<< 08-11-2018 >>
00:17:02krux02Araq: lol
00:17:47krux02isn't `not nil` an unary use of `not`?
00:18:16krux02well good night
00:18:17*dddddd quit (Remote host closed the connection)
00:18:58*krux02 quit (Remote host closed the connection)
00:51:59*citycide quit (Ping timeout: 276 seconds)
00:59:11*wildlander quit (Quit: Konversation terminated!)
01:03:00*Sembei joined #nim
01:03:05*darithorn quit ()
01:04:26*citycide joined #nim
01:05:45*Pisuke quit (Ping timeout: 252 seconds)
01:06:30*theelous3_ quit (Ping timeout: 252 seconds)
01:24:12*zachk quit (Quit: Leaving)
02:09:06*couven92 quit (Quit: Client Disconnecting)
02:14:21FromGitter<zetashift> Why isn't this printing `c` instead of `b`? https://pastebin.com/fKLUkmgv
02:17:08FromGitter<kayabaNerve> @zetashift What is it printing?
02:17:41FromGitter<kayabaNerve> Also, you set the initial value to 1 and only increment it on the second time, yet you return it when it finds 1
02:17:52FromGitter<kayabaNerve> L11 should be == 0
02:18:15FromGitter<kayabaNerve> Or you should move the increment out of the else
02:18:24FromGitter<kayabaNerve> *set the intial value to 0
02:18:34FromGitter<kayabaNerve> Or you should have the initial value set to 1
02:20:56FromGitter<zetashift> @kayabaNerve thanks! it's printing all together `{'a': 3, 'b': 1, 'c': 0, 'd': 0} ⏎ b ⏎ `
02:21:39FromGitter<kayabaNerve> Got it. So did I solve it? :P
02:23:03FromGitter<zetashift> I need to increment it everytime it stumbles on that char again
02:23:12FromGitter<zetashift> but the initial value had to be set to 1 yeah
02:25:49FromGitter<zetashift> this is my code currently and it correctly prints the table with it's values but still the wrong value: https://pastebin.com/WCvGqy3Z
02:28:00FromGitter<kayabaNerve> Why are you using a table btw
02:28:16FromGitter<kayabaNerve> And it's because you don't break out of your for loop
02:28:58FromGitter<zetashift> thought it would be the best structure to handle this.
02:29:05FromGitter<zetashift> maybe a Seq would be faster
02:29:15FromGitter<kayabaNerve> Generally a double for is using
02:29:15FromGitter<zetashift> A Seq containing a Seq*
02:29:57FromGitter<zetashift> thanks that solved it :P
02:30:09FromGitter<zetashift> Here I was thinking I was being nice with the result
02:33:26*k1tt3hk4t joined #nim
02:33:52*darithorn joined #nim
02:37:48FromGitter<kayabaNerve> ... are you supposed to be able to break out of if statements?
02:39:00FromGitter<kayabaNerve> I would've never guessed yes except that's what's happening right now.
02:39:17*cyraxjoe quit (Quit: I'm out!)
02:39:42FromGitter<zetashift> you can do an early return yea that's what I did, not sure if that's good practice though
02:39:55FromGitter<kayabaNerve> Docs say `break` breaks out of its block statement. An if statement is... possibly technically a block statement? Except I'm trying to use the if to decide if I want to break out of the for loop. Because it's breaking out of an if an if statement, the loop still runs
02:40:09FromGitter<zetashift> I just did `return k`
02:40:12FromGitter<kayabaNerve> I can't name a loop so I'd have to put the loop in a named block statement to do this if this is intended
02:40:18FromGitter<kayabaNerve> @zetashift I'm not talking about your code
02:40:31FromGitter<zetashift> ah sorry
02:41:02*cyraxjoe joined #nim
02:45:08FromGitter<kayabaNerve> Nope. I'm just an idiot, as I thought.
02:45:28FromGitter<kayabaNerve> Can't believe I didn't catch this 0_0
02:47:40FromGitter<kayabaNerve> @zetashift
02:47:44FromGitter<kayabaNerve> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5be3a3d047217e07ffea82f4]
02:48:17FromGitter<kayabaNerve> There's the two loop version, and the generic answer to your problem.
02:49:37FromGitter<kayabaNerve> A seq + checking if the seq contains it effectively does the same thing
03:01:51*k1tt3hk4t quit (Quit: WeeChat 2.3)
03:05:23*banc quit (Quit: ZNC - http://znc.in)
03:11:06Demos[m]I want just regular ligatures
03:11:15Demos[m]sorry, response to a month old message
03:11:18Demos[m]please ignore
03:13:03*Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
03:21:30*banc joined #nim
03:39:06martin1_1is there any widely accepted code formatter? I've found a few but I'm not sure which one should I go with
04:39:56*endragor joined #nim
04:42:32*vlad1777d_ quit (Ping timeout: 268 seconds)
05:00:08leorizemartin1_1: there's nimpretty
05:00:50*chemist69 quit (Ping timeout: 252 seconds)
05:02:56*chemist69 joined #nim
05:22:15*endragor quit (Remote host closed the connection)
05:22:27*endragor joined #nim
05:34:29*nsf joined #nim
05:35:06*darithorn quit ()
05:37:46*vlad1777d_ joined #nim
06:01:48FromDiscord_<treeform> nimpretty is the standard
06:10:27*martin1_1 quit (Ping timeout: 240 seconds)
06:53:56FromGitter<tim-st> PMunch: Now with full suggestion support from critbits-tree, I hope the lsp protocol supports import name suggestions: https://gist.github.com/tim-st/df0ce1c4b712abe7d4e2a90f0b8c78a2
06:59:27*mech422__ joined #nim
07:01:32*krux02 joined #nim
07:02:16FromGitter<AchalaSB> Hi, Anyone have idea to solve this issue?
07:02:30FromGitter<AchalaSB> ``````
07:02:39*mech422_ quit (Ping timeout: 244 seconds)
07:04:09FromGitter<AchalaSB> Hi, Anyone have idea to solve this issue? ``` vagrant@bodhi64 /v/Samples> make hello ⏎ nim c --out:hello hello.nim ⏎ /etc/nim.cfg(256, 1) Error: closing " expected ⏎ Makefile:10: recipe for target 'hello' failed ⏎ make: *** [hello] Error 1. ``` [https://gitter.im/nim-lang/Nim?at=5be3dfe892b7d11721563711]
07:11:49FromGitter<krux02> seems to be an illegal makefile
07:18:33*kapil____ quit (Quit: Connection closed for inactivity)
07:26:41*Vladar joined #nim
08:42:02FromGitter<AchalaSB> Makefile is there inside Samples folder
08:46:47*PMunch joined #nim
09:00:52PMunchmartin1_, just merged your PR for nimlsp
09:01:07PMunchHow's the stability by the way? I've been getting some random freezes
09:01:33PMunchI can see in the Sublime console window that it is trying to send requests, but nimlsp seems to be hanging on something
09:02:34PMunchI'm assuming it's either nimsuggest output it's waiting for, or it's something around its input/output streams that locks up
09:26:07Araqthe joys of interprocess communication
09:47:59FromGitter<survivorm> Oh yeah. Freezes in random places, sometimes interwined with something low-level. Joy of debugging interprocess communications is something that couldn't be underestimated :)
10:03:56FromGitter<alehander42> can we add `isEmpty` to the stdlib
10:04:15FromGitter<alehander42> I am sure somebody else suggested it before, but I can't remember if people liked the idea or hated it
10:10:03FromGitter<alehander42> maybe `not isEmpty` looks strange ? I think I prefer it compared to `len > 0` still
10:11:20FromGitter<Varriount> @PMunch How's the language server?
10:12:37FromGitter<alehander42> when do we usually get `len(a) == L` seq modified while iterating over it [AssertionError] ?
10:15:34*craigger_ quit (Read error: Connection reset by peer)
10:16:51*craigger joined #nim
10:18:16*chemist69 quit (Ping timeout: 260 seconds)
10:18:38*elrood joined #nim
10:19:07FromGitter<narimiran> @alehander42 `isEmpty` == `not isAny` ?
10:19:10*chemist69 joined #nim
10:28:56FromGitter<alehander42> do we have isAny
10:29:58FromGitter<narimiran> oh, i meant `anyIt` (IIRC the previous discussion was about `isEmpty` in that context)
10:35:02FromGitter<alehander42> aah nah, I meant a general isEmpty
10:37:27*martin1_ quit (Ping timeout: 240 seconds)
10:40:15*martin1_ joined #nim
10:43:47*dddddd joined #nim
11:03:05*martin1_ quit (Quit: WeeChat 2.2)
11:04:03*martin1_ joined #nim
11:04:28Araqalehander42: I want it too. for packedjson 'len == 0' is expensive. On the other hand, we have TR macros for that, maybe we should use them more
11:05:38martin1_PMunch: the stability is sub-optimal to be honest.. I'm getting some random freezes as well.. I think the issue now is when getting messages from the client, I'll see if I can find the issue this night and hopefuly submit another PR
11:08:34*floppydh joined #nim
11:09:20FromGitter<alehander42> Araq, that's true, some types can overload `isEmpty` with more optimal checks
11:09:45*Vladar quit (Remote host closed the connection)
11:09:48FromGitter<alehander42> well the use of TR macros is impl detail
11:14:44Araqhmm thinking the unthinkable here
11:15:09Araqmaybe every string should really be an Option[string not empty]
11:16:16FromGitter<mratsim> as long as options are”cost-free” abstractions in most cases
11:16:55Araqwell I don't want to write if not x.isEmpty: for y in x.unwrap():
11:17:03FromGitter<mratsim> everytime I see string processing I’m always think “so where will this allocate and how can I avoid this” and the common answer is avoid strings :D
11:17:46Araqavoiding strings is generally a good thing, not only for performance, but also for correctness
11:18:40martin1_I'm pretty much a newcomer to the language so maybe this was already discussed but, given that nim is style insesitive it would be nice if nimsuggest could be configured to enforce a style on the suggestions, wouldn't it?
11:19:02Araqwe're getting 'nim --styleCheck:error' for that
11:19:04FromGitter<mratsim> there is a -nep1 flag to warn
11:19:10FromGitter<alehander42> thankfully, strings are unavoidable in real world
11:19:23Araqmratsim: already renamed to --styleCheck
11:19:26FromGitter<mratsim> the real world is a matrix, and you don’t need strings there :D
11:19:50FromGitter<alehander42> Araq, why is it so important if a string is empty on a type level ?
11:20:13Araqbecause it's usually invalid.
11:20:20Araqempty filename? that's special.
11:20:23FromGitter<alehander42> eh nope
11:20:25Araqempty Nim identifier? special.
11:20:32FromGitter<alehander42> well that's why filenames and idents have their own types
11:20:50Araqan empty SQL query? likely something special.
11:20:57FromGitter<alehander42> it's the same as saying "empty lists are usually invalid"
11:21:11FromGitter<alehander42> sql: again it needs its own type in good code
11:21:15Araqand that would be equally true
11:21:38Araqkeep in mind that you're talking to the only person left who dislikes Option[T]
11:22:25FromGitter<alehander42> :D I see this and still here I think the lesson is different
11:22:32FromGitter<alehander42> if your data is probably invalid for len == 0
11:22:44FromGitter<alehander42> probably you need to `distinct` or wrap it somehow
11:23:29Araqtype SqlQuery = distinct Option[string not empty] ?
11:23:53FromGitter<alehander42> yeah maybe
11:23:56FromGitter<alehander42> but still
11:24:00FromGitter<alehander42> if you think about it
11:24:07FromGitter<alehander42> it just needs a full blown validation assert
11:24:14FromGitter<alehander42> abs paths need to start with `/`
11:24:22FromGitter<alehander42> sql queries need to be valid
11:24:32FromGitter<alehander42> idents need to be \w
11:24:44Araqyeah and what does the validator return?
11:24:44FromGitter<alehander42> the len check is a lot of work for just one edge case
11:25:20FromGitter<alehander42> well, hopefully something that can optionally contain an error message :D
11:25:25Araqit returns an (string not empty) | Error
11:25:45Araq(or it raises an exception, yadda yadda)
11:26:00FromGitter<alehander42> well ok, string not empty seems fine
11:26:51FromGitter<alehander42> but my point is that you can also have domains where string len>=4 or string [a-zA-Z] makes perfect sense
11:27:14Araqsure and also empty lines are usually perfectly fine and valid
11:27:27Araqso it's really lines(): seq[string]
11:27:34Araqnot lines(): seq[string not empty]
11:27:38FromGitter<alehander42> no, what I mean is that in the same vein people might say "I want type string len >= A"
11:27:56FromGitter<alehander42> and know you want dependent types :D
11:28:23AraqI think you don't want dependent types.
11:28:57FromGitter<alehander42> yeah, but string not empty is a hardcoded instance of a dependent type
11:29:22FromGitter<alehander42> I hope that one day somebody will make them magic/ergonomic enough to use in normal langs :D
11:29:28Araqit's a useful hardcoding
11:29:41Araqthere is always the more general solution lurking everywhere
11:30:01Araqso everything ends up in sophistry.
11:30:04FromGitter<alehander42> yeah which reminds me
11:31:06Araqand while we are talking about type thery some javascript kiddie implements Bomberman for the iPhone and becomes a millionaire
11:32:25FromGitter<alehander42> the sink type in destructors, is it basically an affine type or a linear type? does it have to be always consumed?
11:32:49Araqit's affine.
11:33:08AraqI started with "always has to be consumed" and gave up rather quickly
11:33:23FromGitter<alehander42> well good for the kiddie, games are useful too
11:33:29Araqit's a much harder problem, and in addition to that, also annoying
11:33:50FromGitter<alehander42> yeah , so the wiki is a bit older
11:34:06Araqproc put(t: Table; key, val: sink U) # overwrite t[key] if it exists, otherwise create the entry
11:34:10FromGitter<alehander42> yeah it doesn't seem too useful too
11:34:27Araqnote that you want 'key' to be 'sink' but whether it really is consumed varies at runtime
11:35:10Araqwell I suppose you could swap() the key then
11:35:32FromGitter<alehander42> yeah iirc this was the problem: with branched code often you have a case where a var is not consumed
11:35:59Araqneither Rust nor C++ are after linear types, that's telling us something
11:36:03FromGitter<alehander42> yes
11:36:54Araqon the other hand "AST substructure is not added to tree" is a bug linear types would find
11:37:38FromGitter<alehander42> so this affine check mechanism would be general enough to be used for various resources , e.g. modelling that a network message is always consumed max once etc ?
11:37:54FromGitter<alehander42> in the beginning I thought this is something deeply tied to memory ptr types
11:38:26Araqdepending on how you structure the code, 'sink' will find this problem.
11:40:03FromGitter<alehander42> good. and the linear types: what is so hard about checking them, if you have the ability to prove that a value is used in a branch (which you do? if you have affine types?) can't you just check if its provable in all branches?
11:40:44FromGitter<alehander42> actually you just prove that is not used after consumption
11:40:49FromGitter<alehander42> I have to take a look at it
11:41:42Araqevery function call can 'raise', that's why.
11:42:21Araqand before you say "aha, exceptions bad!!!!!", the 'if error: return/goto errHandler' has exactly the same problem
11:42:41FromGitter<alehander42> so if a function calls only functions with checked side effects? is it easier then?
11:42:54FromGitter<alehander42> no, I like exceptions
11:44:00Araqif we map overflows to exceptions (and what would be the alternative anyway?) every function can raise
11:44:01FromGitter<alehander42> yeah you wrote Furthermore the compiler should probably get even smarter in its inference of raises: [].
11:44:25AraqI changed my mind a bit, raises: [] is a myth
11:44:41Araqit's programming, programs run on physical machines, you can always have
11:44:44Araqoverflows
11:44:49Araqout of heap memory
11:44:49FromGitter<alehander42> and if we map overflow, we get exceptions everywhere?
11:44:53Araqout of stack space.
11:45:58FromGitter<alehander42> well, some of those errors are probably "fatal", so one shouldn't guarantee correctness after them
11:47:48Araqthat works but then you cannot recover from non-critical component failures
11:48:00Araqunless you assume an OS with process management
11:49:43Araqand libraries that willy-nilly call exit because "exceptions are bad, mmkay?" are not that convincing
11:50:24Araqthat remains true when it was the Nim compiler that inserted these trap instructions
11:56:06FromGitter<alehander42> hm true
11:56:14FromGitter<alehander42> I have to read more about that
11:56:38FromGitter<alehander42> ok, how do languages with linear types deal with this
11:56:44*theelous3 joined #nim
12:35:16*dabedoc joined #nim
12:35:18dabedochello
12:38:06*dabedoc left #nim (#nim)
12:43:59*Snircle joined #nim
12:45:52FromGitter<narimiran> hello dabedoc
12:52:47*PMunch quit (Ping timeout: 240 seconds)
12:53:19*PMunch joined #nim
12:56:04PMunchVarriount, it's getting there
12:56:18PMunchIt has quite a bit of functionality already, but it's not very stable
12:56:35PMunchI think there might be an issue in the nimsuggest interface layer that makes it freeze up
13:08:21*btbytes joined #nim
13:08:41*btbytes quit (Client Quit)
13:09:07*btbytes joined #nim
13:24:26*endragor quit (Remote host closed the connection)
13:26:08*martin1__ joined #nim
13:26:31*seni joined #nim
13:27:48*PMunch_ joined #nim
13:28:21*PMunch quit (Remote host closed the connection)
13:28:37*PMunch_ quit (Client Quit)
13:28:47*PMunch joined #nim
13:32:44*dom96_w joined #nim
13:39:27*martin1__ quit (Ping timeout: 244 seconds)
13:40:05*martin1__ joined #nim
13:40:46AraqPMunch, how do you call nimsuggest? as a process or do you use it as a library?
13:41:45PMunchAs a process
13:41:55PMunchIs it available as a library?
13:42:18Araqwell kind of, you would need to replicate the program driver
13:42:51PMunchHuh?
13:43:12*Vladar joined #nim
13:43:25PMunchBasically right now I open nimsuggest in stdin mode and ready and write to the process
13:43:32PMunchOne process for each project file
13:43:44*endragor joined #nim
13:43:59Araqnimsuggest.nim uses the compiler as a library
13:44:05Araqyour program could do the same
13:44:16Araqyes, it uses undocumented hooks
13:44:45Araqwe can document these and ensure they are reasonably stable now that the compiler's hardest refactorings are behind us
13:45:31Araqbtw communication over stdin/stdout is much more fragile than sockets IME
13:45:36PMunchSo basically reverse engineer nimsuggest?
13:46:05PMunchYeah I know, but with sockets I would need to find an unused port for each of the nimsuggest procedures I need to have running
13:46:56Araq"revese engineer" 600 lines of code, yes
13:46:59PMunchBut making a library out of nimsuggest would definitely be the best
13:47:34Araqand most of its complexity is the multi-threading that it does
13:47:48PMunchYeah it's not huge, but still there's some work to be done to make it work
13:47:50Araqand that you can hopefully avoid, but I'm not sure
13:47:52PMunchI'll look into it though
13:47:57*endragor quit (Ping timeout: 252 seconds)
13:48:09PMunchBut I'm probably stuck at work all day today, like I was yesterday..
13:48:16*dom96_w quit (Ping timeout: 260 seconds)
13:49:22Araqsure thing, no worries
13:50:26*btbytes quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
13:54:51*Vladar quit (Remote host closed the connection)
13:56:10*Vladar joined #nim
13:59:21FromGitter<Varriount> PMunch Opening a connection on port 0 causes the system to allocate a random free socket.
14:01:58PMunchOh, that's good to know
14:02:06*theelous3 quit (*.net *.split)
14:02:07*vlad1777d_ quit (*.net *.split)
14:02:07*Sembei quit (*.net *.split)
14:02:07*edcragg_ quit (*.net *.split)
14:02:08*rockcavera quit (*.net *.split)
14:02:08*def- quit (*.net *.split)
14:02:08*oprypin quit (*.net *.split)
14:02:08*livcd quit (*.net *.split)
14:02:08*shadowbane quit (*.net *.split)
14:02:09*copygirl quit (*.net *.split)
14:02:09*Summertime quit (*.net *.split)
14:02:09*so quit (*.net *.split)
14:02:09*def-- joined #nim
14:02:09PMunchBut how do I know which port to connect to then?
14:02:17*copygirl joined #nim
14:02:25*vlad1777d_ joined #nim
14:02:30*rockcavera joined #nim
14:02:31*Summertime joined #nim
14:02:35*livcd joined #nim
14:02:38*edcragg_ joined #nim
14:02:41*oprypin joined #nim
14:03:21*rockcavera is now known as Guest324
14:03:28*Guest324 quit (Changing host)
14:03:28*Guest324 joined #nim
14:03:30*livcd quit (Changing host)
14:03:31*livcd joined #nim
14:03:48*Sembei joined #nim
14:04:14*shadowbane joined #nim
14:04:24*gsingh93 quit (Ping timeout: 246 seconds)
14:04:31*so joined #nim
14:04:31*theelous3 joined #nim
14:06:57*gsingh93 joined #nim
14:09:45*eubn joined #nim
14:24:19*btbytes joined #nim
14:26:45FromGitter<Varriount> You could send that port via stdout, or via about connection
14:27:07FromGitter<Varriount> PMunch: ^
14:30:20PMunchHmm I guess
14:30:28PMunchStill think a library is the way to go thoug
14:32:11FromGitter<Varriount> A library is harder to integrate with.
14:32:47PMunchHow so?
14:33:46PMunchA library with "var nimsuggest = initNimSuggest(projectfile = "somefile.nim"); nimsuggest.sug("somefile.nim", "dirtyfile.nim", 10,4)" syntax should be much easier than socket or stdin communication
14:33:55FromGitter<Varriount> Not all languages have easy C interoperability.
14:34:14FromGitter<Varriount> How would we load this library in JS, or Python?
14:34:42PMunchOh yeah, I'm not saying that nimsuggest should be turned into library, just that it should also be available as a library.
14:35:11Araqthe LSP server would use nimsuggest as a library, the editors use the LSP protocol
14:35:20Araqno language integration woes
14:35:32PMunchI imagine the nimsuggest program should just be read from socket/stdin -> Call corresponding library function -> write out over socket/stdout
14:37:18PMunchI think we should maintain the capability of writing direct nimsuggest plugins
14:37:43PMunchJust add the possibility to interact with nimsuggest from Nim without having to go through a socket/stdin
14:43:50FromGitter<zacharycarter> for anyone using vulkan - https://github.com/zeux/volk
14:48:38*btbytes quit (Quit: Textual IRC Client: www.textualapp.com)
14:49:51*btbytes joined #nim
15:09:50*dom96_w joined #nim
15:16:39FromGitter<kaushalmodi> Is `{'b' .. 'z'} + 'a'` supposed to work/
15:16:40FromGitter<kaushalmodi> ?
15:16:57FromGitter<kaushalmodi> ref: https://nim-by-example.github.io/bitsets/
15:17:21FromGitter<kaushalmodi> I get this error: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5be45381bb8878747499174b]
15:19:18FromGitter<pr-yemibedu> Simon Peyton has a video illustrating Haskell with an extension for linear typing through function restrictions and mentions Rust as being one of the few with a linear type system (using borrowing and ownership) https://www.bing.com/videos/search?q=linear+haskell+&&view=detail&mid=8999700116D3709B94838999700116D3709B9483&&FORM=VRDGAR
15:19:59PMunchkaushalmodi, yes + only works between sets
15:20:29FromGitter<pr-yemibedu> In Rust they try to leave out exceptions and prefer the Result type which includes the error and leaving "panic" for exceptional cases.
15:20:31PMunchSee this: https://nim-lang.org/docs/manual.html#types-set-type
15:20:45PMunchSo if you want to add a unit to a set you need to use incl
15:21:05PMunchWhich is the same as adding a set only containing that one unit
15:21:09FromGitter<pr-yemibedu> so it would be ``` {'b' .. 'z'} + {'a'} ``` basically
15:21:33PMunchSo either {'b'..'z'} + {'a'} or {'b'..'z'}.incl 'a'
15:21:56PMunchpr-yemibedu, sounds like Midori
15:22:54FromGitter<kaushalmodi> PMunch: Thanks. I was just confirming
15:23:08FromGitter<kaushalmodi> Was porting that set of Nim by Examples to my notes which I found that issue
15:23:25FromGitter<kaushalmodi> may be `+` was overloaded for `set[T],T` at some point
15:24:03FromGitter<pr-yemibedu> it could be overloaded. but the explicit nature as it stands makes better sense
15:24:14FromGitter<kaushalmodi> @pr-yemibedu I agree :)
15:24:29FromGitter<kaushalmodi> Wanted to check if anyone knew the history of deprecating that older syntax
15:25:02FromGitter<pr-yemibedu> I don't but it makes the algebra more concrete.
15:25:15AraqI don't recall any changes for + on sets
15:26:12FromGitter<kaushalmodi> Araq: I stumbled upon this when reading https://nim-by-example.github.io/bitsets/
15:26:38FromGitter<kaushalmodi> the table there seems a bit outdated compared to one in https://nim-lang.github.io/Nim/manual.html#types-set-type
15:27:26PMunchMake him a PR: https://github.com/flaviut/nim-by-example/blob/master/content/content/bitsets.md
15:30:29FromGitter<pr-yemibedu> or do you want to add the convenience to the language?
15:34:04FromGitter<pr-yemibedu> speaking of, does a person do a type check on a value or variable for conditional processing? ```
15:34:08FromGitter<kaushalmodi> @pr-yemibedu no
15:34:45*narimiran joined #nim
15:34:46FromGitter<kaushalmodi> We can already do incl and + {..}
15:35:24FromGitter<pr-yemibedu> sounds like a document update then for the existing logic.
15:38:32FromGitter<alehander42> @pr-yemibedu ok, but isn't Rust using affine types: it checks if a value is borrowed
15:38:57FromGitter<alehander42> but it's not an error if it's never borrowed ? (AFAIK?)
15:39:40*dom96_w quit (Changing host)
15:39:40*dom96_w joined #nim
15:39:56dom96_wWould be cool to see this in Nim: https://github.com/ixy-languages/ixy-languages (Any volunteers?)
15:42:19FromGitter<alehander42> oh yeah I saw this today, wondered if they thought of Nim
15:42:22*citycide quit (Ping timeout: 268 seconds)
15:44:05FromGitter<pr-yemibedu> we are talking about passing to a function and you are either moving the reference or you are borrowing the reference. Values are copied.
15:44:55FromGitter<pr-yemibedu> it think the ownership part is needed to balance out notion.
15:50:40FromGitter<pr-yemibedu> I am a newbie for both of these languages so pardon my lack of detailed knowledge in the mechanics or how they cross boundaries exactly. Too many syntax and capability feature matrices in my head.
15:50:43*citycide joined #nim
16:04:59*Trustable joined #nim
16:10:12*darithorn joined #nim
16:21:24FromGitter<alehander42> hmm
16:22:18FromGitter<alehander42> I have to watch the video maybe, but in general it seems harder to achieve
16:22:25FromGitter<alehander42> I have no doubt somebody in haskell has it
16:24:01*Vladar quit (Remote host closed the connection)
16:28:00FromGitter<pr-yemibedu> Just realize that Simon gave credit to Rust team and contributors for being first for having a *practical implementation* of a useful constraint to reason about correctness.
16:31:01FromGitter<Varriount> Meh, depends on your definition of practical.
16:33:58FromGitter<mratsim> I thought formal language like Coq or even Idris had that?
16:35:10FromGitter<pr-yemibedu> practical means the implementation can be used for real world applications where the average effort is not more than the total understanding of the concept involved. Since the language uses this all throughout *safe* code, it is simplified its meaning and it also enables the compiler to help you fine straightforward cases to maintain the constraint.
16:35:28FromGitter<alehander42> I agree, not many people use idris
16:35:35FromGitter<alehander42> btw idris is adding uniqueness types iirc
16:35:41FromGitter<alehander42> this should be interesting
16:35:47FromGitter<pr-yemibedu> Idris does not have it just yet
16:35:52FromGitter<mratsim> I’m sure Ada has that stuff and it powers aerospace
16:37:33FromGitter<alehander42> uniqueness types are like an alt of monads
16:37:35FromGitter<pr-yemibedu> Idris does not guarantee that the use of a ref is limited. It may tell you that you type have a structural constraint.
16:37:57FromGitter<mratsim> ref can escape in Idris?
16:38:22FromGitter<alehander42> for side effects: they seem inspired by Clean, I read some criticism of Clean's uniqueness types practicity tho
16:38:48FromGitter<alehander42> @pr-yemibedu http://docs.idris-lang.org/en/latest/reference/uniqueness-types.html
16:38:58FromGitter<alehander42> > A value with a unique type is guaranteed to have at most one reference to it at run-tim
16:39:16FromGitter<pr-yemibedu> oh so nice. thank you.
16:39:17FromGitter<alehander42> but it does seem newer ⏎ ⏎ > They are inspired by linear types, Uniqueness Types in the Clean programming language, and ownership types and borrowed pointers in the Rust programming language.
16:39:25FromGitter<alehander42> so in a way, you're still correct
16:40:29FromGitter<alehander42> unique types in clean seem fun: ftp://ftp.cs.ru.nl/pub/Clean/papers/2008/vrie08-IFL07-UniquenessTypingSimplified.pdf
16:40:57FromGitter<alehander42> in their example e.g. file operation returns a "new" file handler which is a "different type"
16:41:11FromGitter<pr-yemibedu> pdf not found on server.
16:41:13FromGitter<alehander42> which makes the usage "referentially transparent"
16:41:27FromGitter<alehander42> did this mean that it's always the same if called several times?
16:41:43FromGitter<alehander42> ugh gitter breaks it
16:41:44FromGitter<pr-yemibedu> so the Pony programming language does a lot of work to help to allow some of this. Reference capabilities.
16:41:48FromGitter<alehander42> you have to copy the ftp://
16:41:49FromGitter<alehander42> too
16:41:55FromGitter<pr-yemibedu> okay. got it.
16:43:27FromGitter<alehander42> hm I was thinking of idempotent
16:43:50FromGitter<alehander42> this definition of "pure" seems strange to me, the side effect only applies to the input which is never used again
16:43:56FromGitter<alehander42> but it still changes the actual "world"
16:44:29FromGitter<pr-yemibedu> So this allows Pony to compile to nice little binaries and have all kinds of nice features. https://tutorial.ponylang.io/capabilities/reference-capabilities.html
16:44:48FromGitter<alehander42> I haven't analyzed Pony a lot, how are capabilities different than e.g. a ownership checker
16:45:48FromGitter<pr-yemibedu> so when Simon talks about effecting, He mentions as getting new tokens for channels over getting a sequenced token for transitioning to a *new world*.
16:46:02*mrwonko joined #nim
16:48:42FromGitter<alehander42> seems interesting, I have to play with it to make sense of it
16:48:52xaceI have tried variatons of this, but cant I define a proc in a similiar fashion?: proc isSrt(extension:string): bool = if extension == ".srt": return true else: return false
16:48:57FromGitter<pr-yemibedu> So the Rep Cap of Pony has more transitive states for the data than you see with Rust.
16:49:06FromGitter<alehander42> I don't understand very well how the tag cap helps with sharing data
16:49:49FromGitter<alehander42> so I usually work with ref and val in my actor and then write to iso or val and send this other actors?
16:50:15FromGitter<pr-yemibedu> @alehander42 have you seen the little matrix of read and write of the caps? https://soyyomakesgames.wordpress.com/2016/10/10/ponys-capabilities-made-simple/
16:52:41FromGitter<alehander42> I see, this makes more sense
16:52:49FromGitter<alehander42> so basically they have something like erlang
16:52:57FromGitter<alehander42> but instead of sending copies of data around
16:53:16FromGitter<alehander42> they share it internally most of the time but they protect it using the caps
16:53:21FromGitter<pr-yemibedu> as you notice *tag* sounds like a way to identify something
16:54:11FromGitter<pr-yemibedu> so they can gain performance by not having to copy all the data.
16:54:22FromGitter<alehander42> yes exactly
16:54:32FromGitter<alehander42> sounds cool
16:55:05FromGitter<pr-yemibedu> they also help you to have more control over lifetimes of data so each actor can collect garbage without global state checks.
16:55:43*martin1__ quit (Ping timeout: 272 seconds)
16:55:46FromGitter<alehander42> of course, now I wonder if any of the nim actor libs can somehow have similar type-based impl
16:56:02FromGitter<alehander42> hm, that's also nice
16:56:07FromGitter<pr-yemibedu> the beauty comes at a price.
16:56:53FromGitter<alehander42> what are the cons
16:58:27FromGitter<pr-yemibedu> Pony and Rust and Idris have these semantics all the way down.
16:59:44FromGitter<alehander42> well, that's true. but still I can imagine that in languages like Nim you can at least limit some kind of subset in which you can simulate something similar
17:00:06FromGitter<alehander42> like , how C has subset impl-s with very strong static checkers
17:01:03FromGitter<alehander42> I read that the rust guys had a more pony-like design in the beginning https://www.reddit.com/r/rust/comments/34rszb/pony_type_and_memory_safe_language/cqxkmm8/
17:02:09FromGitter<pr-yemibedu> so the love of Nim having options means that many things needs lots of checking over what is implementing which set of features.
17:02:35FromGitter<pr-yemibedu> you can do many of these things in Nim.
17:03:36FromGitter<alehander42> yeah, I agree, I still find it interesting to experiment with
17:03:47FromGitter<pr-yemibedu> so the limits are that you would need type to be able to describe use constraints (of that the functions can give guarantees they will use the data as prescribed).
17:04:05FromGitter<alehander42> maybe the biggest problem with languages like Pony/Erlang is: they seem very cool for their domain, but I wouldn't want to use them for everything
17:04:31*smt joined #nim
17:04:36*theelous3 quit (Quit: Leaving)
17:05:27FromGitter<mratsim> I think you can subvert the effect system to do capabilities ;)
17:05:42FromGitter<mratsim> or “hasCustomPragma” annotations :P
17:05:47FromGitter<pr-yemibedu> I agree. I like the notion of subsets of the language being Rust like or Pony like to help reason and to get Python like quickness of prototyping an idea.
17:06:52FromGitter<pr-yemibedu> that would attach the capabilities to the function site and not guarantee transition with the data.
17:07:14FromGitter<alehander42> you can attach pragmas to variables
17:07:24FromGitter<alehander42> but I have no idea what they can do
17:07:32FromGitter<pr-yemibedu> a new assignment would loose a capability implicitly. Pony forces you to be explicit about this.
17:08:37*martin1__ joined #nim
17:08:49FromGitter<pr-yemibedu> ```let a {. iso .} = 4 ⏎ let b {. ref .} = ref a``` [https://gitter.im/nim-lang/Nim?at=5be46da162866f74736a2410]
17:09:11FromGitter<alehander42> wow I was just writing the same syntax
17:09:28FromGitter<pr-yemibedu> that makes sense, but the compiler needs more machinery if those caps were to be reversed.
17:09:37FromGitter<alehander42> oh yeah
17:09:52FromGitter<alehander42> but I can still imagine hacks
17:10:00FromGitter<alehander42> e.g. you can have an actor pragma
17:10:11FromGitter<mratsim> pragmas, pragmas everywhere
17:10:28FromGitter<alehander42> for a function and analyze all the var usage inside.. but then you need to invoke only other actor functions
17:10:39FromGitter<alehander42> and you basically reimplement effect checking
17:10:46FromGitter<pr-yemibedu> that is just for locals. so when a val is passed into the procedure, the price has to be paid in a lot of spots beyond simple macro transforms.
17:10:51FromGitter<mratsim> I wonder how far monster went with his Actor library https://forum.nim-lang.org/t/3291
17:12:22FromGitter<pr-yemibedu> his last reply literally mentions Pony as a base for further analysis :) :-) .
17:15:16FromGitter<alehander42> I wonder if somebody did a compiled-to-erlang research typed lang
17:15:25FromGitter<alehander42> also possible to find cool ideas there
17:15:50FromGitter<pr-yemibedu> so does Nim have support for traits? That would allow for other provisions on data.
17:17:13FromGitter<alehander42> btw session types seem like something that could be also very cool
17:17:32FromGitter<alehander42> for an actor-based lang , somehow encoding protocol validity
17:18:37FromGitter<alehander42> found the idea here http://ohad.space/blog/Quartz:%20Towards%20a%20statically%20typed%20Erlang.html
17:18:58FromGitter<pr-yemibedu> http://simonjf.com/2016/05/28/session-type-implementations.html
17:19:20FromGitter<pr-yemibedu> and the Simon Peyton video touches on this.
17:19:25FromGitter<mratsim> traits like Rust? imho concepts are more powerful
17:20:15FromGitter<mratsim> what’s the difference between session types and Generics?
17:20:33FromGitter<pr-yemibedu> the too new for prime time language Hackett has many videos. Some engage this talk about them. author Alexis King.
17:21:06FromGitter<mratsim> actually session types can probably be implemented as concepts: ⏎ ⏎ > Session types are a type discipline for communication-centric programming, where programs can be checked to see if they conform to protocols.
17:21:34FromGitter<mratsim> or design by contracts: https://github.com/Udiknedormin/NimContracts
17:24:07FromGitter<pr-yemibedu> the code contracts vs the data contracts involves a lot of extra typing. Very good in its ability to exist but then starts to make have a ref cap palatable.
17:28:04xaceproc isSrt(extension:string): bool = result = if extension == ".srt": true else: false # anyhow i settled on this, my interpretation is that the single line if statement is only allowed in assignment expressions
17:28:45FromGitter<pr-yemibedu> for clarification of names used to describe what a language *allows* versus what is *based* on. https://gankro.github.io/blah/linear-rust/
17:29:40narimiranxace: `proc isSrt(extension: string): bool = extension == ".srt"`
17:29:52narimiranno need for `if true: true` ;)
17:29:57FromGitter<alehander42> @xace try to not use a lot of stuff on one line, becomes harder to read
17:30:02xacenarimiran: ah, thats true
17:30:41xacealehander42: Yeah, i'm currently trying to learn the language, it's just experimentation
17:31:38mrwonkoHow would I set up my nimble file when I have a frontend and backend, and the backend should statically embed the compiled frontend JS in the binary?
17:32:11mrwonkoI'll probably want separate tasks for compiling the frontend and backend, and then nimble install would first compile the backend and then the frontend.
17:32:54*martin1__ quit (Ping timeout: 252 seconds)
17:35:20mrwonkoAnd what should my directory structure look like in that case? Would I have a separate nimble module for the frontend that the backend depends on?
17:37:06*PMunch quit (Quit: Leaving)
17:38:29FromGitter<pr-yemibedu> Does Nim have something like ``` if typeof(A) == type(string): ``` or ``` if A isA string: ``` ?
17:39:16FromGitter<kaushalmodi> yes
17:39:19FromGitter<kaushalmodi> `is`
17:39:27FromGitter<kaushalmodi> assert A is string
17:39:45FromGitter<pr-yemibedu> excellent. This language is amazing in its simplicity.
17:39:52FromGitter<kaushalmodi> ```A = "abc" ⏎ doAssert A is string``` [https://gitter.im/nim-lang/Nim?at=5be474e83102f145216e0e15]
17:40:23FromGitter<kaushalmodi> `doAssert` is preferred over `assert` if you want asserts that can never be turned off
17:40:39FromGitter<kaushalmodi> also look at typetraits stdlib to see how to print/echo type names
17:40:44*mrwonko quit (Quit: mrwonko)
17:40:51FromGitter<mrwonko> I'm still here, just not in IRC.
17:41:29FromGitter<kaushalmodi> @pr-yemibedu https://scripter.co/notes/nim/#check-type, https://scripter.co/notes/nim/#typetraits
17:41:36FromGitter<pr-yemibedu> thank you. the vscode plugin does not help a lot in seeing the type of a hovered binding.
17:49:18FromGitter<mratsim> @mrwonko, you can have a look at https://github.com/Vindaar/NeuralNetworkLiveDemo or https://github.com/brentp/nim-plotly for packages with a frontend+backend
17:51:27FromGitter<alehander42> Mrwonko you can also read the frontend files on compile time
17:51:37FromGitter<alehander42> In the backend
17:52:53FromGitter<alehander42> But you need to serve them yourself, i think there are libs in other langs that use the same api that filesystens have
18:12:49FromGitter<iffy> How do I expand a seq to be passed in as varargs to a proc?
18:13:04FromGitter<iffy> (in python: foo(*mytuple) )
18:13:28narimiran@iffy there is `varargs`
18:13:59narimiranhttps://nim-lang.org/docs/manual.html#types-varargs
18:14:04FromDiscord_<treeform> Is there some thing like asyncproc that allows me to read from stdout and stderr from another process using the asyncdispatch stuff?
18:15:02FromGitter<iffy> narimiran: right, I have a proc expecting varargs and I want to call it like this: someproc(arg1, *my_seq_of_other_args)
18:16:49*nsf quit (Quit: WeeChat 2.3)
18:17:08narimiran@iffy http://ix.io/1rkR/nim
18:17:51FromGitter<iffy> narimiran: and if I don't get to decide how the proc is defined (it only accepts a single varargs argument)?
18:18:22narimiranwhat exactly would you like to do? ;)
18:19:00FromGitter<iffy> There exists a proc: `proc run(args:varargs[string, `$`])` and I want to call it like this: `run("first arg", seq_of_other_args)`
18:19:42FromGitter<iffy> as a workaround, I'm just creating a new seq with "first arg" in front and calling run with the new seq
18:20:48narimiranyeah, i don't think you can mix single arg and seq[args] if proc expects (just) varargs
18:21:08FromGitter<iffy> okay; good enough; thank you!
18:21:28*floppydh quit (Quit: WeeChat 2.3)
18:30:33*Yardanico joined #nim
18:31:51*dada78641 quit (Ping timeout: 252 seconds)
18:32:14*dada78641 joined #nim
18:32:57*tobbez quit (Ping timeout: 252 seconds)
18:34:31*tobbez joined #nim
18:34:33xacelet s:set[char]={'å','ä','ö','Å','Ä','Ö'} # isn't compiling, any suggestions on how to deal with this?
18:35:54*zachk joined #nim
18:36:23narimiranxace: those are not ascii chars, i guess that's the problem
18:37:04*zachk quit (Changing host)
18:37:04*zachk joined #nim
18:38:31FromGitter<mrwonko> @mratsim thanks for the links, but the NeuralNetworkLiveDemo sadly doesn't use nimble at all, and nim-plotly does not seem to contain any backend (or frontend), just the library (which supports (but does not require) the JS backend)
18:39:03FromGitter<mrwonko> @alehander42 Yes, I was planning on reading the compiled frontend js on compile time, that's what I meant by "statically embed"
18:39:43xacenarimiran: yeah, but i figured since their value is <255 it wouldnt be a problem
18:40:22xacedoes this mean that chars are limited to the ascii length?
18:41:36xacehttps://nim-lang.org/docs/system.html#char says they are 8bit. ill try giving the decimal value
18:41:44*dom96_w quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
18:44:54FromGitter<Vindaar> @mrwonko Indeed, the NN demo requires to manually compile and start the JS and C programs independently. Both could be done via `nimble` though. I can try to implement that later, if you like. ⏎ I'm not sure what you mean by "statically embedding" the JS program though.
18:45:10FromGitter<Vindaar> (the compiling, not the starting of course)
18:46:15FromGitter<mrwonko> I mean once the nim frontend has been compiled to JS, that can be read by the backend at compile time and included directly in the binary, so the web server can serve it to the client from memory.
18:47:12FromGitter<Vindaar> Oh, well in principle that should be "easy"
18:47:24FromGitter<mrwonko> I'll probably just add a task that does `nim js` for building the frontend
18:47:44FromGitter<Vindaar> yes, and then just `staticRead` the build `.js` from the server
18:47:57FromGitter<Vindaar> *built
18:48:08FromGitter<mrwonko> yeah, that will be the easy part
18:49:37xaceconst s:set[char]={chr(229),chr(228),chr(246),chr(197),chr(196),chr(214)} # was accepted
18:51:18*dom96_w joined #nim
18:54:13*btbytes quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
18:55:40*dom96_w quit (Client Quit)
18:55:57*miran joined #nim
18:57:54*Guest324 is now known as rockcavera
18:58:08xacewhere can i find the meaning of these {.magic: "Chr".} pragmas? Where do I find the magic "Chr" definition?
18:58:15*jlhouchin joined #nim
18:58:29*btbytes joined #nim
18:58:34*narimiran quit (Ping timeout: 272 seconds)
19:12:13FromGitter<mrwonko> in asynchttpserver.server, how do I define the listen address? It seems to use 0.0.0.0 by default (which gives me a firewall warning), I only want to listen on 127.0.0.1 instead.
19:13:02FromGitter<mrwonko> ah, serve seems to have an optional address argument
19:13:51FromGitter<mrwonko> yeah, that did the trick
19:14:20Araqxace, it's built into the compiler but there are docs for what it does
19:33:48miranwhen i use `asyncdispatch.runForever()` i get "Error: unhandled exception: No handles or timers registered in dispatcher. [ValueError]", what should i do?
19:45:19FromGitter<mrwonko> I'm using xmlgen for some simple html generation (is that the right tool for that, if I want to keep things simple and stick to the stdlib?), how can I prevent injections?
19:45:29FromGitter<mrwonko> sorry, it's called htmlgen
19:46:55FromGitter<mrwonko> xmldom.escapeXml sounds good
19:53:45FromGitter<mrwonko> I'm looking for the os.getEnv documentation. https://nim-lang.org/docs/os.html lists it in the exports, but clicking on it sends me to https://nim-lang.org/docs/ospaths.html#getEnv which is an anchor that does not exist
20:06:10Araqmiran, before you register any handler, runForever would be an endless loop
20:06:23Araqthe ValueError tells you you're using it wrong
20:06:45FromGitter<mrwonko> if I use strformat.& in my asynchttpserver callback, I get some kind of effect that makes the function signature not match, I think it's something about `{.locks: <unknown>.}`
20:08:35*Yardanico quit (Remote host closed the connection)
20:09:01*Yardanico joined #nim
20:09:19*Yardanico_ joined #nim
20:11:15Araqlocks is irrelevant
20:15:00*theelous3_ joined #nim
20:17:25FromGitter<mrwonko> okay give me a second, it's difficult to reproduce
20:25:00FromGitter<mrwonko> building a minimal reproduction actually gave me a significantly more useful error message, "This expression is not GC-safe.": https://gist.github.com/mrwonko/617bdd870dd8c31b6874b51acf39ecfb in the original context I only got a type mismatch
20:26:52FromGitter<tim-st> it's global, you have to put the string inside a main() proc
20:27:04FromGitter<tim-st> (and also the server etc)
20:29:42*Trustable quit (Remote host closed the connection)
20:31:15*Yardanico_ quit (Remote host closed the connection)
20:36:17FromGitter<alehander42> @iffy you can define a macro apply
20:36:29FromGitter<alehander42> which generates a call, but there isnt a builtin way
20:36:37FromGitter<alehander42> (maybe call is a better name? as in js)
20:36:54*NimBot joined #nim
20:36:59FromGitter<iffy> okay, making my own seq is a better alternative in this case, but I'll keep it in mind for the future
20:39:24FromGitter<mrwonko> oh wow that resulted in a nasty compile error inside the generated C code in clang: member reference base type 'tyObject_FuturecolonObjectType__te3W2Tqi7xuJ7rlPtg9al5w *(tyObject_Request_1Wfxxtqzahm0R6lueM9aQmQ *)' is not a structure or union
20:39:29*nsf joined #nim
20:42:44FromGitter<pr-yemibedu> I am using foldl and supplying `{}` as an initial `set[char]` is not enough. so I have ``` ⏎ let ex = "ab" ⏎ let s:set[char] = {} ⏎ let val:set[char]=foldl(ex, a + {b}, s) ⏎ ``` which works but ``` ... [https://gitter.im/nim-lang/Nim?at=5be49fc347217e07fff0ce80]
20:44:59FromGitter<mrwonko> this is as short as I could get the repro, without the & or the case it doesn't seem to trigger: https://gist.github.com/mrwonko/9e59750c13ad411c9945b918bb8c5a11 I think I'll create an issue on github for this?
20:45:25Araqplease do
20:45:46Araqfrom strformat import `&` is wrong btw
20:46:04Araqimport strformat
20:49:32FromGitter<mrwonko> okay, here it is with the fixed import: https://github.com/nim-lang/Nim/issues/9655
20:49:32FromGitter<tim-st> I think the case together with async proc is the problem, the crash on gcc windows
20:49:58FromGitter<tim-st> 1) same crash
20:51:23FromGitter<mrwonko> Okay, I'll leave it at that for today, see you later.
21:04:05*Yardanico quit (Remote host closed the connection)
21:20:39*nsf quit (Quit: WeeChat 2.3)
21:35:38miranAraq: re runForever and using it wrong: that's the only way i was able to reproduce a bug i'm trying to solve
21:38:31*martin1__ joined #nim
21:40:06FromGitter<kayabaNerve> I put runForever on the first line of my async proc so it'd run forever. Now nothing works. Nim is broken! /s
21:41:18*miran quit (Ping timeout: 245 seconds)
21:41:27*btbytes quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
21:42:02*Ven`` joined #nim
21:42:20*btbytes joined #nim
21:44:43*btbytes quit (Client Quit)
21:46:26*Ven`` quit (Client Quit)
21:49:25*dom96_w joined #nim
21:54:56*dom96_w quit (Ping timeout: 244 seconds)
21:55:54*martin1__ quit (Ping timeout: 252 seconds)
22:41:54*krux02 quit (Remote host closed the connection)
23:19:54FromGitter<zacharycarter> This is slightly off topic - but has anyone done any 3d work with Godot? I'm not very much concerned with whether or not you've used Nim, but the complete lack of 3d examples beyond the platformer demo for Godot, leave me scratching my head in regards to how to do a lot of things with the engine
23:19:59FromGitter<zacharycarter> I'm thinking about just using Unity
23:20:15*chemist69 quit (Ping timeout: 252 seconds)
23:23:53*btbytes joined #nim
23:24:25Araqzacharycarter: 3D is fine with Godot
23:24:39Araqno need to switch to Unity for this reason
23:24:59Araqbut well ... I only know so much, not any details
23:28:47FromGitter<zacharycarter> Araq: it's the capabilities, it's the lack of examples / learning resources
23:28:56*btbytes quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
23:28:57FromGitter<zacharycarter> I can look at a unity project and totally understand how to do something with the engine
23:29:06FromGitter<zacharycarter> but finding a comparable example for godot is not such a trivial task
23:29:55FromGitter<zacharycarter> also - godot does this funky node / component bullshit where the nodes themselves are components instead of individual components being added to nodes
23:30:11FromGitter<zacharycarter> so it's not like you can look at a unity project and immediately understand how to port it to godot
23:33:43Araq"fuky node / component bullshit" is pretty standard, Urho used something like this too.
23:33:54Araqand yeah, I dislike it too
23:34:28FromGitter<zacharycarter> well
23:34:45FromGitter<zacharycarter> with unity - you have a hierarchy of nodes and then each node you add components too
23:34:55FromGitter<zacharycarter> with godot, each node in of itself is a component it seems
23:35:11FromGitter<zacharycarter> so a node in unity can have ten different components doing different things
23:35:18FromGitter<zacharycarter> optimally - working together to achieve something
23:35:32FromGitter<zacharycarter> in godot - you have like, a 3d mesh node / component
23:35:37FromGitter<zacharycarter> it's just fundamentally different / weird
23:36:06FromGitter<zacharycarter> the whole scene graph thing is definitely overused - not every game needs it
23:36:14FromGitter<zacharycarter> but now engines bake it in
23:39:10FromGitter<zacharycarter> https://armory3d.org/manual/#/code/wasm - we might have a new option for game engines with Nim
23:39:27FromGitter<zacharycarter> although I have a feeling without a true LLVM backend - this won't work out so well
23:40:06Araqwe have a true LLVM backend, it's called nlvm
23:40:44FromGitter<zacharycarter> I know of it's existence - I just didn't know what it's status was
23:41:11FromGitter<zacharycarter> Araq: do you know if anyone has played with wasm yet via nlvm?
23:41:42FromGitter<zacharycarter> I guess it might be a better question for the maintainer
23:45:14AraqI know people played with it
23:52:10*theelous3_ quit (Ping timeout: 246 seconds)