
17:10:20FromDiscord<nocturn9x> I have nim.cfg
17:10:24FromDiscord<nocturn9x> thought that was enough no?
17:10:30FromDiscord<nocturn9x> (edit) "enough no?" => "enough, is it not?"
17:10:37FromDiscord<nocturn9x> in my project's dir I mean
17:12:20FromDiscord<ninovanhooff> sent a code paste, see https://play.nim-lang.org/#pasty=VkUgrscj
17:14:30FromDiscord<Robyn [She/Her]> In reply to @ninovanhooff "Could I get a": I don't think there's really any performance issues here but you might need some validation and you'd want `SomeSignedInt` to prevent overflow imo
17:19:25FromDiscord<spotlightkid> Yeah, basically the same. `config.nims` makes it obvious, that you can use nimscript (and I'm not sure whether you need to use `nimble` for it to be used).↵(@nocturn9x)
17:19:39*ntat joined #nim
17:19:40FromDiscord<nocturn9x> gotcha :)
18:26:21FromDiscord<mjsdev> sent a code paste, see https://play.nim-lang.org/#pasty=PNOZfNrn
18:40:27*rockcavera quit (Remote host closed the connection)
18:51:39*ntat quit (Quit: Leaving)
19:22:28*coldfeet joined #nim
19:42:25FromDiscord<thepotatochronicler> Is there a way to evaluate a NimNode in a macro into a concrete value?
19:42:43FromDiscord<thepotatochronicler> And no, I don't mean `static[T]` as a param
19:49:31FromDiscord<nasuray> In reply to @spotlightkid "Yeah, basically the same.": You do not need nimble for config.nims support. Nimble uses nimscript but nimscript is builtin with the compiler. You can even make one at ~/.config/nim/config.nims and write tasks that can be invoked by the compiler as subcommands no different than a nimble task
20:02:42FromDiscord<spotlightkid> Nice! Thanks for the confirmation.
20:30:04*coldfeet quit (Quit: Lost terminal)
20:34:06FromDiscord<Elegantbeef> @thepotatochronicler Nope you cannot raise a NimNode to a true type as macros are compiled once
20:34:19FromDiscord<Elegantbeef> Like consider `var a = fromNimNode(if someCond: "test" else: 10)`
20:54:03*Guest70 joined #nim
20:58:02FromDiscord<oranwutan_78823> Hi! I'm trying to make a GUI with owlkettler, I need a ColumnView to be populated with data from an array (lets say a 5 x 5 array with random 0s and 1s), have anybody experience or know how to do it. I have read all the docs and made several (unsuccesful atempts).
20:58:21FromDiscord<oranwutan_78823> (edit) "it." => "it?"
20:59:56FromDiscord<Elegantbeef> Did you refer to https://github.com/can-lehmann/owlkettle/blob/c0c4e55f166dd74f32af0d4fef8c9b7d2f0de2e3/examples/widgets/column_view.nim ?
21:00:30FromDiscord<oranwutan_78823> Yes, they do:
21:00:31FromDiscord<oranwutan_78823> proc viewItem(row, column: int): Widget =↵ result = gui:↵ Label:↵ text = $row & ", " & $column
21:01:16FromDiscord<oranwutan_78823> But that $row and $ column just refer to the value of the index. I need to get access to the index to set the values from the aforementioned array
21:03:13FromDiscord<oranwutan_78823> thet renderse something like:↵0, 0 1,0 2,0 3,0
21:03:32FromDiscord<oranwutan_78823> 0,1 1,1 1,2 1,3
21:03:35FromDiscord<oranwutan_78823> and so on..
21:05:44FromDiscord<Elegantbeef> you have the row and column though
21:05:53FromDiscord<Elegantbeef> `myArray[x + y width]`
21:10:23FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=cQBwpfTl
21:14:53FromDiscord<oranwutan_78823> Ok! Let me take a look, thanks!!!
21:19:33*ensyde joined #nim
21:47:05FromDiscord<arkanoid> I'm experimenting writing code using same error handling pattern as Status code. I like it! It makes exception disappear and brings many benefits the the modules
21:47:24FromDiscord<System64 ~ Flandre Scarlet> In reply to @nocturn9x "this is my nim.cfg:": Wait↵We can enable SIMD?!
21:47:40FromDiscord<nocturn9x> that switch just enabled my own manual SIMD
21:47:48FromDiscord<System64 ~ Flandre Scarlet> Oh, sorry
21:48:01FromDiscord<nocturn9x> I have my engine's evaluation inference implemented using AVX2 intrinsics
21:48:12FromDiscord<nocturn9x> which are a significant speedup over the autovectorized stuff
21:48:37FromDiscord<System64 ~ Flandre Scarlet> No idea if I can take full advantage of AVX2 in my synth
21:48:52FromDiscord<nocturn9x> generally if you're doing a lot of integer ops on sequential data
21:48:57FromDiscord<nocturn9x> you probably can
21:49:13FromDiscord<System64 ~ Flandre Scarlet> my modular synth uses recursion
21:49:34FromDiscord<System64 ~ Flandre Scarlet> a module calls synthesize() in a deeper module and so on
21:57:20FromDiscord<aintea> Isn't there any way to break a line of code in nim ?
21:58:37FromDiscord<aintea> sent a code paste, see https://play.nim-lang.org/#pasty=wuYUyMFi
21:58:53FromDiscord<mjsdev> I think you need to leave the and on the line before
21:58:59FromDiscord<mjsdev> not as nice
21:59:03FromDiscord<mjsdev> but you could also use parens?
21:59:13FromDiscord<mjsdev> I think?
21:59:59FromDiscord<aintea> Parenthesis work, but I want to apply this to operators
22:00:19FromDiscord<aintea> sent a code paste, see https://play.nim-lang.org/#pasty=zlTBMjnv
22:00:27FromDiscord<aintea> and yes I have the dot operator
22:00:38FromDiscord<aintea> but I prefer this one
22:09:25FromDiscord<oranwutan_78823> In reply to @Elegantbeef "`myArray[x + y *": It is a nice example, not exactly what I need, I said an "array" but maybe it is a bad example. The big picture is that I need to render de content of a csv or excel file within the column viewer. So I need something like table[i, j] to be rendered.
22:09:47FromDiscord<oranwutan_78823> HOnestly, the API to columnviewer is hard to understand for me.
22:12:29*beholders_eye joined #nim
22:34:03FromDiscord<arkanoid> Elegantbeef\: hi! your `Guarded` module suggestion you sent me yesterday makes no calls to `initLock` and `deinitLock`. Is this a deliberate action or is it required to do so?
22:38:35FromDiscord<Elegantbeef> You should have a constructor which inits it and a destructor that deinits it
22:42:48FromDiscord<Elegantbeef> Here's the rest of the owl arkanoid https://play.nim-lang.org/#pasty=JUEYkvOF
22:42:59FromDiscord<Elegantbeef> Whoops
22:43:40FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=uaOwvpbD here's the full owl
22:43:56FromDiscord<Elegantbeef> We disable copies so we ensure there is only one instance of these made and you do not accidentally copy them
22:44:30FromDiscord<arkanoid> maybe you need also var for set?
22:44:30FromDiscord<Elegantbeef> We have a destructor so the lock is automatically deinited, and we have a constructor so you don't even need to pretend the lock exists 😄
22:44:42FromDiscord<Elegantbeef> Ah yes good catch
22:45:00FromDiscord<Elegantbeef> Also probably need var for get aswell since it acquires the lock
22:47:15FromDiscord<arkanoid> thanks for the owl
22:50:41FromDiscord<arkanoid> it doesn't like the that the destructor touches data without proper protection `guarded.nim(11, 21) Error: unguarded access: guarded.data`
22:51:29FromDiscord<arkanoid> Id say first destroy data with lock, then deinit lock
22:51:56FromDiscord<Elegantbeef> Right
22:53:04FromDiscord<arkanoid> mh, but acquire requires var, so this makes =destroy require var
22:53:34FromDiscord<arkanoid> but `Warning: A custom '=destroy' hook which takes a 'var T' parameter is deprecated; it should take a 'T' parameter`
22:54:41FromDiscord<arkanoid> also `nim-2.2.0/lib/std/private/syslocks.nim(175, 7) Error: cannot 'importc' variable at compile time; initSysLockAux`
22:55:26FromDiscord<Elegantbeef> Are you doing `const bleh = guard(...)`?
22:56:04FromDiscord<arkanoid> no, the error is given when compiling `guarded.nim` alone as is, with no use (still)
22:56:27FromDiscord<arkanoid> nim check guarded.nim
22:56:44FromDiscord<Elegantbeef> No issue here on 2.20
22:57:04FromDiscord<arkanoid> mh, let me skip my conf then
22:57:10FromDiscord<Elegantbeef> Everything is generic so there would be no instantiations until use so any error is clearly wrong
22:57:17FromDiscord<arkanoid> you're right, it's was my config, not sure what
23:06:54FromDiscord<arkanoid> I just had to add `return` and `var` to the `get` https://play.nim-lang.org/#pasty=cpkMWIyg
23:11:29*Guest70 quit (Quit: Client closed)
23:14:34FromDiscord<arkanoid> sent a code paste, see https://play.nim-lang.org/#pasty=hcnBgutT
23:15:27FromDiscord<Elegantbeef> Why does that matter?
23:15:36FromDiscord<Elegantbeef> Oh god you're doing a result of a guarded
23:15:49FromDiscord<Elegantbeef> That's no bueno as the lock moves iirc
23:17:11FromDiscord<arkanoid> \:(
23:17:32FromDiscord<Elegantbeef> The idea was to pass the state as a pointer then have the fields guarded then you do `.field.get`
23:19:16FromDiscord<arkanoid> that would be the same of having the pointer on the stack available for all
23:19:29FromDiscord<Elegantbeef> Except now you have safeties and no long have to cast gcsafe
23:21:37*jkl quit (Quit: Gone.)
23:27:02FromDiscord<arkanoid> true, but not sure which one is worst\: copying pointers explicitly between threads, or use cast
23:28:40FromDiscord<Elegantbeef> The copying doesn't provide safeties if the data changes from other threads 😄
23:35:50*jkl joined #nim
23:37:28FromDiscord<arkanoid> sure, I'm not saying that guarded is not necessary, I'm saying that copy pointer between thread is unnecessary, as state object is 1 and is global
23:37:29FromDiscord<arkanoid> I need to guard field access, and guarded should be sufficient. (I first need to untangle `Result[Config, ConfigError]` from my code)
23:37:29FromDiscord<arkanoid> I'm not quite sure what happens if a lock is copied
23:38:04FromDiscord<arkanoid> shouldn't be responsibility of Lock object to forbid copy and dup? https://github.com/nim-lang/Nim/blob/version-2-2/lib/core/locks.nim#L1
23:48:25FromDiscord<Elegantbeef> In fact it's preferable to prevent competing over a lock for writing
23:48:25FromDiscord<arkanoid> sure thing, but threads runs how the scheduler wants, without shared locks you can't prevent reader to be in the middle of writer op
23:48:26FromDiscord<Elegantbeef> After all you didn't describe what you're doing and I just pulled the Guarded idea out of my arse
23:48:26FromDiscord<Elegantbeef> I'm assuming there is no issue if a cycle runs with outdated info
23:50:58FromDiscord<arkanoid> it's a simple producer consumer task. Producer can change config, consumers read config and act accordingly. Producer may receive an updated config at any time and change the existing one
23:51:54FromDiscord<arkanoid> Guarded is OK. In my specific context I could use it to guard the whole Config object instead of the single field. This doesn't make a bug difference performance wise
23:55:05*beholders_eye quit (Ping timeout: 260 seconds)
23:55:11FromDiscord<arkanoid> but the issues are the same, copies are done and I need to change how I deal with errors
23:56:06FromDiscord<arkanoid> Btw I feel like i have all the pieces of the solution already, I just need to move them to stop Guarded fight with Result and make Threads happy. Thanks for the help
23:56:26FromDiscord<Elegantbeef> I did it?! 😛
23:59:59FromDiscord<arkanoid> btw I'm googling what happen when locks are copied, still no answers. I'm quite curious