<< 29-11-2019 >>
00:17:15FromDiscord_<onelivesleft> Hey, tangential question: I'm trying to get my nim-generated doc to link to the github file. It looks like the url is correct, and I've pushed the tag to github, but github is giving me a 404.
00:19:55FromDiscord_<onelivesleft> This looks right, right? : https://github.com/onelivesleft/simple_parseopt.git/tree/v1.0.0/src/simple_parseopt.nim#L98
00:21:17FromDiscord_<onelivesleft> https://github.com/onelivesleft/simple_parseopt/releases/tag/v1.0.0
00:21:51shashlicktree should be blob
00:22:56FromDiscord_<onelivesleft> blob doesn't work either, and that url is generated by `nim doc2`
00:27:57FromDiscord_<onelivesleft> ok, something screwy is going on: in one tab it 404s, in another (where I typed in the url) it works
00:30:08FromDiscord_<onelivesleft> n/m, it was the `.git` in the url
00:56:02madpropsis there a library that orders items in columns like the ls command does?
01:00:12*joshbaptiste quit (Ping timeout: 252 seconds)
01:00:35*joshbaptiste joined #nim
01:05:13zedeus!repo asciitables
01:05:14disbothttps://github.com/xmonader/nim-asciitables -- 11simple terminal ascii tables for nim 8 21⭐15 0🍴 7& 1 more...
01:05:30zedeus!repo terminaltables
01:05:31disbothttps://github.com/xmonader/nim-terminaltables -- 11terminal tables for nim 8 16⭐15 1🍴
01:06:27*clyybber quit (Quit: WeeChat 2.6)
01:11:11madpropsoh coool
01:23:33madpropshmm not versioned
01:25:02disruptekyeah, i'm having to point at a git hash for terminaltables.
01:25:33disruptekso annoying that people refuse to tag releases.
01:25:49madpropsthe hash is the commit id right?
01:26:13madpropsnot sure how to do this
01:26:54disruptekyeah.
01:27:04disruptek!repo disruptek/wet
01:27:05disbothttps://github.com/disruptek/wet -- 11render hourly weather report in the console 8 0⭐15 0🍴
01:27:41disrupteki'm using it in golden, too, but wet might be a better demo, crude as it is.
01:27:49madpropsnice
01:31:16madpropshow do i update nimble?
01:32:04madpropsmy older version doesn't work anymore after 1.0.4
01:32:48disruptekdunno. i used to use a fork but now i just use nimph.
01:33:23disruptekdoesn't `build_all` build it?
01:34:54madpropsbuild_all is a command?
01:35:43disruptekbuild_all.sh in the Nim repo.
01:39:45madpropshmm
01:40:03madpropsjust tried going to the choosenim toolchains dir
01:40:07madpropsand doing ./koch tools
01:40:12madpropsbut compilation fails
01:40:26madpropsthat's basically what build_all.sh does
01:40:27disruptekyou're on linux, right?
01:40:31madpropsyea
01:40:44disruptekwhy don't you see if you can build nimph and let me know what you think.
01:40:50disruptek!repo disruptek/nimph
01:40:50disbothttps://github.com/disruptek/nimph -- 11a nim package hierarchy manager from the future 8 2⭐15 0🍴
01:41:18madpropsaight
01:42:47disruptekthought if you can build that, you should be able to build nimble.
01:43:02madpropscan't install dependencies cause nimble is broken lol
01:43:18disruptekthe irony.
01:45:44madpropsim gonna try running build_all.sh entirely
01:49:02disruptekseems weird that nimble cannot be installed with nimble.
01:52:14*thomasross quit (Ping timeout: 268 seconds)
01:52:31*thomasross joined #nim
01:53:27madpropswhen i try to run nimble i geT:
01:53:29madpropsError: Requested executable is missing. (Path: /home/yo/.choosenim/toolchains/nim-1.0.4/bin/nimble)
01:54:03madpropsseems im gonna have to download the nim repo
01:54:06madpropsand run build_all
01:54:09zedeusif you're using choosenim you don't have to build stuff yourself
01:54:28madpropsi tried choosenim update nimble etc
01:54:36zedeus`choosenim update stable`
01:54:39zedeusor devel
01:54:45madpropsi did that
01:54:49madpropsit updated nim to 1.0.4
01:54:52madpropsbut not nimble
01:55:08madpropswell i have to mention in mid compilation my computer froze (im taking care of that issue)
01:55:36zedeusnimble is included in the update
01:55:55disruptekdoesn't choosenim use binaries?
01:56:02zedeussometimes
01:56:22disrupteki don't use it. i think life is better via the repo.
01:56:42zedeusmeh i prefer just running `choosenim update devel` and letting it do its thing
01:58:23madpropsso maybe i could remove 1.0.4 and try installing again
01:58:41madpropsit definitely didn't use binaries. it did a long compilation
02:01:29madpropshmm it fails at: Building tools (nimble, nimgrep, nimpretty, nimsuggest)
02:01:41madpropsit just says it failed it seems
02:01:43disruptekin the repo?
02:01:48madpropswith choosenim
02:02:11madprops.cache/nim/nimfind_r/stdlib_assertions.nim.c.o: file not recognized: File truncated
02:02:20disruptekrm -rf .cache/nim
02:02:53madpropsok here we go again
02:03:41disruptekat some point, you have to ask yourself what your time is worth.
02:05:41madpropsok done
02:06:01madpropsi would prefer to download binaries a lot more of course
02:06:10madpropsinstead of compiling
02:06:16disruptekpffbt, not me.
02:06:51madpropswhat's the benefit of compiling?
02:07:02madpropsif the binaries came from the nim website they should be trustable
02:07:04disruptekit's more likely to work.
02:07:42disruptektakes me <9s to build nim. what's the point of hazarding someone else's toolchain?
02:07:56madpropsit takes me a couple of minutes here
02:08:06disruptekhow is that possible?
02:08:23madpropswell i have an i5 from 2014
02:08:23*bacterio joined #nim
02:08:35madpropsbut if im lucky im gonna have a new amd soon
02:09:03madpropsbtw: equires "terminaltables.git#82ee5890c13e381de0f11c8ba6fe484d7c0c2f19"
02:09:05disbothttps://github.com/nim-lang/Nim/pull/82 -- 5Implemented Lazy[T], Documentation changes, etc @11AzureKitsune
02:09:06madpropsdoesn't seem to work
02:09:10madprops"requires"*
02:09:23disruptektake out the .git
02:09:48madpropswhy do you have .git there?
02:10:01disruptek'cause i'm dumb.
02:10:12madprops:o
02:16:51disrupteki guess i should make the bot ignore that stuff.
02:19:59disruptekterminaltables.git#82ee5890c13e381de0f11c8ba6fe484d7c0c2f19
02:21:04disruptek#12767
02:21:06disbothttps://github.com/nim-lang/Nim/issues/12767 -- 3new $nimbleDir path substitution yields unexpected search paths @11disruptek; snippet at 12https://play.nim-lang.org/#ix=235K 😏
02:26:09madpropsterminaltables works but
02:26:15madpropsi have to figure out how to render the columns
02:26:32madpropsi mean ls does it in alphabetical order going down
02:26:47disruptekyes, the math is up to you.
02:26:53madpropshehe
02:31:11disruptekso build_all.sh worked.
02:31:34disruptekwelcome to the elite.
03:25:48*leorize_ joined #nim
03:25:52*leorize quit (Remote host closed the connection)
03:40:51madpropsyou ready? https://adventofcode.com/
03:55:36*endragor joined #nim
04:11:19disrupteknah.
04:13:07FromDiscord_<yewpad> not for me this year
04:13:50*chemist69 quit (Ping timeout: 276 seconds)
04:15:38*chemist69 joined #nim
04:36:14*nsf joined #nim
04:50:35madpropshow am i supposed to use a function that calls a function above that calls a function below? i tried separating them into files but recursive import problems .. no idea
04:54:52FromDiscord_<WilliamDraco> LIke.. a forward declaration?
04:55:01shashlickyou can create a function prototype above
04:55:15madpropsok i have an idea
05:06:06madpropswait, how do i forward declare a proc
05:06:12madpropsjust write the header up there and then implement it down?
05:06:21disruptekyeah, leave the = ... off.
05:08:26madpropsit says it expects an implementation
05:08:38madpropsoh wait
05:08:39disruptekdid you leave off the =?
05:08:54madpropsnvm
05:09:01madpropsyeah, it had different numbers of arguments
05:10:19madpropsbtw i've had a bit of a problem with for x in 0..somethingThatsZero
05:10:26madpropsit does it at least once sometimes i think
05:11:28disruptekyeah, but only when you aren't watching.
05:11:38disruptekyou have to keep your eye on those ranges.
05:11:38madpropsquantum trickery
05:14:46disruptekbut seriously, ranges are inclusive.
05:14:56disruptekuse 0 ..< 9 to omit 9.
05:26:05madpropsoh, thanks
05:26:13madpropsneed a space after the < ?
05:26:26madpropsi like to compact stuff sometimes
05:39:32madpropsstarting to make this tree mode http://i.imgur.com/IppSgxq.png
05:39:40madprops.git is so huge i just print "git stuff" heh
05:43:07*narimiran joined #nim
05:45:29*dddddd quit (Remote host closed the connection)
05:50:31*uu91 quit (Read error: Connection reset by peer)
05:51:10*uu91 joined #nim
06:09:02FromDiscord_<Rika> madprops, 0..<9 is afaik the preferred writing actually
06:09:30narimiran`0 ..< 9` is the nimpretty way, IIRC
06:11:21yumaikasSo, does Nim have closures a la JS?
06:11:31yumaikasOr nested procs, I guess?
06:11:36narimiranyes
06:12:27yumaikasYou have a link handy for the syntax for them? I'm looking, but it escapes me atm
06:14:09narimiranyumaikas: https://play.nim-lang.org/#ix=2363
06:15:01yumaikasThanks!
06:15:59yumaikasThat trailing = trips me up pretty often
06:16:45yumaikasFor better or worse reasons, I'm currently seeing if I can cook up a smallish spreadsheet application.
06:17:19yumaikasWriting a small little scripting language to use to extend the formulas
06:18:29*Kaynato quit (Quit: Leaving)
06:18:47*Kaynato joined #nim
06:23:07FromDiscord_<Rika> narimiran nimpretty doesnt format my code that way, is it off by default?
06:24:41narimiranthen i misremembered
06:28:27narimiranthis is the only example in NEP-1 with the range and it has spaces, even though it doesn't mention it explicitly: https://nim-lang.github.io/Nim/nep1.html#introduction-coding-conventions
06:30:24*thomasross_ joined #nim
06:30:24*thomasross quit (Killed (wolfe.freenode.net (Nickname regained by services)))
06:30:24*thomasross_ is now known as thomasross
06:31:57yumaikasnarimiran: How goes it?
06:32:53FromDiscord_<Rika> i'll wait for a rigid coding convention, like python has, because i dont have strong opinions on code convention (other than lines before braces, which nim doesnt have anyway)
06:45:53*clyybber joined #nim
06:52:13*thomasross quit (Ping timeout: 250 seconds)
06:53:14*thomasross joined #nim
07:04:38*solitudesf joined #nim
07:05:31*LargeEpsilon joined #nim
07:09:11*lbartoletti joined #nim
07:11:26*lbart quit (Ping timeout: 265 seconds)
07:19:10*uu91 quit (Ping timeout: 265 seconds)
07:20:48*uu91 joined #nim
07:20:55FromGitter<zacharycarter> does anyone know if it's possible to run apple's debugging tools without running your app inside xcode?
07:21:07FromGitter<zacharycarter> for instance - I'd like to use the metal frame capture tool but I don't want to use xcode
07:21:59FromGitter<zacharycarter> oh yeah - you just run instruments
07:22:04FromGitter<zacharycarter> dumb question my bad
07:25:56*solitudesf quit (Ping timeout: 265 seconds)
07:41:44*lbartoletti is now known as lbart
07:42:57*Perkol joined #nim
07:54:17FromGitter<zacharycarter> well apparently instruments cannot do metal frame captures so bleh
07:57:51*PMunch joined #nim
08:00:00*gmpreussner quit (Quit: kthxbye)
08:03:32*Perkol quit (Remote host closed the connection)
08:04:48*gmpreussner joined #nim
08:08:23Araq!github last-recently-used
08:08:24disbothttps://github.com/cleverdeng/LruCache.py -- 11Implements LRU(Last-recently-used) cache algorithm, Support the thread safe, With Python 8 12⭐15 5🍴 7& 29 more...
08:08:42Araq!github LruCache
08:08:44disbothttps://github.com/mohaps/lrucache11 -- 11A header only C++11 LRU Cache template class that allows you to define key, value and optionally the Map type. uses a double linked list and a std::unordered_map style container to provide fast insert, delete and update No dependencies other than the C++ standard library. This is a C++11 remake of my earlier LRUCache project (https://github.com/mohaps/lrucache) The goal was to create a fas
08:09:34*crem quit (Ping timeout: 252 seconds)
08:11:48*crem joined #nim
08:36:15PMunchHmm, looking at this Arduboy library I wonder if it would be possible to check for all instances of "print" in the code and translate all the strings to simple indices into the font buffer and only generate those characters that are actually required. Could save a nice chunk of memory..
08:37:43clyybberwhat about priniting runtime generated strings?
08:37:45PMunchSo say the string "Hello world" would be translated on compile-time to "\0\1\2\2\3\4\5\3\6\2\7" and the font bitmap would only store the characters "Helo wrd".
08:38:01PMunchclyybber, good point..
08:39:10PMunchIt could be added as an optimisation flag
08:39:43PMunchSo you can tell it to either include all, include only those used (no dynamic), or include only the characters in a supplied list.
08:46:34*floppydh joined #nim
08:46:34lqdev[m]PMunch: it wouldn't store the space, though
08:47:30PMunchHuh?
08:47:42PMunchWhy wouldn't it store the space?
08:48:17lqdev[m]because the space is an empty bitmap?
08:48:35PMunchAnd?
08:48:57PMunchOh you mean you could optimise it away so that it didn't store the space?
08:49:01PMunchYeah I guess that's true
08:49:24lqdev[m]exactly
08:51:12PMunchFollowing that logic it could also store a couple of special characters like / as a drawLine operation
08:54:38*njoseph quit (Ping timeout: 276 seconds)
08:58:39*njoseph joined #nim
09:04:53PMunchHmm, I'm looking at how nimterlingua is able to rewrite any text in the source code. And it uses term rewriting templates that it generates based on the INI file of translations. This obviously won't work for me since I don't know all the strings to begin with..
09:05:08PMunchCan term rewriting macros grab any string literal?
09:10:40PMunchApparently you can :)
09:13:59*narimiran quit (Quit: leaving)
09:16:37Araqit's PR friday? tell me about your important PRs
09:17:50FromDiscord_<Rika> it's friday?!
09:18:03Araqapparently
09:18:55*narimiran joined #nim
09:19:32PMunchFun with term rewriting: https://play.nim-lang.org/#ix=236m
09:21:47Araqyeah this is not supported if you read the spec carefully
09:22:05AraqTR macros are only for optimizations, you can turn them off
09:23:22lqdev[m]PMunch: why do `tick = if tick: false else: true` when you can simply do `tick = not tick`
09:23:42Araqand TR macros need some love in general, it's an old design
09:24:06Araqand I happen to believe that I'm a better language designer now... :P
09:25:48Araqin fact, I think I know how to do it properly now...
09:25:59*Araq must resist ... writing another ... RFC ...
09:26:02PMunchlqdev[m], I had it as tick = not tick initially, but it was all in a template with a `when` statement so it didn't work properly. That was just an attempt to make sure I wasn't misremembering how not worked
09:26:23PMunchAraq, well that was just an abuse to see how they worked. I'm actually thinking of an optimisation
09:27:00*solitudesf joined #nim
09:28:19PMunchSo memory is in short supply on microcontrollers. And the Arduboy library has a 1280 byte big array of all 256 extended ASCII characters in a 5x7 font in it. I was wondering if I could use term rewriting macros to extract which characters where actually used, and rewrite them to other offsets so that I could only include the used characters.
09:31:17Araqif you have control over the ascii array why not limit it to 'a'..'z', '0'..'9', ' ' and don't use other characters in your string
09:31:46PMunchBecause that would mean I have to manually specify all the characters I need
09:32:03Araqwell it's 37
09:32:05madpropstoday i learned about forward declaration, should be like the first thing that's taught lol
09:32:21Araq37 "manual" entries, come on
09:32:41Araqsometimes programming involves manual work
09:33:00PMunchI mean with Nim I can already say: generateFontArray({'a'..'z', 'A'..'Z', '0'..'9', ' '}) and have it generate a font array of those characters
09:33:09PMunchAnd write a macro that spits out those characters
09:33:25PMunchBut you'd also need a mapping table from ASCII -> myTable indices
09:33:50Araqeasily done:
09:33:52PMunchAraq, well I'm trying to blow away the dusty C programmers here :P
09:34:06Araqtranslate"look at this string"
09:34:42Araqjust give your strings a prefix, solved.
09:34:58madpropsi still would like f"" lol
09:35:05Araqyou won't have many strings in your program either way, you have no memory to hold them
09:35:05PMunchThey already got exited when I showed them a macro that converted bitmaps into the internal sprite format so that they didn't have to drag it into a web-site and copy-paste the code into their code :)
09:35:20madpropslike you have 'and' and 'or' to avoid hitting that nasty symbols key row. But no f""
09:35:22FromDiscord_<Rika> woah woah woah what do these {a}s and {lit, noalias} things mean PMunch
09:35:39Araqmadprops, &"{x} from strformat
09:35:53PMunchRika, shhhh don't worry about it
09:35:54madpropsyes that's what i do. my point is that that & is hard to type
09:36:00FromDiscord_<Rika> what
09:36:00PMunchRika: https://nim-lang.org/docs/manual_experimental.html#term-rewriting-macros
09:36:07FromDiscord_<Rika> are we back at the hard to type argument again
09:36:08Araqfmt"{x}" # also possiblee
09:36:16FromDiscord_<Rika> oh thanks
09:36:17madpropsnope. fmt is a raw string
09:36:20narimiranmadprops: that again?
09:36:23FromDiscord_<Rika> seems like i skipped over that section
09:36:23madpropslol
09:36:26Araq"{x}".fmt
09:36:33madpropshmm
09:36:38Araq^ works too and now shut up :P
09:36:50madpropsdo you have anything against the letter f though
09:36:50FromDiscord_<Rika> personally i dont like the & thing too, because i dont see the connection between "format this string" and "ampersand"
09:36:51AraqI personally like &"" best
09:37:12AraqRika: it produces a concat expression, that's the connection
09:37:18FromDiscord_<Rika> ok
09:37:22FromDiscord_<Rika> im like
09:37:24FromDiscord_<Rika> negative iq
09:37:33FromDiscord_<Rika> forgive me for not seeing that]
09:37:34madpropsim like negative desire to use the symbols row
09:37:48madpropsmy soft typing skills don't apply there
09:37:49Araqwe also thought about %"" which would produce a format expression
09:37:58Araqwhich is important for i18n
09:38:08narimiranmadprops: so you never use parentheses?
09:38:11Araqbut we don't have it (yet?)
09:38:24madpropsparens are easy because i use them a lot. it's natural but & *& ^%$% etc is unatural
09:38:46narimiranthen we should immediately change $ too
09:38:49FromDiscord_<Rika> maybe you're just not used to it
09:38:51PMunchmadprops, pull a french one and swap numbers an symbols
09:39:09FromDiscord_<Rika> # for comments isnt natural coming from c-likes isnt it
09:39:09narimiranshit, you never do multiplication either
09:39:23madprops#@* are easy
09:39:30madpropsit's the middle section that troubles me
09:39:37PMunchmadprops, what layout do you use?
09:39:37FromDiscord_<Rika> bruh
09:39:41FromDiscord_<Rika> * is in the middle is it not
09:39:48madpropsyea
09:39:50FromDiscord_<Rika> good question pmunch
09:39:50narimiranso * is easy, but one key to the left of it is huge problem
09:40:00madpropsyea lol idk why
09:40:09FromDiscord_<Rika> basically, you're weird
09:40:14madpropsah
09:40:38PMunchI mean I kinda agree with madprops, I can use the symbols I'm used to using, but things like & and $ have me looking at my keyboard to find them..
09:40:54FromDiscord_<itmuckel> Hey guys! What model loader (for example for .obj files) do you use for your OpenGL projects?
09:41:07narimiranPMunch: so the solution is to use &{...} more so you get used to it! :D
09:41:24PMunch&{}?
09:41:27FromDiscord_<itmuckel> I found https://github.com/barcharcraz/nim-assimp/blob/master/assimp.nim but it doesn't compile anymore
09:41:37FromDiscord_<Rika> its an opportunity to never need to look at the keyboard again for any symbol
09:41:40narimiranah, &"...{...}...."
09:41:48PMunchOf course it doesn't help that I use a Norwegian layout so half of the useful keys are behind AltGr and the other half is behind shift
09:41:55Araqitmuckel: please fix it an create a PR, everything should work with 1.0, that's why we have it
09:42:08FromDiscord_<Rika> it still mentions nimrod damn
09:42:16Araq1.0 is for the Nimble packages to catch up
09:42:19FromDiscord_<Rika> is the dev active?
09:42:34AraqI doubt it, but you can always fork it
09:42:38PMunchOoof, if it's from the Nimrod days it's pretty old..
09:42:50FromDiscord_<Rika> sub-0.13?
09:42:56PMunchSomething like that
09:43:10FromDiscord_<itmuckel> I just looked it up, there are three pull requests that address the problems I have with it, but none of them accepted. 🙄
09:43:11PMunchYeah, sub 0.13
09:43:21Araqfork it!
09:43:41FromDiscord_<Rika> fork their pr branch
09:43:51FromDiscord_<Rika> then just use your fork lmao
09:44:04FromDiscord_<Rika> or theirs if the code is all working
09:44:04PMunchI know because that's how far the playground goes back :P I tried to build versions further back but it didn't work as the binary ended up being named "nimrod" so it couldn't find it when it wanted to compile code :P
09:45:23FromDiscord_<Rika> oh wow, a converter
09:45:29FromDiscord_<Rika> are converters still useful
09:45:49FromDiscord_<Rika> or is there something more preferred
09:46:34PMunchWell they can be useful
09:47:06PMunchI find them particularily helpful when working with C libraries that just fuck around with their types..
09:48:11FromDiscord_<itmuckel> Why would I fork it? I could just use a fork. This seems to incorporate the other pull requests, but I'm not sure. What's the best approach when you have an abandoned project like this? A new fork fragments this even more
09:48:11FromDiscord_<itmuckel> https://cdn.discordapp.com/attachments/371759389889003532/649909461338816542/unknown.png
09:48:20FromDiscord_<Rika> i havent encountered a situation where i had to interact with a c library yet
09:48:36PMunchitmuckel, you fork it and take over maintenance :)
09:48:39FromDiscord_<Rika> @itmuckel if it doesnt work, fork the one with most progress
09:48:52narimiranideally you would fork it and send a PR to nimble packages so it now points to your repo
09:49:03FromDiscord_<itmuckel> @PMunch Aaaaaah
09:49:03FromDiscord_<itmuckel> 🙂
09:49:20madpropswhy does choosenim update doesn't download binaries?
09:49:34FromDiscord_<Rika> it shou;d
09:49:38madprops;d
09:49:43PMunchRika, yeah.. I keep doing this thing where I'm thinking about how I can fit the non-shaped liquid Nim into various shaped holes..
09:49:44FromDiscord_<Rika> ree
09:50:00madpropshmm liquid nim
09:50:08*leorize_ quit (Ping timeout: 260 seconds)
09:50:09PMunchRika, problem with your symbol row there?
09:50:16madpropsobviously
09:50:23PMunchmadprops, it's got electrolytes!
09:50:24FromDiscord_<Rika> lmao it's the home row
09:50:39FromDiscord_<Rika> i suck at typing on my right hand
09:50:44PMunchRika, hmm for me ' is one row up from home
09:51:03FromDiscord_<Rika> i hit ;
09:51:16madpropsyeah my l and ; are next to each other
09:51:20FromDiscord_<Rika> besude l
09:51:21PMunchThat for me is top left..
09:51:22FromDiscord_<Rika> fcusgbeyn
09:51:28FromDiscord_<Rika> beside.
09:51:31madpropsproblem i have sometimes is that i hit / instead of ?
09:51:31FromDiscord_<Rika> b e s i d e
09:51:44madpropsbad Shift timing
09:51:49FromDiscord_<Rika> pmunch, you got an odd keyboard
09:51:59FromDiscord_<Rika> madprops, you could say it's a race condition
09:52:01FromDiscord_<Rika> heh
09:52:02PMunchRika, oh you only know the half of it :P
09:52:15FromDiscord_<Rika> this is super off topic now lmao
09:54:18PMunchRika: https://photos.app.goo.gl/c83maeUQokXYjxwU8
09:54:31PMunchThat is my keyboard
09:54:49PMunchAnd some of the buttons don't quite do what they are labeled as
09:54:51madpropsthat Z is a long way from home my dude
09:54:58madpropsactually everything is
09:55:01FromDiscord_<Rika> what the absolute shit is that layout
09:55:01madpropsis this dvorak?
09:55:09FromDiscord_<Rika> i was gonna ask
09:55:12PMunchmadprops, Norwegian Dvorak yes
09:55:34narimiranPMunch: your / is in the "hardest place to type it"
09:55:48madpropsmy / is where your Z is
09:55:50PMunchHold CapsLock and it works like Ctrl, hit it and you get Escape. Hold tab and it works like Super, hit it and you get a tab.
09:56:09PMunchYeah Shift+7 is inconvenient..
09:56:24PMunchBut not more than AltGr + 7 to create {
09:56:25FromDiscord_<Rika> / looks so irritating to use for you
09:56:37FromDiscord_<Rika> ouch.
09:56:54madpropswell at least you have some nice media buttons
09:57:06PMunchYeah.. Norwegian keyboards aren't really made for programming..
09:57:21PMunchHaha, I don't think I've ever hit a single one of those media keys :P
09:57:27madpropsis there a reason you don't use a US kb?
09:57:29FromDiscord_<Rika> programming was made with a qwerty keyboard in mind so...
09:57:34*leorize_ joined #nim
09:57:39FromDiscord_<Rika> maybe hes norwegian i dont know
09:57:40PMunchAnd TBH I don't know if they work on this Linux setup
09:57:49AraqNim was made with the German keyboard layout in mind...
09:57:50PMunchHaha, yeah I'm Norwegian
09:58:06madpropsi mean my first language is spanish and i use US layout
09:58:06Araq# is easy to type, {} is not
09:58:07FromDiscord_<Rika> Araq, TIL
09:58:24FromDiscord_<Rika> {} is still kinda hard to type in qwerty though
09:58:33PMunchWell I kinda need ÆØÅ, which I can't get on US keyboards
09:58:42madpropsproblem is i have to find the ñ on google
09:58:42PMunchAnd I prefer the ISO enter over the ANSI one
09:58:47Araqand äüöß are operators in Nim, instead of x.len you can also type älen
09:59:10AraqI mean äx
09:59:12FromDiscord_<Rika> i'd like to see that in an april fools
09:59:20FromDiscord_<Rika> this heavily reminds me of APL now
09:59:25FromDiscord_<Rika> or J
09:59:39FromDiscord_<Rika> i'm getting kinda tingly because APL is scary
09:59:41PMunchWe have dead-keys madprops, so I hit Alt+Gr + [~] followed by an n to make ñ
10:00:04FromDiscord_<Rika> madprops, can't you just do n + ~
10:00:08madpropsn~
10:00:15FromDiscord_<Rika> us intl allows for stuff like that apparently
10:00:17Araqtrue story: I thought about using # for len
10:00:26FromDiscord_<Rika> sounds like lua
10:00:30narimiranlast character of string `s`? get it with š
10:00:32PMunchI can actually create the german letters more easily than some symbols on this keyboard..
10:00:41madpropslike #names ?
10:00:43PMunchß for example is just AltGr + s
10:00:51Araqmadprops, yeah
10:01:03madpropswhat stopped you?
10:01:06FromDiscord_<Rika> <-- first language has "a letter" with two letters
10:01:15PMunchAnd ä is just the same key to make ñ but without AltGr and then followed by a
10:01:16*abm joined #nim
10:01:16Araqbut by then I had thousands of lines of code where # was used to introduce comments
10:01:39madpropsbtw is the 'result' proc return variable thing you borrowed from other language?
10:01:59FromDiscord_<Rika> feels like its from golang
10:01:59PMunchPascal IIRC
10:02:03PMunchOr Delphi rather
10:02:05FromDiscord_<Rika> i think golang has it i dont remember
10:02:06Araqand # comment vs #array (no space here!) would have been a bitter pill to swallow
10:02:08FromDiscord_<Rika> ah i see
10:02:22PMunchAraq, yeah that would be ugly :P
10:02:24madpropsi mean i kinda feel pleasure in typing stuff.len
10:02:38madpropsit's nice to type
10:02:44Araqanyway, 'result' is in Delphi and Eiffel too
10:02:52FromDiscord_<Rika> ah UFCS is such a godsend
10:02:59Araqand C++ emulates it via NRVO
10:03:04PMunchRika, Nim is older than Golang..
10:03:23FromDiscord_<Rika> PMunch, features can be introduced after the creation of a language
10:03:50Araqso it seemed like a very good idea. it also fits the whole "unstructured control flow is bad" narrative
10:03:50PMunchRika, yeah UFCS is amazing, but it also comes up as a contested topic when Nim is mentioned..
10:04:02madpropssomething i wonder, is the autoindentation in vscode in any way controlled by the nim plugin? because sometimes it's pretty bad
10:04:16Araqit's not and it's terrible
10:04:23narimiransTyLe_iN_SenS__iTiViTy !!!1!111!!!1!
10:04:29PMunchThe one on the playground is pretty bad as well..
10:04:38FromDiscord_<Rika> narimiran, i hate that argument
10:04:39Araqnarimiran, yay...
10:04:56FromDiscord_<Rika> who would ever do that in a real scenario...
10:04:57narimiranAraq: well nobody had mentioned it yet, i couldn't let that happen
10:05:03PMunchnarimiran, that is actually not a valid identifier :P
10:05:19narimiranPMunch: the exclamation part?
10:05:21PMunchCan't have __
10:05:30Araqas I said, you can use \u00AC in Java in identifiers
10:05:31narimiranwow had no idea :D
10:05:31FromDiscord_<Rika> Araq, so it's vscode that sometimes gives me a 4 space indent when there's no reason to and not the plugin?
10:05:38PMunchOnly a single consecutive underscore
10:05:59FromDiscord_<Rika> what really
10:06:02FromDiscord_<Rika> let me try that
10:06:32PMunchI figured that out when I wanted to separate something in a long bit literal IIRC
10:06:32FromDiscord_<Rika> what the fuck its real
10:06:32narimiranyep, it complains about trailing underscore
10:06:33FromDiscord_<Rika> is there a reason?
10:06:36Araqok, one more time: Nothing in C#/C++/Java stops me from using sTyLe_iN_SenS__iTiViTy as an identifier
10:06:38FromDiscord_<Rika> that doesnt sound right
10:06:56Araqit's a valid identifier in these languages, they only enforce that I apply it consistently
10:07:07Araqyay, consistently bad identifiers, great.
10:07:17FromDiscord_<Rika> basically what you're saying is that whatever the buffoons are saying is invalid
10:07:17Araqsee lua's API for a real world example
10:07:18PMunch!eval echo NimVersion
10:07:27NimBot1.0.4
10:07:32narimiranweeeeeeee
10:07:39Araqlua_wecannotuseunderscores(...)
10:07:44PMunchFor some reason the damn cron-job didn't work, again..
10:08:22PMunchThe worst part is that I can see that the cron job ran..
10:09:03PMunchAnd running that script manually works fine
10:09:34FromDiscord_<Rika> bad cron
10:13:43*leorize_ quit (Quit: WeeChat 2.6)
10:13:54Araqcan somebody take over https://github.com/nim-lang/Nim/pull/12506 please?
10:13:55disbot➥ Fix sequtils.delete bug with out of bounds indexes @11GULPF
10:14:18*NimBot joined #nim
10:14:32*uu91 joined #nim
10:14:52Araqyeah, that sounds good, thanks for listening
10:15:23PMunchWe're all rubber ducks on this blessed day!
10:15:53PMunchOh FFS, why is Windows so bad at reporting DLL errors..
10:17:11FromDiscord_<Rika> can it be that `delete` is the official name for procs that delete
10:18:17AraqPMunch, get the "CFF explorer"
10:18:32Araqit's a good tool to inspect DLLs and their deps
10:18:48Araqor even to figure out whether it's a 32 or a 64 bit DLL
10:20:00PMunchIt's 64-bit
10:20:15PMunch"Portable Executable 64" according to CFF Explorer
10:23:36Araqas I said, it also tells you about its deps
10:23:49PMunchAha
10:23:57PMunchIt was indeed missing a dependency
10:25:59PMunchGood thing I have the error message: "The specified module could not be found" without it specifying which fucking module it's trying to load..
10:26:39PMunchOn Linux at least that error message says something like "Couldn't find somerandomlib.so"
10:28:19AraqPMunch, might also be a Nim stdlib problem
10:28:32Araqwhere we don't ask for extended error info
10:29:07PMunchThe error message is from the parent C application
10:29:26PMunchWhich calls some Windows procedure to read out the error message of an error code
10:29:35PMunchSimilar to what you would do with "dlerror" on Linux
10:34:01Araqinteresting but we all know Windows is a piece of junk :-)
10:35:11PMunchThis is the part that is used to log the errors: https://github.com/PMunch/unbound/blob/master/dynlibmod/dynlibmod.c#L11-L32
10:35:43PMunchI guess there might be a way to get more information..
10:43:11*Trustable joined #nim
10:52:27madpropssome feedback about the errors messages .. i often find them too verbose like "got this but expected this or this or this or this oh and you forgot to make it mutable or this or this"
10:52:50madpropsi think rust messages are a bit more effective at this
10:53:22madpropslike they pin point the most probable problem
10:53:38narimiranwe came full circle
10:54:08madpropsyears ago it was the opposite?
10:54:16*lritter joined #nim
10:54:23narimiranwhat difference a year makes
10:56:16FromDiscord_<Rika> narimiran, care to explain?
10:56:53Araqit used to be "the error messages are bad, I dunno where my mistake is"
10:56:58narimiran^
10:57:14FromDiscord_<Rika> madprops said that rust's error messages do better though
10:57:23FromDiscord_<Rika> and those are super verbose arent they?
10:57:33madpropsrust even prints like the line numbers with a big red "the error is here!" lol
10:57:47FromDiscord_<Rika> even makes arrows...
10:57:59AraqRust assumes you use tooling from 1980
10:57:59madpropsbut that said i find the error in nim pretty fast
10:58:09FromDiscord_<Rika> what do you mean?
10:58:12FromDiscord_<Rika> 1980?
10:58:21FromDiscord_<Rika> oh
10:58:24FromDiscord_<Rika> terminal only
10:58:32FromDiscord_<Rika> i mean, it looks pretty thats for sur
10:58:39Araqeditor --> switch to terminal --> look at error messages --> switch back to editor
10:58:44FromDiscord_<Rika> not super pretty but kinda pretty
10:58:54FromDiscord_<Rika> maybe "embezzled" is a better word
10:59:24AraqNim can do that too via some switch but I think it's anachronism
10:59:31madpropssomething i like about rust, is how the vscode checking feels. it quickly highlights/dehighlights. i know if there are no errors shown everything is ok
10:59:55madpropsin nim i pretty much need to compile to check
11:00:01FromGitter<alehander92> Araq i dont agree
11:00:07FromDiscord_<Rika> ah yeah the vscode plugin needs a save before it checks
11:00:08FromGitter<alehander92> why should one context switch 3 times
11:00:12FromGitter<alehander92> to just understand an error
11:00:40FromGitter<alehander92> i agree with having two "modes" with very dense vs more detailed context
11:00:42FromDiscord_<Rika> alehander92, a beginner would need to do that provided they dont install any plugin in their editor
11:01:04FromGitter<alehander92> there are so many cases where one still can build in terminal or look at logs e.g. CI etc
11:01:32FromGitter<alehander92> i think building in terminal is often better
11:01:38FromGitter<alehander92> this terminal might be even inside your editor
11:01:43FromGitter<alehander92> but it still gives you more flexibility
11:02:02FromGitter<alehander92> you can run a more complex file watcher system etc
11:02:11FromDiscord_<Rika> i have a good feeling you misunderstood what he implied
11:02:17FromDiscord_<Rika> but i cannot say for sure
11:02:56*nif quit (Quit: ...)
11:03:06*nif joined #nim
11:03:28FromGitter<alehander92> yeah i am sorry for my tone
11:03:43FromGitter<alehander92> and tooling is nice
11:03:55FromGitter<alehander92> and i agree sometimes detailed messages are overrated etc
11:04:14FromGitter<alehander92> but as an option i think detailed context is ok
11:04:49FromGitter<alehander92> narimiran but i think
11:04:53FromGitter<alehander92> the feedback is different
11:04:57FromGitter<alehander92> its like the info is good
11:05:17*njoseph quit (Ping timeout: 276 seconds)
11:05:22FromGitter<alehander92> it just might be presented more "visually", e.g. with overloads you can just underline / color the diff/mismatch
11:05:39FromGitter<alehander92> which is both dense and informative
11:06:02FromGitter<alehander92> again, very subjective
11:06:12FromDiscord_<Rika> i'd like that more
11:06:17FromDiscord_<Rika> errors on my editor
11:06:40FromGitter<alehander92> but this helps both
11:07:34FromGitter<alehander92> i am talking in the last sentences about what madprops said and rendering the info itself, even for signatures
11:07:51FromGitter<alehander92> you still see this error in your editor
11:11:03madpropscan't say rust tooling is great though
11:11:10madpropsthe vscode plugin is buggy
11:11:17madpropssometimes i need to restart vscode
11:12:33FromDiscord_<Rika> nim vscode editor doesnt update until i save tho, is that an issue or just because
11:13:00madpropswhat rust rls does is wait a few seconds after last thing typed
11:13:07madpropswhich would be better imo
11:14:56FromDiscord_<Rika> yeah, i dont know why nim's plugin doesnt do that
11:15:09narimiranuse autosave?
11:15:58FromDiscord_<Rika> what if i dont wanna save?!
11:16:00madpropsgonna try that. im used to saving on focus out
11:16:02FromDiscord_<Rika> i kid, i will
11:16:09FromDiscord_<Rika> me too, madprops
11:16:35narimiranyoungsters these days....
11:21:18*ng0 joined #nim
11:23:41*tjmac joined #nim
11:31:06*uu91 quit (Read error: Connection reset by peer)
11:31:20*uu91 joined #nim
11:34:35*uu91 quit (Read error: Connection reset by peer)
11:34:48*uu91 joined #nim
11:37:53*thomasross_ joined #nim
11:37:53*thomasross quit (Killed (adams.freenode.net (Nickname regained by services)))
11:37:53*thomasross_ is now known as thomasross
11:43:01*Romanson joined #nim
11:45:32*tjmac left #nim ("-bye")
11:47:32*thomasross quit (Ping timeout: 276 seconds)
11:47:47*thomasross joined #nim
11:51:30PMunchNice, managed to make a docker container that will take a huge C project and a small Nim DLL project, cross-compile them for Windows, then run a Windows application to spit out an installer that sets up and installs everything in working order :)
12:01:47*njoseph joined #nim
12:03:22Zevvwhere are the days that you just put your floppy in drive 1 and press reset to get your code to run
12:04:04Araqyeah it was great, you could write up to 3 times to a floppy until it broke in random ways
12:05:34Araqor the nice copy protection mechanisms where you could install a program exactly once from the floppy
12:06:06Araqand the company which sent you the floppies tried to install it itself and claimed "this version really works now, we tested it!"
12:07:37*nsf quit (Quit: WeeChat 2.6)
12:09:47*sealmove joined #nim
12:17:33FromGitter<alehander92> were those
12:17:36FromGitter<alehander92> the big floppies
12:19:08Araqno
12:19:19FromGitter<alehander92> i wonder
12:19:26FromGitter<alehander92> if gramophone disks
12:19:36FromGitter<alehander92> could have made it into modern computers
12:28:20*Vladar joined #nim
12:29:45Zevvwell, casette tapes did, no technical reasons why a record could not
12:30:46sealmoveZevv: want to tackle left-recursion?
12:32:53FromGitter<alehander92> hmm
12:33:09FromGitter<alehander92> cool
12:33:15FromGitter<alehander92> how is npeg going
12:34:47Zevvsealmove: well, its already there, basically
12:35:14Zevvits bounded by precedence levels, which solves the runaway problem
12:36:06Zevvcame in in 0.20.0
12:37:14Zevvfor really real left recursion the grammar needs to be rewritten, but I lack decent grammar AST for rewriting at that level
12:37:26Zevvalso, I dot understand the problem well enough
12:37:48Zevvbut to answe your question: sure, give it a go! )
12:43:10FromDiscord_<sealmove> Right, if there is no compelling evidence it's useful, then why bother.
12:45:17FromDiscord_<sealmove> precedence is good enough I guess, especially since it's a Nim dsl (where precedence is predefined and natural to the user)
12:46:36Zevvsure
12:47:07*floppydh quit (Quit: WeeChat 2.6)
12:47:31FromDiscord_<sealmove> Btw do we have bnf dsl?
12:50:28Zevvthere is something close in npeg misc
12:50:38FromDiscord_<sealmove> Oh also, Guido is tranfering Python to PEG https://medium.com/@gvanrossum_83706/peg-parsing-series-de5d41b2ed60
12:51:28Zevvhttps://github.com/zevv/npeg/blob/master/misc/mouse2npeg.nim
12:51:31FromGitter<zacharycarter> cool - got things working on osx
12:51:48FromGitter<zacharycarter> (https://files.gitter.im/nim-lang/Nim/uyLR/Screenshot-2019-11-29-at-14.51.11.png)
12:52:39Zevvbtw sealmove, did you get to try the vm register count update I made, does that solve it for you?
12:53:09*uu91 quit (Read error: Connection reset by peer)
12:53:25*uu91 joined #nim
12:53:40FromDiscord_<sealmove> Yes!! :))
12:53:54clyybberrebase and go go go
12:53:56FromDiscord_<sealmove> I am very happy, my parser works just fine!
12:57:41FromDiscord_<sealmove> Had to handle indentation with global state.
12:58:27FromDiscord_<sealmove> Also I ended up using many stacks. I wonder if a tree would make more sense, no idea.
13:00:03Zevvwhatever makes you happy
13:00:29Zevvwhy global stack? get me your parserurl again, let me take a look
13:01:28ZevvAraq: with the vm increase to 64 bits we also increased regs ABC tot 16 bits, so now we seem to allow op to 64k registers, do you foresee any obvious problems with that? If not we can put that in the PR
13:02:38FromDiscord_<sealmove> Here you go https://github.com/sealmove/nimitai/blob/master/nimitai/private/ksypeg.nim
13:03:45*uu91 quit (Read error: Connection reset by peer)
13:04:00*uu91 joined #nim
13:09:14clyybberAraq: Unifying pArg and p to a proc that only has one extra param is quite tricky, since f(g()) with no sink params must be transformed to f(var tmp; =sink(tmp, g()); tmp) finally: =destroy(tmp)
13:09:57clyybberAh, I have the solution. We handle the other case in moveOrCopy
13:10:30AraqZevv, it should be fine
13:11:00Araqthe VM uses a seq already for the register memory
13:11:21Araqso we don't waste anything more than we currently do, except that the VM instructions are bloated by a factor of 2
13:11:34Araqbut my suspicion is that the memory overhead is in the noise
13:11:50Araqwe lose all the RAM in the tree data structures anyway (ropes!)
13:12:00clyybberraptorjit wants to go 96bit
13:12:09FromGitter<alehander92> why not move to strings
13:12:13FromGitter<alehander92> cant be that hard
13:12:21Araqha!
13:12:25FromGitter<alehander92> :)
13:12:30Araqbut it is
13:12:35ZevvAraq: ok thanks
13:12:51FromGitter<alehander92> but seriously
13:12:52Araqthe coding patterns are different, alehander92, you can't optimize FP code.
13:12:54FromGitter<alehander92> it cant be that hard
13:13:12FromGitter<alehander92> but my impression was
13:16:04FromGitter<alehander92> sorry
13:16:08FromGitter<alehander92> well i have to look at it
13:16:17Araqwell replacing Rope by 'string' is easy enough
13:16:23Araqbut then it's slower
13:16:31FromGitter<alehander92> yeah but
13:16:38FromGitter<alehander92> is that because of allocations of many small strings
13:17:30FromDiscord_<onelivesleft> Is there a way to fill out the tags for `nimble publish` automatically?
13:17:31FromGitter<alehander92> i sometimes build code generators which mostly add to the same big string
13:17:33madpropshmm lastWriteTime and lastAccessTime are giving me 1969 dates on some directories
13:17:35FromGitter<alehander92> i think
13:17:48FromGitter<alehander92> but with complicated code generation this is not easy i guess
13:17:55Araqhttps://github.com/nim-lang/Nim/blob/devel/lib/pure/json.nim#L735
13:17:57madpropsi guess directories don't have write date data?
13:18:06Araqthis is the pattern you need to adopt when you use 'string'
13:18:15*krux02 joined #nim
13:18:23Araqand you don't have to do this with ropes so we didn't
13:18:58FromGitter<alehander92> but is this a problem because of allocation?
13:19:03FromGitter<alehander92> or because of many copying
13:19:15FromGitter<alehander92> i havent profiled enough such stuff
13:19:17Araqit's a different big O thing
13:19:42Araqit's O(1) vs O(n) and then inside loops it becomes O(1) vs O(n*m)
13:19:47Araqor with recursion
13:19:57FromGitter<alehander92> but i wonder what is O(n) about strings except for copying
13:20:06Araqthe copying.
13:20:10FromGitter<alehander92> yes
13:20:11FromGitter<alehander92> ok
13:20:25FromGitter<alehander92> but we can still
13:20:33FromGitter<alehander92> there are copy ellision
13:20:47FromGitter<alehander92> a() & "e"
13:21:11FromGitter<alehander92> and you can e.g. maybe just directly fill the result of `a` into the result of the expression
13:21:17FromGitter<alehander92> but i guess this is not general enough
13:21:48Araqwell that's what I said, to "fill the result directly" you switch over to 'result: var string'
13:22:13Araqdifferent coding style. to be applied to over 4000 lines of code, enjoy
13:22:13FromGitter<alehander92> ah but what i imagined was that
13:22:19FromGitter<alehander92> you can do it automatically
13:22:30FromGitter<alehander92> like,
13:22:37FromGitter<alehander92> the function returns string
13:22:38Zevvsealmove: to get rid of your globals, read https://github.com/zevv/npeg#generic-pegs-and-passing-state
13:22:56FromGitter<alehander92> and the compiler auto-passes the target expression as var result if
13:23:02FromGitter<alehander92> this would be ok
13:23:07*Araq sighs
13:23:45Araqthe return value doesn't have append semantics
13:24:00Araqthe transformation is hard
13:24:00FromGitter<alehander92> well you know what i mean is it called rvo
13:24:32Araqno it's not "rvo", you call it this way because everything in your head is conflated into a soup
13:24:33FromGitter<alehander92> but if you have a() & b() the compiler can do
13:24:45*uu91 quit (Read error: Connection reset by peer)
13:25:03*uu91 joined #nim
13:25:18FromGitter<alehander92> expression : string ⏎ a(pass expression) ⏎ b(pass expression from last index) etc
13:25:45FromGitter<alehander92> its more like string slices ok
13:25:57FromGitter<alehander92> yeah i completely agree about the soup, thats why i am asking
13:26:32Araqno it's not "like string slices" either, string construction has nothing to do with slicing which is about cheap read-only views into the string's insides
13:26:41FromGitter<alehander92> but what i mean is
13:26:52FromGitter<alehander92> you pass a writeable slice of the string to b
13:27:06FromGitter<alehander92> so it still fills it in the usual way
13:27:10FromGitter<alehander92> without it needing to know of append
13:27:39Araqso ... it's a more abstract string which supports easier appends.
13:27:45Araqlike a rope?
13:28:08FromGitter<alehander92> well its not a tree
13:28:31FromGitter<alehander92> so its linear in memory
13:29:08Araqanyway, to make it short: it's hard when you consider all the details and it looks easy when you take the shortcuts and avoid thinking about the details. as usual.
13:29:16FromGitter<alehander92> i know that
13:29:20FromGitter<alehander92> but maybe then a Rope with different underlying implementation
13:29:35FromGitter<alehander92> which doesnt allocate so much/uses indexes instead of ptr
13:29:38FromGitter<alehander92> (ref)
13:29:46FromGitter<alehander92> might be ok
13:30:10clyybberwhat differntiates and index from a ptr??
13:30:34clyybberAnd index is just a ptr + the root
13:30:35FromDiscord_<mratsim> an index is relative a pointer is absolute
13:31:04FromGitter<alehander92> well my point was more to not use `ref` and the heap so much
13:31:06Araqalehander42: that kind of works when your ropes API is not all about "immutable" stuff and "easy sharing"
13:31:13FromGitter<alehander92> like not in such nonlinear way
13:31:21Araqbut our ropes are hardcore FP
13:31:26FromGitter<alehander92> use the heap more linearly*
13:31:36Araqand implementation changes will be reflected in the API
13:31:41Araqor result in subtle bugs
13:32:09FromDiscord_<mratsim> FP requires excellent memory management / garbage collector. Though apparently it enables different garbage collection strategy (software transactional memory in Haskell)
13:32:12FromGitter<alehander92> how does rope exposes sharing
13:32:14*solitudesf quit (Quit: Leaving)
13:32:19FromGitter<alehander92> it should be impl detail imo
13:32:28FromGitter<alehander92> but yeah you can pass ropes
13:32:30FromGitter<alehander92> which are the same nvm
13:32:42Araqand to make things worse: once you did that the codegen is still wrong
13:32:57Araqas it should produce a C/Nim AST anyway, not strings
13:33:26FromGitter<alehander92> yeah that doesnt seem so hard
13:33:41Araqit's lipstick on a pig, ropes and strings in general are the wrong idea for Nim's codegen
13:33:50FromGitter<alehander92> but i admit obviously as a person who tried to do that with jsgen: it seems to be not hard, but its just much work
13:33:58FromGitter<alehander92> one needs 1-2 weeks probably
13:34:18Araqwant to optimize '&*x' to 'x'? good luck, you can't, it ended up in a rope
13:34:32FromGitter<alehander92> ok, but what would you use for the final C ast -> code gen
13:34:58Araqfor the final step a simple stream will do
13:35:08FromGitter<alehander92> i wonder if C ast is the best repr or maybe some kind of low lever three address IR
13:35:10Araqyou walk the tree and output to a stream directly
13:35:11clyybberCompile to GENERIC directly
13:35:16FromGitter<alehander92> as maybe the second one can be reused to llvm etc
13:36:17FromGitter<alehander92> ok
13:36:24FromDiscord_<mratsim> C is not the best repr but, you usually don't have to worry about the backend
13:36:35FromDiscord_<mratsim> though I manage to trigger a GCC bugs on wednesday
13:36:39FromGitter<alehander92> well many possible ideas, i still want to see parallel incremental nim one day too
13:37:31FromGitter<alehander92> mratsim yeah clang
13:37:41FromGitter<alehander92> and gcc are most popular, right?
13:38:11FromGitter<alehander92> which one do you prefer
13:38:17Araqsame stuff with LLVM, in the end you do stream into a more linear representation but it helps if it's not an untyped string
13:39:21Araqmratsim: https://github.com/nim-lang/Nim/issues/12722
13:39:22disbot➥ Sink regression on devel: cannot compile the generated C code anymore @11mratsim; snippet at https://play.nim-lang.org/#ix=22P4 😏
13:41:21FromGitter<alehander92> Araq yes
13:41:45FromDiscord_<mratsim> yeah I know, but I wanted to finish my fight against multithreading and memory bugs
13:42:07FromDiscord_<mratsim> I would love to set breakpoints in VS code
13:43:19*solitudesf joined #nim
13:47:16Araqalehander92: there is a different way to optimize ropes though, you can introduce special pattern rope kinds
13:47:30Araqlike $#($#, $#, $#)
13:47:40Araqthen you save the '(' and ',' tokens
13:47:56Araqand C function calls take up less memory
13:48:16FromGitter<alehander92> but what is the bottleneck with ropes currently
13:48:23Araqdo the same for dot expressions and [] expression and assignments and you cover 80-90% of the produced code
13:48:27FromGitter<alehander92> is it that the gc cant clean them up well?
13:48:34FromGitter<alehander92> or just many pointers
13:48:45Araqmany many allocations
13:49:03Araqthe GC can clean up but a GC cycle is not free
13:49:51PMunchHmm, what is the right way to specify GetLastError for a cimport. My code work fine on x86_64, but when trying to cross-compile to 32 bit I get this error: http://ix.io/2376
13:50:35AraqPMunch, give your importc the right calling convention, 'stdcall'
13:50:52Araqand maybe you also want to use 'nodecl' or 'header'
13:52:06FromGitter<alehander92> Araq
13:52:49FromGitter<alehander92> that sounds smart yeah i am not sure if i get it but i imagine you can just directly input all the args in one rope "object" instead of many subnodes
13:53:15FromGitter<alehander92> but ugh yeah sharing
13:53:37FromGitter<alehander92> but does the compiler expect that if you pass a rope `a` somewhere
13:53:38FromGitter<alehander92> and change it
13:54:01FromGitter<alehander92> all the rope "consumers" of it should be also changed
13:54:32PMunchAh thanks Araq, that worked
13:54:33FromGitter<alehander92> isn't it kinda ok to just fix those cases and still introduce a linear non-sharing impl
13:54:43FromGitter<alehander92> 1) expect info from the test suite
13:55:39FromDiscord_<mratsim> @Araq, I'm not sure I will be able to get a minimal example. Seemingly unrelated changes (no change on the type, on the proc, or even in the file they are living in) in some of my latest commits made the problem disappear.
13:55:39FromDiscord_<mratsim>
13:55:39FromDiscord_<mratsim> I'm dropping it lower priority, I expect someone will hit it again at one point.
13:56:08Araqok, well I'm working on more destructor bugfixes anyway
13:56:20Araqand maybe you triggered:
13:56:33Araqhttps://github.com/nim-lang/Nim/pull/12770
13:56:34disbot➥ fixes #12766 @11Araq
13:56:53Araqthinking about it, it looks like this is your bug too?
13:57:03Araqwild sharing because of proc inlining
13:59:26FromDiscord_<mratsim> heh yes good catch
13:59:37FromDiscord_<mratsim> removing inline removed the bug
14:00:02FromDiscord_<mratsim> but I didn't change them in my latest commit
14:01:10FromDiscord_<mratsim> I'll test with the PR
14:02:58*solitudesf quit (Remote host closed the connection)
14:03:57FromDiscord_<mratsim> yes the PR fixes my bug as well
14:04:02*solitudesf joined #nim
14:06:42clyybberAraq: Huh, good job catching that bug
14:16:08*tane joined #nim
14:17:03*nsf joined #nim
14:19:59disruptekshashlick: #12768 looks perfect to me. nice job. 👍
14:20:00disbothttps://github.com/nim-lang/Nim/pull/12768 -- 6Fix #12767 @11genotrance
14:22:37*thomasross quit (Ping timeout: 240 seconds)
14:25:04FromDiscord_<mratsim> Traitor spotted: https://github.com/narimiran/AdventOfCode2019
14:25:14narimiran:P
14:25:30Zevvthere you go narimiran, told you so!
14:25:42Zevv"traitor", that's how it begins
14:25:53narimiranZevv: be quiet there!
14:26:15Zevvwelcome to hotel california
14:26:39disruptek!repo disruptek/bump
14:26:40disbothttps://github.com/disruptek/bump -- 11a tiny tool to bump nimble versions 8 4⭐15 0🍴
14:27:03tanenarimiran, did you settle for `core`?
14:27:32narimirantane: i'll be using ocaml's stdlib + containers, if that's what you're asking
14:27:38taneyep
14:27:42tanecontainers is good too
14:28:01tanethe `gen` package too
14:28:04narimirani like it very much as it doesn't hijack stdlib like janestreet stuff does
14:28:11*Vladar quit (Quit: Leaving)
14:28:32tanewell, I'd `open Containers` too
14:28:52narimiran(should we go to offtopic/private?)
14:29:00FromDiscord_<Rika> yes
14:29:16taneis there an OT channel?
14:29:24narimiranyeah, #nim-offtopic
14:29:27narimiransee you there :)
14:30:22PMunchThat being said it is normally fine to be off-topic here, as long as no-one wants to talk on topic
14:49:40*Hideki_ joined #nim
14:53:08*PMunch quit (Quit: Leaving)
14:55:04*ng0_ joined #nim
14:57:40*ng0 quit (Ping timeout: 260 seconds)
15:09:29*dddddd joined #nim
15:11:58*ng0_ is now known as ng0
15:13:37FromDiscord_<Bub_Lite_63_Jr> Is there a different command for updating Nim than the normal install command? I read that Nim 1.0.4 is out, but using the normal install command doesn't seem to update my Nim from version 1.0.2
15:13:43FromDiscord_<Bub_Lite_63_Jr> ```
15:13:43FromDiscord_<Bub_Lite_63_Jr> curl https://nim-lang.org/choosenim/init.sh -sSf | sh
15:13:43FromDiscord_<Bub_Lite_63_Jr> ```
15:19:46FromDiscord_<Rika> @Bub_Lite_63_Jr `choosenim update stable`
15:21:16*Kaivo quit (Quit: WeeChat 2.6)
15:21:36*Kaivo joined #nim
15:24:16*MarquisdeFalbala joined #nim
15:26:15FromGitter<deech> Am I correct in conceptualizing `mixin` a simple dynamically scoped value?
15:26:18FromDiscord_<Bub_Lite_63_Jr> Thank you @Rika ! I will give it a shot.
15:26:25FromDiscord_<Bub_Lite_63_Jr> Seems to be working!
15:27:23Araqdeech: it's comparable but the "dynamic" has nothing to do with "at runtime"
15:28:19FromDiscord_<Bub_Lite_63_Jr> Thank you!
15:28:40FromGitter<deech> Araq, right I meant more like https://en.wikipedia.org/wiki/Scope_(computer_science)#Dynamic_scoping
15:31:03narimiranBub_Lite_63_jr where did you read that nim 1.0.4 is out, that didn't also state how to update?
15:39:05*uu91 quit (Read error: Connection reset by peer)
15:39:20*uu91 joined #nim
15:43:55disruptekthe problem with package managers modifying a git-tracked nim.cfg willy-nilly is that now lockfiles cannot distinguish between meaningful changes to the repo.
15:44:46disrupteki guess i can parse a diff, but... ugh.
15:45:59disruptekmaybe we just create our own branch for each lock.
15:47:19Araqwell we don't have lockfiles yet
15:47:22disruptekthis feels like the right solution. anyone have a thought?
15:47:48Araqmaybe you argue for *not* submitting the nim.cfg to the git repo
15:47:49disrupteki'll have them by the end of the day. 😉
15:48:14Araqand yeah you're onto something here
15:48:42Araqmaybe like this:
15:48:55FromDiscord_<Rika> @Bub_Lite_63_Jr, someone mentioned you above
15:49:12Araqthe nim.cfg contains a new @include "nimblepaths.cfg"
15:49:20Araqand nimblepaths is never checked in
15:49:35Araqand if the file doesn't exist @include doesn't complain either
15:49:42disruptekis @include a thing?
15:49:50Araqand then these --paths can also simply be absolute paths
15:50:09Araqand then your PM can ignore the upcoming nimble "modecfg"
15:50:22Araqdisruptek, no but we could add it
15:50:26disruptekoh believe me, i'm already ignoring it quite happily.
15:50:31disruptekhave you tried nimph?
15:52:14Araqno, sorry, I'm busy with Nim core development
15:52:33clyybberAraq: How does the new fix on your PR work?
15:52:47Araqclyybber, link?
15:52:47shashlickThis same $nimbleDir thing can be used for lock files
15:53:08clyybberAraq: https://github.com/nim-lang/Nim/pull/12770/commits/b1df02b245af7df0ff877916487867a1a0c5bff0
15:53:09disbot➥ fixes #12766 @11Araq
15:53:17shashlickAlready you are telling Nim the exact version of a package
15:53:49Araqshashlick, most of us don't enjoy commit hashes in nim.cfg files though so lockfiles should be separate
15:53:54shashlickWhat remains is for Nimble to create a new lock file to track the exact versions to download and install
15:54:08Araqbut please think about my @include idea
15:54:09*Romanson quit (Quit: Connection closed for inactivity)
15:54:14shashlickYa we've laid that idea to rest
15:54:26Araqit does give us lockfiles as a side effect?
15:54:45disruptekno, it's just a better way to scope site changes.
15:54:48disruptekcall it site.cfg.
15:54:50Araqanyhow, shashlick, I don't like that --path:$nimblepath/foobar is a gamble when you have 2 nimblepaths
15:54:59disruptekexactly.
15:55:13Araqand I'm sorry that I didn't understand it before
15:55:20shashlickWhen you want lock files, you no longer want any lazy path stuff
15:55:26Araqbut the idea is that Nimble tells Nim *precisely* the paths
15:55:31disruptekand if you have 6 nimblePaths, which you might if you're using 6 distributions, you have 6 new --path statements which point into 6 distributions where new versions could appear.
15:55:37shashlickWhich means clean nimble path and one nimble path
15:55:37disruptekit's nonsensical.
15:55:50Araqand so we must not guess in the compiler
15:55:55Araqwhich means absolute dirs
15:56:14Araqwhich then cannot checked in but so what, you gotta die one death
15:56:20disruptekyep.
15:56:40disrupteki only calculate the best substitution for check-in reasons; this would let me remove that.
15:56:43shashlickThere isn't any guessing when there's only one nimble path
15:56:59disruptekonly one nimblePath is silly.
15:57:21Araqyeah but we do support multiple nimble paths
15:57:21disruptekbut, you can impl your pm that way if you want. 🤷
15:57:25shashlickSo you are saying that lock files and lazy paths should coexist
15:57:32disruptekof course.
15:57:36shashlickWhat sense does that make
15:57:48disruptekthey provide two completely different features.
15:58:17Araqclyybber, we missed a[dependentVarHere] in the dfa :-(
15:58:32Araqthere are more bugs like the fact that we "move" integer variables around
15:58:47Araqbut if we do that, we can at least do it right :P
16:00:22disruptekalso, having site.cfg would let us evaluate only one project-level config which helps my overlay use for interpreting the env of dependencies.
16:01:05disrupteki'd be able to trust that a dependency nim.cfg was only going to affect the local dep itself.
16:01:49disruptekso it shouldn't be an include. it'll just be something that the pm implements.
16:02:00Araq@include can also simplify the nim.cfg logic inside the compiler. we only evaluate project.nim.cfg or $compiler/config/nim.cfg
16:02:19Araqand inside project.nim.cfg you have to include the default config if you need it
16:02:33Araqno more --skipProjectCfg command line options
16:02:42disruptekdoes project.nim.cfg have a special role?
16:03:00Araqwell it gets picked up when you do 'nim c project.nim'
16:03:49disruptekwe can't just impl this in pm because i want the compiler to read the same env. so, there's that.
16:03:54disruptekbut i'm not sure about @include yet.
16:04:28*tklohna quit (Ping timeout: 245 seconds)
16:06:54disruptek--outdir should get pathSub'd.
16:07:11Araqwhat's frustrating to me personally is that everything we end up with is more complex than what I simply do manually
16:07:24disruptekhence nimph.
16:07:50Araqwhich is a projects/ directory and then I can access dep A from module B via ../A/foo.nim
16:08:50Araqmuch of the complexity seems to come from pretending "what if X is not in ../X"
16:09:26Araqbut instead of simply enforcing this we come up with different, more complex rules
16:09:37disruptekdo you want me to screencast some nimph gynastics so you can watch it work?
16:09:45disruptekgymnastics, too
16:10:24Araqactually I was about to live code some solution for PMunch
16:10:48Araqso that let x {.codegenDecl.} = [1, 2, 3] produces C code that isn't stupid
16:19:33*solitudesf- joined #nim
16:19:46Zevv\o/
16:20:43*solitudesf quit (Ping timeout: 245 seconds)
16:20:51ZevvI do not understand azure output. My job is failed and yet it says "Hooray there are no test failures"
16:21:52*tyler569 joined #nim
16:22:04narimiranZevv: azure knows it is friday, so you can freely relax cause "there are no test failures"
16:22:32Zevvyeah, but in our country, if you break the build, you stay late
16:22:37Zevvand that's not something you want on a friday
16:22:49narimiranwe're not in your country
16:28:36*clyybber quit (Quit: WeeChat 2.6)
16:32:34*thomasross joined #nim
16:36:49*nsf quit (Quit: WeeChat 2.6)
16:41:00*letto quit (Quit: Konversation terminated!)
16:43:38*drewr quit (Ping timeout: 245 seconds)
16:52:33*Jjp137 quit (Read error: Connection reset by peer)
16:53:03*Jjp137 joined #nim
17:05:55FromGitter<alehander92> mm
17:06:10FromGitter<alehander92> zevv hope you fixed it
17:07:17*LargeEpsilon quit (Ping timeout: 246 seconds)
17:18:17*disruptek_ joined #nim
17:18:59*disbot_ joined #nim
17:19:03*disruptek quit (Ping timeout: 245 seconds)
17:19:09*disruptek_ is now known as disruptek
17:20:20*disbot quit (Ping timeout: 276 seconds)
17:21:17*disbot_ is now known as disbot
17:22:22*njoseph quit (Ping timeout: 265 seconds)
17:25:02*tane quit (Quit: Leaving)
17:30:39Zevvfixed what?
17:31:57FromGitter<alehander92> the build
17:32:21FromGitter<alehander92> ooh you're already on the road riding a huge bike right
17:32:38FromGitter<alehander92> sorry, couldnt imagine more stereotypical dutch friday evening thing
17:33:00Zevvno I'm having lasagna and wine
17:33:06Zevvvery italian and french
17:33:15Zevvoh the cheese is swiss I just noticed
17:33:44FromGitter<alehander92> nicee
17:33:45Zevvabout the build: I have no clue. narimiran sais its ok, it's just azure whining
17:33:58FromGitter<alehander92> ah i thoughts its work related
17:34:13FromGitter<alehander92> you should try musaka, its like the balkan lasagna
17:34:32narimiranmmm, musaka is great!
17:34:42FromGitter<Willyboar> miam, miam
17:34:44FromGitter<alehander92> i wondered if narimiran would know it
17:34:54FromGitter<alehander92> i thought croatia might have become very western
17:34:56FromGitter<alehander92> :P
17:35:02narimiranyep, we have it and we call it the same
17:35:25FromGitter<alehander92> i still like the cabbage foods even more
17:35:40narimiranwell, i live near italy so we have italian cuisine, but we also have balkan stuff here too
17:36:02FromGitter<alehander92> sour cabbage*
17:36:22FromGitter<alehander92> nice
17:36:37FromGitter<alehander92> i imagine the burek family of foods etc are there
17:36:48narimiransour cabbage — i'm not a fan of it, but it is one of the favourite foods of my girlfriend
17:37:23FromGitter<Willyboar> μουσακάς/mousaka is my favourite food
17:38:04*MarquisdeFalbala quit (Remote host closed the connection)
17:38:09narimiranZevv: now, on a serious note, if you go to azure logs and clik on the "Run CI . Failed" task you will see the log there
17:38:25narimiranZevv: `compiler/vmdef.nim(45, 15) Error: unhandled exception: value out of range`
17:39:10FromGitter<alehander92> @Willyboar you were from greece
17:39:12*rockcavera quit (Remote host closed the connection)
17:39:13FromGitter<Willyboar> @alehander92 maybe you have tried and https://en.wikipedia.org/wiki/Pastitsio
17:39:20FromGitter<alehander92> yeah, i like your food a lot
17:39:41FromGitter<alehander92> i think maybe, it sounds relevant but not sure
17:39:51FromGitter<alehander92> but i have to ask you this: my parents got me this jam of
17:40:30FromGitter<alehander92> um, walnuts
17:40:49FromGitter<alehander92> from greece, and its nice but sounds strange, do you guys have a lot of nutbased jams
17:46:28disruptekshashlick: i'd rather not have docs if the alternative is to be unable to parse nimgit2.nimble via nimscript without determining its requirements first.
17:46:37disruptekchicken and egg.
17:48:32*Trustable quit (Remote host closed the connection)
17:51:04shashlickI understand, we need a better was to do that
17:51:24shashlickBut is it failing in the compiler Nim config parser?
17:51:47shashlickWhy is that gorge not working as expected?
17:51:59disrupteknot the config parser, but in `nim e`.
17:52:11shashlickBecause the nimble command isn't aware of the nimbledir in use?
17:52:27disruptekit fails because i haven't installed nimterop yet.
17:52:35Zevvnarimiran: sure, thats the 32 bit version and that one I expected
17:52:38Zevvbut all the others?
17:52:46shashlickIs nimterop installed on your user nimble dir?
17:52:58narimirani've seen that one in 64-bit CI, iirc
17:53:03disruptekeven if i specify it as a dependency, unless my dependencies are ordered, i won't necessarily install nimterop first.
17:53:48disruptekonce i clone nimterop, it works fine of course.
17:54:05shashlickYa that's why it checks if it is installed, if not, it shouldn't import nimterop docs
17:54:17disrupteki'm trying to track down a weird path problem that i can only repro in a proj with lotsa deps. so i keep running into this.
17:54:26shashlickAre you setting nimble dir in the cfg
17:54:45disrupteki pass it to nimble and i set it in the $env.
17:56:01disruptekthe problem is that your call to nimble finds nimterop (because it doesn't look at the env? or it's cumulative?) but nimscript doesn't.
17:57:01disrupteki guess it's time to finally nuke .nimble.
18:03:20*nsf joined #nim
18:04:12disruptekif nimble cannot create ~/.nimble/pkgs, it crashes. 🙄
18:07:08disruptek`mkdir -p ~/.nimble/pkgs; chmod 0 ~/.nimble/pkgs; chmod 100 ~/.nimble` seems to work for `dump`, at least.
18:10:50disruptekwow, that magically fixed my duplicate lazyPaths issue, too.
18:11:21FromGitter<alehander92> wow this almost convinced me that go error handling is good
18:11:32FromGitter<alehander92> for small cli tools
18:12:25FromGitter<alehander92> but yeah raise tags still should be similar as ergonomics
18:13:40disruptekproblem is that they don't work.
18:19:46FromGitter<alehander92> i thought they do
18:20:03disrupteknah.
18:20:06FromGitter<alehander92> my assumption was that without many stdlib annotations, they might not be so useful
18:20:10FromGitter<alehander92> but maybe i forgot how it works
18:20:29FromGitter<alehander92> i think i did use them
18:20:35FromGitter<alehander92> so you mean they dont work well
18:20:36FromGitter<alehander92> ?
18:21:06disruptekwell, there are a few corner cases (that i don't recall off the top of my head) where they don't work in normal code.
18:21:13disruptekthey don't work properly in async code.
18:21:30disruptekergo, they don't work well enough. no exception handling does, imo.
18:21:45disruptekso i almost never catch exceptions.
18:22:26*letto joined #nim
18:27:10*njoseph joined #nim
18:28:59FromGitter<zacharycarter> hrm - I guess nimterop throws up on unions?
18:30:15FromGitter<zacharycarter> I really want to love and use nimterop - but any time I stumble across an issue like this, I end up having to re-define every type is referenced in the problem type
18:30:24FromGitter<zacharycarter> so like type a has type b in it and type b has a problem
18:30:30FromGitter<zacharycarter> now I have to re-define type a and b
18:30:35*sagax quit (Ping timeout: 246 seconds)
18:30:39FromGitter<zacharycarter> and it just makes the whole process not worth it and then I go back to c2nim
18:31:24*sagax joined #nim
18:34:13*sagax quit (Client Quit)
18:34:42Zevvsealmove: this PR might prove tricky to get through
18:36:17sealmoveseems so, but can't complain, we are trying to tweak an important component of the compiler
18:37:05sealmovei still think fine-grained control through cli or nim cfg is a good idea...
18:37:15sealmovewith a good default of course
18:44:38lqdev[m]@zacharycarter not anymore, that's a thing of the past
18:44:54Zevvsealmove: could you answer krux02s last remark about VM exhaustion still happening with devel
18:44:59lqdev[m]shashlick has relatively recently added a patch that fixes thi
18:45:03lqdev[m]s/thi/this/
18:45:51FromGitter<zacharycarter> ah
18:45:56Zevvand maybe we should split stuff up. First make a PR with refactoring but keeping all the sizes the same
18:46:40sealmoveZevv: ok sure, I will. Give me a couple of hours because I must finish something else first.
18:46:52Zevvoh no hurries
18:56:33*dwdv joined #nim
18:59:03*narimiran quit (Ping timeout: 245 seconds)
19:00:22lbartHi,
19:00:22lbartNew with nim. I read a lot about this language and I like it. My subjects of interest are math, geometry, GIS, CAD.
19:00:22lbartUser and maintainer on FreeBSD , I'll try to port some software with nim to this system.
19:00:22lbartYou may see me soon for questions or improvements for BSD :)
19:00:22lbarteuantor if I'm not wrong you test nim on FreeBSD and OpenBSD?
19:00:32*adeohluwa joined #nim
19:00:54Araqhi lbart, welcome
19:01:03euantorlbart: Correct, I have a currently open PR to add CI support for both too
19:01:15euantorThe FreeBSD CI currently passes, the OpenBSD one does not
19:01:36euantorBut I hope to eventually get time to fix up the stdlib on OpenBSD and get it passing there too
19:01:49Araqeuantor, ping me when it's ready, I keep forgetting about most PRs and PR-friday only allows to catch up a little bit
19:02:02euantorAraq: https://github.com/nim-lang/Nim/pull/12179
19:02:03disbot➥ #12103 - CI for FreeBSD @11euantorano
19:02:11Araqis it ready?
19:02:19euantorFreeBSD CI is ready, the failures on it are on all other platforms except FreeBSD
19:02:40euantorMost recent build is here: https://builds.sr.ht/~araq/job/112890
19:03:28euantoroops, looks like a conflict since I last tested. I'll fix that up right now
19:03:40Araqso ... should I merge it or not?
19:04:21euantorYou won't be able to right now, but once I push this commit in 1 second you can :D
19:08:11euantorOk, it's running now. Just got to wait to ensure it passes still and hasn't broken since my last test
19:11:36*adeohluwa quit (Remote host closed the connection)
19:11:52*adeohluwa joined #nim
19:17:06*Ven`` joined #nim
19:18:29*Ven`` quit (Client Quit)
19:29:34lbartthanks
19:34:27*uu91 quit (Ping timeout: 250 seconds)
19:35:13*uu91 joined #nim
19:40:15*rayman22201 quit (Quit: Connection closed for inactivity)
19:46:45disruptekshashlick: i made a mistake. i wasn't running `dump` with $NIMBLE_DIR set.
19:49:25disruptekseems like it works fine. just another disruptek screwup. 😁
19:55:05*Ven`` joined #nim
19:57:20euantorAraq: Ready to merge https://github.com/nim-lang/Nim/pull/12179 - all checks are passing
19:57:21disbot➥ #12103 - CI for FreeBSD @11euantorano
19:58:45Araq.builds/freebsd.yml ? does this dot file crap ever end?
19:59:13euantorunfortunately not
20:00:11disruptekjust make a symlink to it so you remember it exists.
20:00:14euantorYou can submit builds via the API too, but for the GitHub integration they have to be in the .builds folder
20:01:14Araqmerged it
20:01:32Araqnow every other CI needs to support the .builds/ directory, so nice
20:01:52*Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
20:01:59euantorthanks, closed a few issues with it
20:02:41euantorbuilds.sr.ht supports a variety of Lines too, and OpenBSD. The OpenBSD CI uses the same service
20:26:23*adeohluwa quit (Remote host closed the connection)
20:35:08*tane joined #nim
20:42:30*zyklon joined #nim
20:43:47*uvegbot quit (Ping timeout: 250 seconds)
20:44:31lbartAraq: euantor FYI I added nim on https://wiki.freebsd.org/3rdPartySoftwareCI
20:51:00*thomasross_ joined #nim
20:51:00*thomasross quit (Killed (wolfe.freenode.net (Nickname regained by services)))
20:51:00*thomasross_ is now known as thomasross
20:51:30*rayman22201 joined #nim
20:56:22*thomasross quit (Ping timeout: 250 seconds)
21:02:10shashlick@zacharycarter ya cOverride is smarter now, thanks @lqdev
21:02:30*NimBot joined #nim
21:02:35shashlickUnions are supported
21:02:57shashlick@disruptek on good to know, but let me know if you do see other issues
21:03:14shashlickSeems like nimble should also parse the environment like nimph does
21:03:14disruptekso many issues.
21:04:16*thomasross joined #nim
21:16:26*adeohluwa joined #nim
21:28:42Zevvit's like my life
21:29:41disrupteki was gonna do lockfiles today but i'm mired in some nim.cfg bug.
21:31:12*narimiran joined #nim
21:31:21disruptekjust can't escape .nimble/pkgs no matter what i do.
21:32:37*adeohluwa quit (Ping timeout: 240 seconds)
21:40:37*nsf quit (Quit: WeeChat 2.6)
21:42:03*sagax joined #nim
21:42:52*Ven`` joined #nim
21:52:17*nif quit (Quit: ...)
21:52:27*nif joined #nim
21:53:24shashlickAlways standing on someone's shoulders
21:57:18disruptekbeats kneeling at their feet.
22:00:20disruptekall hail the mighty stacktrace.
22:04:23*narimiran quit (Ping timeout: 276 seconds)
22:14:10sealmoveZevv: do all registers have to fit in 16bit in order to make TInstr [0 .. 2^16] ?
22:22:20disrupteknah, we'll just have a bucket to catch any that overflow.
22:26:52sealmovedisruptek: I see, makes sense
22:26:58sealmovebuckets work nicely
22:28:16sealmoveZevv: sorry I didn't phrase it correctly. Can we keep a register 8bit and still make TInst [0 .. 2^16]?
22:28:29*sealmove quit (Quit: WeeChat 2.6)
22:36:17*MarquisdeFalbala joined #nim
22:37:32*solitudesf- quit (Ping timeout: 276 seconds)
22:38:59FromDiscord_<exelotl> wrote a cheeky nim script at work today
22:42:26FromDiscord_<exelotl> just needed a tool to remove some fields from certain json files. I gave docopt a try for specifying the arguments and holy crap it's nice
22:42:43FromDiscord_<exelotl> super impressive stuff
22:43:09disruptekdocopt?
22:43:18disruptek!repo docopt
22:43:19disbothttps://github.com/docopt/docopt.nim -- 11Command line arguments parser that will make you smile (port of docopt to Nim) 8 131⭐15 16🍴 7& 1 more...
22:43:56disruptekhow does it compare to cligen? any idea?
22:48:28krux02sealmove: do you have a link to you npeg so I can reproduce the register overuse?
22:49:56krux02Sorry for my words on github. I would like to reproduce the problem first and see if there isn't a trivial fix in npeg before we start to change the vm which will have an impact on everybody (potentionally a bad one)
22:51:25madpropsis there a way to have a proc that receives a list of strings of any size, without having to initialize seqs?
22:51:31madpropsof any number of items*
22:51:41disrupteklinked list.
22:52:17madpropsi just want to be able to send ["asdf", "xcvb"]
22:52:30madpropsbut in the proc the array[..] needs to know the length i think
22:52:52disruptekand why don't you want to use a seq?
22:53:02madpropsbecause then the user can't simply pass in a list argument
22:53:06madpropsthey need to init a seq and fill it
22:53:19disruptekyou mean @["like", "this"]?
22:53:46disruptekuse varargs, i guess.
22:53:47krux02madprops: sounds pretty much like ``proc foobar(args: openarray[string])
22:55:01krux02then you call call foobar like this: ``foobar(["asdf", "xcvb", "xvlc", "uiae"])``
22:55:14madpropsthat's cool
22:58:22madpropscan i init empty openarrays?
22:58:48krux02what you you mean?
22:58:58krux02you mean ``foobar([])``?
22:59:31krux02openarray is not a real type like array or seq, it is a special parameter type only.
22:59:36madpropsthing is i have a function wrapper, that sends all arguments to another function
22:59:43madpropsand it receives the same arguments
22:59:51krux02example please
22:59:52madpropsall of them have default values
22:59:57madpropssec
23:00:13krux02default value for openarray, never tried it, but why not should work
23:00:18madpropshttps://play.nim-lang.org/#ix=1LTW
23:00:33madpropswtf is that
23:00:38madpropsi didn't paste that lol
23:00:49madpropsbug..
23:00:57madpropsnvm
23:01:02krux02what did you paste?
23:01:08madpropshttps://play.nim-lang.org/#ix=1LTW
23:01:17madpropsthis is the link the playground gave me
23:01:19madpropsthe ix link
23:01:28madpropsbut it's showing another paste
23:02:00madpropshttps://play.nim-lang.org/#ix=239i this one works
23:02:10madpropsshould notify whoever does the playground
23:04:22krux02http://ix.io/239j
23:04:51disrupteklooks like it's working to me.
23:11:05madpropsyeah it works now
23:18:49*tane quit (Quit: Leaving)
23:28:00*luis_ joined #nim
23:38:04*MarquisdeFalbala quit (Remote host closed the connection)
23:41:47madpropsis break valid in case statements?
23:42:32*luis_ quit (Ping timeout: 276 seconds)
23:42:50disrupteknah.
23:47:50disruptekyou can name your block: and break blocks by name, which i use all the time.