<< 02-01-2019 >>
00:12:44*rect0x51 quit (Quit: WeeChat 2.3)
00:17:18*Notkea quit (Read error: Connection reset by peer)
00:18:26*Notkea joined #nim
00:34:26*xace quit (Remote host closed the connection)
00:34:49*rockcavera joined #nim
00:36:37*xace joined #nim
00:39:49FromGitter<gogolxdong> How to get the value of HTML element with contenteditable attribution, checked some articles says innerText work which isn't included in Karax.
00:42:54*Ven`` joined #nim
00:44:28FromGitter<gogolxdong> tried to update the value of a table cell onblur, and got 'expected 1 real 2 TD nesting 4'
00:47:13*thomasross quit (Remote host closed the connection)
00:47:34*thomasross joined #nim
01:05:07FromGitter<zetashift> How do I match on a Nim variant object?
01:05:19*stefantalpalaru quit (Quit: stefantalpalaru)
01:36:29*d10n-work quit (Quit: Connection closed for inactivity)
01:46:22FromGitter<gogolxdong> ok, it has innerHTML
01:47:20*zachk quit (Quit: Leaving)
02:01:02FromDiscord_<ksen136> Sex Dating > http://discord.amazingsexdating.com
02:18:23*kungtotte quit (Ping timeout: 245 seconds)
02:30:32*Tyresc quit (Quit: WeeChat 2.4-dev)
02:34:32*Ven`` quit (Ping timeout: 250 seconds)
03:00:35*banc quit (Quit: Bye)
03:01:28*kapil____ joined #nim
03:16:13*banc joined #nim
03:30:05*dddddd quit (Remote host closed the connection)
04:04:55*Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
04:05:50*nsf joined #nim
04:54:32*Ven`` joined #nim
05:06:36*Ven`` quit (Ping timeout: 272 seconds)
05:15:08FromDiscord_<AregevDev> Ban
05:15:10FromDiscord_<AregevDev> Reeee
05:24:52*xet7 joined #nim
06:06:33*narimiran joined #nim
06:12:28FromGitter<gogolxdong> I have a table whose td value varies with different options, how to refresh the values onchange the option of select?
06:40:05*kapil____ quit (Quit: Connection closed for inactivity)
07:02:23FromGitter<gogolxdong> http://ix.io/1xjl
07:03:30FromGitter<gogolxdong> the value of kpi varies with selecting, but the table value doesn't change.
07:06:05Araqthat's not how 'sync' works
07:06:41*krux02 joined #nim
07:09:21Araqdon't use innerHtml
07:09:34Araqto attach data to a closure follow Karax's documentation
07:10:05Araqhttps://github.com/pragmagic/karax#attaching-data-to-an-event-handler
07:15:54shashlickAraq: good morning - i'm about to sign off, but windows builds should be working now - both 32-bit and 64-bit
07:16:06shashlickhttps://travis-ci.org/nim-lang/nightlies/builds/474294418
07:16:11Araq:O
07:16:29shashlicki'll deliver this PR and delete the releases posted so far through testing - https://github.com/nim-lang/nightlies/releases
07:16:42shashlickafter an hour or so, new binaries should get posted
07:16:54Araqok, great. Do I have to do anything?
07:17:08FromGitter<gogolxdong> without innerHTML , cannot get element value with contenteditable attribution
07:17:29shashlickwell, i'll just recommend checking all the posted artifacts to verify that they look okay - not sure how they should be tested
07:18:15Araqvia 'koch testinstall
07:18:25FromGitter<gogolxdong> like a cell of an editable table
07:18:25Araqbut that only works on Unix
07:18:57shashlicklet me ask something - is it sensible to generate osx and linux xz files when they all do the same thing?
07:20:07*ChristianWitts joined #nim
07:20:26shashlickalso, can you please disable appveyor on this repo so that successful PRs look green
07:24:24shashlickokay - i've merged my last PR and have cleaned up the releases
07:25:07shashlickonce this build is done, we should get everything posted neatly - https://travis-ci.org/nim-lang/nightlies/builds/474302732
07:26:31Araqwell osx could contain binaries too
07:26:51Araqand maybe also Linux can contain nim binaries
07:27:07shashlicki would like to see that eventually
07:27:23shashlickthen choosenim can install binaries like is possible on windows
07:28:11shashlickwith build stages, we can test the built packages in a separate job - https://docs.travis-ci.com/user/build-stages/
07:30:41shashlickdo you want nightlies to build automatically every night
07:30:47shashlickor do you want to trigger it
07:44:00*ripspin joined #nim
07:45:13ripspinWhere should the nim std libs be installed on linux?
07:48:03shashlickHow are you installing?
07:49:05FromGitter<gogolxdong> you can use innerHTML to get value of any editable element which supported in Html5, and attach data using closure.
07:50:32*ChristianWitts quit (Ping timeout: 244 seconds)
07:50:34ripspinI ran choosenim and it did it great, only thing is I think I should have been told to be in a particular dir before I ran it
07:52:28ripspinits now in my home dir ~/.nimble
07:52:41shashlickChoosenim installs in ~/.choosenim
07:52:52shashlickBinaries and the standard lib
07:53:29shashlickAdditional packages are installed using nimble to ~/.nimble
07:54:02ripspinI ran it in my home dir and it installed it in: ~/.nimble
07:54:24*kapil____ joined #nim
07:55:58ripspinIts not compiling something it should compile so I want to install it again so that it works . where should it be installed
07:57:42ripspinI also have ~/.choosenim
07:57:58shashlickYou need to add ~/.nimble/bin to your path
07:58:20shashlickEverything will start working then
08:00:51shashlickAraq: builds have started posting to the nightlies releases page
08:01:10shashlickFor whatever reason, they aren't being grouped into one release
08:01:30ripspinI have linked all of the binaries in ~/.nimble/bin into /usr/local/bin (PATH)
08:01:46shashlickGuess it depends on timing of 0.19.2 vs 0.19.9 builds completing
08:02:41ripspinIt runs file on other code it fails on a standard lib right at the start
08:07:15FromGitter<gogolxdong> made a workaround as follows :http://ix.io/1xjr
08:07:37FromGitter<gogolxdong> any better solution?
08:10:26*ChristianWitts joined #nim
08:10:28ripspinIt runs fine on other code it fails on a standard lib right at the start
08:15:28ripspinhttps://pastebin.com/n8jNjH3D
08:46:29*ripspin quit (Remote host closed the connection)
08:59:56*Vladar joined #nim
09:01:42avsejlooks like new release tarball is really different: 0.19.0 was just 3MB size, but 0.19.2 is 21MB!
09:02:08avsejalso it is not possible to run ./koch test with 0.19.2
09:02:33avsejinstall_nimble.nims and install_tools.nims missing
09:04:38avsejdom96, I remember size issue in the past, maybe some local scripts are outdated again? or dirty tree when doing release
09:06:50avsejdom96, Araq, 'proc tests' in koch.nim expects to compile 'testament/tester.nim' but it looks like it has been moved into 'tests/tester.nim' without updating koch
09:07:27*floppydh joined #nim
09:08:22avsejor even just the fact 'tests/testament' is missing in distribution tarball for some reason
09:12:34Araqavsej: we know
09:13:47Araqshashlick: building every night is nice
09:13:55Araqit's called 'nightlies'.
09:15:04FromGitter<gogolxdong> contentEditable of kdom.nim only works when its type is cstring instead of string, dir might also be related.
09:15:14Araqgogolxdong: innerHtml is always a big hack
09:20:38FromGitter<gogolxdong> yeah, I am aware of that, but have no better idea, I need to operate on a table cell which has no value attribution to set and get.
09:26:28*rect0x51 joined #nim
09:31:09FromGitter<gogolxdong> attribute
09:36:05*PMunch joined #nim
09:39:18*nc-x joined #nim
09:40:31nc-xAraq, @alehander42 I was wondering if something like https://github.com/nim-lang/Nim/pull/10127 would be possible for case objects so that the field not available error is given at compile time?
09:41:21rect0x51PMunch: Hi! I want to ask about nimlsp
09:42:04PMunchSure, ask away :)
09:42:20*FromGitter quit (Remote host closed the connection)
09:42:39*FromGitter joined #nim
09:42:49avsejAraq, thanks. But 0.19.2 was announced as release. Will there be 0.19.3 with the fix soon?
09:43:31Araqthe timing was excellent, just after 0.19.2 the automated builds started to work...
09:43:36rect0x51First of all... I failed to install it. It's probably some configuration since I got message of stuff being undefined. What is the state of the project? LSP is very important so I'd like to help.
09:44:09PMunchHow did you try to install it?
09:44:21PMunchAnd which editor do you intend to use it with?
09:45:44*theelous3 joined #nim
09:46:11rect0x51Ok first I tried "nimble install nimlsp" -> package not found, then I did "nimble install <link to github>" and it installed dependancies but not nimlsp. Then I clones the repo and used "nimble build" with various options to get it installed, like -d:nimcore and --threads:on, but I kept getting more errors of stuff being undefined...
09:46:18*nc-x quit (Quit: Page closed)
09:46:24Araqavsej: so yeah, we'll do something, stay tuned
09:46:42avsejthanks
09:46:51rect0x51(I intend to use vim)
09:46:59rect0x51(neovim in particular)
09:47:03*ChristianWitts quit (Ping timeout: 250 seconds)
09:47:32PMunchOh yeah, I forgot about that
09:47:43*vlad1777d joined #nim
09:48:18PMunchSo what you need to to us pull the repo and run "nimble debug" in the folder. That should create a debug build of the project which could then be placed in your path.
09:49:26synthmeat(lsp is a really helpful thing for language adoption. i usually end up turning it off later on, when i'm familiar with api surface/stdlib)
09:50:11PMunchIt's also really great for navigating a new project (with go to definition and such)
09:51:15synthmeatand refactoring, especially when you have ignored casing :)
09:52:25rect0x51Just to clarify... nimlsp is not ready for use yet, right?
09:52:26PMunchOh yeah, that's true as well
09:52:37PMunchrect0x51, well I use it
09:52:42PMunchWorks great
09:52:49rect0x51I mean... a casual user can't just do "nimble install" and use it yet
09:53:06*rockcavera quit (Remote host closed the connection)
09:53:10PMunchOh yeah, that's mostly because I still haven't gotten around to fixing the .nimble file
09:53:11Araqshashlick: I removed nightlies from appveyor
09:53:20rect0x51if it's already good, then should we supply it in a convinient way?
09:53:35PMunchDefinitely!
09:53:52PMunchI actually have the file needed to do it at home
09:54:04PMunchJust keep forgetting to add it to the repo..
09:54:37PMunchIt's essentially just the configs that are now passed over command line options in the .nimble file
10:00:35rect0x51Nice, "nimble debug" just works
10:01:02PMunchYeah, that's what numble build is supposed to do as well :P
10:02:06rect0x51So what is it supposted to do in vim? I added the viml lines but when I open a .nim file nothing happens
10:02:59PMunchIf you start typing something and hit <tab> it should give you a list of options
10:03:17PMunchOr you can use the commands that vim-lsp define like :LspDefinition
10:03:40PMunchgithub.com/prabirshrestha/vim-lsp#supported-commands
10:04:42rect0x51"Retrieving definition not supported for nim"
10:05:32PMunchThat means it is not set up correctly
10:05:41rect0x51yeah obviously
10:05:44PMunchDo you have nimlsp in your path?
10:05:58rect0x51I put it where nim binary is
10:07:24PMunchHmm, that should work. But you could try with a direct path like I do in my config: http://ix.io/1xjI/
10:08:25FromGitter<SolitudeSF> PMunch, `nimble debug` gives `nimlsppkg/suggestlib.nim(5, 8) Error: cannot export: Suggest` for me
10:08:42PMunchWhich Nim version?
10:09:09PMunchOh and you need to init the git submodule for it to work
10:09:24rect0x51SolitudeSF: make a folder in src/nimlsppkg called Nim, then copy nimsuggest folder from Nim/ in src/nimlsppkg/Nim
10:09:54FromGitter<SolitudeSF> i have submodule fetched
10:09:57PMunchOr do that, if you don't want a complete copy of Nim :P
10:10:05*kapil____ quit (Quit: Connection closed for inactivity)
10:10:17PMunchSolitudeSF, hmm that's strange
10:11:59FromGitter<SolitudeSF> i symlinked nimsuggest folder from my local nim install and it worked
10:12:27PMunchHmm, weird that the submodule didn't work
10:12:36PMunchMaybe it had pulled the wrong version?
10:14:06rect0x51Sadly it doesn't work even when I put the explicit path like you showed :(
10:14:10*lritter joined #nim
10:15:07PMunchHmm, strange
10:15:26PMunchIs there anything in the nimlsp log?
10:15:31PMunchOr the vim-lsp log?
10:15:59PMunchLog should be in /tmp/nimlsp by the way
10:16:29rect0x51it's empty
10:16:41rect0x51oh wait sorry
10:16:54rect0x51vim-lsp is not empty
10:17:22PMunchIf /tmp/nimlsp is empty it means that vim never actually started it
10:17:44rect0x51vim-lsp has logged the calls I made
10:18:10rect0x51give me a sec
10:18:49PMunchDo you have something like this in the vim-lsp file: Wed 02 Jan 2019 10:52:16 CET:["lsp#register_server","server registered","nimlsp"]
10:19:38rect0x51yes
10:19:57*ripspin joined #nim
10:20:26rect0x51"Unable to find \"config/nim.cfg\" in \"/home/stef/Desktop/nimlsp/src/nimlsppkg/Nim\". Supply the Nim project folder by adding it as an argument.", ""]]"
10:21:34rect0x51/home/stef/Desktop/nimlsp/src/nimlsppkg/Nim only has nimsuggest in it
10:21:41PMunchAh yes
10:22:01PMunchThat's because you only copied the nimsuggest folder instead of using the git module
10:22:16rect0x51ok, i'll do it properly, w8
10:22:53PMunchBasically nimsuggest needs to find a Nim project so that it knows about the stdlib and such
10:23:45PMunchBut if you have nim cloned somewhere you can add the argument for vim to launch it
10:23:59rect0x51I have it, how to do that?
10:23:59ripspinIt runs fine on other code it fails on a standard lib right at the start, https://pastebin.com/n8jNjH3D How do I start a project?
10:25:18PMunchrect0x51, you should be able to just add the path to the server_info array in your vim config
10:27:17*nc-x joined #nim
10:27:24nc-xripspin: Which nim version are you on?
10:27:52ripspin0.19.2
10:28:11nc-xIIRC this aporia issue is fixed on devel but the fix was not backported to 0.19.2
10:28:46nc-xhttps://github.com/nim-lang/Nim/issues/9889
10:30:06ripspinSo it was fixed in nim but not in aporia?
10:31:27nc-xNo. It should work fine if you compile latest nim from devel branch on github.
10:31:41nc-xBut the fix is not available on the newly released v0.19.2
10:31:43rect0x51PMunch: Thanks so much for the help, I got it running.
10:32:06PMunchGreat!
10:32:25PMunchShould make some kind of FAQ I think..
10:32:41PMunchBut first get it to work with just "nimble install nimlsp" :)
10:32:45rect0x51Yeah, was just thinking about that, I can write it for you
10:32:50ripspinOK, can I have the devel ver also ?
10:33:04rect0x51PMunch: Exactly
10:33:26nc-xripspin: if you use choosenim, you can install devel easily
10:33:44nc-xelse follow these steps https://github.com/nim-lang/Nim#compiling
10:33:49PMunchrect0x51, great :) PR it to the readme!
10:33:56ripspinWill it install over the stable ver?
10:34:11nc-xchoosenim allows you to switch between versions
10:34:17PMunchripspin, it installs side by side, but only one will be in your path
10:34:28rect0x51PMunch: Ok!
10:34:30nc-xinstalling manually, you can install it to any folder, just make sure the version you want is in the path
10:34:31PMunchSo you can quickly change between versions while having multiple ones installed :)
10:34:40ripspinOK i will give it a go
10:34:54*seni joined #nim
10:36:29ripspin"the latest current commit of Nim" would that do it?
10:36:49*seni left #nim (#nim)
10:37:43nc-x`choosenim devel` is the command I think (I don't use choosenim so not sure)
10:37:47krux02Happy new Year
10:38:22PMunchHappy new year!
10:38:42*nc-x quit (Quit: Page closed)
10:41:16rect0x51Happy new year
10:42:47PMunchGot an e-mail regarding my Nim lightning talk for FOSDEM by the way
10:42:50ripspinnc-x: choosenim devel looks good so far
10:42:52PMunchIt's accepted :)
10:43:00narimiranPMunch: congrats!!
10:43:12PMunchMetaprogramming with Nim | H.2215 (Ferrer) | 2019-02-03 | 11:20:00
10:43:13rect0x51ohhh!! nice!!
10:43:17*stefanos82 joined #nim
10:43:30rect0x51stefanos82: from Greece?
10:43:48stefanos82rect0x51: Cyprus
10:44:06rect0x51stefanos82: Nice to meet you, I'm from Greece and we have same name ;p
10:44:36stefanos82yes, I know. I saw your ascii video yesterday that demonstrated the '\ issue
10:44:45stefanos82nice to meet you too
10:46:59ripspinnow i get: https://pastebin.com/PZvA5pdh when i compile with devel (aporia)
10:47:18rect0x51PMunch: I am testing the vim commands now. LspDefinition seems to never return. I will do proper testing and report.
10:47:47PMunchI'm not sure why, but sometimes it takes quite a long while to find it
10:48:22rect0x51PMunch: pretty sure it's stucked, my test.nim is 5 lines and it's been running for minutes now...
10:48:27rect0x51stuck*
10:48:35PMunchOuch, yeah that's not normal
10:49:35*ChristianWitts joined #nim
10:54:35stefanos82one thing that i have noticed on my debian testing 64-bit system is that an aporia dependency is considered orphan package during a system apt upgrade
10:54:55stefanos82and when i remove it, aporia does not compile successfully
10:58:59*dom96_w joined #nim
11:02:01*theelous3 quit (Ping timeout: 250 seconds)
11:04:06rect0x51PMunch: It's weird. nimlsp.log shows that nimlsp successfully found the definition of the proc call. nimlsp.log ends with this line: Got valid Notification message of type textDocument/didOpen
11:06:04PMunchHmm, that means vim probably sent a didOpen notification for the file it was supposed to open for the definition
11:06:05PMunchStrange
11:06:16PMunchMight be a bug with vim-lsp
11:06:27PMunchCould you paste the nimlsp log somewhere?
11:06:37PMunchI'll look at it when I get back from lunch
11:07:15rect0x51should I just paste the link here? or open an issue?
11:09:38*ChristianWitts quit (Ping timeout: 245 seconds)
11:10:48rect0x51PMunch: I'll open an issue, bon appetit
11:16:40*noonien joined #nim
11:27:17*NimBot joined #nim
11:31:10*theelous3 joined #nim
11:31:37*theelous3_ joined #nim
11:34:33rect0x51PMunch: https://github.com/PMunch/nimlsp/issues/17
11:59:13PMunchrect0x51, replied
12:04:47FromGitter<kaushalmodi> shashlick: πŸ‘ Awesome work! Thanks for making nightlies pass!
12:04:51rect0x51PMunch: nimlsp keeps running yes
12:06:02FromGitter<kaushalmodi> shashlick: So you can now just delete the Appveyor config file from there? @Araq?
12:07:21PMunchrect0x51, so then it hangs for some reason..
12:09:47*vlad1777d quit (Ping timeout: 240 seconds)
12:10:35rect0x51yeah
12:10:38*narimiran quit (Remote host closed the connection)
12:11:07*narimiran joined #nim
12:11:08*narimiran quit (Remote host closed the connection)
12:15:27*nsf quit (Quit: WeeChat 2.3)
12:15:27*narimiran joined #nim
12:15:53*narimiran quit (Remote host closed the connection)
12:17:12*narimiran joined #nim
12:22:54*gangstacat quit (Quit: Ĝis!)
12:28:18*ChristianWitts joined #nim
12:29:09FromDiscord_<PusiteGA> ```
12:29:09FromDiscord_<PusiteGA> if paramCount()<2:
12:29:09FromDiscord_<PusiteGA> echo "Need 2 arguments, filename and size";
12:29:09FromDiscord_<PusiteGA> return;
12:29:09FromDiscord_<PusiteGA> ```
12:29:17FromDiscord_<PusiteGA> why it says return not allwed here
12:29:26FromDiscord_<PusiteGA> i just want to terminate my function
12:29:41FromDiscord_<PusiteGA> i just want to terminate my script
12:29:56PMunchYou're looking for `quit`
12:30:08FromDiscord_<PusiteGA> i just want to terminate my script, hmm meybe cuz i am not in fuction
12:30:19FromDiscord_<PusiteGA> ok will try that
12:30:20PMunch`return` is only for procedures and such
12:30:43FromDiscord_<PusiteGA> yep got it
12:30:59FromDiscord_<PusiteGA> i have do discard quit?
12:31:09PMunchNope
12:31:34rect0x51PMunch: I copied the whole Nim folder into nimlsp/src/nimlsppkg and recompiled with the "nimble debug" command. Then everything works fine! Sorry for all this trouble, but readme.md is seriously misleading...
12:31:38FromDiscord_<PusiteGA> quit(0)
12:31:47FromDiscord_<PusiteGA> works
12:32:02PMunchYeah, `quit 1` would give the program an exit code of 1
12:32:12FromDiscord_<PusiteGA> got it πŸ˜ƒ
12:32:32PMunchrect0x51, that's what the git submodule is for.
12:33:12rect0x51PMunch: yeah, it seems this is the only correct way to install nimlsp atm, so we should just fix readme.md atm to explain that.
12:48:14*dddddd joined #nim
13:03:47FromDiscord_<PusiteGA> do i have to create a openarray from paramStr to pass to startProcess or ist there some faster way?
13:05:17PMunchPusiteGA, you can get all of them with `commandLineParams`
13:05:24PMunchSo you don't have to create it yourself
13:06:09FromDiscord_<PusiteGA> whats posix, that mac or what?
13:06:12PMunchopenarray isn't a specific type by the way: https://nim-lang.org/docs/tut1.html#advanced-types-open-arrays
13:07:18PMunchWell, POSIX is a standard that many Linux and Linux-like systems adhere to
13:07:26PMunchMacOS being one of them
13:08:37*Snircle joined #nim
13:12:40*rockcavera joined #nim
13:13:58*kapil____ joined #nim
13:15:49*Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
13:15:50FromDiscord_<PusiteGA> so i am on linux and it says that commandLineParams mey not work on POSIX
13:16:12FromDiscord_<PusiteGA> whatewer will just forloop and create an seq/array
13:16:25*Snircle joined #nim
13:18:36rect0x51PMunch: ... it wasn't an installation problem after all. nimlsp indeed hangs. Can you check your setup? Does vim work properly with nimlsp?
13:18:45*salewski joined #nim
13:19:06PMunchPusiteGA, it's only if you try to use it from a DLL that it won't work
13:19:32PMunchrect0x51, that is strange
13:19:44PMunchI'm using it every day and it works fine for me
13:20:03rect0x51what version of vim do you have?
13:20:07PMunchIt does sometimes do some strange stuff, but I haven't had it just not work like you're describing
13:20:15PMunch8.1
13:20:31rect0x51could it be a problem with neovim?
13:20:37PMunchCould be..
13:20:51PMunchDo you have another editor that supports LSP which you can try it on?
13:21:13rect0x51I can download one, maybe VSCode?
13:21:20rect0x51which ones do you want me to test?
13:21:32PMunchWell I've only tried vim and sublime
13:21:45rect0x51ok I will try normal vim (not neovim) and sublime
13:22:50*nsf joined #nim
13:22:53salewskiDoes the search entry field in https://nim-lang.org/docs/lib.html work for you?
13:23:14salewskiNot for me currently, may it be an issue with 0.19.2 docs?
13:23:27PMunchWorks for me
13:23:31FromDiscord_<PusiteGA> ```
13:23:31FromDiscord_<PusiteGA> Error: unhandled exception: ; Additional info: "Could not find command: \'usr/bin/ffprobe\'. OS error: No such file or directory" [OSError]
13:23:31FromDiscord_<PusiteGA> ```
13:24:02FromDiscord_<PusiteGA> why is it "'usr/bin/ffprobe\' i newer added \ on end?
13:24:19*Vladar quit (Remote host closed the connection)
13:24:21FromDiscord_<PusiteGA> let pr = startProcess("ffprobe","",arguments ,nil,{poStdErrToStdOut});
13:24:42FromDiscord_<PusiteGA> let pr = startProcess("usr/bin/ffprobe","",arguments ,nil,{poStdErrToStdOut});
13:24:48*Vladar joined #nim
13:24:53salewskiWell maybe it is my Firefox, will try another browser. Bye.
13:24:57*salewski quit (Quit: WeeChat 2.3)
13:25:11leorizePusiteGA: you should add poUsePath to options
13:25:18leorizeit's easier that way
13:25:36leorizethe `\` is because `'` need to be escaped, that's all
13:27:28FromDiscord_<PusiteGA> ok anywya if frogot / ifont usr , but ye with poUsePaths ffprobe works np, dident complain about \
13:29:13rect0x51PMunch: I feel stupid that I wasted your time now... I used a clean .vimrc and it works. Must be some kind of conflict with the other vim configurations of mine.
13:30:29PMunchHaha, no problem :P
13:30:42FromDiscord_<PusiteGA> accepting you stupit is a first step forward elightment xD
13:30:52PMunchNext time some has this issue I will be ready to tell them what to do :)
13:34:12*gangstacat joined #nim
13:42:30rect0x51PMunch: 'zah/nim.vim' is a dependancy
13:43:01rect0x51?
13:43:20PMunchUhm, not 100% sure tbh
13:43:44PMunchMight be just to get syntax highlighting, but I think that is what makes vim realise that a file is Nim or not
13:43:58PMunchI know Sublime requires the NimLime plugin
13:44:26rect0x51I see, yeah, without it nimlsp doesn't start.
13:45:33rect0x51Do you use zah/nim.vim?
13:45:54PMunchYes
13:46:35rect0x51I am not sure what is happening anymore... I'll post more info at the issue.
13:47:44FromGitter<zacharycarter> https://github.com/google/flutter-desktop-embedding
13:47:54FromGitter<zacharycarter> I wonder if flutter will become a viable alternative to electron
13:48:08FromGitter<zacharycarter> or SDL2 / GLFW3
13:52:13rect0x51PMunch: I wrongly thought for a moment that it worked. https://pastebin.com/R3Qjx31j with Vim 8.1 still same behavior.
13:55:33rect0x51Must be the way I installed nimlsp after all. I will try the git module method as you suggested at the issue
13:56:42FromDiscord_<PusiteGA> so how how pull only part of stream i am interested, in my case
13:56:47FromDiscord_<PusiteGA> ```
13:56:47FromDiscord_<PusiteGA> compatible_brands: qt
13:56:47FromDiscord_<PusiteGA> encoder : Lavf58.20.100
13:56:48FromDiscord_<PusiteGA> Duration: 00:02:00.02, start: 0.000000, bitrate: 371 kb/s
13:56:48FromDiscord_<PusiteGA> Stream #0:0(eng): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1880x1008 [SAR 1:1 DAR 235:126], 355 kb/s, 30 fps, 30 tbr, 15360 tbn, 60 tbc (default)
13:56:49FromDiscord_<PusiteGA> Metadata:
13:56:51FromDiscord_<PusiteGA> ```
13:57:16narimiranuse some pastebin-like site for code-pasting, please and thank you
13:57:17FromDiscord_<PusiteGA> i just want to pull out value of Duration:
13:57:40FromDiscord_<PusiteGA> oh sorry i frogot this irc too
13:57:57FromDiscord_<PusiteGA> anyway how to rape from text what i need only
13:57:59FromDiscord_<PusiteGA> Duration: 00:02:00.02, start: 0.000000, bitrate: 371 kb/s
13:58:11FromDiscord_<PusiteGA> so i want to get value of duration only
13:58:37FromDiscord_<PusiteGA> and i cant be sure it will write it on same line , or at least i duno if it does πŸ˜ƒ
14:00:46FromDiscord_<PusiteGA> i can meybe go read all lines unitl i found 1 line with duration then just take this time part with substring, or is there a better way?
14:02:42FromDiscord_<PusiteGA> or be smarter and just add more parementars to get exact data πŸ˜ƒ
14:02:53FromDiscord_<PusiteGA> or be smarter and just add more parementars to get exact data πŸ˜ƒ from ffprobe
14:03:58PMunchIf possible I'd go with options to only get the single value
14:04:18PMunchOr get output in some format that is intended to be machine-readable
14:04:53PMunchBut you can of course just read line by line until the line starts with "Duration: " and then grab the the timestamp
14:04:54FromDiscord_<PusiteGA> i found it allreay i get exactly what i want just output in seconds
14:05:05PMunchProbably things in strutils that can help you out there
14:05:57*whaletechno quit (Quit: ha det bra)
14:10:38PMunchIf you have it in seconds you can use initDuration and toParts from times to create a timestamp similar to the one it outputs
14:15:20rect0x51I am at a loss :( "git submodule update --recursive --remote" doesn't do anything when run
14:17:27PMunchYou need to run git submodule init first
14:17:39PMunchgit-scm.com/book/en/v2/Git-Tools-Submodules
14:17:46PMunchCloning a Project with Submodules
14:19:01FromDiscord_<PusiteGA> ty pmuch but i need it in seconds anyway πŸ˜ƒ
14:21:23rect0x51PMunch: Now I get what the other guy said some hours before, "Error: cannot export: Suggest"
14:21:54PMunchHold on, I'll try to do the whole process myself
14:22:02rect0x51thanks
14:35:41PMunchAha
14:35:59PMunchThe submodule must be set up to track the "devel" branch
14:38:11FromDiscord_<PusiteGA> strutils.nim parseFloat
14:38:11FromDiscord_<PusiteGA> Error: unhandled exception: invalid float: 120.022000
14:38:39PMunchrect0x51, try to pull nimlsp again and reinitialize the git submodule
14:39:00FromDiscord_<PusiteGA> hmm shoudent this be parsable, or meybe i have some whitespaces ?
14:39:07PMunchPusiteGA, make sure there is no whitespace
14:40:00*Marmumoto joined #nim
14:40:29FromDiscord_<PusiteGA> yep my bad
14:45:27rect0x51PMunch: ok it builds now, but what I described at the issue remains same :|
14:47:13PMunchHold on, I'll try to let vim use my newly built version
14:52:46*rockcavera quit (Ping timeout: 268 seconds)
14:54:49*Ven`` joined #nim
14:56:32*narimiran quit (Remote host closed the connection)
15:08:37shashlick@kaushalmodi - please try out koch testinstall when you get a chance
15:09:01shashlickWe also need to figure out how to group the release files together on http://github.com
15:09:46shashlickFor some reason, 0.19.2 and 0.19.9 get grouped together even though they should have different titles
15:09:57Araqshashlick, "github Page build failure" but I don't know what it means
15:10:17shashlickThe appveyor file can remain if it has anything useful
15:10:31shashlickWhere do you see that Araq
15:10:48Araqgot an email from github
15:10:59AraqI'm sure it has something to do with the release
15:11:14Araqbut that's all I know
15:12:03*narimiran joined #nim
15:13:38FromGitter<kaushalmodi> shashlick: I am not sure what you mean by grouping
15:13:44*ChristianWitts quit (Ping timeout: 250 seconds)
15:13:46FromGitter<kaushalmodi> the asset file name is derived here: https://github.com/nim-lang/nightlies/blob/12b31a73924f56e967d710533b1d2928e2f24e51/.travis.yml#L162
15:14:21FromGitter<kaushalmodi> it should be creating a separate archive file for each DEPLOY_VERSION
15:15:00FromGitter<kaushalmodi> I see that you add ARCH var too
15:15:24FromGitter<kaushalmodi> so looks like ARCH needs to be made a part of ASSETFILE too
15:16:32shashlickArch is part of suffix
15:16:49shashlickAnd only relevant to windows
15:17:08FromGitter<kaushalmodi> ah, ok. I see that now
15:17:24shashlickI'm talking about the releases page in http://github.com how it shows multiple releases
15:17:51shashlickhttps://github.com/nim-lang/nightlies/releases/tag/untagged-16d01ddd109555351010 has a mix
15:18:06*abbiya joined #nim
15:18:22FromGitter<kaushalmodi> I don't know how to separate that
15:18:39FromGitter<kaushalmodi> if you know a project that does that, then may be we just need to look at that project's travis config
15:19:18shashlickBut this and three others are isolated
15:19:22shashlickhttps://github.com/nim-lang/nightlies/releases/tag/untagged-99a315c692ccde8f0a8d
15:19:29shashlickJust weird
15:20:02shashlickBut more importantly, all these builds need to be verified
15:20:17FromGitter<kaushalmodi> shashlick: I think I know to separate
15:20:18FromGitter<kaushalmodi> see https://github.com/nim-lang/nightlies/blob/master/.travis.yml#L137
15:20:22shashlickThen we can fix the cosmetics
15:20:30FromGitter<kaushalmodi> NIMVERSHORT needs to be recalculated in the matrix too I think
15:20:43*ChristianWitts joined #nim
15:20:55FromGitter<kaushalmodi> then we should have unique titles for different versions from https://github.com/nim-lang/nightlies/blob/12b31a73924f56e967d710533b1d2928e2f24e51/.travis.yml#L176
15:21:14FromGitter<kaushalmodi> > But more importantly, all these builds need to be verified ⏎ ⏎ yes, `koch testinstall`
15:22:18shashlickWell I'd rather see two releases, one for devel with 4 assets and a second for stable with 4 assets
15:22:53shashlickAnd this method should work since I deleted all cache and kicked off this build
15:22:55FromGitter<kaushalmodi> shashlick: yeah.. do you want to open an issue for this? This is cosmetic think as you said
15:23:00FromGitter<kaushalmodi> let's get the testinstall to pass
15:23:17FromGitter<kaushalmodi> about about question about why I cache ..
15:23:24shashlickI think we need to use build stages
15:23:39shashlickAnd have all builds pass first
15:23:42FromGitter<kaushalmodi> it just optimization, we do not need to rebuild the deploy assets if the git hash hasn't changed overnight
15:23:57FromGitter<kaushalmodi> that's mostly true for the stable deploys
15:24:05PMunchrect0x51, seems to work fine for me
15:24:08shashlickThen do a test install before deploy
15:24:25FromGitter<kaushalmodi> shashlick: that's the plan
15:24:27shashlickI agree
15:24:32FromGitter<kaushalmodi> I did that but then removed it because it always failed
15:24:52shashlickHave you done build stages before in Travis?
15:25:15FromGitter<kaushalmodi> probably not, because I don't know what that means :)
15:25:23shashlickhttps://docs.travis-ci.com/user/build-stages/
15:25:40shashlickProbably not as trivial since we have a matrix as well
15:25:58FromGitter<kaushalmodi> hello_musl and nightlies are the most complex travis config I have dealt with
15:25:59rect0x51PMunch: Very strange. I will investigate more. Thanks for testing.
15:26:07shashlickAnyway I'll open an issue for the releases page and look into that
15:26:19shashlickPlease proceed as originally planned
15:26:33shashlickI need to fix my choosenim pr as well
15:34:27*PMunch quit (Remote host closed the connection)
15:37:17*ChristianWitts quit (Remote host closed the connection)
15:39:10*Ven` joined #nim
15:40:44*Ven` quit (Client Quit)
15:41:16*Sembei joined #nim
15:41:25*Trustable joined #nim
15:42:19*Marmumoto quit (Quit: Page closed)
15:42:46*Pisuke quit (Ping timeout: 250 seconds)
15:43:02*abm joined #nim
15:44:58*jxy quit (Quit: leaving)
15:46:29*rect0x51 quit (Quit: WeeChat 2.3)
15:56:42narimiranhttps://nim-lang.org/docs/lib.html β€” does the search work for you guys?
15:58:22shashlickyep just tested on chrome - searched split and got popup, clicked link and it went to right location
15:59:01narimiranhmmm, nothing happens on vivaldi (chrom(ium) based)
15:59:01*abbiya left #nim (#nim)
15:59:12shashlickcheck the console
16:00:11narimiran"lib.html:1227 Uncaught ReferenceError: search is not defined at HTMLInputElement.onkeyup (lib.html:1227)"
16:15:22narimiranhmmm, for me it doesn't work on firefox either
16:16:18FromGitter<kaushalmodi> narimiran: Hmm, it works for me on firefox
16:16:34FromGitter<kaushalmodi> Firefox 64.0 Win 10 64-bit
16:17:19FromGitter<kaushalmodi> narimiran: Does it not work for *any* search string? E.g. I typed "hash" and it showed the results
16:17:45narimiranwhatever i type, nothing happens, not even after i press enter
16:18:12FromGitter<kaushalmodi> narimiran: yeah, that search doesn't need enter to be pressed
16:18:34narimirani know, and it worked until the end of 2018 quite alright :)
16:18:40FromGitter<kaushalmodi> may be try clearing your cache
16:18:54FromGitter<kaushalmodi> can you Ctrl+F5 the page and try again (I did that, and it still works)
16:19:06shashlick@kaushalmodi, @Araq: https://github.com/nim-lang/nightlies/issues/14
16:20:50narimiran@kaushalmodi still the same, in both browsers
16:21:32narimiranhttp://nim-lang.github.io/Nim/lib.html works fine!
16:34:12*vlad1777d joined #nim
16:37:38*quipa joined #nim
16:45:30Araqhttps://forum.nim-lang.org/t/4519 can anybody reproduce this?
16:46:43FromGitter<kaushalmodi> Araq: No, it works great on Firefox (Win 10, Android)
16:47:48narimirancannot reproduce, but ever since the new design, nim forum is the only site which won't load in the background when i e.g. open several topics in new tabs
16:47:55shashlickthanks for the info @kaushalmodi
16:48:21narimirandon't know if that's somehow connected with what oyster describes
16:48:50FromGitter<kaushalmodi> shashlick: You are welcome. You can see that I am really inexperienced with Travis :P
16:48:59FromGitter<kaushalmodi> I got stuff working after a lot of hit and miss
16:49:56shashlicktrust me I'm bumbling through it as well
16:50:07shashlicki don't see you pushing the tag once you create it
16:50:11shashlickin your history
16:50:20shashlickwere tags created for releases ever?
16:52:03FromGitter<kaushalmodi> shasklick: yes, I then did a cleanup of all that mess
16:52:21shashlickI don't see a git push --tags in your history
16:52:21Araqsorry about the mess
16:52:24FromGitter<kaushalmodi> I never got the git tagging to happen *only when* the hash changed
16:52:35FromGitter<kaushalmodi> Araq: it wasn't you :)
16:52:48FromGitter<kaushalmodi> we are talking about the mess I created when experimenting with git tagging the releases
16:52:59FromGitter<kaushalmodi> I cleared the mess before anyone can see it
16:53:08*floppydh quit (Quit: WeeChat 2.3)
16:53:21shashlickwell, so unless you pushed the tag, github would still simply create untagged- releases
16:53:29shashlicksince the tag only exists on the local copy
16:53:51FromGitter<kaushalmodi> shashlick: I might be misremembering but the pushes did happen
16:54:00shashlickalso you need to set TRAVIS_TAG or it doesn't know how to match
16:54:28FromGitter<kaushalmodi> I think the issue was when travis tried to apply the same git tag multiple times (because nim devel didn't have any new commit)
16:55:07shashlickthen that shouldn't post anything - you already block that right?
16:55:10FromGitter<kaushalmodi> shashlick: It could be a compound issue; not sure
16:55:14FromGitter<kaushalmodi> I took the easy way out
16:55:32shashlickok let me experiment with my apparent confidence πŸ˜‰
16:55:38shashlicklooks like you added most of the code already
16:55:45FromDiscord_<PusiteGA> how to rape extension from file, is there some secure way or should i just read <<< until i hit . and remove it?
16:55:46shashlickhow long does testinstall take
16:56:07FromGitter<kaushalmodi> it takes quite a bit, an hour or more
16:56:19FromGitter<kaushalmodi> but probably that was before the tester optimizations
16:56:48shashlickokay then we might have to do it in a separate step
16:56:53FromGitter<kaushalmodi> well, it couldn't be hour or more .. because Travis doesn't allow running 1+ hr jobs
16:56:59shashlicksince linux already takes 20+ minutes to package
16:57:02FromGitter<arnetheduck> @Araq re RVO / placing `result` as a parameter instead of return value (in the generated code), did you ever consider doing it as a transformation step instead doing it in cgen?
16:57:43FromGitter<kaushalmodi> shashlick: I am running testinstall in my fork's master: https://travis-ci.org/kaushalmodi/nightlies/builds/474484665
16:57:46FromGitter<kaushalmodi> let's see how it goes
16:59:46*nc-x joined #nim
17:00:26shashlickcool
17:01:09shashlicki presume this commit worked for tagging - https://github.com/kaushalmodi/nightlies/commit/08bd48adaa64edec008c7c6c71ef497b26dec4d6
17:01:27nc-xshashlick, kaushalmodi: well you can use azure ci, it has a configurable 180 minutes timeout iirc
17:01:36shashlickcan you let me know what you used for GIT_TAG_USER_NAME and GIT_TAG_USER_EMAIL
17:01:38nc-xbut that would require to redo the work
17:01:44shashlickor set that in travis please
17:01:49FromGitter<kaushalmodi> shashlick: it can be anything you want
17:02:14FromGitter<kaushalmodi> set it in *your* travis config, commit to your fork master and see the results in your Travis
17:02:20*nc-x quit (Client Quit)
17:02:35FromDiscord_<PusiteGA> how to rape extension from file, is there some secure way or should i just read <<< until i hit . and remove it? or is there some file info or some class like that?
17:02:39FromGitter<kaushalmodi> basically git needs username and email to associate with the git tagging
17:03:22FromGitter<kaushalmodi> nc-x: I am cringing at the idea of reworking this :P
17:03:47narimiran@PusiteGA https://nim-lang.github.io/Nim/os.html#splitFile%2Cstring ?
17:04:09FromGitter<kaushalmodi> shashlick: ⏎ ⏎ > i presume this commit worked for tagging ⏎ ⏎ No, that commit is where I get tired of the git tagging idea and get rid of it [https://gitter.im/nim-lang/Nim?at=5c2cef09d945b96882343a34]
17:04:27shashlickwell, i mean did it create the tag?
17:04:33shashlickand push it correctly
17:04:42shashlickforget the other issues
17:05:03FromDiscord_<PusiteGA> ty
17:05:07FromGitter<kaushalmodi> shashlick: I believe it did push but git puked on trying to set the same tag twice or something
17:05:18shashlickbut you have a check for it
17:05:19FromGitter<kaushalmodi> sorry, don't remember much from that experiment
17:05:22Araqarnetheduck: Yes, I dream of an AST->AST codegen, remember?
17:05:27Calinouhey, I wonder if there's a way to specify compile flags when installing a package using nimble
17:05:31Araqalso I looked at your travis.yml
17:05:33FromGitter<kaushalmodi> shashlick: yeah, but something still didn't work
17:05:34Calinou(such as --passc:-flto)
17:06:19Araqarnetheduck: how about this. A nim-lang/releases repo that simply lists the commit hashes of the releases
17:06:31shashlickokay cool
17:06:53Araqrelease building updates it automatically, no need to unzip stuff
17:07:01FromGitter<arnetheduck> yeah, ast-ast is one thing - but barring that, I'm curious why the return-value-to-arg transformation is done in cgen - if there was a compelling reason to put it there? basically, it means I must redo the same logic, as well as the "isresultassigned" logic in `nlvm`
17:07:20Araqand no freaking commits in koch.nim either.
17:07:46FromGitter<kaushalmodi> shashlick: about testinstall, I am not developing on Windows env, but I doubt if this same would work on windows too: https://github.com/kaushalmodi/nightlies/commit/6a04cfd550b52e7abb62641d2ee7d1da93719fae#diff-354f30a63fb0907d4ad57269548329e3R125
17:07:58FromGitter<kaushalmodi> because I see that you are not doing the koch xz step for windows
17:08:22Araqcompelling reason? not really. It was a C specific problem though.
17:08:33dom96_wa repo that contains commit hashes?
17:08:35dom96_wwhat?
17:08:51dom96_wThis is why you have git tags
17:10:18*Ven`` quit (Ping timeout: 250 seconds)
17:12:21shashlick@kaushalmodi: ya, Araq said testinstall is for linux/osx only
17:12:39shashlickwe may have to create a separate test for the windows build
17:12:47shashlickwhere no build happens again
17:14:52FromGitter<arnetheduck> separate repo sounds.. tricky. It's better than nothing at all, but it feels similarly loose to git tags.. for example, if I'm following the 0.19 branch of nim, how do I know which nimble to pick with that?
17:19:49FromGitter<arnetheduck> at that point, it might be that it's just easier for us to fork nimble and track it the same as nim.. then we can revisit it when the new nim release process with point releases and all has settled down a bit, to see if there's a workflow that's convenient for everyone.. it's really not something worth complicating nim development over
17:21:24FromGitter<arnetheduck> will you take a patch that allows the specification of a hash to `build_all.sh` and `koch tools` (as command line param)?
17:25:27shashlick@arnetheduck: this is why I want nimble to evolve separately from the compiler - it should support multiple versions of compilers and update like pip and co
17:27:07FromGitter<arnetheduck> @shashlick well yeah, that's nice, but my goal is to find *one* version that *just works* :) (and be able to reproduce that 6 months later, should I need to)
17:28:56shashlickright and that can happen if nimble released independenty
17:29:11shashlickyou can download an older version and stay with if you preferred
17:30:07FromGitter<arnetheduck> nimble is merely the means - we have no real interest in nimble except as a tool to solve the actual problem we're solving.. nimble however is kind of crucial to the nim "experience" overall - and given it's tight integration with the language and compiler (library lookup paths etc), it's important that we find versions that work well together
17:31:17FromGitter<arnetheduck> since upstream is selecting such matching versions on a regular basis, we'd love to just follow that and get back to doing our stuff
17:33:18shashlickthat's why I did a POC to evaluate if you can run nimble files without bundling in the entire compiler
17:33:40shashlickit worked well enough so it is possible to create a nimble that simply runs nim e as a subprocess
17:35:26FromGitter<arnetheduck> haha, well I'd actually prefer it if nimble files went back to being declarative so they wouldn't require a full language at all - then it would be easier to integrate them in other tooling as well that could do smart stuff with them, and they'd have fewer dependencies and moving parts generally
17:35:30*nc-x joined #nim
17:36:34shashlickya well, I think it might be hard going that far back, but removing the internal compiler dependency is possible
17:36:36FromGitter<arnetheduck> in addition to that, they could/should have extension points for those custom scenarios, but have the 95% cases covered by declarative
17:36:49nc-xIMO nimble should be merged with nim. That way we also have one less configuration language. And whichever nim version you build you will have a working package/build manager.
17:36:54shashlickof course, if you use nimscript that an older compiler doesn't support in the VM, there might be issues
17:37:43shashlicknc-x: i'm proposing the opposite, a package manager should move faster than a compiler
17:37:44FromGitter<arnetheduck> for example, for wrappers that would be awesome - you'd have a custom nim script that builds the underlying dependency - but the rest of the nimble file would be discoverable and predictable
17:38:32shashlicki would think that's possible with nimble today - that's what i'm kind of doing with nimgen wrappers
17:38:51FromGitter<arnetheduck> the custom part would merely inform the declarative part of the outcome (which libs to link etc), but the whole experience overall would be much less messy
17:39:22nc-xand no i don't mean adding nimble repo to nim repo. I mean deprecating nimble and adding its features to nim itself. So nimscript file would contain the dependencies as well and `nim build` would do the dependency downloading/checking etc
17:39:46shashlickok so that's more step two, are there any concerns with pulling out the compiler from nimble and just using an installed nim?
17:40:03nc-xthe 3 configuration languages for nim is IMO one of the worst things about nim
17:40:43shashlickwell, so nimble is just nimscript with some macros and in my POC, i could treat nimble files as a nimscript file that `nim e` could run just fine
17:41:22shashlickit would be cool if nim e allowed specifying an entry point but even that can be managed somewhat
17:42:14*nc-x quit (Quit: Page closed)
17:43:08*dom96_w quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
17:43:48Araqarnetheduck: alternative we can make the tag commit messages contain the used C sources and Nimble tags
17:44:33shashlickwhy does a package manager version become important for reproducibility
17:45:04Araqshashlick, I don't know tbh. that travis.yml file from Status wasn't enlightening
17:45:09FromGitter<arnetheduck> how's that different from a submodule update (except needing a custom parser)
17:45:34AraqI had really bad experience with git submodules in the past
17:45:36FromGitter<arnetheduck> well, it's what builds the project - 3 weeks ago, someone updated nimble and our builds broke as a result
17:46:03shashlicki'm not aware of this travis file
17:46:17Araqand even assuming these issues have been fixed by now it remains true that I don't understand git submodules well enough. And I don't use what I don't understand.
17:46:38FromGitter<arnetheduck> yeah, but think about it - a submodule is really nothing but a text file with a hash in it.. putting it in the commit message is.. very similar
17:46:55Araqit doesn't show up in 'git diff' though
17:46:59Araqit remains meta-data
17:47:06Araqthat's why it's good. ymmv.
17:47:19FromGitter<arnetheduck> it does show up in git diff - at least when I used it in the past
17:47:20FromDiscord_<PusiteGA> hmm how do i close process if i do an 1 liner
17:47:22FromDiscord_<PusiteGA> let wholeVideoDuration : float = startProcess("ffprobe","",getLengOfVideoArgs,nil,{poStdErrToStdOut,poUsePath}).outputStream().readAll().strip().parseFloat();
17:47:34*Tyresc joined #nim
17:48:37FromGitter<arnetheduck> ```code paste, see link``` ⏎ ⏎ this is nlvm updating the nim version it's tied to [https://gitter.im/nim-lang/Nim?at=5c2cf97593cce97d3bd7d66a]
17:49:20FromGitter<arnetheduck> oh. you want it to not show up?
17:49:52shashlicki guess I don't understand the issue - i feel a package manager, like choosenim should evolve in parallel and that ongoing improvement should help iron out compatibility issues
17:49:55*druonysus joined #nim
17:49:55*druonysus quit (Changing host)
17:49:55*druonysus joined #nim
17:50:04shashlicknimble should have an expanded test framework testing against older versions of nim
17:50:36AraqI would merge Nimble into Nim's repo but that's ultimately dom96's decision
17:50:37shashlickbut this is far easier if nimble doesn't package in nim altogether since it only carries the compiler, not the stdlibs
17:51:03shashlicki disagree with that - it still ties nimble heavily with the compiler - no good reason for that
17:51:05FromGitter<arnetheduck> shashlick, that's fine by me as well, but it should be reflected in the nim release process then
17:51:26Araqalso, side point but 'nimble dump' or similar exists and outputs the crucial information for others
17:51:45shashlickagreed - nim can provide a nimble binary on ship but nimble update nimble should be used to help keep things moving
17:52:14Araqarnetheduck: I want it to *not* show up, yes.
17:52:55Araqand my suggestion makes sense, it's a git tag, it might as well contain useful information
17:54:58FromGitter<arnetheduck> I don't think I've ever used or seen an annotaded git tag - it's certainly novel :)
17:56:40*theelous3_ quit (Ping timeout: 250 seconds)
17:56:41shashlicklooks like we already have a way to update nimble - `koch nimble` does what we need
17:56:51Araqbut I got the basic problem right, right? All you want is the (nim commit, nimble commit, csources commit) triple?
17:56:57shashlickonly thing is that koch is not in the bin directory so it isn't accessible - say if nim is installed with choosenim
17:58:18FromGitter<arnetheduck> hm, more accurately: given a nim commit, I want nimble+csources commits that work with it
17:58:36FromGitter<kaushalmodi> shashlick, Araq: Hmm Travis times out after just 50 mins, trying to finish running `koch testinstall`: https://github.com/nim-lang/nightlies/issues/4#issuecomment-450934475
17:58:42*vivus joined #nim
17:59:15shashlickI don't think testinstall can be done in the same job - it needs to be run in a separate step after releases are created but before releases are posted.
17:59:19Araqkaushalmodi: 'koch testinstall' can be trimmed down to only test e.g. 'megatest'
17:59:21FromGitter<arnetheduck> shashlick, it doestn't do what I need it to do: if I checkout nim 0.17 and run `koch nimble`, it will give a random version that does not work with nim 0.17
18:00:09shashlickwhy wouldn't you use the xz files? that has an exact snapshot
18:00:32Araqhe he, shashlick is following my footsteps
18:00:35FromGitter<arnetheduck> doh.
18:01:06shashlick@kaushalmodi, @Araq - looks like tags are working properly now - see my releases page - https://github.com/genotrance/nightlies/releases
18:01:31Araqshashlick, awesome :-)
18:01:44FromGitter<kaushalmodi> shashlick: awesome!
18:01:53FromGitter<arnetheduck> circles, circles.. because we occasionally patch nim and want to use git tools exclusively to help in that process - for example to make it easier to submit patches to upstream
18:01:57FromGitter<kaushalmodi> minor nitpick.. may be title the stable releases without hash?
18:02:29narimiranshashlick: very nice!
18:02:32FromGitter<kaushalmodi> "vx.y.z-hash nightly build" and "vx.y.z stable release"?
18:02:40Araqis it the stable release though? or the version-0-19 branch?
18:02:50Araqthese are not the same thing...
18:02:51FromGitter<arnetheduck> also, because git provides a secure hash as part of its tooling. with tarballs, we have to recreate that experience doing manual hashing and stuff
18:02:56shashlickstable is still a snapshot, assuming version-0-19 continues to evolve
18:03:04FromGitter<kaushalmodi> Araq: good point
18:03:11FromGitter<kaushalmodi> yeah, then the hash makes sense
18:03:23shashlickok so osx builds are still not up yet
18:03:44shashlickany way to speed up winrelease? it does koch boot twice as discussed earlier
18:03:56FromGitter<kaushalmodi> shashlick: yeah, osx lags behind by 10-15 mins usually
18:04:06shashlickwell, that's because there's 8 jobs
18:04:16shashlickand windows takes 35 minutes whereas linux/osx take 18 mins
18:04:17FromGitter<arnetheduck> if we were not contributing to nim or if it was less important to us, we'd use the tarballs - but given that we're involved in developing it, git is *the* tool that makes it convenient
18:04:26FromGitter<kaushalmodi> you might also want to reword the release description, now that we have Windows assets too
18:04:39shashlickI have
18:04:56*Snircle quit (Ping timeout: 246 seconds)
18:05:03FromGitter<kaushalmodi> oh yes, you have! :P
18:06:14FromGitter<kaushalmodi> .. so the key was to use TRAVIS_TAG
18:06:14FromGitter<kaushalmodi> nice
18:06:24Araqcan we come to a conclusion now? I'd like to automate 'git tag' with 'koch' anyway, I can make the commit message contain the C source and Nimble hashes, problem solved in ~5 lines of Nim code and automated
18:06:41Araqyes/no/maybe?
18:07:05shashlick@kaushalmodi: yep - looks like you went 95% down so was super easy
18:07:58shashlickone concern is that in the middle of the builds, someone could deliver something to devel and you'd get a different hash for some jobs
18:08:09shashlickless likely with stable but still possible
18:08:53narimiranwell, it is NIGHTLY β€” built it at 3am CET, you should be safe :D
18:09:06FromGitter<kaushalmodi> shashlick: yeah, but that shouldn't matter much
18:09:27shashlickya but the tags won't look good! πŸ˜‰
18:09:36FromGitter<kaushalmodi> it's updated every night, so one platform might have a max latency of 2 days compared to other platform
18:10:19FromGitter<kaushalmodi> shashlick: this is a bad rabbit hole, up to you :P
18:10:36shashlick@narimiran: 3am CET is 10pm for me
18:11:10shashlickor 8pm rather
18:11:21shashlickanyway, small victories
18:11:59narimiranbut there won't be pushes to devel at that time ;)
18:12:17shashlickbigger ones will be when linux/osx nightlies are binaries and then when choosenim can install a nightly binary
18:12:38FromGitter<arnetheduck> no/maybe. doesn't solve the branch use case (we patch nim but want to use the same nimble). it's good if it's combined with build_all and koch tools allowing to specify the hash to use, so we can use those to build nim.. a key objective is that we can use upstream build process to produce nim+nimble
18:13:42FromGitter<arnetheduck> in a way, that would accommodate @shashlick's use case as well, in case `nimble` ever gets more attention than `nim` - because then there's a common and supported way to build a new nimble for an old nim
18:14:54shashlickso tagging nim to a particular nimble release should be fine, even if nimble becomes more independent of compiler
18:16:45shashlickhowever, nimble also has releases and tags so in theory, you could do nimble install nimble@version - curious if @dom96 already enabled that
18:16:58Araqyou lost me again. If you checkout Nim v0.17 why would you keep the current Nimble version?
18:17:42Araqit's unlikely to work... You need the nimble that belongs to 0.17 as you said it yourself
18:17:51FromGitter<arnetheduck> well, now that we've talked it through at least twice, I see it more clearly :) there are two separate problems: ⏎ ⏎ 1) there's no way to specify an exact version of nimble and csources to use when building (the official build instructions just take a random version that happens to be the latest)
18:18:22FromGitter<arnetheduck> 1) the record of which versions were used are loose / voluntary
18:19:48FromGitter<arnetheduck> @Araq - yes, it's unlikely to work. but it's the exact case that I want to enable. and actually, not just for 0.17 specifically, but for every commit in between 0.17 and 0.18 for example, and also for my private branch of nim
18:21:09shashlickso I went to my 0.18.0 install using choosenim, nimble -v shows 0.8.10
18:21:22shashlickthen I ran nimble install [email protected] and it built successfully
18:21:58shashlickit is installed to ~/.nimble/pkgs though and does not replace the nimble installed by choosenim and linked to from ~/.nimble/bin
18:22:02FromGitter<arnetheduck> I'm using 0.17 here as an example which is ridiculously old, but keep in mind that we'll be in the same spot a a year from now, talking about 0.19.. by that time, nimbus will have had an official release, and we will want to create patch releases for it. to do that, we will need to replicate the exact build environment we used to produce it in the first place, meaning we'll be going back in time a year, every time
18:22:02FromGitter... we want to fix a bug
18:22:03Araqbtw, csources version tags are Nim version tags
18:22:24shashlickso in theory, you could checkout an old nim and use nimble to build a specific tag or hash of nimble
18:22:46shashlickbut ya, you still need to know what nim tag links to what nimble tag
18:23:15Araqshashlick, also since the tar.xz contains Nimble's source code, you can find out its git commit hash. It's messy, but the information is there
18:23:35Araqyou merely have to compare all the file contents ;-)
18:24:07*rect0x51 joined #nim
18:25:00*vivus quit (Remote host closed the connection)
18:25:00*zachk joined #nim
18:25:24FromGitter<arnetheduck> > btw, csources version tags are Nim version tags ⏎ ⏎ ok, but `build_all.sh` does not use those tags - that's a missing link
18:25:38shashlick@kaushalmodi - looks like osx also was uploaded correctly - am pushing this commit to master
18:25:58FromGitter<kaushalmodi> shashlick: of course! thanks
18:26:37*zachk quit (Changing host)
18:26:37*zachk joined #nim
18:26:49rect0x51Hey, what's the best plan in getting a working Nim REPL? Are we out of ideas?
18:27:00shashlickam also going to delete the other stuff on the releases page to clean things up
18:27:11leorizerect0x51: for starters, there's one: nim secret
18:27:24leorizeand it sorta works
18:27:42leorizehttps://github.com/nim-lang/Nim/issues/8927
18:27:47rect0x51leorize: I know, it can't import important module though, right?
18:27:51shashlickthere's also inim
18:27:53shashlickhttps://github.com/AndreiRegiani/INim
18:27:53rect0x51modules*
18:27:59FromGitter<kaushalmodi> Araq: About the megatest and travis timing out, will you update koch.nim in Nim devel?
18:28:07FromGitter<kaushalmodi> Should I open an issue for that?
18:28:39AraqI'm thinking about it but I think nightlies could use its own koch.nim called build.nim
18:29:05Araqunless you enjoy this bash script hacking so much :P
18:29:58rect0x51Oh ok, so AndreiRegiani made an one after all :P I'll try it out
18:30:01FromGitter<kaushalmodi> Araq: this won't be in bash script land, right?
18:30:11FromGitter<kaushalmodi> the testinstall target in Nim core needs to be updated
18:30:25Araqdoes it though?
18:30:42Araqit works, travis times out so travis needs a 'koch shorttestinstall'
18:30:55Araqand it's not clear that's koch.nim's business
18:32:59shashlickI don't think we should do this - rather run the full test suite in a separate job
18:33:18FromGitter<kaushalmodi> shasklick: makes sense, but I don't know how
18:33:30shashlickneed to use build stages
18:33:33FromGitter<kaushalmodi> may be this comes back to the build phases you mentioned earlier
18:33:34FromGitter<kaushalmodi> yes
18:33:59rect0x51does inim depend on nimscript VM, or is it independent?
18:34:49leorizeit put everything into a file
18:34:59leorizeif it compiles, keep the added lines
18:35:02leorizeif not, remove it
18:35:14leorizethe code is reexecuted for every input
18:36:00rect0x51mmm i see, not a true REPL then
18:36:47leorizesee the issue I linked above
18:36:50rect0x51since nimscript has some limitations (not designed for this), I guess a proper REPL would require a different VM?
18:37:05leorizethe biggest limitation is that C FFI is not supported
18:37:17Araqrect0x51, the VM got FFI support in a branch
18:37:43AraqI don't want it but I won't fight it too heavily since everybody else wants it.
18:37:53*natrys joined #nim
18:37:54rect0x51Araq: ohh
18:38:24shashlickThat's fantastic news
18:38:44shashlickAraq do we still have the copyMem limitation though?
18:39:06Araqwhat is the copyMem limitation?
18:39:45shashlickWhen you import some modules it complains that you cannot cause the module allocates memory
18:39:56shashlickBeen a while, need to check back and see
18:40:19Araqthat is still there yes
18:40:38shashlickOk
18:40:50FromGitter<Clyybber> Araq: Why do you not want it? The patch is pretty simple, isn't it?
18:41:16Araqit's a can of worms.
18:41:39FromGitter<kaushalmodi> actually before looking into timeouts, I see that koch testinstall has many failures because of: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5c2d05e337975e7ca962cf9a]
18:42:16AraqFFI calls have "marshalling" semantics, pass a pointer, some blob is updated but its symolic VM representation isn't
18:43:00Araqhow do you debug this thing if it crashes and it will crash
18:43:37Araqcan you run async in the VM? threads?
18:44:24FromGitter<Clyybber> Maybe its time to use a C interpreter for the VM
18:44:33*krux02 quit (Remote host closed the connection)
18:44:36*d10n-work joined #nim
18:44:41FromGitter<Clyybber> or a fast C compiler
18:44:52*krux02 joined #nim
18:44:56shashlick@kaushalmodi need to install node
18:44:58Araqthat would make much more sense IMO
18:45:10Araqor finally embrace LLVM/nlvm
18:45:17shashlickTcc
18:45:25FromGitter<kaushalmodi> shashlick: yes, reading through https://docs.travis-ci.com/user/languages/javascript-with-nodejs/
18:45:36FromGitter<kaushalmodi> i wish they gave one golden way to do it
18:45:50shashlickNo just install it with apt
18:46:16AraqI'd like to see more action on nlvm to give us a REPL and a better debugging experience
18:47:05AraqI consider TinyC and libFFI to be deadends. And on my bad days Nim's C's backend too.
18:47:49FromGitter<kaushalmodi> shashlick: ok, let's see how the next iteration of testinstall fares
18:48:30FromGitter<arnetheduck> > better debugging experience ⏎ ⏎ anything specific?
18:48:54shashlick@kaushalmodi: interestingly nim-lang/nim doesn't install node
18:49:03shashlickin its travis file
18:49:26FromGitter<Clyybber> Would it make sense to upstream nlvm to get better exposure and development?
18:49:28Araqin theory and in practice C code is impossible to generate. Thanks to #define, you never know what symbols are #defined and these cannot be used as struct fields etc.
18:49:51FromGitter<kaushalmodi> shashlick: hmm wonder how the js tests are running on the nim repo then
18:50:28shashlickAraq: i've had decent luck with -dD to get defines
18:50:56FromGitter<kaushalmodi> ah! because of `dist: trusty`?
18:50:57FromGitter<kaushalmodi> https://docs.travis-ci.com/user/reference/trusty/#javascript-and-nodejs-images
18:51:53FromGitter<kaushalmodi> *all of this feels like deja vu to me*
18:54:43rect0x51nlvm compiles nim directly to llvm IR?
18:54:51Araqyup
18:56:07Araqand it's a real IR, not a clusterfuck of broken stuff like "strict aliasing rules that betray C's spirit" or "you need to make your locals volatile when using setjmp"
18:56:32FromGitter<Clyybber> Is it planned to be upstreamed?
18:57:47Araqit's not scheduled but some day I'll do it
18:57:58FromGitter<Clyybber> Nice
18:58:52Araqand of course my favourite "we won't tell you the contents of your overflow flags"
18:59:20shashlickAraq: is it possible to speed up the dochack step - perhaps like megatest
19:00:39shashlickperhaps merge compile all tutorials/examples
19:03:10Araqdo you mean the runnable tests?
19:03:15shashlicklooks like we are doing koch doc0 in linux/osx but winrelease runs dochack which takes much longer
19:03:17shashlickya
19:03:30Araqdon't do 'doc0'
19:03:34Araqdo 'doc' instead
19:03:36FromGitter<Clyybber> @arnetheduck Is {.importc.} working in nlvm?
19:04:29*quipa quit (Remote host closed the connection)
19:05:03shashlickAraq: so https://github.com/nim-lang/nightlies/blob/master/.travis.yml#L126 should change to koch doc?
19:05:30*quipa joined #nim
19:05:45Araqyeah, doc0 is for me only
19:06:01shashlickok will change that
19:07:00shashlicklooks like winrelease effectively does a doc0
19:07:07shashlicksince it passes in gaCode
19:07:30Araqhum, but when I wrote that it has some purpose
19:09:07shashlickok np will update linux / osx to do koch doc
19:09:15shashlickstill wondering why linux/osx are so much fater
19:09:16shashlickfater
19:09:19shashlickfassster
19:09:20*Snircle joined #nim
19:09:41shashlicki know winrelease does koch boot again but travis has no timestamps
19:10:24*quipa quit (Excess Flood)
19:11:59*quipa joined #nim
19:12:55*quipa quit (Read error: Connection reset by peer)
19:15:56shashlickAraq: given we don't post nsis installers anymore and nim is already compiled with the correct arch, do we really need to rebuild in winrelease
19:15:56shashlickhttps://github.com/nim-lang/Nim/blob/devel/koch.nim#L351
19:19:30*rockcavera joined #nim
19:22:24*kapil____ quit (Quit: Connection closed for inactivity)
19:22:45Araqwinrelease needs to know its bits
19:22:56Araq when sizeof(pointer) == 4:
19:22:56Araq winReleaseArch "32"
19:22:57Araq when sizeof(pointer) == 8:
19:22:59Araq winReleaseArch "64"
19:23:41shashlickthat is because winReleaseArch() recompiles everything, my question is why
19:24:14shashlickso in travis, i install 32bit mingw, then compile csources, then nim so i know everything is 32bit
19:24:30shashlickthen when winrelease is compiled, it is 32bit, along with everything else - so why recompile again?
19:30:45FromGitter<arnetheduck> @Clyybber `importc`: yes, `header`: no
19:31:15FromGitter<Clyybber> Nice, gonna try it out now
19:41:33FromGitter<timotheecour> @araq for your comment here https://github.com/nim-lang/Nim/pull/10150#issuecomment-450714131 on FFI: ⏎ `{.push experimental: "allowFFI”.}` doesn’t work ; `--experimental:allowFFI` works though
19:45:02FromGitter<arnetheduck> actually @Araq, "you need to make your locals volatile when using setjmp" is still true in llvm - that or some other barrier is needed to not reorder stuff
19:45:58FromGitter<timotheecour> When the VM starts reaches this point: `if allowFFI notin c.config.features`, even though a prior semantic pass correctly set `c.config.features`, by that time, `c.config.features` is back to empty
19:47:59*nsf quit (Quit: WeeChat 2.3)
19:51:00FromGitter<Clyybber> @arnetheduck When running `make compare` building fails: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5c2d162492cf4d22423403da]
19:52:24FromGitter<arnetheduck> yeah, that's the `pie` bug - your distro forces `pie` while `nlvm` builds non-`pie` objects.. https://stackoverflow.com/questions/43367427/32-bit-absolute-addresses-no-longer-allowed-in-x86-64-linux
19:53:04FromGitter<arnetheduck> need to add `--no-pie` or something like that to the link flags in `nim.cfg`
19:53:33FromGitter<Clyybber> alright, thanks
19:53:53*natrys quit (Quit: natrys)
19:54:29shashlickok - latest nim nightly releases are posted - https://github.com/nim-lang/nightlies/releases
19:54:31*ripspin quit (Quit: Leaving)
19:54:41FromGitter<Clyybber> @arnetheduck When I run make compare it builds Nim + nlvm with nlvm? Is that right?
19:54:44shashlick@kaushalmodi - already ran into the issue of different hashes πŸ˜„
19:55:00FromDiscord_<PusiteGA> so why i cant in catch block echo?
19:55:53FromDiscord_<PusiteGA> so why i cant in catch block echo? , my bad
19:55:55FromGitter<arnetheduck> it builds `nlvm` with nim, then builds `nlvm.self` with `nlvm`, then compares the `llvm IR` that the two produce when building `nlvm.nim`
19:57:13FromGitter<Clyybber> Ah, makes sense thanks
19:57:42FromGitter<arnetheduck> there's actually some small bug in there right now I think.. I mostly look at `make test` - some 10% are failing right now - some because of real bugs and some because of unsupported features (emit, header, c++, that kind of stuff)
19:59:56FromGitter<Clyybber> Why does nlvm build non-pie objects? Is it for better utilizing available registers?
20:00:18FromGitter<arnetheduck> it's the default and it works for me
20:01:34FromGitter<arnetheduck> ie the project hasn't gotten far enough to take a stance on such details.. generally, there's a small performance cost to PIE due to extra indirections in some cases, but I haven't really studied it
20:02:12FromGitter<arnetheduck> if you make a patch to change it to pie, I'll accept
20:02:30FromGitter<Clyybber> Alright, will do
20:15:41FromDiscord_<PusiteGA> yey i made my first github Nim repo πŸ˜ƒ , just a script but its a progress
20:15:41FromDiscord_<PusiteGA> https://github.com/kodkuce/SplitVideo-Nim
20:16:19FromDiscord_<PusiteGA> go lought if want πŸ˜ƒ
20:16:20FromGitter<kaushalmodi> PusiteGA: Why are you using semicolons?!
20:16:23narimiransemicolons?
20:16:52FromGitter<kaushalmodi> also committing binary to git is not a good idea
20:17:18FromDiscord_<PusiteGA> i am just used to writing semicolons
20:17:26narimiranalso give your code some breathing space
20:17:40FromGitter<kaushalmodi> PusiteGA: If you make your repo a valid nimble project, users can simply do `nimble install https://github.com/kodkuce/SplitVideo-Nim`
20:18:15FromGitter<kaushalmodi> you can read more about that on https://github.com/nim-lang/nimble#creating-packages
20:18:58FromDiscord_<PusiteGA> ye i know about nimbe but, why would anyone need this for a project its just a simple script for terminal
20:19:13FromDiscord_<PusiteGA> tought i guess somone could use it for project too
20:20:22FromGitter<kaushalmodi> PusiteGA: In that case, you can looking into using Travis to generate Release assets
20:20:44FromGitter<kaushalmodi> i.e. each time you commit to github, the corresponding binary gets updated in your GitHub repo's releases section
20:20:59FromGitter<kaushalmodi> though, I can understand, setting that for the first time can be very time consuming
20:21:11FromGitter<kaushalmodi> but once you do it for one Nim project, it's the same for all others
20:21:59FromDiscord_<PusiteGA> duno meybe i try some other time atm baby steps πŸ˜ƒ
20:22:54FromGitter<kaushalmodi> PusiteGA: I understand that. But binaries shouldn't be committed to git.
20:25:49FromDiscord_<PusiteGA> removed πŸ˜ƒ
20:26:13FromDiscord_<PusiteGA> removed πŸ˜ƒ , anyway whats reson i cant leave binary on git?
20:26:20rect0x51PusiteGA: O.O semicolons! (sorry I had to comment)
20:26:44FromDiscord_<PusiteGA> ok will remove them but i like semi πŸ˜ƒ
20:27:06FromGitter<kaushalmodi> PusiteGA: If you stop typing semicolons, you will improve your coding speed significantly :P
20:27:40narimiranwhile you're at it, add some spaces after ',' and around operators, etc.
20:28:05FromGitter<kaushalmodi> PusiteGA: you can google for "why not to commit binaries to git" and will find many posts
20:28:15FromGitter<kaushalmodi> here's an arbitrary post from those results: https://robinwinslow.uk/2013/06/11/dont-ever-commit-binary-files-to-git/
20:28:27narimiranbtw, nimpretty can fix some inconsistencies, try it out
20:28:41FromGitter<kaushalmodi> summary: your repo size will bloat, git is great for text stuff; you lose on diff, blame, etc. for binary file commits
20:29:08FromGitter<kaushalmodi> also if you have a "bad" binary committed, you risk someone running that by reverting to an older commit containing that binary
20:29:35FromGitter<kaushalmodi> in short: use git only for stuff *you* create, not what Nim or any other tool creates
20:29:43*dom96_w joined #nim
20:30:04FromGitter<kaushalmodi> also that binary won't work for everyone
20:30:19FromGitter<kaushalmodi> platform dependencies, glibc dependencies, etc
20:30:23rect0x51I really want to work on nlvm and repl but I'm 95% sure it's out of my reach :|
20:30:45*vivus joined #nim
20:30:54*dom96_w quit (Client Quit)
20:34:33*Vladar quit (Remote host closed the connection)
20:35:11FromGitter<zacharycarter> there's already WIP for hot code reloading
20:35:42FromGitter<zacharycarter> rect0x51: https://github.com/nim-lang/Nim/issues/8927
20:36:27rect0x51zacharycarter: thanks, leorize linked it earlier
20:38:20FromDiscord_<PusiteGA> i removed semicolons, but readed late about spaces around operators xD
20:38:49FromDiscord_<PusiteGA> tought i noticed when i have semicolons i mix them somethins with : so i think its better to not write them
20:39:28*kungtotte joined #nim
20:40:10FromDiscord_<PusiteGA> nice i killed my script while cleaning up xD
20:40:21FromDiscord_<PusiteGA> should test before commit
20:40:23FromGitter<zacharycarter> ah, my bad - I think a lot of the work is already done
20:40:37FromGitter<zacharycarter> at least it appears so if you examine the fork / branch
20:41:47*endragor joined #nim
20:55:54*narimiran quit (Quit: Leaving)
20:55:58FromDiscord_<PusiteGA> hmm, 1 quesiton when i run process if i dont do outputStreem.ReadAll() it continues code even it dident finish process
20:56:07FromDiscord_<PusiteGA> hmm, 1 quesiton when i run process if i dont do outputStreem.ReadAll() it continues code even it dident finish process, that normal?
21:00:28*rockcavera quit (Ping timeout: 245 seconds)
21:02:07*Trustable quit (Remote host closed the connection)
21:10:17*abm quit (Quit: Leaving)
21:30:10*xet7 quit (Quit: Leaving)
21:30:31*sagax joined #nim
21:35:13rect0x51What the heck is a jupyter kernel? I don't get it.
21:41:18FromGitter<kaushalmodi> Feels great to not have to `import typetraits` just to print a variable type name ( https://gitlab.com/kaushalmodi/kaushalmodi.gitlab.io/commit/eb38c453fddba99192b3a4e554056bd0f7bd571e <-- expand diff for the `nim.org` file). Thanks @timotheecour!
21:44:38*thomasross quit (Remote host closed the connection)
21:45:05*thomasross joined #nim
22:08:07rayman22201@rect0x51 https://jupyter.org/ It's a popular live coding environment
22:09:06*MyMind joined #nim
22:11:13*Sembei quit (Ping timeout: 268 seconds)
22:15:28*MyMind quit (Ping timeout: 245 seconds)
22:15:34*Pisuke joined #nim
22:17:36rect0x51rayman22201: Like an online REPL or something?
22:20:12*lritter quit (Ping timeout: 272 seconds)
22:24:17rayman22201It's a weird thing to explain. It has "REPL" like features, but it's more like literate coding. You have a browser based WYSIWYG interface where you can write notes, but also you can make little snippets of code that run along with your notes.
22:24:42rayman22201It's popular with some data science people
22:25:14rayman22201the best way to understand it is to try it: https://jupyter.org/try
22:28:33rayman22201The snippets run by sending the little code snippets from the browser to a server that compiles and runs the code, and then sends the result back to browser. The server that actually runs the code is called a "jupyter kernel" and you can swap out different languages / compilers.
22:29:23*dddddd quit (Ping timeout: 244 seconds)
22:30:32rect0x51rayman22201: wow it sounds powerful
22:34:14rect0x51do you know which branch is the one where VM got FFI support?
22:34:50FromGitter<Clyybber> rect0x51: https://github.com/nim-lang/Nim/pull/10150
22:38:11rayman22201Jupyter is ok. Depends on your use case. It has limits. It's good for some research projects lol.
22:39:25rect0x51Clyybber: thanks!
22:39:51FromGitter<Clyybber> np :)
22:52:51*dddddd joined #nim
22:54:09*Ven`` joined #nim
22:54:48*cspar joined #nim
23:11:17*Geezus42 joined #nim
23:17:15stefanos82jupyter has become the standard tool for data scientists and AI / ML / DL researchers
23:17:37stefanos82I use it for teaching myself ML
23:23:55*stefanos82 quit (Remote host closed the connection)
23:26:06shashlickWhat is the popular choice for IPC in Nim
23:26:06FromGitter<Clyybber> @arnetheduck Fixed the issue with PIE being enabled by default: https://github.com/arnetheduck/nlvm/pull/10
23:27:22FromGitter<Clyybber> shashlick: dbus always works, doesn't it?
23:29:03FromGitter<Clyybber> shashlick: Assuming you are not on windows, which I now remember you are :P
23:30:15FromGitter<Clyybber> But I guess pipes or sockets should do
23:31:17FromGitter<Clyybber> dbus should also work on windows
23:34:32*rect0x51 quit (Quit: WeeChat 2.3)
23:43:40*kinkinkijkin joined #nim