<< 23-01-2020 >>
00:07:30FromDiscord_<Tak> Or it it best to use a normal object with a [T] parameter?
00:08:34FromGitter<alehander92> this sounds like a generic return type
00:08:38FromGitter<alehander92> somethig like
00:09:06FromGitter<alehander92> proc get(a: YourType, T: type): T = ..
00:09:13FromGitter<alehander92> or more exactly
00:09:56FromGitter<alehander92> maybe this would works
00:09:58FromGitter<alehander92> work*
00:10:26*dawkot quit (Ping timeout: 240 seconds)
00:10:28FromGitter<alehander92> disruptek i tried to think of a example maybe macro pragma would also be possible? not sure
00:13:50FromDiscord_<Tak> I will post a playground but my internet just died so bear with
00:24:30FromDiscord_<Tak> https://play.nim-lang.org/#ix=286m
00:24:59FromDiscord_<Tak> Apologies if bad formatting, I wrote on phone
00:37:17leorizeAraq: can you review this? https://github.com/nim-lang/Nim/pull/13201
00:37:17disbot➥ Make file descriptors from stdlib non-inheritable by default
00:37:30leorizethe test failure is due to timeout and I'm too lazy to restart the ci
00:37:33leorizebut it should pass
00:40:52*icebattle quit (Ping timeout: 268 seconds)
01:10:02*lritter quit (Ping timeout: 240 seconds)
01:11:02*lritter joined #nim
01:20:40*jholland__ quit (Quit: Connection closed for inactivity)
01:44:17disruptekalehander92: i tried a macro pragma; it has two problems: 1) a pragma with no arguments doesn't get picked up by `macros.pragma()`. 2) the macro pragma does not compile when applied to a proc because `Error: expression has no type`.
01:44:35madpropstrying to find out how to find connected usb (pen) drives (linux)
01:44:54madpropsthe idea is to make a tool to put iso images to usb drives, like rufus/unetbootin
01:45:12madpropsso the first step is detecting available device targets
01:49:23FromDiscord_<demotomohiro> !repo libusb
01:49:24disbothttps://github.com/nimious/libusb -- 9libusb: 11Nim bindings for libusb, the cross-platform user library to access USB devices. 15 6⭐ 4🍴
01:50:14madpropscool
01:51:27FromDiscord_<demotomohiro> I don'T
01:52:19FromDiscord_<demotomohiro> i dont know whether it really can do what you want to do
01:53:26disruptekshould work fine.
01:54:52shashlickSo px2 now has the same command line syntax as curl - just pulled in that code
01:57:12shashlickAlso supports curlrc
02:03:00disruptekkinda cool, if you need such an animal.
02:04:20disruptekso you can just drop-in px2 instead of curl? i wonder if you could just use it to replace some symbols from libcurl via a PRELOAD; then you could just have curl work, magically.
02:07:09*arecaceae quit (Remote host closed the connection)
02:07:28*arecaceae joined #nim
02:24:07FromGitter<deech> Is https://github.com/alehander92/gara the best pattern matching library currently available?
02:30:04shashlickWell the idea is to use curl as a proxy server so you might as well have the same configurability
02:30:43*cgfuh quit (Quit: WeeChat 2.6)
02:33:51shashlickYou use curl as a url downloader, px2 as a proxy to go thru auth proxies that most command line tools cannot talk to
02:35:46shashlickA lot of my work with px was getting thru different proxy auth mechanisms and implementations and I didn't have access to most scenarios
02:36:05FromGitter<iffy> leorize: I'll admit that using Dbus is kind of a pain, but trying to get Nim to use libsecret (because it uses glibc, I think) was too hard for me. Someone with glibc experience could probably get it going easily
02:36:13shashlickWith curl, someone has already done the work and is maintaining it
02:45:08*endragor joined #nim
03:03:21leorize@iffy: uhmm nim uses the libc, glibc has nothing to do here :P
03:04:13leorizeyou probably mean gtk
03:19:57*lritter quit (Quit: Leaving)
03:51:00*muffindrake quit (Ping timeout: 248 seconds)
03:53:04*muffindrake joined #nim
04:20:01*Guest94638 quit (Ping timeout: 265 seconds)
04:25:50*dddddd quit (Remote host closed the connection)
04:36:26*dadada joined #nim
04:36:48*dadada is now known as Guest67266
04:39:07*nsf joined #nim
04:48:29*ljoonal quit (Read error: Connection reset by peer)
04:48:44*ljoonal joined #nim
04:52:54*ljoonal quit (Client Quit)
04:53:12*ljoonal joined #nim
05:15:00*marmotini_ joined #nim
05:23:09*ltriant quit (Quit: leaving)
05:42:17*gmpreussner quit (Ping timeout: 258 seconds)
05:43:26*gmpreussner joined #nim
05:45:55*narimiran joined #nim
05:51:22TangerWhoa! linenoise is part of the stdlib?
06:06:43Yardanicowell yeah, it's a very simple binding to the C lib
06:07:01Yardanicohttps://github.com/nim-lang/Nim/blob/devel/lib/wrappers/linenoise/linenoise.nim
06:07:29TangerBeautiful
06:07:45TangerAlmost makes it hard to believe the lib is 20k+ lines
06:07:54Yardanicoit's not :)
06:08:15Yardanico"linenoise.c -- guerrilla line editing library against the idea that a line editing lib needs to be 20,000 lines of C code."
06:08:25Yardanicoit's around 1.1kloc
06:08:33TangerOh! Haha, I can't believe I completely missed that word XD
06:22:42*marmotini_ quit (Remote host closed the connection)
06:23:16*marmotini_ joined #nim
06:27:54*marmotini_ quit (Ping timeout: 265 seconds)
06:39:26*marmotini_ joined #nim
06:52:01*solitudesf joined #nim
06:54:55*Pqzcih5 quit (Remote host closed the connection)
07:11:27*zyklon quit (Ping timeout: 260 seconds)
07:23:06*ptdel quit (Remote host closed the connection)
07:25:05*marmotini_ quit (Remote host closed the connection)
07:29:54*mjsir911 quit (Ping timeout: 246 seconds)
07:30:34*mjsir911 joined #nim
07:58:28*salotz[m] quit (*.net *.split)
07:58:28*aurielmp quit (*.net *.split)
07:58:29*nergal[m]1 quit (*.net *.split)
07:58:30*Demos[m] quit (*.net *.split)
07:58:38*vqrs quit (*.net *.split)
07:58:40*hohlerde quit (*.net *.split)
07:58:40*crem quit (*.net *.split)
07:58:43*GitterIntegratio quit (*.net *.split)
07:58:43*tobbez quit (*.net *.split)
07:58:44*lkw quit (*.net *.split)
07:58:52*stefantalpalaru quit (*.net *.split)
07:58:53*JStoker quit (*.net *.split)
07:58:53*Cadey quit (*.net *.split)
07:58:59*muffindrake quit (*.net *.split)
07:59:04*skelett quit (*.net *.split)
07:59:06*d-nice2[m] quit (*.net *.split)
07:59:07*oprypin quit (*.net *.split)
07:59:08*ronny quit (*.net *.split)
07:59:08*skrylar[m] quit (*.net *.split)
07:59:10*BitPuffin quit (*.net *.split)
07:59:18*d10n quit (*.net *.split)
07:59:19*Ckat quit (*.net *.split)
07:59:20*d10n_ joined #nim
07:59:21*PMunch joined #nim
07:59:21*theelous3 quit (Ping timeout: 265 seconds)
08:01:08*planetis[m] quit (Ping timeout: 252 seconds)
08:01:09*silvernode[m] quit (Ping timeout: 245 seconds)
08:01:09*pigmej quit (Ping timeout: 245 seconds)
08:01:09*leorize[m] quit (Ping timeout: 245 seconds)
08:01:14*Jjp137 quit (Ping timeout: 240 seconds)
08:01:17*watzon[m] quit (Ping timeout: 246 seconds)
08:01:17*unclechu quit (Ping timeout: 246 seconds)
08:01:38*Yardanico[m] quit (Ping timeout: 260 seconds)
08:01:47*gmpreussner quit (Ping timeout: 268 seconds)
08:01:54*oswin[m] quit (Ping timeout: 250 seconds)
08:01:59*aeverr[m] quit (Ping timeout: 245 seconds)
08:01:59*lqdev[m] quit (Ping timeout: 245 seconds)
08:02:12*vycb[m] quit (Ping timeout: 260 seconds)
08:02:14*k0mpjut0r quit (Ping timeout: 252 seconds)
08:03:02*muffindrake joined #nim
08:03:02*skelett joined #nim
08:03:02*oprypin joined #nim
08:03:02*Ckat joined #nim
08:05:13*gmpreussner joined #nim
08:05:15*gsingh93 quit (Ping timeout: 240 seconds)
08:07:50*gsingh93 joined #nim
08:28:08*kungtotte quit (Ping timeout: 260 seconds)
08:35:35*floppydh joined #nim
08:45:51*vqrs joined #nim
08:45:51*crem joined #nim
08:45:51*tobbez joined #nim
08:45:51*lkw joined #nim
08:45:51*stefantalpalaru joined #nim
08:45:51*JStoker joined #nim
08:45:51*Cadey joined #nim
08:48:39*dom96 quit (Ping timeout: 268 seconds)
08:51:24*dom96 joined #nim
08:51:24*dom96 quit (Changing host)
08:51:24*dom96 joined #nim
09:05:13FromGitter<zacharycarter> @mratsim - if you're around at some point, I'd like to poke you about some questions I have regarding weave
09:12:05FromGitter<alehander92> @deech
09:12:17FromGitter<alehander92> well i tried to take what i can from the todo of patty
09:12:34FromGitter<alehander92> and try to support it (not all )
09:12:59FromGitter<alehander92> but i didnt add the variantp/variant macro ( https://github.com/alehander92/gara/issues/5 )
09:12:59disbot➥ Macro to generate variant objects
09:13:47FromGitter<alehander92> and the other remaining thing is some kind of compile time exhaustiveness checking which didnt seem very hard for many cases
09:14:02FromGitter<alehander92> but i just havent done it yet
09:14:31*krux02 joined #nim
09:14:46FromGitter<alehander92> https://github.com/krux02/ast-pattern-matching is the best one for macro stuff (NimNode-s)
09:14:58FromGitter<alehander92> but it is targetted only for that iirc
09:22:15FromGitter<alehander92> so did you solve it disruptek
09:24:32*Vladar joined #nim
09:37:41ZevvOooh I got a nice idea. Matching Nim AST with npeg. Just because it can.
09:37:59ZevvThat is sooooo meta
09:40:57Yardanicowoah, finally made a somewhat working ADB protocol implementation (client side, only remote TCP connection for now) in Nim
09:41:21YardanicoADB is poorly documented really, had to edit a implementation in nodejs to get logs to figure out stuff :D
09:44:57Yardanicoalso, is it pretty light, simplest async shell implementation (well, it also uses threadpool to take user input) uses less than 10mb ram
09:49:36Zevvurl?!
09:50:14Yardanicoi didn't upload it to github yet :P
09:50:28YardanicoI'll in a few hours maybe, need to make it a bit better so you can actually use it from other files
09:51:50Zevvsweet
09:55:47PMunchUgh, I'm starting to miss metaprogramming again :(
09:56:23Zevvprogramming BASIC again, eh?
09:56:33Zevvjust stop that dude
09:56:39Zevvthat's baaad mkkaaayy
09:57:16PMunchProgramming C#..
09:57:24PMunchAnd it's not by choice
09:57:34Zevvit's all by choice
09:57:43Zevvyour life is yours
09:58:21PMunchHaha, fair point. But doing this one project in C# so I can go back to program things in whatever I like seems worth it
09:59:09Zevvthere you go :)
09:59:33PMunchI just meant that I didn't choose to do this project in C#
09:59:41Yardanicothis is how the usage of ADB lib looks like: https://play.nim-lang.org/#ix=287r
09:59:54Yardanicothe spawn stdin trick is from https://github.com/dom96/nim-in-action-code/blob/master/Chapter3/ChatApp/src/client.nim#L38
10:01:18YardanicoI will probably split it into the low-level implementation (so you can handle ADB messages on your own), and this "high-level" one
10:01:19Zevvnice. I now pipe stuff but having native would be cool
10:01:49Yardanicowell, I wanted to implement AdbStream based on streams from stdlib, but failed really :(
10:02:09*Jjp137 joined #nim
10:02:12Yardanicosince I use recv and stuff, and there are effects like TimeEffect which are incompatible with StreamObj proc definitions
10:07:37FromDiscord_<mratsim> @zacharycarter ask away
10:11:51FromGitter<zacharycarter> mratsim: thanks! so I am trying to use the parallel task functionality (which I think is the simplest to wrap my head around) - basically what I'm doing is spawning a bunch of calls to a proc that performs astar and storing the returned FlowVars in a sequence, then I sync them
10:13:00FromGitter<zacharycarter> the astar method is using generics in its signature for the parameters - it's signature looks like this: `proc path*[G: Graph, N: Node, D: Distance](graph: G, start, goal: N): seq[N]`
10:13:20FromGitter<zacharycarter> so the first error I run into is: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5e2971c0e177666195b401a2]
10:13:35FromGitter<zacharycarter> I believe because of the return type of `seq[N]`
10:14:11FromGitter<zacharycarter> but if I replace `seq[N]` with `seq[tuple[x, y: int]]` I believe I get past that error - the code is at home but I'm pretty sure that's what I did
10:14:24FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5e2972009ad22d5bd5c62318]
10:15:14FromGitter<zacharycarter> I guess that's because my `Spaceship` object has some member that is a sequence?
10:15:35FromDiscord_<mratsim> looking
10:16:06FromGitter<zacharycarter> thanks
10:16:30FromDiscord_<mratsim> for the second error, in the Weave devel, i removed the GC constraint
10:16:40FromGitter<zacharycarter> oh okay awesome
10:16:45FromDiscord_<mratsim> I don't guarantee that it will work,
10:17:03FromGitter<zacharycarter> well - I will certainly give it a shot and let you know what I find out
10:17:05FromDiscord_<mratsim> but supportsCopyMem(T) was lumping together both ref types and types with custom destructors
10:17:17FromDiscord_<mratsim> custom destructors types should work
10:18:35FromDiscord_<highDevGuy> Is there a way to delay execution of some code without blocking the thread?
10:18:35FromDiscord_<highDevGuy>
10:18:43FromDiscord_<highDevGuy> I'm using nigui for the GUI and I have to execute a certain proc every second. If I make the call to that proc recursive with a bit of sleep(1000). it freezes everything. I explored threads and spawn but it seems like I have to rewrite almost everything because I have global variables used both by the GUI on the main thread and the background thread for the script. Basically I was wondering if there is a way to delay execution of certai
10:18:43FromDiscord_<highDevGuy>
10:18:46FromDiscord_<highDevGuy> I hope that make sense and thank you for your time 🙂 Also, this is basically my 3rd day with nim so please go easy on me if I'm asking dumb questions 😄
10:18:49FromDiscord_<mratsim> for the first error, it's strange that the line points to your code, I expect it's related to my assertiosn code: https://github.com/mratsim/weave/blob/master/weave/instrumentation/contracts.nim#L36-L39
10:19:24FromDiscord_<mratsim> if you compile with assertions:off and you don't have compile-time error that would be prove it
10:21:10FromGitter<zacharycarter> okay - I'll try both of those suggestions when I get home tonight - thanks Mamy!
10:21:15Yardanico@highDevGuy well, you have to use threads anyway, you might have some success with other GCs (you can try boehm which has shared thread heap, or --gc:arc if you're on devel) or rewriting your code a bit and using locks and stuff
10:21:33FromDiscord_<mratsim> for storing a seq of FLowvar then syncing them, you can check my parallel nqueens code: https://github.com/mratsim/weave/blob/master/benchmarks/nqueens/weave_nqueens.nim#L119-L132
10:21:42FromGitter<zacharycarter> oh awesomeeeee
10:22:05FromDiscord_<mratsim> if the array is temporary and does not escape its scope, alloca will bring tremendous performance benefits by avoiding memory fragmentation
10:22:13FromGitter<zacharycarter> perfect
10:22:36FromDiscord_<mratsim> allocation and fragmentation*
10:22:42FromDiscord_<mratsim> but allocation is the main issue
10:23:12Yardanico!repo nim-adb
10:23:13disbothttps://github.com/Yardanico/adbscan -- 9adbscan: 11A @nim-lang utility for scanning IPs for unprotected ADB Android devices connected to the internet 15 5⭐ 0🍴 7& 1 more...
10:23:16FromGitter<zacharycarter> I can use ALL the performance benefits :D
10:23:22YardanicoZevv: https://github.com/Yardanico/nim-adb
10:23:27Yardanicoit's pretty barebones but it kinda works :P
10:23:40Yardanicoand ofc only TCP (remote connection to ADB) for now, no fancy USB stuff
10:24:22FromDiscord_<mratsim> @highDevGuy, you can use asyncDispatch, create an async procedure with an asyncSleep timer
10:24:44Yardanico@mratsim , well, he uses NiGui, so he'll have to either poll inside the GUI loop for like 50ms, or use threads anyway
10:24:52FromDiscord_<mratsim> you await in that asyncProc that the sleep is elapsed and then trigger whetever you need to do
10:25:05FromDiscord_<mratsim> ah, well never used NiGui :p
10:25:20ZevvYardanico: thanks, will look into that!
10:30:49FromDiscord_<highDevGuy> thank you... going to try asyncDispatch if I figure it out as it seems to be something that would require a lot of code rewriting 😄
10:31:02FromDiscord_<highDevGuy> thank you... going to try asyncDispatch if I figure it out as it seems to be something that wouldn't require a lot of code rewriting 😄
10:31:09Yardanicowell, asyncdispatch wouldn't require you to rewrite a lot of code actually
10:31:27Yardanicoyou can define some async proc to run something over a timer, and poll asyncdispatch global dispatcher in your GUI loop
10:31:35Yardanicofor like 5-10ms each time your GUI updates
10:32:23Yardanico@highDevGuy, what about https://github.com/trustable-code/NiGui/blob/master/examples/example_08_timers.nim ?
10:32:29Yardanicoit seems to have timers just like you want
10:33:47FromDiscord_<highDevGuy> omg, how did I miss that
10:33:51FromDiscord_<highDevGuy> thanks ❤️
10:39:29FromDiscord_<highDevGuy> works perfectly. Thank you guys
10:41:36*Hideki_ joined #nim
10:46:26*Hideki_ quit (Remote host closed the connection)
10:53:41*Hideki_ joined #nim
10:55:35*ng0 joined #nim
10:57:42*abm joined #nim
10:57:52*Hideki_ quit (Remote host closed the connection)
11:05:36FromGitter<alehander92> zevv do it!
11:05:50FromGitter<alehander92> i still plan to use npeg
11:05:57FromGitter<alehander92> one day in this bigger toolkit!
11:17:30*dawkot joined #nim
11:24:56*dddddd joined #nim
11:26:14*cgfuh joined #nim
11:28:24FromDiscord_<Rika> just noticed std/asyncstreams, shame that it's still a tiny module
11:39:57*nergal[m] joined #nim
11:48:49*Guest67266 is now known as dadada
11:49:42dadadahttps://github.com/nim-lang/Nim/blob/devel/compiler/semexprs.nim#L2363
11:49:53dadadapair.add(n[j])
11:49:57dadadapair.add(x[1])
11:51:23dadadawould be great if stuff like that was chainable always
11:51:37dadadalike this: pair.add(n[j]).add(x[1])
11:51:59Yardanicodadada: https://github.com/nim-lang/Nim/pull/13092
11:52:00disbot➥ added operateOn to sugar.nim to give Nim the chaining mechanism it de…
11:53:09Araqthis is old code, result.add newTree(nkTupleConstr, n[j], x[1])
11:53:27Araqwould be better
11:59:46dadadaAraq: is "on" reserved for something else? if not, why not shorten operateOn to just "on"... that feature is seriously useful, and since almost anything inside a program is a operation one way or the other, "operate" seems does help little to understand the code
12:01:10Araqchaining itself isn't all that common in Nim, the "builder pattern" is a workardound for missing named arguments in the core language, but Nim has named args
12:03:55Araqbut more importantly, I doubt I got it right
12:04:02Araqinstead of
12:04:15Araqon x: add("a"); add("b")
12:04:18dadadaI'm thankful you're working on this feature
12:04:24Araqpeople seem to want
12:04:39Araqx.on.add("a").add("b")
12:04:56dadadaI have to disagree with those on the github issue that say chainOn would be nice ... it's too long as well
12:05:15Araqfor dot chaining 'on' doesn't read well does it
12:05:17dadadaif it got to contain "chain" it might as well drop the "On" part
12:05:42Araqlet's first decide on dots vs blocks
12:05:51dadadaI thought the block idea was good
12:06:11dadadaon x: add(a).add(b)
12:06:22Araqhuh?
12:06:26Araqmixing both?
12:07:37dadada1) we tend to be used to chaining from functional langs or JavaScript, and there you get something like doA().thenB().thenC()
12:08:14dadadaof course being used to something doesn't have to mean it's the best paradigm
12:08:34Zevvoi Araq. Does the alloc cleanup PR need more work or is it good to go afayc?
12:08:47*Demos[m] joined #nim
12:08:47*GitterIntegratio joined #nim
12:08:47*unclechu joined #nim
12:08:47*Yardanico[m] joined #nim
12:08:48*k0mpjut0r joined #nim
12:08:48*lqdev[m] joined #nim
12:08:48*BitPuffin joined #nim
12:08:48*salotz[m] joined #nim
12:08:48*ronny joined #nim
12:08:49*leorize[m] joined #nim
12:08:51AraqNim's unofficial motto is "Copying bad design is not good design"
12:08:53*hohlerde joined #nim
12:08:53*silvernode[m] joined #nim
12:08:54*aurielmp joined #nim
12:08:54*vycb[m] joined #nim
12:08:54*aeverr[m] joined #nim
12:08:54*planetis[m] joined #nim
12:08:54*grantmwilliams joined #nim
12:08:55*oswin[m] joined #nim
12:08:55*skrylar[m] joined #nim
12:08:55*watzon[m] joined #nim
12:08:55*pigmej joined #nim
12:08:55*d-nice2[m] joined #nim
12:09:02Yardanicooh, matrix bridge is up agan :D
12:09:04Yardanicoagain*
12:09:21AraqZevv, what did you end up with regarding leak detection?
12:09:39dadadaAraq: could we use a semicolon? on x: add(a); add(b)
12:09:51Yardanicodadada: that's already possible with the PR I think
12:10:01Yardanicosince semicolon exists in Nim :)
12:10:05Zevvit's almost all clean now, I had some double bookkeeping. Two tests have a few les deallocs because the =destroys are done at the end of NimMainModule, so I can't report those properly
12:10:25ZevvThat either needs hooks or stuff from C, which I both don't like. afaik this is good as it is now, as the tests are the only cases using this for now.
12:10:50Araqhmm but you could use the subtraction method to ensure nothing inside main() did leak
12:11:00dadadaon pair: add(x[i]); add(y[i]) # is this readable?
12:11:16ZevvNot for everything, some stuff ends up in NimMainModule still
12:11:34dadadahow about using @? is that used yet by anything in Nim?
12:11:36Araqmost tests use a main proc for this reason
12:11:46dadada@pair: add(x[i]); add(y[i])
12:11:48Zevvaraq: not enough. Let me check the details
12:12:46dadadaI like @ because there won't be a conflict with anything else that might now be using "on"
12:13:13dadadait's also short syntax, downside might be that the code would very mildly start to look like perl
12:14:01Araq@ is used for @[1, 2, 3]
12:14:01dadadacould also support @pair.add(x[i]).add(y[i])
12:14:02*endragor quit (Remote host closed the connection)
12:14:50ZevvAraq: exceptions seem to cause this for example: http://ix.io/287J
12:15:31Araqyeah well
12:15:51Zevvyeah well indeed
12:17:45dadadaAraq: is ° free to use?
12:18:38dadadadoesn't come up anywhere in manual.html
12:18:51dadada°pair: add(a); add(b)
12:19:33narimirandadada: honest question: are you a troll?
12:20:08Yardanicodadada: I don't think using characters which can't be typed using a normal keyboard is okay
12:20:24Yardanicothere's https://www.emojicode.org/ for example :P
12:20:28narimiraneven & is "hard to reach", according to some authors :D :D
12:21:18dadadanarimiran: no, just throwing some ideas out there, Yardanico: if that's true then it's my bad, ° is easy to type on my keyboard, as is & and @
12:21:27Yardanicowell, on most keyboards it's not :)
12:22:06Yardanicomost keyboards have this layout https://en.wikipedia.org/wiki/QWERTY
12:22:19Yardanicothe US or UK one
12:22:51Yardanicoeven in Russia we have keyboards which have letters for russian layout and US qwerty layout
12:23:13FromDiscord_<Rika> APL flashbacks
12:23:19narimirandadada: if i'm not confusing you with someone else (sorry if i am), all your ideas seem to be "oh, i can A in nim. nah, B feels better to me, lets change the whole language to accomodate for B" or "oh, somebody proposed C. nah, i like D more because of reasons"
12:23:32*endragor joined #nim
12:24:09PMunchnarimiran, dadada seems fine, I think you're confusing him with someone else :P
12:24:27narimiranthen i'm really sorry
12:24:41narimiranbut recently there were some discussions like i described
12:25:00narimirani guess i forgot by whom
12:25:43PMunchHmm, I can't seem to type ° by any combination..
12:26:01Araqdadada, it's universally agreed upon that only Leibniz should have been allowed to invent new notation or operators
12:26:50dadadanarimiran: it might have been me in the last couple of days (before is unlikely), however I never dissed Nim, I just voice my ideas and opinions, you can think whatever you want, but please don't place thought restrictions on me
12:27:10ZevvAraq: so it's fine for now? I'll go look for other things then.
12:27:57FromGitter<Vindaar> @narimiran: I think you didn't see my message from yesterday. I was wondering whether the only dll dependencies we allow for windows are those, which come with the `dlls.zip` from the nim download section. Is that correct?
12:28:29*endragor quit (Ping timeout: 272 seconds)
12:29:25PMunchWell, I can do Ctrl+Shift+u then B0<enter>, but that's cheating
12:29:59dadadaPMunch: stop, I dropped the idea now I know that my layout is more special than anticipated
12:30:14PMunchHaha :P What layout do you use by the way?
12:30:30dadadaGerman
12:33:28lqdev[m]PMunch: alt gr+shift+0 doesn't work for you?
12:34:02PMunchNope..
12:34:13PMunchBut then again I've customised my keyboard pretty hard..
12:35:07dadadanarimiran: maybe you an put me on an ignore list, block me or anything like that, I didn't make any demands here, just brainstorming and with me only recently learning the intricates of Nim my perspective on it obviously differs from those who have used it many years
12:35:31lqdev[m]add it into your X keymap ;)
12:35:32PMunchlqdev[m], next time I reboot it will :)
12:35:39PMunchJust did :P
12:36:00PMunchdadada, don't worry you're welcome to stay :)
12:36:33PMunchWe've just had some unfortunate trolls in the past, so people are a bit on edge against trolls
12:37:08PMunchWhere does the german layout put ° by the way?
12:37:22dadadaright above tab, shift+keyabovetab
12:37:41PMunchAh, that yields a § for me
12:37:44FromGitter<alehander92> Araq but isnt on x: op, op very similar to the other one
12:37:53FromGitter<alehander92> like, if we ignore the op; op vs op.op
12:38:01FromGitter<alehander92> its only method call syntax ina ction
12:38:05PMunchSo you have pipe(|) and then with shift you get degree symbol?
12:38:10FromGitter<alehander92> on x: op vs x.on(op)
12:38:13YardanicoI have ` or ~ (with shift) above the tab :)
12:38:24FromGitter<alehander92> btw .. yeah why not chain with pipes
12:38:29FromGitter<alehander92> like elixir
12:38:41dadadaPMunch: no there is ^ and ° (with shift)
12:38:51PMunchWait, where do you have pipe then?
12:39:19dadadaright side alt + key right to left side shift
12:39:20FromGitter<alehander92> ok its not the same semanthic at all, but at least the syntax is an alternative
12:39:30FromGitter<alehander92> but yeah somethingOn is probably cleaner
12:39:51PMunchOuch, so pipe is a bit cumbersome for you then?
12:40:07dadadano | is perfectly awesome
12:40:19PMunchOh, I can actually get | by doing Altgr + that key
12:40:30dadadaleft hand pinky and right hand thumb
12:40:31Yardanicofor me it's just shift + thatkeyafter ]
12:40:33PMunchThat key is normally Ø for me :P
12:40:44PMunchYardanico, but where is ]?
12:40:52dadadamuscle memory makes | super easy
12:40:52PMunchFor me ] is AltGr + 9
12:41:08YardanicoPMunch: like on this layout https://en.wikipedia.org/wiki/QWERTY#/media/File:KB_United_States.svg
12:41:16Yardanicoi have the same layout as this one
12:41:22Yardanicowell, almost
12:41:39YardanicoI have "FN" key instead of windows key on the right side
12:43:04PMunchAh, right
12:43:40PMunchMine is like this: https://upload.wikimedia.org/wikipedia/commons/e/ec/Dvoraktastatur.png
12:43:42PMunchMore or less
12:44:03Yardanicooh, dvorak?
12:44:07Yardaniconeat
12:44:09PMunchI have some custom bindings to stuff, but for the most part that's the one I use
12:44:13PMunchYeah, Norwegian dvorak
12:45:03narimiran@vindaar yeah, sorry, i missed that message. yeah, it seems like that - it is similar to what is in appveyor.yml
12:45:33narimirancan we test only a part of ggplotnim, which doesn't need extra dll?
12:46:11AraqZevv, link to your PR please
12:52:34Araqnever mind, found it
12:57:07FromGitter<Vindaar> @narimiran: yeah, then there's no way around. Give me another day and I'll change the test files such that every test that actually draws stuff is in its own file and I add a `-d:noCairo` compile option. Then at least the plot construction can be tested
13:08:34*marmotini_ joined #nim
13:08:56*kungtotte joined #nim
13:09:33*narimiran thumb up
13:10:50*nsf quit (Quit: WeeChat 2.7)
13:11:23FromGitter<iffy> leorize: yes, GTK sounds familiar (GObjects and such?). Anyway, if someone submits a PR to use libsecret, I wouldn't turn it down
13:15:23*dawkot quit (Ping timeout: 258 seconds)
13:16:51*dawkot joined #nim
13:16:58FromDiscord_<Clyybber> Araq: Maybe we should really introduce some operator for outplace?
13:19:11*marmotini_ quit (Remote host closed the connection)
13:20:20*Pqzcih5 joined #nim
13:20:41Araqmeh, nobody likes custom operators, they are "confusing"
13:21:04AraqI've yet to find a single thing in programming that is not "confusing" for somebody but whatever
13:21:12FromGitter<alehander92> its not about confusing
13:21:15FromGitter<alehander92> its about balance
13:21:22FromGitter<alehander92> its cool to have custom operators
13:21:26FromGitter<alehander92> but if one has 10 custom operators
13:21:32FromGitter<alehander92> it becomes a whole new notation
13:22:03AraqZevv, this *Impl stuff is excessive, is it really necessary?
13:22:58*marmotini_ joined #nim
13:25:20FromDiscord_<Clyybber> I agree operators should be introduced very sparingly
13:25:51FromDiscord_<Clyybber> But with chainOn and outplace I feel like nobodys gonna use them if they are "big words"
13:26:19AraqIMO outplace is fine and chainOn needs a better name
13:26:29Araq'on' would work, I tested it
13:26:37Araqbut it's too short for my taste
13:26:47FromDiscord_<Clyybber> Maybe overload outplace?
13:28:01*marmotini_ quit (Remote host closed the connection)
13:28:09ZevvAraq: Well, it is if you want to catch all allocs through one place and dispatch to the proper implementation
13:28:40ZevvThere is only *one* `alloc(size: int)` proc.
13:29:34ZevvOtherwise the accounting has to be implemented in 4 places
13:29:41Araqyeah it's pretty nice
13:29:51Araqtoo bad it doesn't use the old -d:nimTypeNames mechanism
13:30:00Zevvdon't know about that?
13:30:00Araqbut it would be hard to do that
13:30:17Araqread gc_common.nim, it's pretty neat
13:30:28Araqyou can get stats on the number of objects *per type*
13:31:12Zevvoh right, that would be sweeet. But I guess that doesn'nt have to go in with this PR, right?
13:31:21ZevvI'd like to get this out of the way before adding more stuf and divering further
13:31:23FromGitter<alehander92> so maybe offtopic, but now what is needed to do for a bare os : to redefine `alloc` or to just link to a custom mini-libc?
13:31:43Zevvalehander92: compile with --os:any and implement malloc()/free()/realloc()
13:31:57Zevvand -d:useMalloc
13:32:08Zevvalehander92: https://github.com/zevv/nim-linux-kernel/blob/master/src/stubs/stubs.c
13:32:23Zevvthat is all I need to stub to run nim in the linux kernel
13:32:28narimiran`chainOn` is worse name than `operateOn`, because you can have only one argument (ok, why would you, but still), and there's no chaining there, but it is operating
13:32:48Zevvthanks araq
13:32:49FromGitter<alehander92> good note
13:32:57FromGitter<alehander92> great zevv awesome
13:35:49narimiran"great zevv awesome" +1
13:36:06*marmotini_ joined #nim
13:36:49ZevvI guess that stuff should be documented. The 'embedded' section in the compiler guide still talks about os:standalone, should that be replaced?
13:37:54FromDiscord_<Clyybber> Afaict operateOn/cascade/with all accomplish roughly the same thing
13:40:02*endragor joined #nim
13:41:18AraqZevv, yeah, rewrite the docs
13:42:01AraqI looked at 'cascade', it's well done. it does a little bit too much for my taste
13:42:47Araqbut our guidelines need to be expanded, we need a rule like
13:43:07Araq"Common syntactic sugar should be in the stdlib"
13:44:09*zyklon joined #nim
13:44:10*marmotini_ quit (Read error: Connection reset by peer)
13:51:05*marmotini_ joined #nim
13:58:19*marmotini_ quit (Read error: Connection reset by peer)
13:59:20FromGitter<alehander92> does this include pattern matching
14:01:42*marmotini_ joined #nim
14:02:14Araqit does.
14:02:49AraqPM is much more important than chaining
14:02:56Araqfor some reason I forgot about it
14:04:03Zevvhttps://github.com/zevv/Nim/blob/zevv-embedded-docs/doc/nimc.rst#nim-for-embedded-systems
14:04:57*marmotini_ quit (Remote host closed the connection)
14:06:30narimiranshashlick: are you around maybe?
14:17:19shashlickSup
14:18:15FromGitter<alehander92> ok good to know
14:18:39FromGitter<alehander92> zevv nice, i have to try it
14:20:40narimiranshashlick: you don't see private messages here, right? will telegram do?
14:31:35dadadato cat(enate) means the same as to chain sth, is well known from the POSIX cat command
14:31:49dadadacat x: add(a); add(b)
14:32:02dadadait's also pretty short
14:44:06*lritter joined #nim
14:45:31ZevvAraq: what is the purpose of your C++ exceptions branch? I'd like to target cpp on the avr atmel for PMunch's experiments, but at this time this does not work because of AVR limitations with exceptions - maybe your work will help here somehow
14:49:49*marmotini_ joined #nim
14:50:43FromGitter<alehander92> dadada i wouldnt advertsice posix naming
14:50:44FromGitter<alehander92> to Araq :P
14:51:13FromGitter<alehander92> otherwise i admit concatenating is a well known termin
14:51:40FromGitter<alehander92> btw have you seen concepts, i guess they would be interesting for you to comment
14:53:10dadadaalehander92: which concepts?
14:57:37*floppydh quit (Read error: Connection reset by peer)
14:58:26dadadaok, you probably mean those https://nim-lang.org/docs/manual_experimental.html#concepts
14:58:38*floppydh joined #nim
14:59:44*ptdel joined #nim
15:00:19dadadahmm, looks nice
15:02:13*PMunch quit (Quit: Leaving)
15:06:18narimiranwhat's with the obsession about short names? nobody here uses autocomplete?
15:08:16disrupteknot for my name.
15:11:55lqdev[m]disr
15:12:03lqdev[m]damn, my autocomplete didn't work
15:12:54lqdev[m]narimiran: imo the only abbreviations that one should make in code are common words
15:13:02lqdev[m]eg. gfx for graphics
15:13:09lqdev[m]when dealing with a graphics project
15:13:11lqdev[m]etc.
15:15:03FromGitter<alehander92> narimiran well shorter names are good for primitives imo
15:15:26FromGitter<alehander92> but i agree we might overdo it sometimes!
15:15:35*marmotini_ quit (Remote host closed the connection)
15:16:13FromGitter<alehander92> dadada yes, they're relatively experimental, i think most related to c++ concepts
15:16:29FromGitter<alehander92> you might also take a look at the nimz3 idea
15:16:34FromGitter<alehander92> if you're interested in such stuff
15:16:49FromGitter<alehander92> but thats much more type-stuff-y
15:17:51*floppydh quit (Quit: WeeChat 2.7)
15:17:55*ng0_ joined #nim
15:19:10*nsf joined #nim
15:21:30*ng0 quit (Ping timeout: 265 seconds)
15:41:08AraqZevv, for me C++ based exception handling is just broken
15:41:31Araqbut for you ... you should make --exceptions:goto work for C++
15:41:39Araqwe talked about the fix
15:44:46FromGitter<Varriount> narimiran: If you have that opinion, try to avoid Go
15:46:24YardanicoGo code looks extremely verbose imo, big tabbing and "simple" syntax, I'm used to 2-space indents in Nim :)
15:46:48FromDiscord_<Rika> go indents arent important anyway afaik?
15:46:48*dddddd quit (Ping timeout: 260 seconds)
15:46:56FromDiscord_<Rika> so you can always convert to 2-space
15:47:16FromGitter<Varriount> Yardanico: Part of that is due to a lack of language constructs and/or generics
15:48:21FromGitter<Varriount> Want to check if an element is in a slice (sequence)? Guess what, you get to write that `for` loop on your own
15:48:39Yardanicoreally?
15:48:45FromGitter<Varriount> Yes
15:48:50YardanicoLMAO https://golangcode.com/check-if-element-exists-in-slice/
15:48:57Yardanicohttps://stackoverflow.com/questions/15323767/does-go-have-if-x-in-construct-similar-to-python
15:49:13Yardaniconice DRY
15:49:25*icebattle joined #nim
15:49:27FromGitter<Varriount> There are technically ways to do it, but you have to use Go's equivalent of a void pointer (an empty interface)
15:49:39Yardanicoalso since there's no generics you're gonna write a new one for each type ?
15:49:46FromGitter<Varriount> Yep
15:50:02FromDiscord_<gingerBill> Not to defend Go but I rarely check for elements in a slice and when I do, I usually do need a for loop at any rate because I want to do other checks.
15:50:21FromDiscord_<gingerBill> So for me personally, it was rarely a problem.
15:50:22FromDiscord_<Rika> void pointer eugh
15:50:54FromGitter<Varriount> Gingerbill: Fair point
15:52:12FromGitter<Varriount> However it makes creating constructs (like, say, wrappers and decorators) more difficult than it should be
15:52:17FromDiscord_<gingerBill> So doing `found := false; for _, e := range slice { if e == to_find { found = true; break } }` is not a huge issue in practice.
15:53:31FromDiscord_<gingerBill> So doing `found := false; for _, e := range slice { if e == toFind { found = true; break } }` is not a huge issue in practice.
15:53:39FromDiscord_<gingerBill> Even if you want to do `toFind in slice`
15:54:29FromDiscord_<gingerBill> And if I want to do `in` on something, I usually want the index too not just a boolean check.
15:55:03FromDiscord_<gingerBill> `index := -1; for i, e := range slice { if e == toFind { index = i; break } }`
15:55:38FromDiscord_<gingerBill> So if Go had some form of metaprogramming (that is not go gen), this would be nice.
16:07:34Araqyeah you can simply always write much more code, it never gets tiresome.
16:07:50*endragor quit (Remote host closed the connection)
16:11:13FromDiscord_<gingerBill> It only gets tiresome if you have to write it a lot, but as I said, in my personal experience, I rarely wanted this particular construct.
16:11:35FromDiscord_<gingerBill> There other idioms I have wanted in Go but have lived without (to my minor annoyance)
16:13:17FromGitter<Varriount> gingerBill: You might like https://github.com/hexdigest/gowrap
16:14:35FromDiscord_<gingerBill> That's a cool metaprogramming tool!
16:29:16*Kaivo quit (Quit: WeeChat 2.7)
16:39:19*icebattle quit (Ping timeout: 258 seconds)
16:40:16*jholland__ joined #nim
16:52:08*samuell joined #nim
16:52:31*samuell quit (Read error: Connection reset by peer)
16:53:53*dawkot quit (Ping timeout: 260 seconds)
16:54:14*marmotini_ joined #nim
16:58:07*marmotini_ quit (Remote host closed the connection)
16:58:12FromGitter<alehander92> Araq why cant i access let
16:58:16FromGitter<alehander92> in func
16:58:36*NimBot joined #nim
17:01:45disrupteks/on/do-on/ s/do-on/doon/
17:04:08*dawkot joined #nim
17:04:14Yardanicowell, `let` means it still can change between different runs right?
17:04:21Yardanicoso func results for same arguments might be different
17:04:24Yardanicothat's how I see it at least
17:07:08leorizeAraq: I've looked into using OS-native functions for io
17:07:30leorizeit's not possible to implement io.nim as it is right now with os-native functions
17:08:06leorizein particular io.reopen(), since that requires access to File internals, which we can't do
17:09:42*icebattle joined #nim
17:17:40*ng0_ is now known as ng0
17:21:31*Trustable joined #nim
17:24:33*seni joined #nim
17:35:31cronanyone see zolk3ri here recently?
17:39:45*abm quit (Quit: Leaving)
17:41:42Araqleorize, I'm not following
17:42:54FromDiscord_<Recruit_main_70007> is there something i can use in nim similar to libbtc?
17:42:54FromDiscord_<Recruit_main_70007> https://github.com/jonasschnelli/libbtc
17:46:12leorizeAraq: so you told me you wanted to use native win32 api for io
17:46:45leorizeI've done some research and found that we would need our own File structure to do so
17:47:09leorizebecause we can't implement reopen() with C's FILE since we don't have access to its internals
17:47:11shashlickmratsim: going to try weave instead of threadpool for px2
17:47:34leorizebut that would break backward compatibility, so it can't be done atm
17:47:58Araqsure it would require our own File object, what's wrong with that?
17:48:14leorizebreaking existing C FFI code that rely on File
17:50:26leorizebut we will have a lot to gain with our own File object, for one it reduces our dependency on libc :) and it'll be possible to have deterministic behavior across different systems
17:52:51Araqyup
17:53:16AraqRecruit_main_70007: you could wrap it but I wouldn't use it, looks like bad code to me
17:53:57leorizeI will try to draft out an io.nim with custom File, and see how much would that break things
17:54:53shashlickmratsim: waah - processClient has arguments managed by GC (ref/seq/strings), they cannot be distributed across threads.
17:55:22leorize--gc:arc? :P
17:55:53FromDiscord_<Recruit_main_70007> araq, i found a nim wrapper for secp256k1, (which is basiacally what libbtc is based on)
17:55:57FromDiscord_<Recruit_main_70007> thanks
17:57:14*theelous3 joined #nim
17:58:49shashlickokay all my arguments are now pointers except the callback but still same error
18:01:33shashlickokay, made the callback an unsafeAddr and it compiles but crashes on runtime
18:02:21shashlickError: unhandled exception: /home/gt/.nimble/pkgs/weave-0.1.0/weave/memory/memory_pools.nim(625, 8 ) `sizeof(ChannelMpscUnboundedBatch[ptr MemBlock]) == 272` MPSC channel size was 512 [AssertionError]
18:02:39shashlickmratsim ^^^ when you get a chance
18:02:55*nsf quit (Quit: WeeChat 2.7)
18:04:56*Pqzcih5 quit (Remote host closed the connection)
18:05:40*Pqzcih5 joined #nim
18:27:37FromDiscord_<Recruit_main_70007> https://twitter.com/Phemex_official/status/1217808220703473664
18:27:37FromDiscord_<Recruit_main_70007>
18:27:37FromDiscord_<Recruit_main_70007> i leave these here, someone could get it
18:34:39*seni quit (Quit: Leaving)
18:37:00*marmotini_ joined #nim
18:37:08*jwm224 quit (Ping timeout: 268 seconds)
18:41:29*jwm224 joined #nim
18:52:18ZevvCan anyone test this for me with --gc:arc: http://ix.io/289A
18:52:58disrupteksegv
18:53:13Zevvright
18:53:14Zevvthanks
18:53:42Araqwait what?
18:53:55*edubart joined #nim
18:54:07disrupteks2 is unalloc
18:54:34Araqhmmm
18:54:53*edubart quit (Remote host closed the connection)
18:55:27Yardanicohad some fun with https://github.com/belamenso/v https://i.imgur.com/7Mxp1r1.png
18:57:39AraqZevv, is that a regression?
18:58:46Zevvbisecting
18:59:16Zevvbut I think no
19:02:04Zevvwell how far back should I go
19:02:58Yardanicofun fact: you can name your .nim file as "🤔.nim" and even import it successfully, but nimble doesn't like that package name :(
19:04:33ZevvAraq: it was broken in december already
19:04:59disruptekyardanico: what does nimph do with it?
19:05:07Yardanicodisruptek: 🤔
19:08:03*dawkot quit (Ping timeout: 260 seconds)
19:08:37*dawkot joined #nim
19:12:01AraqZevv, I meant a regression compared to Nim before your PR was merged, obviously
19:20:02ZevvOh you were blaming *me* :)
19:20:23Yardanicodamn this is too meta, you can bootstrap this lib in itself http://ix.io/289G
19:20:25ZevvDid you really think I would come out with this without checking for that
19:20:49Araqso ... 'koch boot -d:release --gc:arc' is a totally a thing now, right?
19:21:22disruptekit's fine as long as you init all your v2 strs/seqs.
19:22:00ZevvAraq: I guess I should make an issue for that crash?
19:22:54Zevvor is disruptek already fixing it :)
19:23:11Araqof course, it's a showstopper bug
19:23:12disruptekeh i cannot even build head due to a hard styleCheck error.
19:23:56disrupteks/compilerProc/compilerproc/ in memtracker
19:26:54Araqmemtracker? is that still a thing?
19:29:42*dddddd joined #nim
19:34:10Araquh oh
19:34:23Araqbootstrapping fails on my machine
19:34:31Araqthe error message is 'FAILURE'
19:34:42Araqand nothing else
19:36:07*Pqzcih5 quit (Remote host closed the connection)
19:36:21AraqZevv, the "peak mem" reporting seems to broken
19:36:26Araq*to be
19:36:36*Pqzcih5 joined #nim
19:36:45*Pqzcih5 quit (Remote host closed the connection)
19:37:03*Pqzcih5 joined #nim
19:40:02*cron quit (Ping timeout: 240 seconds)
19:42:25AraqZevv, yay, for me Nim's bootstrapping now fails
19:42:33Araqwhat did you do to our allocations?
19:43:18Zevvdang, that was not part of CI of course :/
19:43:32Araqhmm maybe it's my patch
19:43:38Araqlet's see
19:44:35Araqoh yeah
19:44:37Zevvwell I fiddled in one of the most essential core components of the ecosystem, what can go wrong, right
19:44:39Araqfalse alarm
19:44:45AraqI think.
19:45:19Araqbut the "peak mem" bug might be yours
19:45:42Zevvlikely. What is used to get that number?
19:45:45Araqyou can easily check this, if it reports < 400 MB
19:45:59Araqfor 'koch boot' then you misreport it
19:46:12AraqgetMaxMem proc or similar
19:46:44Zevvok
19:49:15Araqhmm ok it was my bug
19:49:24Araqand peak mem reporting seems fine too, sorry for the noise
19:50:09Zevvno problem. I just will remember the false accusations for the decades to come.
19:50:13*jwm224 quit (Ping timeout: 258 seconds)
19:52:58*narimiran quit (Ping timeout: 260 seconds)
19:53:26*Ven`` joined #nim
19:53:41*narimiran joined #nim
19:55:02*drewr quit (Ping timeout: 260 seconds)
19:55:27*cron joined #nim
19:56:44*jwm224 joined #nim
19:58:35*cron1 joined #nim
19:59:48*cron quit (Ping timeout: 258 seconds)
20:04:05Zevvthat was a nice red herring
20:04:18Zevvthere's a nimZeroMem happening on a const struct
20:04:50Zevvvalgrind has a hard time explaining that, funnily enough
20:06:02disrupteki have a hard time explaining where my memory went, too.
20:06:29euantorWas glancing at the `parsecfg source, and noticed the case for the `[` token sets the literal to `]`. It does seem to wok, but was a little surprising.- is it intended? https://github.com/nim-lang/Nim/blob/devel/lib/pure/parsecfg.nim#L357
20:07:06Yardanicolol
20:07:40Yardanicomaybe it works because the tok.kind is set correctly
20:07:48Yardanicoand code doesn't really need tok.literal of that kind
20:08:09Yardanicoah no, actually it is used https://github.com/nim-lang/Nim/blob/devel/lib/pure/parsecfg.nim#L425
20:08:38Yardanicoah nvm it's not, this is the only place where it's used and the code gets the closing bracket in that branch
20:09:19YardanicoI guess this is a result of ctrl+c; ctrl+v :P
20:10:02Yardanicothat code was commited 12 years ago
20:11:21*cron1 quit (Quit: Leaving.)
20:12:06*drewr joined #nim
20:12:08disruptekso it's not even old.
20:12:26Zevvok boomer
20:12:31Yardanicolmao
20:12:39disruptektoo funny.
20:13:07*Vladar quit (Quit: Leaving)
20:13:21FromDiscord_<DeltaPHC> Man, Nim is that old?
20:13:24Yardanicoeuantor: so I guess since you found out that typo you have the rights to make a PR for it :P
20:13:44Yardanico@DeltaPHC well yes, first commit in nim repo is https://github.com/nim-lang/Nim/commit/405b86068e6a3d39970b9129ceec0a9108464b28
20:14:09Yardanicobut AFAIK it was developed since 2006 by Ar*q, but he only published it in 2008
20:14:12YardanicoI might be wrong though
20:14:14euantorYeah, I had a glance and couldn't see anything that would break by changing it, but haven't actually tested that theory yet
20:14:41Yardanicowell that's why we have CI :P
20:14:49FromDiscord_<DeltaPHC> I do recall when it was called Nimrod, but I didn't quite realize it was so long ago
20:15:20FromDiscord_<DeltaPHC> That name change was a good decision, lol
20:15:45Yardanicoenglish is not my native language so there wasn't any real difference for me :P
20:15:50Yardanicoalthough a shorter name is a bit nicer I guess
20:16:30euantorhere goes nothing, let's see if the CI goes mental :)
20:16:46Yardanicoinb4 it breaks the compiler somehow xd
20:17:03Yardanicobut really I think it shouldn't since rawGetTok is not even exported
20:18:47*ng0_ joined #nim
20:19:14*ng0 quit (Ping timeout: 265 seconds)
20:21:51*opal quit (Changing host)
20:21:51*opal joined #nim
20:22:41Yardanicoguess I need to leave nim repo, i'm collaborator for like 2 years now but I didn't do anything meaningful :P and I'm kinda always scared I would press a "merge pull request" accidentally xd
20:23:11AraqZevv, https://github.com/nim-lang/Nim/pull/13244
20:23:13disbot➥ make goto based exceptions available for 'nim cpp'
20:23:23AraqYardanico, you could also start to contribute
20:23:31YardanicoAraq: yeah, I know, but I would do it with PRs anyway :D
20:24:05*marmotini_ quit (Remote host closed the connection)
20:24:37ZevvAraq: ah right, evalConstExprAux was my problem, thanks
20:25:26ZevvI assume that fixed the 'can only raise ref object' error?
20:25:57Araqno, I didn't do any testing whatsoever
20:26:35Zevvoh ok, let me check
20:26:44Araqer if FTP is dead how come I keep getting PRs about it?
20:27:47*cron joined #nim
20:28:59*sagax quit (Read error: Connection reset by peer)
20:29:05Zevv/home/ico/external/Nim/lib/system/excpt.nim(571, 12) Error: only a 'ref object' can be raised
20:30:04ZevvIt tries to raise a C exception, so I guess the 'except StdException' has to go if :goto is enabled
20:30:09Zevvbut I couldn't find a flag for that yet
20:31:55*marmotini_ joined #nim
20:35:27*sagax joined #nim
20:46:39*cron quit (Quit: Leaving.)
20:47:54*ng0_ is now known as ng0
20:51:39FromGitter<zetashift> FTP?
20:52:00*narimiran quit (Ping timeout: 268 seconds)
20:52:12*Ivan_ joined #nim
20:52:49*Ivan_ quit (Client Quit)
20:55:23Zevvthis arc crash is remarkably tricky
21:05:13dadadaYardanico: just discovered your nympylib, it's awesome and it really blows me away that you even can emulate python style classes with it
21:05:28*marmotini_ quit (Remote host closed the connection)
21:06:01*marmotini_ joined #nim
21:08:51Zevvaraq can you explain something to me? I don't know much about the internals, but what I see is that with arc if I add object A to a seq, A will be memZero'd out after - this happens after a eqsink so I assume the compler deduces the sink here. The zeroed out object is a 'static const' struct, which gcc sometimes puts in a readonly page, and sometimes not - not sure what this depends on
21:09:33ZevvI can 'fix' this by not emitting a 'const' struct, but that's probably too simple
21:10:30*marmotini_ quit (Ping timeout: 268 seconds)
21:14:14leorizeor you can also not emit a static :P
21:15:06leorizewe could give up a bit of potential performance / less memory usage for determinism imo
21:18:20*Ven`` quit (*.net *.split)
21:18:20*dawkot quit (*.net *.split)
21:18:20*dom96 quit (*.net *.split)
21:18:21*Yardanico quit (*.net *.split)
21:18:21*jxy quit (*.net *.split)
21:18:21*bozaloshtsh quit (*.net *.split)
21:18:21*EvergreenTree quit (*.net *.split)
21:18:22*livcd quit (*.net *.split)
21:18:23*lbart quit (*.net *.split)
21:18:23*Tanger quit (*.net *.split)
21:21:53*Trustable quit (Remote host closed the connection)
21:26:14*jxy joined #nim
21:26:15*livcd joined #nim
21:27:59FromGitter<hlaaftana> Lol, I made the FTP PR, the bug i fixed is a pretty bad one if you look at it, the fact that it was around since 2016 is a pretty good indicator that no one uses FTP
21:28:30disrupteknice; new bugs, old code. i love it.
21:30:54disruptekZevv: so the way it works is, we snoop the nim repo. we create issues with snippets at a foobar repo. foobar is where we enact all the automation and store all the traffic about a bug. yeah, we get full history with whatever github behavior is enabled.
21:31:40disruptekthe bot can snoop foobar repo and pull out code for testing. it'll actually put all of those automated tests in a single package, maintained separately.
21:32:37disruptekthe first comment it makes will be to confirm that someone audited the issue and marked it as legit.
21:33:27disruptekthat'll provide the necessary link. other than that, it won't need to spam the nim repo at all. i think maybe adding flags might not be too disruptive, but i don't think it's strictly necessary.
21:35:31*cron joined #nim
21:36:33disruptekit'll just fill out the first comment of the post with a complete postmortem; anything we can think of to put in there. we can regenerate it whenever we touch it, so we don't need to spam ourselves with myriad tiny updates.
21:37:24disruptekof course, you'll be able to use this in any project, but the compiler will be the best testbed.
21:43:17Zevvso you want to exctract snippets automatically
21:43:51disruptekyeah, but i'm already doing that.
21:43:55Zevvbut you can't expect these to make sense? often it's not a complete runnable snippet
21:43:58disruptekjust ignore the uninteresting bits.
21:44:16disruptekyou can make new tickets in the repo that are "sensical" if you want.
21:44:30disruptekyou can also simply fix the existing ticket.
21:44:41Zevvhm right
21:45:56disruptekyou can start with the existing tests, for example.
21:48:28Zevvthere's no way to add some kind of metadata to an issue probably
21:48:38Zevvoh you can re-update the same comment, that's what you mean?
21:48:38disruptekbasically, we have all this already. the bits i expect to build are a new ci engine and some glue to connect it to github.
21:48:43disruptekright.
21:48:55disruptekthe data can go into html that's hidden one way or another.
21:49:20disruptekit'll write an email version, then edit/replace it with the sexy github version that also includes the data.
21:49:26Zevvor you can put it somewhere in a git repo and just add a link to the commit in a comment
21:49:43Zevviirc git commits can cross reference over repos
21:49:46disruptekyep. it will be writing the tests repo anyway.
21:49:56Zevvso, is it done yet?
21:50:15disruptekwe'll just put all the data in the test. we can even use the doc generator to generate nice web-pages from it that include the original explanation, etc. etc.
21:50:38disrupteki'm relying upon you to do your usual devil's advocate thing.
21:51:12ZevvI can see this work, but not without significant manual intervention
21:51:49disrupteklike what?
21:52:11ZevvI might not understand all of it, but you want to take snippets from issues and make tests out of those?
21:52:22disruptekevery issue is a test, yes.
21:52:34disruptekif the issue is open, the test is failing. if it's closed, the test passes.
21:52:41Zevvoooh right
21:52:45disruptekwe only open one issue at a time.
21:53:08disruptekso it's not like you have 1200 open issues.
21:53:21Zevv1,466
21:53:41disruptekwell, it'll be less when the bot runs and confirms that some bugs are fixed.
21:54:58Zevvok, I was on my way to bed, I'll let this sink in. You'll have to fill me in on the details with a beer next week
21:55:08disrupteksleep tight, zevv.
21:55:14Zevvo/
22:00:38*icebattle quit (Ping timeout: 268 seconds)
22:01:42*Ven`` joined #nim
22:01:42*dom96 joined #nim
22:01:42*Yardanico joined #nim
22:01:42*bozaloshtsh joined #nim
22:01:42*EvergreenTree joined #nim
22:01:42*lbart joined #nim
22:01:42*Tanger joined #nim
22:04:11*marmotini_ joined #nim
22:05:59disruptekyou'll leave a comment like "!mwe" and some code in the nim-side ticket to add/update a list of examples that reproduce the problem. of course, these can turn into your tests that demonstrate a fix, etc.
22:07:31disrupteki guess you could even link the docs against the tickets/tests. or you could use it to get a heatmap of instability across the language. 😁
22:11:14*filcuc joined #nim
22:16:38*krux02 quit (Remote host closed the connection)
22:18:02*solitudesf- joined #nim
22:18:27disruptekclearly you want performance graphs, too. those can be regenerated whenever the test inputs change. hmm, maybe this should run as a fork. then it will work seamlessly for smaller projects that don't mind integrating ci into their issues.
22:20:45*solitudesf quit (Ping timeout: 265 seconds)
22:22:51*filcuc quit (Ping timeout: 240 seconds)
22:26:22*abm joined #nim
22:30:18*ptdel quit (Ping timeout: 245 seconds)
22:31:55*icebattle joined #nim
22:33:12*abm quit (Quit: Leaving)
22:33:42disruptekwhat we're really talking about here is just an integration of the local development environment and github. when people share their environments, they share this testing load. in fact, we could even do crazy stuff, like when we see that you built xyz with q+abc+e and we have a test for abc, we can build abc incrementally and provide the result cheaply.
22:35:45Yardanicoi wonder what's the most viewed nim forum thread 🤔
22:36:59*icebattle quit (Ping timeout: 265 seconds)
22:37:21*blackbeard420 quit (Quit: ZNC 1.7.5 - https://znc.in)
22:37:22*hpyc9 quit (Quit: ZNC 1.7.5 - https://znc.in)
22:39:02*lritter quit (Quit: Leaving)
22:39:18*solitudesf- quit (Ping timeout: 268 seconds)
22:41:01*blackbeard420 joined #nim
22:41:30*hpyc9 joined #nim
22:42:12*ptdel joined #nim
22:48:38disruptekthis system can learn the fastest way to test the project, but it also knows what's related. so it needs to project entry from any file, or really all files in a commit.
22:57:56disruptekwhich really means that we operate commit-by-commit. we just monitor the rest of the repo and act on any merge that doesn't have our performance data in it. we measure that data and write it into a new commit.
22:59:49disrupteknow your repo has test data in it from real user environments. anyone who wants to help test can do so and automatically create prs to upload their data. you also know which file changes affect which tests.
23:00:27disruptekknowing the fastest way to test your project means you know the most brittle bits. but you know that for any package across the ecosystem that adopts such tests.
23:00:46disruptekyou can estimate your risk across the whole project, an for any given commit.
23:00:55disruptekin any dependency.
23:03:07disruptekyou get that data, hosted locally in your git clone, regardless of whether you participate in testing.
23:08:20AraqZevv, if it's a NIM_CONST it's a "blueprint" for the object and should be copied into the object
23:08:30Araqelse it's not NIM_CONST and we can mutate it
23:08:59Araqdoesn't seem hard to fix but probably you should just look at my PR tomorrow
23:14:47*ltriant joined #nim
23:33:21disrupteki always liked ncurses because i never had trouble building it or linking it.
23:33:32disruptekbut why is it so large?
23:39:49Araqdunno.
23:51:32disruptekif they didn't want me to get drunk on hand sanitizer, why did they make it so tasty?
23:53:42Araqwhy would anybody watch "Two and a half men" with Ashton Kutcher.
23:54:00disruptektwo questions for the eternities.
23:59:21*dadada quit (Ping timeout: 265 seconds)