<< 11-10-2018 >>
00:09:35*Tortice quit (Remote host closed the connection)
00:09:58*owner_ joined #nim
00:10:16*owner_ is now known as Guest71257
00:12:27*mr_yogurt quit (Ping timeout: 252 seconds)
00:14:35*wildlander quit (Quit: Konversation terminated!)
00:14:55*mr_yogurt joined #nim
00:32:14*druonysus joined #nim
00:32:14*druonysus quit (Changing host)
00:32:14*druonysus joined #nim
00:34:51*zachk quit (Quit: Leaving)
00:48:26*vlad1777d quit (Ping timeout: 244 seconds)
00:49:50FromGitter<timotheecour> @kaushalmodi are you still here?
00:51:57*chemist69 quit (Ping timeout: 252 seconds)
00:52:55*chemist69 joined #nim
01:05:48*vqrs quit (Ping timeout: 252 seconds)
01:06:54*Yardanico quit (Ping timeout: 252 seconds)
01:07:26*planetis[m] quit (Ping timeout: 276 seconds)
01:07:26*stisa[m] quit (Ping timeout: 276 seconds)
01:07:39*MrAxilus[m] quit (Ping timeout: 250 seconds)
01:07:39*TheKing[m] quit (Ping timeout: 250 seconds)
01:07:39*Demos[m] quit (Ping timeout: 250 seconds)
01:08:48*stisa[m] joined #nim
01:08:49*planetis[m] joined #nim
01:09:20*MrAxilus[m] joined #nim
01:09:21*TheKing[m] joined #nim
01:09:56*flyx quit (Ping timeout: 260 seconds)
01:10:36*flyx joined #nim
01:11:18*Yardanico joined #nim
01:11:30*Demos[m] joined #nim
01:12:24*vqrs joined #nim
01:12:33FromGitter<timotheecour> Never mind, cc-d you here: https://github.com/nim-lang/Nim/issues/9301
01:24:42*Yardanico quit (Ping timeout: 252 seconds)
01:25:03*abm quit (Ping timeout: 245 seconds)
01:34:29AlexMaxshashlick: How do I prepend lines to the start of the file?
01:34:34AlexMaxlike #mangle?
01:37:12*planetis[m] quit (Ping timeout: 252 seconds)
01:37:21*vqrs quit (Ping timeout: 252 seconds)
01:37:30FromGitter<kayabaNerve> include? What?
01:37:47FromGitter<kayabaNerve> I have no idea what mangle is and no idea about the prior debate.
01:37:56*petersjt014[m] quit (Ping timeout: 252 seconds)
01:37:57*zacharycarter[m] quit (Ping timeout: 252 seconds)
01:38:09*ryuo quit (Ping timeout: 252 seconds)
01:38:18*dyce[m] quit (Ping timeout: 252 seconds)
01:38:53*vqrs joined #nim
01:40:07AlexMax@kayabaNerve I'm talking about nimgen
01:40:29*ryuo joined #nim
01:43:00*planetis[m] joined #nim
01:44:17FromGitter<kayabaNerve> Never mind. :P
01:44:27*dyce[m] joined #nim
01:44:55*zacharycarter[m] joined #nim
01:45:05*petersjt014[m] joined #nim
01:52:27AlexMaxI'm having complicated issues with --nep1
01:52:31AlexMaxit's frustrating
01:56:37*Demos[m] quit (Ping timeout: 250 seconds)
01:57:53*Demos[m] joined #nim
01:58:52*Yardanico joined #nim
02:00:47*Tongir quit (Remote host closed the connection)
02:01:16*thomasross quit (Remote host closed the connection)
02:01:39*thomasross joined #nim
02:04:25*banc quit (Quit: ZNC - http://znc.in)
02:04:53*stefanos82 quit (Quit: Quitting for now...)
02:04:57FromGitter<yyyc514> ok mac client
02:04:59FromGitter<yyyc514> that’s better
02:05:10FromGitter<yyyc514> is there a way to get rid of the “type here” help text?
02:05:12FromGitter<yyyc514> i know where to type
02:06:16FromGitter<yyyc514> sorry will take this to gitter
02:07:33*PrimHelios joined #nim
02:08:30*thomasross quit (Ping timeout: 252 seconds)
02:08:40*Tongir joined #nim
02:19:34*thomasross joined #nim
02:20:01*banc joined #nim
02:33:07shashlick@AlexMax: prepend
02:34:27shashlickI've not used --nep1 so far since it results in wrappers that are too messed up
02:34:50shashlickAnd need too much repair to get functional
02:40:09AlexMaxshashlick: I tried that, but I was ending up with a single quote at the start of the .nim file
02:40:42shashlickCan you post the cfg
02:40:47AlexMaxnah i threw it away
02:40:59AlexMaxI was trying to --nep1 the config file
02:41:09AlexMaxcan you use just "prepend" by itself and it'll prepend to the start of the file?
02:41:16shashlickYes
02:41:28AlexMaxokay, not sure why that didn't work then
02:41:34AlexMaxmaybe it was because i used single quotes?
02:41:37AlexMaxeither way, moot point
02:42:02shashlickYa I'm not sure parsecfg likes single quotes
02:42:36AlexMaxI wish I knew enough about c2nim to try and fix its issues
02:42:54AlexMaxbut I'm not sure I want to go _that_ far down the layer cake
02:44:03shashlickSame here, I've opened several issues with c2nim but it's not gotten much attention
02:44:28shashlickI'm not sure I want to go that far, will be good for me but impossible to find time these days
02:44:42shashlickKeep https://github.com/nim-lang/nimble/issues/549 in mind with your nimble file
02:44:51shashlickWhat are you wrapping by the way
02:45:07AlexMaxI was trying to fix nimnuklear to respect --nep1
02:45:29shashlickYep I tried that myself but it's a can of worms
02:47:34AlexMaxdid you see? the most bizare issue I had was that some identifiers just wouldn't get capitalized
02:47:38AlexMaxno idea why
02:47:49AlexMaxI figured out why some (int, char, etc.) weren't
02:48:00AlexMaxbut others like "cursor" wouldn't either
02:48:34AlexMaxsorry for any onlookers, I do complain about c2nim and friends a lot, but only because Nim has been such a pleasure in most other respects
02:49:35AlexMaxya know
02:49:40AlexMaxthere's also cimgui
02:50:44AlexMaxcimgui seems like it was designed to be generated
02:52:29shashlickWell none of this would be possible without c2nim
02:53:30shashlickThat being said, I'd love to see Nim interop with C directly
02:53:30AlexMaxhttps://github.com/cimgui/cimgui/tree/master/generator/output
02:53:55AlexMaxThis is very interesting
02:56:11shashlickIt's only for imgui?
02:57:35AlexMaxYes. But I wonder how difficult it would be to parse those .json files and turn them into Nim bindings
02:58:05AlexMaxIt might allow skipping of c2nim entirely
02:58:14AlexMax(although as a downside, imgui seems much more complicated)
02:59:53AlexMaxor I could suck it up and keep using nimnuklear in its current state :P
03:00:03AlexMaxdespite not respecting NEP1
03:00:09shashlickI can give it a spin
03:00:09shashlickI can give it a spin
03:00:23shashlickBut didn't we have an imgui wrapper already
03:02:55AlexMaxno idea
03:03:07AlexMaxshashlick: I don't want to add more to your plate if you're already busy :P
03:03:28AlexMaxespecially considering that right now there is a "good enough" solution already in place
03:08:02FromGitter<kayabaNerve> Does newString(1) return [0] or [0, 0]?
03:08:24FromGitter<kayabaNerve> (do I need to include the null terminator byte in the count)
03:10:24FromGitter<kayabaNerve> It appears to exclude the null terminator in the count.
03:12:53shashlickhttps://github.com/lmariscal/nimgl/blob/master/src/nimgl/imgui.nim
03:13:55leorizekayabaNerve: here's the internal on `newString(1)`: `{Sup = {len = 1, reserved = 7}, data = someAddress ""}`
03:14:05leorizegot it through gdb
03:14:23shashlickRepo is active so seems it should be functional
03:17:55FromGitter<Varriount> @kayabaNerve The terminator byte should be thought of as an implementation detail.
03:20:00*chemist69 quit (Ping timeout: 252 seconds)
03:22:08*chemist69 joined #nim
03:27:27FromGitter<kayabaNerve> @Varriount Meaning...?
03:27:55FromGitter<Varriount> Meaning that procedures in the standard library ignore it.
03:27:59FromGitter<kayabaNerve> Got it
03:28:24FromGitter<Varriount> `newstring(1)` returns `"\0"`
03:29:13FromGitter<Varriount> The fact that the underlying string structure is actually `"\0\0"` is an implementation detail meant for optimization.
03:44:29AlexMaxshashlick: dang, wow
03:57:09*PrimHelios quit (Quit: Leaving)
04:34:11*endragor joined #nim
04:35:26FromGitter<xmonader> @kaushalmodi good job (y)
04:37:27FromGitter<xmonader> Guys i tried loads of combinations and don't seem to be able to get it work ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ libssl.a exists in /usr/lib/musl/lib and all the headers are in /usr/lib/musl/include so what am I doing wrong? [https://gitter.im/nim-lang/Nim?at=5bbed38764cfc273f9ba2aad]
04:38:53leorizeuse `--passL:'-static -lssl -lcrypto'`
04:39:33leorizewhen you pass `lib.a`, you're refering to the file `lib.a` in the current location, not in the search path
04:45:17FromGitter<xmonader> i see let me try thanks @leorize
04:45:21FromGitter<kaushalmodi> @xmonader yeah, pass the absolute path to the .a files
04:45:33FromGitter<kaushalmodi> See the .nims for pcre support
04:45:53FromGitter<kaushalmodi> May be you copy paste that for ssl and send a PR?
04:46:18FromGitter<kaushalmodi> I used passC for setting the include path
04:47:44FromGitter<kaushalmodi> @timotheecour I saw my mention in issue https://github.com/nim-lang/Nim/issues/9301. Thanks for the analysis.
04:54:52FromGitter<xmonader> @kaushalmodi i tried
04:55:09FromGitter<xmonader> ```code paste, see link``` ⏎ ⏎ breaks with lots of undefined references... [https://gitter.im/nim-lang/Nim?at=5bbed7adf659e67772968f2f]
04:57:33*narimiran joined #nim
04:57:48leorizeif there're undefined refs then probably you're missing some libraries
04:58:20leorizeunlike dynamic linking, when you link with static archive you have to manually link with all of its dependencies
05:02:39FromGitter<xmonader> @leorize mainly ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbed96fbbdc0b25051f0749]
05:03:12FromGitter<xmonader> openssl_init sounds like the first thing should be in ssl headers i include no?
05:04:11leorizeweird. Have you tried to link it my way?
05:07:24FromGitter<xmonader> @leorize exactly the same ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbeda8c38449236611c33fd]
05:10:00FromGitter<gogolxdong> Is there a library wrapped Vulkan?
05:15:04FromGitter<gogolxdong> What's the comparison of glm and ArrayMancer?
05:18:41FromGitter<kaushalmodi> @xmonader you have two --passL in your last snippet, not sure if that matters
05:19:55FromGitter<kaushalmodi> I can trying hacking with this tomorrow when at computer. But hopefully you or leorize can get to it.
05:20:32FromGitter<kaushalmodi> I have no experience with C other than what I got via Nim :)
05:21:13*chemist69 quit (Ping timeout: 252 seconds)
05:22:21*chemist69 joined #nim
05:26:19FromGitter<xmonader> can't have multiple passL? to the linker? not sure, today will be a busy morning i'll try to investigate this afternoon, please if anyone has an idea ping me
05:29:21FromGitter<bung87> guess you may specify a path for compiler search header files
05:30:28FromGitter<xmonader> @bung87 isn't that achieved here? ⏎ ⏎ ```--passC:'-I/usr/lib/musl/include'``` [https://gitter.im/nim-lang/Nim?at=5bbedff4c08b8b3067fd546b]
05:36:06FromGitter<juancarlospaco> ```code paste, see link```
05:36:28FromGitter<juancarlospaco> https://pastebin.com/ESBepPJe
05:37:19FromGitter<juancarlospaco> I dunno were error comes from, is from installing `httpauth` and its dependencies.
05:39:43FromGitter<bung87> does your lib header files also there?
05:42:05FromGitter<juancarlospaco> Im on linux, arch 64bit, stable 0.19.0, happens on any path, the weird is that Path kinda hardcoded there.
05:42:39*nsf joined #nim
05:47:11FromGitter<timotheecour> @kaushalmodi you’re welcome :)
05:49:18FromGitter<bung87> if so ,I dunno then
05:51:01FromGitter<kayabaNerve> Thoughts on Nim compiler options to detect unused imports/things that could be `func` but aren't?
05:52:24*darithorn quit ()
05:59:55*krux02 joined #nim
06:03:14FromGitter<timotheecour> hi @krux02 are you here?
06:03:42*Tongir quit (Remote host closed the connection)
06:04:22krux02timo the courious: yes I am
06:05:13FromGitter<bung87> @xmonader have you tried --cincludes:DIR ?
06:05:58FromGitter<timotheecour> :) actually yes indeed i’m curious, on high level, is your PR almost ready or just failing on https://github.com/krux02/Nim/blob/sizeof-alignof/tests/typerel/trectuples.nim#L6 ?
06:06:49*Tongir joined #nim
06:12:48*narimiran quit (Quit: Konversation terminated!)
06:27:29hohlerdeshould I close the client socket in the async callback proc of AsyncHttpServer? or lets say, is it safe to do so?
06:32:18krux02timotheecour: I fixed the test and that is the only test failing
06:33:14krux02but that is a hard PR, whenever I fix something here something that shouldn't be related to sizeof breaks
06:33:30krux02computesizealign was inteded to compute the size and the alignment of a type
06:33:40Araqthat is the nature of the compiler
06:33:40krux02but now it is much more than that
06:33:59Araqit's complex stuff
06:34:43leorizexmonader: This is important: What is the version of openssl you have installed there?
06:35:12Araqoh and here is another "easy" thing for the hacktober
06:35:19krux02it is responsible to detect illegal type recursions, even those that are unrelated to the computation of the size (like in the last failed test), it is resposible to handle an "unknown size" state for imported types
06:35:35Araq"clean up the tests, merge tiny tests into one"
06:36:49krux02and Araq: I would really like to merge tiny tests into one, but for the tests that expect a compilation error that is not possible afaik
06:37:33krux02would be nice to have a "try catch" for compiler errors
06:37:48krux02but I am not sure if it is worth it, it would be for test cases only
06:37:52Araquse 'nimout' and 'cmd: "nim check"'
06:38:07Araqand that's cool because "nim check" needs more testing too
06:38:57krux02tests/async/tasyncssl.nim failed
06:38:59krux02:/
06:39:28Araqcrash?
06:39:31*xomachine[m] quit (Ping timeout: 260 seconds)
06:39:39*xomachine[m] joined #nim
06:40:33hohlerdethis code crashes with nodejs using the fetch library to do the http call, but works with curl. the culprit seems to be the client.close(). I am hesitating to file a github issue, as I am not sure if I miss something with async procs.
06:40:34hohlerdehttps://privatebin.tarakis.de/?c8de7db969cb8a99#t2Jq+rlSihj13gDM8xKIpC12rzjjadiPSeKP2aYwH4Q=
06:40:39krux02assert fail in the test case
06:42:33FromGitter<timotheecour> *<krux02>* and Araq: I would really like to merge tiny tests into one, but for the tests that expect a compilation error that is not possible afaik ⏎ ⏎ it is, if https://github.com/nim-lang/Nim/issues/8803 is accepted (I or others could implement it; shouldn’t be hard)
06:43:24Araqtimotheecour: I just explained how to do it without more compiler magic
06:43:46Araqand integration tests are better than unittests
06:44:07Araqhohlerde, https://nim-lang.org/docs/asynchttpserver.html doesn't close anything
06:45:04hohlerdeAraq: yes, it is the close in my code at the end of the callback proc, using Request.client.close()
06:45:35FromGitter<narimiran> @Araq if you want to delegate the boring job of cleaning up the stale branches, my yesterday's offer still stands (i need to be a maintainer for a half an hour)
06:46:50Araqwhat's your github name?
06:47:14FromGitter<narimiran> also narimiran
06:48:37*PMunch joined #nim
06:48:41Araqdone
06:49:52FromGitter<kayabaNerve> Araq If you're just handing those privs out...
06:51:27leorizehohlerde: maybe it's because of a double-close https://github.com/nim-lang/Nim/blob/master/lib/pure/asynchttpserver.nim#L275
06:53:25leorizeyep, it is
06:53:36leorizereproduced with `curl --http1.0`
06:54:10leorizebasically, don't close the connection in the callback, or patch `asynchttpserver` to detect closed connection
06:58:52Araqthis will be so much better with destructors...
07:00:43hohlerdeleorize: thanks a lot. you're right. checked the nodejs call. it is using http1.1 and setting "Connection: close" in header, which results in a double close
07:03:06FromGitter<narimiran> ok, we're now at 35 branches (down from 107). there are several more that could be maybe deleted, but i don't want to do it before consulting here
07:04:35FromGitter<narimiran> `araq-devel`, `araq-misc`, `araq-better-docgen`are all listed as active, but they are all behind the devel, with no commits of their own. keep them?
07:04:52PMunchHmm, if you try to use a unicode character in a character literal you get a "missing closing ' for character literal" error. That error message should be improved
07:07:06FromGitter<kayabaNerve> Destructors are broken as hell right now.
07:08:06Araqdefinitely for `araq-better-docgen`
07:08:08FromDiscord<cipharius> Talking about char literals, same error is thrown by `\n` character. Why is it not accepted in first place, while `\t` is fine
07:08:23Araqnot sure about araq-misc and araq-devel but these should have been merged too
07:08:35FromGitter<kayabaNerve> Nope. Thinking of something else.
07:08:38FromGitter<kayabaNerve> NVM me.
07:08:43FromGitter<narimiran> besides those branches, there are several stale branches from 3+ years ago with just one commit. i also left them, as i'm not sure what to do
07:09:05Araqcipharius, \n is valid since Nim 0.18, update your Nim
07:09:33FromDiscord<cipharius> Ah, i hadn't tried it in 18.0 yet, as i remembered that being broken
07:09:35FromGitter<narimiran> Araq maybe you can take it from here, as you'll know better what (not) to delete. you can remove me now from the nim-team
07:11:31AraqI'll keep you, I trust you
07:11:53FromGitter<narimiran> shit, now i'm under pressure!
07:13:04FromGitter<narimiran> can i use these privileges to close those old issues i have commented where nim now successfully compiles/runs without an error?
07:14:00Araqusually it's better to close issues with a PR that adds a testcase in order to prevent regressions
07:14:25Araqbut it's a tradeoff because we want the testsuite to remain "speedy" (cough...)
07:14:51Araqand some bugs are simply too improbable to ever return
07:14:57FromGitter<narimiran> ok, will work on testcases for (some of) them :)
07:17:36FromGitter<timotheecour> while we’re on the topic of long needed cleanups, how about https://github.com/nim-lang/Nim/issues/8576#issuecomment-411928292 ? ([git] rename git branch `devel` to `master` and `master` to `stable` )
07:20:00Araqtimotheecour, after 0.19.2 is out
07:20:09FromGitter<timotheecour> k
07:20:18Araq0.19.2 will backport the most pressing bugfixes to the 0.19 branch, let's hope for the best
07:20:50Araqwith the upcoming sizeof() changes and all the other stuff that's coming this trunk-based development can't work :-/
07:21:21FromGitter<timotheecour> I’m so very happy to hear that. Branch based is so much cleaner.
07:22:36FromGitter<kayabaNerve> Why would devel ever be master and stable ever not be master?
07:23:31leorizebecause convention™
07:23:35FromGitter<timotheecour> because 99.999% of projects on github do that
07:23:50FromGitter<kayabaNerve> Sounds like 99.999% of projects are stupid as hell.
07:24:01FromGitter<timotheecour> talk to them ;-)
07:24:03FromGitter<kayabaNerve> In my opinion, ofc, but still. :thinking:
07:24:07FromGitter<kayabaNerve> I'm talking to one right now
07:24:11Araqmost projects on github are not a fullblown programming language
07:24:11FromGitter<kayabaNerve> Literally right now
07:34:35FromGitter<kayabaNerve> I feel that master should be THE master branch. The releases are the confirmed entries into the Nim Compiler, there to stay or be updated later, and declared... stable (hard to say that before 1.0)
07:34:50FromGitter<kayabaNerve> devel means cutting edge features with more bugs.
07:35:39FromGitter<kayabaNerve> I personally feel logic states master should be stable. Devel should be explicitly defined as so.
07:44:57FromGitter<alehander42> @kayabaNerve yeah but that doesn't work well if you want to maintain more than one releases (I guess?)
07:45:20FromGitter<alehander42> at least that's how i understand it after searching for both models
07:47:37FromGitter<kayabaNerve> branch 0.18
07:47:41FromGitter<kayabaNerve> branch 0.18
07:47:46FromGitter<kayabaNerve> stabe
07:47:57FromGitter<kayabaNerve> devel
07:48:57FromGitter<alehander42> well why would devel be not master then
07:49:52FromGitter<timotheecour> exactly @alehander42 ’s point, and the 99.999% argument above.
07:51:20FromGitter<kayabaNerve> By stable I meant master. Sorry. ⏎ ⏎ @alehander42 Because that doesn't explicitly state you're getting the version with more bugs.
07:56:05PMunchA little terminal snake: http://ix.io/1oUb/Nim
07:56:50PMunchJust playing a bit more with sub-character terminal drawings
07:57:19PMunchThis uses the 1/2 block to essentially draw squares at 2x the lines in your terminal
08:01:55*onionhammer quit (Ping timeout: 244 seconds)
08:03:35*gmpreussner quit (Ping timeout: 268 seconds)
08:04:02*gmpreussner joined #nim
08:14:21*floppydh joined #nim
08:16:14*onionhammer joined #nim
08:16:41FromGitter<alehander42> @kayabaNerve well stable explicitly states you're getting the version with less bugs
08:18:07FromGitter<kayabaNerve> But you have to ask to be safe. Why not ask to be unsafe?
08:21:40leorizewhy not ask for both? you can certainly remove the master branch AFAIK
08:23:03FromGitter<narimiran> bikeshedding?
08:23:10leorizewhat color is it?
08:25:40*Zevv left #nim (#nim)
08:32:58PMunchI vote for colouring the master branch blue
08:33:06PMunchAnd the devel branch green
08:45:39*craigger quit (Read error: Connection reset by peer)
08:46:18*craigger joined #nim
08:46:33FromGitter<alehander42> no love for yellow
08:49:49FromGitter<narimiran> stable = black, devel = orange ---> orange is the new black!
08:49:49*abm joined #nim
08:50:52*craigger quit (Read error: Connection reset by peer)
08:51:20*craigger joined #nim
08:54:59*krux02 quit (Remote host closed the connection)
08:55:59FromGitter<kayabaNerve> Why does varargs not works with async?
09:04:00AraqNim doesn't know how to put to the heap and/or the copies would be oh-so-expensive
09:04:15FromGitter<kayabaNerve> K
09:04:22FromGitter<kayabaNerve> Is it issue worthy or na?
09:05:22FromGitter<kayabaNerve> It seems like it should have an issue but also its a PITA and it is already known
09:05:49FromGitter<timotheecour> @ PMunch your link prints weird: "import math import strutils import sequtils im…"
09:06:05leorizeI think ix.io is being broken
09:06:14leorizeso just trim the `/nim` part from the URL
09:06:29FromGitter<narimiran> @timotheecour here's a better version of that link: http://ix.io/1oUb
09:07:47FromGitter<timotheecour> Thanks! (why not point to a GitHub/similar link instead?)
09:08:31leorizebecause ix.io is lightweight?
09:08:44leorizeand you can paste from command line with just `curl`?
09:08:46FromGitter<mratsim> ^, this, I’m always wondering when people use strange pastebins
09:09:33Araqso ... was IUP ported to OSX finally?
09:10:06*craigger_ joined #nim
09:10:48leorizealso your typical pastebin wouldn't let you host testresults.html ;)
09:11:15*craigger quit (Ping timeout: 250 seconds)
09:11:35leorizeAraq: is this expected on 32bit OS? ```Error: unhandled exception: /home/tumble/Nim/lib/pure/math.nim(717, 14) `log(4.0, 3.0) == ln(4.0) / ln(3.0)` [AssertionError]```
09:15:57*dddddd joined #nim
09:15:59Araqleorize, the tests are not green on a 32bit OS for unknown reasons
09:16:10AraqI think appveyor 32 is disabled :-/
09:16:53*craigger_ quit (Ping timeout: 250 seconds)
09:18:50*endes[m] quit (Ping timeout: 276 seconds)
09:18:58*endes[m] joined #nim
09:20:08*petersjt014[m] quit (Ping timeout: 276 seconds)
09:20:20*petersjt014[m] joined #nim
09:20:26leorizeSome tests fail due to overflow and/or size differences between types of different architectures
09:21:25leorizebut weirdly, some fails because the `Hint:` lines weren't filtered by testament
09:22:47Araqsome tests need to be fixed and probably also some things in the compiler and libraries
09:24:22*banc quit (Ping timeout: 268 seconds)
09:24:23leorizeThere's https://github.com/nim-lang/Nim/issues/9138 with a 10 days old log attached, if anyone want to try fixing 32 bit bugs
09:26:23*banc joined #nim
09:28:26*craigger joined #nim
09:31:20*couven92 joined #nim
09:38:24FromGitter<xmonader> @leorize OPENSSL_VERSION=1.0.2o
09:51:14leorizexmonader: that's why it didn't work
09:52:28leorizetry `-d:openssl10`
09:52:36floppydhhey, I've been absent for a while, is dynamic-linking still "the-way" to go for nim? - I remember wanting to statically-link my stuff but I remember that it was very uncommon
09:52:55leorizefloppydh: you can easily static link stuff now
09:52:59PMunchtimotheecour, leorize, seems like ix.io doesn't like to create line numbers in my paste..
09:53:19floppydhleorize: how? any keywords?
09:53:31leorize--dynlibOverride
09:53:39leorizethen pass the link flags yourself
09:53:48leorizebetween modules then it has always been static-linked
09:54:12PMunchmratsim, timotheecour, because I have a keyboard shortcut that grabs whatever I have selected, creates a ix.io paste out of it, and puts the resulting link in my clipboard
09:54:26PMunchJust a lot easier than copy-pasting it into a web-page and copying the link back out
09:54:43FromGitter<xmonader> @leorize thank you will try it
09:54:45FromGitter<mratsim> I just use gitter :P
09:55:35floppydhleorize: oh I see, but I have to deal with getting the static-libs and then passing their flags all by myself no?
09:55:45*vlad1777d joined #nim
09:55:46leorizexmonader: if that doesn't work also then update your openssl to 1.1.0+
09:56:21FromGitter<xmonader> @leorize that's perfect will try to do the whole musl flow now at work for pcre and openssl
09:56:23FromGitter<xmonader> thank you
09:56:50leorizefloppydh: yes, but at least it's universally appliable to all wrappers now ("assuming they use {.dynlib.} pragma")
09:57:04floppydhleorize: I see
09:58:22FromGitter<timotheecour> for me the problem with stuff like http://ix.io/1oUb is there are no “back-links” (pointing to author or github parent repo etc) or ways to write comments, etc, unlike something like https://gist.github.com/
09:59:20leorizeix.io is use for simple snippet sharing, and usually used with a chat service, like IRC
09:59:33PMunchWell I don't want all that stuff @timetheecour, I just want a simple way to share snippets with people on IRC
09:59:36leorizean alternative to pasting code on gitter for IRC user
10:05:06FromGitter<alehander42> i just paste stuff on gitter too
10:05:34FromGitter<alehander42> besides, i think there are similar plugins for gist, there was one for sublime
10:08:51PMunchWith my system I can copy anything from anywhere into a paste and immediately have it on my clipboard as a link. No plugins to set, no nothing
10:12:34*craigger quit (Read error: Connection reset by peer)
10:13:42*craigger joined #nim
10:16:16*craigger_ joined #nim
10:18:13*craigger quit (Ping timeout: 252 seconds)
10:19:53FromGitter<tim-st> Is an RFC for Concept std types needed? https://github.com/nim-lang/Nim/issues/9289
10:24:29PMunchhttps://gist.github.com/PMunch/4455921b1c35772f8a59d1e1711e43a5
10:25:03PMunchCleaned the code up a bit, fixed some bugs, and made the timing better
10:25:19planetis[m]how can I compile runnableExamples? nim c -r module doesn't run them
10:25:39FromGitter<timotheecour> nim doc foo.nim
10:26:31FromGitter<kaushalmodi> planetis: nim doc ..
10:27:08planetis[m]thx
10:31:11planetis[m]How to name a mutable string concept add `Mutable`, `Mut` as a prefix or postfix?
10:31:23planetis[m]@tim-st do you want to make an RFC?
10:32:39Araqthe convention is an M prefix, 'mitems'
10:32:49Araqso it could be MStringLike
10:33:03planetis[m]ok
10:33:19Araqbut I don't know what this means :P
10:33:31AraqAddableString ?
10:34:10Araqmutability is usually already encoded in the 'var'ness of the parameter type
10:37:46*hoijui joined #nim
10:38:25FromGitter<mratsim> unless it’s a ref
10:38:43FromGitter<mratsim> but yeah, why do we need a concept for var string?
10:39:37planetis[m]so I will add the `c.add(char)` directly to the StringLike concept
10:40:14FromGitter<mratsim> sounds like architecture astronauting to me :P beyond strings what kind of type would use that?
10:41:21FromGitter<mratsim> If you define the concept properly you can accept a `var MyConcept` as input
10:42:06planetis[m]sure I think it is simpler and nicer
10:42:45Araqplanetis[m], which proc requires the MutableString?
10:42:48FromGitter<mratsim> Actually I think we might need something for Utf strings
10:43:01FromGitter<mratsim> but probably `add` should be part of the concept definition
10:43:18planetis[m]none it was suggested here: https://github.com/nim-lang/Nim/issues/9289
10:44:56planetis[m]i will close the Pr so it doesn't rebuild each time
10:45:18Araqseems to me that return type concepts will be lots of fun...
10:45:56Araqthe input arguments are always easy, it can be an Indexable or an Iterable
10:46:17FromGitter<tim-st> Araq: all procs that take `var string` should get the mutable StringLike I think
10:46:24*elrood joined #nim
10:46:38Araqbut how the heck am I supposed to know the return type in the new shiny concept world?
10:46:58FromGitter<tim-st> planetis[m]: no, I didnt want to write one (I have no good plan for the inheritance and how many types)
10:47:12FromGitter<tim-st> Araq: return types always string?
10:47:20Araqyeah, great :P
10:47:33FromGitter<tim-st> for StringLike at least
10:48:11Araqso ... we pretend to not know the input type but the output type is string and everything ever written in Nim uses 'string'
10:48:14planetis[m]algorithm procs took openarray and returned seq. so nothing changes
10:48:33Araqok that's fair
10:49:05Araqalso, the RFC said, algorithm and sequtils are affected
10:49:17Araqnot to patch 'string' everywhere else
10:49:48FromGitter<tim-st> should the concept type definition be in system.nim ?
10:50:00Araqno.
10:50:21Araqthe std / baseconcepts or something
10:50:39planetis[m]at least parseutils no?
10:50:46FromGitter<tim-st> and strutils
10:51:33FromGitter<narimiran> strlikeutils :P
10:51:36Araqno in my opinion, but it's another fight I'll lose
10:52:02FromGitter<mratsim> let’s remove strings, close 1000 issue
10:52:20FromGitter<tim-st> I dont see the use case of a type StringLike when the procs for strings dont work with it
10:52:43AraqI don't see the use case of StringLike when every return type stays 'string'
10:53:14FromGitter<narimiran> what is StringLike besides string?
10:53:25FromGitter<narimiran> array[char] doesn't count
10:53:34FromGitter<mratsim> unicode string
10:53:45*Vladar joined #nim
10:53:46Araqyeah... in fact
10:53:50FromGitter<mratsim> those probably need some care
10:53:58Araqthe 'char' aspect is what can be abstracted over
10:53:59planetis[m]I think strutils shouldn't be touched
10:54:52FromGitter<mratsim> type StringLike = string or seq\[Rune\]
10:56:00Araq"Rune" is wrong too, it needs to be whatever Unicode says is most letter-like
10:56:30FromGitter<tim-st> I would use something like this for StringLike :) https://github.com/status-im/nim-ranges/blob/481a3d308684cdda4ca37cf3b43239a1a69029b3/ranges/memranges.nim#L5
10:56:41FromGitter<narimiran> @timotheecour what's the point of #9304, after it was decided that nimblepkglist.nim should stay?
10:57:12FromGitter<mratsim> @tim-st this is a workaround to toOpenArray not being a value type
10:57:31FromGitter<alehander42> ok, would the implementation of add be the same for strings and for unicode ?
10:57:33FromGitter<alehander42> i doubt it
10:57:36Araqnarimiran: move nimblepkglist.nim to tools/
10:57:48FromGitter<alehander42> concepts should be used where you can reuse an algorithm
10:57:53FromGitter<kaushalmodi> @narimiran Can you commit a quick changelog edit about the xlmdom and xlmdomparser moved out to graveyard?
10:58:11FromGitter<alehander42> add seems like something that would just need overloading
10:58:21FromGitter<mratsim> Yeah, I don’t want Nim to become the Java of procedural programming language
10:58:21FromGitter<tim-st> @mratsim does `toOpenArray` has a length (len) ?
10:58:32FromGitter<kaushalmodi> (or should I open an issue for that? It's kinda part if that cleanup issue)
10:58:37FromGitter<mratsim> @tim-st `toOpenArray` is implemented as pointer + len
10:59:13FromGitter<tim-st> so when there was a type it would be the same as array[runtimeSize, Type] ?
10:59:59*xet7 joined #nim
11:00:02FromGitter<tim-st> if that works, then openArray[char] is enough to be StringLike
11:00:21FromGitter<narimiran> @kaushalmodi changelog.md is still about 0.19.0, from what i can tell?
11:00:25FromGitter<mratsim> more like tuple\[p: ptr UncheckedArray\[T\], len: len\]
11:00:25FromGitter<alehander42> but you wouldn't represent unicode string as openArray[char] ?
11:00:50FromGitter<timotheecour> @narimiran feel free to close #9304 if nimblepkglist.nim must stay
11:00:57FromGitter<mratsim> openarray\[char\] shouldn’t match string.
11:01:02FromGitter<mratsim> or vice-versa
11:01:10FromGitter<tim-st> @mratsim it does match string
11:01:13FromGitter<tim-st> in 0.19.0
11:01:16Araqbut it does and it was added after a feature request...
11:01:30FromGitter<tim-st> and I like it :)
11:01:32FromGitter<mratsim> but it was removed after a freature request after 0.17.2 no?
11:02:03AraqI don't know, we changed that array[char] is compatible with cstring
11:02:12FromGitter<mratsim> I think we should provide a toString with a `lent` something
11:02:17FromGitter<mratsim> ah
11:02:28FromGitter<kaushalmodi> If the removal happened on devel branch, the changelog on devel branch should reflect that imo, @Araq? May be add 0.19.2 header?
11:02:51planetis[m]well sheep `StringLike` overly restrictive doesnt even fit my use case
11:03:03Araqas I said, take the 0.19 tag, apply/cherry-pick bugfixes, release 0.19
11:03:10Araq*0.19.2
11:03:45FromGitter<mratsim> @Araq: https://github.com/nim-lang/Nim/issues/6350
11:04:33Araqyeah, implicit conversions can hurt
11:05:24FromGitter<narimiran> @timotheecour there's another solution: to move it to tools, i'm working on it
11:06:09FromGitter<timotheecour> lol ya, i just posted that on that issue
11:07:49FromGitter<mratsim> I’m gonna write a RFC “The meaning of string” :P
11:07:51FromGitter<timotheecour> `more like tuple\[p: ptr UncheckedArray\[T\], len: len\]` => openArray should just be that
11:08:36FromGitter<mratsim> with a lifetime and “lent”-checker attached to it :P
11:09:06Araqwrite an RFC instead, "how to program"
11:09:56FromGitter<timotheecour> I’ve said it before but string should be implemented in terms of seq, with string litterals being special in that they add a trailing 0 for cstring compatibility just in that case. So we only need to define seq, and string is a regular `seq[char]`
11:10:25Araqyou've said it before, I disagreed back then and still do :P
11:10:32FromGitter<mratsim> seq\[char\] and string should be separate
11:10:38FromGitter<narimiran> bug-FIXES, guys! ;)
11:10:46FromGitter<mratsim> it’s semantically quite different.
11:11:08FromGitter<mratsim> worse is conflating string with seq\[byte\]
11:11:10Araqyup and with Unicode it became hopeless to abstract over and keep performance
11:11:12*gangstacat quit (Quit: Ĝis!)
11:12:19FromGitter<timotheecour> for unicode all we just need iterators, eg `byGrapheme` etc
11:13:52Araqa grapheme itself is of variable length, no?
11:15:17FromGitter<timotheecour> 1 grapheme = variable number of characters indeed
11:22:33Araqtimotheecour, any opinion on my compiler/pathutils? I think we should move large parts from it to ospaths
11:25:02FromGitter<timotheecour> how about ospaths2 according to “tradition” for other modules ?
11:30:01Araqhuh? we would patch ospaths to use my PathIter primitive
11:30:26Araqand I would keep the strongly typed variants for the compiler
11:30:55Araqbecause there would be endless discussions about these 4 distinct string types
11:31:18FromGitter<timotheecour> which ones? AbsoluteFile … RelativeDir?
11:31:23Araqyes
11:31:58FromGitter<timotheecour> as a user I’d like to have an entirely typesafe alternative to ospaths (eg ospaths2)
11:32:40FromGitter<timotheecour> likewise with stuff like writeFile(path:Path, contents: string) which would be impossible to type in wrong order (unlike what we have now)
11:33:35AraqI agree but this is a larger change and also related to the "remove system.nim RFC"
11:34:11FromGitter<timotheecour> so basically it deserves an RFC
11:34:20FromGitter<timotheecour> (also its 4am here, need to crash…)
11:34:28Araqoh sorry, good night
11:34:35Araqno RFC please :P
11:34:44Araqjust an implementation change for now
11:36:05FromGitter<timotheecour> well can u just write a blurb in an issue of what’s ur intent with moving compiler/pathutils into ospaths? will look tmrw
11:36:18FromGitter<timotheecour> ok thanks gtg
11:38:40*krux02 joined #nim
11:40:59*vlad1777d_ joined #nim
11:41:43*vlad1777d quit (Ping timeout: 245 seconds)
11:47:49krux02it seems like implementing sizeof in the compiler was much more work than I ever thought it would be.
11:48:11krux02but it works now, tests run through
11:51:40FromGitter<mratsim> @tim-st: https://github.com/nim-lang/Nim/issues/9312
11:52:07FromGitter<mratsim> @timotheecour as well, for good dreams :P
11:53:10FromGitter<mratsim> @araq, I think contributing.rst should be top level
11:55:54Araqno, I can set that in the github settings somewhere, I think
11:56:34FromGitter<mratsim> you should also add “add an empty line at the end of your file”, this can be configured in your code editor
11:57:01Araqargh... moar RFCs, that's it
11:57:11Araqgonna create nim-lang/rfcs
11:57:59FromGitter<mratsim> nim-lang/specs. Once a RFC is accepted it becomes a spec ;)
11:58:05FromGitter<mratsim> welcome to bureaucracy!
11:58:33FromGitter<mratsim> though if you require those specs to be written in RST instead of Markdown I’m sur it will cut those in half :P
11:59:14Araqsounds like a feature to me
12:07:23FromGitter<kaushalmodi> (and pandoc downloads shoot through the roof)
12:24:47FromGitter<narimiran> re: code owners: i want to see more of LemonBoy there!
12:28:15FromGitter<narimiran> sweet jesus, there are 89 open issues labeled RFC!
12:33:50*fangs joined #nim
12:38:07*stefanos82 joined #nim
12:40:56*fangs left #nim (#nim)
12:41:24FromGitter<kayabaNerve> So I found a bug.
12:41:32FromGitter<kayabaNerve> It's weird as hell.
12:42:07FromGitter<kayabaNerve> I try not to directly link my project but I have this file which works fine. https://github.com/EmberCrypto/Ember/blob/master/src/MainGlobals.nim
12:42:36*gangstacat joined #nim
12:42:57FromGitter<kayabaNerve> I changed the last part to https://gist.github.com/kayabaNerve/1aac9f97cf7964fa0936bfe0a29eab29
12:43:02FromGitter<kayabaNerve> Generally, I would say I f***ed up and move on. Here's the error though.
12:43:15FromGitter<kayabaNerve> MainGlobals.nim(24, 13) template/generic instantiation from here ⏎ MainGlobals.nim(26, 9) Error: unreachable statement after 'return'
12:43:47*Yardanico quit (Ping timeout: 252 seconds)
12:43:52FromGitter<kayabaNerve> I'm 99.9% sure I don't have any return statement there.
12:44:23FromGitter<kayabaNerve> It only appeared when I switched to async. I assume I should create a GH issue for this but honestly I would love validation this isn
12:44:34FromGitter<kayabaNerve> *isn't a me-problem and is actually a problem with the language.
12:50:23FromGitter<kayabaNerve> I minimized it; definitely asyncdispatch; creating an issue; NVM.
12:50:28Araqasync maybe injects a return for you
12:50:46Araqbut sure, report it
12:50:58FromGitter<kayabaNerve> It's because of quit in async
12:51:02FromGitter<kayabaNerve> If I remove that, it's fine
12:56:07*Yardanico joined #nim
13:11:41*endragor quit (Remote host closed the connection)
13:13:24FromGitter<bung87> @mratsim https://github.com/nim-lang/Nim/issues/9312#issuecomment-428930747 I commented.
13:13:45FromGitter<kaushalmodi> I am trying to do something like this: ⏎ ⏎ ```proc mapconcat*[T](s: openArray[T]; op: proc(x: T): string = `$`; sep = " "): string =``` [https://gitter.im/nim-lang/Nim?at=5bbf4c89ae7be940163b66fd]
13:14:01FromGitter<kaushalmodi> what's a valid way so that the `op` defaults to `$`
13:14:02FromGitter<kaushalmodi> ?
13:14:25FromGitter<mratsim> the propername is concatmap by the way
13:14:40FromGitter<kaushalmodi> ah!
13:14:46FromGitter<kaushalmodi> that's why I couldn't find it :P
13:14:55FromGitter<kaushalmodi> wait
13:14:58FromGitter<kaushalmodi> does that exist?
13:15:02FromGitter<mratsim> Here you go: https://github.com/mratsim/Arraymancer/blob/v0.1.0/src/arraymancer/utils/functional.nim#L88-L90
13:15:16FromGitter<mratsim> that was like 1.5 years ago, I’ve removed it since then
13:15:25FromGitter<kaushalmodi> cool
13:15:27FromGitter<kaushalmodi> why?
13:15:39FromGitter<kaushalmodi> I was missing elisp functions like those in nim: https://www.gnu.org/software/emacs/manual/html_node/elisp/Mapping-Functions.html
13:15:48FromGitter<kaushalmodi> it's `mapconcat` in elisp
13:15:58FromGitter<mratsim> it’s called concatmap in Haskell
13:16:03FromGitter<kaushalmodi> that name makes sense because you first map through all elements and then concat them
13:16:04krux02kaushalmodi: overloading
13:16:08FromGitter<mratsim> mostly because I didn’t need it
13:16:23FromGitter<kaushalmodi> krux02: what do you mean by that comment?
13:16:35FromGitter<mratsim> function are done from inner to outer concat(map(…))
13:16:45krux02the op = `$`
13:16:55FromGitter<kaushalmodi> that's what I did
13:16:59FromGitter<kaushalmodi> doesn't compile
13:17:17FromGitter<kaushalmodi> here's the rudimentary code that I expected to compile: https://ptpb.pw/vpXm/nim
13:17:26krux02I think mapconcat is a stupid name
13:17:28*tdc quit (Ping timeout: 268 seconds)
13:17:33FromGitter<kaushalmodi> @mratsim I am looking for a way to set a default proc for the `op` param
13:17:36FromGitter<mratsim> can you try adding a generic `$` that is unexported?
13:18:10FromGitter<mratsim> proc `$`T (x: T): string = system.`$`(x)
13:18:17krux02it is for string creation
13:18:32krux02with mapconcat I would assume that it also works for Seq type
13:18:43krux02but that isn't the case it can only produce a string
13:18:52FromGitter<kaushalmodi> @mratsim thanks trying it out
13:19:12krux02you can look in system and see how $ on seq is implemented
13:19:47FromGitter<mratsim> otherwise there is this weird ```varargs[T, `$`]``` syntax
13:20:07FromGitter<kaushalmodi> @mratsim it's not yet clear how to set that proc as the default in the signature
13:20:12FromGitter<kaushalmodi> I get: Error: '$' cannot be passed to a procvar
13:20:20FromGitter<kaushalmodi> so what can I pass instead?
13:20:21FromGitter<mratsim> Like here: https://nim-by-example.github.io/varargs/
13:20:39FromGitter<mratsim> you have to use a template then magics cannot be passed as proc param
13:21:16krux02the problem is `$` is more than a syngle proc
13:21:27FromGitter<kaushalmodi> krux02: hmm, ok
13:21:46FromGitter<kaushalmodi> @mratsim I don't understand "magics". I'll hold off on this till I understand all that
13:22:00FromGitter<kaushalmodi> problem with that varargs example is that it is hard-coding the `$`
13:22:07krux02magic means it is implemented in the compiler.
13:22:09*vegax87 quit (Changing host)
13:22:09*vegax87 joined #nim
13:22:09*vegax87 quit (Changing host)
13:22:09*vegax87 joined #nim
13:22:23FromGitter<mratsim> magic are proc that are not really Nim proc but directly implemented in the compiler or in C
13:22:26FromGitter<kaushalmodi> in above proposed mapconcat sig, I want `op` to be any proc that take T and returns string
13:22:28FromGitter<mratsim> like + -
13:22:38FromGitter<kaushalmodi> hmm
13:23:28FromGitter<bung87> that would depends whether T has a $ proc implement.
13:23:45FromGitter<kaushalmodi> it would be in general
13:24:16krux02kaushalmodi, I don't think nim has a sytax to get a function pointer by name and signature
13:24:52FromGitter<bung87> a type T has repr proc is general.
13:25:15FromGitter<mratsim> you can cast a proc() to a ByteAddress or an int
13:27:33krux02kaushalmodi: this is the problem http://ix.io/1oVn
13:31:25krux02http://ix.io/1oVo
13:31:28krux02with output
13:32:36*endragor joined #nim
13:32:39FromGitter<Vindaar> @kaushalmodi like this? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf50f7600c5f64238ec6e3]
13:33:57*elrood quit (Quit: Leaving)
13:35:51FromGitter<kaushalmodi> @Vindaar woh! yes :)
13:36:53*endragor quit (Ping timeout: 250 seconds)
13:37:19FromGitter<Vindaar> :)
13:37:29FromGitter<kaushalmodi> thanks
13:37:39FromGitter<kaushalmodi> wonder if this simple proc can make into sequtils
13:39:19FromGitter<kaushalmodi> krux02: I *still* don't understand macros.. I don't know what I should expect out of that snippet
13:40:13FromGitter<kaushalmodi> @Vindaar I am starting to collect emacs-lisp fns in a sep nim lib
13:40:33FromGitter<kaushalmodi> This was the first fn that I sorely missed
13:41:06FromDiscord<deech> I'm trying to grok concepts and I can't understand why the example in this issue doesn't work. https://github.com/nim-lang/Nim/issues/9289
13:41:07FromDiscord<deech> I reduced the working example to:
13:41:07FromDiscord<deech> ```
13:41:07FromDiscord<deech> type
13:41:07FromDiscord<deech> StringLike*[T] = concept c
13:41:07FromDiscord<deech> c[int] is T
13:41:07FromDiscord<deech>
13:41:09FromDiscord<deech> proc test(s: StringLike) =
13:41:11FromDiscord<deech> echo s[0] == '2'
13:41:11FromDiscord<deech>
13:41:12FromDiscord<deech> test("Hello")
13:41:13FromDiscord<deech> ```
13:41:14FromGitter<kaushalmodi> I didn't like manually making sure that spaces were present in snippets like these: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf52fac7bf7c3662f26940]
13:41:15FromDiscord<deech> and the failing one to:
13:41:16FromDiscord<deech> ```
13:41:17FromDiscord<deech> type
13:41:19FromDiscord<deech> StringLike = concept c
13:41:20FromDiscord<deech> c[int] is char
13:41:21FromDiscord<deech>
13:41:23FromDiscord<deech> proc test(s: StringLike) =
13:41:24FromDiscord<deech> echo s[0] == '2'
13:41:25FromDiscord<deech>
13:41:26FromGitter<kaushalmodi> woh, deech!
13:41:27FromDiscord<deech> test("Hello")
13:41:28FromDiscord<deech> ```
13:41:30FromDiscord<deech> The only difference is the failing one doesn't take a type parameter. What's the difference?
13:41:57FromGitter<tim-st> @mratsim thanks! btw I didnt meant that openArray[char] matching was reintroduced in 0.19.0 I just meant that it works in the current version
13:41:59FromGitter<Vindaar> @kaushalmodi oh, that's a good idea! :) do you have that online already?
13:42:14FromGitter<kaushalmodi> no, I just did nimble init
13:42:45FromGitter<kaushalmodi> I got into a rabbit hole
13:42:55FromGitter<kaushalmodi> was working on something else and NIm errored: ⏎ ⏎ > ../../../stow/pkgs/nim/devel/lib/system/nimscript.nim(237, 7) Error: unhandled exception: FAILED: ./configure --prefix=/home/kmodi/sandbox/nim/hello_musl/openssl/1.1.1 no--sharedno-zlib-fPIC
13:43:18FromGitter<kaushalmodi> turns out I forgot to manually put space between "no-shared" and "no-zlib" and "-fPIC"
13:43:30FromGitter<kaushalmodi> and made me wish that Nim had mapconcat
13:43:31FromGitter<kaushalmodi> :P
13:43:42*vlad1777d_ quit (Ping timeout: 252 seconds)
13:47:33*SenasOzys quit (Ping timeout: 245 seconds)
13:49:19krux02kaushalmodi: you know emacs lisp, maybe you should look at macros there
13:49:29krux02for example the rx macro
13:49:33krux02and then see what it does
13:49:51krux02and then you can try to reproduce something similar with nim macros
13:50:16FromGitter<narimiran> @Araq when you said this morning "clean up the tests, merge tiny tests into one" (https://irclogs.nim-lang.org/11-10-2018.html#06:35:35), did you mean something like this: https://github.com/narimiran/Nim/commit/fe3b5ec5a8f08bd9d744c68a5b2dc999a706f5a9
13:50:42FromGitter<narimiran> i've done that for three folders, but would like to know if this is the way to go before i continue
13:51:43FromGitter<kaushalmodi> krux02: Yes, I will eventually learn the Nim macros, but am just procrastinating because I haven't yet needed to tinker with it
13:51:49krux02@narimiran: in his name I can say yes
13:51:52Araqnarimiran: seems ok
13:52:02krux02ah you are back
13:52:08FromGitter<narimiran> thanks guys, will continue to work on these before sending PR
13:52:17FromGitter<kaushalmodi> krux02: Even without macros, I spend lot of time tinkering in other Nim things :P
13:52:33FromGitter<alehander42> ah the fact that you need to escape with `""` in `"""` strings is so confusing
13:53:37FromGitter<alehander42> surprises are bad :(
13:53:50Araqhuh? there is no need for that
13:54:36FromGitter<alehander42> strange
13:54:57FromGitter<alehander42> i tried it normally the first time
13:55:57FromGitter<alehander42> ah i am sorry
13:56:08FromGitter<alehander42> i meant in <stuff>".." not in triple " strings
13:56:28krux02let myRawStr = r" Hello""World!\"
13:57:02krux02this is where you have to escape the " with "
13:57:18*Snircle joined #nim
13:58:13FromGitter<alehander42> ok, this makes sense
13:58:28FromGitter<alehander42> and i need `\` to escape in `&".."`
13:58:38FromGitter<alehander42> but then i need `""` to escape in `sql".."`
14:00:46*Snircle quit (Client Quit)
14:01:21*abm quit (Read error: Connection reset by peer)
14:01:56leorizealehander42: `"..".sql`
14:02:03krux02alehander42: I have sql syntax highlighting in the sql string literals :P
14:02:49krux02did you ever try to do: sql"""
14:03:14FromGitter<alehander42> yes, but for shorter queries i typically use sql".."
14:03:29krux02doesn't sql support single quotes as well?
14:03:37FromGitter<alehander42> "..".sql is a ok workaround, it's just i don't get the inconsistency
14:04:23krux02“single quotes are to ‘quote’ within double quote without creating nesting confusion”
14:05:09krux02alehander42: It is super consistent
14:05:27FromGitter<alehander42> i mean i get why r".." might take `\` literally
14:05:28krux02it's a raw string made for regular expressions
14:05:40FromGitter<alehander42> but all other <stuff>".."
14:05:44FromGitter<alehander42> usually don't need it
14:06:24FromGitter<alehander42> so I guess because originally r".." syntax was for regexes and later generalized, sql".." fmt".." etc also escape differently?
14:06:26leorizealehander42: here is an another way: `sql ".."` (mind the space between sql and "..")
14:06:50krux02especially for the other stuff it is important to have the raw string literals
14:07:22FromGitter<alehander42> the confusing thing is
14:07:27FromGitter<alehander42> `&".."` escape with `\`
14:07:45krux02echo"Hello\nWorld"
14:07:48krux02mind the space
14:08:05leorizebecause `&` is a operator
14:08:37leorizethe rule for raw literals only exists for identifier AFAIK
14:08:48*UxerUospr joined #nim
14:08:53*Snircle joined #nim
14:08:53Araqyup, that's also clearly covered by the spec
14:09:08FromGitter<narimiran> can anybody explain why we have these: https://github.com/nim-lang/Nim/tree/devel/tests/benchmarks ? their syntax is quite wrong, i guess they are not run
14:09:18FromGitter<alehander42> i think 90% of new users would just try different things until their escape compilers
14:09:48Araqthat's ok as long as the other 10% become valuable contributors
14:09:51Araq:P
14:10:01Araqand as long as the absolute numbers are high
14:10:07FromGitter<alehander42> valuable contributors need users :D
14:10:18FromGitter<alehander42> otherwise they don't find it valuable to contribute
14:10:28leorizenarimiran: they don't start with `t` so I assume they aren't run
14:10:51FromGitter<kaushalmodi> @Vindaar This is just the beginning https://github.com/kaushalmodi/nimy_lisp
14:10:58FromGitter<alehander42> ok, at least it would be cool to improve the error somehow
14:11:00*SenasOzys joined #nim
14:11:11FromGitter<alehander42> as currently it doesn't even seem to have anything to do with the string
14:11:17FromGitter<alehander42> but that might be hard
14:12:05FromGitter<narimiran> leorize: then another question is: even if they were ok (which they are not), why are they in the tests folder?
14:12:48leorizeAraq: ^
14:13:10FromGitter<alehander42> @krux02 postgres supports single quotes for strings, but double for identifiers (afaik)
14:13:26*jhorwitz_ joined #nim
14:13:36*jhorwitz_ quit (Client Quit)
14:13:51*jhorwitz_ joined #nim
14:15:13FromGitter<kaushalmodi> Araq: Would would be good answers to these questions? https://github.com/nim-lang/Nim/issues/9288#issuecomment-428907045
14:15:22FromGitter<kaushalmodi> the PR that closed that issue doesn't answer those
14:15:53*jhorwitz_ quit (Client Quit)
14:16:40leorizekaushalmodi: the manual describes the spec, so it doesn't specify the exact number
14:16:55FromGitter<kaushalmodi> apart from the number..
14:17:00FromGitter<kaushalmodi> the other 2 questions
14:17:13FromGitter<kaushalmodi> so the TRM just dies because it inflooped the first time?
14:19:29Araqyeah I'm looking at it, it's a bug, I think
14:19:50Araqbut the bailout should be based on scoping, hmmm
14:20:39FromGitter<kaushalmodi> thanks. can you please reopen that issue then?
14:21:51*krux02_ joined #nim
14:22:07*krux02_ quit (Remote host closed the connection)
14:23:49*hoijui_ joined #nim
14:27:42*hoijui quit (Ping timeout: 252 seconds)
14:30:00*NimBot joined #nim
14:30:18*PMunch quit (Quit: Leaving)
14:35:47FromGitter<stvnfx> Just asking a dumb question here but if I interface with a cpp static library file, nimsnappy in my case and I use the c backend the linker cant fine the symbols. is there anyway around this?
14:37:29FromGitter<stvnfx> I am sure this has been asked but maybe my google fu is a bit lacking :)
14:38:36FromGitter<stvnfx> It does work when I use the nim cpp backend when I staticaly link it
14:52:36FromGitter<mratsim> {.passC:”-DSYMBOL”.} or {.passl:”-LSYMBOL”.} is the great escape hatch
15:01:20*floppydh quit (Quit: WeeChat 2.2)
15:04:03FromGitter<stvnfx> hmmm let me read more about that, I did something similar where I added {.passL:"alibsnappy.a”.} when I compile to a static lib file but the linker would complaint about missing symbols. I was hoping there was some sort of way to demangle the cpp snappy lib file. I am not the best with cpp and its the only dependency that is compiled with cpp
15:04:59FromGitter<stvnfx> @mratsim thanks for the response 😄
15:06:38FromGitter<stvnfx> just making one fat executable for work to ship easily
15:06:55FromGitter<mratsim> you probably need a dynlibOverride somewhere
15:07:38FromGitter<tim-st> @stvnfx do you use this library? https://github.com/NimCompression/nimsnappyc
15:07:56FromGitter<mratsim> Here is how I statically link to Intel MKL in my library: https://github.com/mratsim/Arraymancer/blob/master/nim.cfg#L72-L82
15:08:06FromGitter<bung87> is there pragma `{.this: self.}` for initObj ? has many properties to init
15:08:26FromGitter<mratsim> "this" is deprecated
15:08:28leorizeI think it's being deprecated
15:08:55leorizeso `result = Obj(field1: val1, field2: val2)`
15:09:14FromGitter<stvnfx> @tim-st no, I found this https://github.com/dfdeshom/nimsnappy I did not see that one
15:09:39FromGitter<tim-st> that should fix your problem I assume :)
15:10:12FromGitter<stvnfx> @tim-st yes defenitly :)
15:10:13FromGitter<bung87> hmm if I use `result = ` should I also need `{.noinit.}` ?
15:10:25FromGitter<stvnfx> thanks @tim-st @mratsim
15:10:55FromGitter<mratsim> @bung87 noInit is for arrays mainly
15:10:57FromGitter<tim-st> @bung87 I think `{.noInit.}` goes to the return type definition of the proc
15:11:09FromGitter<tim-st> (when you want it for result)
15:12:03FromGitter<tim-st> maybe not useful anymore for strings?
15:12:12FromGitter<tim-st> and seq
15:12:15FromGitter<bung87> ok I accept that
15:12:34FromGitter<mratsim> Tracked in that issue: https://github.com/nim-lang/Nim/issues/8745
15:12:53FromGitter<mratsim> but it’s been seriously improved for all heap types
15:13:29*narimiran joined #nim
15:21:23ryuoI've been watching Nim for awhile and I was curious. What makes it worthwhile to people here compared to more mainstream languages?
15:22:06FromGitter<rokups> any idea why nim would fail to execute vccexe.exe on windows? https://pastebin.com/ZNCR7Kx5
15:23:38leorizeryuo: it's simple, clean
15:24:10FromGitter<mratsim> it’s complex, with lots of architecture astronauting potential ;)
15:24:39FromGitter<mratsim> Also it looks good in Meetups and other social gatherings ;)
15:25:31FromGitter<stvnfx> and it makes tiny binaries :)
15:26:37ryuoto date i've primarily used C and few other scripting languages. it's weird how I knew exactly what I wanted back then but now I feel kinda... paralyzed by all the choices available.
15:27:12FromGitter<mratsim> Pralysis of choice is a real thing, it’s like “better alone than in bad company"
15:27:44FromGitter<mratsim> anyway if you used C you’ll be right at home there. No OOP madness, sane proc.
15:28:06FromGitter<mratsim> memory representation is the same as well, C interop is easy.
15:28:30ryuoyea... but i figure i'll have to learn a language like Python at some point for other reasons.
15:28:48ryuoC... very common yet hardly a skill that people ask for.
15:28:58FromGitter<mratsim> You can even get goodness like OpenMP or computed gotos
15:29:26FromGitter<mratsim> my company is looking into C and C++ devs for Nim positions.
15:29:35ryuoDoesn't nim inherit the same concept of whitespace being meaningful for scope?
15:29:56ryuoHuh. Interesting. I've spent like 8 years with C, primarily with userspace work.
15:30:09ryuoI learned a great deal writing some distribution software in the past.
15:30:32FromGitter<mratsim> yes, scope is white-space aware
15:30:45*stevenfx joined #nim
15:30:59ryuoI've also used C++, but i'm hardly as fluent it in as with C.
15:31:26ryuoIt's like with modern C++ they decided to throw in the kitchen sink for good measure.
15:32:24ryuomratsim: Is your company still looking then?
15:33:00FromGitter<mratsim> You bet: https://status.im/open-positions.html?gh_jid=1196154
15:33:41FromGitter<mratsim> Networking and/or low-level cryptography experience are what we’re looking for the most.
15:33:43*Snircle quit (Read error: Connection reset by peer)
15:34:12FromGitter<mratsim> Embedded devices experience as well.
15:34:31*Snircle joined #nim
15:35:29ryuoMost of my networking experience is with configuring networks and sometimes when I wrote web applications... And I've only used cryptography libraries.
15:36:25ryuomratsim: thanks for the information.
15:38:28*stevenfx quit (Quit: stevenfx)
15:42:00FromGitter<iffy> Can I get a sanity check? Am I barking up the wrong tree here? https://forum.nim-lang.org/t/4300#26770
15:45:47FromGitter<mratsim> @iffy, not really. When you use quote do you get back into the “runtime” domain and Nim compiler has to do a symbol resolution pass again so you get symbols and not identifiers
15:46:00FromGitter<mratsim> You can use the replaceNodes proc I mentionned here: https://github.com/nim-lang/Nim/pull/8531#issuecomment-410436458
15:46:11FromGitter<mratsim> or don’t use quote.
15:46:38FromGitter<mratsim> Build the AST by hand or use a DSL like breeze: https://github.com/alehander42/breeze
15:46:50krux02mratsim: sometimes I wished that ``quote do`` wouldn't do symbol resolution
15:47:18FromGitter<mratsim> 100% agree: https://github.com/nim-lang/Nim/issues/7719
15:47:26krux02it is quite annoying when local identifiers that are not supposed to be used in the DSL get injected there
15:47:36FromGitter<mratsim> I basically have this `replaceNodes` proc in all my projects
15:48:35FromGitter<iffy> I read somewhere that breeze doesn't work on newer versions (but I didn't actually verify)
15:48:45*FromGitter * iffy looks at replaceNodes
15:51:07krux02mratsim: That's interesting
15:51:22FromGitter<iffy> hehe... it literally just replaces the nodes I don't like. Why didn't I think of that :)
15:51:55FromGitter<iffy> I'll give it a shot in my program; thank you @mratsim !
15:52:03krux02I think you can create a quote that creates a `newTree` expression and has an untyped argument
15:52:20FromGitter<kaushalmodi> leorize: Any help with that ssl static linking. I tried it on ssl 1.1.1:
15:52:23FromGitter<kaushalmodi> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf71b7600c5f64238fe2b6]
15:52:50FromGitter<kaushalmodi> I have the ssl compilation .nims online here: https://github.com/kaushalmodi/hello_musl/blob/ssl/config.nims#L96-L106 if you want to review
15:52:50krux02iffy, you can comment in that thread and give some feedback
15:53:02leorizekaushalmodi: link with libcrypto also
15:53:08FromGitter<kaushalmodi> I did
15:53:14FromGitter<kaushalmodi> see that snippet
15:53:31FromGitter<iffy> krux02: 7719 or 8531?
15:54:46leorizekaushalmodi: have you checked if `COMP_CTX_free` exists in `libcrypto.a`?
15:55:05FromGitter<iffy> 7719
15:55:37FromGitter<kaushalmodi> let me grep
15:56:10FromGitter<kaushalmodi> that's in include file
15:56:12FromGitter<kaushalmodi> ```1.1.1/include/openssl/comp.h ⏎ 29:void COMP_CTX_free(COMP_CTX *ctx);``` [https://gitter.im/nim-lang/Nim?at=5bbf729c600c5f64238fe8ef]
15:56:28FromGitter<kaushalmodi> and I am adding `switch("passC", "-I" & sslIncludeDir) `
15:56:54leorizeno, check for it in the archive
15:56:59leorizeuse `nm libcrypto.a | grep ...`
15:57:16FromGitter<kaushalmodi> TIL about `nm`.. trying it out
15:57:55FromGitter<kaushalmodi> seems to be there: `00000000000000d0 T COMP_CTX_free`
15:58:26FromGitter<kaushalmodi> if you are on a GNU/Linux system and have musl installed, you can try it out
15:58:31leorizecan you make Nim prints out the command it runs?
15:58:34FromGitter<kaushalmodi> i.e. cloning the repo
15:58:45FromGitter<kaushalmodi> sure, one min
15:59:24FromGitter<kaushalmodi> wait, I don't know how, if you meant expanding the pass\* switches
15:59:52FromGitter<iffy> @mratsim After adding this to your `replaceNodes` proc it works! ⏎ ⏎ ```of nnkOpenSymChoice: ⏎ return inspect(node[0])``` [https://gitter.im/nim-lang/Nim?at=5bbf7378600c5f64238ff38e]
16:00:13leorizekaushalmodi: like making Nim prints out the command it uses in `Link: `
16:00:20FromGitter<iffy> I don't really know what I'm doing, though (who knows if choosing the first node is always the right thing to do)
16:01:16FromGitter<kaushalmodi> leorize: sorry, I am at loss there.. let me see if increasing verbosity prints those
16:01:46FromGitter<mratsim> @iffy you’re not choosing actually, the compiler will redo the symbol resolution pass
16:02:31FromGitter<iffy> I wish I knew enough about compilers to know what that meant :) My C experience is years ago
16:02:39leorizekaushalmodi: hang on, I'm testing your `config.nims` here
16:02:44leorizethis might take awhile...
16:03:22FromGitter<mratsim> @iffy you have “identifier” —> “existing symbol” —> “overloading resolution” —> “calling the correct proc"
16:03:32FromGitter<mratsim> OpenSymChoice is step 2
16:03:38FromGitter<mratsim> replace nodes reverts to step 1
16:03:56FromGitter<iffy> oh, that makes sense
16:04:02FromGitter<iffy> each of those things is operating on the AST, right?
16:04:08FromGitter<mratsim> yes
16:04:21FromGitter<iffy> okay, that's a very clear explanation; thank you!
16:05:00FromGitter<kaushalmodi> leorize: may be this error helps? https://ptpb.pw/49jV/text
16:05:04FromGitter<kaushalmodi> thanks for testing
16:05:28FromGitter<kaushalmodi> leorize: clone the repo, checkout ssl branch, run `nim musl -d:ssl src/hello_musl_ssl.nim`
16:05:52leorizedoing it atm, but openssl build is kinda slow
16:06:19FromGitter<kaushalmodi> musl task will build automatically
16:06:23FromGitter<kaushalmodi> takes about a min or 2
16:06:23*SenasOzys quit (Ping timeout: 276 seconds)
16:06:36leorizedepends on what machine you're building on
16:06:56FromGitter<kaushalmodi> I am in no hurry. I am glad you opted to help out
16:08:01FromGitter<stvnfx> does anyone have a nimble task/command that they use to package a mac .app folder?
16:09:24FromGitter<mratsim> you probably can reuse how people package C/C++ with xcode cli tools :?
16:09:35FromGitter<mratsim> or compile to obj-C :P
16:09:38leorizekaushalmodi: `async.c:(.text+0x36): undefined reference to `setcontext'`
16:09:42leorizethat's from libcrypto
16:10:02leorizesetcontext should be provided by the libc
16:10:09leorizebut musl might have got rid of them
16:10:44FromGitter<stvnfx> @mratsim yes very true, just wondering if anyone uses something they added to their nimble builds :)
16:11:12leorizekaushalmodi: https://www.openwall.com/lists/musl/2016/02/04/5
16:12:12FromGitter<kaushalmodi> leorize: Thanks for the investigation. So musl + ssl seems like a no-go :(
16:12:18leorizenot really
16:12:24leorizeyou just need to build openssl correctly
16:12:28*Trustable joined #nim
16:12:54FromGitter<kaushalmodi> i.e. build ssl using gcc and then the nim bin using musl-gcc?
16:13:37leorizeno of course
16:14:17leorizeI'll take a look at this
16:14:50*SenasOzys joined #nim
16:14:58FromGitter<kaushalmodi> thanks, I am really at the edge of mindless hacks that I can do with this without understanding that root cause
16:20:47*darithorn joined #nim
16:24:03leorizekaushalmodi: first thing first: move `-lcrypto` to after `-lssl`
16:24:44leorizesecond thing: use http://www.libressl.org/index.html
16:24:54leorizesince it's saner and works
16:25:18*hdias joined #nim
16:26:09*hdias quit (Client Quit)
16:26:20FromGitter<kaushalmodi> leorize: thanks, and the build command stays the same?
16:26:28leorizeyep
16:26:34FromGitter<kaushalmodi> ` sslConfigureCmd = ["./Configure", sslSeedConfigOsCompiler, "no-shared", "no-zlib", "-fPIC", "--prefix=" & sslInstallDir]`
16:26:37leorizeactually not
16:26:45FromGitter<kaushalmodi> will look at INSTALL
16:26:47leorizeit uses `autoconf` so it's more friendly
16:27:02leorizeexactly like pcre
16:27:15*rokups joined #nim
16:28:18leorizekaushalmodi: `sslConfigureCmd = ["./configure", "--prefix=" & sslInstallDir, "--disable-shared"]`
16:28:23leorizesimple as that
16:29:10leorizealso, you might want to parallelize your `make` invocation
16:29:26leorizeuse `make -j<number of processor cores>`
16:31:29*zama_ joined #nim
16:32:00FromGitter<kaushalmodi> thanks! trying it out
16:34:03leorizesmall tips:
16:34:13leorizebuild only the components that you use
16:34:15leorizelike this
16:34:21leorize`make -C crypto`
16:34:25leorize`make -C ssl`
16:34:29*zama quit (Remote host closed the connection)
16:34:33*zama_ is now known as zama
16:34:56leorizethen install them with `make -C crypto install`
16:35:01FromGitter<kaushalmodi> ok, trying that once I see the current attempt work
16:37:47*craigger joined #nim
16:38:27*craigger_ quit (Ping timeout: 250 seconds)
16:39:03FromGitter<kaushalmodi> leorize: I am getting a different error this time: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf7ca7ae7be940163d038a]
16:39:34*Vladar quit (Remote host closed the connection)
16:40:04leorizecongrats, you've built libressl with glibc instead of musl
16:40:28leorizehave you set `CC=musl-gcc` before running `./configure`?
16:40:53FromGitter<kaushalmodi> :facepalm:
16:41:28leorizespeaking of which, your pcre might have been built with glibc instead
16:41:33leorizealso openssl...
16:42:28FromGitter<arnetheduck> haha, cross compiling is always such a PITA.. I used to think it was a viable solution in the past, but I've started warming to the idea that `chroots` or `dockers` are much easier, in terms of getting things right.. there are so many small details needed to avoid leaks between host and target that it becomes an almost inhuman task
16:42:29FromGitter<kaushalmodi> pcre build was fine.. I was like I am not compiling until configure finishes.. so just let me move that setenv
16:42:48FromGitter<kaushalmodi> I did that move yest
16:43:48leorizeconfirmed that openssl were built with glibc, since you can't build openssl without patches on musl :P
16:46:11FromGitter<kaushalmodi> I don't know if I should pursue this further (could be some other stupid mistake): ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf7e5382893a2f3bcb7471]
16:46:40*platoff joined #nim
16:46:51leorizehave you moved `-lcrypto` to after `-lssl`?
16:47:34leorizethis is because of `ld` dependency resolution rules
16:47:42leorizeit loads libssl
16:48:05leorizethen try to find required symbols from libraries loaded **after** it
16:48:51FromGitter<kaushalmodi> it compiles! but..
16:48:59FromGitter<kaushalmodi> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf7efaef4afc4f28585050]
16:49:06FromGitter<kaushalmodi> 😆
16:49:31FromGitter<kaushalmodi> such an informative error
16:49:37leorizeclean the cache and everything
16:51:52FromGitter<kaushalmodi> got the same error.. but trying everything from scratch starting with the libressl download
16:52:10FromGitter<kaushalmodi> btw does it work for you locally?
16:52:18*elrood joined #nim
16:52:25leorizeit does
16:52:36FromGitter<kaushalmodi> that's comforting at least :)
16:52:57leorizewhat kernel are you on?
16:53:04FromGitter<kaushalmodi> I have committed the latest config.nims to the ssl branch
16:53:35FromGitter<kaushalmodi> `uname -r`: 2.6.32-642.6.2.el6.x86_64
16:53:38FromGitter<kaushalmodi> it's RHEL 6.8 ..
16:54:12FromGitter<xmonader> guys i got it to work
16:54:36FromGitter<kaushalmodi> leorize: same issue after clearing cache, rebuilding, etc.
16:54:49FromGitter<xmonader> ```code paste, see link``` ⏎ ⏎ I'm sure there's some useless flag here [https://gitter.im/nim-lang/Nim?at=5bbf8059bbdc0b2505243061]
16:54:51FromGitter<kaushalmodi> @xmonader awesome! looks like my machine is then to blame
16:55:13leorizeI do know that libressl might SIGKILL if your system is too bad :P
16:55:16FromGitter<xmonader> i will go to my german course now and later will try to cleanly build against newer version from ssl
16:55:36FromGitter<kaushalmodi> leorize: Oh well, let me see how Travis fares
16:56:03FromGitter<xmonader> can you check if the flags makes sense @kaushalmodi ? @leorize?
16:56:18leorizeno, sorry :)
16:56:37leorize`musl-gcc` already include those paths, so you may omit them
16:56:38FromGitter<xmonader> Also, my demo went fine and the project compiled `dynamically linked` and people were happy.. and have 20 issues in my backlog now
16:56:56FromGitter<xmonader> yeah i was trying to be more explicit about it :D
16:57:14leorizeand that `--out` is actually passed to the program instead of the compiler
16:57:29leorizekaushalmodi: working here
16:57:43leorizeyour latest commit
16:57:58*aziz joined #nim
16:59:36FromGitter<kaushalmodi> leorize: thanks for confirming!
16:59:48FromGitter<kaushalmodi> waiting to see this work on Travis.. just pushed out that commit
16:59:53leorizeif you still want to use openssl
17:00:02leorizepass `"no-async"` to the configure
17:00:08leorizethat should take care of it
17:02:56*aziz quit (Ping timeout: 276 seconds)
17:05:36*craigger_ joined #nim
17:05:58*craigger_ quit (Read error: Connection reset by peer)
17:06:20FromGitter<alehander42> @iffy ah i haven't looked at breeze, but even if it's broken, it should be easy to fix for 0.19
17:06:57*couven92 quit (Read error: Connection reset by peer)
17:07:05FromGitter<iffy> @alehander42 wait, didn't you write breeze?
17:07:08*craigger quit (Ping timeout: 245 seconds)
17:07:35FromGitter<alehander42> yes, I meant I haven't tested if it's compatible with newest versions of Nim
17:08:09FromGitter<iffy> oh, gotcha
17:08:20*craigger joined #nim
17:09:06FromGitter<alehander42> did your CLI params macro work out btw
17:09:48FromGitter<iffy> It's working out slowly :) My questions today were about that, too.
17:10:16FromGitter<bung87> if a cpp function return const pointer how to implement in nim?
17:10:29FromGitter<iffy> My first working version is going to be garbage, but it will work. Here's what I have so far: https://github.com/iffy/nim-argparse
17:10:36leorizebung87: it returns a pointer
17:11:40FromGitter<bung87> what if a cpo function return ref
17:12:24FromGitter<bung87> cpp
17:12:30leorizeI don't think a cpp function can return a `ref`
17:12:53*craigger_ joined #nim
17:13:13FromGitter<bung87> sorry I mean address
17:13:20*craigger quit (Ping timeout: 276 seconds)
17:13:55leorizeaddress is just a pointer so... `ptr something` or `pointer`
17:15:36*craigger joined #nim
17:15:54*aziz joined #nim
17:16:01FromGitter<bung87> ah so that means I can actually make it work but loss many compiler check?
17:16:11leorizeyes
17:16:31leorize`const pointer` means a pointer to a constant
17:16:51FromGitter<bung87> wow thanks !helpful
17:17:33*craigger_ quit (Ping timeout: 245 seconds)
17:18:15*aziz quit (Client Quit)
17:19:19FromGitter<kaushalmodi> leorize: building openssl, this time correctly using musl-gcc, showed me another issue
17:19:20*craigger quit (Read error: Connection reset by peer)
17:19:26FromGitter<kaushalmodi> this is a never ending rabbit hole
17:19:37FromGitter<kaushalmodi> workaround: https://github.com/openssl/openssl/issues/7207#issuecomment-420814524
17:19:39leorizehave you passed `no-async` to configure?
17:19:44FromGitter<kaushalmodi> yes
17:20:04FromGitter<kaushalmodi> this time, I got "missing linux/mman.h"
17:20:21FromGitter<kaushalmodi> .. and travis build fails while building libressl
17:20:29leorizeconfirmed libressl is saner
17:20:36*craigger joined #nim
17:20:38FromGitter<kaushalmodi> https://travis-ci.org/kaushalmodi/hello_musl/builds/440247816#L1414
17:20:43FromGitter<kaushalmodi> said too soon :P
17:21:19FromGitter<kaushalmodi> it worked!!!!!!!!!!!!!!!!!!!!!!!
17:21:31FromGitter<kaushalmodi> openssl with that insecure memory workaround
17:21:34FromGitter<kaushalmodi> on my machine
17:21:47leorizeofc, given how old RHEL is...
17:21:55leorizeI'm not surprised that libressl doesn't work
17:22:05leorizesomeone complained abt it once IIRC
17:22:48*craigger quit (Read error: Connection reset by peer)
17:22:51FromGitter<kaushalmodi> I cannot believe my eyes!
17:22:55FromGitter<kaushalmodi> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf86eff659e677729bf52b]
17:23:14leorizeif you want to compile stuff with musl libc, probably you'll have better luck with running an actual cross-compiler
17:23:28FromGitter<kaushalmodi> leorize: I owe you your `(favBeverage or "beer")` if we ever meet
17:23:28leorizebtw, `ldd` isn't the best way to check
17:23:35*craigger joined #nim
17:23:49FromGitter<kaushalmodi> then?
17:23:59leorize`file`
17:24:03leorizeor `readelf -d`
17:24:13leorize`ldd` is a security hell
17:24:34FromGitter<kaushalmodi> `file` is nice, tells exactly what it is
17:24:35FromGitter<kaushalmodi> thanks
17:25:07FromGitter<kaushalmodi> you deserve so many thanks! There were great TIL moments today.
17:25:13FromGitter<kaushalmodi> all due to your patience
17:25:54leorizenp :)
17:26:03*craigger_ joined #nim
17:28:27*craigger quit (Ping timeout: 268 seconds)
17:30:31*craigger joined #nim
17:30:31*craigger_ quit (Ping timeout: 252 seconds)
17:34:29*nsf quit (Quit: WeeChat 2.2)
17:39:44FromGitter<kaushalmodi> leorize: If I may bother you one more time.. that build works locally but fails on Travis: https://travis-ci.org/kaushalmodi/hello_musl/builds/440264561#L1367
17:40:05FromGitter<kaushalmodi> I'd have to guess `openSslSeedConfigOsCompiler = "linux-x86_64"` doesn't apply for openssl builds on Travis?
17:41:13FromGitter<kaushalmodi> may be it's as simple as this: https://github.com/mirage/mirage-block-unix/issues/45#issuecomment-196892988
17:41:43FromGitter<Varriount> Anyone know if -fPIC is a default option for GCC?
17:42:02leorizekaushalmodi: you need `linux-libc-dev` installed for Ubuntu
17:42:07leorizeVarriount: it depends
17:42:32leorizebut most of the time, no
17:43:15leorizeunless the distributor enable `-pie` by default
17:43:55FromGitter<Varriount> Why not? Isn't position-independent code more memory efficient?
17:44:12FromGitter<Varriount> (At least for dynamic libraries)
17:44:47leorizebecause you don't build executable with `-fPIC`
17:44:54leorizeunless, `-pie` is used
17:46:12leorizeback in the time an executable won't run if you build it with `-fPIC`
17:47:26*opi_ quit (Ping timeout: 260 seconds)
17:47:45*opi_ joined #nim
17:50:06*Death916 quit (Ping timeout: 252 seconds)
17:50:06*EastByte quit (Ping timeout: 252 seconds)
17:51:30*Araq quit (Ping timeout: 264 seconds)
17:52:29*Death916_ joined #nim
17:52:42*EastByte joined #nim
17:53:09*Araq joined #nim
17:59:17FromGitter<kaushalmodi> leorize: that package didn't help, tried a linux-header pkg, that too didn't work.
17:59:19FromGitter<kaushalmodi> https://travis-ci.org/kaushalmodi/hello_musl/builds/440273887#L1364
18:00:17leorizeno idea how ubuntu does this kind of thing
18:00:35FromGitter<kaushalmodi> wrong line: https://travis-ci.org/kaushalmodi/hello_musl/builds/440273887#L1377
18:00:35leorizeyour problem is that `gcc` ignored `/usr/include` as an system include path
18:01:08FromGitter<kaushalmodi> hmm.. so probably that needs to be added manually? you mean `musl-gcc`?
18:01:32leorizeno, the `musl-gcc` _should_ already did it
18:01:39leorizeadding it manually will make it worse
18:01:58leorizesince glibc symbols will override musl ones
18:02:27FromGitter<kaushalmodi> ah, yes
18:03:40leorizeI think I looked at this the wrong way
18:06:30leorizethose headers aren't accesible with `musl-gcc` on my machine either
18:07:14*flaviu quit (Ping timeout: 250 seconds)
18:07:54FromGitter<kaushalmodi> hmm, I wonder how it got built on our local machines then
18:07:58leorizekaushalmodi: it doesn't compile on mine machine with openssl
18:08:00*TheLemonMan joined #nim
18:08:07FromGitter<kaushalmodi> ?!
18:08:20leorizeI tested with libressl the last time, remember?
18:08:34leorizedied at the same spot as travis
18:08:37FromGitter<kaushalmodi> so the old RHEL 6.8 is special in some way?
18:09:21leorizetry this `echo '#include <linux/version.h>' | musl-gcc -x c -`
18:09:30leorizedoes it spits out any include error?
18:09:37leorizeignore the undef ref to main
18:10:18FromGitter<kaushalmodi> got: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf920a384492366121d859]
18:10:43TheLemonManI just saw I now own parts of the compiler, yay
18:10:43FromGitter<kaushalmodi> musl version mismatch?
18:10:57*flaviu joined #nim
18:11:01leorizewow, your gcc actually able to find that file
18:11:09leorizetry this `echo '#include <linux/version.h>' | musl-gcc -x c -v -`
18:11:14narimiranTheLemonMan: and, if you ask me, you should "own" more things ;)
18:11:24leorizekaushalmodi: I would like to know your include paths
18:11:51TheLemonMannarimiran, land, money and beer
18:12:00narimiranhaha
18:12:46FromGitter<kaushalmodi> leorize: https://ptpb.pw/P9iZ/text
18:13:31leorizelooks like you got away due to how RHEL people configure the compiler
18:14:08narimiranbtw, TheLemonMan, where are you from? (my guess is italy)
18:15:19TheLemonManyep, how did you find that out?
18:15:55FromGitter<kaushalmodi> leorize: so RHEL has some backup include paths to look into if the usual ones fail?
18:16:16leorizeyes, completely breaking how gcc treats `-nostdinc`
18:17:50leorizekaushalmodi: here's a "hack" to make libressl work
18:18:29leorizeadd `#undef SYS__sysctl` to the beginning of `libressl-src/crypto/compat/getentropy_linux.c`
18:18:58leorizeI've no idea where that symbol is actually defined, and how come Ubuntu's gcc define it
18:19:08leorizebut that hack should work™
18:23:13narimiranTheLemonMan: the way you wrote somewhere 1st, 2nd, etc. -> you used that small circle (like for degrees)
18:24:08TheLemonMannarimiran, hah it was WhiteDuke who wrote that test case
18:25:31FromGitter<kaushalmodi> leorize: you mean work on RHEL? The executable simply dies on my machine.
18:26:02leorizeI mean it should make libressl work on travis
18:26:38FromGitter<kaushalmodi> ok, will try that next
18:26:44leorizef
18:27:07leorizefor your RHEL, can you strap a debugger to it and see why was it killed?
18:29:24FromGitter<kaushalmodi> as you might have guessed, my gdb "skills" are almost non-existent as I am not a C coder: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf9683f659e677729c6fd3]
18:29:56FromGitter<kaushalmodi> and bt gives "no stack"
18:30:10FromGitter<kaushalmodi> *may be should run non-release build*
18:30:16leorizeoh yea, cus your `config.nims` kindly stripped the entire executable...
18:33:31FromGitter<kaushalmodi> no luck without striping or any other optimization too
18:33:36FromGitter<kaushalmodi> "no stack"
18:33:49leorize`--debugger:native`?
18:34:25FromGitter<kaushalmodi> "gdb: unrecognized option '--debugger:native'"
18:34:38FromGitter<kaushalmodi> I think I should take a break today and allow you one too
18:35:02leorize:P
18:49:56FromGitter<kaushalmodi> ok, fwiw, that switch didn't work with nim
18:50:35FromGitter<kaushalmodi> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5bbf9b7bc7bf7c3662f4d542]
19:04:15*craigger_ joined #nim
19:05:47*craigger quit (Ping timeout: 250 seconds)
19:06:28*abm joined #nim
19:08:09*rnrwashere joined #nim
19:08:29*rnrwashere quit (Client Quit)
19:08:52*rnrwashere joined #nim
19:11:26*craigger_ quit (Ping timeout: 268 seconds)
19:12:22*craigger joined #nim
19:21:48*xylef joined #nim
19:22:47FromDiscord<An unconspicuous looking friend> what's the easiest way to check the return type of a function call?
19:22:54FromDiscord<An unconspicuous looking friend> some macro like: result_type(foo(int, string))
19:23:02FromDiscord<An unconspicuous looking friend> where you don't need to put in the actual values
19:23:11FromDiscord<An unconspicuous looking friend> just the types
19:24:16FromGitter<kaushalmodi> leorize: I forgot to implement the hack you suggested for libressl on travis. But I am getting the same error after that too: https://travis-ci.org/kaushalmodi/hello_musl/builds/440311276#L1420, https://github.com/kaushalmodi/hello_musl/commit/399a2752f69f340c82afb054da4bd53fdec87591
19:26:35*TheLemonMan quit (Quit: "It's now safe to turn off your computer.")
19:29:59FromGitter<zetashift> @An unconspicuous looking friend maybe `type.name`? https://nim-lang.org/docs/typetraits.html#name%2Ctypedesc
19:32:00*flaviu quit (Remote host closed the connection)
19:32:26*flaviu joined #nim
19:36:46*rokups quit (Quit: Connection closed for inactivity)
19:48:28FromGitter<stvnfx> is there a way to prevent nimble from adding —noNimblePath
19:48:48FromGitter<stvnfx> it seems to be adding it to my tasks when I use setCommand
19:50:48FromGitter<yyyc514> anyway to do a postfix operation like “unless” in ruby with macros?
19:50:58FromDiscord<An unconspicuous looking friend> zerashift: I need something like this to run:
19:50:59FromDiscord<An unconspicuous looking friend> func foo(a: int, b: string): float = discard
19:50:59FromDiscord<An unconspicuous looking friend> func foo(a: int, b: float): string = discard
19:50:59FromDiscord<An unconspicuous looking friend>
19:50:59FromDiscord<An unconspicuous looking friend> type T = type(foo(int, string))
19:51:08FromDiscord<An unconspicuous looking friend> I suppose I'll just need to create a macro
19:51:09FromGitter<yyyc514> “kill(alien) unless alien.super()
19:51:11FromDiscord<An unconspicuous looking friend> shouldn't be too hard
19:51:22dom96An unconspicuous looking friend: Don't paste code in here please
19:51:37FromDiscord<An unconspicuous looking friend> ok
19:51:46FromGitter<yyyc514> unless as an if like operator was pretty easy
19:51:56dom96stvnfx: if you don't want --noNimblePath then execute `nim` with ``exec``
19:52:47FromGitter<stvnfx> @dom96 does nimble add it automatically? I was looking at the nimble test files https://github.com/nim-lang/nimble/blob/master/tests/nimscript/nimscript.nimble as examples :)
19:53:21UxerUosprleorize: dumb question... What does RHEL stand for in this context?
19:53:39FromGitter<stvnfx> I really like the `—` template, very cool
19:53:47dom96stvnfx: yes, that's the point of calling `nimble` :)
19:53:57dom96Why do you not want --noNimblePath?
19:56:18FromGitter<stvnfx> Unless I am wrong but thats where the compiler is getting some of the imports I use. I could be completely wrong here. I just stumbled on it now
19:56:45dom96Nimble reads your .nimble file and passes dependencies explicitly
19:57:07dom96This ensures others can build your package
19:57:28dom96if you just use `nim` then you can use any package you have installed
19:57:41dom96which won't work for others if they haven't installed them
19:58:47FromGitter<stvnfx> ah ok I see, that makes sense, I have been doing a lot of static linking of small libs and trying to reduce the size of a app for work. just spending some time making the project build with nimble
19:59:06FromGitter<stvnfx> so mostly I have been using a make file
19:59:33*darithorn quit ()
20:01:52FromGitter<stvnfx> Btw, your talk on youtube was amazing. I hope in the next month or two if time allows it to give a dev talk on it at work. right now we have nim in one app
20:02:24FromDiscord<deech> Any help with this appreciated, or even juat a pointer to the code in the compiler that does concept resolution. https://discordapp.com/channels/371759389889003530/371759389889003532/499939505487347715
20:03:11FromGitter<yyyc514> shorter way to write `result = result & rna©` ?
20:03:20FromGitter<yyyc514> hahaha, silly markdown
20:03:31FromGitter<kaushalmodi> `result.add(..)`
20:03:41FromGitter<yyyc514> ah &= works too
20:03:54FromGitter<kaushalmodi> I've seen `.add` to be more canonical
20:08:26FromGitter<yyyc514> thanks
20:10:45*vlad1777d_ joined #nim
20:17:20*hoijui_ quit (Quit: Leaving)
20:22:51*narimiran quit (Ping timeout: 252 seconds)
20:26:08dom96stvnfx: Thanks! Glad you enjoyed it :)
20:26:24*Trustable quit (Remote host closed the connection)
20:26:33shashlickdom96: I think you missed my point on the nimble issue
20:26:41dom96Oh?
20:27:22shashlickWhat you said works for installing the files but I still have the nimgen dependency issue that's been there since the beginning
20:27:32shashlickAfter install solved both
20:31:43FromGitter<yyyc514> "mgetOrPut - either way returning a value which can be modified” am I misunderstanding this? modifying the returend value doesn’t seem to update the table
20:42:50FromGitter<xmonader> @kaushalmodi why r u doing libressl? didn't openssl already works fine for you?
20:43:20FromGitter<xmonader> for me now i get ⏎ ⏎ ```ldd /tmp/dmdm 18:58:09 ⏎ not a dynamic executable``` [https://gitter.im/nim-lang/Nim?at=5bbfb5e8c7bf7c3662f5a4c4]
20:43:37FromGitter<kaushalmodi> Doesn't work on Travis
20:43:51FromGitter<kaushalmodi> Neither does openssl work on Travis
20:43:53FromGitter<xmonader> it works on my archlinux
20:44:13FromGitter<kaushalmodi> Yeah some issue with Ubuntu on Travis
20:44:16FromGitter<xmonader> doesn't travis use a pretty old machines?
20:44:22FromGitter<xmonader> ah i see
20:44:30FromDiscord<An unconspicuous looking friend> Is type `None` a thing?
20:44:33FromGitter<xmonader> i would suggest u can boot docker in travis instead
20:45:02FromDiscord<An unconspicuous looking friend> because that's what I get when I do `type foo` if foo is an overloaded `proc`
20:45:02FromGitter<kaushalmodi> Can you provide a PR?
20:45:28FromGitter<xmonader> i can tinker around with it tonight yes
20:45:45FromGitter<xmonader> this would be amazing for people to have @kaushalmodi !
20:46:09FromGitter<kaushalmodi> See the latest Travis logs, I'm stuck with if_alg.h not found
20:46:29FromGitter<kaushalmodi> It was already a pain to get the version.h found
20:46:45*rnrwashere quit (Remote host closed the connection)
20:47:16dom96shashlick: yes, so if it solved both there is no other problem, right?
20:47:21dom96What am I missing?
20:49:05*rnrwashere joined #nim
20:51:59FromGitter<yyyc514> "mgetOrPut - either way returning a value which can be modified” am I misunderstanding this? modifying the returend value doesn’t seem to update the table
20:58:42FromGitter<dom96> Oooh https://twitter.com/travisci/status/1050475566770589697?s=19
21:01:27*craigger_ joined #nim
21:02:27*craigger quit (Ping timeout: 252 seconds)
21:03:12*SenasOzys__ joined #nim
21:03:42shashlickdom96: I thought you said `after install` was a workaround and it wasn't recommended
21:04:39*SenasOzys quit (Ping timeout: 252 seconds)
21:05:14*UxerUospr quit (Quit: Lost terminal)
21:09:43FromGitter<yyyc514> ok this syntax confuses me a little: `result.mgetOrPut(word, 0) += 1`
21:09:44FromGitter<yyyc514> but ti works
21:09:54FromGitter<yyyc514> is that the only way to get at the returned value’s reference?
21:10:10*darithorn joined #nim
21:11:18*zama quit (Ping timeout: 245 seconds)
21:12:16*zama joined #nim
21:15:22*flaviu quit (Remote host closed the connection)
21:17:19*flaviu joined #nim
21:20:49Calinouhttps://blog.travis-ci.com/2018-10-11-windows-early-release
21:24:01federico3oh that's good
21:24:30dom96shashlick: huh? What makes you think that?
21:26:33FromGitter<yyyc514> `v = addr result.mgetOrPut(word, 0)`
21:26:39FromGitter<yyyc514> just betting that isn’t idiomatic thought :)
21:30:53FromGitter<mratsim> use result\[word\].unsafeAddr
21:31:16FromGitter<yyyc514> or should i just write `result.mgetOrPut(word, 0) += 1`
21:31:30FromGitter<mratsim> if you want to increment in place yes
21:32:06FromGitter<yyyc514> does that work in C if you return a &ref ?
21:32:44FromGitter<yyyc514> just reads different than what i’m used to… where i’m from you’d pass an update proc or something to set/update the value :)
21:33:12FromGitter<mratsim> reusing your code you can do `v[] += 1` afterwards but that is definetely not idiomatic
21:33:29FromGitter<mratsim> `v[]` is the dereference operator
21:33:32FromGitter<yyyc514> yeah i figured that syntax out :)
21:34:09*Sembei joined #nim
21:34:10FromGitter<yyyc514> acutally i guess you can do that in C.. isn’t that how they write [] functions? return a ref to value?
21:34:25FromGitter<yyyc514> array[x] = 32; // array[x] being a &int
21:50:39*platoff quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
21:57:02*elrood quit (Quit: Leaving)
21:58:47*craigger joined #nim
22:00:54*craigger_ quit (Ping timeout: 252 seconds)
22:01:34*rnrwashere quit (Remote host closed the connection)
22:07:10*gmpreussner_ joined #nim
22:07:43*gmpreussner quit (Ping timeout: 252 seconds)
22:10:27*rnrwashere joined #nim
22:18:22*rnrwashere quit (Remote host closed the connection)
22:19:01FromGitter<zacharycarter> does anyone use kakoune with Nim?
22:26:43*rnrwashere joined #nim
22:30:57*craigger_ joined #nim
22:31:23FromGitter<yyyc514> do we have anything like ruby’s IOString?
22:31:55*craigger quit (Ping timeout: 252 seconds)
22:35:01FromGitter<citycide> in a macro, how would I constrain a typedesc parameter to one of 2 types, and then get the name of whichever type is given?
22:38:39FromGitter<mratsim> @citycide `macro foo(T: typedesc[int or uint]): untyped = echo $T; result = newStmtList()`
22:39:11FromGitter<mratsim> that’s assuming the macro is called directly with idents
22:39:36FromGitter<mratsim> if it’s called with foo(type x) you will have to use getType
22:39:58FromGitter<mratsim> @yyyc514 you would have to explain what Ruby’s IOString are
22:42:30FromGitter<citycide> I'm aware of `typetraits.name` but if I constrain like `typedesc[A or B]` I just get `A or B` when I want to know which of the two it was
22:42:37FromGitter<mratsim> going to sleep though
22:43:09FromGitter<mratsim> @citycide you should get the actual type with `getTypeInst` at the very least
22:43:25FromGitter<mratsim> it will give you the instantiated type
22:43:50FromGitter<mratsim> if it’s generic it will do it after generic symbol resolution
22:43:59FromGitter<citycide> @mratsim thanks, let me try it out
22:45:35FromGitter<mratsim> For example: https://github.com/status-im/nim-stint/blob/master/stint/private/datatypes.nim#L172-L175
22:45:54FromGitter<mratsim> the input is “typed” but it would work with generics as well
22:48:20*xet7 quit (Quit: Leaving)
22:51:07*krux02 quit (Remote host closed the connection)
22:53:16*rnrwashere quit (Remote host closed the connection)
22:53:24FromGitter<yyyc514> what’s the new syntax for `for x in 0 .. <i:`?
22:55:59FromGitter<kaushalmodi> `for x in 0 ..< i:`
22:56:09FromGitter<kaushalmodi> See the change in space
22:57:02FromGitter<citycide> @mratsim nice thanks, ended up being `getTypeInst(typ)[1].strVal.toLowerAscii`
22:57:34FromGitter<yyyc514> @kaushalmodi NICE thanks
22:57:42*rnrwashere joined #nim
22:57:45FromGitter<kaushalmodi> https://scripter.co/notes/nim/#space-around-dot-dot
22:57:47FromGitter<citycide> well, the `toLowerAscii` is specific to my use case
23:00:13*rnrwashere quit (Remote host closed the connection)
23:00:50*rnrwashere joined #nim
23:14:39*seni quit (Quit: Leaving)
23:28:37shashlickdom96: here you commented about the `after install` and `before uninstall` method
23:28:38shashlickhttps://github.com/nim-lang/nimble/issues/549#issuecomment-428757516
23:33:47*rnrwashere quit (Remote host closed the connection)
23:34:07*rnrwashere joined #nim
23:44:22*craigger joined #nim
23:45:24*craigger_ quit (Ping timeout: 252 seconds)
23:45:25*stefanos82 quit (Quit: Quitting for now...)
23:47:40*craigger_ joined #nim
23:48:55*craigger quit (Ping timeout: 252 seconds)