<< 25-05-2024 >>
00:00:40FromDiscord<Robyn [She/Her]> Fair
00:01:28*beholders_eye quit (Read error: Connection reset by peer)
00:03:42FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=dphLlISJ
00:04:01FromDiscord<Elegantbeef> To me enum branches are existential types and not just runtime types so they should exist and be operated on
00:07:09FromDiscord<Robyn [She/Her]> Aah okay I understand that now
00:07:41*beholders_eye joined #nim
00:08:14FromDiscord<Robyn [She/Her]> Though, does that require explicit conversion or do converters handle the conversion in a nice way?
00:09:03FromDiscord<Robyn [She/Her]> In reply to @chronos.vitaqua "Though, does that require": Seems like converters handle it nicely
00:09:08FromDiscord<Elegantbeef> There is upcasting to `Shape` but no down casting
00:09:20FromDiscord<Elegantbeef> Shape cannot be safely downcasted it's a type narrowing operation
00:09:49FromDiscord<Elegantbeef> All circles are shapes but not all shapes are circles
00:10:21FromDiscord<Robyn [She/Her]> I don't think I understand? Wouldn't you just rely on the `kind` stored in the defined sum type to handle the converter code?
00:10:32FromDiscord<Elegantbeef> Converters are statically dispatched
00:10:41FromDiscord<Robyn [She/Her]> Ah
00:11:22FromDiscord<Robyn [She/Her]> Then, the user would have to explicitly do `Shape.Circle` if they wanted to pass a circle to `area`?
00:11:27FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=mDqJQBmm
00:11:29FromDiscord<Elegantbeef> Whoops
00:11:35FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=dvJakDrR
00:11:42FromDiscord<Elegantbeef> Well that's where an `unpack` would come in
00:11:45FromDiscord<Robyn [She/Her]> (edit) "Then, the user would have to explicitly do `Shape.Circle` if they wanted to pass a circle ... to" added "that was upcasted to a shape"
00:12:14FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=TdJbBkBe
00:12:31FromDiscord<Elegantbeef> So you'd do like `shape.unpackIt(it.area())`
00:12:53FromDiscord<Robyn [She/Her]> Yeah okay that makes sense, I'm assuming the only reason it isn't implemented is because you don't use fungus often? Or just didn't think of it
00:13:23FromDiscord<Elegantbeef> Just never got to it
00:13:53FromDiscord<Robyn [She/Her]> Makes sense, I might add this feature if I procrastinate my project for longer
00:13:56FromDiscord<Elegantbeef> https://forum.nim-lang.org/t/10836 has a bunch of ramblings from me
00:14:08FromDiscord<Elegantbeef> Specificaally https://forum.nim-lang.org/t/10836#72416
00:15:05FromDiscord<Robyn [She/Her]> In reply to @Elegantbeef "https://forum.nim-lang.org/t/10836 has a bunch": Oh wow I actually like that proposed syntax a lot
00:15:31FromDiscord<Elegantbeef> It's literally fungus but inside a typesection 😄
00:17:18FromDiscord<Robyn [She/Her]> Yeah exactly!
00:17:49FromDiscord<Robyn [She/Her]> Why isn't it just called `sum` instead of `case` or `enum`?
00:18:23FromDiscord<Elegantbeef> No clue, but sum is a common word
00:18:32FromDiscord<Elegantbeef> `case` is less of a common word and is already a keyword
00:18:58*greaser|q quit (Changing host)
00:18:58*greaser|q joined #nim
00:19:00*greaser|q is now known as GreaseMonkey
00:19:31FromDiscord<Robyn [She/Her]> That's a good point
00:20:20FromDiscord<Elegantbeef> Hey I pretend to know some things about design
00:38:52FromDiscord<Robyn [She/Her]> Posted a new comment in the thread
01:07:24*def- quit (Quit: -)
01:08:41*def- joined #nim
02:03:41*beholders_eye quit (Ping timeout: 240 seconds)
02:30:54*krux02 quit (Remote host closed the connection)
04:00:55FromDiscord<morgan> In reply to @Elegantbeef "<@277133333773811712> are you attempting": i set up tuples with math operations as 2d points and using them to index into an array object (bounds plus 1d ptr uncheckedarray)
04:05:50FromDiscord<morgan> In reply to @Elegantbeef "https://forum.nim-lang.org/t/10836 has a bunch": what does this actually gain over variant types? it certainly isn’t a replacement as it imposes one axis of variability
04:33:22*rockcavera quit (Remote host closed the connection)
04:35:05*rockcavera joined #nim
04:36:31FromDiscord<Elegantbeef> @morgan It's a more structured approach to variant types
04:38:26FromDiscord<danielsokil> Hi all
04:38:41FromDiscord<danielsokil> sent a code paste, see https://play.nim-lang.org/#pasty=tSBUdvff
04:39:02FromDiscord<danielsokil> sent a code paste, see https://play.nim-lang.org/#pasty=iQyiHnLu
04:39:19FromDiscord<Elegantbeef> You're not invoking `check`
04:39:23FromDiscord<Elegantbeef> You're attempting to add ap roc
04:39:35FromDiscord<danielsokil> Yes, silly me, lol just noticed
04:40:25FromDiscord<danielsokil> In reply to @Elegantbeef "You're not invoking `check`": Thanks (;
04:42:27FromDiscord<danielsokil> What is Nims alternative to Go coroutines (goroutines)?
04:42:40FromDiscord<danielsokil> From my understanding, async is single threaded?
04:42:43FromDiscord<Elegantbeef> It does not have a good threading story
04:43:34FromDiscord<danielsokil> Can I do async across multiple threads, safely?
04:43:49FromDiscord<Elegantbeef> Nope futures must be completed on the thread that made them
04:45:19FromDiscord<Elegantbeef> So if you want multi threaded async you practically need to send across channels what the thread needs to create a request
04:46:40FromDiscord<danielsokil> And, these are lightweight threads right? runtime managed
04:46:58FromDiscord<Elegantbeef> No they're proper threads
04:52:10FromDiscord<danielsokil> In this example: https://nim-by-example.github.io/parallelism/↵↵`createThread` creates a system thread?
04:52:26FromDiscord<Elegantbeef> Yes
04:57:34FromDiscord<nnsee> In reply to @Elegantbeef "You're not invoking `check`": am i missing messages here? to me it looks like he just said hi and you already knew the solution to his problem
04:57:58FromDiscord<nnsee> https://media.discordapp.net/attachments/371759389889003532/1243790140497858600/Screenshot_20240525-075736_Firefox2.png?ex=6652c155&is=66516fd5&hm=36d9698d12a35ec8136e44c7aa5f9ae163fe8733bd0be564a58f44ed885e2cad&
04:58:06FromDiscord<nnsee> something like this happened yesterday as well
04:58:14FromDiscord<Elegantbeef> They did the wrong thing and they deleted their messages
04:58:19FromDiscord<nnsee> oh
04:58:20FromDiscord<Elegantbeef> People are bad people
05:00:22FromDiscord<danielsokil> 😝
05:01:31FromDiscord<nnsee> some day i _will_ catch beef in the act of clairvoyance
05:02:35FromDiscord<danielsokil> What is the default GC in Nim2?
05:02:44FromDiscord<Elegantbeef> No matter how many crystals I put in my pooper I never get any magic
05:02:54FromDiscord<Elegantbeef> Nim 2.0 uses Orc
05:03:11FromDiscord<Elegantbeef> Which is move semantics with destructors and a mark and sweep cycle collector
05:06:57FromDiscord<nnsee> arc is just orc without the mark & sweep, right?
05:07:34FromDiscord<odexine> pretty much
05:08:20FromDiscord<danielsokil> sent a code paste, see https://play.nim-lang.org/#pasty=czvJJleu
05:08:49FromDiscord<Elegantbeef> You need to do `let ipAddress = ipAddress` before to remove the `lent` and copy it
05:08:57FromDiscord<Elegantbeef> It's an annoying part about `lent T`
05:09:44FromDiscord<danielsokil> what is a lent string?
05:09:48FromDiscord<danielsokil> In there a lender?
05:10:07FromDiscord<morgan> In reply to @Elegantbeef "<@277133333773811712> It's a more": ah ok
05:10:12FromDiscord<danielsokil> (edit) "In" => "Is"
05:10:45FromDiscord<Elegantbeef> `lent` is a non copy reference
05:11:07FromDiscord<danielsokil> Okay, a pointer
05:11:16FromDiscord<Elegantbeef> Well no
05:11:19FromDiscord<Elegantbeef> Cause we have pointers 😄
05:12:10FromDiscord<Elegantbeef> But yes the implementation is that it's a safe pointer
05:12:20FromDiscord<danielsokil> In reply to @Elegantbeef "You need to do": So with this, we can explicitly copy
05:12:54FromDiscord<Elegantbeef> Well a copy is made whenever you make `lent` a "R value"
05:13:33FromDiscord<danielsokil> Could I get the underlying pointer to `ipAddress`?
05:13:58FromDiscord<Elegantbeef> `addr` does that
05:15:54FromDiscord<danielsokil> What about pntr value?
05:16:37FromDiscord<Elegantbeef> Huh?
05:18:02FromDiscord<danielsokil> Ah, `repr` gets it done
05:18:30FromDiscord<Elegantbeef> Ah do you mean "how does one dereference"? 😄
05:19:48FromDiscord<danielsokil> Yes
05:20:03FromDiscord<Elegantbeef> `[]`
05:20:50FromDiscord<danielsokil> sent a code paste, see https://play.nim-lang.org/#pasty=kDLMewzz
05:20:58FromDiscord<morgan> ptr and ref are pointers (ref is just managed). lent afaik is an annotation to handle compilation
05:21:34FromDiscord<Elegantbeef> Well we have references of three variety and they're all type annotations
05:22:12FromDiscord<Elegantbeef> `lent T` is valid only as a return type, `var T` is valid as parameters and return type. `ref T` is a heap allocated `T` that is automatically managed
05:22:42FromDiscord<morgan> ah interesting i didn't know lent was only as returns
05:24:04FromDiscord<danielsokil> Do we have a decent AES lib, specifically one that supports ECB mode
05:25:04FromDiscord<danielsokil> Block padding, etc
05:29:51FromDiscord<morgan> In reply to @morganalyssa "i set up tuples": ok so copying the procs into the file makes it compile
05:34:31*SchweinDeBurg quit (Quit: WeeChat 4.3.0-dev)
06:00:39*SchweinDeBurg joined #nim
06:30:46*beholders_eye joined #nim
06:35:32*beholders_eye quit (Ping timeout: 268 seconds)
06:40:09FromDiscord<codic> is it possible to stop nimble from checking all the dependnceies of a project upon compilation
06:40:17FromDiscord<codic> and printing them
06:46:39FromDiscord<kots> `nimble setup` and then use `nim` directly
06:46:46FromDiscord<kots> nimble users might be able to give you a "better" answer
07:52:48*derpydoo joined #nim
08:01:53FromDiscord<dawidek.2137> nim doc about bind statement:↵"But a bind is rarely useful because symbol binding from the definition scope is the default."↵↵So when is it actually useful if it's the default?
08:04:42FromDiscord<dawidek.2137> also I don't see how it's the default... I need to import `tables` to use my template that calls `tables.contains` from another file, so I'm getting confused
08:07:21FromDiscord<dawidek.2137> ok, I see, the compiler only chimps out if I call `notin` which implicitly refers to `tables.contains`, if I call `contains` directly there is no problem
08:07:32FromDiscord<dawidek.2137> is it necessary to avoid notin etc?
08:11:01FromDiscord<Elegantbeef> `bind` is default when there is no overloaded symbol
08:12:54FromDiscord<dawidek.2137> makes sense, but what about the `notin` problem?
08:13:31FromDiscord<dawidek.2137> tables.`[]=` also doesn't get picked up in my template call unless I import tables in the file where the template gets called
08:13:46FromDiscord<Elegantbeef> `mixin`
08:14:05FromDiscord<Elegantbeef> Templates are expanded late so you need to either use `contains` or `mixin`
08:15:57FromDiscord<dawidek.2137> doesn't mixin do the opposite of what I want?
08:16:45FromDiscord<dawidek.2137> which is to bind notin to whatever it was defined as in the file where the template was defiend
08:18:05FromDiscord<Elegantbeef> You do not want to bind `notin` you want to bind `contains`
08:18:22FromDiscord<Elegantbeef> `mixin` accepts both scope of declaration and instantiation scope
08:18:52FromDiscord<Elegantbeef> `notin` is a template that does `not(contains(x, y))`
08:19:06FromDiscord<Elegantbeef> Though it's untyped so should be exanded at declaration
08:25:17FromDiscord<dawidek.2137> sent a code paste, see https://play.nim-lang.org/#pasty=AGKRKrJu
08:25:30FromDiscord<dawidek.2137> also `mixin contains` etc
08:25:53FromDiscord<Elegantbeef> Do you have a min repro?
08:26:59FromDiscord<dawidek.2137> I'll try to make one
08:28:59FromDiscord<dawidek.2137> sent a code paste, see https://play.nim-lang.org/#pasty=ftezKBQj
08:30:25FromDiscord<Elegantbeef> Oh you're using a template
08:31:46FromDiscord<Elegantbeef> Because this is a template it does not expand `notin` until it's instantiated, as such your `bind` or `mixin` contains goes on deaf ears
08:32:48FromDiscord<Elegantbeef> Symbol resolution and symbol opening is a fun thing with templates
08:38:52FromDiscord<dawidek.2137> 💩
08:40:07FromDiscord<Elegantbeef> I can only say "Only use a template where it makes sense"
09:36:13*derpydoo quit (Ping timeout: 268 seconds)
09:53:35*ntat joined #nim
10:21:58*derpydoo joined #nim
10:25:33*ntat quit (Quit: Leaving)
11:15:35*def- quit (Quit: -)
11:15:52*def- joined #nim
11:27:31*derpydoo quit (Ping timeout: 260 seconds)
11:30:21FromDiscord<tauruuuuuus> Soooo I'm not sure I'm doing something wrong there, but it seems like I can't evaluate untyped stuff in macros at compile time? Is it right?
11:33:21*def- quit (Quit: -)
11:34:35*def- joined #nim
11:34:57*dtomato quit (Quit: The Lounge - https://thelounge.chat)
11:35:26*dtomato joined #nim
11:37:14FromDiscord<dawidek.2137> I think so
11:38:42FromDiscord<dawidek.2137> (besides "evaluating" it in a quote as a compile time expression, but you obviously can't get the result as anything but a NimNode)
11:39:43*def- quit (Quit: -)
11:40:12*def- joined #nim
11:48:56FromDiscord<ninovanhooff> Is float multiplication faster, slower or equal to float division? ie. is there a performance difference between `a0.5` and `a/2.0`?
11:51:24FromDiscord<odexine> mult is faster but often is not the main bottleneck
12:01:59*def- quit (Quit: -)
12:04:03FromDiscord<Phil> In reply to @tauruuuuuus "Soooo I'm not sure": Define "evaluating".↵If you want to access the types etc. you'll need to use typed
12:04:07FromDiscord<Phil> (edit) removed "the"
12:21:49*def- joined #nim
12:29:29*dtomato quit (Quit: The Lounge - https://thelounge.chat)
12:34:37*dtomato joined #nim
13:07:23FromDiscord<dmisener> Fungus
13:08:24FromDiscord<dmisener> Please ignore above - finger fumble
13:23:22*def- quit (Quit: -)
13:24:24*def- joined #nim
13:47:00*dtomato quit (Quit: The Lounge - https://thelounge.chat)
13:47:28*dtomato joined #nim
14:19:11*beholders_eye joined #nim
14:19:52FromDiscord<System64 ~ Flandre Scarlet> Is it possible to get Nim autocompletion, errors and so on in neovim?
14:20:55FromDiscord<kots> Yes, I use nimlangserver with nvim-lspconfig
14:24:55FromDiscord<Robyn [She/Her]> In reply to @dmisener "Fungus": Fungus is a good Nim library :)
14:37:47*beholders_eye quit (Ping timeout: 264 seconds)
14:43:02*lucasta joined #nim
15:06:17FromDiscord<goerge_lsd> http://marcio.io/2015/07/calculating-multiple-file-hashes-in-a-single-pass/
15:06:35FromDiscord<goerge_lsd> Can this be easily done in nim ? Calculate various checksums in a single pass
15:07:05FromDiscord<dawidek.2137> silly idea, but can I magically make my enums be backed up by strings on js backend without modifying the compiler? 🤫
15:07:23FromDiscord<dawidek.2137> i want to more easily see what's up when printing objects in browser console
15:07:32FromDiscord<Robyn [She/Her]> In reply to @dawidek.2137 "silly idea, but can": What do you mean backed up by strings?
15:07:36FromDiscord<Robyn [She/Her]> Do you mean a string representation for it?
15:07:42FromDiscord<dawidek.2137> they are numbers in js after compilation
15:07:52FromDiscord<dawidek.2137> when you console.log from the console etc
15:09:15FromDiscord<dawidek.2137> I know I can use use typed constants to fake enum semantics etc but then I lose exhaustiveness checks in case statements
15:13:02FromDiscord<juancarlospaco> Enum can have string values.
15:14:17FromDiscord<juancarlospaco> (edit) "values." => "values, or make a $ for it."
15:14:29FromDiscord<goerge_lsd> why is md5 proc for string https://nim-lang.org/docs/md5.html#getMD5%2Cstring while sha1 is for openArray[char] ? https://nim-lang.org/docs/sha1.html#secureHash%2CopenArray%5Bchar%5D
15:15:36FromDiscord<juancarlospaco> openarray char takes string too.
15:15:39FromDiscord<goerge_lsd> Didn't these docs used to have links to proc source on github ?
15:15:47FromDiscord<goerge_lsd> i know, but why the inconsistency
15:16:36FromDiscord<juancarlospaco> both take string.
15:17:32FromDiscord<juancarlospaco> openarray sometimes can be used for some optimizations in the code.
15:17:44FromDiscord<goerge_lsd> @juancarlospaco I am looking to see if you can create various checksums of a file in one pass, without reaing the file from disk many times
15:18:06FromDiscord<goerge_lsd> found an example in go, somehow that is possible there, because of the common interface used by the checksum code
15:18:25FromDiscord<goerge_lsd> (edit) "reaing" => "reading"
15:18:51FromDiscord<Robyn [She/Her]> In reply to @juancarlospaco "Enum can have string": That doesn't seem to implement it for the JS backend, which is what dawidek wants
15:18:59FromDiscord<goerge_lsd> I got an idea of a project, to make an overlay fs-like interface, to access file on disk based on their checksum, instead of path
15:19:18FromDiscord<Robyn [She/Her]> @dawidek.2137 you can emit code for the JS backend specifically if wanted, why not try that?
15:19:20FromDiscord<goerge_lsd> so you can move things aroun and your app will still work
15:19:38FromDiscord<goerge_lsd> (edit) "aroun" => "around" | "aroundand your app will still work ... " added "(if it uses such an interface)"
15:19:55FromDiscord<juancarlospaco> In reply to @chronos.vitaqua "That doesn't seem to": You can use cstring if you want to see the actual literal string, otherwise is an array of int in the JS.
15:20:12FromDiscord<goerge_lsd> guess I need to wait for the bot with dog head avatar to come back : P
15:20:13FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "I got an idea": Issue with that is obviously files can change
15:20:27FromDiscord<Robyn [She/Her]> In reply to @juancarlospaco "You can use cstring": Ah that is a good idea yeah
15:20:57FromDiscord<goerge_lsd> @Robyn [She/Her] of course, it would need to be run as a deamon, to record when last change was saved, use inotify to track changes on fs etc
15:21:53FromDiscord<goerge_lsd> but for what I want to use it, files won't really change. Sort of like an IPFS storage, content addressable. Only it would also work locally
15:22:23FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "<@524288464422830095> of course, it": Fair enough, but I still don't see a practical application of it, since if you're using inotify anyway, you can simply map files to a virtual path using fuse
15:22:52FromDiscord<goerge_lsd> I don't want to use FUSE though, it's not really portable to windows
15:23:01FromDiscord<Robyn [She/Her]> Ah it isn't? Didn't know that
15:23:09FromDiscord<goerge_lsd> there's some fuse-like thing on windows.. dokan ?
15:23:23FromDiscord<Robyn [She/Her]> Either way, it can still be done without fuse with a custom API that you would've implemented anyway
15:23:28FromDiscord<goerge_lsd> although i'd like to look into your idea as well, didn't consider this
15:23:58FromDiscord<goerge_lsd> but let's say daemon didn't run while some change to FS was done, and you renamed a file..
15:24:16FromDiscord<goerge_lsd> can you still access that file based on a known.. file node id or whatever it's called
15:24:19*beholders_eye joined #nim
15:24:30FromDiscord<goerge_lsd> inode
15:25:19FromDiscord<Robyn [She/Her]> I have no idea, but if you could, wouldn't that still make a checksum-based API for file lookup useless?
15:26:04FromDiscord<goerge_lsd> I guess it would for this use case
15:26:11FromDiscord<goerge_lsd> well, no, it wouldn't
15:26:23FromDiscord<goerge_lsd> you'd ideally want to map known files with already existing stuff on your disk
15:26:40FromDiscord<saint.___.> In reply to @sys64 "Is it possible to": Ya you can! Although it's not the best atm
15:26:58FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "you'd ideally want to": Fair enough
15:36:20*beholders_eye quit (Read error: Connection reset by peer)
15:36:29*def- quit (Quit: -)
15:38:07FromDiscord<goerge_lsd> so back to the original issue, any design ideas on how to make a single pass file read, multiple checksums compute ?
15:40:56FromDiscord<goerge_lsd> looks like checksum algos can be run sequantlly on the input data?
15:41:11FromDiscord<goerge_lsd> https://github.com/nim-lang/checksums/blob/7ff0b762332d2591bbeb65df9bb86d52ea44ec01/src/checksums/md5.nim#L312
15:41:51*beholders_eye joined #nim
15:43:28FromDiscord<goerge_lsd> `proc md5Update(c: var MD5Context, input: openArray[uint8]) =`↵`proc update(ctx: var Sha1State, data: openArray[char]) =`
15:43:57FromDiscord<goerge_lsd> naming inconsistencies, data type inconsistencies.. or is `openArray[uint8]` perfectly equivalent to `openArray[char]`? don't think so
15:47:45FromDiscord<threefour> I imagine they'd use the 8 bits the same way though.
15:49:15FromDiscord<System64 ~ Flandre Scarlet> In reply to @saint.___. "Ya you can! Although": And does it still work well?
16:13:28*def- joined #nim
16:15:02FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=jBJIZLdg
16:15:48FromDiscord<goerge_lsd> kinda ambiguous if you wanna import it
16:18:50FromDiscord<goerge_lsd> can you rename it somehow ?
16:20:39FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "naming inconsistencies, data type": A char is just a byte, and a byte is just a uint8 in Nim
16:21:33FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=lwJOWgEx
16:21:40FromDiscord<goerge_lsd> this works, but can it be done in a shorter way ?
16:21:49FromDiscord<goerge_lsd> not needing a separate import maybe
16:23:33FromDiscord<juancarlospaco> In reply to @goerge_lsd "naming inconsistencies, data type": https://github.com/nim-lang/Nim/blob/daad06bd07ff11cc2b0c74b604ff82899e923d59/lib/system/basic_types.nim#L22
16:24:44FromDiscord<goerge_lsd> anyway, how would you suggest defining an object that can hold various hash digests/checksums ?
16:26:58FromDiscord<juancarlospaco> Use a simpler hash for file checksums, like just a CRC.
16:27:29FromDiscord<goerge_lsd> bcrypt digest is array of uint8, md5 is uint8, sha1 is uint8, sha2 is array of char, sha3 also array of char
16:27:31FromDiscord<juancarlospaco> Implement with that, then you can scale up complexities.
16:27:44FromDiscord<goerge_lsd> @juancarlospaco thanks but that's not what I want to implement
16:28:28FromDiscord<goerge_lsd> I want to make a small lib that can hash different algos in one run
16:35:09*def- quit (Quit: -)
16:35:37*lucasta quit (Quit: Leaving)
16:37:22FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=QQfNvnas
16:37:40FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=eFPWVKrv
16:37:52FromDiscord<goerge_lsd> I think this would need the object holding the value to be generic, no ?
16:53:24FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=grTFOmeJ
16:53:45FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "anyway, how would you": You could use a case object if wanted
16:54:07FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "bcrypt digest is array": Why not just store them all as `seq[byte]`?
16:54:26*def- joined #nim
16:54:47FromDiscord<goerge_lsd> sent a code paste, see https://paste.rs/hGgDx
16:55:02FromDiscord<goerge_lsd> (edit) "https://paste.rs/sIGMW" => "https://play.nim-lang.org/#pasty=jemzHsUg"
16:55:45FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=wPnuibSo
16:56:38FromDiscord<goerge_lsd> In reply to @chronos.vitaqua "Why not just store": I was just showing how the code is in the checksums module.. very inconsistent naming
16:56:43FromDiscord<goerge_lsd> (edit) "In reply to @chronos.vitaqua "Why not just store": I was just showing how the code is in the checksums module.. very inconsistent naming ... " added "and types used"
16:57:06FromDiscord<goerge_lsd> (edit) "used" => "used. Maybe different devs worked on it and there was no consensus on the common interface"
16:57:46FromDiscord<Robyn [She/Her]> That could be a possibility, yeah
16:57:58FromDiscord<Robyn [She/Her]> You could make your own wrapper over those interfaces, perhaps?
16:58:31FromDiscord<goerge_lsd> not a big problem at the moment, I think
16:59:04FromDiscord<goerge_lsd> In reply to @chronos.vitaqua "You could use a": yes but I can have an arbitrary numbers of types of checksums stored there, variant objects don't help here ?
16:59:24FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=TviSTpJV
17:00:15FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "yes but I can": Not sure what you mean, I don't use hashing algorithms so I don't know the specifics besides different types being annoying
17:00:54FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=YqVHxzBX
17:01:57FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=kxuLKlyJ
17:04:46FromDiscord<goerge_lsd> well, I guess, but that doesn't help my problem.. how to store arbitrary number of them
17:05:38FromDiscord<goerge_lsd> also how to "cast" a `distinct string` as `string`?
17:05:51FromDiscord<goerge_lsd> `Path` type doesn't have a `$` proc, so annoying
17:07:33FromDiscord<goerge_lsd> repr ?
17:08:34FromDiscord<odexine> `string`
17:12:14FromDiscord<goerge_lsd> maybe I should use a table instead of object, to store `key: HashType`, `value: seq[byte]` ?
17:13:03FromDiscord<goerge_lsd> although that is kind of overkill for jus 11 possible key values
17:14:00FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=qjPZPTxb
17:14:36*Guest8054 joined #nim
17:14:49*def- quit (Quit: -)
17:15:02*def-- joined #nim
17:15:26*def-- is now known as def-
17:20:05FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=zaAwQbfe
17:23:13*xutaxkamay quit (Quit: ZNC 1.8.2+deb3.1 - https://znc.in)
17:29:41FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=QUqcDKaO
17:29:58FromDiscord<@@prestosilver> got a stupid question, are all refs the same size (like pointers) or do they store more than just a pointer and gc stuff
17:30:16FromDiscord<@@prestosilver> like is it safe to do something like this
17:31:31FromDiscord<griffith1deadly> In reply to @prestosilver "got a stupid question,": ref itself just pointer, but managed
17:31:42FromDiscord<griffith1deadly> so sizeof(ref) == sizeof(ptr)
17:31:44FromDiscord<@@prestosilver> o ok
17:32:07FromDiscord<@@prestosilver> good thing you knew that bc idk how to write the example i was gonna give lmao
17:32:21FromDiscord<Robyn [She/Her]> In reply to @griffith1deadly "so sizeof(ref) == sizeof(ptr)": It's `pointer` not `ptr` but yeah
17:32:47FromDiscord<griffith1deadly> In reply to @chronos.vitaqua "It's `pointer` not `ptr`": in that context u also can't use ref
17:32:53FromDiscord<griffith1deadly> (edit) "ref" => "`ref`"
17:33:00FromDiscord<@@prestosilver> essentially i just wanted make an enum array of refs, but i need to mix types is that possible even lol
17:33:13FromDiscord<@@prestosilver> without inheratance bc that would break stuff in this contexrt
17:33:16FromDiscord<@@prestosilver> (edit) "contexrt" => "context"
17:33:36FromDiscord<@@prestosilver> like mixed refs
17:33:56FromDiscord<@@prestosilver> eh might be cleaner to just use an object, ignore that
17:35:12FromDiscord<Robyn [She/Her]> In reply to @griffith1deadly "in that context u": True, it'd be `ref object`
17:35:36FromDiscord<Robyn [She/Her]> In reply to @prestosilver "essentially i just wanted": Enum array of refs? What do you mean?
17:35:53FromDiscord<Robyn [She/Her]> In reply to @prestosilver "eh might be cleaner": Have you looked at object variants? Might fit the usecase
17:36:30FromDiscord<griffith1deadly> i love Nim so much that I use it to write the backend for my thesis, despite all the pain of using nimsuggest :nim1:
17:37:25FromDiscord<Robyn [She/Her]> Nim is a great language :)
17:39:01FromDiscord<@@prestosilver> yea, its hard to explain without just explaining the whole thing, basically writing a hardcoded plugin management library (essentially dlc but for programs) and I wanted a way of storing a global reference to each plugin (a .so/.dll library) instance in the main program, I already have a enum with each plugin id but I need a better way of storing the refs to each plugin.
17:39:17FromDiscord<@@prestosilver> I think inheratance might endup being the way to go
17:39:26FromDiscord<@@prestosilver> (edit) "endup being" => "have to be"
17:40:19FromDiscord<Robyn [She/Her]> In reply to @prestosilver "yea, its hard to": A table of ref objects could work fine here as long as you remember to cast them to the appropriate type for each object
17:40:45FromDiscord<@@prestosilver> is ref object a valid thing outside of object definitions
17:40:51FromDiscord<Robyn [She/Her]> You can store `ref object` I believe, not sure how safe it is though
17:41:03FromDiscord<Robyn [She/Her]> In reply to @prestosilver "is ref object a": Yeah it is, it's used in generics often, for example
17:41:09FromDiscord<@@prestosilver> ah
17:41:12FromDiscord<@@prestosilver> might try that
17:41:21FromDiscord<Robyn [She/Her]> Good luck!
17:41:36FromDiscord<griffith1deadly> In reply to @chronos.vitaqua "Nim is a great": but nimsuggest..
17:41:42FromDiscord<@@prestosilver> In reply to @chronos.vitaqua "Good luck!": lol im gonna need it
17:42:06FromDiscord<mikaelashipha> Hey guys, I'm kinda new to the language and wanted to start with it, but having trouble of finding text editor for it
17:42:07FromDiscord<griffith1deadly> In reply to @griffith1deadly "but nimsuggest..": even vs intellisence better
17:42:11FromDiscord<Robyn [She/Her]> In reply to @griffith1deadly "but nimsuggest..": Yeah it probably needs a rewrite
17:42:24FromDiscord<Robyn [She/Her]> In reply to @mikaelashipha "Hey guys, I'm kinda": Have you tried Visual Studio Code? It's my go to
17:42:40FromDiscord<mikaelashipha> Yea, no.. anything but vs code
17:42:43FromDiscord<griffith1deadly> visual studio code & sublime text
17:43:06FromDiscord<mikaelashipha> I checked sublime lsp language servers and didn't saw nim there
17:43:14FromDiscord<mikaelashipha> Is there support for it in sublime?
17:43:18FromDiscord<abaer> vscodium is what i use.
17:43:36FromDiscord<Robyn [She/Her]> In reply to @mikaelashipha "Is there support for": There's Nimline but idk how up-to-date it is
17:44:34FromDiscord<mikaelashipha> So currently it's only vs code?
17:44:46FromDiscord<griffith1deadly> https://github.com/nim-lang/langserver?tab=readme-ov-file#sublime-text
17:45:41FromDiscord<Robyn [She/Her]> In reply to @mikaelashipha "So currently it's only": There's plugins for multiple editors ofc, but afaik most here use VSC↵Here's the official Sublime plugin for Nim: https://github.com/nim-lang/NimLime
17:46:14FromDiscord<abaer> Pretty sure vscodium also has an extension for nim as well.
17:46:56FromDiscord<Robyn [She/Her]> In reply to @abaer "Pretty sure vscodium also": It's the same one used for VSC, it's just not on the OpenVSX registry
17:46:56FromDiscord<mikaelashipha> In reply to @chronos.vitaqua "There's plugins for multiple": Hmmm, for some reason the sublime lsp language support list didn't mention this, well give it a shot
17:47:26FromDiscord<Robyn [She/Her]> Good luck!
17:47:35*def- quit (Quit: -)
17:47:46*def- joined #nim
17:47:49FromDiscord<abaer> Yeah. Hate the damn ide but geany lacks most functionality to sanely manage and code large projects.↵(@Robyn [She/Her])
17:48:01FromDiscord<mikaelashipha> I just came here base on people's recommendation about cli languages and it seems nim is suitable one too
17:48:50FromDiscord<mikaelashipha> Im really into cli programming and I don't really like the popular options and rather to use less popular ones and less known languages
17:49:05FromDiscord<Robyn [She/Her]> In reply to @abaer "Yeah. Hate the damn": That's unfortunate, oh well
17:49:24FromDiscord<Robyn [She/Her]> In reply to @mikaelashipha "Im really into cli": Understandable, you may find libraries like cligen pretty nice in that case
17:50:00FromDiscord<Robyn [She/Her]> https://github.com/c-blake/cligen
17:50:06FromDiscord<nnsee> In reply to @mikaelashipha "Im really into cli": welcome!
17:50:13FromDiscord<Robyn [She/Her]> There's also argparse and parseopt
17:50:36FromDiscord<nnsee> Nim is very powerful for cli programming, I highly recommend cligen as well
17:51:48FromDiscord<mikaelashipha> Before that, while u guys are here might I ask them right now, nim is compiled language right? It produces binary and executable?
17:52:06FromDiscord<mikaelashipha> Also does it have an option for cross compiling like go and zig?
17:52:12FromDiscord<sOkam! 🫐> In reply to @mikaelashipha "Before that, while u": yes
17:52:14FromDiscord<nnsee> yes and yes
17:52:17FromDiscord<griffith1deadly> In reply to @mikaelashipha "Also does it have": u can use zigcc as compiler
17:52:23FromDiscord<mikaelashipha> Or it's more of traditional approach of compiling on target os?
17:52:26FromDiscord<sOkam! 🫐> In reply to @mikaelashipha "Also does it have": yes, but tricky. I created `confy` for that reason, if you want to take a look
17:52:29FromDiscord<sOkam! 🫐> it cross compiles easily
17:53:10FromDiscord<mikaelashipha> As long as it makes the cross compiling easy and I can simply produce windows binary or Mac or vice versa
17:53:12FromDiscord<sOkam! 🫐> confy uses the `zigcc` package, but automates the process. otherwise you can also use zigcc in your nimble file
17:53:25FromDiscord<sOkam! 🫐> In reply to @mikaelashipha "As long as it": yes, definitely doable. I do it all the time
17:53:58FromDiscord<mikaelashipha> In reply to @mikaelashipha "Or it's more of": I mean windows binary on Mac and vice versa
17:54:29FromDiscord<sOkam! 🫐> In reply to @mikaelashipha "Or it's more of": it uses GCC or CLANG, and it provides tools to define the CPU and OS. but then you need a backend to cross compile the resulting backend easily↵that is what the zigcc package simplifies, as you may know from zig
17:54:42FromDiscord<mikaelashipha> In reply to @heysokam "yes, definitely doable. I": Then I guess I'm officially a nim user
17:54:53FromDiscord<sOkam! 🫐> so definitely not the "traditional" way you might be thinking from C
17:55:39FromDiscord<sOkam! 🫐> In reply to @mikaelashipha "Then I guess I'm": welcome aboard! nim is great for many many reasons. you will see for yourself 🙂
17:56:14FromDiscord<sOkam! 🫐> the cleanliness of the syntax, even with big messy codebases, is just 🤤
17:56:58FromDiscord<mikaelashipha> In reply to @heysokam "the cleanliness of the": I will be the judge of that 😛
17:57:13FromDiscord<sOkam! 🫐> fair point 🙂
17:57:25FromDiscord<the_real_hypno> https://tenor.com/view/wolf-of-wall-street-jordan-belfort-leonardo-di-caprio-one-of-us-jonah-hill-gif-5441859
17:57:38FromDiscord<sOkam! 🫐> but yea, its really hard to make nim look ugly/unreadable, even when actively trying
17:58:01FromDiscord<the_real_hypno> In reply to @heysokam "but yea, its really": Hold my beer
17:58:49FromDiscord<sOkam! 🫐> In reply to @the_real_hypno "Hold my beer": relax, warrior. its already done. any of my codebases will trigger the hell out of everyone here 😂
17:58:53FromDiscord<sOkam! 🫐> hahaha
17:59:07FromDiscord<abaer> i assure you as a terrible programmer i can make nim look ugly and like an 8 year old first learned how to type on a keyboard.↵(@sOkam! 🫐)
17:59:15FromDiscord<the_real_hypno> Na just kidding, the editor will make me suffer big times 😅
17:59:19FromDiscord<Robyn [She/Her]> The Nim compiler is not having fun with ADT enums and my code... Oof
17:59:51FromDiscord<abaer> have you tried asking nicely and offering a blood sacrifice?↵(@Robyn [She/Her])
18:00:09FromDiscord<sOkam! 🫐> ^ also try rebooting your computer first
18:00:15FromDiscord<sOkam! 🫐> it fixes all
18:00:19FromDiscord<sOkam! 🫐> :kappa:
18:00:35FromDiscord<the_real_hypno> You just gotta give Araq a Schnitzel
18:00:41FromDiscord<the_real_hypno> No blood needed
18:02:09FromDiscord<morgan> so i have a proc which takes a ptr cstring (ptr ptr char in the c code it interfaces with). i need to set the internal pointer to that of a constant. would i be able to do that just by giving the constants the byref pragma or should i make it take ptr ptr char instead or what's the best way to do that?
18:02:56FromDiscord<Robyn [She/Her]> In reply to @abaer "have you tried asking": Nimsuggest
18:03:10*def- quit (Quit: -)
18:03:11FromDiscord<Robyn [She/Her]> Rebooting VSC fixes it until I edit the code again
18:03:21FromDiscord<Robyn [She/Her]> It's like, a sticky error in Nimsuggest
18:03:24FromDiscord<morgan> closing and reopening the window has worked for me
18:03:36FromDiscord<morgan> but i did just killall nimsuggest and restart it
18:03:39FromDiscord<the_real_hypno> In reply to @morganalyssa "so i have a": You sure about the ptr cstring?
18:03:53FromDiscord<Robyn [She/Her]> In reply to @morganalyssa "closing and reopening the": Yeah but I'm in the process of editing the code and pkill isn't enough here
18:04:11FromDiscord<morgan> In reply to @the_real_hypno "You sure about the": i am working with c library code, and its comments say to do that
18:04:58FromDiscord<the_real_hypno> No clue about the code but my spiderman sense tells me it wants cstring[0].addr
18:05:20FromDiscord<morgan> oh that looks like it might do it
18:05:57FromDiscord<the_real_hypno> And it would be ptr char then, which would make more sense
18:06:03FromDiscord<the_real_hypno> Atleast with the given info
18:06:08FromDiscord<morgan> prob will be a bit before i can run it and test it
18:06:31FromDiscord<morgan> sent a code paste, see https://play.nim-lang.org/#pasty=zcLqpVAC
18:06:49FromDiscord<the_real_hypno> I feel you, since i started interfering with C i start crying if someone uses strings
18:07:13FromDiscord<the_real_hypno> Uh, it actually is ptr ptr char
18:07:16FromDiscord<morgan> im writing an abstraction layer (offbeat) and inside that i do use normal strings
18:07:43FromDiscord<morgan> isn't ptr char the same as cstring?
18:07:52FromDiscord<morgan> or is cstring const ptr char?
18:07:54FromDiscord<Robyn [She/Her]> Not in Nim, iirc
18:07:58FromDiscord<the_real_hypno> It is
18:08:03FromDiscord<the_real_hypno> Well
18:08:09FromDiscord<Robyn [She/Her]> It's `UncheckedArray[cchar]` or something
18:08:31FromDiscord<the_real_hypno> Cstring[0].addr is a pointer to the first char of the array
18:08:38FromDiscord<morgan> yeah
18:08:41FromDiscord<Robyn [She/Her]> In reply to @the_real_hypno "It is": It is with even worse type safety :p
18:08:41FromDiscord<the_real_hypno> But thats ptr char
18:08:45FromDiscord<the_real_hypno> Not ptr otr
18:08:54FromDiscord<the_real_hypno> (edit) "otr" => "ptr"
18:09:18FromDiscord<morgan> i wrote it as ptr cstring
18:09:33FromDiscord<the_real_hypno> Ye would be correct
18:09:45FromDiscord<morgan> sent a code paste, see https://play.nim-lang.org/#pasty=UURBCEKG
18:09:53FromDiscord<the_real_hypno> Somehow my brains wants to advice ptr cstring[0].addr
18:10:00FromDiscord<the_real_hypno> With a cast
18:10:20FromDiscord<the_real_hypno> But i wouldn't listen to that fluffy piece of meat
18:10:26FromDiscord<morgan> well later on when everything else is in enough to test i can see if that satisfies it or not
18:11:09FromDiscord<dawidek.2137> is there a type for a js array in std/jsffi?
18:11:14FromDiscord<dawidek.2137> or wherevr
18:14:23FromDiscord<demotomohiro> sent a code paste, see https://play.nim-lang.org/#pasty=CrwfOGOG
18:16:21FromDiscord<Robyn [She/Her]> Nimsuggest is hating my code so much right now
18:16:26FromDiscord<morgan> ah cool thanks
18:16:42FromDiscord<Robyn [She/Her]> It refuses to launch, it keeps crashing
18:18:43FromDiscord<Robyn [She/Her]> Ah it hated my union 😔
18:18:47FromDiscord<Robyn [She/Her]> Is it a union?
18:18:58FromDiscord<Robyn [She/Her]> `Identifier | NumLit` (both are `distinct AstNode`
18:19:01FromDiscord<Robyn [She/Her]> (edit) "AstNode`" => "AstNode`)"
18:22:58*def- joined #nim
18:31:27FromDiscord<Robyn [She/Her]> Ugh, native support would be so much better for ADT enums
18:32:27FromDiscord<goerge_lsd> there is no native support ?
18:32:33FromDiscord<goerge_lsd> (edit) "no" => "non"
18:33:43FromDiscord<Robyn [She/Her]> Yeah, via libraries such as https://github.com/metagn/skinsuit or https://github.com/beef331/fungus
18:34:38FromDiscord<Robyn [She/Her]> They have some limitations though, I was using `fungus` because it looked much nicer, but unfortunately it's a bit painful because it uses a macro for the syntax, which doesn't work with Nim type sections
18:35:09FromDiscord<Robyn [She/Her]> I wrote a macro to join a type section and ADT enum from `fungus` into one, but it's not working that great
18:38:21FromDiscord<michaelb.eth> In reply to @goerge_lsd "there is non native": discussion here, also in some linked issues: https://github.com/nim-lang/RFCs/issues/548
18:38:50FromDiscord<michaelb.eth> for native, I mean
18:39:01FromDiscord<morgan> welp guess im going back to fixing up this regex https://media.discordapp.net/attachments/371759389889003532/1243996761782227165/image.png?ex=665381c4&is=66523044&hm=bdc3ceb9e8c641f7a905bfa14e476d4e1778605bfec4d8b3c48e26bd62e5e73d&
18:39:03FromDiscord<goerge_lsd> is this implemented somewhere or just talk ?
18:39:22FromDiscord<michaelb.eth> In reply to @goerge_lsd "is this implemented somewhere": just talk at this point
18:40:09FromDiscord<Robyn [She/Her]> https://forum.nim-lang.org/t/10836 This is a forum thread discussing it too
18:40:15FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "is this implemented somewhere": Just talk, currently
18:40:25FromDiscord<goerge_lsd> so we gotta wait for nim3
18:40:43FromDiscord<goerge_lsd> or nimskull 1.0 ? : )
18:41:07FromDiscord<Robyn [She/Her]> Probably would be implemented as an experimental feature in pre-3.0, if there even is a 3.0 planned
18:44:59FromDiscord<dawidek.2137> sent a code paste, see https://play.nim-lang.org/#pasty=ESxbJniD
18:45:04FromDiscord<the_real_hypno> In reply to @morganalyssa "welp guess im going": Where is the regex? 🤔
18:45:43FromDiscord<Robyn [She/Her]> In reply to @dawidek.2137 "can i force an": Not sure if you can do that, but why not try placing a doc comment in the function instead?
18:46:24FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=LBNwCjIh
18:46:51FromDiscord<morgan> In reply to @the_real_hypno "Where is the regex?": in the nim.json in the nim vscode extension
18:47:09FromDiscord<morgan> the regex to highlight emits with a list of strings and variables
18:47:33FromDiscord<the_real_hypno> Oooooh, i was trying really hard to find a regex string in your code 🥲
18:48:44FromDiscord<morgan> lol sorry
18:52:20FromDiscord<dawidek.2137> In reply to @chronos.vitaqua "Not sure if you": I want my typescript tooling to recognize the types and putting the comment inside doesn't seem to make it work
18:52:59*krux02 joined #nim
18:53:25FromDiscord<goerge_lsd> so, now I need to fix the inconsistency of nim's checksums module when it comes to byte type used (char or uint). Any ideas ?
18:53:52FromDiscord<goerge_lsd> I want to use a single proc to upate any kind Hash context with buffer data, which can be openArray[char] or uint8
18:54:24FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=vLUrpbbK
18:57:40FromDiscord<sOkam! 🫐> In reply to @morganalyssa "so i have a": `cstringArray`
18:59:12FromDiscord<goerge_lsd> hm what about variant objects with same variable name, were they experimental or planned for future?
18:59:29FromDiscord<morgan> In reply to @heysokam "`cstringArray`": no it only points to one
18:59:30FromDiscord<sOkam! 🫐> (edit) "`cstringArray`" => "`cstringArray`↵if its not an array of `char `, then please specify the C type you want to generate"
18:59:39FromDiscord<sOkam! 🫐> oh, i just edited it
18:59:48FromDiscord<sOkam! 🫐> so whats the type you want in C? no nim syntax
19:00:38FromDiscord<morgan> sent a code paste, see https://play.nim-lang.org/#pasty=QbxNqPMq
19:01:21FromDiscord<sOkam! 🫐> is that `char const `, `char const ` or `char const`?
19:01:36FromDiscord<morgan> that is exactly what is in the c code
19:01:39FromDiscord<sOkam! 🫐> remember that right-const-rule is fucky
19:02:24FromDiscord<sOkam! 🫐> In reply to @morganalyssa "that is exactly what": if im not mistaken, thats `char const `
19:02:40FromDiscord<sOkam! 🫐> which would be indeed `ptr cstring`
19:03:04FromDiscord<sOkam! 🫐> but if you want the `const` part mandatory... then you might need to wrap it and use that instead
19:03:12FromDiscord<sOkam! 🫐> let me search, i had issues with this before. sec
19:04:11FromDiscord<sOkam! 🫐> @morgan https://github.com/heysokam/wgpu/blob/b88fb2dcf80eda0f94b77c407bf6b3ba6afb95a5/src/wgpu/types.nim#L10-L13
19:04:44*def- quit (Quit: -)
19:04:52FromDiscord<sOkam! 🫐> that's the only way to be truly compatible with C. cstring does not map well sometimes
19:05:11*def- joined #nim
19:05:36FromDiscord<sOkam! 🫐> then you do `ptr CString` and it will be mapped to `const char`
19:06:26FromDiscord<morgan> i so far haven't had issues with cstring anywhere but i'll try to remember it if i do
19:08:13*def- quit (Client Quit)
19:08:57*def- joined #nim
19:08:57FromDiscord<sOkam! 🫐> In reply to @morganalyssa "i so far haven't": you won't if you dont set the C compiler to strict mode. if you do, then coercing `char ` to `char const ` won't be enough
19:09:09FromDiscord<morgan> ah
19:09:48FromDiscord<morgan> do i gain anything from strict mode?
19:11:08FromDiscord<sOkam! 🫐> type safety and UB protection. and if you ever use `zigcc`, it comes with all of that enabled
19:11:35FromDiscord<morgan> ah ok
19:14:42*krux02_ joined #nim
19:17:12*krux02 quit (Ping timeout: 260 seconds)
19:17:12*beholders_eye quit (Read error: Connection reset by peer)
19:17:50FromDiscord<Elegantbeef> @goerge_lsd you can do `data: openArray[byte or char]`
19:18:11FromDiscord<goerge_lsd> I was just thinking about you Beef, if you maybe want some homework 😄
19:19:08FromDiscord<goerge_lsd> ok, so I read from file (buffer) as `string`.. what's the best way to conver this to a type that fits in `openArray[byte or char]` ?
19:19:20FromDiscord<goerge_lsd> (edit) "conver" => "convert"
19:19:38FromDiscord<Elegantbeef> string is implicitly convertible
19:19:45FromDiscord<Elegantbeef> `openArray[char]` accepts `string`
19:20:51FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=kXpduHZj
19:21:09FromDiscord<goerge_lsd> it matched it to cstring ?
19:21:39FromDiscord<Elegantbeef> `uint8` is not char
19:22:21FromDiscord<goerge_lsd> yeah but what can I do about that.. checksums lib has incosistent data types :\
19:22:32FromDiscord<goerge_lsd> some use uint8, others char
19:22:36FromDiscord<Elegantbeef> `string` to `openArray[uint8]` requires explicit conversion using `str.toOpenArrayByte(0, str.high)`
19:22:45FromDiscord<goerge_lsd> alright then
19:22:48FromDiscord<Elegantbeef> PR to fix it to use `openArray[char or uint8]`
19:23:01*beholders_eye joined #nim
19:23:05FromDiscord<goerge_lsd> @ElegantBeef , what I am trying to do, is write a module that can hash a file or data in a single run, to multiple hasing algos.
19:23:18FromDiscord<goerge_lsd> Can you take a look and tell me if the code so far makes sense, or could have been done better
19:24:19FromDiscord<goerge_lsd> https://play.nim-lang.org/#pasty=XEOVPJbw
19:24:32*xet7 quit (Quit: Leaving)
19:26:54FromDiscord<goerge_lsd> of course ieally it would be multi threaded, but trying to figure it out for single thread before doing that
19:27:01FromDiscord<goerge_lsd> (edit) "ieally" => "ideally"
19:28:00FromDiscord<Elegantbeef> Personally I do not like the usage of the object variant as it means it's not arbitrary hash procedures
19:28:35FromDiscord<goerge_lsd> what would ElegantBeef do ?
19:29:04FromDiscord<Elegantbeef> Probably use a tuple of tuples
19:29:29FromDiscord<Elegantbeef> `((Sha128Context, sha128HashProc), (Md5Context, md5HashProc))`
19:29:38FromDiscord<Elegantbeef> Then iterate over those doing `hashProc(ctx, data)`
19:29:54FromDiscord<goerge_lsd> I thought of literally everything but tuples lol
19:30:50FromDiscord<Elegantbeef> Tuples are just lovely
19:33:06*xet7 joined #nim
19:33:50FromDiscord<the_real_hypno> What would ElegantBeef do :kekw:
19:34:53FromDiscord<goerge_lsd> blah, but `toOpenArrayByte` is not compatible with `openArray[char]`
19:35:15FromDiscord<Elegantbeef> Right
19:35:21FromDiscord<Elegantbeef> Just use `openArray[byte or char]`
19:35:24FromDiscord<Elegantbeef> Or just byte
19:36:05FromDiscord<goerge_lsd> well I am using that in the proc param, but the hash functions expect either openArray[uint8] or openArray[char]
19:36:12FromDiscord<goerge_lsd> (edit) "openArray[uint8]" => "`openArray[uint8]`" | "openArray[char]" => "`openArray[char]`"
19:37:05FromDiscord<goerge_lsd> this is what I hate in nim, this string,byte,uint8 nonsense.. files should be seq of uint8, not char or strings
19:37:35FromDiscord<Elegantbeef> You're complaining about a non issue
19:37:49FromDiscord<goerge_lsd> Yeah, skill issue
19:38:04FromDiscord<goerge_lsd> I think in Go you don't have this problem
19:38:10FromDiscord<Elegantbeef> You can cast between to get `openArray[char]`
19:38:43FromDiscord<goerge_lsd> cast at parameter when calling the hash update function ?
19:38:58FromDiscord<goerge_lsd> default to a type, and cast to the other when needed
19:41:03*xutaxkamay joined #nim
19:41:50Guest8054exit
19:41:54Guest8054quite
19:42:46Guest8054sorry I failed.
19:42:49*Guest8054 quit (Quit: WeeChat 4.1.1)
19:43:46FromDiscord<the_real_hypno> No problem
19:47:09FromDiscord<Elegantbeef> But yea I'd do something like https://play.nim-lang.org/#pasty=OmczhayR
19:51:17FromDiscord<Elegantbeef> Actually probably can do `field.Y is char`
19:54:32FromDiscord<Robyn [She/Her]> Hey Beef, any clue why https://play.nim-lang.org/#pasty=JKIiSPmF causes this error?↵`/home/chronos/Projects/Napkins/src/napkinspkg/parser/types.nim(30, 37) Error: undeclared identifier: 'AstNode'`↵Relevant line is `ProtoIDPair = tuple[proto: ref AstNode, packet: ref AstNode]`, I can't figure out a minimal reproducible example though...
19:55:18FromDiscord<Elegantbeef> `--expandMacro:adtEnum`
19:55:58FromDiscord<Elegantbeef> Actually sorry
19:56:03FromDiscord<Elegantbeef> `--expandMacro: join`
19:56:14FromDiscord<Elegantbeef> I love that you're asking me to debug your macros, as if I do not have enough
19:57:17FromDiscord<the_real_hypno> What would ElegantBeef do!
19:57:51FromDiscord<Robyn [She/Her]> Lol
19:58:37FromDiscord<Robyn [She/Her]> It's not printing out the AST :/
19:58:43FromDiscord<griffith1deadly> you are a macros god
19:58:47FromDiscord<griffith1deadly> beef
20:02:10FromDiscord<Elegantbeef> Well robyn time to stop playing around with `join` and use https://play.nim-lang.org/#pasty=CLcBwHxi
20:02:13FromDiscord<Elegantbeef> type section macro time
20:13:35FromDiscord<Robyn [She/Her]> I had to step away for a bit so I couldn't modify my macro to print the AST
20:14:09FromDiscord<Robyn [She/Her]> In reply to @Elegantbeef "Well robyn time to": Oh damn, that'd actually be nice
20:14:13FromDiscord<Robyn [She/Her]> Issue however:
20:14:16FromDiscord<Robyn [She/Her]> No converters?
20:15:14FromDiscord<Robyn [She/Her]> And I'm pretty sure it can't define multiple types at once (the enum and the type itself)
20:15:39FromDiscord<goerge_lsd> `array[char]` to `seq[uint8]` ?
20:18:08FromDiscord<Robyn [She/Her]> In reply to @goerge_lsd "`array[char]` to `seq[uint8]` ?": You can do `@myArray` to get a `seq[char]` if wanted
20:18:15FromDiscord<goerge_lsd> that I already knew..
20:18:40FromDiscord<Robyn [She/Her]> Then what's the problem? Have you not done `seq[char or uint8` like Beef suggested?
20:18:50FromDiscord<the_real_hypno> Are you to talking about different topics?
20:18:58FromDiscord<the_real_hypno> (edit) "to" => "two"
20:19:33FromDiscord<goerge_lsd> this is different part of the code. I am doing a finalize proc to get the digests, but haven't decided how to store the result yet, so i am reverting to seq[uint8] for everything to check if it's the correct checksum
20:19:55FromDiscord<goerge_lsd> sent a code paste, see https://play.nim-lang.org/#pasty=DhAKBKhb
20:20:22FromDiscord<Robyn [She/Her]> So you want to convert a char array to a uint8 seq?
20:20:27FromDiscord<goerge_lsd> indeed
20:21:38FromDiscord<Elegantbeef> You can define multiple types at once Robyn
20:22:21FromDiscord<the_real_hypno> Isnt char defined as uint8?
20:22:30FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=WQYQpxTw
20:22:33FromDiscord<Elegantbeef> No
20:22:42FromDiscord<Elegantbeef> char is a distinct type to `uint8`
20:23:09FromDiscord<sOkam! 🫐> In reply to @the_real_hypno "Isnt char defined as": char is `int8`
20:23:18FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=UiguJgfV
20:23:30FromDiscord<Robyn [She/Her]> This just requires an explicit conversion
20:23:44FromDiscord<Robyn [She/Her]> Though casting can be done, it's not advised for reasons I can't remember
20:24:11FromDiscord<Elegantbeef> No `char` is `char` 😄
20:24:42FromDiscord<Robyn [She/Her]> In reply to @Elegantbeef "You just need to": Oh damn, I didn't know that! Does this also define `Circle` in a new scope? And can you define templates and converters like this?
20:25:37FromDiscord<griffith1deadly> sent a code paste, see https://play.nim-lang.org/#pasty=vtQkqgYF
20:25:42FromDiscord<griffith1deadly> `copyMem`
20:26:09*def- quit (Quit: -)
20:26:49*def- joined #nim
20:27:52FromDiscord<Robyn [She/Her]> Copy mem doesn't work on the JS backend though, so that's something to keep in mind
20:28:46FromDiscord<Elegantbeef> Yes you can robyn
20:29:21FromDiscord<Elegantbeef> The issue is that generics do not work as one would intend
20:30:06FromDiscord<the_real_hypno> `char↵Built-in 8 bit character type (unsigned). `
20:30:22FromDiscord<Robyn [She/Her]> In reply to @Elegantbeef "The issue is that": Oh? What do you mean?
20:30:39FromDiscord<Elegantbeef> They do not work
20:30:50FromDiscord<Robyn [She/Her]> Ah
20:30:52FromDiscord<Elegantbeef> Generics defined like the above just explode
20:32:07FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=xrKKFQyk
20:32:23FromDiscord<Elegantbeef> No it just explodes 😄
20:33:40FromDiscord<the_real_hypno> Ok, for real, i totally get that it probably is a distinct type but uint8 and char is still the same!?
20:33:45FromDiscord<Elegantbeef> https://github.com/beef331/constructor/pull/14 refer to this
20:38:30FromDiscord<the_real_hypno> ` cuchar {.importc: "unsigned char", nodecl, deprecated: "use `char` or `uint8` instead".} = char`
20:39:05FromDiscord<the_real_hypno> char or uint8 aren't distinct?
20:39:34FromDiscord<Elegantbeef> They are
20:39:37FromDiscord<Elegantbeef> They're magic
20:39:56FromDiscord<the_real_hypno> Is that implying distinction?
20:40:31FromDiscord<Elegantbeef> It means the compiler handles their implementation
20:41:11FromDiscord<Robyn [She/Her]> https://play.nim-lang.org/#pasty=FrYPjSjn this won't even compile for me
20:41:20FromDiscord<Robyn [She/Her]> `/home/chronos/test/main.nim(4, 17) Error: expected: ')', but got: 'OptionKind'`
20:42:11FromDiscord<Elegantbeef> Well yea it has to be done through a macro
20:42:26FromDiscord<Elegantbeef> or template
20:42:27FromDiscord<the_real_hypno> Meh. Guess I'll just do what ElegantBeef would do
20:43:48FromDiscord<Robyn [She/Her]> Ah
20:50:33FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=tZapwvwy
20:50:48FromDiscord<Elegantbeef> Depend on your code
20:51:46FromDiscord<Robyn [She/Her]> My code is in the link
20:52:14FromDiscord<Elegantbeef> I didn't see it
20:52:14FromDiscord<Elegantbeef> \:d
20:52:30FromDiscord<Robyn [She/Her]> Oh!
20:52:32FromDiscord<Robyn [She/Her]> This works!
20:52:33FromDiscord<Robyn [She/Her]> https://play.nim-lang.org/#pasty=RuwikRbf
20:52:43FromDiscord<Elegantbeef> "didnt"
20:53:44FromDiscord<Elegantbeef> Your code works fine if you do it properly
20:53:49FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=PBPqTeNL
20:54:02FromDiscord<Robyn [She/Her]> https://play.nim-lang.org/#pasty=wyqGlYyH
20:54:11FromDiscord<Robyn [She/Her]> Yeah :P
20:54:17FromDiscord<Robyn [She/Her]> So then, how does it blow up?
20:54:55FromDiscord<Elegantbeef> I showed you an example and you just disregarded it
20:54:56FromDiscord<Elegantbeef> So....
20:55:37FromDiscord<Elegantbeef> `constructor.defaults` fails to work with generics
20:56:17FromDiscord<Robyn [She/Her]> Did some messages get skipped on the bridge? Hold on
20:58:01FromDiscord<Robyn [She/Her]> No it didn't, I don't know where you showed an example tho
21:01:02FromDiscord<the_real_hypno> https://github.com/beef331/constructor/pull/14
21:01:27FromDiscord<the_real_hypno> He referred this
21:01:59FromDiscord<the_real_hypno> In reply to @Elegantbeef "https://github.com/beef331/constructor/pull/14 refe": ^
21:02:47FromDiscord<goerge_lsd> @ElegantBeef http://marcio.io/2015/07/calculating-multiple-file-hashes-in-a-single-pass/ check this out, quite short code to do this in Go
21:02:48FromDiscord<Robyn [She/Her]> Aaaah
21:03:01FromDiscord<Elegantbeef> Cool it's Go
21:03:12FromDiscord<Elegantbeef> It also is hardcoded hash procedures
21:03:24FromDiscord<Elegantbeef> Want to add a new has procedure, too bad
21:03:28FromDiscord<goerge_lsd> yeah, no variable hashes
21:03:46FromDiscord<Elegantbeef> Using a tuple of objects like I showed is more sensible imo
21:04:08FromDiscord<goerge_lsd> I didn't quite wrap my mind .. care to add a runnable example to that code ? :\
21:04:31FromDiscord<Elegantbeef> Given I do not know how hash functions generally work... I can try
21:05:51*def- quit (Quit: -)
21:06:03FromDiscord<goerge_lsd> well i had no idea as well, but by looking at their runnableExamples, i could see that you can add varying amounts of data to an already computed context, and it will give the right result, once finalized
21:06:29FromDiscord<goerge_lsd> they all have the same interface in that regard
21:06:45FromDiscord<goerge_lsd> init/update n times/finalize
21:06:54*def- joined #nim
21:09:49*def- quit (Client Quit)
21:10:44FromDiscord<goerge_lsd> https://play.nim-lang.org/#pasty=mXuzWCFH
21:10:55FromDiscord<goerge_lsd> well, the results are correct.. time to beautify code
21:11:02FromDiscord<goerge_lsd> then thread it
21:11:30FromDiscord<Robyn [She/Her]> I'm gonna name my lib `cordyceps` and make the pragma `spore` :)
21:13:01*def- joined #nim
21:21:00FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=jSCyLynh
21:21:05FromDiscord<Elegantbeef> there you go @goerge_lsd
21:22:46FromDiscord<Elegantbeef> I do realise now you do not want to run them on top of each other, but that's also easy to do
21:25:08FromDiscord<goerge_lsd> are you sure this is absolutely the most elegant nim, given the circumstances ? : D
21:25:15FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=auPItMgE
21:25:19*def- quit (Quit: -)
21:25:25FromDiscord<Elegantbeef> It's quite elegant as it does not rely on maintaining an object upstream
21:25:35*def- joined #nim
21:25:39FromDiscord<Elegantbeef> As such you can just define a bunch of these `consts` and use them
21:27:02FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=xrTtnvrE made it even cleaner
21:27:25FromDiscord<Elegantbeef> The simple fact is that `checksums` is not written to be generic so this is required
21:28:39FromDiscord<goerge_lsd> hm but your code does not do the hashing in chunks..
21:30:43FromDiscord<Elegantbeef> I also value reusable code more than specialised code
21:30:43FromDiscord<Elegantbeef> Having arbitrary support for hash procedures is just great
21:30:45FromDiscord<Elegantbeef> What?
21:30:45FromDiscord<Elegantbeef> You asked for an example
21:30:45FromDiscord<Elegantbeef> If you want to make it more elaborate, do that
21:30:46FromDiscord<Elegantbeef> Is it so hard to add a `finishProc` and to invoke that at the end of `hashIt`
21:30:46FromDiscord<Elegantbeef> Methinks not
21:33:13FromDiscord<goerge_lsd> finishProc to the Hasher Object
21:33:39FromDiscord<Elegantbeef> technically we could remove `Hasher` and just use generic procedures
21:33:39FromDiscord<Elegantbeef> That involves shadowing all of the procedures into a single shape
21:33:39FromDiscord<Elegantbeef> Which you're practically doing now
21:34:59FromDiscord<goerge_lsd> not sure I follow, remove Hasher ?
21:35:25FromDiscord<goerge_lsd> procedures are already generic
21:35:46FromDiscord<Elegantbeef> No they're not
21:39:51FromDiscord<goerge_lsd> well, if you think that's more elegant, would like to see a sample : P
21:40:20FromDiscord<goerge_lsd> can't get as elegant as that Go code though, most likely, we don't have generic writeres and multiplexers
21:40:28FromDiscord<Elegantbeef> Ok final thing I will contribute
21:40:29FromDiscord<goerge_lsd> (edit) "writeres" => "writers"
21:40:29FromDiscord<Elegantbeef> https://play.nim-lang.org/#pasty=MikNqHxj
21:40:39FromDiscord<Elegantbeef> Paint the rest of the owl thanks
21:41:49FromDiscord<Elegantbeef> I mean I think this code is more elegant as it does not rely on runtime interfaces, but ymmv
21:43:30FromDiscord<Elegantbeef> Could also use a macro so you could take in `"bleh".hashIt(tuple[md5: Md5Context, sha1: Sha1State])` then get back a `tuple[md5: string, sha1: string]`
21:43:47FromDiscord<Elegantbeef> Then you can do `echo "bleh".hashIt(tuple[md5: Md5Context, sha1: Sha1State]).md5`
21:44:32FromDiscord<Elegantbeef> But this works with arbitrary field names and types so is wonderful all without having to hoist data to dynamically dispatch
21:44:41FromDiscord<Elegantbeef> Which I find more elegant 😛
21:45:12FromDiscord<goerge_lsd> never touched macros :\
21:45:48FromDiscord<Elegantbeef> Well this would be a very simple one 😄
21:49:45FromDiscord<sOkam! 🫐> simple macros are a lie. they don't exist
21:49:56FromDiscord<Elegantbeef> They really do
21:50:14FromDiscord<sOkam! 🫐> they are always like santa, big fat and full of surprises
21:55:06FromDiscord<Robyn [She/Her]> sent a code paste, see https://play.nim-lang.org/#pasty=zGIesieb
21:55:56FromDiscord<Elegantbeef> Now now sokam this is simple https://play.nim-lang.org/#pasty=PTRHJVnH
21:56:25FromDiscord<Elegantbeef> Technically we can use `typetraits.isNamedTuple` to remove the error from the macro
21:57:31FromDiscord<Elegantbeef> \To finish this off we could even remove the string conversion, and emit a `tuple[name: typeof(state.finish())]`
21:57:47FromDiscord<Elegantbeef> Which means no pointless allocations!
21:58:24FromDiscord<Elegantbeef> I did forget to do `static: i = 0` at the start of `hashIt`
22:00:55FromDiscord<goerge_lsd> don't understand the macro version, and previous one with overloaded procs.. doesn't look so great to me
22:01:20FromDiscord<goerge_lsd> anyway, what thread lib would you use to pass the buffer to each thread handling a hash function ?
22:01:20FromDiscord<Elegantbeef> Well you're blind to the glory that is generic interfaces and reusable code
22:01:51FromDiscord<Elegantbeef> The benefit of the overloaded procs is you can use any hash procedure that you want without having to make a PR or fork the library this comes from
22:02:04FromDiscord<Elegantbeef> It's endlessly extensible without relying on runtime dispatch
22:02:31FromDiscord<goerge_lsd> I think you have to pass the pointer to it, that would be safe since it's read only, wait for all threads to.. yield? or how do you call it, finish work without kiling the thread ?
22:02:33FromDiscord<Elegantbeef> The tuple version simply emits a tuple instead of an array
22:25:32*beholders_eye quit (Read error: Connection reset by peer)
22:28:18FromDiscord<Robyn [She/Her]> In reply to @Elegantbeef "Which means no pointless": Rustacean energy
22:31:21*beholders_eye joined #nim
22:33:40FromDiscord<zumi.dxy> Is there any way I can do "subsets" of cases? https://play.nim-lang.org/#pasty=lVWEaruF
22:37:30FromDiscord<polylokh_39446> rejected in <https://github.com/nim-lang/Nim/issues/11086>
22:38:18FromDiscord<polylokh_39446> but also, an example of a subrange in that
22:39:58FromDiscord<polylokh_39446> <https://play.nim-lang.org/#pasty=zERWJZzL>
22:40:26FromDiscord<polylokh_39446> requires the subrange to be continuous though
22:44:14FromDiscord<zumi.dxy> aight, I've got this now: https://play.nim-lang.org/#pasty=mXOBMvrd
22:44:43FromDiscord<zumi.dxy> can I extract the "range" out of the type it's contained in?
22:45:05FromDiscord<zumi.dxy> hmm...
22:45:13FromDiscord<polylokh_39446> what do you mean by extract?
22:45:26FromDiscord<Robyn [She/Her]> Can you not do `of {A, B, C}: ...`?
22:45:55FromDiscord<zumi.dxy> In reply to @zumi.dxy "can I extract the": ok almost forgot about this https://play.nim-lang.org/#pasty=NuOFQikj
22:46:11FromDiscord<Robyn [She/Her]> What you're asking is confusing tbh
22:46:41FromDiscord<polylokh_39446> the likeliest problem to come up with code like this is that it doesn't work with non-contiguous values
22:48:47FromDiscord<zumi.dxy> In reply to @chronos.vitaqua "What you're asking is": I like my exhaustive case checks and would like to not use dodgy if-else inside of a case covering multiple values↵like I have a `of A, B, C:`, so I know `D` and `E` won't ever be touched
22:49:28FromDiscord<zumi.dxy> so my code will assume likewise, but I just have an `if…else`↵but what if I add `D` to that `of`-clause?
22:49:48FromDiscord<zumi.dxy> I want that to be checked
22:50:11FromDiscord<zumi.dxy> like "you added `D` to that case but you didn't account for it"
22:50:21FromDiscord<Robyn [She/Her]> You could use `else`?
22:50:41FromDiscord<Elegantbeef> I would do explict range cases and `assert false`
22:51:11FromDiscord<Robyn [She/Her]> https://play.nim-lang.org/#pasty=FSbXktqx
22:51:44FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=DJEFazEG
22:52:42FromDiscord<Elegantbeef> Could even do `of A..pred(E)` and `succ(D) .. TheWholeSet.high`
22:52:53FromDiscord<Elegantbeef> Whoops `pred(D)`
22:58:46FromDiscord<zumi.dxy> In reply to @polylokh_39446 "the likeliest problem to": and casting `something` to a set instead of a range doesn't seem to work 🤔
22:59:28FromDiscord<polylokh_39446> casting preserves memory layout and sets are completely differently laid out
22:59:46FromDiscord<zumi.dxy> https://play.nim-lang.org/#pasty=hVjvJTAf
23:00:45FromDiscord<Elegantbeef> `range[D..E](something)`
23:01:00FromDiscord<polylokh_39446> you might raise an issue to remove the need for the assert falses there. Type inference has been improved in this direction before, for discriminated unions
23:01:44FromDiscord<Elegantbeef> Well araq will probably say "Until NIR is used ...."
23:01:46FromDiscord<zumi.dxy> In reply to @Elegantbeef "`range[D..E](something)`": yeah that's for continuous types↵I can handle that tbh
23:02:00FromDiscord<zumi.dxy> In reply to @polylokh_39446 "you might raise an": This could use an RFC even
23:02:13FromDiscord<zumi.dxy> but I have no clue on how to make a good one
23:03:01FromDiscord<Elegantbeef> https://github.com/nim-lang/RFCs/issues/429
23:03:53FromDiscord<zumi.dxy> ah, there it is↵scope narrowing
23:04:05FromDiscord<zumi.dxy> yeah that's what I meant
23:08:45FromDiscord<Robyn [She/Her]> Uuuuh how do I rename an identifier? `myIdent.strval = "a"` doesn't work
23:08:57FromDiscord<Elegantbeef> `myIdent = ident"a"`
23:09:07FromDiscord<Robyn [She/Her]> I don't want to create a new identifier because I'm purposefully relying on ref semantics
23:09:15FromDiscord<Elegantbeef> So then dereference it?
23:09:29FromDiscord<Robyn [She/Her]> Wait I can do that?
23:09:41FromDiscord<Elegantbeef> Why not?
23:09:42FromDiscord<Robyn [She/Her]> So `myIdent[] = (ident"a")[]`?
23:09:51FromDiscord<Robyn [She/Her]> I just didn't think of it honestly
23:10:16FromDiscord<Elegantbeef> pointer emulation is sorta a surprise from the NimVm
23:10:38FromDiscord<Elegantbeef> I had `ptr array[..., proc]` working with the VM and was amazed
23:11:29FromDiscord<Robyn [She/Her]> It's pretty neat honestly
23:21:55FromDiscord<Robyn [She/Her]> Micros probably has some useful stuff I should look into
23:40:43*xet7 quit (Ping timeout: 256 seconds)
23:42:49FromDiscord<Elegantbeef> I wish it was more complete but there is some nice stuff