<< 29-01-2022 >>
00:01:18FromDiscord<Elegantbeef> https://github.com/marcomq/nimview yea there is stuff like this
00:01:55anddamwhoa
00:01:58anddamthat spot on
00:02:05anddamx/that/ a/'s/
00:02:29FromDiscord<Elegantbeef> https://github.com/oskca/webview https://github.com/juancarlospaco/webgui are also stuff to look at
00:02:38FromDiscord<Elegantbeef> webgui relies on gtk though
00:09:35NimEventerNew thread by Dom96: NimForum 2.2.0 - Fixes a big CVE!, see https://forum.nim-lang.org/t/8852
00:21:58*noeontheend joined #nim
00:38:12*rp1 quit (Ping timeout: 250 seconds)
00:38:30*rp1 joined #nim
00:39:21*jjido quit (Quit: My MacBook Air has gone to sleep. ZZZzzz…)
00:44:26FromDiscord<xx_ns> In reply to @NimEventer "New thread by Dom96:": hey i'm mentioned 🙂
00:44:39FromDiscord<xx_ns> feelsgoodman.jpg
00:45:30FromDiscord<ajusa> are there any good dotenv parsers for Nim? https://github.com/euantorano/dotenv.nim seems to require quoting file paths for some reason
00:46:28FromDiscord<Elegantbeef> https://nimble.directory/search?query=env
00:47:04FromDiscord<ajusa> In reply to @Elegantbeef "https://nimble.directory/search?query=env": thank you, maybe I shouldn't have search `dotenv` and `.env` lol
00:48:00FromDiscord<Elegantbeef> You could always fork the otherone if that was your only issue
00:50:02FromDiscord<ajusa> I tried doing a simple change to get it to work and that didn't help, so I gave up. Looks like xmonader's doesn't have that issue, but it doesn't handle quoted strings properly. The envconfig one isn't a proper dotenv reader, it's for Nim configuration only.
00:50:11FromDiscord<ajusa> I'll just put stuff in double quotes lol
00:51:01FromDiscord<xx_ns> sent a code paste, see https://play.nim-lang.org/#ix=3NPi
00:51:42FromDiscord<xx_ns> @ajusa
00:51:45FromDiscord<ajusa> uh
00:52:21FromDiscord<ajusa> I guess clearing the environment variables becomes easier if you read it in? But yeah I see your point
00:52:41FromDiscord<ajusa> (eg it doesn't pollute a terminal with a bunch of vars potentially)
00:53:09FromDiscord<xx_ns> sent a code paste, see https://play.nim-lang.org/#ix=3NPj
00:53:30FromDiscord<xx_ns> once `./mybinary` exits and the subshell exits, the env vars are cleared
00:53:33FromDiscord<ajusa> Thanks! That's a much simpler solution
00:57:33anddamwhat even is https://github.com/marcomq/nimview#minimal-python-example ? installing "nimview" pip package results in a single nimview.pyd: PE32+ executable (DLL) (console) x86-64, for MS Windows into site-packages
00:57:55anddamhow does that cope with Platforms: Windows, Linux, MacOS, Android, Cloud / Web
00:58:04anddam(I am on Linux)
00:59:07FromDiscord<Elegantbeef> Perhaps it's an oversight?
00:59:27FromDiscord<Elegantbeef> I mean why are you even installing the pip package?
01:03:36*noeontheend quit (Ping timeout: 268 seconds)
01:16:14*krux02 quit (Remote host closed the connection)
01:24:49FromDiscord<Patitotective> What is the closest GUI library to sublime text?
01:24:59FromDiscord<Elegantbeef> None
01:25:36FromDiscord<xx_ns> sublime uses a custom UI framework
01:26:02FromDiscord<xx_ns> something based on gtk iirc
01:26:37FromDiscord<xx_ns> so something like https://github.com/StefanSalewski/gintro may be what you're looking for?
01:28:00FromDiscord<Patitotective> In reply to @xx_ns "so something like https://github.com/StefanSalewski": Hm, yea, but it isn't highly customizable like qt
01:28:05FromDiscord<Patitotective> i think
01:29:45FromDiscord<Elegantbeef> Gtk is used for many different desktop environments
01:29:46FromDiscord<Elegantbeef> So it's pretty customizable
01:31:04FromDiscord<Patitotective> In reply to @Elegantbeef "Gtk is used for": hm, you're right, idk why i thought it wasnt
02:19:36*neurocyte0917090 quit (Ping timeout: 256 seconds)
02:52:47FromDiscord<Patitotective> is there any application built with _gintro_?
03:02:13FromDiscord<haolian9 (高浩亮)> nimforum reports `Couldn't retrieve posts.` every time i click the `load more posts` on [this page](https://forum.nim-lang.org/t/7983)
03:14:54*noeontheend joined #nim
03:50:00FromDiscord<JSONBash> In reply to @haolian9 (高浩亮) "nimforum reports `Couldn't retrieve": I get the same
03:52:27FromDiscord<JSONBash> how do you pass an immutable variable as a reference?
03:52:39*arkurious quit (Quit: Leaving)
03:53:57FromDiscord<Elegantbeef> What are you trying to do?
03:55:32FromDiscord<JSONBash> https://play.nim-lang.org/#ix=3NPG
03:57:58FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=3NPH
03:58:39FromDiscord<Elegantbeef> There is no way to make pass it as `var string` without making a `var` copy
04:05:09FromDiscord<JSONBash> so cannot pass an immutable variable as a reference then?
04:07:04FromDiscord<Elegantbeef> well impliclity yes
04:07:14FromDiscord<Elegantbeef> Large types will be passed as reference
04:07:54FromDiscord<JSONBash> is there a size where it does that? and is that automatic?
04:08:06FromDiscord<Elegantbeef> You can also do `type MyType {.byref.} = object`
04:08:12FromDiscord<Elegantbeef> 3 float iirc
04:08:22FromDiscord<Elegantbeef> > = 3 floats will automatically be passed by ref
04:10:33FromDiscord<JSONBash> could I do this then? https://play.nim-lang.org/#ix=3NPL
04:11:07FromDiscord<Elegantbeef> distinct string might work
04:11:11FromDiscord<Elegantbeef> busy now so cant check
04:11:20FromDiscord<JSONBash> something equivalent to 3 floats in size is passed by reference?
04:11:30FromDiscord<Elegantbeef> more than
04:13:41FromDiscord<JSONBash> oh so I shouldn't spend much time worrying about this then because nim does all the heavy lifting for me?
04:13:48*rienske quit (Quit: WeeChat 3.4)
04:13:48FromDiscord<Elegantbeef> Yes
04:14:01FromDiscord<Elegantbeef> It has optimizations for copy vs ref passing
04:14:02FromDiscord<JSONBash> also is this what you meant? https://play.nim-lang.org/#ix=3NPM
04:14:16FromDiscord<JSONBash> In reply to @Elegantbeef "It has optimizations for": nice, that's even better!
04:14:22FromDiscord<Elegantbeef> Yea
04:32:06FromDiscord<noow> um how do I troubleshoot if I have an import, a public symbol from that import and it is not recognizing it
04:32:40FromDiscord<Elegantbeef> What's the compiler say
04:32:42FromDiscord<noow> sent a code paste, see https://paste.rs/nCa
04:33:03FromDiscord<noow> oh wait
04:33:06FromDiscord<noow> I can't read
04:33:12FromDiscord<noow> it's a recursive module dependency
04:33:55FromDiscord<noow> why do my generics not work tho 😭
04:34:24FromDiscord<noow> the previous setup had a different setup and I was calling a function on a generic, and another module had the implementation of that function for that type, yet it couldn't find it
04:34:45FromDiscord<noow> honestly I will just cat all my nim files together, put function definitions on top and call it a day
04:35:12FromDiscord<that_dude> Does anyone have a good place for me too look to understand the difference between exception and defect? They seem to be used interchangeably
04:35:12FromDiscord<Elegantbeef> You could also fix the import order 😜
04:35:20FromDiscord<noow> but it's a recursive dependency
04:35:25FromDiscord<noow> it is meant to be recursive
04:35:33FromDiscord<Elegantbeef> delayed imports allow that
04:35:38FromDiscord<noow> huh?
04:35:39FromDiscord<Elegantbeef> Is the code open?
04:35:44FromDiscord<noow> yes but not pushed
04:35:55FromDiscord<Elegantbeef> Push to a branch and i'll see if i can help
04:36:03FromDiscord<noow> okay gimme a min
04:36:03FromDiscord<Elegantbeef> Will be in a bit though
04:36:08FromDiscord<noow> i appreciate any help
04:38:40FromDiscord<noow> https://github.com/prod2/nondescript just view latest commit on the main branch
04:39:18FromDiscord<noow> fnv1a for NdValue is defined in types/value, while the generic using it is in types/hashtable.nim
04:39:26FromDiscord<noow> and value depends on hashtable through an intermediary
04:40:36FromDiscord<noow> you can `./build.sh` or `nim c --gc:arc -d:danger main`
04:53:27FromDiscord<Elegantbeef> Ok i'll take a look now
04:56:17*noeontheend quit (Ping timeout: 240 seconds)
04:57:12FromDiscord<Elegantbeef> your `findEntry` you need to `mixin` your procedures
04:57:25FromDiscord<noow> mixin?
04:57:30FromDiscord<noow> do I ctrl+f the nim manual?
04:57:34FromDiscord<Elegantbeef> ` mixin fnv1a, equal`
04:57:39FromDiscord<Elegantbeef> It forces the symbols open inside of a generic
04:57:55FromDiscord<Elegantbeef> So it reasons them at instantiation
04:58:15FromDiscord<noow> i'll try that, thank you
04:58:25FromDiscord<Elegantbeef> But there is another issue
05:00:14FromDiscord<noow> i pushed the mixin and a small fix to vm
05:00:31*noeontheend joined #nim
05:01:24FromDiscord<noow> okay it compiles and works
05:01:27FromDiscord<noow> thanks a lot!!!
05:01:30FromDiscord<Elegantbeef> No problem
05:01:41FromDiscord<Elegantbeef> last issue was doing `not x == b` instead of `x != b`
05:01:42FromDiscord<noow> i should read the nim manual from top to bottom sometime 🙂
05:02:10FromDiscord<noow> In reply to @Elegantbeef "last issue was doing": yeah, i just changed a bunch of code at once with no compilations
05:02:19FromDiscord<noow> things like that creep in when i do that
05:02:25FromDiscord<Elegantbeef> Well anywho there you go no big file
05:02:33FromDiscord<noow> 🥲
05:02:46FromDiscord<noow> no need for reinventing C headers
05:03:32FromDiscord<Elegantbeef> Even though some people do importing is fine in most cases
05:03:37FromDiscord<noow> yay nondescript calculates 37th fib now under 6.5 seconds
05:03:47FromDiscord<Elegantbeef> Let's get it faster!
05:03:54FromDiscord<noow> python does it under 9, lua under 3
05:04:35FromDiscord<noow> In reply to @Elegantbeef "Let's get it faster!": yes, I ran out of ideas of optimization just by looking at the code
05:04:38FromDiscord<noow> time to bring out the profiler
05:04:54FromDiscord<Elegantbeef> How do i run the benchmark?
05:05:09FromDiscord<noow> time ./nds benchmarks/fib.nds
05:05:34FromDiscord<noow> after ./build.sh which uses arc and danger
05:05:42FromDiscord<Elegantbeef> Also you know you can use nimble to make a cross platform build task
05:06:04FromDiscord<Elegantbeef> then you could do `nimble buildFast` and it'd output nds
05:06:16FromDiscord<noow> I don't like having a src folder
05:06:32FromDiscord<Elegantbeef> You dont need one
05:06:35FromDiscord<Elegantbeef> Nimble just bitches for no reason
05:07:01FromDiscord<noow> oh? then i will look into it
05:07:24FromDiscord<Elegantbeef> Nico for instance doesnt have a `src` and uses a nimble file with tasks
05:08:33FromDiscord<noow> ah thanks, I will add it to my todo list
05:09:05FromDiscord<noow> might be useful when I start building a test suite
05:13:46FromDiscord<Elegantbeef> No hashing to access it or creating it
05:13:47FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=3NPU
05:14:26FromDiscord<noow> oh, you're right
05:15:03FromDiscord<noow> wow... with all this help, would you like a mention in the commit message, or something
05:18:32FromDiscord<Elegantbeef> @lakmatiol\: https://github.com/nim-lang/Nim/pull/19467 when/if this gets merged you'll be able to use you code on devel
05:20:00FromDiscord<Elegantbeef> I mean you do you i didnt do much, I already have many contributions 😛
05:33:26FromDiscord<noow> sent a code paste, see https://play.nim-lang.org/#ix=3NPV
05:34:02FromDiscord<noow> time for expandArc or something 🤔
05:34:30FromDiscord<Elegantbeef> Perhaps a double free?
05:34:59FromDiscord<Elegantbeef> Ah nevermind i'm dumb
05:35:04FromDiscord<xflywind> try also with `-d:useMalloc`?
05:35:11FromDiscord<Elegantbeef> smoke bomb
05:38:52FromDiscord<noow> sent a code paste, see https://play.nim-lang.org/#ix=3NPW
05:39:05FromDiscord<Elegantbeef> Dont think so it should error
05:39:29FromDiscord<noow> nevermind the type is uint, it's just being cast to ptr
05:39:37FromDiscord<noow> so destroying an uint shouldn't do anything
05:40:16*Gustavo6046 joined #nim
05:40:45FromDiscord<noow> In reply to @flywind "try also with `-d:useMalloc`?": malloc: invalid size
05:40:57FromDiscord<noow> oh no something must be very wrong with my code
05:44:30FromDiscord<Rika> you cant have a destructor for a ptr object afaik
05:44:51FromDiscord<noow> it was a var uint
05:45:16FromDiscord<noow> im just casting heavily cause NaN boxing
05:45:32FromDiscord<Rika> mmmmmmmm thanks FPU
05:46:55FromDiscord<noow> im also munmap_chunk(): invalid pointer
05:47:01FromDiscord<Rika> is the broken code pushed
05:47:31FromDiscord<Rika> oh wow i kinda hate this code xd procs in procs ngl cursed
05:47:33FromDiscord<Rika> anyway
05:47:55FromDiscord<noow> In reply to @Rika "oh wow i kinda": 😦
05:48:00FromDiscord<Rika> is the broken code pushed
05:48:02FromDiscord<noow> yes
05:48:08FromDiscord<Rika> branch? master?
05:48:09FromDiscord<noow> loops are broken specifically
05:48:12FromDiscord<noow> In reply to @Rika "branch? master?": yes
05:48:13*noeontheend quit (Ping timeout: 240 seconds)
05:48:21FromDiscord<noow> noone is relying on my code so 🤷🏻‍♂️
05:49:15FromDiscord<noow> var i = 0; while (i < 3) {print i; i = i + 1;};↵triggers it
05:49:18FromDiscord<Rika> what exam
05:49:18FromDiscord<Rika> okay
05:50:00FromDiscord<Rika> i dont get the error
05:50:07FromDiscord<noow> do ./debug.sh
05:50:07FromDiscord<Rika> wait, ver of nim?
05:50:12FromDiscord<noow> 1.6.2
05:50:21FromDiscord<Rika> neither
05:50:29FromDiscord<noow> -d:useMalloc tho
05:50:32FromDiscord<noow> oh
05:50:33FromDiscord<Rika> im on 1.6.0
05:50:46FromDiscord<Rika> another regression?
05:51:01FromDiscord<Rika> let me check
05:51:02FromDiscord<noow> 1.6.0 is fine
05:51:21FromDiscord<Rika> no issue on 1.6.2 for me
05:51:57FromDiscord<noow> well
05:52:11FromDiscord<noow> 1.6.0 works fine on my system, 1.6.2 doesn't
05:53:01FromDiscord<Rika> neither build or debug break on 1.6.2
05:53:14FromDiscord<Rika> maybe you have a fucked 1.6.2 somehow? that would be extremely starnge
05:53:16FromDiscord<Rika> (edit) "starnge" => "strange"
05:53:25FromDiscord<noow> I'll redownload
05:53:26FromDiscord<Rika> are you using a strange system?
05:53:37FromDiscord<noow> just a normal linux distro
05:53:38FromDiscord<noow> 😭
05:54:01FromDiscord<noow> although
05:54:17FromDiscord<Rika> ?
05:54:27FromDiscord<noow> try running benchmarks/efib.nds
05:54:40FromDiscord<Rika> got it
05:54:42FromDiscord<Rika> got the issue
05:55:22FromDiscord<Rika> same with 1..6.0
05:55:23FromDiscord<Rika> okay
05:55:34*rienske joined #nim
05:57:36FromDiscord<Rika> its a current bug it seems, gonna koch
05:57:42FromDiscord<Rika> actuall
05:57:45FromDiscord<noow> 1.4.8 also breaks
05:57:49FromDiscord<Rika> koch wont help, wonder whats wrong then
05:58:00FromDiscord<Rika> would expand arc help i dunno
05:59:07FromDiscord<noow> yeah im looking into that
06:00:08*rienske quit (Client Quit)
06:00:13FromDiscord<noow> okay
06:00:20FromDiscord<noow> it's an out of bounds indexing
06:01:52FromDiscord<Rika> where?
06:01:59FromDiscord<noow> line 184 in vm
06:02:22FromDiscord<Rika> do you not have bounds checks on ^^;
06:02:40FromDiscord<noow> I reimplemented seq's for optimization, but yes I turned my own bouds checks on
06:03:39FromDiscord<Bung> In reply to @Bung "<@!145405730571288577> `Error: cannot open": @ElegantBeef have you check this ?
06:03:55FromDiscord<Elegantbeef> 1.0.11 should fix it
06:04:22FromDiscord<Rika> use this https://nim-lang.org/docs/system.html#compileOption%2Cstring `compileOption("boundChecks")`
06:04:51FromDiscord<Bung> ok let me try
06:04:56FromDiscord<Elegantbeef> May i ask what optimizations you made to your seq?
06:05:24FromDiscord<noow> instead of defining with a length and a start, I am defining it in terms of pointers to the bottom and top object
06:05:39FromDiscord<noow> and capacity
06:05:41FromDiscord<Rika> idk how thats an optimisation but ok
06:06:04FromDiscord<noow> the improvements in performance could be just a different growing rate
06:06:20FromDiscord<Elegantbeef> You can implement your own growth rate using `newSeqWithCap` 😛
06:06:25FromDiscord<Rika> not really
06:06:30FromDiscord<Rika> itll grow at the same rate
06:06:35FromDiscord<Rika> just a different cap at the start
06:06:47FromDiscord<noow> In reply to @Rika "idk how thats an": also, I can drop the last n objects by decrementing the top pointer
06:06:57*rienske joined #nim
06:07:05FromDiscord<Rika> same thing with a length? just reduce the length
06:07:09FromDiscord<Elegantbeef> You can drop the last N objects by just decrementing `len` 😀
06:07:28FromDiscord<Rika> not sure if the bridge was slow this time
06:07:35FromDiscord<Elegantbeef> Nah i was
06:07:55FromDiscord<noow> well, that little math of adding len or just having a top directly probably doesn't make that much of a difference true
06:08:15FromDiscord<Rika> yeah most of the time is spent on the alloc and copies
06:08:41FromDiscord<Rika> so yeah newSeqWithCap would help if you knew the init.size or the max size
06:08:48FromDiscord<noow> I just straight out double the length everytime I run out and don't shrink right now
06:09:00FromDiscord<Elegantbeef> Nim seq's dont shrink eeither
06:09:34FromDiscord<Elegantbeef> seq's also have a doubling growth rate iirc until a point where they do a larger growth
06:09:58FromDiscord<Rika> you should be fine as long as you do a good/not stupid initial capacity
06:10:28FromDiscord<noow> well, sometihng improved the performance 🤷🏻‍♂️ or i just suck at benchmarking
06:10:54FromDiscord<Elegantbeef> Profiling is a gift
06:10:55FromDiscord<Rika> could be both
06:11:01FromDiscord<noow> probably the drop-n tho
06:11:09FromDiscord<Bung> @ElegantBeef it seems it does not add current project src to search paths
06:11:14FromDiscord<noow> since I was for countup discard popping
06:11:31FromDiscord<Elegantbeef> of course not↵(@Bung)
06:11:33FromDiscord<noow> i didn't find a easy way to drop the last n with seqs
06:11:58FromDiscord<Elegantbeef> `yourSeq.pop`
06:12:14FromDiscord<Elegantbeef> `yourSeq.setLen(yourSeq.len - n)`
06:12:16FromDiscord<Rika> well yeah thats one good point, afaik nim doesnt just dec. the length when the del is at the end
06:12:19FromDiscord<Rika> ah yeah setlen
06:12:20FromDiscord<Rika> lmao
06:12:28FromDiscord<noow> oh
06:13:22FromDiscord<Elegantbeef> Remember bung Nimscripter is a general purpose interop tool, it doesnt make many assumptions about where nim files should be
06:14:03FromDiscord<Elegantbeef> If you dont give it a `NimScriptPath` it doesnt have 0 inkling any folder other than `stdlib` is in the path
06:15:36FromDiscord<Elegantbeef> So what i'm saying is supply `getCurrentDir()` into the search paths
06:23:45FromDiscord<noow> nevermind xD
06:23:49FromDiscord<noow> it was a stupid typo
06:23:53FromDiscord<Rika> Lol
06:23:56FromDiscord<noow> i mean i did a pop instead of peek
06:23:57FromDiscord<Elegantbeef> Dont do that
06:24:01FromDiscord<Rika> Lol
06:24:06FromDiscord<noow> this happens when it's 7 am and i haven't slept all night
06:24:27FromDiscord<Bung> oh , I get new error that without idea
06:24:40FromDiscord<Bung> sent a code paste, see https://play.nim-lang.org/#ix=3NQ8
06:26:10FromDiscord<Elegantbeef> code you're using write's a stack trace
06:26:24FromDiscord<noow> @Rika @ElegantBeef well thanks for all the willingness to help, i hope you forgive a stupid thing or two i've done today
06:26:49FromDiscord<Elegantbeef> Hey i help people i dont care how many stupid things they do aslong as they're sincerely trying 😛
06:27:55FromDiscord<noow> i wonder if languages with string interning GC strings 🤔
06:29:19FromDiscord<Elegantbeef> I dont know what you just said
06:30:26FromDiscord<noow> im just crying out loud because i have no idea how I will implement a GC
06:31:27FromDiscord<Elegantbeef> mark and sweep isnt too difficult afaik
06:32:04FromDiscord<noow> yeah i'll figure it out somehow
06:32:25FromDiscord<noow> i will probably have to keep a reference to every single object on the heap though
06:33:22FromDiscord<noow> or keep track of all allocated memory and check which I can access from the root
06:33:43FromDiscord<noow> well i'll go sleep now xD thank a lot again
06:33:51FromDiscord<Elegantbeef> No problem
06:34:33FromDiscord<Rika> I was doing something
06:34:42FromDiscord<Rika> String interning is strange to think about I'd say
06:35:36FromDiscord<Bung> do know which part use writeStackTrace, my code does not use it directly
06:36:12*nrds joined #nim
06:36:31FromDiscord<Elegantbeef> I havent got a way to figure it out yet
06:36:44FromDiscord<Elegantbeef> but `fileIndex` should let us get the file
06:37:20FromDiscord<Elegantbeef> https://github.com/beef331/nimscripter/blob/master/src/nimscripter.nim#L28-L31 feel free to figure it out if you have the time/effort
06:37:21FromDiscord<Bung> ` (8, 6): 'writeChars' [proc declared in /Users/bung/.choosenim/toolchains/nim-1.6.2/lib/system/io.nim(214, 6)]` you mean this hint , I also checked
06:38:08FromDiscord<Elegantbeef> No i mean we have the file index from `info` and we should be able to get what file, but I havent got around to it
06:38:42FromDiscord<Bung> okay , I'll try get some info from there
06:43:30*jmdaemon quit (Ping timeout: 250 seconds)
06:44:26*jmdaemon joined #nim
06:47:34FromDiscord<that_dude> Sorry to ask again, but what is the difference between exception and defect, and which one should I use. They seem to be used interchangably
06:47:47FromDiscord<Elegantbeef> Defects are for logic errors and not meant to be caught
06:47:57FromDiscord<Elegantbeef> Exceptions are exceptional code which can be caught
06:48:34FromDiscord<that_dude> So I would use defect if I did a wrong conversion in code, but exception if I couldn't load the needed data from a website?
06:49:41FromDiscord<that_dude> or defect is "kill through error" and exception is "give another chance through error"
06:49:52FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=3NQa
06:52:18FromDiscord<that_dude> ok I think I got it
06:52:45FromDiscord<Elegantbeef> You shouldnt often use defects unless it's truely unhandleable
06:53:52FromDiscord<that_dude> Fair enough
06:57:23FromDiscord<Bung> @ElegantBeef https://github.com/beef331/nimscripter/pull/10
06:57:51FromDiscord<Elegantbeef> Can you bump to 1.0.12?
06:59:06FromDiscord<Bung> sure
06:59:54FromDiscord<Bung> done
07:00:31FromDiscord<Elegantbeef> There you go, thanks!
07:03:05*nrds quit (Remote host closed the connection)
07:03:08FromDiscord<Bung> you can still use https://github.com/bung87/crown_ui/tree/nimscripter for reproduce , it report nimyaml use `writeStackTrace`
07:03:37*nrds joined #nim
07:03:38FromDiscord<Elegantbeef> Yea I cant do anything here, dont add files you cant use in nimscript
07:05:13FromDiscord<Bung> is there simple rule I can no which lib I cant use with nimscript ?
07:05:54FromDiscord<Elegantbeef> Well in this case the issue is a lacking flag
07:06:26FromDiscord<Elegantbeef> So perhaps i need to expose yet another parameter
07:06:29FromDiscord<Elegantbeef> Or make an object
07:07:02FromDiscord<Elegantbeef> https://github.com/flyx/NimYAML/blob/a72966374fe9c88e0974036131a36a8b4201e20d/yaml/private/internal.nim#L29-L37 i think the issue here
07:09:05FromDiscord<Bung> in there which code cant be used with nimscript ?
07:10:03FromDiscord<Elegantbeef> Any code that relies on low level code will not work
07:10:28FromDiscord<Elegantbeef> Many things will work if they dont rely on importC or similar
07:12:46FromDiscord<Bung> but here I see use nim builtin procs
07:15:32FromDiscord<Elegantbeef> `when defined(nativeStacktrace) and nativeStackTraceSupported:`
07:15:40FromDiscord<Elegantbeef> NimVm doesnt have stacktrace
07:16:39FromDiscord<Bung> okay, thanks for explanation
07:41:11euantorano@ajusa I’m the author of the dotenv library you mentioned earlier. I’m not sure I understand what you mean by “it seems to require quoting file paths for some reason”, but I’m happy to try and fix any issues you’re having. Can you please open an issue on GitHub?
07:45:29*rienske quit (Quit: WeeChat 3.4)
07:45:48*rienske joined #nim
07:46:21FromDiscord<Elegantbeef> I believe they're talking about `MyVar=/somePath/`
07:46:33FromDiscord<Elegantbeef> do you not make it so you have to do \`MyVar="/somePath/"\~
08:01:52*jmdaemon quit (Ping timeout: 256 seconds)
08:05:10*jmdaemon joined #nim
08:23:23*jjido joined #nim
08:46:38*jmdaemon quit (Ping timeout: 256 seconds)
08:53:13*jjido quit (Quit: My MacBook Air has gone to sleep. ZZZzzz…)
09:00:10*jjido joined #nim
09:15:23haakonIs there any difference between separating proc parameters with comma and semicolon? Which should I prefer?
09:17:43FromDiscord<Illusive> Is there a good text editor on Nim?
09:18:06FromDiscord<BhamidipatiNikhil> In reply to @Illusive "Is there a good": VScode
09:19:47arkanoidhaakon, comma. Semicolon is used for special cases, when compiler can't find by itself what you want
09:20:31haakonarkanoid: God it, thanks
09:20:38haakongot*
09:24:31arkanoidI have kinda naive question. I see code calling dealloc(data) when data is "UncheckedArray[T]". Con can free memory if array size is not there?
09:25:25arkanoidFor example in myseq example here: https://nim-lang.org/docs/destructors.html
09:28:26arkanoidDoes nim keeps the size of each allocated heap size for me?
09:30:52FromDiscord<Phil> ... I think I'm missing something regarding reversing a string
09:33:46FromDiscord<Phil> Never mind, I kept using "let" strings with reverse
09:35:14FromDiscord<Phil> now how to turn a seq[char] into a string...
09:35:34FromDiscord<Phil> (edit) "string..." => "string in a simple manner that isn't me writing a loop on that array manually..."
09:40:10FromDiscord<Rika> arkanoid: i would recommend semicolon over comma once you're an "advanced user" slash you're using `using`
09:40:31*jjido quit (Quit: My MacBook Air has gone to sleep. ZZZzzz…)
09:41:24FromDiscord<Rika> phil: choose, loop or memory copy or unsafe cast 😛
09:47:05FromDiscord<mratsim> In reply to @arkanoid "Does nim keeps the": That's the role of an allocator yes, malloc/free are the same
09:50:35FromDiscord<Phil> In reply to @Rika "phil: choose, loop or": Loop it is for I am afraid of the unsafe
09:50:46FromDiscord<Phil> Namely of having to debug stuff I don't comprehend that well
09:54:06FromDiscord<Phil> ... man, that moment when you realize you can use nim aliases on 10 types and just use them by defining a proc for every single type in the alias
09:54:13FromDiscord<Phil> (edit) "a" => "an overloaded"
09:54:34FromDiscord<Rika> ?
09:57:28FromDiscord<Phil> So I have a full text search using sqlite's fts5 table
09:57:45FromDiscord<Phil> And the contents of that search table come from 10 other tables, for all of which I have model-types (using norm)
09:58:17FromDiscord<Phil> Now I still need procs that handle inserting and updating entries to and from that searchTable
09:58:40FromDiscord<Phil> And I'm not going to write 10 procs per functionality I need called "insertCharacter", "insertCreature", "insertblablabla"
10:01:46FromDiscord<Phil> sent a long message, see https://paste.rs/f4i
10:02:28FromDiscord<Phil> I maybe should add that the way to build the text and title that should go into the search table are different for every of the 10 tables
10:02:40FromDiscord<Rika> i mean if they all share mainly the same proc body that works yes
10:02:53FromDiscord<Phil> Anyway, I just realized I could do this as I was halfway of writing the very convoluted way
10:03:01FromDiscord<Phil> When Aliases and how useful they are hit me in the head
10:03:38FromDiscord<Phil> the proc body of "insertSearchEntry" is always the same, only the way of how you get the text for body and title of an entry differs
10:28:32FromDiscord<Phil> Is there a way for me to ascertain at compile time that all types of an alias have a proc with a given signature?
10:31:02FromDiscord<Rika> With concepts or compiles()
10:31:04FromDiscord<Rika> I guess
10:31:32FromDiscord<Phil> concepts are slowly but surely reaching the point of "most awaited feature for nim stable"
10:31:50FromDiscord<Phil> (edit) ""most" => "being my most" | "stable"" => "stable"
10:31:59FromDiscord<vindaar> yeah, it sounds like you want concepts. But I don't understand what you mean by "all types of an alias"
10:32:50FromDiscord<Rika> All types of a type class
10:33:14FromDiscord<vindaar> sent a code paste, see https://play.nim-lang.org/#ix=3NQW
10:34:07FromDiscord<Phil> sent a code paste, see https://paste.rs/xlY
10:34:10FromDiscord<vindaar> ah. But in that case why place them inside the type class if they don't fulfill the requirement already?↵(@Rika)
10:34:37FromDiscord<vindaar> so this is for users of some library?
10:34:46FromDiscord<Rika> Concepts is the solution pretty much…
10:35:06FromDiscord<Phil> This is for myself, see the debate above about me having a search table with content of 10 other tables.↵Yeah, concepts seem like the way to go here
10:35:27FromDiscord<vindaar> well, it's a partial solution for this. Because it will only cause type `B` to fail resolving some other proc that requires the type `MyConcept`. Nothing more
10:36:36FromDiscord<Rika> You can put outside the proc “if not compiles(…): {.error:….}
10:37:46FromDiscord<Phil> That might actually be nicer in the meantime as, while I'm eager to work with concepts, I'm hesitant to go for an experimental feature if the implementation might change on me while I'm in the middle of writing my project
10:38:06FromDiscord<Rika> Compiles is just as unstable as concepts I would say.
10:38:22FromDiscord<Phil> It is? dangit!
10:38:25FromDiscord<vindaar> I did read what you wrote above now. Still not quite following. In the snippet with `MyAB` you posted. Just having the code written like this already causes a CT error for `getNum` if you hand a `B`
10:38:54FromDiscord<Rika> But only if you hand a B
10:38:59FromDiscord<vindaar> concepts aren't actually that unstable. They can be a bit annoying if they don't match what you think / match too much. Or if you need recursive concepts
10:39:03FromDiscord<Rika> I assume they want it to error even if not
10:39:14FromDiscord<Phil> Pretty much
10:39:20FromDiscord<Phil> I want the guarantee at compile time that the procs are there
10:39:28FromDiscord<Phil> Not an error at runtime once the passing to B happens
10:39:36FromDiscord<Rika> You could just write tests 😛
10:40:07FromDiscord<Phil> ....but....but
10:40:20FromDiscord<vindaar> that cannot work however. At least not with concepts. A concept doesn't "know" anything about all other concepts that match. It's only a way to restrict generics. That doesn't apply here. Your generic is perfectly restricted using the `MyAB` alias↵(@Rika)
10:41:10FromDiscord<vindaar> if you really want that, you'd have to write a macro that iterates the type alias and checks if the proc exists with some hack (i.e. either by calling it with a dummy or emitting `compiles` statements for each type)
10:41:16FromDiscord<Rika> In reply to @vindaar "that cannot work however.": Concepts aren’t a match in this case yes, but what I said (compiles) should be with the right predicate
10:44:09FromDiscord<vindaar> and at the same time I don't see how `getNum` with the type alias will ever match a `getNum` call for any explicit type in the alias. The explicit version of the proc will always be called instead I think
10:44:56FromDiscord<Rika> I think it’s just an example here
10:45:00FromDiscord<vindaar> of course, the other `getNum` can be something like `getNumImpl` or whatever
10:45:09FromDiscord<Phil> Ah, maybe bad example, the getNum(t: MyAB) could also be called "getANumber(t: MyAB)" for my purposes
10:45:15FromDiscord<Rika> Actually that’s kinda brilliant
10:45:28FromDiscord<Phil> Ohh I like the way of impl
10:45:30FromDiscord<Rika> If you want a proc to be guaranteed for a type in a type class
10:45:39FromDiscord<Rika> You make a proc using the type class that’s an error
10:45:56FromDiscord<Phil> I don't think I understand
10:46:08FromDiscord<Rika> Actually no
10:46:18FromDiscord<Rika> That’s still compile time error only on usage
10:46:26FromDiscord<Rika> Dude just write tests dammit hahaha
10:46:33FromDiscord<Phil> But lazyness!
10:48:12FromDiscord<vindaar> ohh, that's actually pretty smart if you want a custom error message↵(@Rika)
10:48:25FromDiscord<vindaar> but still only works on usage
10:48:46FromDiscord<vindaar> but then again, if you never need the missing proc, arguably it's not a problem if your program doesn't error at CT
10:51:27FromDiscord<Phil> ~~I hate it when people notice that I'm trying to use the compiler as my todo list reminder~~
10:51:33FromDiscord<Phil> (edit) "~~I hate it when people notice that I'm trying to use the compiler as my todo list reminder~~ ... " added "😄"
10:52:10FromDiscord<Rika> No I get it I don’t like it either when CTEs I make only trigger on usage
10:58:31FromDiscord<Phil> Hmmm
10:58:36FromDiscord<Phil> I could use some advise organization wise
10:58:45FromDiscord<Rika> Write tests
10:58:48FromDiscord<Rika> Lol
10:58:54FromDiscord<Phil> Dang it Rika!
11:00:17FromDiscord<Phil> I organize my application by domain. ↵So I've got src/applications and inside that I've got on dir for each creature, character, location, item etc.↵Now I have one that is "search", which makes use of stuff in creature/character/location/item etc., mainly the knowledge on how to turn an instance of a character/bla into a string
11:00:44FromDiscord<Rika> What’s the issue?
11:00:48FromDiscord<Phil> I'm just contemplating whether "getSearchBody(article: Character)" should be located in the "character" folder or the "search" folder
11:00:49FromDiscord<vindaar> sent a code paste, see https://play.nim-lang.org/#ix=3NR5
11:00:56FromDiscord<Rika> In reply to @Isofruit "I'm just contemplating whether": Search
11:01:07FromDiscord<Rika> It’s not inherent to characters it’s inherent to searching
11:01:30FromDiscord<Phil> Makes sense to me, thanks!
11:02:24FromDiscord<Phil> In reply to @vindaar "well, you asked for": This is the type of thing that makes me scared of macros 😄
11:03:12FromDiscord<vindaar> if not for the `when compiles` in there, I'd almost consider it usable. But the purpose doesn't warrant it imo
11:04:23FromDiscord<Phil> Thanks for the proof of concept though! If I'm interpreting this correctly the "explode" macro turns into code that tries to do a call for every combination of a type in MyAB with "getNumImpl" and errors out if it's not there, leaving no code behind even if it compiles
11:04:27FromDiscord<Phil> Am I reading that correctly?
11:05:55FromDiscord<vindaar> exactly↵(@Phil)
11:06:15FromDiscord<vindaar> ah wait, not quite
11:06:29FromDiscord<vindaar> nvm
11:06:51FromDiscord<vindaar> sorry 😅 I was about to say it's not quite right it doesn't leave anything behind, but I was confused for a second
11:30:48*jjido joined #nim
11:46:49haakonIs it not possible to foldl over HashSets? `@[1,2,3].toHashSet().foldl(a + b)` gives me "template/generic instantiation of `foldl` from here"
11:47:21FromDiscord<Rika> How do you iterate over a hash set
11:47:29FromDiscord<Rika> There's no order
11:49:04haakonThe order would be arbitrary. You can iterate over a HashSet with `for`
11:49:48FromDiscord<Rika> In reply to @haakon "Is it not possible": What is the whole error
11:49:53FromDiscord<Rika> That is only one line of many
11:50:45FromDiscord<enthus1ast> well there is https://nim-lang.org/docs/sets.html#OrderedSet
11:51:16FromDiscord<Bung> @ElegantBeef do you have idea how that could be fixed ?
11:51:24haakonThe whole error is https://paste.rs/VEI
11:51:48haakon(from the inim repl)
11:52:11FromDiscord<vindaar> most (all?) procedures / templates in `sequtils` assume your input data is a sequence↵(<@709044657232936960_haakon=5b=49=52=43=5d>)
11:52:33FromDiscord<Rika> In reply to @vindaar "most (all?) procedures /": I think it's more "sequence like"
11:52:58haakonvindaar: yeah I guess that's the problem. Folding over sets like this is simply not supported
11:53:49FromDiscord<vindaar> that's what it should really be, but the implementations aren't exactly very generic↵(@Rika)
11:54:30FromDiscord<vindaar> you can use a `toSeq` in there though. Of course means you have an additional iteration over your set. If that's fine `theSet.toSeq.foldl(a + b)` works though↵(<@709044657232936960_haakon=5b=49=52=43=5d>)
11:54:51haakonYeah that's what I'm falling back on :-)
12:01:21haakonBut toSeq is failing in the same way? `@[1,2,3].toHashSet().toSeq()` -> "template/generic instantiation of `toSeq` from here"
12:03:26FromDiscord<Rika> Sorry but again we need more context
12:04:35haakonAh right. Here: "type mismatch: got 'int' for 'x`gensym35' but expected 'HashSet[system.int]"
12:05:46FromDiscord<Phil> I keep forgetting... but does nim have any remotely close equivalent to elvis operator?
12:11:21haakonEh, rewrote it in a less functional way, works
12:11:26FromDiscord<Rika> No, nim does not have truthiness
12:12:10FromDiscord<Rika> Unless you only want an elvis operator for booleans, in which that's just "or"
12:17:07emeryhas anyone considered porting nim to solo5?
12:30:27FromDiscord<Rika> Tobwhatv
12:30:30FromDiscord<Rika> To what
12:35:27FromDiscord<Phil> In reply to @Rika "Unless you only want": Pretty much, I often used to do stuff like : ↵`let bla = procThatReturnsBool() ? valueA : valueB`
12:35:51FromDiscord<Phil> Or rather the equivalent of that in JS/Python
12:37:50FromDiscord<Rika> That’s not Elvis
12:37:55FromDiscord<Rika> That’s ternary
12:38:20FromDiscord<konsumlamm> you can just use an if expression theb
12:38:21FromDiscord<Rika> Ternary is inline if statement in Nim
12:38:24FromDiscord<konsumlamm> (edit) "theb" => "then"
12:38:31FromDiscord<Rika> Statement? Expression
12:46:12*mahlon quit (Ping timeout: 250 seconds)
13:32:30*lumo_e joined #nim
13:53:55arkanoidmratsim: (late) thanks
14:06:42*pro joined #nim
14:14:34*PMunch joined #nim
14:15:29FromDiscord<BhamidipatiNikhil> Guys, i am stuck... Any idea why the 30th line is showing indentation error? https://media.discordapp.net/attachments/371759389889003532/936987919925796915/unknown.png
14:17:19*Gustavo6046 quit (Quit: Leaving)
14:33:13PMunch@BhamidipatiNikhil, I can't see the image Discord sent..
14:44:37*lumo_e quit (Ping timeout: 240 seconds)
14:51:45*arkurious joined #nim
14:52:42FromDiscord<BhamidipatiNikhil> I don't want to read the whole line as input... just the first element i want as input and ignore the rest.... How should i go about it?
14:53:27FromDiscord<Bung> http://debuggable.com/posts/parsing-file-uploads-at-500-mb-s-with-node-js:4c03862e-351c-4faa-bb67-4365cbdd56cb it seems this is right solution of a http multipart parser ?!
14:57:57PMunch@BhamidipatiNikhil, just read until you're happy and stop?
15:18:14FromDiscord<BhamidipatiNikhil> sent a long message, see http://ix.io/3NSs
15:18:43FromDiscord<BhamidipatiNikhil> i cannot control how i give input, because i am attempting a problem in an onlinejudge...
15:18:52*pro left #nim (WeeChat 3.4)
15:19:58PMunchRead the full line, then use "align" from strutils with "0 " as your pattern perhaps?
15:20:27PMunchAh, it only takes a character, not a string
15:29:04FromDiscord<tandy> is there a way to get the first digit of an int without casting?
15:29:16FromDiscord<tandy> i want to check whether http error codes start with `2`
15:29:22FromDiscord<tandy> [Edit](https://discord.com/channels/371759389889003530/371759389889003532/937006441187917874): is there a way to get the first digit of an int without casting to string?
15:29:33Zevvtandy: waht about div 100
15:29:36FromDiscord<tandy> i guess i can just do range check
15:30:11Zevvright, that's nicest. `val in 200..299`
15:30:13FromDiscord<Rika> In reply to @Zevv "<@878733865919844433>: waht about div": you mean div
15:30:17FromDiscord<Rika> oh wait
15:30:19FromDiscord<Rika> what
15:30:24FromDiscord<Rika> why did i read mod somehow?
15:30:31FromDiscord<Rika> i am going crazy
15:30:35Zevvdiv, mod, es ist mir alles egal
15:30:36FromDiscord<Rika> this is it guys
15:31:18Zevvremember, life is just a simulation. You're running on a cpu that does div and mod in one instruction. easy to confuse these kind of things
15:31:34FromDiscord<Rika> @_@
15:34:16FromDiscord<tandy> oh thats clean, didnt know u could do that↵(<@709044657232936960_=5aevv=5b=49=52=43=5d>)
15:34:18FromDiscord<tandy> ty
15:35:06Zevvtechnically, there's nothing wrong with a >= 200 and a <= 299, of course. that's what the range check will boil down to udner the hood anyway
15:52:56FromDiscord<Patitotective> Idea: `..` first parameter should be 0 as default, so you can write `..10`
15:57:37FromDiscord<haxscramper> In reply to @Patitotective "Idea: `..` first parameter": this was implemented but then deprecated
15:57:42FromDiscord<Patitotective> why?
15:57:53FromDiscord<Patitotective> (edit) "why? ... " added "😕"
16:13:00FromDiscord<haxscramper> https://github.com/nim-lang/RFCs/issues/311
16:14:08*noeontheend joined #nim
16:15:43FromDiscord<Patitotective> :o
16:21:25FromDiscord<Patitotective> Is it possible to build a free-dependencies _gintro_ application? Like, self-contain all the dependencies required
16:29:03FromDiscord<auxym> In reply to @Patitotective "why? 😕": you lose a explicitness to save a single character
16:29:14FromDiscord<auxym> would be my guess
16:32:43FromDiscord<Patitotective> yea, i just realize is not clear what it means
16:32:51FromDiscord<Patitotective> (edit) removed "i just realize"
16:33:46*jmdaemon joined #nim
16:48:58anddamElegantbeef: still there?
16:49:12anddamanyway I checked out the python binding since it's what I am most accustomed to
16:49:30*PMunch quit (Quit: leaving)
16:49:32anddamand the readme suggested it could be used straight away
16:50:18anddamdifferent topic, I have 1.6.0 installed on this computer, is there a nimble way to upgrade to 1.6.2 or should I just paint over it with choosenim (like I did in first place to install nim)?
17:27:39anddamoh choosenim has me served already
17:29:22FromDiscord<Phil> sent a long message, see http://ix.io/3NU1
17:30:10FromDiscord<Phil> (edit) "http://ix.io/3NU1" => "http://ix.io/3NU2"
17:30:45FromDiscord<Phil> (edit) "http://ix.io/3NU2" => "http://ix.io/3NU3"
17:40:50anddamI am possibly doing something really wrong, I installed webgui package, then a simple "import webgui" in test.nim and `nim r test.nim` results in "/webgui-git/examples/test.nim(1, 8) Error: cannot open file: webgui"
17:41:54anddamthe example.nim at https://github.com/marcomq/nimview/tree/main/examples has "import ../src/webgui" but isn't the point of installing a package just that you get the package namespace regardless of the source wd?
18:02:10FromDiscord<ajusa> In reply to @Isofruit "Wait, I'm designing myself": That's a wall of text but have you tried includes?
18:02:32FromDiscord<ajusa> Like have one models.nim file that just includes all the other files that you've split it into
18:03:08FromDiscord<ajusa> Unless you have mutually dependent types, then I think the compiler forces you to declare those in the same type section
18:08:20FromDiscord<Goel> I though the primitive type `int` was `int32` on x86 archit and `int64` in x64 archit, but it's not like that? Or maybe it takes that size but it remain of type `int`
18:09:01FromDiscord<Phil> In reply to @ajusa "That's a wall of": That would sort of work for models, as I tend to intentionally do multiple model-specific sub-models (e.g. Encounter needs a Diaryentry model, so they get a custom EncounterDiaryEntry Model which has only a subset of the data)
18:09:45FromDiscord<Phil> That does still leave the question open of the repositories though, where encounterRepository has a proc that depends on diaryEntryRepository, while diaryEntryRepository has its own proc that depends on a separate proc on encounterRepository
18:10:02FromDiscord<Goel> (edit) "I though the primitive type `int` was `int32` on x86 archit and `int64` in x64 archit, but it's not like that? Or maybe it takes that size but it remain of type `int`" => "sent a code paste, see https://play.nim-lang.org/#ix=3NUr"
18:10:39FromDiscord<Phil> Or wait, would "include" workon that to? I need to read up on that keyword
18:11:17*k0mpjut0r quit (Ping timeout: 252 seconds)
18:11:20FromDiscord<Phil> Oh wow, yeah that might work, if it "copy pastes" the contents of one file into another file
18:11:29FromDiscord<Phil> I could have one mega repository file and a mega model file
18:12:32FromDiscord<konsumlamm> In reply to @Goel "I though the primitive": it's not about x86 vs x64 (those are two specific architectures), it's about 32 bit vs 64 bit architectures
18:12:56FromDiscord<konsumlamm> but the answer is that `int` and `int32`/`int64` have the same size, but they're not literally the same type
18:13:25FromDiscord<konsumlamm> whereas `float` is always an alias for `float64`
18:13:26FromDiscord<Goel> Wait what, why int32 and int64 have the same size?
18:13:34FromDiscord<konsumlamm> they don't
18:14:07FromDiscord<konsumlamm> (`int` and `int32`) or (`int` and `int64`)
18:16:00FromDiscord<Phil> Wait, that wouldn't give me access in encounterRepository.nim to diaryEntryRepository.nim, hmmm
18:16:08FromDiscord<Phil> (edit) "diaryEntryRepository.nim," => "diaryEntryRepository.nim and vice versa,"
18:21:14*jmdaemon quit (Ping timeout: 256 seconds)
18:23:15*mahlon joined #nim
18:32:36FromDiscord<Phil> All SO gives me is "rethink your problem domain and make repositories for that", gna
18:35:40*jmdaemon joined #nim
18:41:15FromDiscord<gcao> Today I run into this fork of Nim called Nimskull (https://github.com/nim-works/nimskull). does anyone know anything about it?
18:41:58FromDiscord<haxscramper> Aside from plans in the readme etc.?
18:42:10FromDiscord<Phil> Ah, hax is here, I can save myself the typing
18:42:49FromDiscord<haxscramper> If you want the short rant-free version then it is a fork of nim for now
18:43:19FromDiscord<Phil> Wasn't the key differentiator your focus would be more on stability and less buggs?
18:43:52FromDiscord<haxscramper> yes, that hasn't changed
18:44:07FromDiscord<Phil> Check, just wanted to see whether I recalled that correctly
18:44:36FromDiscord<haxscramper> well, right now the focus is on the massive refactoring of the compiler impl. because it is barely maintainable
18:45:37FromDiscord<haxscramper> with some intermediate steps like minor https://github.com/nim-works/nimskull/pull/214 and major https://github.com/orgs/nim-works/projects/4 tooling improvement
18:47:23FromDiscord<lpha3ch0> sent a code paste, see https://play.nim-lang.org/#ix=3NUO
18:48:16FromDiscord<haxscramper> `&` can be an overloaded prefix oeprator
18:48:22FromDiscord<haxscramper> what is the full context of this code?
18:48:38FromDiscord<lpha3ch0> In reply to @haxscramper "what is the full": https://gist.github.com/khchen/5cd03eab742517fa20ccc688e6b1a1a6
18:50:46FromDiscord<haxscramper> https://github.com/khchen/winim/blob/master/winim/utils.nim#L43 someone can't use `addr`
18:50:50FromDiscord<haxscramper> (edit) "https://github.com/khchen/winim/blob/master/winim/utils.nim#L43 someone can't use `addr` ... " added "I guess"
18:51:02FromDiscord<haxscramper> this is a C-like addr version
18:51:56FromDiscord<Rika> Yikes
18:52:38FromDiscord<lpha3ch0> That's what it looked like to me but I was still confused because I was aware of it's application to concatenate strings and saw it was overloaded but couldn't find any references to how it was used other than concatenation.
18:52:43FromDiscord<haxscramper> In reply to @haxscramper "with some intermediate steps": btw if someone is interested in what they can theoreticallly get when implicit system import happens https://media.discordapp.net/attachments/371759389889003532/937057687169007667/151627106-713d4d46-ea9d-4694-a886-4d261eaa9d9b.png
18:53:23FromDiscord<lpha3ch0> It must be overloaded somewhere in winim. Thanks for your help.
18:54:46FromDiscord<haxscramper> '> somewhere↵↵https://github.com/khchen/winim/blob/f658fc523a7c8960b2f808d774b0f77cecec86db/winim/utils.nim#L43-L52 - this is where it is overloaded
18:55:33FromDiscord<haxscramper> In reply to @Isofruit "Ah, hax is here,": I probably should wait a little next time before jumping in
18:58:34arkanoidAny idea how to solve this issue when using OOP and finalizers? Have to fix without relying on declaring =destroy for backward compatibility: https://play.nim-lang.org/#ix=3NUW
19:08:39*k0mpjut0r joined #nim
19:27:22FromDiscord<vindaar> https://play.nim-lang.org/#ix=3NVi
19:27:23FromDiscord<vindaar> that's what I'd do
19:40:29FromDiscord<krisppurg> In reply to @Rainbow Asteroids "does anyone know how": first things first, I think you need to get member from guild (because I think message.member doesnt have as much data). Then do something like member.computePerms(guild.channels[m.channel_id])
19:41:11FromDiscord<krisppurg> https://krisppurg.github.io/dimscord/dimscord/helpers.html#computePerms%2CGuild%2CRole
19:42:40FromDiscord<tandy> love it when you can see posts on nim forum but not actually visit them
19:54:55arkanoidvindaar, I don't understand your, example, it gives me the same error as before
19:55:15arkanoiddo I have to define "gcDestructors"?
19:55:25FromDiscord<SecureThisShit> Stupid question. I have an byte sequence and would like to get a pointer to the first byte of it. How can I do that?
19:55:47arkanoidoh, you mean just branch code with and without destructors
19:55:58FromDiscord<vindaar> oops, the two branches are exchanged. The `defined(gcDestructors` should be the one that can define `=destroy` hooks. It's a variable that's automatically defined when compiling with ARC↵(<@709044657232936960_arkanoid=5b=49=52=43=5d>)
19:57:01FromDiscord<vindaar> if `s` is your `seq[byte]` (I assume that's what you mean by a byte sequence?) it's just `s[0].addr` (or possibly `unsafeAddr` depending on the variable. That'll change soon though...)↵(@SecureThisShit)
19:57:55FromDiscord<SecureThisShit> In reply to @vindaar "if `s` is your": Nice, that worked thanks!
19:58:08FromDiscord<SecureThisShit> Easy ^^
19:58:17arkanoidvindaar, ok, let me test this with a larger example
19:59:08*lumo_e joined #nim
20:02:11arkanoidwhat's happening to nim forum?
20:02:30*vicfred joined #nim
20:02:47FromDiscord<vindaar> what do you mean?↵(<@709044657232936960_arkanoid=5b=49=52=43=5d>)
20:03:44arkanoidI was receiving cloudflare 502 bad gateway a minute ago, but now it loads
20:04:05arkanoidnow again
20:04:21arkanoidmaybe just some pages? https://forum.nim-lang.org/t/4132
20:04:47arkanoidshould I report it?
20:05:25FromDiscord<vindaar> oh, indeed
20:13:37arkanoiddone
20:14:03arkanoiddo I have to declare the non-ref type to declare it's =destroy?
20:15:15arkanoidlike `ParentObj = object of RootObj` + `Parent = ref ParentObj` + `proc `=destroy`(self: Parent)` ?
20:15:46arkanoidsorry, third should be `self: ParentObj`
20:21:04FromDiscord<demotomohiro> You can also define destructor for ref object type like ``=destroy`(self: Parent()[].typeof)`
20:21:42FromDiscord<demotomohiro> `=destroy`(self: Parent()[].typeof)
20:21:49arkanoidvindaar, btw the solution you suggested is not possible to apply in my context as it would break existing code that use the `proc delete(x: T) = x.ParentT.delete` pattern
20:23:23FromDiscord<vindaar> existing code in the sense of users being dependent on it or you having to change your code?↵(<@709044657232936960_arkanoid=5b=49=52=43=5d>)
20:27:00arkanoidthe first one
20:28:41FromDiscord<Patitotective> What are you working on? @Arkanoid :p
20:28:47FromDiscord<vindaar> hm \:/
20:29:23FromDiscord<vindaar> I'd still work toward a new major release with a breaking API, because `=destroy` hooks aren't going away
20:30:24FromDiscord<vindaar> after all, it should really only be code that must be removed for the user.↵You might as well add dummy `delete` procedures for ARC, that simply don't do anything, but write a deprecation message.That is of course, unless the `delete` has a purpose beyond being a finalizer
20:32:19arkanoidlong story short: I am not the author of the lib, but an user that is trying to fix a memory leak problem when using ARC/ORC (finalizer not called, issue already posted). Everything started be removing the root of the problem, that is remove all forward declarations of the finalizers, that triggered chain reaction: had to move all constructor in same module as type declaration, had to replace include
20:32:21arkanoidwith import to decouple, had to move imports to late imports (thanks beef!) and merge some modules to avoid recursive dependency issue, but now I'm facing this show stopper: https://play.nim-lang.org/#ix=3NUW
20:34:15arkanoidthis is the library I'm trying to improve, but the "code smell" (just include and no imports) is actually not so wrong considering all the workarounds that I'm facing to improve it
20:34:17arkanoidhttps://github.com/filcuc/nimqml
20:35:18FromDiscord<krisppurg> sent a code paste, see https://play.nim-lang.org/#ix=3NVH
20:36:57*elves-upstate joined #nim
20:38:10FromDiscord<demotomohiro> Adding --passC:-Ic:\path\to\your\include\dir?
20:38:21arkanoidthis is the root cause that makes refc code silently leaking when using arc/orc https://github.com/nim-lang/Nim/issues/19402
20:40:28emeryis there a way to trace which modules import which? I find myself importing unusable modules
20:43:49arkanoidemery, good question, I've been facing kinda same issue when trying to deal with multiple recursive dependency issues. At least when that kind of issue is found the compiler gives you the import chain root cause. I had to write a graphviz diagram by hand
20:45:52FromDiscord<vindaar> from skimming this\:↵the `var T` issue isn't an issue in practice. Just define a regular base `object` that the thing that has the current name is a `ref object` of. Then you can just define the `=destroy` for the object (which then takes `var T`). But as one never calls `=destroy` manually (well, sometimes from a different destructor), you don't really have to change anything in the code (aside from the removal of `delete`)↵(<@709044
20:47:09emeryi'm trying to import asyncdispatch for a platform without sockets or file descriptors, i think thats my problem
20:49:04arkanoidvindaar: the library currently asks the user to extend base class, setup new finalizer, and use it to call parent finalizer: https://github.com/filcuc/nimqml/blob/04888273ce01038334e7c89b379432180acefedb/examples/abstractitemmodel/mylistmodel.nim#L20
20:49:21arkanoidemery, which platform?
20:49:28emeryits a suprise
20:51:35arkanoid:O
20:52:48emeryI think I should remove genode before I add another esoteric platform
20:53:05*elves-upstate quit (Remote host closed the connection)
20:53:22*elves-upstate joined #nim
20:53:35arkanoidbtw a think a simple program that walks the project folder and greps "import" and prints "srcmodule -> dstmodule" can be turned into a graphviz digraph easily
20:54:08arkanoidquestion is that this could be done also with a project level macro, or module level macro, but not sure about their existence
20:54:19*elves-upstate quit (Remote host closed the connection)
20:56:29arkanoiddetomotohiro: thanks for the MyType()[].typeof, works nicely! and also makes a lot of sense
20:59:43FromDiscord<krisppurg> In reply to @demotomohiro "Adding --passC:-Ic:\path\to\your\include\dir?": no?
21:03:13*noeontheend quit (Ping timeout: 250 seconds)
21:04:57*lumo_e quit (Ping timeout: 240 seconds)
21:07:50*lumo_e joined #nim
21:11:31FromDiscord<demotomohiro> @krisppurg Do you have the header file that gcc says no such file? If so, you need to pass the directory to gcc with -I option.
21:16:20*jmdaemon quit (Ping timeout: 256 seconds)
21:25:17FromDiscord<noow> can I disable the "install" and "build" commands for a nimble package
21:25:24FromDiscord<noow> I only want custom tasks
21:26:36FromDiscord<tandy> anyone here know stuff / maintain karax? i wonder if its possible to add better async support
21:26:40FromDiscord<noow> or alternatively, how do I add custom flags to the custom build
21:26:52FromDiscord<noow> (edit) "custom" => "nimble" | "nimblebuild ... " added "command"
21:28:26FromDiscord<Elegantbeef> you can make a `config.nims` inside your project with the flags you want
21:31:00FromDiscord<noow> but then i'll have to --skipProjCfg and --skipParentCfg everywhere
21:31:32arkanoidcan you tell me why I'm not getting destructors being called here with --gc:arc? https://play.nim-lang.org/#ix=3NWd
21:32:03FromDiscord<krisppurg> In reply to @demotomohiro "<@222794789567987712> Do you have": no
21:32:12FromDiscord<krisppurg> (edit) "In reply to @demotomohiro "<@222794789567987712> Do you have": no ... " added "i dont have the header file"
21:37:39FromDiscord<Elegantbeef> Expanding arc shows that it calls destroy for the a, but i dont get why it doesnt destroy the inner object
21:38:57arkanoidhi beef!
21:39:07FromDiscord<Elegantbeef> Hello
21:39:07Amun-Rahmm, same with orc
21:39:23FromDiscord<Elegantbeef> Even minimized it doesnt work
21:39:44arkanoidoh no :(
21:40:39FromDiscord<oldteen> Hello, noob question, what do you use as language server with visual studio code ?
21:40:53FromDiscord<Elegantbeef> I use saem's extension
21:41:00FromDiscord<Elegantbeef> https://play.nim-lang.org/#ix=3NWh this works arknoid
21:41:07arkanoidI also use saem's extension
21:41:16FromDiscord<Elegantbeef> I think the issue is that ref's dont have destructors so it doesnt attempt to call it?
21:41:43FromDiscord<Elegantbeef> Well it attempts to call it but `=destroy(aRefObject)` is just "no call so we dont destruct"
21:42:02FromDiscord<Elegantbeef> I dont recall why ref's dont have destructors
21:43:32*lumo_e quit (Ping timeout: 256 seconds)
21:44:31arkanoid{.borro: `.`.} ... efficient, expressive, elegant
21:45:01*lumo_e joined #nim
21:45:11FromDiscord<Elegantbeef> Lol It's the least toxic borrow method
21:45:17FromDiscord<Elegantbeef> Borrowing anything else is hell on earth
21:46:35arkanoidso this means destructors + arc + ref objects are a no?
21:46:39FromDiscord<vindaar> https://play.nim-lang.org/#ix=3NWj
21:46:53FromDiscord<vindaar> same as on non arc you still have to do something to turn that ref into a non nil
21:47:12FromDiscord<Elegantbeef> Oh i didnt even notice that
21:47:17FromDiscord<Elegantbeef> Nice vindaar 😀
21:51:26arkanoidoh! right, but is it aways necessary? I don't recall calling `new` in my constructors returning ref type all the times
21:51:53FromDiscord<Elegantbeef> Well `Type()` also `new`s internally
21:52:19FromDiscord<Elegantbeef> So you can either `new result` or do `MyType(field: someVal, ... )`
21:52:51arkanoidmmm ok
21:54:10FromDiscord<Elegantbeef> But yes you need to allocate that memory somehow 😛
21:54:27FromDiscord<Elegantbeef> Is life making anymore sense in regard to the delayed imports
21:55:18arkanoidsure, but calling proc that returns a result of ref type should alloc it automatically, just like result for non ref returning procs
21:56:21FromDiscord<Elegantbeef> It 0's memory
21:56:23FromDiscord<Elegantbeef> Nim just 0's memory
21:56:28FromDiscord<Elegantbeef> And for a reference `0` is nil
21:56:35FromDiscord<Elegantbeef> Otherwise you have unintended allocations
21:56:42arkanoidElegantbeef, this question of mine is still about that very same project. I've been applying delayed + merging when necessary and it seems going, problem is that I'm now facing another show stopper: https://play.nim-lang.org/#ix=3NUW
21:57:07FromDiscord<Elegantbeef> Say you're searching a list for a specific value and you return nil if you dont find it, it's bad if the result is intialized
21:57:45arkanoidElegantbeef, there's Option[type] for that
21:58:06*vicfred quit (Quit: Leaving)
21:58:44FromDiscord<Elegantbeef> I mean sure, but that doesnt change the point
21:58:55FromDiscord<Elegantbeef> Intializing the result of something that returns ref is an allocation you cannot stop
21:59:10arkanoidtrue
22:00:06FromDiscord<Elegantbeef> I assume in your world `var a: MyRef` should also allocate 😛
22:00:31arkanoidmmm no :D
22:06:01arkanoidElegantbeef, with that project I have to deal with existing code that's calling `child.Parent.delete()' in their finalizers, this makes useless all my efforts to move constructions and finalizers on top of modules just after type declaration, as ARC will try to declare a =delete before the finalizer itself
22:06:27arkanoidjust like in https://play.nim-lang.org/#ix=3NWt
22:07:50FromDiscord<demotomohiro> In reply to @krisppurg "no i dont have": Then, you need to get the header file. Read readme of module that cause the error or google library name.
22:26:38*carce joined #nim
22:26:42carceHello]
22:26:46carce:D
22:26:56FromDiscord<Elegantbeef> Hello
22:27:21arkanoidhi!
22:29:27arkanoidso there's no way to call parent finalizer/destructor with ARC without using gcDestructors definition or triggering "Error: cannot bind another '=destroy' to: Child:ObjectType;"?
22:30:44FromDiscord<Elegantbeef> Do not know
22:32:34arkanoidthanks for the feedback. Well. I hope the original author will accept to introduce breaking changes, or at least raise compile error with ARC, otherwise code will leak memory
22:33:56FromDiscord<TurtleP> hey, I remember someone telling me I don't need to `staticRead(path/to/nimble).fromNimble(field)` to get the data I want from that, but I was never told how to do it, anyone know?
22:34:16FromDiscord<TurtleP> I tried to import `system/nimscript` but `echo(nimscript.author)` is empty
22:34:25FromDiscord<TurtleP> (edit) "I tried to import `system/nimscript` but `echo(nimscript.author)` is ... emptystring" added "printing an" | "printing anempty ... " added "string"
22:34:51FromDiscord<Elegantbeef> that stuff is only set inside the nimble VM not the compiler VM
22:35:24FromDiscord<TurtleP> right, but is there a way to do the thing I'm trying to do?
22:35:48FromDiscord<Elegantbeef> use strdefines, or parse the nimscript file
22:36:08FromDiscord<TurtleP> you mean the nimble file?
22:36:19FromDiscord<Elegantbeef> Same difference
22:38:51FromDiscord<TurtleP> though I'm confused--since I was told a while back I could just not need to read the nimble file
22:40:03FromDiscord<Elegantbeef> I mean it's interpreted so you can just read it, but in very small chances it'll not work
22:40:25FromDiscord<Elegantbeef> I dont get why you need to know the author but you do you 😀
22:40:36FromDiscord<TurtleP> I mean it's an example
22:40:50FromDiscord<TurtleP> and it's better than having to redefine something already defined
22:41:00FromDiscord<TurtleP> e.g. outputting program version in `help`
22:43:25FromDiscord<gcao> In reply to @haxscramper "well, *right now* the": Thanks for your reply. I sent that message and went for shoveling so couldn't engage with the discussion. Are you the core contributor to Nimskull? What are the features planned for Nimskull but not in Nim? I tried to search that but couldn't find it.
22:45:55FromDiscord<haxscramper> In reply to @gcao "Thanks for your reply.": There are no planned new features for now aside from the improvements listed on the roadmap, and associated requirements for fixing different parts of the compiler/tooling
22:47:01FromDiscord<gcao> sorry for being blatant. that doesn't sound intriguing for developers to give it a try though.
22:47:51FromDiscord<Elegantbeef> Well it's the first step of features
22:47:57FromDiscord<Elegantbeef> Making a solid foundation that can be expanded
22:52:00*lumo_e quit (Ping timeout: 250 seconds)
22:54:00FromDiscord<gcao> A big refactoring or rewrite usually does not end well. Unless you have very good reason it won't fall to that trap.
22:54:01*elves-upstate joined #nim
22:54:33FromDiscord<haxscramper> Well, believe me, we do have a reason, that one I can say with absolute certainty
22:54:44FromDiscord<Elegantbeef> Well the good reason is tech debt that needs addressing
22:55:16FromDiscord<gcao> I guess Araq doesn't feel that way or it's not his priority. so the fork is born?
22:55:17FromDiscord<haxscramper> After spending last two months digging through the current compiler and testament, I would say ton of the code has never seen anyone who thought about long-term sustainability
22:55:44FromDiscord<haxscramper> For example error messages that I talked about for couple months prior
22:55:53FromDiscord<Elegantbeef> To the contrary he actually has commented multiple times about the current state of code and that things could be done better↵(@gcao)
22:55:55FromDiscord<haxscramper> RFCs #323, #324 and #325
22:56:19FromDiscord<haxscramper> I claim that with current architecture the compiler is unable to produce good error messages
22:56:42FromDiscord<haxscramper> This is fundamentally impossible, this thing needs to be refactored with a scorched earth approach
22:57:05FromDiscord<haxscramper> One example of a refactor that I can talk in details about
22:57:29FromDiscord<haxscramper> IT has already been done by me, so far there are no major abnormalities
22:57:52FromDiscord<haxscramper> (edit) "IT" => "It"
22:58:20FromDiscord<gcao> I don't mean to question your decision, just out of pure curiosity from technical perspective.↵If the fork didn't get blessing from Araq, where will the project land? A language work like Nim but is not called Nim ?
22:58:34FromDiscord<Elegantbeef> Yes
22:58:57FromDiscord<huantian> Nimish
22:59:13FromDiscord<Elegantbeef> A language with many similarities to Nim but cares more about a technically correct compiler, and working features
22:59:36FromDiscord<Elegantbeef> For instance look at how many generic bugs exist
22:59:37FromDiscord<gcao> And it's not planned to be compatible with Nim in the long run
22:59:59FromDiscord<Elegantbeef> Well it's not Nim 😀
23:01:53FromDiscord<gcao> If it doesn't get adoption, technically correct compiler won't matter. Do you foresee it gets same adoption as Nim which is only a niche language right now?
23:02:08FromDiscord<gcao> And it'll probably only get adoption from current Nim developers
23:02:32FromDiscord<Elegantbeef> I dont think that's presently a concern
23:02:41*jmdaemon joined #nim
23:02:56FromDiscord<Elegantbeef> The people that contribute to Nimskull want a Nim that works, cause the love the language, but hate it's short comings
23:03:24FromDiscord<Elegantbeef> You dont dedicate time making a fork a compiler for a language you find abhorrent 😀
23:04:06FromDiscord<haxscramper> Right now main concern is to have the working implementation that is understandable to the contributors, there are so many small bits of shifting sand in this right now, that talking about grand future vision is next to useless
23:05:10FromDiscord<gcao> Agree. I just hope the people who work on the fork and who work on current compiler can join force, or the fork gets blessing from Araq. But it's just a hope.
23:05:12FromDiscord<Elegantbeef> Should clarify this is atleast my reason for backporting my fixes to Nimskull
23:07:02FromDiscord<haxscramper> In reply to @gcao "Agree. I just hope": I have my doubts
23:08:40FromDiscord<gcao> When Nim is big enough, with enough resources and backing from big companies, forks will come inevitably. But if at the early stage, lots of core contributors left to work on a fork. that doesn't help Nim to grow
23:09:21FromDiscord<haxscramper> But unmaintainable codebase with ossified garbage all over the place certainly helps nim to grow, right
23:09:24FromDiscord<gcao> Didn't mean to stop you guys doing what you think is right 🙂
23:10:31FromDiscord<gcao> I'll end my comments here. Now I think I got the big picture
23:10:49FromDiscord<haxscramper> Not like we haven't worked on nim before, krux, xybber were core devs, saem made a vscode extension, leorize did vim one, I worked on pattern matching
23:11:26FromDiscord<haxscramper> Nimble RFCs, error message RFCs, there is also a cps that was a joint effort by several people at nimskull
23:11:27FromDiscord<Elegantbeef> How hax i very well feel included!
23:11:28FromDiscord<haxscramper> And so on
23:13:20FromDiscord<tandy> nim is in no position to receive more sponsors really↵(@gcao)
23:13:24FromDiscord<tandy> and thats why it hasnt
23:14:07FromDiscord<tandy> biggest proof that nim has failed in marketing and growth is rust
23:14:27FromDiscord<haxscramper> No, let's be honest here, rust is a bad comparison
23:14:33FromDiscord<ynfle> In reply to @tandy "biggest proof that nim": Lol. That's backed my mozzilla
23:14:33FromDiscord<haxscramper> And go as well
23:14:34FromDiscord<Elegantbeef> Eh there is a minor difference with Rust and Nim that Rust had it's origin mozilla
23:14:40FromDiscord<Elegantbeef> Nim vs Zig is more equivlent
23:14:43FromDiscord<Elegantbeef> Or Nim vs Odin
23:14:54FromDiscord<tandy> eh, zig is way lower level
23:15:00FromDiscord<haxscramper> You just can't compete with 100kk USD per month poured into the language
23:15:12FromDiscord<Elegantbeef> They're both system languages
23:15:15FromDiscord<Elegantbeef> Nim can be used everywhere Zig can be
23:15:27FromDiscord<haxscramper> In reply to @tandy "eh, zig is way": We are talking marketing here mostly
23:15:37FromDiscord<haxscramper> Well, the default rant about
23:15:42FromDiscord<haxscramper> What was it
23:15:46FromDiscord<Elegantbeef> Zig is a system language with origins aside from a large corporation
23:15:46FromDiscord<gcao> Agree. Nim didn't get Rust's adoption mainly because there is no company like Mozilla behind it.
23:15:51FromDiscord<tandy> rust wasnt backed by mozilla from the start↵(@ynfle)
23:15:54*pro joined #nim
23:15:56FromDiscord<Elegantbeef> Nim is a system language with origins aside from large corportation
23:16:04FromDiscord<Elegantbeef> > The language grew out of a personal project begun in 2006 by Mozilla employee Graydon Hoare
23:16:12FromDiscord<Elegantbeef> I mean it's closely related to mozilla
23:16:25FromDiscord<Elegantbeef> > Mozilla began sponsoring the project in 2009
23:17:30FromDiscord<gcao> Let's not attribute the low adoption to Nim's marketing because it's not the language developer's job to push it to Rust's popularity
23:18:04FromDiscord<Elegantbeef> It's an interesting look to see how many developers of Nim have left vs. Zig or Odin
23:18:14FromDiscord<Elegantbeef> Zig has picked up an ex Nim compiler developer
23:18:17FromDiscord<haxscramper> Anyway, it is the "same shit different day" scenario, so to save the time, tldr for my rant is:↵- document things↵- enabling people to do the work is always more productive than BDFLing your opinion in↵- be more open with your plans, talk to people, show your roadmap
23:18:51FromDiscord<gcao> those I do agree
23:18:53FromDiscord<haxscramper> Put UX high on the list, irritated people don't do anything good
23:18:57FromDiscord<Elegantbeef> My view is simpler and why i still often make PRs to Nim, a more competent compiler is a better compiler
23:19:16*lumo_e joined #nim
23:19:22FromDiscord<Elegantbeef> I dont care which compiler works more, i just want fewer "X should work but doesnt"
23:19:28FromDiscord<haxscramper> Basically that's what I stand for in nimskull and that's what I completely failed to find in mainline
23:19:55FromDiscord<Elegantbeef> I'm too easily amused 😀
23:20:00FromDiscord<haxscramper> Repeatedly, on multiple topics
23:20:13FromDiscord<gcao> is Nimskull compatible with Nim 1.6 or 1.4 or some lower version?
23:20:20*pro quit (Client Quit)
23:20:38FromDiscord<Elegantbeef> Nimskull is mostly compatible with 1.6 as that's the forking point, though async is removed and some other behaviours have changed
23:20:57FromDiscord<haxscramper> I might not be the best manager in the world, to be honest I wouldn't trust myself to run a web scraper peoject, but I at least try to follow these principles on nimskull
23:21:12FromDiscord<haxscramper> When I get the real feedback we can iterate from there
23:21:34FromDiscord<haxscramper> In reply to @gcao "is Nimskull compatible with": We still have the full test suite, largely intact in place
23:21:48FromDiscord<haxscramper> And added some 4.5k sloc of spec on top of that
23:22:06FromDiscord<haxscramper> But considering nims stability is a duct tape solution, we can't guarantee anything
23:22:21FromDiscord<haxscramper> Despite trying not to cause pointless regressions
23:23:11FromDiscord<ynfle> In reply to @Elegantbeef "Zig has picked up": Who?
23:23:14FromDiscord<Elegantbeef> There will be eventually an overhaul of the types which will cause regressions almost certainly
23:23:19FromDiscord<Elegantbeef> Lemonboy
23:23:40FromDiscord<Elegantbeef> But the regression will be due to things that arent supposed to work working
23:24:04FromDiscord<haxscramper> In reply to @Elegantbeef "There will be eventually": Yes, but for now the general plan is to stay stable for as long as possible, other we get another layer of sand under our feet
23:24:30FromDiscord<gcao> Thank you @haxscramper @ElegantBeef for taking the time to reply and explain things. I wish both Nim and Nimskull best of luck
23:24:44FromDiscord<Elegantbeef> To be fair though every time i fix a generic issue i wonder if i should attempt to improve Nim's types myself
23:25:17FromDiscord<Elegantbeef> Would take so much longer and probably wouldnt be accepted, but man so many issues with generics
23:25:46FromDiscord<auxym> huh, first I hear of nimskull. You guys plan on keeping C as a compiler target?
23:26:03FromDiscord<haxscramper> Yes, of course
23:26:07FromDiscord<Elegantbeef> Think Saem floated the idea of using Zig possibly, but C is it for now
23:26:09FromDiscord<haxscramper> C, C++, js
23:26:31FromDiscord<Elegantbeef> Think Saem was more interested in the Zig compiler as a portable GCC alternative
23:26:54FromDiscord<haxscramper> I mean, I want to reiterate on that again - we are not going to take a hammer and smash everything we see because idk, someone though it is a good idea
23:27:11FromDiscord<tandy> when backends are more pluggable im interested in setting up a typescript backend \:)↵(@auxym)
23:27:22FromDiscord<auxym> I just really like the whole "targets any platform gcc does" thing
23:27:25FromDiscord<Elegantbeef> I mean arent backends pretty pluggable?
23:27:45FromDiscord<haxscramper> There might be some back and forth discussions about different things, but in general things put on the readme roadmap are what we are going to do
23:27:59FromDiscord<haxscramper> In reply to @Elegantbeef "I mean arent backends": No
23:28:43FromDiscord<haxscramper> For example sem analyzes code little differently for expression, depending on whether we compile to c or cpp
23:29:13FromDiscord<haxscramper> Backends can do the error reports that should've been in the sem layer
23:29:15FromDiscord<Elegantbeef> Ah apparently i've never seen that before
23:29:56*elves-upstate left #nim (#nim)
23:30:03FromDiscord<haxscramper> Also there is a a `backend: RootRef # minor hack` in the module graph that is used by all backends to store their global data
23:30:16FromDiscord<haxscramper> Although that one can be called pluggable probably
23:30:34FromDiscord<tandy> `# minor hack` makes me laff
23:30:54FromDiscord<Elegantbeef> It's a bit better than `pointer`
23:30:59FromDiscord<haxscramper> A bit
23:31:59FromDiscord<haxscramper> Anyway, it's 02:30 here, so I'm going to sleep now
23:32:08FromDiscord<tandy> gnight hax
23:32:14FromDiscord<Elegantbeef> Buh bye
23:32:15FromDiscord<haxscramper> Hope inwont come back to 1.2k messages of the following nimskull discussion
23:32:35FromDiscord<haxscramper> (edit) "inwont" => "I won't"
23:32:40FromDiscord<haxscramper> That I missed
23:32:55FromDiscord<Elegantbeef> Read the roadmap 😀
23:39:08*carce quit (Remote host closed the connection)
23:58:18*jmdaemon quit (Ping timeout: 250 seconds)