<< 20-06-2023 >>
00:00:03FromDiscord<elegantbeef> Give it a few minutes or check matrix to see the response now :d
00:00:13FromDiscord<JJ> lol will do
00:05:19FromDiscord<JJ> wait, hold on, can you not refer to the concept type itself in the current concept syntax?
00:19:23FromDiscord<Elegantbeef> Nope
00:19:34FromDiscord<Elegantbeef> Not really
00:19:35FromDiscord<Elegantbeef> `echo a + b` would fail due to using procedures outside of `Product`
00:19:36FromDiscord<JJ> oh that's a serious limitation, huh.
00:19:36FromDiscord<Elegantbeef> It's unimplemented
00:19:36FromDiscord<Elegantbeef> So yes it fails to compile inside of a generic instantiation presently, but it's supposed to fail at invokation
00:19:37FromDiscord<Elegantbeef> If you treat concepts as feature constrained types, if a procedure is called not inside of the concept it's an error
00:19:37FromDiscord<JJ> mmm, i still don't understand
00:19:37FromDiscord<Elegantbeef> Generics are declared right?
00:19:38FromDiscord<JJ> well you can use `Product[T]` sure
00:19:38FromDiscord<Elegantbeef> The entire point is it fails to compile in declaration
00:19:38FromDiscord<Elegantbeef> Instead of at invocation
00:19:39FromDiscord<Elegantbeef> Hell `echo`could also fail
00:19:39FromDiscord<Elegantbeef> Plus your concept doesnt make any sense really 😄
00:19:39FromDiscord<Elegantbeef> If you have a concept you constrain the operations that type is allowed to your concept
00:19:40FromDiscord<JJ> i don't think i understand what you mean by declaration vs. invocation
00:19:40FromDiscord<Elegantbeef> Since that requires a `$` implemented which is not apart of your product type
00:19:43FromDiscord<JJ> bridge go brr
00:19:54FromDiscord<JJ> sent a long message, see https://paste.rs/8Jnmc
00:30:09FromDiscord<michaelb.eth> anyway to avoid collision between parameter names of a template and invoking a proc within the template with named arguments that match up?
00:32:00FromDiscord<Elegantbeef> So in the above case `echo` and `a + b` are invalid operations since `Product` doesnt require them
00:32:01FromDiscord<Elegantbeef> Declaration is where the generic procedure is written
00:32:02FromDiscord<Elegantbeef> `proc doThing(a, b: Product)` is a generic declaration
00:32:03FromDiscord<JJ> uh i don't know what you mean by declared
00:32:03FromDiscord<Elegantbeef> `doThing(10, 20)` is an invocation
00:32:03FromDiscord<Elegantbeef> It doesnt correctly catch that
00:32:04FromDiscord<JJ> right, and the current concepts impl correctly catches that
00:32:04FromDiscord<Elegantbeef> remove your usage of the generic
00:32:05FromDiscord<Elegantbeef> Cause it doesnt do it until invocation
00:32:05FromDiscord<JJ> i mean it fails on the `echo a + b` line, that&#x27;s where i would expect it to fail
00:32:07FromDiscord<Elegantbeef> I mean it's like a trait
00:32:09FromDiscord<JJ> yeah it does rn
00:32:09FromDiscord<Elegantbeef> Not at invocation, not at usage, not at instantiation, at declaration
00:32:10FromDiscord<Elegantbeef> No it doesnt
00:32:10FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4yE9
00:32:10FromDiscord<Elegantbeef> Whoops declaration checking
00:32:12FromDiscord<Elegantbeef> They map to eachother 1\:1
00:32:12FromDiscord<Elegantbeef> At line 5 the compiler would error
00:32:12FromDiscord<Elegantbeef> The above compiles
00:32:13FromDiscord<JJ> wait but that's new style concepts right?
00:32:13FromDiscord<Elegantbeef> Even with old style it compiles
00:32:14FromDiscord<Elegantbeef> It'll still compile just fine
00:32:14FromDiscord<Elegantbeef> The idea is to treat concepts as more existential types and erroring inside of generic procedures when a usage of the concept is not covered by the concept
00:32:14FromDiscord<Elegantbeef> They're checked to ensure the operations they require are inside the concept
00:32:16FromDiscord<JJ> oh, what the fuck, weird
00:32:18FromDiscord<Elegantbeef> What?
00:32:18FromDiscord<Elegantbeef> Exactly
00:32:18FromDiscord<Elegantbeef> What is the compiler error?
00:32:30FromDiscord<JJ> i am now confused as fuck about a number of other things
00:32:30FromDiscord<JJ> i&#x27;m surprised that concepts are even implementable at all without generic instantiation
00:32:31FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/TirTO
00:32:33FromDiscord<Elegantbeef> It's identical to a trait when implemented in the new style way
00:32:35FromDiscord<JJ> i&#x27;m more confused about araq&#x27;s "Empower concepts to make generic code type-checkable at declaration time, not only at instantiation time" because it seems like generics are fine and everything else is broken
00:32:38FromDiscord<Elegantbeef> > generic code type-checkable at declaration time
00:32:40FromDiscord<Elegantbeef> That's what traits are
00:32:41FromDiscord<Elegantbeef> Rust traits are generic monomorphic functions, that use their logic to do instantiation checking
00:32:43FromDiscord<Elegantbeef> When you say "This thing implements x, y, z" it's identical to the concept that requires those procedures
00:32:43FromDiscord<JJ> sent a code paste, see https://paste.rs/hyYDo
00:32:44FromDiscord<Elegantbeef> Concepts arent broken
00:32:47FromDiscord<Elegantbeef> Declaration checking was never apart of concepts
00:32:48FromDiscord<JJ> i mean that code compiles and runs and fails to throw warnings
00:32:51FromDiscord<JJ> oh
00:32:52FromDiscord<JJ> then how does concepts being totally broken in this example relate?
00:33:33FromDiscord<Elegantbeef> I mean it couldnt be implemented JJ
00:33:33FromDiscord<Elegantbeef> Old style concepts are declarative requirements of a type you can do `c.len == 30` if you want
00:34:02FromDiscord<JJ> wow the bridge is fucked my god
00:34:13FromDiscord<Elegantbeef> @michaelb.eth not really without some type intermediate macro
00:34:25FromDiscord<Elegantbeef> Seems it's fine again
00:34:38FromDiscord<JJ> okay imma go eat and come back to this i think i'm misunderstanding something fundamental
00:35:39FromDiscord<Elegantbeef> Old style concepts were just declarative static code that had to compile for the type to match, just like C++'s.↵New style concepts are declarative procedure headers that have to match for the type to match, there is a plan to have declarative checking for them.
00:36:18FromDiscord<Elegantbeef> The new styled concepts are also easier to make runtime interfaces like rust's `dyn trait`
00:58:30*FromDiscord quit (Server closed connection)
00:59:12*FromDiscord joined #nim
01:00:17FromDiscord<etra> so I have a DLL that would get injected/deinjected multiple times, is there a way to tell nim's compiler 'in this specific function you'd loose control so free everything'? like in `DLL_PROCESS_DETACH `
01:01:03FromDiscord<Elegantbeef> There is not presently a way to deallocate everything afaik
01:01:15FromDiscord<etra> or at least give it a hint that its control would terminate of some sort
02:02:01FromDiscord<arathanis> In reply to @Elegantbeef "Old style concepts were": is there something that is recording the progress towards implemntation of new styl concepts?
02:02:16FromDiscord<arathanis> i know its not all the way there yet, but seeing where its at might help us use it
02:49:08FromDiscord<JJ> aight i&#x27;m back and i understand nothing
02:52:03FromDiscord<JJ> 1. old style concepts are checked at instanciation rather than declaration. this means they check correctly in generics, but incorrectly in normal functions. huh?
02:52:23FromDiscord<Elegantbeef> Old style concepts are type constraints
02:52:24FromDiscord<JJ> oops didn't mean to send that yet, oh well
02:52:34FromDiscord<Elegantbeef> They're checked in type matching yes
02:52:54FromDiscord<Elegantbeef> their usages are unchecked, just like their new style usage
02:54:37FromDiscord<JJ> okay, i&#x27;m confused by "checked in type matching" because they&#x27;re not checked in ex. the `a + b where a, b are Product` case
02:54:54FromDiscord<JJ> i would think that resolves to a function call and then would error out there but it doesn't
02:55:21FromDiscord<Elegantbeef> Right there are two different thing shere
02:55:34FromDiscord<Elegantbeef> We have concepts as type constraints `proc (a, b: Product)`
02:55:44FromDiscord<Elegantbeef> and we have concepts being used `a + b`
02:55:57FromDiscord<Elegantbeef> Constrains are of course impossible to check at declaration, cause we do not know the user types
02:56:07FromDiscord<Elegantbeef> We only want to check that the usages abide by the concept
02:56:10FromDiscord<JJ> oh. i see
02:56:58FromDiscord<JJ> wait, this shouldn&#x27;t be true
02:57:23FromDiscord<arathanis> its structural typing, where if the given type contains at least the structure defined by the concept that its type is considered to be that concept
02:57:30FromDiscord<arathanis> its kind of like an implicit interface
02:57:50FromDiscord<Elegantbeef> So in the product concept case we want to go "No you cannot use `+` as it's not defined for Product"
02:57:50FromDiscord<Elegantbeef> Concepts are practically user defined duck typing and that's it presently
02:57:51FromDiscord<Elegantbeef> That always will be true
02:57:52FromDiscord<JJ> `proc (a, b: Product)` is impossible to check?
02:57:53FromDiscord<JJ> ah
02:57:53FromDiscord<Elegantbeef> You give the shape you want for a type and let the generic instantiation error if something else is needed
02:57:53FromDiscord<Elegantbeef> We only know the that it'll quack like a duck, not what duck it'll be
02:57:55FromDiscord<Elegantbeef> The declaration is the procedure itself
02:57:55FromDiscord<JJ> sorry, where is the declaration again?
02:58:53FromDiscord<Elegantbeef> `proc name(a, b: Product) = body` is a declaration
02:59:07FromDiscord<Elegantbeef> It's not structural typing
03:00:07FromDiscord<Elegantbeef> It's compile time userdefined duck typing
03:00:08FromDiscord<Elegantbeef> In our example all types are product types if a `` is defined
03:00:09FromDiscord<Elegantbeef> We did not define a `+` so it should error statically inside the generic before going to any other code
03:00:10FromDiscord<Elegantbeef> The old style concepts would make this prohibitively difficult
03:00:10FromDiscord<JJ> okay right sure, but i don&#x27;t understand how when we know `a` and `b` are of type `Product` and we know `+` is not defined on type product that the compiler can&#x27;t find out it shouldn&#x27;t compile
03:00:11FromDiscord<JJ> yeah
03:00:11FromDiscord<Elegantbeef> since you can use any static code
03:00:13FromDiscord<Elegantbeef> Of course we can reason this
03:00:15FromDiscord<JJ> oh i get i
03:00:26FromDiscord<Elegantbeef> Is completely valid
03:00:27FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4yEs
03:00:51FromDiscord<Elegantbeef> Old style concepts being unstructured means that you cannot simply check them
03:00:55FromDiscord<JJ> yeah, and that's not verifiably checkable except at runtime
03:01:08FromDiscord<Elegantbeef> No it's compile time checkabvle
03:01:09FromDiscord<Elegantbeef> arrays have static lengths
03:01:20FromDiscord<Elegantbeef> I can implement a type that has a `len` that is static aswell
03:01:34FromDiscord<JJ> ah. okay never mind i don't get it
03:02:15FromDiscord<Elegantbeef> Ok so what are you stuck on?
03:02:52FromDiscord<JJ> what does this mean?
03:03:21FromDiscord<Elegantbeef> Old style concepts allowed any static code it doesnt have to be a procedure
03:03:47FromDiscord<JJ> okay i get it for real. thanks for forcably explaining this through my dense head
03:04:12FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/WTvpQ
03:04:32FromDiscord<Elegantbeef> You cannot map this to new style concepts
03:04:34FromDiscord<JJ> so in order to safely ensure stuff, proc declarations and field accesses have to be easily known to the compiler.
03:04:57FromDiscord<JJ> actually, is there anything else besides associated procs and fields?
03:05:06FromDiscord<Elegantbeef> Well it needs to be structured so you can do assurances that it's accessible
03:05:11FromDiscord<Elegantbeef> Araq plans on making fields a proc aswell
03:05:12*greaser|q quit (Changing host)
03:05:12*greaser|q joined #nim
03:05:14FromDiscord<Elegantbeef> So there is only procs
03:05:25*greaser|q is now known as GreaseMonkey
03:05:39FromDiscord<JJ> i hate this fwiw
03:05:42FromDiscord<Elegantbeef> Really it doesnt matter if it's a proc, field, template, or macro
03:05:57FromDiscord<arathanis> is there a good record of old vs new style concepts and how they differ? and how far along the new style is in its implementation?
03:06:06FromDiscord<Elegantbeef> Nope
03:06:10FromDiscord<arathanis> dang
03:06:14FromDiscord<arathanis> i was afraid of that
03:06:15FromDiscord<Elegantbeef> I use old style cause they're more usable
03:06:27FromDiscord<JJ> i would rather he special-case routine syntax, field syntax, and then keep a generic old-style concepts part
03:06:27FromDiscord<arathanis> you can pick between them???
03:06:28FromDiscord<Elegantbeef> new style have bugs and are much more limited
03:06:34FromDiscord<arathanis> how does that work?
03:06:44FromDiscord<Elegantbeef> one is `concept c` the other is `concept`
03:06:52FromDiscord<arathanis> ohh got it
03:06:53FromDiscord<Elegantbeef> I say c but it can be any symbol
03:07:00FromDiscord<arathanis> old style is the
03:07:07FromDiscord<arathanis> `concept c, type C` thing
03:07:10FromDiscord<JJ> [@arathanis](https://matrix.to/#/%40_discord_136570191038513152%3At2bot.io)\: old concepts are documented in the experimental documentation, new concepts are documented in an rfc on github
03:07:31FromDiscord<arathanis> In reply to @JJ "[@arathanis](https://matrix.to/#/%40_discord_136570": its unfortunately not accurate the examples dont work
03:08:41FromDiscord<arathanis> befe, is there a way to define an any type in old style concepts yet? in the docs it uses `auto T`
03:08:47FromDiscord<arathanis> or was it `var T`?
03:09:08FromDiscord<JJ> `type C` should work?↵(@arathanis)
03:09:34FromDiscord<JJ> the identifiers following `concept` in old-style concepts are generic
03:10:22FromDiscord<arathanis> sent a code paste, see https://play.nim-lang.org/#ix=4yEv
03:10:31FromDiscord<Elegantbeef> No clue what you mean arath
03:10:53FromDiscord<arathanis> in regards to which message/
03:10:55FromDiscord<arathanis> (edit) "message/" => "message?"
03:12:15FromDiscord<Elegantbeef> the auto
03:12:46FromDiscord<arathanis> yeah it was just poor memory, the docs actually have `type T = auto`
03:13:03FromDiscord<JJ> yeah, `c` is a generic instance, `C` is a generic type
03:13:10FromDiscord<JJ> which you can further constrain ofc
03:13:14FromDiscord<arathanis> yes
03:13:47FromDiscord<JJ> oh i thought that's what you meant by an any type
03:14:16FromDiscord<arathanis> i just meant they were placeholders in the concept body for an instance of and the type of the concept resspectively
03:15:15FromDiscord<Elegantbeef> People try to use concepts in FPy ways which hurts my head
03:16:41FromDiscord<arathanis> sent a code paste, see https://play.nim-lang.org/#ix=4yEA
03:16:54FromDiscord<arathanis> (edit) "https://play.nim-lang.org/#ix=4yEA" => "https://play.nim-lang.org/#ix=4yEB"
03:17:12FromDiscord<arathanis> (edit) "https://play.nim-lang.org/#ix=4yEB" => "https://paste.rs/ENHAR"
03:17:19FromDiscord<arathanis> i should play with concepts more to solidify their use
03:17:37FromDiscord<arathanis> In reply to @Elegantbeef "People try to use": yeah that is one of the big draws of it i think, to implement some useful haskell-esque semantics
03:18:05termerI'm trying to pass in a pointer to a cdecl proc (a pure function pointer) as a callback for a C library, but I'm running into issues figuring out how to capture data around it
03:18:13FromDiscord<Elegantbeef> Meanwhile me a chimp using it to constrain tuples
03:18:23termerI can't use a closure because it needs to be a normal function pointer, not a nimcall
03:18:46FromDiscord<Elegantbeef> You could use a global variable to hold onto your state inside the proc
03:18:50FromDiscord<Elegantbeef> Not clean but it'd work
03:19:03termerBasically I've got a for loop that iterates over an item
03:19:06termer*over items
03:19:13termerand for each, it creates a button and sets the callback
03:19:30termerit's a bit difficult to do it for that reason
03:19:34FromDiscord<arathanis> In reply to @Elegantbeef "Meanwhile me a chimp": i see nothing wrong with this, got an example?
03:22:43FromDiscord<JJ> also beef, curious if you know if general iterator semantics can be overridden
03:22:58FromDiscord<Elegantbeef> The above link
03:23:00FromDiscord<JJ> er. never mind that's a stupid statemet
03:23:09FromDiscord<arathanis> In reply to @termer "I can't use a": use rawEnv and rawProc?
03:23:46FromDiscord<arathanis> https://nim-lang.org/docs/system.html#rawEnv%2CT↵https://nim-lang.org/docs/system.html#rawProc%2CT↵as suggested by Araq here: https://forum.nim-lang.org/t/1488
03:24:11FromDiscord<arathanis> might work, and it would be cool to know if it does
03:26:11FromDiscord<JJ> curious if you can provide an extremely generic iterator, so that you could take an iterator, transform it into a custom iterator with say `iterator.peekable()`, and then define a `.next()` and `.peek()` on that iterator
03:26:49FromDiscord<Elegantbeef> https://github.com/beef331/slicerator/blob/master/src/closures.nim#L91-L108 this is sketchy as hell
03:26:59FromDiscord<Elegantbeef> we already have next it's just `myIter()`
03:27:07FromDiscord<JJ> this is hammering rust style semantics into nim ofc. but i&#x27;m kinda curious if there&#x27;s a good way to get the ergonomics of having an arbitrary peekable iterator without needing allocations when not needed
03:27:24FromDiscord<Elegantbeef> There is no way to do that
03:28:16FromDiscord<Elegantbeef> You cannot really have a peekable iterator with present Nim
03:28:27FromDiscord<JJ> oh, i mean you'd have to explicitly mark it as peekable i think
03:28:27FromDiscord<JJ> hmm
03:29:01FromDiscord<JJ> what you've got there is fucky and interesting thanks
03:29:11FromDiscord<Elegantbeef> You have to store state to have a peekable iteratorr afaik
03:29:32FromDiscord<Elegantbeef> like consider an iterator on a socket how do you peek
03:29:49termerI did it
03:29:53termerMy horrible solution
03:30:02termerso the function takes in a UI component handle
03:30:19termerso I have a global table with a key of that handle
03:30:25termerand its value is a pointer to a nim closure
03:30:29termermy only concern here is memory leaks
03:31:10termerwhatever cleanup thing is required for the C library for these components will need to also clear the entries in the table
03:31:15termerbut for now I'm happy
03:31:30termerI was also wondering, does Nim have a weak reference it can use
03:31:39termerThat might help
03:31:43termerlike a weak reference map
03:31:50termerThe UI library is iup by the way
03:32:17FromDiscord<JJ> what i&#x27;m hoping could work is a `PeekableIterator[T] = object[iterator: Iterator, buffer: seq[T]`, where you could somehow override the iterator implementation to push to a buffer on peek and pull from the buffer first on next. but that i don&#x27;t think there&#x27;s a way to do that yeah
03:34:09FromDiscord<JJ> i don&#x27;t think there are weakrefs. i kinda wish there were though, but orc is good enough for me
03:37:52FromDiscord<Elegantbeef> Well we do have weakrefs but they're nothing fancier than `ptr T`
03:38:00FromDiscord<Elegantbeef> `{.cursor.}` exists
03:39:09FromDiscord<JJ> oh right!
03:39:31FromDiscord<JJ> it's such a weird name. i'm sure it makes sense internally though
03:40:02termerwhat is .cursor.
03:40:26FromDiscord<Elegantbeef> Nim's sorta weakref
03:41:23termeris it documented lol
03:41:41termeror is it only documented in the mind of Beef and Araq
03:41:45FromDiscord<JJ> yeah\: https://nim-lang.org/docs/destructors.html
03:41:49termerthank god
03:42:15termeroh, I remember reading about this
03:42:24termerI completely forgot about it
03:43:31FromDiscord<JJ> i wish zevv would finish http://zevv.nl/nim-memory/ and http://zevv.nl/nim-memory/nim-arc.html
03:44:35FromDiscord<JJ> i also wish i knew enough about nim to fill out some sections \:-P
03:45:04FromDiscord<Elegantbeef> "Beef write more please" 😄
03:45:59FromDiscord<JJ> beef write more please
03:46:01FromDiscord<JJ> haha
03:46:21FromDiscord<michaelb.eth> yes please, many of us will pay in $ and/or beanie babies and/or garbage pail kids
03:46:39termerI will pay real money (monero and cash)
03:46:54FromDiscord<michaelb.eth> (edit) "$" => "USD"
03:54:25*NimBot joined #nim
04:08:48FromDiscord<JJ> bluh.nim https://media.discordapp.net/attachments/371759389889003532/1120565892061876315/bluh.nim
04:08:49FromDiscord<JJ> any idea why this code does not compile?
04:11:18FromDiscord<Elegantbeef> seq does not have a `pop` that returns `seq`
04:11:59FromDiscord<Elegantbeef> `concept x, y` binds two symbols to the type you're checking
04:12:14FromDiscord<Elegantbeef> you want `type Stack[T] = concept s`
04:12:22FromDiscord<JJ> oh! i see
04:12:37FromDiscord<JJ> i thought it was an arbitrary object for some reason.
04:13:02FromDiscord<JJ> and then type T is also the type you're checking?
04:13:04FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4yEJ
04:13:06FromDiscord<Elegantbeef> Yes
04:13:52FromDiscord<JJ> ah. yeah of course it's literally always been this way. i broke my brain understanding the problem with old concepts or smth
04:43:44*azimut quit (Ping timeout: 240 seconds)
04:44:24FromDiscord<heysokam> @elegantbeef what does graffitti do if you ask it to send a new tag, but the version is the same?
04:44:46FromDiscord<Elegantbeef> If a tag exists it should do nothing
04:45:48FromDiscord<Elegantbeef> It uses git diff to find versions to tag
04:45:52FromDiscord<Elegantbeef> If there is no diff there is no tag
04:46:31FromDiscord<heysokam> kk. so it should be safe to ask it to do something everytime you compile the package?
04:46:57FromDiscord<heysokam> i was just thinking of how to stop having to remember to update the task, and only having to remember changing the version instead
04:46:58FromDiscord<Elegantbeef> Probably
04:47:20FromDiscord<heysokam> (edit) "task," => "tag,"
05:06:56*ntat joined #nim
05:16:06*rockcavera quit (Remote host closed the connection)
05:34:50FromDiscord<heysokam> @elegantbeef might be helpful to also take the body of the commit into the tag message 🤔 https://media.discordapp.net/attachments/371759389889003532/1120587539879833621/image.png https://media.discordapp.net/attachments/371759389889003532/1120587540219568138/image.png
05:35:07FromDiscord<heysokam> also, tysm for this tool, i love it 🙏
05:35:38FromDiscord<Elegantbeef> Perhaps, I made it in like 2 days and it's intended to be as simple as possible
05:36:22FromDiscord<heysokam> definitely worth the 2d for me, will be using this everywhere from now on
05:36:47FromDiscord<Elegantbeef> Well if you want to add anything feel free to, it's very simple
05:37:41FromDiscord<heysokam> oh 56 lines. damn, thats tiny
05:38:55FromDiscord<heysokam> i figure this is where the message is created? https://media.discordapp.net/attachments/371759389889003532/1120588570416140308/image.png
05:39:12FromDiscord<Elegantbeef> Yes
05:39:53FromDiscord<heysokam> what does quoteshell do?
05:40:06FromDiscord<Elegantbeef> rtfm 😄
05:40:15FromDiscord<heysokam> manual is not very helpful https://media.discordapp.net/attachments/371759389889003532/1120588905226440725/image.png
05:40:35FromDiscord<Elegantbeef> https://nim-lang.org/docs/os.html#quoteShell%2Cstring
05:40:50FromDiscord<heysokam> oh
05:41:17FromDiscord<heysokam> oh so you just literally take the first id of the line[] array?
05:41:54FromDiscord<Elegantbeef> What?
05:42:15FromDiscord<heysokam> im trying to understand that line, but i don't really get how you get the message body
05:42:25FromDiscord<heysokam> i just want to extend that so that it also adds the rest fo the commit message
05:42:29FromDiscord<Elegantbeef> The message is the commit got from `git log`
05:43:01FromDiscord<Elegantbeef> ` gitCommand = "git -C $1 log --format=\"%H %s\" -- $2"`
05:43:24FromDiscord<Elegantbeef> That returns `hash` and a `commit message`
05:44:04FromDiscord<heysokam> how do you change it so that the rest of the message gets added too?
05:44:31FromDiscord<heysokam> i just dont understand how you are filtering it, so i cannot add anything to it because im not getting how it works
05:44:58FromDiscord<heysokam> usual sokam obtuse style, probably 🐠 , but i really dont get how to extend that
05:45:16FromDiscord<Elegantbeef> I then check for each commit the diff of the nimble to see if the version was changed
05:45:48FromDiscord<Elegantbeef> No clue that's on you to look at `git log --format`
05:46:06FromDiscord<JJ> hmm. trying to figure out what i think about new style concepts. i think zah is right on the money where there is a strong distinction between "arbitrary type predicates" and "supported syntax requirements"
05:46:18FromDiscord<Elegantbeef> `git log` allows you to format, I chose a format that prints `hash commit message`
05:46:30FromDiscord<JJ> where old concepts are the first, and new concepts are the second
05:46:38FromDiscord<Elegantbeef> If you want more you'll need to change that
05:46:53FromDiscord<Elegantbeef> They're two different things yes
05:47:05FromDiscord<Elegantbeef> One is user defined duck typing, the other is interface like
05:47:28FromDiscord<JJ> i think the "supported syntax requirements" can also be distinguished into functions defined on the type, and what type the function is
05:47:40FromDiscord<heysokam> sent a code paste, see https://play.nim-lang.org/#ix=4yEW
05:47:57FromDiscord<Elegantbeef> Makes parsing more complicated
05:48:34FromDiscord<Elegantbeef> Personally for me I found format fine, so I personally do not have much interest in looking into this
05:48:54FromDiscord<Elegantbeef> I won't decline a PR of course but just look at the `git log` manpage
05:50:22FromDiscord<heysokam> i add a crap ton of extra messages in my git messages usually, the patchnotes are almost always there. so i would be losing a lot of info
05:50:31FromDiscord<JJ> anyway i&#x27;m trying to figure out if there&#x27;s some way to unify this. i almost want to say keeping old style concepts, renaming new style concepts to use the reserved `interface` keyword, and adding some way to denote what type the interface is would work. but these would then have to be defined for the same type? and that&#x27;s kinda bad
05:50:56FromDiscord<Elegantbeef> I don't know how to format the git messages in an easily parsable way so look into that 😄
05:51:18FromDiscord<JJ> also araq wants to just make fields referentiable as procs but that&#x27;s pretty gross and doesn&#x27;t scale across different or more complex types
05:51:27FromDiscord<Elegantbeef> It scales mostly fine afaik
05:52:17FromDiscord<JJ> it won't work for variant objects or static arrays at least
05:52:23FromDiscord<JJ> sure it does work for seqs but
05:52:30FromDiscord<Elegantbeef> Why not arrays?
05:52:36FromDiscord<Elegantbeef> Why not variant objects?
05:53:07FromDiscord<JJ> arrays\: no way to say "i know this array is six elements long, so array[5] is safe"
05:53:55FromDiscord<JJ> variant objects\: i guess it does, ugh, but that&#x27;s gross
05:54:03FromDiscord<Elegantbeef> Well `array[5]` is statically checked so it's not a big deal
05:54:11FromDiscord<heysokam> `git log --format="%H %s %b`
05:54:28FromDiscord<heysokam> literally your syntax, but with `%b` added, so the body is not cut
05:54:33FromDiscord<Elegantbeef> If you do an explicit index it's an statically caught issue, otherwise it's runtime
05:54:37FromDiscord<Elegantbeef> Well PR it sokam
05:54:39FromDiscord<JJ> well it wouldn't be statically checked within functions though right
05:54:48FromDiscord<Elegantbeef> These are generic
05:54:56FromDiscord<JJ> for the same reason as old concepts compiling weird code earlier
05:54:57FromDiscord<heysokam> PR a two characters change? 🤔
05:55:07FromDiscord<Elegantbeef> you also need to handle it in the tag don't you?
05:55:32FromDiscord<Elegantbeef> Also people have PR'd worse, they've PR'd typo fixes in comments 😄
05:56:28FromDiscord<heysokam> the main problem, beef, is that you are like 1000million times smarter than i am, and i struggle to understand your code like an idiot
05:56:38FromDiscord<heysokam> i have no clue how to change it so that the body is not lost
05:57:13FromDiscord<Elegantbeef> Well you've added the body back to the message so now if we run it it should push the body as the tag message
05:57:52FromDiscord<heysokam> didnt you say that it needs to be handled in the tag?
05:57:55FromDiscord<Elegantbeef> Is that where all that information is supposed to go?
05:58:16FromDiscord<heysokam> wdym?
05:58:16FromDiscord<Elegantbeef> I don't know if it's spposed to go to `-m` inside git tags 😄
05:58:32FromDiscord<heysokam> whats `-m`?
05:58:55FromDiscord<Elegantbeef> ` tagCommand = "git -C $1 tag -a v$2 $3 -m \"$4\""` right now everything goes to the message of the `tag`, which means it'd put it all as the tag message
05:59:25FromDiscord<heysokam> yeah, just like the title, right?
05:59:28FromDiscord<Elegantbeef> Is that correct for these large messages, would that behave properly?
05:59:38FromDiscord<heysokam> right now you are only adding the title, the idea is to also add the %b body
06:00:05FromDiscord<heysokam> are tags capped to a certain message size or something?
06:00:23FromDiscord<Elegantbeef> I don't think so
06:00:45FromDiscord<Elegantbeef> So this seems like all you want me to do is append `%b`, do me a favour nuke your tags and run graffiti with that change
06:01:49FromDiscord<heysokam> you mean you updated it? or that i should clone and test with the change locally? or what does that mean?
06:02:00FromDiscord<Elegantbeef> Yes test locally with that change
06:02:07FromDiscord<Elegantbeef> I want to ensure it works before I push it
06:02:17FromDiscord<heysokam> how do i edit the package? its not a lib, so nimble didn't download the src
06:02:19FromDiscord<Elegantbeef> I do not have any large repos with large messages
06:02:33FromDiscord<Elegantbeef> clone it like a normal git repo
06:02:38FromDiscord<heysokam> k
06:02:58FromDiscord<Elegantbeef> to delete all your tags you're going to have to refer to your favourite search engine
06:04:03FromDiscord<heysokam> sent a code paste, see https://play.nim-lang.org/#ix=4yF0
06:04:08FromDiscord<Elegantbeef> Seems fine
06:07:00FromDiscord<heysokam> was easy enough. `git push --delete origin tagname` https://media.discordapp.net/attachments/371759389889003532/1120595636690104371/image.png
06:09:10FromDiscord<heysokam> i don't think this is working https://media.discordapp.net/attachments/371759389889003532/1120596182083838042/image.png
06:09:21FromDiscord<heysokam> but i don't know if it used the correct version, or if it downloaded the git version again
06:09:52FromDiscord<heysokam> it also re-added the 0.0.2, which shouldn't have done because it got deleted
06:11:09FromDiscord<Elegantbeef> If you didnt delete your local tags pushing will add them back
06:11:25FromDiscord<Elegantbeef> It adds all tags that were there though
06:11:33FromDiscord<Elegantbeef> That's part of the point of it
06:11:54FromDiscord<heysokam> so the git push delete thing was incorrect?
06:12:09FromDiscord<Elegantbeef> No you have to also delete local I think... I do not recall
06:12:17FromDiscord<Elegantbeef> I just copied some commands that deleted all tags
06:14:06FromDiscord<heysokam> definitely didn't work https://media.discordapp.net/attachments/371759389889003532/1120597427183951914/image.png
06:14:20FromDiscord<heysokam> added more text, but not the entire thing
06:15:14FromDiscord<Elegantbeef> Well i'm splitting by lines so likely it is multilined
06:15:35FromDiscord<heysokam> yep
06:15:52FromDiscord<Elegantbeef> So this now becomes a case of rewriting the entire program 😄
06:16:39FromDiscord<heysokam> forget it, i will just suck it and stay with the title exclusively. i don't understand the code anyway, so i cannot debug it
06:17:05FromDiscord<Elegantbeef> It now needs to be changed to `--format="Commit: %H \n %s \n %b\n"`
06:20:22FromDiscord<Elegantbeef> This also could be done with `--format="%H"` and going back to the shell to get the body
06:27:51FromDiscord<Elegantbeef> To explain the code in hopes to motivate you. The complex stuff isnt even really the code it's the commands. `gitCommand` returns all the commits that the `nimble` file was modified in with it's message. `diff` returns the diff of that commit in that file. We use those two in combination to find all the commits that the nimble was modified and where that modification changed `version =`
06:28:43FromDiscord<Elegantbeef> Everything else is just working on that using string operations, it's `commit yourMessageHere` so we know the commit is the left of a space
06:50:16*PMunch joined #nim
07:04:00*oldpcuser quit (Remote host closed the connection)
07:06:33*oldpcuser joined #nim
07:28:24FromDiscord<dlesnoff> sent a long message, see http://ix.io/4yFb
07:31:02FromDiscord<dlesnoff> I have seen a package to do CPS-styled optimizations so I assumed that Nim just do not do this by default.
07:33:00FromDiscord<dlesnoff> I could test with the implementations whether some optimizations are done but I am not sure how to check memory usage at runtime.
07:33:00FromDiscord<Elegantbeef> Nim inline iterators are just templates, it's closure iterators are lambda lifted with very few optimisations
07:35:18FromDiscord<demotomohiro> In reply to @jiezron "Hello! I have a": It seems such optimization depends on backend C compiler: https://github.com/nim-lang/Nim/issues/21844
07:38:32FromDiscord<JJ> what's up with these weird concepts here: https://github.com/mashingan/anonimongo/blob/9b65c833cef22d942fbcd97b2d83ab6c2208a5a7/src/anonimongo/core/streamable.nim#L5 ? they look like a mashup of old and new concepts
07:39:50FromDiscord<Elegantbeef> Never seen those before
07:39:51FromDiscord<Elegantbeef> Doubt that works
07:42:08PMunchIsn't that the old concept syntax? I seem to remember stuff like that working
07:42:42FromDiscord<JJ> there's an old-old concept syntax?
07:42:47FromDiscord<Elegantbeef> Right but there's the pointles proc headers
07:42:48FromDiscord<Elegantbeef> Like what's even the point of those
07:42:54FromDiscord<Elegantbeef> No there are just old and new
07:43:44FromDiscord<Elegantbeef> It seems like they might need those procs to exist for their generics to compile so they include them there
07:43:49FromDiscord<Elegantbeef> But that doesnt make any sense
07:47:09FromDiscord<dlesnoff> In reply to @demotomohiro "It seems such optimization": Thanks @demotomohiro ! Tail call optimizations, that's it!
07:48:40FromDiscord<JJ> In reply to @Elegantbeef "But that doesnt make": hmm. it's one weird case out of everything on github so good chance it just doesn't work
07:48:57FromDiscord<Elegantbeef> I mean it probably compiles but it's pointless
07:49:09FromDiscord<Elegantbeef> You're introducing a procedure so other stuff compiles defeating the purpose of it
07:49:38FromDiscord<JJ> also, i've been going through and it's striking how concepts are basically used for exclusively one-of an arbitrary structural type or as an interface
07:50:14FromDiscord<JJ> does nim have any kind of structural equality among structs again?
07:50:26FromDiscord<Elegantbeef> tuples do
07:50:59FromDiscord<JJ> not with objects, though? hmm
07:52:11FromDiscord<Elegantbeef> Indeed
07:52:33FromDiscord<Elegantbeef> I have thought about a row polymorphic type matching done like `object[x, y: int]`
07:52:46FromDiscord<Elegantbeef> Concepts are great but they're all the same level
07:53:16FromDiscord<Elegantbeef> It'd be nice to be able to say `proc doThing(vec: object[x, y, z: float32])` and `proc doThing(vec: object[x, y: float32])`
07:53:31FromDiscord<Elegantbeef> Then instead of inheritance you could use the above and get the lovely semantics
07:54:21FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4yFe
08:34:33FromDiscord<arkanoid> I am receiving errors while browsing nimble.directory website↵↵https://termbin.com/shqj
08:37:49FromDiscord<voidwalker> it's written in nim
08:43:21PMuncharkanoid, seems like it tries to parse a version number and expects a number
08:43:35PMunch@federico3 is still the maintainer of the directory I believe
08:55:34FromDiscord<mratsim> In reply to @jiezron "Hello! I have a": you're looking for tail call and they are on only with -O1 in C compilers
09:08:20FromDiscord<yodacbt> hi everyone
09:08:26FromDiscord<yodacbt> can someone help me with something i am coding
09:08:36FromDiscord<Andreas> sure, show us smth
09:08:57FromDiscord<yodacbt> sent a long message, see http://ix.io/4yFo
09:09:08FromDiscord<yodacbt> ok it says execution of a external program failed
09:09:53FromDiscord<Andreas> In reply to @yodacbt "ok it says execution": sry. never tried nimx..
09:11:34FromDiscord<yodacbt> its ok
09:11:36FromDiscord<yodacbt> i am new
09:11:41FromDiscord<yodacbt> i have no idea what i am doing
09:12:04FromDiscord<Andreas> In reply to @yodacbt "i have no idea": can you run one of the provided examples or tests ?
09:21:49FromDiscord<odexine> In reply to @yodacbt "ok it says execution": What is the full error
09:38:30*lumidify quit (Server closed connection)
09:38:46*lumidify joined #nim
10:07:24FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#ix=4yFy
10:27:16*lucasta joined #nim
10:56:31FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#ix=4yFD
11:17:10*void09 quit (Server closed connection)
11:17:26*void09 joined #nim
11:41:22FromDiscord<jmgomez> what about `quit`?
11:47:16FromDiscord<graveflo> you can `raise new(AssertionDefect)` for example
11:49:03FromDiscord<graveflo> I've only seen `error` used like that in macros. For runtime errors `raise` seems to work better
11:49:35FromDiscord<jmgomez> He wants to error at compile time though
11:50:21FromDiscord<graveflo> oh well I'm stumped then. It does show up as an error with nimsuggest so I guess that counts as a in between or something.
11:54:36*sagax quit (Remote host closed the connection)
11:54:42FromDiscord<graveflo> I think an error at compile time would be pretty difficult to make happen because that would cause `foo` to fail compiling in general. I could be wrong but I am still leaning towards raising the exception to do this sort of thing
12:01:16FromDiscord<demotomohiro> @jmgomez @graveflo Thank you!. Both `quit()` and `raise` worked.↵`raise` looks better as it prints call stack.
12:02:28FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#ix=4yFV
13:04:06*azimut joined #nim
13:50:17FromDiscord<chri> hi a question about cpp code import
13:51:02FromDiscord<chri> i think nim may be one of the best languages for interoperability with cpp
13:51:13FromDiscord<chri> since you can just compile down to it
13:51:50*PMunch quit (Quit: Leaving)
13:52:07FromDiscord<chri> and you can inline cpp code, then wrapping functions becomes very easy, isnt it?
13:52:48FromDiscord<chri> however cpp has very complex type system
13:53:10FromDiscord<chri> for example, wrapping std::vector in nim sounds hard to me, isnt it?
13:53:15FromDiscord<chri> maybe im wrong
13:53:49FromDiscord<chri> but cpp objects may do a lot of stuff under the hood
13:55:27FromDiscord<chri> sent a code paste, see https://play.nim-lang.org/#ix=4yGh
13:56:08FromDiscord<chri> am i supposed to wrap all the copy stuff and destructor and stuff also in nim?
13:56:58FromDiscord<chri> and how do i wrap stdvec in nim? do i need to replicate all the internal structure of stdvec or maybe just its size?
14:02:55FromDiscord<chri> sent a code paste, see https://play.nim-lang.org/#ix=4yGl
14:08:06FromDiscord<spotlightkid> It very much depends on how you want to use the Cpp type in Nim. If you just just receive a pointer to it from a Cpp function and then pass it on to another Cpp function, you can just use a `ptr` to a `distinct object`.↵(@chri)
14:14:28FromDiscord<mratsim> In reply to @chri "for example, wrapping std::vector": - https://github.com/mratsim/agent-smith/blob/master/third_party/std_cpp.nim#L26-L31↵- https://github.com/Clonkk/nim-cppstl/blob/master/cppstl/std_vector.nim
14:15:25FromDiscord<chri> In reply to @spotlightkid "It very much depends": and what abou more complex types
14:15:34FromDiscord<chri> like a big nested struct
14:15:44FromDiscord<chri> if i don't need the fields in nim
14:16:03FromDiscord<chri> do i need to replicate the same size?
14:16:36FromDiscord<spotlightkid> Depends on who allocates the object, Nim or Cpp.
14:16:42FromDiscord<chri> like distinct int[40bytes
14:16:51FromDiscord<mratsim> In reply to @chri "like a big nested": https://github.com/SciNim/flambeau/blob/master/flambeau/raw/bindings/data_api.nim#L78-L253
14:16:59FromDiscord<chri> In reply to @spotlightkid "Depends on who allocates": cpp constructor
14:17:04FromDiscord<chri> but nim wraps it
14:19:23FromDiscord<spotlightkid> Then probably Cpp allocates it and if you don't need access to its fields, you probably don't need to care about its size. It's hard to give any concrete advice, though, without knowing your exact use case, e.g. the signatures of the classes/methods you want to wrap.
14:20:52FromDiscord<chri> In reply to @spotlightkid "Then probably Cpp allocates": ok thanks, so sizeof[CppVector] will be pretty useless
14:21:27FromDiscord<chri> im gonna need to wrap the whole object structure in nim to make it work
14:35:07FromDiscord<jmgomez> You dont need to fake the fields if you use `importcpp` as the size is known by the cpp compiler. You only need to import what you use. if you want some inspiration here you can find a lot of examples: `https://github.com/jmgomez/NimForUE` Probably it's the largest Nim based C++ project out there. Just search for `importcpp`
14:35:17FromDiscord<jmgomez> (edit) "`https://github.com/jmgomez/NimForUE` Probably" => "https://github.com/jmgomez/NimForUE`Probably"
14:35:22FromDiscord<jmgomez> (edit) "https://github.com/jmgomez/NimForUE`Probably" => "https://github.com/jmgomez/NimForUE Probably"
14:36:01FromDiscord<jmgomez> (edit) "inspiration here" => "inspiration,"
14:48:12*lucasta quit (Quit: Leaving)
14:50:29*ntat quit (Ping timeout: 240 seconds)
14:59:37FromDiscord<chri> In reply to @jmgomez "You dont need to": but not known to nim compilee
14:59:56FromDiscord<chri> the nims sizeof i meant, not the cpp one
15:03:41FromDiscord<jmgomez> The Nim compiler cant use `importc` types at compile time though. Nim `sizeof` does know the C++ size
15:05:38*ntat joined #nim
15:11:14FromDiscord<jmgomez> sent a code paste, see https://play.nim-lang.org/#ix=4yGF
15:27:29FromDiscord<that_dude.> @chri take a loot at this blog post. https://scripter.co/binding-nim-to-c-plus-plus-std-list/ It may be interesting to you
15:48:18*om3ga quit (Ping timeout: 265 seconds)
15:54:06*oldpcuser quit (Remote host closed the connection)
15:55:01*oldpcuser joined #nim
15:55:38*oldpcuser quit (Remote host closed the connection)
15:56:03*oldpcuser joined #nim
16:01:49FromDiscord<chri> In reply to @jmgomez "The Nim compiler cant": yes this is what i'm saying
16:02:35*oldpcuser quit (Read error: Connection reset by peer)
16:03:09FromDiscord<jmgomez> Point is, what do you care about the size of a type at Nim compile time?
16:03:19*oldpcuser joined #nim
16:03:24FromDiscord<jmgomez> (edit) "what" => "why"
16:03:27FromDiscord<chri> the point is not the size actually
16:03:39FromDiscord<chri> the point is that you can't interact from nim with the object
16:03:47FromDiscord<chri> because that's just an alias for generation
16:03:51*oldpcuser quit (Remote host closed the connection)
16:03:51FromDiscord<chri> (edit) "because that's just an alias for ... generation" added "code"
16:04:16*oldpcuser joined #nim
16:04:27FromDiscord<chri> so you need to wrap all the object if you need to interact with it in a full way (like, meta programming or something)
16:04:33FromDiscord<chri> but wrapping it fully is annoying
16:04:41FromDiscord<chri> that's it, just a dubt
16:04:52FromDiscord<chri> (edit) "dubt" => "duobt"
16:05:04FromDiscord<chri> (edit) "duobt" => "dubbio"
16:05:17FromDiscord<chri> (edit) "dubbio" => "doubt"
16:05:29*oldpcuser quit (Remote host closed the connection)
16:06:07*oldpcuser joined #nim
16:06:39FromDiscord<jmgomez> You may have a miss conception somewhere or I dont understand what you are trying to do. You want to modify the AST of an imported type?
16:07:55FromDiscord<chri> no, it doesn't even have to do with nim
16:07:59FromDiscord<chri> just a nim like language
16:08:08FromDiscord<chri> sorry for confusion
16:08:12FromDiscord<chri> (edit) "sorry for ... confusion" added "creating"
16:32:14FromDiscord<Ras> when where you when nim 3 dropped ??
17:04:29FromDiscord<asviest> does anyone know how clib and clibuserdefine work in c2nim ?
17:14:17*rockcavera joined #nim
17:22:33*ntat quit (Quit: Leaving)
17:33:55NimEventerNew thread by Traveler: Multithread with js target, see https://forum.nim-lang.org/t/10288
17:43:47*ntat joined #nim
17:56:32FromDiscord<voidwalker> `when (compiles do: import morelogging): discard` what does this do ?
18:01:08FromDiscord<demotomohiro> In reply to @asviest "does anyone know how": https://github.com/nim-lang/c2nim↵> c2nim is a tool to translate ANSI C code to Nim. The output is human-readable Nim code that is meant to be tweaked by hand after the translation process.
18:02:25FromDiscord<demotomohiro> In reply to @voidwalker "`when (compiles do: import": seems checking if `morelogging` module can be imported.
18:04:07FromDiscord<voidwalker> somebody did a small PR to my repo lol. first time. I really must learn to use git to keep that shit in sync
18:17:35*ntat_ joined #nim
18:21:00*cedb quit (Ping timeout: 240 seconds)
18:21:01*Ekho quit (Ping timeout: 240 seconds)
18:21:01*systemdsucks quit (Ping timeout: 240 seconds)
18:21:02*ntat quit (Ping timeout: 240 seconds)
18:21:10*FromDiscord quit (Ping timeout: 240 seconds)
18:21:16*FromDiscord joined #nim
18:21:42*systemdsucks joined #nim
18:21:45*cedb joined #nim
18:21:57*ntat_ is now known as ntat
18:24:34*Ekho joined #nim
19:06:46FromDiscord<oa3> Is there a comparable thing to rusts structs in nim?
19:10:48FromDiscord<funnybananas> sent a long message, see http://ix.io/4yHD
19:11:31FromDiscord<funnybananas> (module may be the wrong terminology, still learning Nim)
19:14:29FromDiscord<deech4592> That's really weird. What happens if you compile it with the mingw toolchain on Windows?
19:17:11FromDiscord<funnybananas> hmm i haven't tried using mingw on Windows or another version of mingw
19:17:15FromDiscord<funnybananas> i can try that next to see
19:30:33FromDiscord<Elegantbeef> objects↵(@oa3)
19:32:58FromDiscord<Andreas> @elegantbeef could you pls. repost your article on concepts from a couple of months ago - i can't find it .. thx
19:33:11FromDiscord<Elegantbeef> https://www.jasonbeetham.com/codereuse.html
19:33:18FromDiscord<Elegantbeef> One day I'll make it accessible on my website
19:35:32FromDiscord<funnybananas> @deech4592 Just tested, no change in behavior
19:37:28FromDiscord<funnybananas> that project contains an example binary that if I build from source, continues to work. However, if I remove the "import nimview" line from the example binary, it gets flagged by AV. Very bizzare and not really sure how to debug
19:37:56FromDiscord<funnybananas> Ruled out mingw versions, attempted older and newer Nim versions both release, nightly, and built from src
20:25:32*derpydoo quit (Ping timeout: 240 seconds)
20:34:36*azimut_ joined #nim
20:37:44*azimut quit (Ping timeout: 240 seconds)
21:09:45*ntat quit (Quit: Leaving)
21:22:31*oldpcuser quit (Ping timeout: 260 seconds)
21:29:51*derpydoo joined #nim
22:13:59FromDiscord<deech4592> In reply to @funnybananas "<@271498588981297157> Just tested, no": I'm honestly not sure how to approach this because I don't know what AV uses to flag a binary. I wonder if it still happens if you use `vcc` to compile it.
22:20:39FromDiscord<Elegantbeef> mingw does seem to raise more false positives, but generally speaking it seems all nim software can be flagged
22:21:57FromDiscord<michaelb.eth> what about the clang toolchain for MSYS2 rather than the gcc one?
22:22:03FromDiscord<michaelb.eth> (edit) "what about ... the" added "using"
22:22:17FromDiscord<michaelb.eth> I'd be curious if that makes a difference
22:22:59FromDiscord<Elegantbeef> Could try zigcc in that case, no clue what the host machine is here
22:23:02FromDiscord<michaelb.eth> it certainly makes a difference in the performance of multithreaded Nim code running on Windows
22:23:15FromDiscord<michaelb.eth> (edit) "it" => "clang toolchain"
22:26:16FromDiscord<michaelb.eth> anyway, point being that it's producing a binary that's not exactly the same as the one produced with mingw gcc, and maybe it's different enough to get around whatever it is the virus scanner is looking for
22:32:25*xet7 joined #nim
22:44:09*oldpcuser joined #nim
23:47:02FromDiscord<JJ> sent a code paste, see https://play.nim-lang.org/#ix=4yIq
23:47:03FromDiscord<JJ> sent a code paste, see https://paste.rs/xKZe6
23:47:23FromDiscord<Elegantbeef> `var Self`
23:47:24FromDiscord<JJ> (edit) "https://play.nim-lang.org/#ix=4yIs" => "https://play.nim-lang.org/#ix=4yIr"
23:47:49FromDiscord<JJ> (edit) "https://play.nim-lang.org/#ix=4yIr" => "https://play.nim-lang.org/#ix=4yIt"
23:47:49FromDiscord<Elegantbeef> You do not have any procedures implemented for `step`
23:47:52FromDiscord<Elegantbeef> I mean `Program`
23:51:06FromDiscord<JJ> i fixed some things in my example paste
23:51:33FromDiscord<JJ> does var Program mean program.main is not mutable?
23:52:23FromDiscord<Elegantbeef> Oh whoops didnt realise wehre the mismatch was
23:52:26FromDiscord<Elegantbeef> No it's mutable
23:55:44FromDiscord<JJ> well i'm confused: i was hoping the struct of concepts meant that nim could tell that program.main is of type MainMemory and accordingly must have that get function defined on it
23:57:56FromDiscord<JJ> oh. never mind, i see why this doesn't work.
23:59:54FromDiscord<Elegantbeef> `MainMemory` is a generic