<< 23-12-2022 >>
00:03:10FromDiscord<Elegantbeef> Use `--os:any` instead
00:03:38FromDiscord<impbox [ftsf]> hmm that gives↵`Error: OS module not ported to your operating system!`
00:03:45FromDiscord<Elegantbeef> Arrays and seqs are identical minus the heap allocation↵(@nqeron)
00:03:53FromDiscord<Elegantbeef> Yep you cannot use os on that OS
00:04:21FromDiscord<impbox [ftsf]> ahh... didn't think I was using os... but maybe it's included somehow...
00:05:10FromDiscord<Elegantbeef> https://github.com/PMunch/ratel if you arent using it
00:05:26FromDiscord<deech> sent a code paste, see https://paste.rs/Qyq
00:05:57FromDiscord<Elegantbeef> No offence intended but i do enjoy when people say "common" and i've literally never needed it 😄
00:06:38FromDiscord<impbox [ftsf]> hmm ratel sounds handy!
00:06:57FromDiscord<deech> Maybe it's cultural. It's all over the place in FP ecosystems. 🤷
00:07:23FromDiscord<Elegantbeef> I dont write conventional software so we'll blame that
00:17:50NimEventerNew thread by inv2004: Ttop - System monitoring service tool with tui and historical data, see https://forum.nim-lang.org/t/9751
00:20:18FromDiscord<geekrelief> Is there a way to get the generated C++ symbol name for a genSymed identifier? I'm trying to genSym a variable and pass it's name to an emit pragma which calls a C++ macro.
00:20:27FromDiscord<geekrelief> (edit) "it's" => "its"
00:22:51FromDiscord<geekrelief> sent a code paste, see https://play.nim-lang.org/#ix=4jjE
00:24:22FromDiscord<Elegantbeef> `signatureHash` i think
00:25:07FromDiscord<geekrelief> thanks beef!
00:27:36FromDiscord<impbox [ftsf]> ahh std/random imports os
00:28:44FromDiscord<geekrelief> hmm `signatureHash` isn't working, but maybe there's something else in macros
00:38:07FromDiscord<geekrelief> sent a code paste, see https://play.nim-lang.org/#ix=4jjI
00:38:16FromDiscord<geekrelief> (edit) "https://play.nim-lang.org/#ix=4jjI" => "https://play.nim-lang.org/#ix=4jjJ"
01:38:49FromDiscord<Hoss> anyone using nim on nixos? How should I work with nimler? Should I use the unwrapped package?
01:39:28FromDiscord<huantian> I use Nim on NixOS but I’ve never heard of nimler
01:40:10FromDiscord<huantian> nim (wrapped) == nim-unwrapped + nimble-unwrapped + other bundled Nim tools
01:53:58FromDiscord<Hoss> sorry yeah I meant nimble
01:54:27FromDiscord<huantian> Yeah you want just `nim` then
01:54:34FromDiscord<Hoss> so the regular nim package provides nimble, got it
01:54:34FromDiscord<huantian> That should come with the compiler and nimble
02:07:59FromDiscord<nqeron> `let (tileParsed, tileErr) = tileStr.parseTile↵↵ if tileErr: ↵ return (default(Game), false)` : Error: unhandled exception: index out of bounds, the container is empty [IndexDefect]
02:08:25FromDiscord<Elegantbeef> Ensure your index is in the collcetion you're indexing
02:08:29FromDiscord<nqeron> I've debugged it, and I know tileErr is getting a bool value
02:08:30FromDiscord<jtv> @ElegantBeef Thanks again for the help yesterday. As I mucked around with it all, I learned that most of my troubles were more likely from not remembering that I can't dot anything before the type parameter's bracket. And whenever I did it, all I got was an extraordinarily unhelpful "T cannot be instantiated", and generally recursion when I debugged 🙂
02:09:09FromDiscord<Elegantbeef> Your index defect will give you a line number
02:09:13FromDiscord<jtv> But I'm all good now. What's the reason why the member access before the type parameters is ambiguous?
02:09:27FromDiscord<Elegantbeef> What do you mean?
02:09:28FromDiscord<nqeron> In reply to @Elegantbeef "Ensure your index is": I don't know where it thinks it's trying to access an index - it's pointing to the if statement
02:09:39FromDiscord<Elegantbeef> Are you sure?
02:09:51FromDiscord<nqeron> can I share a repo?
02:09:56FromDiscord<Elegantbeef> Legally?
02:09:58FromDiscord<Elegantbeef> Yes
02:10:04FromDiscord<Elegantbeef> Ethically? Probably
02:10:13FromDiscord<nqeron> it's my repo - for myself
02:10:13FromDiscord<jtv> If I do something like: foo.func[typeParam]() it usually (though not always) says "Cannot instantiate T"
02:10:31FromDiscord<Elegantbeef> `foo.func[: typeParam]()`
02:10:38FromDiscord<nqeron> it has an MIT license - I meant would you actually be able to look at it?
02:10:40FromDiscord<Elegantbeef> UFCS is ambiguous with generics
02:10:46FromDiscord<Elegantbeef> I was joking nqeron
02:10:49FromDiscord<Elegantbeef> Yes it's fine
02:11:05FromDiscord<Elegantbeef> `foo.func[typeParam]` is parsed as `[](foo.func, typeParam)`
02:11:09FromDiscord<jtv> Oh, cause it might be an array access
02:11:10FromDiscord<jtv> Yeah
02:11:19FromDiscord<jtv> Ahhh
02:11:28FromDiscord<jtv> Interesting
02:11:31FromDiscord<Elegantbeef> The `: type` syntax overrides that and you get merry UFCS with generics
02:11:59FromDiscord<nqeron> https://github.com/nqeron/DitakticBot/tree/feature/display_parse
02:12:30FromDiscord<jtv> Great, I'll remember that. Might be nice to optionally support easy to type unicode brackets for type params too like « and »
02:13:31FromDiscord<Elegantbeef> "Easy to type"
02:13:50FromDiscord<Elegantbeef> It could be considered Nim already has unicode operators
02:13:58FromDiscord<jtv> On my keyboard (Mac) it's Option-\ and Option-|
02:14:06FromDiscord<jtv> So yeah, once you know they're there 🙂
02:14:12FromDiscord<Elegantbeef> I use linux and have 0 clue how to type unicode
02:14:37FromDiscord<Elegantbeef> nqeron how do you reproduce the issue?
02:14:48FromDiscord<Elegantbeef> I compiled and ran your main and no issue
02:14:55FromDiscord<Elegantbeef> Also you're pushing `.exe`s
02:15:33FromDiscord<jtv> Depends on what editor you use, emacs? I can give you some elisp for that
02:15:43FromDiscord<Elegantbeef> I use kate/nvim
02:16:10FromDiscord<jtv> Ah, I'm not help there 🙂
02:16:17FromDiscord<nqeron> In reply to @Elegantbeef "nqeron how do you": I'm running from nimble run
02:16:42FromDiscord<Elegantbeef> Oh sorry i need to be on a branch
02:17:10FromDiscord<nqeron> In reply to @Elegantbeef "Also you're pushing `.exe`s": damn - I'll deal with that
02:17:59FromDiscord<Elegantbeef> Odd the error moves with `--mm:orc`
02:29:34FromDiscord<Elegantbeef> Very go code 😄
02:33:26FromDiscord<Elegantbeef> I do not see why this error appears where it does
02:36:17FromDiscord<nqeron> I think it's with how I'm initializing the board
02:36:45FromDiscord<nqeron> something to do with setting the default for nested sequences
02:38:26FromDiscord<nqeron> the default(Board) gives me, well I'm not sure what
02:38:47FromDiscord<Elegantbeef> An empty sequence
02:39:09FromDiscord<nqeron> but I think I want to initialize my sequences so I have a seq[N, seq[N, Tile]]
02:39:23FromDiscord<Elegantbeef> I know it doesnt help your issue whatsoever but consider looking at `std/options`
02:39:29FromDiscord<nqeron> because it's trying to access board[][]
02:40:12FromDiscord<Elegantbeef> Yes you will want to do `import std/sequtils; myBoard = newSeqWith(N, newSeq[T](N))`
02:40:57FromDiscord<nqeron> ah - thanks
02:42:03FromDiscord<Elegantbeef> Some of your code is needlessly complex
02:42:14FromDiscord<Elegantbeef> `(self.ply == 0 and color == black) or (self.ply == 1 and color == white)` is just `(self.ply == ord(color)`
02:42:26FromDiscord<Elegantbeef> or `1 - ord(color)`
02:45:10FromDiscord<JeysonFlores> is IntelliJ IDEA Nim Plugin better than Nimsuggest?
02:45:26FromDiscord<Elegantbeef> Doubtful
02:45:54FromDiscord<Elegantbeef> Jetbrain's addon doesnt do any compilation on the code so macros/templates are not suggested for afaik
02:46:47FromDiscord<Elegantbeef> nqeron this is speed related but doing `for x in a[0..^2]` copies the seq so it's faster to do `for x in a.toOpenArray(0, x.high - 1)`
02:47:02FromDiscord<JeysonFlores> In reply to @Elegantbeef "Jetbrain's addon doesnt do": I see
02:47:15FromDiscord<Elegantbeef> I've never used the plugin to be clear
02:47:28FromDiscord<Elegantbeef> I just know they dont use nimsuggest and they clearly havent build a compiler from scratch
02:47:40FromDiscord<Elegantbeef> So their tooling is going to be lacking for any code that uses macros
02:47:42FromDiscord<Elegantbeef> Which is a fair bit
03:01:25FromDiscord<nqeron> In reply to @Elegantbeef "or `1 - ord(color)`": it's a work in progress - I plan on cleaning some of this up
03:02:00FromDiscord<nqeron> if I have a string with values of "x
03:02:08FromDiscord<Elegantbeef> If i dont needlessly give suggestions who will?!
03:02:55FromDiscord<nqeron> (edit) ""x" => ""x(\d)" and I want to replace them with "x,x,x" where x is repeated that number of times, what's the easiest way to do this?"
03:12:24FromDiscord<nqeron> could not load: pcre64.dll ?
03:12:44FromDiscord<Elegantbeef> your install doesnt have pcre in the path
03:12:58FromDiscord<Elegantbeef> It's probably better to use `nim-regex` or not to use regex at all
03:13:19FromDiscord<nqeron> as opposed to std/re?
03:13:26FromDiscord<Elegantbeef> Yes
03:13:44FromDiscord<jtv> Why are people so obsessed w/ regexps 🙂
03:13:48FromDiscord<Elegantbeef> `std/re` and `std/nre` both rely on the pcre library
03:13:52FromDiscord<Elegantbeef> I do not know jtv
03:13:57FromDiscord<Elegantbeef> I only use them when they make sense
03:14:02FromDiscord<Elegantbeef> User define searches for instance
03:14:37FromDiscord<ringabout> There is also a tinyre https://github.com/khchen/tinyre
03:15:07FromDiscord<ringabout> I'm going to link them in the documentation.
03:15:09FromDiscord<Elegantbeef> When you need a regex library it's nice to have but many people seem to shoehorn regex into a project
03:15:54FromDiscord<jtv> Yeah, and basic globbing in particular. But for any actual parsing, even if it can get the job done, it's just far less clear and more error prone than just using even a parser generator
03:18:14FromDiscord<Elegantbeef> How else do I ensure a string only contains letters?! 😜
03:19:18FromDiscord<jtv> Ha
03:20:00FromDiscord<Rika> parseutils parseutils parseutils parseutils parseutils parseutils parseutils
03:20:41FromDiscord<Elegantbeef> Hmmm do you hear that?
03:20:46FromDiscord<Elegantbeef> Sounds a lot like parseutils
03:20:50FromDiscord<jtv> It's amazing how little "senior" engineers know about strings in general. I know plenty that would still do that with [a-zA-Z], and if you expect them to know anything about Unicode beyond "stuff is generally utf-8 and I don't have to worry..."
03:21:23FromDiscord<Elegantbeef> hey depending on what i'm doing i'd do `str.allInSet({'a'..'z', 'A'..'Z'}`
03:21:28FromDiscord<Rika> unicode is hard to understand though so id say understandable
03:21:49FromDiscord<Elegantbeef> Luckily we have `isAlpha` and friends
03:22:21FromDiscord<nqeron> this is what I came up with, but it doesn't seem to be working as expected: boardStr.replace(re"x(\d)", newSeqWith(($1).parseInt, "x").join(","))
03:22:36FromDiscord<Elegantbeef> Regex t oreplace
03:22:48FromDiscord<jtv> Well, if you want to understand every nuance of the unicode database sure, but the basics CAN be presented pretty easily
03:23:04FromDiscord<jtv> Like, understanding what a codepoint is would be a fine start
03:23:05FromDiscord<Elegantbeef> What is the input string and what's the desired output string?
03:24:03FromDiscord<Rika> In reply to @jtv "Like, understanding what a": i think the names for those could be improved on
03:24:22FromDiscord<Rika> codepoint? glyph? grapheme? very confusing for most
03:24:57FromDiscord<Elegantbeef> Just make strings work with unicode operations
03:25:05FromDiscord<Elegantbeef> `a[30]` goes brrr
03:25:32FromDiscord<nqeron> input: "2,x2,1,x2/x4,1,x/x2,2,12,1,x/x2,12C,1,21C,x/x2,2,1S,12,x/1,x,2,x,1,x 2 12"
03:26:01FromDiscord<Elegantbeef> Yea just use parse utils
03:26:24FromDiscord<nqeron> I'll look into it
03:30:13FromDiscord<jtv> Sure, but as a programmer, glyphs and graphemes are rarely important, but knowing the difference between the abstract representation and the encoding is a big thing
03:31:11FromDiscord<jtv> But you're right. Codepoint's not awesome, but it doesn't help that there's not wide adoption of a single term. "Rune" isn't bad, but it's also not amazing. Codepoint's more clear, if more verbose.
03:33:45FromDiscord<Elegantbeef> Just call it fancy character and move along 😛
03:33:53FromDiscord<jtv> Lol yup
03:34:12FromDiscord<etra> when you use x86 and do `case`, how does 'optimally' compare strings? is there a intrinsic or something?
03:34:22FromDiscord<jtv> Frankly, if we don't care about memory anymore, just make it all UTF-32 and be done with all the subtle errors 🙂
03:34:24FromDiscord<etra> does it builds a jump table?
03:34:38FromDiscord<Elegantbeef> It's supposed to hash but from what i hear it doesnt
03:34:47FromDiscord<etra> oh nim is in godbolt
03:34:49FromDiscord<etra> nice
03:35:13FromDiscord<Elegantbeef> Ah yes the one standard that only one company uses 😛↵(@jtv)
03:35:53FromDiscord<jtv> Is that part of the nim compiler, or left to the underlying pipeline? I'd assume it just ends up in switch() statements and LLVM or gcc takes care of the rest?
03:36:22FromDiscord<Elegantbeef> Does C have switch statements for strings?
03:36:28FromDiscord<jtv> Oh, I was not serious
03:36:51FromDiscord<jtv> No, it doesn't.
03:36:54FromDiscord<Elegantbeef> It's supposed to be a precomputed hash and comparison
03:37:17FromDiscord<Elegantbeef> Yea i didnt think it did, but i've never written C so no clue if it actually did and i was living in ignorance
03:38:26FromDiscord<jtv> But I'm not sure what good hashing the strings does you in terms of speed-up for most cases with fairly small numbers of strings involved
03:39:01FromDiscord<Elegantbeef> Well it allows you do to fast removal when large amount of strings are involved
03:39:16FromDiscord<Elegantbeef> But yes indeed for smaller ones it's likely faster to just compare directly and skip the useless int comparison
03:39:45FromDiscord<jtv> What do you mean by fast removal? Thinking in the context of a case statement I don't get what you're saying
03:40:15FromDiscord<Elegantbeef> You have an O(1) comparison which is beneficial for many branches
03:40:26FromDiscord<Elegantbeef> So you only need the O(n) comparison on the hash hits
03:40:37FromDiscord<emanresu3> any news on Nim 2.0?
03:40:41FromDiscord<Elegantbeef> RC1 is out
03:41:52FromDiscord<jtv> Using hashing for amortized O(1) access is good for larger amounts of data, sure. I'm sure at some point you could get value from a jump table that uses a form of hashing (you'd have to go one byte at a time, I forget what the kind of hash is called, basically a tree hash)
03:41:56FromDiscord<etra> (snip) oh smort, it first compares the lengths of the cases to the length of the input string https://media.discordapp.net/attachments/371759389889003532/1055691697914400908/image.png
03:42:29FromDiscord<Elegantbeef> Much cheaper than hashing 😄
03:42:34FromDiscord<emanresu3> what was the tool to download different nim versions? I forgot I never installed it \:p
03:42:41FromDiscord<jtv> choosenim
03:42:46FromDiscord<Elegantbeef> Ironically i dont think the rc1 supports choosenim
03:42:55FromDiscord<jtv> Lol
03:42:56FromDiscord<emanresu3> thanks
03:46:07FromDiscord<jtv> BTW, Elegantbeef, I have done a lot on lock free and wait free data structures in C in the past, will eventually port it all to Nim. Don't agree with Araq's comments in the locking table code; in my experience, I can make some classes of data structures slightly more efficient than the best locking versions I can find; they're basically the same performance wise, except that threads holding locks get swapped out occasionally.
03:46:11FromDiscord<jtv> See: https://github.com/viega/hatrack.git
03:46:38FromDiscord<Elegantbeef> > Don't agree with Araq's comments in the locking table code↵> No clue what this means
03:46:48FromDiscord<Elegantbeef> > Don't agree with Araq's comments in the locking table code↵No clue what this means\ 😛
03:47:04FromDiscord<jtv> Oh, there's a table in the stdlib, and a comment in it saying, "While I'd accept lock free contributions, if you need that you're doing it wrong"
03:47:25FromDiscord<Elegantbeef> Ok so i have no clue why you're talking to me about it
03:47:40FromDiscord<Elegantbeef> I do multi thread code pretty much never
03:47:56FromDiscord<jtv> Ah, well there you go then, sorry 🙂
03:48:17FromDiscord<Elegantbeef> Unacceptable!
03:48:40FromDiscord<jtv> Just a lot of people I know who care about high scalability and performance are interested in many dimensions of it, sorry for presuming!
04:07:38FromDiscord<Bung> In reply to @jtv "BTW, Elegantbeef, I have": when you done this please also provide document
04:07:48*ltriant quit (Ping timeout: 272 seconds)
04:08:10FromDiscord<jtv> Will do, certainly. Not sure when I'll get to it... whenever I need it next 🙂
04:09:05FromDiscord<Bung> that's nice
04:11:00FromDiscord<jtv> It's not a small project, in large part since I'd have to deeply understand all the options for memory management in Nim and how it impacts what I need to provide. But I want to find the time at some point!
04:16:29FromDiscord<Rika> In reply to @Elegantbeef "Ah yes the one": microsoft uses 16 not 32 iirc
04:20:59FromDiscord<jtv> Yes. Worst of both worlds!
04:21:24FromDiscord<jtv> Have both memory expansion AND problems when people think random-access indexing is okay!
04:34:46FromDiscord<elcritch> Random question, what is the proper way to get a data pointer from an openArray? I don't think `addr someOpenArray` works.
04:35:04FromDiscord<Elegantbeef> `addr someOpenarray[0]`
04:53:09FromDiscord<Elegantbeef> `addr someOpenarray` is fickle it sometimes returns what you expect and other times doesnt
04:59:44*Figworm quit (Remote host closed the connection)
05:00:53*Figworm joined #nim
05:05:04FromDiscord<elcritch> Yah wouldn't an `openArray` with len 0 throw an index exception though?
05:05:40FromDiscord<Elegantbeef> An openarray with 0 data has no data pointer
05:09:36FromDiscord<elcritch> Hmmm, would that just be a nil pointer? I guess doing a length check isn't worse than check if the data pointer is nil. Maybe I was thinking too C like.
05:09:52FromDiscord<elcritch> @ElegantBeef didn
05:10:14FromDiscord<elcritch> (edit) "didn" => "didn't you write a remoteref type library? I can't seem to find it on your github"
05:10:30FromDiscord<Elegantbeef> https://github.com/beef331/nimtrest/blob/master/remoterefs.nim
05:10:33FromDiscord<Elegantbeef> Never moved it to it's own
05:27:50FromDiscord<elcritch> ok, I just want to point it out in https://forum.nim-lang.org/t/9745#64166
05:28:27FromDiscord<elcritch> We should split it out too! It'd be handy package in EmbeddedNim.
05:34:01FromDiscord<Elegantbeef> Sadly it doesnt work with `ref`
05:43:07FromDiscord<elcritch> How so?
05:43:24FromDiscord<Elegantbeef> You cannot make a `ref` that works with the CustomRef
05:44:03FromDiscord<Elegantbeef> Same with `seq` and `string`
05:49:33FromDiscord<elcritch> Sure you can, `ref CustomRef` right? 😉
05:50:01FromDiscord<Elegantbeef> That uses Nim's allocator
05:52:07FromDiscord<elcritch> Oh, you mean to original posters idea. No, but you can wrap a foreign Ref with your RemoteRef right and just treat C's malloc/free?
05:55:06FromDiscord<Elegantbeef> https://github.com/beef331/nimtrest/blob/master/remoterefs.nim#L18-L21 correct
05:56:30FromDiscord<Elegantbeef> I more mean you cannot have a customref with a `ref`, `string` or `seq`
05:56:46FromDiscord<Elegantbeef> You technically can, but it doesnt work how you'd want since it doesnt use those allocators for the type
05:57:18FromDiscord<Elegantbeef> So like if you use SPI allocators which afaik are persistent if you use a Nim GC'd memory type you lose said memory
05:57:55FromDiscord<Elegantbeef> That reason is why i disallow anything but stack
05:59:08FromDiscord<Elegantbeef> I do have a god awful syntax idea that could be used for adding support for `ref` `seq` and `string`
05:59:11FromDiscord<Elegantbeef> https://forum.nim-lang.org/t/9617#63238
06:02:32FromDiscord<Elegantbeef> If you think about allocators like contexts in Zig/Odin you ideally want to use the same allocators for a given type unless otherwise specified
06:13:23FromDiscord<JeysonFlores> Can't you assign default values for generic params? something like:↵`proc promptInput[T: int | float | string](msg: string, default: T = int: 0 | float: 0 | string: ""): T =↵ # stuff here`
06:13:40FromDiscord<Elegantbeef> Nope
06:13:51FromDiscord<JeysonFlores> sh1t
06:13:55FromDiscord<JeysonFlores> thanks
06:14:05FromDiscord<Elegantbeef> It's best to make 3 variations
06:14:22FromDiscord<JeysonFlores> yep, it seems like it. Thanks
06:14:38FromDiscord<Elegantbeef> The benefit of that is you really get a generic interface now
06:14:49FromDiscord<Elegantbeef> So you can make the same procedure for anytype easily
06:17:01FromDiscord<JeysonFlores> what about doing something like:↵`proc promptInput[T: int | float | string](msg: string, default: Option[T]): T =↵ # check if default is None and declare a variable with the default value according to its type`
06:17:33FromDiscord<Elegantbeef> You can always do this cleaner
06:20:07FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4jkr
06:20:39FromDiscord<Elegantbeef> Remove the `: int | float | string` for the most flexible api that allows more types dynamically
06:22:07FromDiscord<Elegantbeef> If we use concepts(a bit 'experimental' but work well enough for things like this) we can make this even relatively documented
06:22:10FromDiscord<Elegantbeef> https://play.nim-lang.org/#ix=4jks
06:23:03FromDiscord<Elegantbeef> This allows you to plug in your own static defaults or supply a default
06:24:24FromDiscord<JeysonFlores> holy crap I didn't know about the `default` function
06:24:44FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4jkt
06:25:02FromDiscord<Elegantbeef> `default(T)` returns a 0'd(in 2.0 this isnt true) version of the value
06:25:26FromDiscord<Elegantbeef> But even there it doesnt really give the most usable API for this
06:25:47FromDiscord<Elegantbeef> shit that should be `promptInput[Confirmation]`
06:26:18FromDiscord<Elegantbeef> Point being depending on how nice you want this API you may want to use more generics and define a generic API to parse the inputs to your specialised types to make your life possibly easier
06:27:17FromDiscord<Elegantbeef> This might be over-engineering for your simple problem though 😛
06:28:41FromDiscord<JeysonFlores> haha indeed, I was close to just give up and make one function per type. Thanks!
07:31:49*rlitp left #nim (#nim)
08:51:24*pusewicz joined #nim
08:55:01*pusewicz quit (Client Quit)
08:55:07*xet7 quit (Remote host closed the connection)
09:44:03FromDiscord<luteva> is there any (non-linear) video editor written in nim?
09:46:16FromDiscord<Elegantbeef> I doubt there is even a linear video editor written in nim 😄
10:24:44FromDiscord<demotomohiro> @luteva ↵mock up: A framework for developing video editing software↵https://github.com/mock-up/mock-up
10:25:40FromDiscord<demotomohiro> I dont know whether it is linear or non-linear.
10:25:40Amun-Radoes pragma (ie. push) have file scope only?
10:26:58*ltriant joined #nim
10:33:18FromDiscord<Bung> that's amazing project "video editor written in nim"
10:33:48FromDiscord<Bung> wait, is there video player in nim ?
10:38:16FromDiscord<luteva> In reply to @demotomohiro "<@954521401073754212> mock up:": thx! Now i only have to learn japanese, it seems 😄↵anyways: can someone add that project to "awesome-nim"? Maybe some more people will see it, use it and help.
10:39:30FromDiscord<luteva> e.g. there are no examples and the muml page is written in japanese 🙂↵but anyway great!
10:41:55FromDiscord<luteva> did someone already used mock-up? Just for a first example on how to use it....
10:42:49FromDiscord<luteva> or does: "Mutsuha Asada - momeemt" is here in the chat?
10:50:06FromDiscord<luteva> In reply to @Bung "wait, is there video": https://github.com/nim-lang-cn/video-player
10:50:45FromDiscord<Bung> I've checked that before it use `ffplay`
11:06:05FromDiscord<Phil> Out of curiosity, what is so experimental about the void type?
11:07:57*xet7 joined #nim
11:32:23FromDiscord<Bung> it will not produce a c type, it's nil in compiler when get the proc return type
11:33:11FromDiscord<Bung> well in explictly generic instaiation it is needed
11:34:01FromDiscord<Bung> so it's sssssssssssspecial handled in compiler compare to other type
12:00:27*genpaku quit (Read error: Connection reset by peer)
12:04:42*genpaku joined #nim
12:37:13*oddish quit (Quit: nyaa~)
12:38:07*oddish joined #nim
12:43:27FromDiscord<Phil> TIL that due to the removal of db_ from the standard libs it breaks the `ndb` package, which breaks `norm`, which breaks my application.↵So off to fix ndb and deal with weird docker issues, yaaaaaay
13:03:14FromDiscord<kaddkaka> sent a code paste, see https://play.nim-lang.org/#ix=4jlK
13:19:04FromDiscord<kaddkaka> `type Map[N: static[int], T] = array[N, array[N, T]]` seems to do the trick
13:20:19FromDiscord<chri> In reply to @Elegantbeef "*Imagine in the future": what is distinct?
13:21:41FromDiscord<kaddkaka> In reply to @chri "what is distinct?": https://nim-lang.org/docs/manual.html#types-distinct-type
13:27:17FromDiscord<Phil> In reply to @kaddkaka "`type Map[N: static[int], T]": Likely because array stuff needs to all be known as compile time and int is not inherently compile time, static blocks are though
13:34:23FromDiscord<kaddkaka> Sure I can understand that, the error message is shit tho 😛
13:44:05FromDiscord<Phil> Yeeeee fair
13:46:34NimEventerNew thread by mantielero: Question about memory management, see https://forum.nim-lang.org/t/9753
13:53:25FromDiscord<kaddkaka> Hmm, nimsuggest (which I think is the tool that supplies me with lint errors) reports a type mismatch, but the code compiles just fine o.O
13:59:54*redj quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
14:03:33*jmdaemon quit (Ping timeout: 252 seconds)
14:20:03FromDiscord<hotdog> In reply to @kaddkaka "Hmm, nimsuggest (which I": I’ve found this kind of thing happens with versions mismatched between the nimsuggest binary and the nim compiler
14:20:24FromDiscord<hotdog> Like if your editor is picking up an old nimsuggest on your path
14:20:55FromDiscord<hotdog> Or can happen when there are different flags (e.g backend) between what nimsuggest is using and what you are compiling with
14:32:00*oprypin joined #nim
14:41:13FromDiscord<ringabout> Is there someone familiar with the chrome driver? It really tortures me => https://github.com/nim-lang/nimforum/pull/337 Could someone more experienced please take control of this PR? I'm too inexperienced to work on it.
14:42:45NimEventerNew thread by c4UlMu: How to move variable to other thread, see https://forum.nim-lang.org/t/9754
14:43:34FromDiscord<kaddkaka> I just do `:make --run --showAllMismatches:on` with `makeprg=nim c --listFullPaths:on $ %`↵Both nim and nimsuggest seem to be 1.6.6 from 2022-05-05 🤔
14:48:39FromDiscord<Phil> In reply to @ringabout "Is there someone familiar": What is the chrome used for? Selenium?
14:49:17FromDiscord<Phil> (edit) "In reply to @ringabout "Is there someone familiar": What is the chrome ... used" added "driver"
14:49:27FromDiscord<ringabout> In reply to @Isofruit "What is the chrome": It is used for testing nimforum using webdriver implemented by Nim.
14:49:44FromDiscord<Phil> Ah, so selenium equivalent
14:49:51FromDiscord<ringabout> Yes
14:51:59FromDiscord<ringabout> It is https://github.com/dom96/webdriver
14:52:51FromDiscord<ringabout> I cannot even test it on the wsl2 since some bugs as to wsl installing chromium.
14:52:55FromDiscord<ringabout> (edit) "wsl" => "wsl2"
14:58:15FromDiscord<kaddkaka> sent a code paste, see https://paste.rs/4yn
14:58:27FromDiscord<kaddkaka> (but compilation is fine)
14:58:54FromDiscord<Phil> could ensure that N is static by also throwing a static around it
14:59:08*arkurious joined #nim
14:59:17FromDiscord<Phil> in the generics definition of the `[]` proc I mean
15:01:27FromDiscord<kaddkaka> In reply to @Isofruit "could ensure that N": Still the same type mismatch errors
15:01:31FromDiscord<kaddkaka> (edit) "In reply to @Isofruit "could ensure that N": Still the same type mismatch ... errors" added "lint"
15:01:52FromDiscord<Phil> In that case no idea.↵I have to admit that I do not really use nimsuggest because for me it barely faster than just running the compiler
15:01:59FromDiscord<Phil> So I typically just compile on save and be done with it
15:05:43FromDiscord<kaddkaka> strange it is ..
15:08:03FromDiscord<Phil> I just read through the newest thread in the saga of antivirus vendors falsely flagging nim binaries: https://forum.nim-lang.org/t/9744↵Reading through the links posted in there, 2 thoughts come to mind:↵1) It seems dumb as hell that basically you're forced to change the APIs you use because AVs just categorically flag this particular API↵2) Awesome work from you ringabout!
15:26:14FromDiscord<Phil> Hmm how do I stringify an identifier?
15:26:31FromDiscord<Phil> Not the value of the identifier, the identifier itself, for an error message
15:29:19FromDiscord<Phil> sent a code paste, see https://paste.rs/dkS
15:31:35FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jmn
15:32:51NimEventerNew thread by RodSteward: Foreign thread callbacks causes memory leak, see https://forum.nim-lang.org/t/9755
15:42:55NimEventerNew thread by wilsonywx2: Uncatchable JsonParsingError when encountered in a macro , see https://forum.nim-lang.org/t/9756
15:46:34FromDiscord<Phil> sent a code paste, see https://paste.rs/Fbv
15:47:32FromDiscord<HitBlast> Could somebody explain why am I getting this error while trying to read the file? https://media.discordapp.net/attachments/371759389889003532/1055874306804031538/Screenshot_2022-12-23_at_9.47.00_PM.png
15:47:47FromDiscord<jtv> From a macro there's a field in the node you can grab
15:48:30FromDiscord<jtv> strVal
15:48:43FromDiscord<Phil> In reply to @jtv "From a macro there's": I'm very low newbie levels of macro knowledge, so the statement is that an identifier is accessible as a macro-node in this kind of template?
15:48:54FromDiscord<jtv> One minute, let me get you something to do it
15:50:43FromDiscord<Rika> In reply to @HitBlast "Could somebody explain why": Parse file is relative to the binary and not the source file
15:52:58FromDiscord<HitBlast> In reply to @Rika "Parse file is relative": What should I put as path since the json file itself (which I'm trying to parse) is located inside a private folder.
15:53:06FromDiscord<jtv> sent a code paste, see https://play.nim-lang.org/#ix=4jmu
15:58:29FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jmw
16:00:16FromDiscord<Phil> Arrrgh doesn't work nicely within a template though
16:03:23FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jmz
16:10:06FromDiscord<Phil> ElegantBeef, I shall denounce your supreme intelligence status!↵For the stringification of identifiers in templates is but a fever dream
16:37:44*om3ga quit (Quit: Bye)
16:41:09*kenran joined #nim
16:41:14*kenran quit (Remote host closed the connection)
16:49:51FromDiscord<HitBlast> @Phil Sorry for the ping but could you help me with this?↵↵I'm trying to turn a `JsonNode` into a set of items. Take the `AVRO_VOWELS` variable as an example, it returns a string of characters if printed. Is there a way to make the object iterable? https://media.discordapp.net/attachments/371759389889003532/1055889989222809681/Screenshot_2022-12-23_at_10.48.23_PM.png
16:50:20FromDiscord<HitBlast> (edit) "make" => "convert" | "convertthe object ... iterable?" added "into an"
16:50:58FromDiscord<Phil> What is Avro Vowels to begin with?
16:51:07FromDiscord<Phil> Just a JS object with multiple key value pairs inside of it?
16:52:10FromDiscord<Phil> Either way, have you already looked at the iterators section of std/json ?↵https://nim-lang.org/docs/json.html#15
16:52:14FromDiscord<HitBlast> In reply to @Isofruit "What *is* Avro Vowels": just a pair of key and value https://media.discordapp.net/attachments/371759389889003532/1055890591269003345/Screenshot_2022-12-23_at_10.51.57_PM.png
16:52:22FromDiscord<HitBlast> with a string as the value
16:52:59FromDiscord<HitBlast> I'd like to separate each character and convert it into an iterable
16:55:10FromDiscord<Phil> So are you asking for an iterator over a string or for an iterator over all keys of a jsonnode?
16:55:32FromDiscord<HitBlast> iterator over a string
16:55:35FromDiscord<HitBlast> like for each character
16:55:52FromDiscord<HitBlast> in Python I'd just convert it into a list using `split()`
16:57:11FromDiscord<Phil> Have you, yknow, tried simply writing a forloop as if the string itself was a iterable like a seq[char] 😛 ?
16:57:49FromDiscord<Luckayla> I was going to say you don't need to split a string in python to do that
16:57:57FromDiscord<Luckayla> You can just iterate the string itself
16:58:03FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jmO
17:02:07FromDiscord<HitBlast> sent a code paste, see https://paste.rs/oj9
17:09:02*pro joined #nim
17:12:40*pro left #nim (#nim)
17:33:31FromDiscord<14k> Where can I find packages to import
17:33:45FromDiscord<14k> I remember there was a website with tons of them like pipy for python, but I forgot the url
17:35:24FromDiscord<Phil> In reply to @14k "Where can I find": The generally fastest way is just using nimble search in the terminal
17:35:32FromDiscord<Phil> If you want a webpage, nimble.directory exists
17:35:49FromDiscord<14k> Thanks, that's exactly what I wanted
17:39:23FromDiscord<14k> https://nimble.directory/docs/ down?
17:39:41FromDiscord<HuskyHacks> Can anyone point me in the direction of the docs on compiling a Windows program with a specified icon? Something like embedding an icon as a resource and setting it as the Windows icon? ty
17:40:08FromDiscord<14k> In reply to @HuskyHacks "Can anyone point me": Just compile into c and use recource hacker to add an icon
17:40:53FromDiscord<HuskyHacks> Yeah I was a little confused as to why the icon doesn't show up in Resource Hacker with a native Nim build
17:43:05FromDiscord<Phil> In reply to @14k "https://nimble.directory/docs/ down?": I don't think I've ever seen that, https://nimble.directory/ is on as normal
17:43:26FromDiscord<huantian> Nimble directory docs have been dead for a while
17:43:34FromDiscord<14k> In reply to @Isofruit "I don't think I've": Yes, the main site works, but whenever I try accessing anyone's docs it just doesn't work
17:43:53FromDiscord<14k> In reply to @huantian "Nimble directory docs have": ddos attack or shitty infrastructure?
17:45:51FromDiscord<Phil> For me its hit or miss.↵Some of those links link to external documentation which works as normal, others route to nimble based pages and those are borked as far as i can see
17:52:16FromDiscord<jtv> @Phil I'd assume it would work fine if you leave the macro returning untyped
17:52:32FromDiscord<jtv> Typing the macro means it can't run as early as you want it to
17:54:46FromDiscord<jtv> sent a code paste, see https://play.nim-lang.org/#ix=4jmV
17:55:12FromDiscord<jtv> arg.strVal returns a string, not a Nim Node
17:56:13FromDiscord<jtv> The thing about macros is you're writing code to operate on parse tree nodes that runs at compile time. But you still are running normal old code, the only big oddity is the semantics of the macro signature (the fact that you get nim nodes as inputs for instance).
17:56:59FromDiscord<Phil> sent a code paste, see https://paste.rs/L6K
17:57:10FromDiscord<Phil> That's on 1.9.1 (devel)
17:57:51FromDiscord<jtv> Ok one sec
17:58:06FromDiscord<Phil> Were you running 1.6? I'll try it out there
17:58:21FromDiscord<jtv> No, hang on
17:58:22FromDiscord<Phil> 1.6.10 prints the same
17:58:26FromDiscord<Phil> Alrighty
17:59:07*om3ga joined #nim
18:04:49FromDiscord<jtv> It's going to take more work to get it working with the dotexpr, so just give me 2 mins
18:08:16FromDiscord<jtv> sent a code paste, see https://paste.rs/yOL
18:08:44FromDiscord<jtv> Tho it isn't quite giving the behavior I'd expect
18:09:14FromDiscord<jtv> I'm going to refine it, one sec
18:10:05FromDiscord<Phil> I don't think it's reaching the error pragma, it seems to blow up somewhere in the macro beforehand ^^
18:10:45FromDiscord<Phil> Though providing almost exactly the intended behaviour for this particular step: Blow up when identifier is not there.↵Next step is "Blow up when identifier is there, but does not match a specified type"
18:12:43FromDiscord<jtv> sent a code paste, see https://paste.rs/OjM
18:14:34FromDiscord<jtv> I've been using Nim less than a month, but have done a lot of compiler work over the past 30 years, so the macros aren't dark magic for me. I was thinking at some point I should write some intro articles on interesting topics like that; the lack of docs and the poor tooling is def. an issue for Nim; I know many people who looked at it and bounced out quickly when they hit roadblocks
18:16:20FromDiscord<Phil> In reply to @jtv "I've been using Nim": I'd like to get into macros but it's a kind of "no usecase no desire to bite my way through the frustration and lack of understanding. But due to no knowledge I can't see the usecase" kind of thing.↵So far I'm of the opinion that just providing dozens of examples and explaining them would already be a massive improvement as they'll inevitably touch all the individual pieces of macro stuff that come
18:16:34FromDiscord<Phil> (edit) "In reply to @jtv "I've been using Nim": I'd like to get into macros but it's a kind of "no usecase ... no" added " =>"
18:16:43FromDiscord<jtv> Yup that's what is needed, far and wide through the language
18:17:04FromDiscord<Phil> ~~I'm doing what I can with the stuff I participate in~~
18:17:16FromDiscord<leorize> that macro of yours can be done with a template
18:17:36FromDiscord<Phil> In reply to @leorize "that macro of yours": The template I tried blew up on me
18:17:36FromDiscord<jtv> Like reading the docs on concepts in the page on experimental features, it didn't show a single example of using one, and didn't explain why you might or might not want to have `concept x, var n`, etc. Stuff that's basic
18:18:04FromDiscord<14k> how could I convert an int to a string?
18:18:17FromDiscord<Phil> Why on earth are the backticks needed in that though?
18:18:25FromDiscord<jtv> In the template?
18:18:27FromDiscord<Phil> In reply to @14k "how could I convert": The answer you are looking for is described within the docs of std/strutils
18:18:30FromDiscord<jtv> Or the macro?
18:18:34FromDiscord<14k> In reply to @Isofruit "The answer you are": aight
18:18:40FromDiscord<Phil> In reply to @jtv "Or the macro?": ` {.error: "'" & `fullLit` & " does not exist!".}`
18:18:44FromDiscord<Phil> In the error pragma
18:18:45FromDiscord<14k> also why tf is the output of execShellCmd an int and not a string
18:18:52FromDiscord<Phil> Is that a way of stringifying nimnodes?
18:18:58FromDiscord<jtv> So the quote() function makes it all work like a template more or less
18:19:23FromDiscord<jtv> It will substitute local variables, though those variables HAVE to be nim nodes
18:19:27FromDiscord<jtv> And only works on locals
18:19:37FromDiscord<jtv> Anything else will give you random crashes
18:19:42FromDiscord<leorize> I'd avoid strformat tbh↵(@Phil)
18:19:49FromDiscord<leorize> and use `system.astToStr`
18:20:14FromDiscord<Phil> In reply to @leorize "I'd avoid strformat tbh": when working with macros maybe, in general in procs and generics I'll use that over string concatenation any day of the week
18:20:21FromDiscord<jtv> How's that differ from macro.toStrLit()
18:20:53FromDiscord<leorize> you can use it outside of a macro
18:21:11FromDiscord<jtv> I don't understand the subtlety of where fmt"x" is different than"x".fmt() yet btw, so might be the issue
18:21:18FromDiscord<jtv> Ah, got it, thanks
18:21:45FromDiscord<leorize> `fmt"x"` interprets `"x"` as a raw string
18:21:50FromDiscord<Phil> In reply to @jtv "So the quote() function": So basically this works because in terms of evaluation order it... first creates the "fullLit" and fills it with a value, then turns the template defined with quote do into actual code and runs it, capturing "fullLit" in a sense?
18:22:07FromDiscord<Phil> No wait, there's no way fullLit gets evaluated first
18:22:14FromDiscord<leorize> it's the same as `fmt(r"x")` if you may
18:22:18FromDiscord<jtv> I don't think quote() is the same implementation of substitution as templates. It's similar but different
18:22:19FromDiscord<14k> sent a code paste, see https://paste.rs/ofJ
18:22:29FromDiscord<jtv> Got it, that's very helpful, thanks, leorize
18:23:10FromDiscord<Phil> sent a code paste, see https://paste.rs/aQM
18:23:26FromDiscord<14k> whenever someone visits the website at localhost:5000, it executes a command and displays the output on the website
18:26:05FromDiscord<jtv> Nim giving full access to the parse tree is amazing, and the tooling makes it really pretty easy, but in a way that leads to unreadable code because it was all auto-generated. I think quote do: is better for readability where possible, if you know about how to work around its limitations. Ideally would eventually want to build something like it, that allows you to get node references to parts of the content where you're going to want to make
18:26:57FromDiscord<jtv> Otherwise I end up building things kind of in reverse order to try to keep the code as clear as possible
18:27:20FromDiscord<leorize> sent a code paste, see https://paste.rs/DjN
18:27:31FromDiscord<leorize> sorry for the short names, I'm typing on phone
18:28:20FromDiscord<Phil> In reply to @14k "whenever someone visits the": I would generally want to question as to why you want to do that in a one-liner.↵As for this though, your issue is that just because you're throwing a comma around it doesn't mean it's concatenated.↵You want to do ` resp "Result: " & $execShellCmd @"cmd"`↵($ is the general nim symbol for stringifying something, basicall nim's version of "toString")
18:28:51FromDiscord<Phil> In reply to @leorize "sorry for the short": The fact you're trying on phone at all is impressive, I just give up on even trying to type examples while on phone, thanks!
18:29:16FromDiscord<jtv> Is there a language rationale anywhere? Haven't been able to find one, and often get curious, e.g., why they chose & for string concat
18:29:41FromDiscord<leorize> `&` because it's not `+` \:p↵(@jtv)
18:30:04FromDiscord<14k> In reply to @Isofruit "I would generally want": thanks
18:30:07FromDiscord<Phil> In reply to @leorize "`&` because it's not": Why exactly is + for string concat so frowned upon?
18:30:21FromDiscord<jtv> Well, a) Why not +? Most languages use it. And b) Why not | which is string concatenation in many formal settings for strings
18:30:35*elph_ joined #nim
18:30:35FromDiscord<leorize> more serious answer\: it's the same as array concat, leaving `+` open for other math stuff you can do with arrays
18:31:12FromDiscord<Phil> So in case I want to view and treat a string as a seq[char] and have better operators for that the `+` symbol is left undefined for that?
18:31:25FromDiscord<leorize> also, Araq really hates `+` on seqs, so that might have been more important \:p
18:31:37FromDiscord<Phil> That's... maybe I just don't write enough low-level code to get that one
18:31:45FromDiscord<Phil> "low-level" in the sense of "close to the compiler"
18:32:00FromDiscord<leorize> just consistency I'd say↵(@Phil)
18:32:00FromDiscord<jtv> Okay, got it. It's one of the few choices in the language I don't love, and one of the first things my co-workers I've been forcing to use it have complained about
18:32:12FromDiscord<leorize> all concat is `&` regardless of containers
18:32:18FromDiscord<Phil> In reply to @leorize "just consistency I'd say": I'd argue that strings are different from seqs in enough ways that they deserve special treatment
18:32:41FromDiscord<leorize> there's always something to complain about when you switch languages \:p↵(@jtv)
18:33:22FromDiscord<jtv> It's not a major thing for me, more a curiosity. People can get over it, and that rationale makes sense. I like to understand why things are the way they are.
18:33:46FromDiscord<jtv> I think it is helpful for people coming from other langs to understand the WHYs when things are going to be significantly unexpected
18:34:03FromDiscord<leorize> wait until they try to do threading and you will never hear the end of it
18:34:49FromDiscord<Phil> ~~See, this is where I'm safe, I let the webservers provided by frameworks I use do the threading so I never have to deal with it~~
18:35:00FromDiscord<jtv> Yeah I need to port over my lock-free, wait-free data structures from C, I am somewhat dreading to wading into the threading model and mem management, there's clearly a lot there (lots of options too)
18:37:07FromDiscord<14k> does nim compile cross platform?
18:37:17FromDiscord<14k> like can I compile on mac and run it on linux
18:38:40FromDiscord<jtv> I don't know the compiler toolchain well but it clearly does, since the compilers (and often llvm underneath them) do. Generally cross-compiling is a bit of an anti-pattern these days, you can generally cheaply automate building natively on all common platforms
18:38:58FromDiscord<14k> In reply to @jtv "I don't know the": nice 👍
18:39:22FromDiscord<jtv> The default flow is nim gets compiled down to C code and fed to your C compiler
18:39:26FromDiscord<14k> this language is amazing ngl
18:39:35FromDiscord<14k> dunno why more ppl don't talk abt it
18:39:48FromDiscord<jtv> So if there's not a flag in the nim compiler (I'm also new to the language), you can def stop it when it generates C
18:40:33FromDiscord<jtv> Yeah, I was just talking about it. It's been getting some buzz lately, but the lack of handholding intro docs def causes many people to bounce out who would otherwise love it
18:41:01FromDiscord<jtv> And also the tooling ecosystem isn't that advanced or that well documented either
18:41:12FromDiscord<jtv> But otherwise, I agree, amazing, amazing language
18:41:40FromDiscord<jtv> I am throwing away C and Python as my two gotos for 30 years
18:41:57FromDiscord<14k> it's a good combo of the 2
18:43:48FromDiscord<14k> All I wanted was a pretty looking compiled language
18:47:17FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jn4
18:48:43FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jn5
18:50:17FromDiscord<leorize> I guess you can try if `is` work here
18:51:03FromDiscord<leorize> `someSymbol is typ`, basically
18:51:21FromDiscord<leorize> or wrap whatever you have in a `compiles()`
18:51:45FromDiscord<jtv> Sorry I can help, just got stuff going on w/ my wife and holiday
18:52:32FromDiscord<Phil> In reply to @jtv "Sorry I can help,": All good, I'm experimenting a bit for now
18:52:35FromDiscord<jtv> But I agree, "is" will probably allow you compare the two types there
18:54:51FromDiscord<Phil> sent a code paste, see https://paste.rs/vTX
18:55:09FromDiscord<Phil> `/home/philipp/dev/playground/src/playground.nim(10, 14) Error: 'x' exists but there is no overload with type 'proc (a: int): int'`
18:55:27FromDiscord<leorize> btw, you can use `elif`
18:55:53FromDiscord<Phil> In reply to @leorize "btw, you can use": I had some wonky behaviour at one point with elif in the past so I got out of the habit of using it with when, fair point though
18:55:54FromDiscord<leorize> and exploit `compiles` if you must
18:55:55FromDiscord<jtv> I think you would need typeof(someSymbol)
18:56:09FromDiscord<jtv> I can play around w/ it later
18:56:39FromDiscord<14k> sent a long message, see https://paste.rs/yph
18:56:42FromDiscord<14k> /usr/lib/nim/pure/concurrency/threadpool.nim(21, 10) Error: Threadpool requires --threads:on option.
18:56:45FromDiscord<14k> What's this?
18:56:59FromDiscord<Phil> In reply to @14k "/usr/lib/nim/pure/concurrency/threadpool.nim(21, 10": A compiler flag
18:57:13FromDiscord<leorize> pass `--threads:on` to `nim c`
18:57:32FromDiscord<pouriya.jamshidi> sent a code paste, see https://play.nim-lang.org/#ix=4jna
18:57:32FromDiscord<14k> ight thanks
18:58:04FromDiscord<Phil> In reply to @pouriya.jamshidi "Hi could anyone": You willing to use an ORM?
18:58:42FromDiscord<Phil> And if so, are you using sqlite/postgres?
18:59:09FromDiscord<pouriya.jamshidi> In reply to @Isofruit "You willing to use": I am more interested to know how to keep/utilize the same structure for the Database class.↵Although answers with ORM are welcome
18:59:15FromDiscord<pouriya.jamshidi> MS SQL
18:59:45FromDiscord<Phil> Errrr okay, in that case the ORM question goes out the window. There is one (norm) but it only covers postgres and sqlite for now
19:00:12FromDiscord<pouriya.jamshidi> I see
19:01:25FromDiscord<Phil> As for ms sql, I don't know if we have a wrapped driver for it
19:01:28FromDiscord<jtv> There's definitely a package for Python programmers that makes things like `class` available for you so that the syntax is more familiar to you
19:01:55FromDiscord<Phil> In reply to @jtv "There's definitely a package": the ms sql driver may be a much larger problem than just the class syntax
19:02:15FromDiscord<jtv> I know, just trying to answer the other part of his question 🙂
19:03:48FromDiscord<Arathanis> might have to use raw `db_odbc`, yeah?
19:03:55FromDiscord<pouriya.jamshidi> that class is just holding statements in it. nothing fancy and we do not do OO here 😄
19:04:12FromDiscord<jtv> The language does have inheritance
19:04:25FromDiscord<jtv> It's just not emphasized, and rightly so 🙂
19:04:35FromDiscord<Arathanis> OOP is good
19:04:39FromDiscord<Arathanis> but like all good tools
19:04:41FromDiscord<Arathanis> you can overuse it
19:04:50FromDiscord<Arathanis> important to know when and wear to use tools
19:04:59FromDiscord<Arathanis> (edit) "wear" => "~~were~~ where"
19:05:01FromDiscord<Phil> In reply to @pouriya.jamshidi "MS SQL": Yeah can't find an MS-SQL driver package.↵You could tackle that by wrapping it on your own or db_odbc if that even works for this
19:05:02FromDiscord<Arathanis> also important to know how to spell
19:05:19FromDiscord<Arathanis> (edit) "~~were~~" => "~~wear~~"
19:05:56FromDiscord<Phil> In reply to @pouriya.jamshidi "that class is just": You can use OO or not, both works. If you want to use OO, you'll want to use `method` instead of `proc`
19:05:58FromDiscord<jtv> Sure, but it's super widely used, and the use cases where inheritance is the right abstraction isn't all that large
19:06:05FromDiscord<Arathanis> In reply to @Isofruit "Yeah can't find an": it mentions MS SQL, albiet its by saying it hasn't been "extensively tested against it"
19:06:25FromDiscord<Phil> In reply to @Arathanis "it mentions MS SQL,": You googled and CTRL+F'ed better than I did
19:06:39FromDiscord<Arathanis> In reply to @jtv "Sure, but it's super": google fu strong
19:06:46FromDiscord<Arathanis> google fu strong
19:07:11FromDiscord<Arathanis> In reply to @jtv "Sure, but it's super": if your inheritance trees go over depth 2, you should take a hard look at what you are doing. If they go over depth 3, you should stop and rethink what you are doing
19:07:26FromDiscord<Arathanis> composition will save you
19:07:50FromDiscord<Arathanis> dislaimer: my opinions
19:07:53FromDiscord<Arathanis> (edit) "dislaimer:" => "disclaimer:"
19:08:14FromDiscord<pouriya.jamshidi> In reply to @Arathanis "if your inheritance trees": I used to work at a place with inheritance levels that would go 40~60 levels deep... and they were super proud of it... academics 🙂↵I still have PTSD from that time
19:08:21FromDiscord<jtv> In general, dynamic dispatch often ends up obfuscating control flow too much leading to fragile base classes, etc
19:08:27FromDiscord<Arathanis> In reply to @pouriya.jamshidi "I used to work": 😬
19:08:48FromDiscord<pouriya.jamshidi> best part is, it was their own DSL, so no IDE help either
19:08:55FromDiscord<Arathanis> oh god lol
19:08:58FromDiscord<pouriya.jamshidi> haha
19:09:05FromDiscord<Phil> Honestly the main reason why I'm playing around with this kind of template-y thing that I've been throwing around is because I want to have "module-wide interfaces".↵Basically an idea lifted from OO with Object interfaces.↵There are like 4+ different packages to basically turn modules into objects, which also would resolve any issue regarding mocking of procs for testing purposes, but that's an OO-style solution and I'd like to see the "
19:09:08FromDiscord<Arathanis> im super excited for concepts to mature, love me some structural pattern matching
19:09:35FromDiscord<Arathanis> (edit) "concepts" => "`concepts`"
19:09:55FromDiscord<Arathanis> its a little too beta right now, i check in on it every few months
19:10:08FromDiscord<Arathanis> i think work has stagnated, and thats fine. its rather niche and experimental
19:10:10FromDiscord<Arathanis> i just want it
19:10:12FromDiscord<Arathanis> lol
19:13:56FromDiscord<Arathanis> In reply to @Isofruit "Honestly the main reason": if what you are trying to do is not too complex, `concepts` might support your use case despites its currently rough edges.
19:14:16FromDiscord<Arathanis> lets you declare a structural contract and anything that meets that contract can be used in place of the concept type.
19:14:47FromDiscord<Phil> In reply to @Arathanis "if what you are": I am aware of concepts, even tried to use them at one point but fell off when it was more work than I wanted to invest into it
19:15:00FromDiscord<Arathanis> my exerpience to a T, that or just unsupported thinsg
19:15:09FromDiscord<Arathanis> once they get `each T` working its gonna open the flood gates.
19:15:19FromDiscord<Arathanis> (edit) "exerpience" => "experience"
19:15:26FromDiscord<Phil> As for this one in particular, it's really just a simple "ensure that this module has these procs with these type definitions
19:15:36FromDiscord<jtv> The docs for concepts suck but they work well
19:16:04FromDiscord<Phil> And maybe a bit of "ensure that this module has this exported variable"
19:16:06FromDiscord<jtv> They're basically interfaces but more generic
19:16:11FromDiscord<Arathanis> In reply to @jtv "The docs for concepts": i find it 50/50. theres a mismash of what the docs support and what the still open rework github issue specifies.
19:16:21FromDiscord<jtv> Sounds cool Ohil
19:16:25FromDiscord<Arathanis> doesn't help that the docs are straight up not accurate last i checked.
19:16:42FromDiscord<Arathanis> the examples do not work
19:16:47FromDiscord<Arathanis> not all of them at least
19:16:57FromDiscord<Arathanis> lookin' at you Functor.
19:17:14FromDiscord<Phil> In reply to @jtv "Sounds cool Ohil": I'm honestly not sure there's a usecase outside of databases. Just with databases you so often have absolutely identical code and it only differs in the type of connections being built
19:19:04FromDiscord<Arathanis> In reply to @Isofruit "I'm honestly not sure": maybe a macro?
19:19:23FromDiscord<Arathanis> like a initializer / verifier
19:19:28FromDiscord<Arathanis> (edit) "a" => "an"
19:20:08FromDiscord<Phil> In reply to @Arathanis "maybe a macro?": If I have to dive into guts that deep I'd like to pull some tooling out of it that allows anyone to solve this kind of problem for themselves.
19:20:14FromDiscord<Phil> (edit) "In reply to @Arathanis "maybe a macro?": If I have to dive into ... guts" added "compiler"
19:20:34FromDiscord<Phil> So that nobody else has to
19:21:06FromDiscord<jtv> Yeah, the examples for concepts don't even show USAGE
19:21:39FromDiscord<jtv> And there's no explanation without reading the code about when you should declare a var of something on the concept line, etc
19:22:13FromDiscord<Arathanis> yeah
19:22:18FromDiscord<Arathanis> and you have to cross reference with this https://github.com/nim-lang/RFCs/issues/168
19:22:28FromDiscord<Arathanis> current implementation is a mishmash of these
19:22:34FromDiscord<jtv> Yup
19:23:09FromDiscord<Arathanis> sent a code paste, see https://play.nim-lang.org/#ix=4jnk
19:23:21FromDiscord<jtv> Getting the most out of the type system is tough if you are expecting good docs and error messages
19:23:28FromDiscord<Arathanis> (edit) "https://paste.rs/qC0" => "https://paste.rs/fMl"
19:24:38FromDiscord<jtv> It took me forever to realize that an error that was just "cannot instantiate T" was not a recursion issue, it was the whole UFCS ambiguity. Then ElegantBeef told me about [:
19:25:15FromDiscord<Arathanis> you either solve the issue yourself or wait long enough for beef to tell you how you messed up
19:25:49FromDiscord<jtv> Yup he was great help
19:25:58FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jnm
19:26:59FromDiscord<Phil> Or rather I need to first get how to even check if a single specific proc is defined with correct type
19:27:14FromDiscord<Arathanis> i think definitely macros
19:27:40FromDiscord<Arathanis> ive done similar things before with macros, it definitely works
19:27:52FromDiscord<Phil> Maybe I should just go with https://github.com/itsumura-h/nim-interface-implements
19:27:59FromDiscord<leorize> figure the data structure first, and the rest will just follows
19:28:15FromDiscord<leorize> and doesn't beef have a macro that does it for you?
19:29:32FromDiscord<Phil> ~~Beef gave me the first borked template~~
19:29:51FromDiscord<arkanoid> What's the Nim composition-based idiom/pattern that replaces OO pattern where the user is requested to extend a library class and implement some methods?
19:30:47FromDiscord<jtv> mixin?
19:31:15FromDiscord<jtv> Not sure, I've seen the keyword and mixins were often that in other language envs in the past
19:31:20FromDiscord<Phil> Mixin is for generics
19:31:40FromDiscord<Phil> generics typically try to draw in symbols that are defined within the module that the generic itself is defined in
19:31:55FromDiscord<leorize> concepts was supposed to be it
19:32:01FromDiscord<leorize> it never worked for too long, though
19:32:11FromDiscord<Phil> By using "mixin" in that generic you tell it to look for that symbol in the module that calls the generic
19:32:49FromDiscord<@thatrandomperson5-6310e3b26da03> Im getting the issue described here\: [nim-lang/Nim#4555](https://github.com/nim-lang/Nim/issues/4555) on nim 1.6.10
19:33:57FromDiscord<@thatrandomperson5-6310e3b26da03> please help
19:34:28FromDiscord<leorize> please provide a your broken code sample
19:34:31FromDiscord<Phil> So you can have a generic proc in module B that relies on a proc "flyingPotato" to be around. But there is no flying potato defined in module B or any module it imports.↵Module A imports module B.↵Module A does have flyingPotato defined.↵Due to the usage of mixin, the "flyingPotato" proc from module A is used and module B leaves that symbol "free" or "opens it up" until that code gets to module A, or whatever that is actually called.
19:34:43FromDiscord<jtv> That's helpful, thanks, Phil
19:35:09FromDiscord<Phil> Generics are like the one feature of nim that I use very extensively
19:35:55FromDiscord<jtv> Cool, will def come to you when I am having future challenges with their semantics 🙂
19:36:17FromDiscord<Phil> Luckily it's not that hard ^^
19:36:21FromDiscord<Phil> But feel free
19:36:31FromDiscord<Phil> (edit) "^^" => "^^, so I doubt there'll be many questions"
19:37:56FromDiscord<jtv> Oh, generics in general I'm down with, just have had a lot of obtuse error messages.
19:38:36FromDiscord<jtv> It's all very well done and powerful.
19:38:59FromDiscord<jtv> Though I do wish people would try some syntax for type parameters other than making it look like an array access
19:39:19FromDiscord<Phil> Except for when you go like 3-4 layers of generics down and suddenly the error tells you that this code can't be generated because there's a type mismatch somewhere but it doesn't tell you where >_>
19:39:34FromDiscord<Phil> Nor what types are mismatching >_>
19:39:35FromDiscord<jtv> Like how about putting them as actual parameters but with "type" in their name or something like that
19:39:46FromDiscord<jtv> Yeah, that's what kept happening, I was using recursion
19:40:08FromDiscord<jtv> And I thought I was doing something wrong, and was just getting "Can't instantiate T" with NO diagnostics
19:40:32FromDiscord<Arathanis> i guess interface recommendation right now is a tuple of procs? interesting
19:40:36FromDiscord<jtv> And kept trying 100000 different things, and I eventually figured out it was the UFCS ambiguity, which I think will bite a LOT of novices
19:40:57FromDiscord<Phil> Solved that one in particular by just assigning those generic proc parameters (often times procs themselves) to a variable with a (generic) type definition first so that this assignment would fail and give me a useful error message
19:41:06FromDiscord<jtv> Concepts are good for that
19:41:20FromDiscord<jtv> Ohhh good idea Phil
19:41:35FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4jnv
19:42:35FromDiscord<jtv> Thanks 🙂
19:43:25FromDiscord<Phil> Happy to help with what I've learned so far
20:06:18*jmdaemon joined #nim
20:17:39*Lord_Nightmare quit (Ping timeout: 256 seconds)
20:47:00*droidrage quit (Remote host closed the connection)
20:47:00*qwestion quit (Remote host closed the connection)
20:47:00*ixmpp quit (Remote host closed the connection)
21:11:33FromDiscord<! Nilts> sent a code paste, see https://play.nim-lang.org/#ix=4jnQ
21:12:17FromDiscord<! Nilts> (edit) "https://paste.rs/evo" => "https://paste.rs/GRL"
21:23:56Amun-Rado you write types first, consts next or the other way?
21:28:27FromDiscord<Elegantbeef> `astToStr(name)`↵(@Phil)
21:36:14*Lord_Nightmare joined #nim
21:48:03*dimi joined #nim
21:49:39*dimi quit (Client Quit)
21:50:54FromDiscord<! Nilts> In reply to @not logged in "As my gitter account": I'll need to go in 10 min, if anyone has an answer ping me
21:58:08*ixmpp joined #nim
21:59:39*ltriant quit (Ping timeout: 256 seconds)
22:07:21*redj joined #nim
22:11:46FromDiscord<JeysonFlores> What's the difference between a template and a macro in Nim? Both execute code at compile time right?
22:12:01FromDiscord<Elegantbeef> templates are code subsitution
22:12:03FromDiscord<Elegantbeef> Macros run code on the vm
22:13:24FromDiscord<JeysonFlores> VM?
22:13:44FromDiscord<JeysonFlores> I thought that was only used for NimScript
22:17:39*wallabra quit (Quit: ZNC 1.8.2 - https://znc.in)
22:18:11*redj quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
22:19:08*redj joined #nim
22:20:11*wallabra joined #nim
22:27:30Amun-Ramacros tinker with ast, templates work like cpp macros
22:28:23FromDiscord<arkanoid> https://nim-lang.org/docs/manual.html#generics-mixin-statement↵↵I don't understand mixin statement. The documentation doesn't say much about it
22:29:09Amun-Raarkanoid: have you read https://en.wikipedia.org/wiki/Mixin ?
22:29:58FromDiscord<arkanoid> Nope! Never used them in OOP. Sorry I didn't know it was an OOP concept. Reading now
22:30:18FromDiscord<Anna Aurora 🏳🌈🏴☠🕳> ugh, Why do I get pinged from this room all the time even though you pals aren't even mentioning me?
22:30:57Amun-RaI did, Python-wise
22:37:34FromDiscord<arkanoid> When trying to convert the software patterns I've learned during my OOP time into nim, I'm finding myself quite lost. I know that OOP in Nim is not as advanced as other programming languages, but I fail to match the corresponding alternative using object variants, or other patterns available at compile time
22:38:23FromDiscord<arkanoid> I quite need an OOP to procedural manual, leveraging on Nim tools
22:39:51FromDiscord<Elegantbeef> Mixins in Nim are used in generics to say "Consider the scope of instantiation for this symbol"
22:39:52FromDiscord<Elegantbeef> Consider a serialiser you might have `serialiseData` and you want to allow overriding it per user type
22:39:52FromDiscord<Elegantbeef> This allows you to define a generic procedure but implement parts of it inside user code
22:40:08FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/sOw
22:40:37FromDiscord<Elegantbeef> Or whatever you want
22:40:37FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4joh
22:41:00FromDiscord<Elegantbeef> I generally call this a generic interface as it's dispatch at compile time but behaves like a interface in other languages
22:42:37FromDiscord<arkanoid> Is mixin preferred to method+inheritance in Nim? I mean, is it more idiomatic or gives more advantages?
22:42:44FromDiscord<Elegantbeef> Depends what you're doing
22:42:47FromDiscord<Elegantbeef> Mixins are dispatched statically
22:42:52FromDiscord<Elegantbeef> So if you can do what you want statically, yes
22:43:04FromDiscord<Elegantbeef> If you cannot then you cannot replicate it with method + inheritance
22:44:19FromDiscord<Elegantbeef> @albassort\: gintro now should work on your distro
22:44:54FromDiscord<albassort> Do you send me this like 3 weeks after I wrote all the Gtk code
22:45:05FromDiscord<Elegantbeef> If only you updated the issue
22:45:13FromDiscord<albassort> I don't want to bother them
22:45:16FromDiscord<Elegantbeef> Stefan fixed it relatively quickly
22:45:16FromDiscord<arkanoid> This seems a big pro
22:45:34FromDiscord<albassort> Wait do we have nim mixins
22:45:36FromDiscord<albassort> What did I miss
22:45:53FromDiscord<albassort> How many years ago did this happen
22:46:09FromDiscord<Elegantbeef> Nim has had mixins forever
22:46:15FromDiscord<Elegantbeef> It's apart of the generic dispatch model
22:46:20FromDiscord<Elegantbeef> It's not mixin like in java
22:46:22FromDiscord<Elegantbeef> They're generic mixins
22:46:52FromDiscord<albassort> Yeah I have no idea what that means
22:47:28FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/e9C
22:47:33FromDiscord<Elegantbeef> Whoops mispelled echo
22:47:40FromDiscord<albassort> It's a bit annoying that my job is making me do stuff involving Microsoft's SQL wrapper and not letting me write SQL into the database because there is no efficient way to do what I'm trying to do and I don't have authorization to make a new view
22:47:48FromDiscord<albassort> :notlikethis:
22:47:51FromDiscord<Elegantbeef> But anyway you can define generic interfaces using mixins as it takes in instantiation scope into account when calling a generic
22:48:07FromDiscord<albassort> Why Microsoft why
22:48:28*elph_ quit (Quit: Connection closed for inactivity)
22:48:34FromDiscord<Elegantbeef> the antonym of `mixin` is `bind` which only takes the generics declaration scope into account
22:48:49FromDiscord<albassort> sent a code paste, see https://play.nim-lang.org/#ix=4joi
22:48:49FromDiscord<albassort> I might have to read the manual
22:50:25FromDiscord<arkanoid> If I replace \:auto with T, would it work anyway? Auto as param is confusing me
22:50:27FromDiscord<Elegantbeef> Of course
22:50:31FromDiscord<Elegantbeef> Auto is just a shorthand for that
22:50:56FromDiscord<Elegantbeef> It literally just says "for this symbol consider the scope of instantiation and this declaration scope "
22:51:01FromDiscord<Elegantbeef> It's not that complicated
22:51:11FromDiscord<albassort> I understand bind but bind is a bit silly because templates already assume that the variables exist until they're statically evaluated
22:51:11FromDiscord<Elegantbeef> So when you call `myProcedure` it prefers the current scope that you call it in
22:51:24FromDiscord<Elegantbeef> notice how there is no `doThing` before `myProcedure`
22:51:29FromDiscord<Elegantbeef> bind isnt a bit silly
22:51:41FromDiscord<arkanoid> Yeah I do understand now. It smells like dependency injection, or function registration at runtime
22:51:49FromDiscord<Elegantbeef> Bind closes the symbol which forces the procedures tobe used that the generic procedure knows
22:51:52FromDiscord<arkanoid> But at compile time is just better
22:51:54FromDiscord<Elegantbeef> In some APIs this is desired
22:52:00FromDiscord<albassort> Maybe I should use mixins to reduce the insane amount of circular dependencies
22:52:02FromDiscord<Elegantbeef> Cause you do not want to allow the user to overload a procedure
22:52:03FromDiscord<albassort> Or something
22:52:19FromDiscord<albassort> I feel like knowing about this can only make my code worse
22:52:33FromDiscord<Elegantbeef> I mean it's pretty much rust traits
22:52:47FromDiscord<albassort> Yeah and name 6 good thing in Rust
22:52:53FromDiscord<albassort> (edit) "thing" => "things"
22:52:56FromDiscord<Elegantbeef> So if having a generic procedure that can work based off just knowing procedures exists can reduce your cyclical dependancy it might
22:53:00FromDiscord<arkanoid> I'm just considering then for a library that in OOP world would require the user to extend a provided class and overridde/implement some methods
22:53:12FromDiscord<Elegantbeef> Traits, everything is an expression, borrowing....
22:53:15FromDiscord<Elegantbeef> Fuck only got 3
22:53:35FromDiscord<albassort> Now name something that Nim doesn't have
22:53:39FromDiscord<albassort> In those three
22:53:45FromDiscord<Elegantbeef> Traits, borrowing
22:53:51FromDiscord<albassort> We need to add those
22:54:03FromDiscord<Elegantbeef> Nah i actually dont like traits that much
22:54:11FromDiscord<albassort> What exactly is borrowing
22:54:27FromDiscord<arkanoid> I prefer Nim much more than rust. Rust code is not a joy to write, Nim is.
22:54:29FromDiscord<Elegantbeef> Safely aliasing pointers without leaking or dangling
22:54:50FromDiscord<Elegantbeef> Fuck
22:54:51FromDiscord<albassort> I don't know how we could easily do that in them
22:54:55FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/ASc
22:55:05FromDiscord<albassort> In nim
22:55:09FromDiscord<albassort> I wouldn't really feel safe aliasing a pointer
22:55:13FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4joj
22:55:16FromDiscord<Elegantbeef> We have the experimental views
22:55:20FromDiscord<Elegantbeef> So we already have the start
22:55:30FromDiscord<albassort> But if we could alias pointers then c interop would be perfect
22:55:35FromDiscord<Elegantbeef> Views are borrowing
22:55:38FromDiscord<albassort> You probably wouldn't even notice you're using C code
22:55:42FromDiscord<Elegantbeef> What?
22:55:52FromDiscord<albassort> Perhaps my brain has changed what you said into something it isn't
22:55:57FromDiscord<Elegantbeef> How does aliasing a pointer have anything todo with C
22:56:13FromDiscord<Elegantbeef> giving a pointer a new name isnt related to C afaict
22:57:00FromDiscord<albassort> Just make a ref object of the pointer
22:57:03FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/Rc7
22:57:08FromDiscord<Elegantbeef> Lol
22:57:16FromDiscord<Elegantbeef> You cannot raise memory to a reference
22:57:24FromDiscord<albassort> Well I tried
22:57:41FromDiscord<albassort> I like your funny words magic man
22:57:46FromDiscord<Elegantbeef> Anyway arkanoid hope the generic interface pans out
22:57:54FromDiscord<Elegantbeef> I really like generic interfaces, it makes code so much nicer
22:58:50FromDiscord<arkanoid> Just to be sure, are you still talking about mixins?
22:58:58FromDiscord<Elegantbeef> Yes
22:59:08FromDiscord<Elegantbeef> I dont like calling them mixins as that's not very descriptive
22:59:56FromDiscord<Elegantbeef> Especially given the definition of mixins deals with OOP
23:02:29FromDiscord<arkanoid> I'll, try to wrap my head around this. It's an important point to convert my OOP head into procedural and static dispatching, that I like. It's like learning how things used to me done, but with better language
23:03:14FromDiscord<Elegantbeef> Yea you can sort of think of mixins in Nim like interfaces, but that are dispatched statically 😄
23:05:10FromDiscord<PunchCake> hello
23:05:33FromDiscord<PunchCake> what does stop the world mean in this context https://media.discordapp.net/attachments/371759389889003532/1055984537382953000/image.png
23:06:26FromDiscord<leorize> that means all of your code stop running until GC is done
23:07:18FromDiscord<PunchCake> so all the gc's with stop the world equal to No means that my program wont pause at all for GC cycles? thus making it suitable for embedded systems?
23:07:31FromDiscord<Elegantbeef> Correct
23:07:36FromDiscord<Elegantbeef> Arc is really embedded's' best friend
23:07:40FromDiscord<Elegantbeef> Unless you need cycles of course
23:07:54FromDiscord<PunchCake> it sounds too good to be true. whats the downsides
23:08:00FromDiscord<Elegantbeef> Actually sorry it will do GC processing ofcourse
23:08:28FromDiscord<leorize> threading is a bit more complicated
23:08:51FromDiscord<PunchCake> thats it?
23:08:55FromDiscord<PunchCake> im sold
23:09:57FromDiscord<leorize> and you're trading throughput for latency as well
23:10:10FromDiscord<leorize> which happens to be what you want almost all the time
23:10:47FromDiscord<PunchCake> so is it worth it to use the arc GC on embedded instead of manual memory management?
23:10:55FromDiscord<Elegantbeef> Yes
23:10:56FromDiscord<leorize> yea
23:11:06FromDiscord<Elegantbeef> If you're using Nim and use no MM you lose most of Nim's stdlib
23:11:12FromDiscord<PunchCake> huh nim can do it all
23:11:12FromDiscord<leorize> it's indistinguishable to manual for the most part
23:11:24FromDiscord<leorize> the distinction comes when you do threading
23:11:31FromDiscord<PunchCake> i am convinced nim is the only language i'll ever need
23:11:50FromDiscord<PunchCake> In reply to @leorize "the distinction comes when": yeah i see
23:12:36FromDiscord<PunchCake> and is it advised to use the arc GC on regular apps such as gui apps or grpc apis
23:12:50FromDiscord<Elegantbeef> Orc is the suggested GC now
23:12:53FromDiscord<Elegantbeef> Arc for embedded
23:12:58FromDiscord<Elegantbeef> Refc for legacy code
23:13:09FromDiscord<leorize> I'd use orc everywhere tbh
23:13:12FromDiscord<PunchCake> whats the diff between arc and orc?
23:13:19FromDiscord<Elegantbeef> Orc can collect cycles
23:13:27FromDiscord<leorize> just avoid it for hard realtime stuff
23:13:53FromDiscord<PunchCake> so pretty much the arc gc is suitable for everything right
23:14:03FromDiscord<Elegantbeef> Nope arc cannot collect cycles
23:14:18FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/rZF
23:14:20FromDiscord<leorize> but if you're doing hard realtime you would be mixing manual for the select spots where you really need it anyway
23:14:40FromDiscord<Elegantbeef> Hey girvo pretty much doesnt use any allocations
23:14:58FromDiscord<PunchCake> so how arc works it detects memory leaks first then fixes it i dont understand
23:15:19FromDiscord<leorize> arc simply keep a counter with your pointer
23:15:30FromDiscord<leorize> copy adds one
23:15:36FromDiscord<leorize> remove remove one
23:15:44FromDiscord<leorize> then when it reaches zero it's gone
23:15:53FromDiscord<Elegantbeef> It's a scoped based ref counter practically
23:16:10FromDiscord<Elegantbeef> At end of a scope if a reference is dead decrement the count and if the count is 0 call destructor
23:16:18FromDiscord<leorize> same as Rust `Ref` afaict
23:16:18FromDiscord<PunchCake> so let me get this straight↵Embedded -> Arc↵anything else -> orc (and thats the default GC)
23:16:30FromDiscord<Elegantbeef> orc is default in 2.0
23:16:32FromDiscord<leorize> everything -\> orc
23:16:45FromDiscord<Elegantbeef> 2.0 isnt out yet it's soon
23:16:45FromDiscord<leorize> if you can't use it you would know
23:16:57FromDiscord<PunchCake> so whats default now? and is
23:17:00FromDiscord<PunchCake> (edit) removed "and is"
23:17:10FromDiscord<Elegantbeef> Yea just use Orc
23:17:10FromDiscord<Elegantbeef> refc
23:17:32FromDiscord<PunchCake> does orc have much overhead?
23:17:56FromDiscord<leorize> a tiny bit for the cycle collector
23:18:28FromDiscord<leorize> benchmark says it's pretty comparable to manual code iirc
23:18:51FromDiscord<PunchCake> yeah i gotta make a bash function to put all the args for me im too lazy to keep typing -mm:orc every time
23:19:07FromDiscord<leorize> you can use a config file if that fancy you
23:19:07FromDiscord<PunchCake> (edit) "-mm:orc" => "--mm:orc"
23:19:16FromDiscord<PunchCake> bash function sounds easier
23:19:23FromDiscord<PunchCake> its like 20 secs
23:19:57FromDiscord<PunchCake> plus why so many GC's?
23:20:26FromDiscord<leorize> the config is easier than bash \:p
23:20:43FromDiscord<leorize> and you don't have to manage it for every other Nim tool you use
23:20:57FromDiscord<Elegantbeef> Well no one uses anything but refc, orc, arc
23:21:08FromDiscord<Elegantbeef> The other GCs exist due to historical reasons mostly
23:21:20FromDiscord<leorize> go and Boehm are for threading when they were introduced
23:21:29FromDiscord<leorize> I'm pretty sure go was for go interop
23:21:30FromDiscord<PunchCake> so now they are obsolete?
23:21:40FromDiscord<Elegantbeef> Plus it's relatively easy to add new GCs
23:21:43FromDiscord<Elegantbeef> Well allocators
23:21:56FromDiscord<Elegantbeef> Unless you were doing go interop, go's was already
23:22:22FromDiscord<Elegantbeef> Boehm is mostly replaced by orc
23:22:28FromDiscord<PunchCake> why the hell would i choose to interop with go from all of the existing languages
23:22:31FromDiscord<Elegantbeef> Though multithreading with orc/arc is still quite sad for refc
23:22:37FromDiscord<leorize> why not?
23:22:59FromDiscord<Elegantbeef> for refs\ 😄
23:23:22FromDiscord<PunchCake> isnt it much more efficient to interop with rust? at least i can make use of some of the gui libs, maybe even tauri!
23:23:33FromDiscord<leorize> rust didn't exist then
23:23:39FromDiscord<Elegantbeef> Also rust doesnt have GC
23:23:41FromDiscord<PunchCake> so it must be very old
23:23:58FromDiscord<Elegantbeef> The point of using Go's GC is so you can seemlessly send data between the languages
23:24:11FromDiscord<PunchCake> huh
23:24:18FromDiscord<Elegantbeef> With rust interop you just use the FFI and treat it like a C library, it's the best you can really do
23:24:47FromDiscord<PunchCake> yeah
23:25:08FromDiscord<Hoss> is nim the best lang out there for c++ interop?
23:25:13FromDiscord<PunchCake> is there any framework for nim js that helps to make frontend apps
23:25:18FromDiscord<PunchCake> In reply to @Hoss "is nim the best": python is
23:25:23FromDiscord<Elegantbeef> Karax
23:25:32FromDiscord<Elegantbeef> It's pretty much the only language aside from D↵(@Hoss)
23:25:38FromDiscord<Elegantbeef> No carbon doesnt count
23:25:48FromDiscord<leorize> crystal have pretty banger interop too
23:26:02FromDiscord<PunchCake> In reply to @Elegantbeef "Karax": is it similar to python flask?
23:26:19FromDiscord<leorize> they wrapped the entirety of Qt, which is my benchmark for C++ interop
23:26:21FromDiscord<Elegantbeef> No clue
23:27:39FromDiscord<Elegantbeef> Hmph i wonder how they did it cause afaik they dont have any special C++ interop
23:28:43FromDiscord<leorize> they just have people that are really good at C++ ABI
23:29:08FromDiscord<Elegantbeef> https://github.com/Papierkorb/bindgen ah this is the key
23:31:26FromDiscord<PunchCake> interop with java when
23:31:30FromDiscord<Elegantbeef> So it seems to work akin to futhark, where it uses LLVM/CLang?
23:31:38FromDiscord<Elegantbeef> Nim already has it
23:31:40FromDiscord<leorize> see\: jnim
23:31:47FromDiscord<Elegantbeef> Nimpy for python
23:32:00FromDiscord<PunchCake> no way this is real
23:32:26FromDiscord<Elegantbeef> Nimscripter for Nimscript(Yes a scripting subset of nim, dont use it though 😛 )↵nimjl for julia
23:32:35FromDiscord<PunchCake> ah made by the one and only yglukhov
23:32:44FromDiscord<PunchCake> man he is clearly top g
23:33:16FromDiscord<Elegantbeef> Huh, bindgen seems to work how one expects, relies on llvm/clang to get the output one wants
23:33:28FromDiscord<Elegantbeef> Though afaict this means crystal can only use non templated classes and the like
23:34:31FromDiscord<Elegantbeef> I'd joke "Hey it doesnt even work on windows" but that's too low hanging
23:34:36FromDiscord<leorize> that project knows how to instantiate templates for containers, though
23:34:41FromDiscord<PunchCake> plus why is everyone bot here
23:34:46FromDiscord<Elegantbeef> Matrix
23:34:52FromDiscord<PunchCake> whats that
23:35:04FromDiscord<leorize> you see, chatgpt is only a demo of Nim discord
23:35:05FromDiscord<Elegantbeef> A open protocol
23:35:14FromDiscord<PunchCake> its like open source discord?
23:35:21FromDiscord<Elegantbeef> Sorta
23:35:30FromDiscord<Elegantbeef> It's a modern open standard with many open source clients
23:35:43FromDiscord<Elegantbeef> image.png https://media.discordapp.net/attachments/371759389889003532/1055992128376938596/image.png
23:35:43FromDiscord<Elegantbeef> Element is kinda like discord 😛
23:36:02FromDiscord<PunchCake> how much ram does it use
23:36:05FromDiscord<ajusa> In reply to @Elegantbeef "image.png": uh oh, I see the forbidden icon
23:36:05FromDiscord<Elegantbeef> Though cinny is more like discord
23:36:18FromDiscord<Elegantbeef> Meh i'm mostly there to joke around
23:36:38FromDiscord<Elegantbeef> They dont take me seriously there!
23:37:08FromDiscord<Elegantbeef> There are multiple clients so depends↵(@PunchCake)
23:37:20FromDiscord<PunchCake> whats it built with
23:37:25FromDiscord<Elegantbeef> Element is electron based↵Cinny is tauri based
23:37:31FromDiscord<Elegantbeef> There are GTK and QT clients
23:37:38FromDiscord<Elegantbeef> Even TUI clients
23:38:07FromDiscord<Elegantbeef> As a stargate fan i do enjoy 'Tauri' as a name
23:38:25FromDiscord<PunchCake> tauri sucks
23:38:37FromDiscord<PunchCake> i perfer to make gui aps with something like GTK or wxwidgets
23:38:49FromDiscord<PunchCake> at least it wont be a memory hog
23:38:55FromDiscord<Elegantbeef> There are many clients all at different stages of development
23:39:12FromDiscord<PunchCake> wait is there nim bindings for gtk?
23:39:20FromDiscord<leorize> tauri is just electron with another name
23:39:23FromDiscord<Elegantbeef> There is gintro
23:39:32FromDiscord<Elegantbeef> Hey tauri is a little lighter on resources
23:39:34FromDiscord<Elegantbeef> Negligably
23:39:36FromDiscord<Elegantbeef> But hey
23:39:42FromDiscord<PunchCake> In reply to @leorize "tauri is just electron": to be fair its much more light weight
23:39:53FromDiscord<PunchCake> bring back the days where nearly all apps where in C++ or C
23:40:04FromDiscord<PunchCake> no more memory hog hidden webbrowsers
23:40:28FromDiscord<Elegantbeef> This is why matrix is cool imo
23:40:33FromDiscord<leorize> the browsers rose because it was dead easy to write GUI on them
23:40:35FromDiscord<Elegantbeef> You can take those words and go throw together a client 😛
23:41:02FromDiscord<leorize> btw if you like gtk checkout Rust's Relm
23:41:10FromDiscord<PunchCake> In reply to @leorize "the browsers rose because": pretty much html + js even my grandma can make a gui with it
23:41:23FromDiscord<PunchCake> with tailwind as well jeez
23:41:50FromDiscord<leorize> relm brings the elm model to gtk, makes it a lot easier to program
23:42:20FromDiscord<PunchCake> wtf is elm model
23:42:42FromDiscord<leorize> elm-lang.org
23:43:08FromDiscord<PunchCake> oh that
23:43:49FromDiscord<PunchCake> tbh the desktop app scene is pretty much dead
23:44:09FromDiscord<PunchCake> wasm is the future
23:53:46*sagax joined #nim
23:54:36FromDiscord<<She>Horizon</Her>> Disagreed
23:56:24FromDiscord<<She>Horizon</Her>> I very much like WebAssembly, and it does have a place in the web, and even outside of it (some uses I've seen and tried experimenting with is using it for scripting in games, i know Beef did that), but WebAssembly will never be as fast as native code, and it also has larger sizes than if you wrote something in JS instead
23:56:44FromDiscord<Elegantbeef> Larger sizes?
23:56:52FromDiscord<<She>Horizon</Her>> Larger file sizes
23:57:01FromDiscord<Elegantbeef> Wasm is byte code though
23:57:03FromDiscord<<She>Horizon</Her>> Sick rn so not making the most sense
23:58:30FromDiscord<Elegantbeef> Is a sufficiently large program smaller in JS than Wasm?
23:58:36FromDiscord<<She>Horizon</Her>> In reply to @Elegantbeef "Wasm is byte code": I was under the impression that larger projects would end up bigger than equivalent JS code? Didn't do much research myself but was told that was a disadvantage over WASM
23:58:50FromDiscord<Elegantbeef> Well without evidence i dismiss it
23:58:57FromDiscord<Elegantbeef> It's possible, but i dont know if it's true
23:59:13FromDiscord<Elegantbeef> Even minified JS is a lot of data
23:59:13FromDiscord<<She>Horizon</Her>> I'll try looking to see if there's actually any backing to that now
23:59:28FromDiscord<leorize> I mean, wasm from C is smaller than a JS program from C