<< 21-06-2018 >>
00:44:17*dddddd quit (Remote host closed the connection)
02:01:58*sbdchd joined #nim
02:20:39*sbdchd quit (Remote host closed the connection)
02:28:47FromDiscord<Kayaba> nasusiro dom96 Ara-q also created America and capitalism and actually good socialized healthcare and stopped slavery
02:28:54FromDiscord<Kayaba> He didn't just create Nim
02:29:34FromDiscord<Kayaba> Though he did admit Nim was transpiled...
02:29:34FromDiscord<Kayaba>
02:29:34FromDiscord<Kayaba> I should just start writing "The bullshit history of Nim" instead of texting here at this point.
02:55:54*endragor joined #nim
03:08:06Demos[m]is that vtbl feature coming along?
03:11:15FromGitter<gogolxdong> How to change a string value to its enum name?
03:16:35FromGitter<kaushalmodi> @gogolxdong Have a look at strutils.parseEnum: https://nim-lang.org/docs/strutils.html#parseEnum,string
03:26:21FromGitter<gogolxdong> yes, good to know, it works ,I tried bindSym ,it works too, is that right?
03:46:51*Hotkeys joined #nim
04:24:43*xkapastel quit (Quit: Connection closed for inactivity)
04:27:11FromGitter<codenoid> how i can do this in nim https://stackoverflow.com/questions/50960344/disable-mouse-click-and-change-pointer-icon
04:28:38*thomasross joined #nim
04:28:58*arecaceae quit (Remote host closed the connection)
04:29:22*arecaceae joined #nim
04:37:07*gsingh93_ quit (Quit: Connection closed for inactivity)
04:53:50*Vladar joined #nim
04:57:28FromGitter<Varriount> @codenoid Same way as how you would do it in C - write or wrap a library to interface with the desktop environment.
05:22:24*nsf joined #nim
05:43:18*zahary joined #nim
05:46:09*rauss quit (Ping timeout: 248 seconds)
05:48:46*jjido joined #nim
06:01:26*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
06:20:16*jjido joined #nim
06:26:54*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
06:55:25*mgdelacroix quit (Read error: Connection reset by peer)
07:00:46*msmorgan quit (Ping timeout: 260 seconds)
07:01:43*mgdelacroix joined #nim
07:02:31*mgdelacroix quit (Max SendQ exceeded)
07:09:19*yglukhov[i] joined #nim
07:09:27*yglukhov[i] quit (Remote host closed the connection)
07:10:04*yglukhov[i] joined #nim
07:12:43*mgdelacroix joined #nim
07:12:54*yglukhov_ joined #nim
07:14:34*yglukhov[i] quit (Ping timeout: 260 seconds)
07:14:58*msmorgan joined #nim
07:18:37*PMunch joined #nim
07:47:26*cspar joined #nim
07:48:20*yglukhov_ quit (Remote host closed the connection)
07:48:59FromGitter<mratsim> Nim squad, assemble! https://www.reddit.com/r/programming/comments/8snyn3/the_state_of_type_hints_in_python/
07:49:11*Vladar quit (Quit: Leaving)
08:04:02PMunchmratsim, did you see this: https://instagram-engineering.com/let-your-code-type-hint-itself-introducing-open-source-monkeytype-a855c7284881
08:06:56PMunch"Types are great, let's hack them into Python by force!"
08:07:03*yglukhov[i] joined #nim
08:07:17FromGitter<mratsim> No, but the people at Ethereum have something similar, when running their test suite it adds a comment about types in their code.
08:08:18PMunchWhat do you mean, in what language?
08:12:22FromGitter<mratsim> Python
08:12:43FromGitter<mratsim> @PMunch: https://www.reddit.com/r/programming/comments/8snyn3/the_state_of_type_hints_in_python/e11dw9r/
08:16:54PMunchHmm, what was that UI library by trustetcode(?) called?
08:17:08PMunchAh NiGui
08:18:12FromGitter<mratsim> ok, adding it and IUP
08:18:31PMunchOh yeah IUP
08:21:55FromGitter<mratsim> btw, do you plan to use webview with gene?
08:22:57PMunchWell the idea is to make it extensible
08:23:28PMunchSo anyone can add a new target without adding it directly to the main project
08:24:00PMunchHmm, would be interesting to try webview with Karax
08:24:13FromGitter<mratsim> Ok, that’s because I had troubles yesterday with Cocoa web view: https://github.com/oskca/webview/issues/4
08:24:31FromGitter<mratsim> I wanted to use Webview + Karax ;)
08:24:45PMunchNot a huge fan of this web-page-as-an-app trend, but it's got some merit to it
08:25:18FromGitter<mratsim> Well, for me I have 2 potential project and one absolutely require JS
08:25:31FromGitter<mratsim> 1 is for a UI for a Go Database program, can be done in anything
08:26:17FromGitter<mratsim> The other is an UI for Arraymancer for plots/data analysis, I absolutely need D3.JS https://github.com/d3/d3/wiki/gallery
08:27:00PMunchOooh, that looks nice
08:27:29livcdWhat does the Go database program do and why Go ? :)
08:27:32PMunchWell genui already supports Karax which is JS
08:27:49PMunchNot sure how hard it will be to wrap that in webview, but probably fairly trivial
08:27:54PMunchlivcd, Go the game, not the language
08:28:02FromGitter<mratsim> I have about 100k+ go games (because i was a competitive Go player)
08:28:05ldlework<3 Baduk
08:28:40FromGitter<mratsim> but the most of the current programs are unmaintaned or outdated (Python 2 …)
08:28:47*dddddd joined #nim
08:28:51FromGitter<mratsim> or Delphi* :P
08:29:16FromGitter<mratsim> I’d like something open-source that runs on everything.
08:30:19FromGitter<mratsim> Besides the UI, the main issue I have is 2D pattern search in 100k+ games, that can happen at any point of the game (about 200 moves)
08:30:47FromGitter<mratsim> I tried to look for spatio-temporal indexing for databases but came-up empty :/
08:30:54PMunch2D pattern search?
08:31:00FromGitter<mratsim> yes
08:31:08PMunchAs in looking for a pattern of pieces?
08:31:19dom96mratsim: You don't need JS for D3.js, just use Nim :)
08:31:40FromGitter<mratsim> @dom96 I’m all ears
08:32:27FromGitter<mratsim> @PMunch, for example: https://senseis.xmp.net/?CornerShapes, and those can be mirrored or rotated
08:33:41FromGitter<mratsim> I’ve been looking into R-Tree (supported by SQLite), Kd-Trees, Z-order curve, Hilbert curve (both can transform N dimensions into a 1-d index and are used in Amazon Aurora and Google G2)
08:33:52FromGitter<mratsim> but I’m still fumbling on the temporal part.
08:33:58dom96mratsim: 1. Wrap D3.js in Nim 2. Use the wrapper. 3. ???? 4. Profit
08:34:05FromGitter<mratsim> lol
08:34:15PMunchAaah I see
08:36:19PMunchmratsim, did you see my PM on Gitter?
08:36:34PMunchHmm, I feel like this should be a solved problem..
08:37:04FromGitter<mratsim> no I don’t see your PM
08:37:25PMunchI mean the Russians had an insane volume of Chess games recorded, someone must've tried to solve this at some point
08:37:35FromGitter<mratsim> ah, I resterted gitter and now I see
08:39:38*noonien joined #nim
08:44:46FromGitter<tim-st> when I define `type t1 = wrapObject` and `type t2 = wrapObject` can I make it that my procs only make one of them?
08:44:57FromGitter<tim-st> *match
08:46:26FromGitter<ephja> type hints in a dynamic language. brilliant
08:46:51FromGitter<tim-st> what do you mean?
08:47:38FromGitter<tim-st> yes, it is from a dynamic language
08:48:25FromGitter<tim-st> In worst case I make a new type and add an enum with specific type only to make it differ, but can this be easier?
08:49:06FromGitter<tim-st> like `type t1 = object -> o: wrapObject; id: Enum`
09:02:29*zahary quit (Quit: Leaving.)
09:14:25*zahary joined #nim
09:16:17dom96amazed that this works without any compiler crashes https://www.reddit.com/r/programming/comments/8si75b/how_an_engineering_company_chose_to_migrate_to_d/e11fm3q/
09:16:35dom96I guess you guys have been stressing static[N] for a while :)
09:16:54*Arrrr joined #nim
09:17:16FromGitter<mratsim> more than stressing
09:17:54FromGitter<mratsim> I’ve been building recursive static types: https://github.com/status-im/nim-stint/blob/master/stint/private/datatypes.nim
09:18:25FromGitter<mratsim> Uint256 is build upon 2xUint128 which is built upon 4xUint64
09:18:43FromGitter<mratsim> 256, 128, etc is a static parameter
09:18:54FromGitter<mratsim> we tried with up to 2048 without any issue
09:25:07dom96nice
09:25:48FromGitter<mratsim> There will be an article about it, maybe in July because I think it’s a very nice piece of tech
09:26:09FromGitter<mratsim> and speed-wise the library is within 10% of the fastest C++ + Assembly library we could find
09:26:18FromGitter<mratsim> but it’s portable ;)
09:26:38FromGitter<mratsim> And we have compile-tim evaluation of Uint of arbitrary size which is nifty
09:39:02*floppydh joined #nim
09:39:27Araqmratsim: awesome, just awesome.
09:40:11FromGitter<mratsim> ^^
09:40:59*yglukhov[i] quit ()
09:42:06FromGitter<mratsim> It uses your ForLoopStmt macro as well: https://github.com/status-im/nim-stint/blob/master/stint/private/as_words.nim#L164
09:42:43FromGitter<mratsim> and usage: https://github.com/status-im/nim-stint/blob/master/stint/private/uint_bitwise_ops.nim#L13
09:42:58*yglukhov[i] joined #nim
09:42:58Araqyou need to add {.experimental: "forloopmacros".} to this soon
09:43:18Araqas new features should start in the .experimental state
09:43:32Araq(new policy...)
09:43:53Araqthat doesn't mean I'll remove it again, I won't.
09:45:23FromGitter<tim-st> Does somebody knows an anwer to my question above? maybe with Concept -> self of wrapObject and self.typeName.asString == "NimTypName" ?
09:51:27*xkapastel joined #nim
09:53:55FromGitter<mratsim> @tim-st, ephja was not referring to your comment, but to my article about type hints in python
09:55:18FromGitter<mratsim> I don’t really understand your question: ⏎ ⏎ > when I define `type t1 = wrapObject` and `type t2 = wrapObject` can I make it that my procs only make one of them? ⏎ ⏎ re you looking for distinct types? [https://gitter.im/nim-lang/Nim?at=5b2b7606d2abe466887c1804]
09:55:24FromGitter<mratsim> are*
09:55:40FromGitter<mratsim> type t1 = distinct wrapObject
10:00:13*Vladar joined #nim
10:03:11FromGitter<tim-st> @mratsim oh, ok^^, thanks for the answer. I get objects from a dynamic language and only at runtime the type is known, so at nim side I can only define the most general type `wrapObject`. now when I define the equivalent nim types, I do: `type A = wrapObject` <- because I cannot define the generic object better. I also do `type B = wrapObject` and `type C = wrapObject`. Now this works good. But the problem is, that on
10:03:11FromGitter... nim side all procs match for the types `A` , `B` and `C` because they are defined all on `wrapObject` like proc test(a: A) = shoudnt match test(b) but it does, so I thought a Concept could check on nim side if the type name of nim type fits
10:03:21FromGitter<tim-st> distinct doesnt work here, I tried
10:05:44FromGitter<tim-st> also in my quote "make" should be replaced by "match"
10:07:17FromGitter<tim-st> the problem is only about matching proc signatures. It's not a big problem, but for now too many procs do match for all types, and when a dev calls the proc on the wrong type it's only seen at runtime that the proc doesnt match
10:11:32FromGitter<tim-st> https://gist.github.com/tim-st/c5e2144d7c621d94996e777d54a7e890
10:16:26FromGitter<tim-st> well distinct goes in the right direction but I get an error about nonFinal or something so I cannot use it; I think concept could fix it
10:20:46FromGitter<Yardanico> @notTita are you around here somewhere?
10:21:48FromGitter<tim-st> Maybe getting this into real nim code would exactly the fix:
10:28:42FromGitter<andreaferretti> I have the following in my nimble file
10:28:55FromGitter<andreaferretti> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2b7de6d2abe466887c2d38]
10:29:11FromGitter<andreaferretti> it makes nimble crash on my MacOS with the message
10:29:34FromGitter<andreaferretti> `Error: unhandled exception: VM is not built with 'gorge' support [ERecoverableError]`
10:29:46FromGitter<andreaferretti> this happens since a few days
10:30:10FromGitter<andreaferretti> In theory `when defined(nimdistros)` should be false on a mac
10:30:24FromGitter<andreaferretti> Does anyone know why this can happen?
10:32:02Araqandreaferretti: I need to fix Nimble ....
10:32:11Araqnimble needs to be compiled with -d:nimcore
10:32:20FromGitter<andreaferretti> I see this happen with nim as well
10:32:29FromGitter<andreaferretti> ```when defined(nimdistros): ⏎ echo "hi"``` [https://gitter.im/nim-lang/Nim?at=5b2b7ebdfc606203806ede1e]
10:32:47FromGitter<andreaferretti> If I run just `nim e foo.nims`
10:33:00FromGitter<tim-st> Wow I got it working: https://gist.github.com/tim-st/8e416d2054346b898afd412d55483407
10:33:09Araqhttps://github.com/nim-lang/Nim/blob/devel/compiler/nim.cfg#L8
10:33:21FromGitter<andreaferretti> it seems to be true on a mac
10:35:24FromGitter<andreaferretti> It is just a matter of adding this flag in nimble?
10:35:36Yardanicoto nimble .cfg probably
10:38:57FromGitter<tim-st> Can I turn off compiler type checks for a small code range?
10:43:27FromGitter<mratsim> congrats, seems like a super concept hack the c.type.name part lol
10:43:57FromGitter<mratsim> but no you can’t turn type checks off, if you want to erase types, use inheritance, or concepts or object variants
10:45:03FromGitter<tim-st> Everything works, but now at creating of these alias types at the newT() procs it fails :(
10:45:22FromGitter<tim-st> I tried ` {.push checks: off.}` but nothing happens
10:45:47FromGitter<tim-st> *not
10:46:12FromGitter<tim-st> although of cource I can prove nothing bad happens there
10:46:53FromGitter<mratsim> checks is about overflow, bounds/index checks
10:47:11FromGitter<tim-st> hm, ok, I need to hack my hack^^
10:47:57FromGitter<tim-st> Ah, I have a solution: I use and not type name in [..] where these are the concurren types :)
10:48:15FromGitter<tim-st> Now I made pythons lib type safe
10:49:42FromGitter<mratsim> you should generate that with a template probably to avoid copy-pasting everywhere
10:50:20FromGitter<tim-st> yes, you're right, thanks!
10:55:35*rauss joined #nim
10:59:13*elrood joined #nim
11:06:44*Hotkeys quit (Quit: Connection closed for inactivity)
11:06:49*themagician joined #nim
11:17:13Yardanicodom96, I'm trying to remove my post on forum and it says "You cannot delete this post", what can be the reason of this?
11:19:14FromGitter<mratsim> I’d like to delete the “Hey Mom” forum post as well ...
11:21:17*zahary quit (Quit: Leaving.)
11:21:17Yardanicohttps://forum.nim-lang.org/t/191 this is history :)
11:21:55*bevo quit (Ping timeout: 260 seconds)
11:22:07*zahary joined #nim
11:25:00ArrrrNobody could predict the shitstorm and butthurt that decision would produce eventually
11:25:48Arrrr!echo NimVersion
11:26:02*yglukhov[i] quit (Ping timeout: 245 seconds)
11:27:26*yglukhov[i] joined #nim
11:27:50Yardanico!eval echo NimVersion
11:27:52NimBot0.18.0
11:41:39FromGitter<bevo009> You guys seen this? https://www.spiria.com/en/blog/web-applications/webassembly-nim
11:43:07*Bevo joined #nim
11:44:14PMunchYeah I read that the other day
11:44:21PMunchGoes the route through C though
12:02:13FromGitter<tim-st> is it intended that primitive alias type get an own typename and object get not? https://gist.github.com/tim-st/00388cc9b325674f8532a74cb8d339a5
12:02:25*birdspider joined #nim
12:03:40FromGitter<kaushalmodi> Can the bot also do a nightly devel build? :) Then we could have "!evalDevel" (or may be "!evilDevil", or may be just "!evil")
12:04:22*PMunch quit (Ping timeout: 268 seconds)
12:06:51FromGitter<mratsim> @kaushalmodi the repo is here: https://github.com/zacharycarter/nim-playground
12:07:27*PMunch joined #nim
12:08:18FromGitter<alehander42> or you can just call it !evalNightly
12:08:50FromGitter<mratsim> !IFeellucky
12:09:21FromGitter<kaushalmodi> @mratsim oh, so the "!eval" keyword uses Nim Playground?
12:10:09FromGitter<mratsim> Yes, unless it changed recently
12:11:16FromGitter<kaushalmodi> Answering to self.. yes.. see `evalCode` in https://github.com/nim-lang/nimbot/blob/master/src/playground.nim
12:14:39FromGitter<alehander42> is UniqueInteger different than distinct int
12:16:27FromGitter<ephja> @alehander42 it's a type alias
12:16:50FromGitter<tim-st> the string output, yes. I think that's expected, at least that's why people use the module `typetraits`, I want this for the object type, too
12:17:13FromGitter<tim-st> but there it doesnt work, that's not consistent
12:17:34*nsf quit (Quit: WeeChat 2.1)
12:19:30FromGitter<ephja> inconsistent*
12:20:26FromGitter<tim-st> that's the same
12:37:55Araqmy all-time favorite word, "inconsistent"
12:46:48Araqandreaferretti: does -d:nimcore make any difference?
12:59:59FromGitter<Varriount> @kaushalmodi Interestingly enough, the Windows 10 command prompt has at least some support for virtual terminal color codes.
13:01:14FromGitter<kaushalmodi> I know, but looks like Nim supports only legacy windows console API. Supporting Windows 10 would need a dedicated Windows 10 dev :)
13:03:36FromGitter<mratsim> https://medium.com/@cscalfani/goodbye-object-oriented-programming-a59cda4c0e53
13:03:44FromGitter<Varriount> Well, terminal.nim appears to use this: https://docs.microsoft.com/en-us/windows/console/setconsolemode in conjunction with ENABLE_VIRTUAL_TERMINAL_INPUT
13:04:30FromGitter<mratsim> > The problem with object-oriented languages is they’ve got all this implicit environment that they carry around with them. You wanted a banana but what you got was a gorilla holding the banana and the entire jungle.
13:23:07FromGitter<andreaferretti> @Araq you are right, it was enough to put `-d:nimcore`
13:23:14FromGitter<andreaferretti> I tested by putting it in koch
13:23:33Araqok, thanks
13:23:35FromGitter<andreaferretti> but maybe it makes more sense to put into some nimble configuration
13:23:45Araqbut then why did Nim have the same problem?
13:23:49Araqit's in Nim's config...
13:24:00FromGitter<andreaferretti> it didn't, my fault
13:24:05Araqok
13:24:06*endragor quit (Remote host closed the connection)
13:37:17FromGitter<Varriount> @kaushalmodi It appears that some of the terminal code enables vt100 escapes: https://github.com/nim-lang/Nim/blob/devel/lib/pure/terminal.nim#L858
13:40:21FromGitter<kaushalmodi> Yes, I see. That's only if the true color is enabled, right? And that only supports setting the true colors, not the styles.
13:51:12*zahary quit (Quit: Leaving.)
13:57:26*birdspider quit (Quit: Leaving)
14:15:44*PMunch quit (Quit: Leaving)
14:23:08FromGitter<mratsim> @dom96, this one is for you: https://www.reddit.com/r/programming/comments/8snyn3/the_state_of_type_hints_in_python/e11oiif/
14:29:54*zahary joined #nim
14:30:52FromGitter<mratsim> Today I Learned that Nim was a braces language: https://www.reddit.com/r/programming/comments/8snyn3/the_state_of_type_hints_in_python/e11onxc/
14:32:28FromGitter<tim-st> Well, this is exactly what I told before: If you put non standard nim examples on the nim start page, people even think nim is a "braces language"
14:33:05FromGitter<tim-st> I mean this `for line in stdin.lines:`
14:33:28FromGitter<tim-st> when I first saw this I decided to not install nim
14:33:40FromGitter<ephja> huh
14:33:46FromGitter<mratsim> there is no braces?
14:34:23FromGitter<mratsim> Where I find a lot of braces is for the pragma for C/C++ in the Features page
14:34:50FromGitter<tim-st> yes, braces is in fact very weird, but I think many python people like to see easy python like code on the first view, especially no namespaces should be "just there"
14:34:57FromGitter<tim-st> "stdin.lines"
14:35:39FromGitter<mratsim> this is method call syntax, not braces though
14:36:15FromGitter<tim-st> I know, maybe it was the json thing
14:36:31FromGitter<ephja> the "namespace" being the global "stdin" File instance or what?
14:36:33Yardanicowell, people should already know what "stdin" is
14:36:37FromGitter<tim-st> the json syntax looks offensive, I agree
14:36:48FromGitter<mratsim> I think it’s the pragma at the bottom of this page: https://nim-lang.org/features.html
14:37:43FromGitter<tim-st> I think these are good examples. but not as first examples. Maybe as 5. example
14:38:04FromGitter<tim-st> the first three should show how easy nim is without pragma thinks and stuff
14:38:12FromGitter<tim-st> *things
14:40:24FromGitter<mratsim> depends on the audience
14:40:40FromGitter<mratsim> Nim also targets people looking for a better C or C++
14:46:50*sz0 joined #nim
14:47:06*zahary quit (Quit: Leaving.)
14:58:42Araqthe "Easy C, C++ and Objective C wrapping" example could be improved
14:59:01Araqit actually looks quite tedious and can be simplified quite a bit
14:59:11FromGitter<mratsim> ^
14:59:19FromGitter<mratsim> maybe pick something else than SFML
14:59:32FromGitter<mratsim> though it’s bonus point for game devs
15:00:51*Trustable joined #nim
15:02:35FromGitter<mratsim> an image or a sound library instead?
15:04:43Araqgraphics.nim
15:05:35*cspar quit (Ping timeout: 240 seconds)
15:07:32*yglukhov[i] quit (Ping timeout: 276 seconds)
15:11:32*PMunch joined #nim
15:13:10*yglukhov[i] joined #nim
15:13:45*nsf joined #nim
15:21:10*tsoernes joined #nim
15:22:22tsoernesWhat's the equivalent of Python's namedtuple? I know regular tuples can be given field descriptors/names in nim, but I'm looking to create a tuple constructor with fixed names
15:23:39AraqI don't know what Python's namedtuple is
15:24:41tsoernes"Point = namedtuple('Point', 'x, y'); p = Point(1, 2); print(point)" >>> Point(x=1, y=2)
15:25:03tsoernescan be accessed "p.x" or "p[0]"
15:25:31FromGitter<kaushalmodi> tsoernes: How is that different from Nim tuples with named fields?
15:26:18tsoernesYou have to specify the field names on instantiation
15:26:26Araqlib/pure/times.nim(794, 29) Error: undeclared identifier: 'CTime'
15:26:26AraqError: call to nim compiler failed
15:26:26AraqError: execution of an external program failed: './tools/niminst/niminst --var:version=0.18.1 --var:mingw=none csource --main:compiler/nim.nim compiler/installer.ini'
15:26:40AraqI love Ansi Escape sequences.
15:27:00FromGitter<kaushalmodi> Araq: heh
15:27:03AraqThey keep ruining my outputs.
15:27:19FromGitter<kaushalmodi> there can be a switch for nocolor
15:27:36Araqthere is a switch for nocolor
15:27:38FromGitter<kaushalmodi> that renders styledEcho to behave just like echo
15:27:51FromGitter<kaushalmodi> oh, then terminal.nim should start respecting that?
15:27:57Araqyeah, I need to update my .travis
15:28:10Araqwell no, --colors:off is a Nim compiler switch
15:28:33Araqnot a terminal.nim thing
15:29:21FromGitter<kaushalmodi> ah, OK. I thought you were referring to the current changes in terminal.nim
15:30:08Araqanyhow it's so annoying. now every tool needs to grow a --colors:auto|on|off switch, it's the typical Unix crap. forcing this friction onto every program I write.
15:30:14FromGitter<kaushalmodi> tsoernes: You don't need to specify fields on instantiation (but doing so is a good practice). https://scripter.co/notes/nim/#named-tuples
15:30:42FromGitter<kaushalmodi> *specify field names
15:31:18Araqis it signal handler safe? do I check for istty(stdout)? is it a deamon or not? do I use stdout or stderr?
15:31:39tsoernesthanks, that's what I was looking for.
15:33:02Araqit's pointless features forced down my throat. Contrast this with C++'s design where I can occasionally really ignore a feature's existence. Not that I like C++ but the Unix fans surely have the objectively worse design
15:33:38FromGitter<tim-st> Araq: is it possible in nim to get the type alias name?
15:37:46Araqand then they dance around and spread FUD about how exceptions are bad and that an assignment that actually respects an object's invariants leads to "subtle" code. Hey, your favourite OS embraces asyncronous exceptions. they are called signals and are much worse than regular exceptions.
15:39:13Araqtim-st: I don't know. use a 'distinct' type if the type distinction is so important for you.
15:39:56FromGitter<tim-st> the whole problem is, that the distinct type doesnt work for nonFinal types, that why I do this "quirky" thing
15:40:34FromGitter<tim-st> In fact as I have shown before if I only have the alias name I solved that distinct also works for nonFinal
15:40:57FromGitter<tim-st> but alias name only works for primitive don't know why...
15:41:19Araqthe compiler needs to resolve object type aliases early because you know
15:41:29Araqthe compiler needs to determine if obj.foo is valid
15:41:44FromGitter<tim-st> ah, ok, hm...
15:41:53Araqprobably it's easy to fix but then I need to ensure these aliases are always respected everywhere
15:41:59Araqand that's a pita
15:42:07FromGitter<tim-st> ok, thanks
15:44:53tsoernesAny better way to define matrices? This is pretty verbose "var neighs1: array[rows, array[cols, array[7, array[2, int]]]]"
15:45:48*lain0n joined #nim
15:52:56FromGitter<mratsim> @tsoernes, use Arraymancer (disclaimer I’m the author)
15:54:06FromGitter<mratsim> https://mratsim.github.io/Arraymancer/tuto.first_steps.html
16:21:03FromGitter<kaushalmodi> How can I package an optional related .nim file in a nimble package. The main nimble file does not need that extra.nim, but a user might import it. That extra.nim is useless by itself, so not worth a separate package..
16:22:21FromGitter<kaushalmodi> Right now, nimble gives a warning for keeping that extra.nim in the nimble structure. I can also not skip it, but I want `nimble install` to install it.
16:22:30FromGitter<kaushalmodi> s/but/because
16:24:42FromGitter<mratsim> @kaushalmodi like here: https://github.com/status-im/nim-stint/tree/master/stint
16:24:51*rauss quit (Ping timeout: 240 seconds)
16:25:00FromGitter<mratsim> lenient_stint is optional and can be imported with `import stint/lenient_stint
16:25:38FromGitter<mratsim> basically root/packagename/optionalpackages.nim
16:25:54FromGitter<mratsim> and you have a root/packagename.nim
16:26:35FromGitter<mratsim> alternatively you have a `root/src/packagename.nim` and `root/src/packagename/optionalpkg.nim `
16:27:55FromGitter<kaushalmodi> @mratsim Thanks! I will follow that
16:28:08*rauss joined #nim
16:28:14FromGitter<kaushalmodi> The nimble README suggests putting everything in an `src/` dir
16:28:21FromGitter<kaushalmodi> so things were getting quite messy
16:28:51*lain0n quit (Quit: Leaving)
16:30:22*zahary joined #nim
16:31:25*floppydh quit (Quit: WeeChat 2.1)
16:34:02FromGitter<kaushalmodi> @mratsim Confirming that that structure works. Thanks!
16:34:24FromGitter<mratsim> It took me a year to get it right ;)
16:35:54*nsf quit (Quit: WeeChat 2.1)
16:39:29*tsaka__ joined #nim
16:40:10tsaka__How can you get the nim compiler to output a file with extension so that compiled files can be ignored with gitignore?
16:42:26*yglukhov[i] quit (Remote host closed the connection)
16:49:36*drs joined #nim
16:53:29*yglukhov[i] joined #nim
16:55:55*sz0 quit (Quit: Connection closed for inactivity)
16:58:31*yglukhov[i] quit (Ping timeout: 268 seconds)
16:59:03FromGitter<Quelklef> @Varriount ooh yikes hi I'm a little late but what do you think of Lilt? It still needs work but I never really found a use for it a kind of dropped it :/
17:11:05*tsaka__ quit (Ping timeout: 245 seconds)
17:12:12FromGitter<Quelklef> What's the reason for giving command calls, `()` calls, infix calls, prefix calls, and accQuoted calls each their own AST node?
17:12:32FromGitter<Quelklef> I know it'd help with error messages and such but is there any, I guess, functional reason?
17:13:58FromGitter<Quelklef> I'd have expected them all to be syntactic sugar for some overarching "call node"
17:15:41FromGitter<kaushalmodi> If I need the compiled C to contain `#include "foo.h"`, what do I need to put in the Nim code?
17:17:03FromGitter<kaushalmodi> For example: this does not create `#include "emacs-module.h"` in the generated C files: https://github.com/yuutayamada/nim-emacs-module/blob/master/emacs_module.nim
17:18:14FromGitter<kaushalmodi> Looking at its .nimble, it was probably created when nim 0.14.x was out.. looking at that code, does something need to be changed for it to work on nim 0.18.1 (devel)
17:22:19FromGitter<Quelklef> why exactly are you trying to do this?
17:23:15FromGitter<kaushalmodi> @Quelklef I am trying to use the Emacs Dynamic Modules.. It's a learning exercise.. with this working, I can call procs defined in Nim from Emacs.
17:23:53FromGitter<kaushalmodi> References: http://diobla.info/blog-archive/modules-tut.html, https://mrosset.github.io/emacs-module/
17:28:17*yglukhov[i] joined #nim
17:28:57*rockcavera quit (Ping timeout: 245 seconds)
17:33:27*yglukhov[i] quit (Remote host closed the connection)
17:33:39*yglukhov[i] joined #nim
17:43:33*rockcavera joined #nim
17:44:16*drs left #nim (#nim)
17:49:11*noonien quit (Quit: Connection closed for inactivity)
17:56:11AraqQuelklef: good question. The idea is that in DSLs you can distinguish between these different syntaxes
17:57:23*yglukhov[i] quit (Remote host closed the connection)
18:00:38*tsaka__ joined #nim
18:00:39FromGitter<Quelklef> Ah, cool!
18:06:19FromGitter<kaushalmodi> Araq: I finally got that Nim emacs module compiling.. though on v0.18.0. Here is the diff of the generated .c: https://gist.github.com/kaushalmodi/77697e1f345d5ba5adb399fdbd0aaff5
18:06:45FromGitter<kaushalmodi> Note that the #includes and much more goes missing in the devel version compile.
18:07:04FromGitter<kaushalmodi> Should I open an issue to track this?
18:07:16*yglukhov[i] joined #nim
18:09:46Araqno, you're holding it wrong.
18:10:01Araq0.18 with --deadCodeElim:on will produce something quite similar, I bet
18:10:07FromGitter<kaushalmodi> ?
18:10:15Araqyou need to use the pragmas differently
18:10:29AraqNim is allowed to optimize away anything you don't use
18:11:50*yglukhov[i] quit (Ping timeout: 255 seconds)
18:12:11FromGitter<kaushalmodi> Sorry, I am out of depth when it comes to C code. I was hoping to use that library to use Nim written procs in Emacs. Can you hint to the changes needed in https://github.com/yuutayamada/nim-emacs-module/blob/master/emacs_module.nim and https://github.com/yuutayamada/nim-emacs-module/blob/master/emextra.nim? Or, are the needed changes too involved?
18:13:21shashlickdom96: moderated forum posts show up in the RSS feed - is this intentional?
18:16:29FromGitter<kaushalmodi> Araq: Or on the flip side, any way to disable dead code elimination on devel? I see that `--deadCodeElim:off` doesn't work on devel
18:16:53Yardanicono
18:16:59Yardanicoit should and will be always enabled.
18:17:08Yardanico(well, that's Araq's opinion but I agree with him)
18:17:22Yardanicowhy would you ever want to disable dead code elimination?
18:17:49*Trustable quit (Remote host closed the connection)
18:18:15FromGitter<kaushalmodi> OK, then I will need help to fix the nim-emacs-module library to work on nim devel, else I would be stuck using 0.18.0 for that :(
18:19:01FromGitter<kaushalmodi> Yardanico: See https://gitter.im/nim-lang/Nim?at=5b2be91b960fcd4eb9175615 for why I asked for disabling dead code elim
18:19:35YardanicoI see that, but that happens because you don't use the types exported by that .nim file
18:20:10*craigger quit (Quit: bye)
18:20:51Araqkaushalmodi: the huge .emit looks suspicous, other parts look good
18:21:09Araqcan't be hard to fix
18:22:24*craigger joined #nim
18:22:49FromGitter<kaushalmodi> Yardanico: ⏎ ⏎ > I see that, but that happens because you don't use the types exported by that .nim file ⏎ ⏎ I wish I knew how to fix that :) [https://gitter.im/nim-lang/Nim?at=5b2becf9960fcd4eb9176210]
18:27:09YardanicoYou would ALWAYS include these headers (at global level), not in the type section
18:27:50Yardanicohm
18:29:47FromGitter<kaushalmodi> Yardanico: What about these missing functions starting here? How do I retain those?: https://gist.github.com/kaushalmodi/77697e1f345d5ba5adb399fdbd0aaff5#file-diff_nim_0_18_0__devel-diff-L196
18:31:06Yardanicowhy would you keep them if you don't use them?
18:31:09YardanicoI don't really get it
18:31:21YardanicoI understand that you want to have types and headers anyway
18:31:56*bozaloshtsh joined #nim
18:32:55FromGitter<kaushalmodi> With those functions removed, that C file fails compilation: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2bef575862c35f47bbc9de]
18:35:05bozaloshtshWhen I try compiling the program "import typeinfo" with --debugger:native, I get "error: unknown type name 'TNimType'"
18:35:20FromGitter<kaushalmodi> I don't understand this 100%, but dead code elimination should not make the C code not compile at all
18:35:39*lainon joined #nim
18:40:56FromGitter<kaushalmodi> Yardanico: ⏎ ⏎ > You would ALWAYS include these headers (at global level), not in the type section ⏎ ⏎ How do I do that? `{. header: <emacs-module.h> .}` ? [https://gitter.im/nim-lang/Nim?at=5b2bf1375862c35f47bbcf72]
18:41:05YardanicoI'll try to do that now
18:41:55FromGitter<kaushalmodi> Yardanico: If you like to reproduce this compile error (does not need Emacs to be installed), see https://github.com/kaushalmodi/nim-emacs-module#usage
18:43:35*lompik joined #nim
18:44:32*ketralnis quit (Quit: Coyote finally caught me)
18:45:54FromGitter<kaushalmodi> nope.. gives this error: "Error: attempting to call undeclared routine: 'header'"
18:46:53Yardanicowell, you can just do {.emit: "c code".}
18:47:07Yardanicoso you can emit the #include statements
18:51:49FromGitter<kaushalmodi> looks like I need to add the emit pragma in all the Nim files.. a sample.nim imports the emacs_module.nim, then I need to add the emit in both. Even then, I am stuck with that same compile error: ⏎ ⏎ > Error: execution of an external compiler program 'gcc -c -w -I../include/emacs25/ -std=gnu99 -fPIC -I/home/kmodi/stow/pkgs/nim/devel/lib -o /home/kmodi/stow/pub_dotfiles/e
18:51:49FromGitter... macs/dot-emacs.d/misc/nim-emacs-module/test/nimcache/emacs_module_sample.o /home/kmodi/stow/pub_dotfiles/emacs/dot-emacs.d/misc/nim-emacs-module/test/nimcache/emacs_module_sample.c' failed with exit code: 1 ⏎ ⏎ /home/kmodi/stow/pub_dotfiles/emacs/dot-emacs.d/misc/nim-emacs-module/test/nimcache/emacs_module_sample.c: In function ‘emacs_module_init’: ... [https://gitter.im/nim-lang/Nim?at=5b2bf3c5479ca26689735c10]
18:53:15FromGitter<kaushalmodi> This is the function in that to-be-compiled sample.so that I would want to call from Emacs
18:53:33FromGitter<kaushalmodi> So there is no caller of that function yet
18:53:47Yardanicodid you do {.exportc.} ?
18:55:36FromGitter<kaushalmodi> In sample.nim, I have: ⏎ ⏎ ```emacs.defun(mod_test_return_t, 1): ⏎ env.intern(env, "t".cstring)``` [https://gitter.im/nim-lang/Nim?at=5b2bf4a8aeb8fa0c0734b7e5]
18:56:02FromGitter<kaushalmodi> In that package's, emextra.nim, I have: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2bf4c27da8cd7c8c614b4b]
18:57:07FromGitter<Yardanico> can you try to add {.exportc.} pragma to fsym?
18:58:42FromGitter<kaushalmodi> I believe it should look like this? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2bf5625862c35f47bbda70]
18:58:52FromGitter<Yardanico> well, I'm just experimenting
18:59:03FromGitter<kaushalmodi> no it doesn't
18:59:09FromGitter<Yardanico> @kaushalmodi and did you know that you can actually specify template arguments?
18:59:29FromGitter<Yardanico> So you don't need to put argument types in docs :)
18:59:47FromGitter<kaushalmodi> I am not the original author of this library.. I don't understand the Nim<->C interface in this lib
19:00:13FromGitter<kaushalmodi> I'm just trying to revive https://github.com/yuutayamada/nim-emacs-module on Nim devel
19:00:34FromGitter<Yardanico> ok, another attempt - add .stdcall,exportc,dynlib. pragmas
19:01:02FromGitter<Yardanico> and are you compiling with --app:lib if you want to create a .so file?
19:05:38FromGitter<kaushalmodi> Yes, I am using --app:lib. I tried what you said, but the same error.
19:05:45FromGitter<kaushalmodi> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2bf709467bd7268c27f537]
19:09:15*lompik quit (Ping timeout: 268 seconds)
19:10:21FromGitter<kaushalmodi> Yardanico: While it still fails.. the diff of the exported C seems to be improved than before: https://gist.github.com/cda626a8896c5628f12c36f2e387fdeb
19:11:05*lainon quit (Ping timeout: 245 seconds)
19:11:25FromGitter<kaushalmodi> The includes now appear without using emit pragmas
19:12:54FromGitter<kaushalmodi> May be this gives you a hint of the issue (before: 0.18.0; after: devel): ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2bf8b65862c35f47bbe77d]
19:13:12FromGitter<kaushalmodi> Note that the `nimEmacs_sample_mod_test_return_t` symbol is lost
19:14:05FromGitter<tim-st> Can nimble also support libs written with nimpy -> require python ?
19:14:58Yardanicowell, it supports nimscript
19:15:20Yardanicoso if you can check that python is installed via nimscript, you can do that with nimble
19:16:27*Vladar quit (Quit: Leaving)
19:18:09FromGitter<tim-st> Ok, thanks, and I wondered what people think about it. wrapping some big python libs is quite easy and if they import c libs it's not much slower than wrapping the c libs, but it's seems much less pure nim, so I wasnt sure if it's ok to have in nimble
19:19:23FromGitter<tim-st> maybe with prefix `nimpy` so people see it directly
19:23:05shashlick@tim-st: i was kind of thinking in similar lines
19:23:31shashlickto write a tool that converts python wrappers of C code to be called by nim directly
19:25:53shashlicki did something similar with nimkerberos which wraps the C code of the winkerberos python module
19:26:29shashlickbut of course, it just exposes the C, rather another tool that can convert the python portions into pure nim which calls into the C through the nim wrapper
19:28:27FromGitter<tim-st> I just thought the python ecosystem is so huge that it often makes more sense to wrap these libs especially if the integration is natively. If there later exists a pure nim package the code can be replaced quite easy. I implemented the python nlp system spacy, and everything works fine, and in python I know that know memory leaks appear, in c i dont know
19:31:22*pigmej quit (Ping timeout: 260 seconds)
19:31:26FromGitter<tim-st> (this means wrapping c, not nims code^^)
19:33:49FromGitter<rayman22201> Hi! I have some random questions. First. Does anybody know if there was progress with this thread? https://forum.nim-lang.org/t/3886 ⏎ I am just really curious if there is a way to make Nim put a struct def in the RO section of a binary.
19:35:06*pigmej joined #nim
19:36:59*Arrrr quit (Remote host closed the connection)
19:37:23*vegax87 quit (Ping timeout: 276 seconds)
19:44:44FromGitter<Varriount> @Quelklef I was going to use it for a parser, but the documentation is a bit lacking... Also, I don't know how to do things like "I need to match this rule or that rule".
19:45:50FromGitter<Quelklef> @Varriount If you're still interested, I'd be happy to work with you and work on it. Especially since with an actual use case, it'll be much more directed development
19:46:05FromGitter<Quelklef> I admit the documentation is iffy, it's my first "real" project with real docs :P
19:48:14*nsf joined #nim
19:48:22FromGitter<rayman22201> lol. Crickets on that one. Second question. Mainly directed at @Araq: On the github issue for the new deterministic runtime, you talk about improving the C++ target support because of exception handling performance. I'm all for that, because it's good to have all the targets well tested and supported, but why can't the Nim home-rolled exception handling be made fast, by doing the "zero-cost" exception thing that
19:48:22FromGitter... LLVM does? This is may be a stupid question, but I don't know a lot about how exception handling systems are implemented. ⏎ see: http://llvm.org/docs/ExceptionHandling.html
19:52:16FromGitter<kaushalmodi> Yardanico: Finally this worked! ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2c01f06ceffe4eba264bd5]
19:52:31FromGitter<kaushalmodi> the `exportc` placement is order dependent!
19:52:50FromGitter<kaushalmodi> earlier I was putting it after that `extern`, and that did not work
19:53:31FromGitter<Yardanico> well, can you run the module from emacs tho?
19:54:47FromGitter<kaushalmodi> hmm.. haven't reached there yet.. give me a min :)
19:55:38FromGitter<kaushalmodi> works! ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2c02ba0168e70c08e1ae96]
19:58:46*vegax87 joined #nim
20:09:09*jjido joined #nim
20:14:23*Bevoss joined #nim
20:15:58Araqrayman22201: pure C has no option to support zero-cost exception handling, that's a C++ invention.
20:16:37Araqwe could map exceptions to propagating 'if' statements though and that would likely be better than the current setjmp based implementation
20:17:01Araqthat's a planned AST to AST pass, scheduled for v1.x
20:18:14Araqexportc is not order dependent, you need to export what should be visible by other C code.
20:18:29Araqand it only applies to procs really.
20:19:34*cspar joined #nim
20:24:06shashlickwho are the moderators for the forum?
20:24:58FromGitter<kaushalmodi> Araq: here is partial Nim snippet: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b2c099ad2abe466887ddbf3]
20:26:13Yardanicoshashlick, Araq and dom96 are admins afaik
20:26:21Yardanicowell, AFAIK moderator = admin on the forum
20:26:26Yardanicobut I might be wrong :)
20:26:36Yardanicomaybe Varriount is a moderator too
20:27:52FromGitter<rayman22201> @Araq, couldn't you emulate zero-cost exception handling effectively by building the exception tables at compile time and using gotos? That's what it seems like LLVM is doing at the IR level. A nested if AST pass sounds much easier to implement, and more has more idiomatic Nim flavor. "mmmm delicious TM".
20:28:37Demos[m]I'm getting a lot of errors from system modues when I run nim doc2
20:29:23FromGitter<ephja> you should release 2.0 before 2025 because AI will write all our software by then, surely ;)
20:30:30FromGitter<rayman22201> all hail our new robot overlords?
20:33:25FromGitter<ephja> Nim 2.0 ... AGI ... YotLD
20:34:10YardanicoDemos[m], I don't get any errors
20:34:12Yardanicojust some warnings
20:46:19Araqrayman22201: these tables would need to be indexed by the 'pc' register or some equivalent which is not available in C
20:48:27AraqC is not really much of low level language, its benefits over the JVM are mostly the support for packed data structures (array of structs of arrays without pointer indirections) and the lack of RTTI
20:51:49FromGitter<rayman22201> inline asm for all the architectures! Should be no problem a weekend project :-P
20:52:15Yardanicoall?
20:52:30Yardaniconim is made to be very portable
20:53:22shashlicksome moderated messages on the forum, need to get published
20:54:45FromGitter<rayman22201> @Yardanico, Nim wouldn't be so portable if you had to implement an index to a table in the pc or segment register. Which is the point @Araq is making.
20:55:09Yardanicowell, I'm talking about the same when you mention asm :)
20:55:41FromGitter<rayman22201> That was sarcasm. It was a bad joke.
20:55:46Yardanicook, my bad
20:57:30*data-man joined #nim
20:59:51Araqshashlick: thanks, fixed
21:03:38FromGitter<rayman22201> As soon as @Araq explained the need to keep an index I realized it would be infeasible. (Unless you have the resources / years of work of a big project like LLVM or GCC). It's also not really in the spirit of Nim. Makes me a sad panda because I like the C target, and I would like Nim to continue to be very fast, but I understand the issue. The nested IF statement pass sounds interesting. It's a very clever idea
21:03:38FromGitter... actually. I'm curious to see how it will compare performance wise. It will definitely be better than setjmp/longjmp. Time will tell!
21:10:57shashlickstruggling to understand this thread: https://forum.nim-lang.org/t/3961, bet Nim supports all they are asking for but not sure
21:12:05dom96Ugh, so somebody removing their repo has caused all the package repo PRs to start failing :(
21:12:22dom96I'll need to figure out a way to only test the changes
21:12:30dom96Anyone know if travis supports that?
21:14:43Yardanicoshashlick, well, in my opinion it's not really worth to answer all these questions if the person didn't try to read the manual and code something in the language
21:14:53FromGitter<zetashift> @shashlick he wants some mixture of C# and Python and JS. I don't think he needs Nim/ expected C# with Python syntax
21:15:01Araqdom96: you don't want to hear it but ...
21:15:28Yardanicohow can he compare C# and JS with Nim? C# is run via a VM (.NET framework and .net core), JS is run via a VM too
21:15:34Yardanicoand nim is compiled to machine code
21:16:15AraqI think turning packages.json into a directory with one file per nimble package entry would be valuable in itself and also make this 'diff' thing easier to do
21:16:26FromGitter<zetashift> I think he's comparing the design choices(e.g. OOP support/single inheritance) vs nim's procedural approach
21:16:39FromGitter<zetashift> not the implementation stuff
21:16:57YardanicoAraq, oh, yeah, packages.json will need to be split into different files anyway at some point
21:17:36AraqYardanico, shashlick just show him Nim's forum, frontend and backend in the same language, leveraging DSLs everywhere (Karax, Jester). that's impossible with C#
21:18:24Araqin fact, eventually it could also use Ormin for the zero overhead superb SQL query creation
21:18:29FromGitter<zetashift> I think he'll just jump from question to question
21:18:35Araqthat had to be built into C#
21:19:16dom96I'm happy to delete that thread btw guys
21:19:16Araqand not into Nim, because Nim focuses on the principles and C# on programming hypes.
21:19:45dom96Seems like a trolling session starting
21:20:34FromGitter<Yardanico> https://github.com/pragmagic/godot-nim/issues/17#issuecomment-399244808
21:20:49FromGitter<zetashift> just show him the nim-way of his unformatted code and let it be an example for future seekers?
21:21:06FromGitter<zetashift> python + c# = f# thoooo
21:23:55dom96Ugh, seriously. Why delete a perfectly good repo?
21:24:03dom96This is why we need a package directory
21:36:29*nsf quit (Quit: WeeChat 2.1)
21:40:59Araqdom96: true but you could also register a github nick that forks every registered Nimble package :-)
21:41:06Araqautomatically.
21:41:19dom96hrm, not a bad idea
21:41:39dom96it would be a pain to keep these forks updated though
21:45:51FromGitter<kayabaNerve> Has nimble.io been registered?
21:46:15dom96yes
21:46:19FromGitter<kayabaNerve> Great
21:46:42dom96Not by us :P
21:47:00FromGitter<kayabaNerve> God damn it
21:47:10FromGitter<kayabaNerve> Has it been registered? Yes. By us? No
21:47:20FromGitter<kayabaNerve> We were so close
21:47:30FromGitter<kayabaNerve> We could just mirror what NPM does
21:47:44FromGitter<kayabaNerve> I mean, I don't like all of NPM but is the biggest code package manager ever built
21:47:56FromGitter<kayabaNerve> Unless you want to consider apt-get *-dev. No idea then...
21:56:55*elrood quit (Quit: Leaving)
22:10:41FromDiscord<treeform> Araq, not sure if this is bike shedding or real nimpretty bug: https://gist.github.com/treeform/c0a3a0f5606d8e3ce5a06584017cfb0e
22:11:13FromDiscord<treeform> The thing does not realize that the list is already wrapped and it wraps it its own way which looks odd.
22:12:22Araqline wrapping is a tough problem and the current implementation is
22:12:36Araq"I'm gonna wrap it so you aware it needs wrapping"
22:12:53Araqin other words, it's mostly a hint that you should reformat the code
22:12:54FromDiscord<treeform> ok, yeah other line wrapping issue:
22:12:55FromDiscord<treeform> https://gist.github.com/treeform/a6f2cac5a145c0df9afb298333d2655d
22:13:08FromDiscord<treeform> I see.
22:14:45FromDiscord<treeform> Good news though it does not produce compiler errors now, and my tests pass. So I am 98% confident it got it right.
22:15:12Araqyour other gist looks fine to me
22:15:48Araqonce nimsuggest integration arrives I'll use it regularly
22:15:54Araqer
22:15:59FromDiscord<treeform> it should probably drop the whole peekToken( down to next line...
22:16:01AraqI mean VSCode integration.
22:16:52Araqtreeform: that's easy to change but in general we need a much more powerful algorithm for wrapping
22:17:01AraqI left a link in the sources about how to do it
22:17:12FromDiscord<treeform> yeah got it. I will not report line wrapping for now.
22:17:28FromDiscord<treeform> Its working, I will try to use it.
22:17:42Araqcool, thanks
22:18:02FromDiscord<treeform> you have a day were you run this over the whole standard library?
22:18:08FromDiscord<treeform> and there will be tons of changes?
22:19:41FromDiscord<treeform> like most of the standard library uses proc (,,,) not proc (;;;) for example.
22:25:58FromGitter<ephja> I've always been breaking before brackets if necessary, but I dunno if I mind breaking as late as possible (possibly in-between brackets) the way nimpretty does it
22:26:20*Bevoss quit (Ping timeout: 260 seconds)
22:29:09Araqtreeform: maybe. I need to work on more important things though.
22:30:27*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
22:30:47Araqalso the ; vs , implementation will probably be changed
22:31:37*yglukhov[i] joined #nim
22:31:43*yglukhov[i] quit (Remote host closed the connection)
22:32:19*yglukhov[i] joined #nim
22:32:51*zahary quit (Ping timeout: 240 seconds)
22:34:38*zahary1 joined #nim
22:35:08FromDiscord<treeform> how can some thing be more important then nimpretty
22:36:27*yglukhov[i] quit (Ping timeout: 240 seconds)
22:42:24*PMunch quit (Quit: leaving)
22:44:57*yglukhov[i] joined #nim
22:49:22*yglukhov[i] quit (Ping timeout: 245 seconds)
22:52:57*yglukhov[i] joined #nim
22:57:17*yglukhov[i] quit (Ping timeout: 245 seconds)
23:18:18*lainon joined #nim
23:20:47*tsoernes quit (Remote host closed the connection)
23:32:15tsaka__how can I add "--define:blas=openblas" to nim.cfg?
23:51:33FromGitter<kayabaNerve> Just like that :thinking:
23:52:04*zahary1 quit (Quit: Leaving.)