<< 25-05-2018 >>
00:06:45*dddddd quit (Remote host closed the connection)
00:10:15FromGitter<data-man> Is there an official Nim Telegram channel? I asked because https://github.com/VPashkov/awesome-nim/pull/29 was merged.
00:30:14*skellock joined #nim
00:33:25*skellock quit (Client Quit)
00:36:05leorizedata-man: I don't see it mentioned in https://nim-lang.org/community.html
00:39:30FromGitter<data-man> @leorize: Exactly! I'm sure that such PRs can only be done by @Araq and @dom.
00:40:04*SenasOzys__ joined #nim
00:40:35*SenasOzys quit (Ping timeout: 240 seconds)
00:40:47leorizeAFAIK all official real time communication channel is linked to IRC, so if telegram is also an official channel, it should be linked to IRC as well, but it doesn't seem to be the case
00:50:45FromGitter<gogolxdong> Is there any example code showing programming without GC？
00:53:01*Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
00:57:40*CodeVance joined #nim
01:04:11*cavariux joined #nim
01:07:28FromGitter<data-man> @gogolxdong: https://github.com/pascalmouret/assyria
01:09:43*cavariux left #nim ("WeeChat 2.1")
01:11:06FromGitter<data-man> https://hookrace.net/blog/nim-binary-size/
01:22:49*saml__ joined #nim
01:28:06*saml__ quit (Remote host closed the connection)
01:30:22*skrylar joined #nim
02:01:21*CodeVance quit (Quit: Leaving.)
02:03:36*tefter quit (Remote host closed the connection)
02:16:04FromGitter<kayabaNerve> How many people here would be interested in a job offer to code in Nim?
02:19:55*SenasOzys__ quit (Remote host closed the connection)
02:20:14FromGitter<gogolxdong> @data-man means have to count the allocate size on our own?
02:20:33FromGitter<gogolxdong> @Varriount Do you mean ``` ⏎ proc `==`*(x, y: MemSlice): bool = ⏎ ## Compare a pair of MemSlice for strict equality. ⏎ proc memcmp(a, b: pointer, n:int):int {.importc: "memcmp",header: "string.h".} ⏎ result = (x.size == y.size and memcmp(x.data, y.data, x.size) == 0) [https://gitter.im/nim-lang/Nim?at=5b0772f1160c567d16d67d5a]
02:20:42FromGitter<gogolxdong> ``````
02:21:00FromGitter<gogolxdong> ```code paste, see link``` ⏎ ⏎ in memfiles.nim? [https://gitter.im/nim-lang/Nim?at=5b07730b4eaffb692d5379ee]
02:27:22*yglukhov[i] quit (Read error: Connection reset by peer)
02:27:55*yglukhov[i] joined #nim
02:56:10*xkapastel joined #nim
03:06:05*gokr quit (Ping timeout: 245 seconds)
03:35:50FromGitter<gogolxdong> What will happen with --gc:none with acctually referenced to the GC'd memory?
03:41:19skrylari would assume dangerous things would happen; you'd be poking memory without altering the hidden reference counts/roots
03:53:20*endragor joined #nim
03:56:44TangerHey guys, I'm trying to use copyMem any I'm (unsurprisingly) hitting a SIGSEGV from it. https://pastebin.com/50fhpM2G <- Code. When I check the pointer addresses of source and destination in my example (using cast[uint32](pointer)), they look like valid allocated memory (like this: Dest: 3739193392 | Source: 3739168852). Any ideas?
04:02:08leorizeTanger: what is the definition of the source and destination type?
04:04:21*leorize quit (Quit: WeeChat 2.1)
04:05:20FromGitter<data-man> @Tanger: Is a platform x32?
04:07:49FromGitter<data-man> Use ```cast[ByteAddress]``` instead of ```cast[uint32]```
04:09:33FromGitter<gogolxdong> I got this error : ``` ⏎ lib/pure/lexbase.nim(156, 13) Error: VM is only allowed to 'cast' between integers of same size
04:10:05FromGitter<data-man> Yes.
04:11:02FromGitter<gogolxdong> Anyone knows this error?
04:13:22FromGitter<gogolxdong> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b078d61016ae21a4ce12e8e]
04:14:38FromGitter<gogolxdong> It used to compile before I updated nim.
04:16:54FromGitter<gogolxdong> and never used lexbase directly.
04:19:37shashlickdata-man, did you try the untar package? It crashes still on read Str
04:20:46shashlicknewTarFile("xyz.tar.gz").extract("directory")
04:21:23shashlickWorks fine on 0.18.0, crashes on #head
04:22:34FromGitter<data-man> @shashlick: No, I don't use it.
04:23:02shashlickThat's why I opened the bug
04:23:12shashlickChoosenim uses untar
04:23:52FromGitter<data-man> untar's implementation isn't very good. :)
04:23:52shashlickI'm working on choosenim features, it doesn't pass tests on #head since untar crashes in readStr
04:25:00shashlickhttps://pastebin.com/cRwjCAu8
04:28:15FromGitter<data-man> streams.nim(254) is ```result = newString(length).TaintedString``` for me.
04:29:27shashlickIt's before your last commit
04:30:14FromGitter<gogolxdong> Can I rollback nimcrypto' verion requirement?
04:30:16shashlickThe line addr(string(result[0]))
04:30:33shashlickShould become cstring(result)
04:31:27FromGitter<data-man> I figured, bug for readStr(0)
04:31:49FromGitter<gogolxdong> It requires nimcrypto>0.18.0 and `Error: undeclared field: 'bits'` for 0.18.0, while 0.18.1 devel `Error: VM is only allowed to 'cast' between integers of same size` , how to fix this?
04:31:52shashlickLine 256 and 263
04:33:10shashlickAlso the other procs I mentioned might need review
04:33:38shashlickNo more [x] address of strings
04:47:27FromGitter<data-man> @shashlick: Fixed now, I hope.
04:50:42Tangerdata-man: Oooh, thanks. I'll give that a try, should give me a better idea of what's happening
04:51:30TangerAlso data-man: 64 bit
04:51:34TangerOh yeah lol
04:51:38TangerSo dumb
04:51:44FromGitter<data-man> :)
04:58:39TangerYou legend data-man. That was the problem!
04:59:37FromGitter<data-man> lol You're embarrassing me.
05:05:21FromGitter<data-man> And it isn't true. ;)
05:15:10FromGitter<data-man> @shashlick: Can you provide a problematic tarball, please?
05:19:03FromGitter<gogolxdong> `Error: module names need to be unique per Nimble package; module clashes with /mnt/d/third_party/src/github.com/nim-lang/Nim/lib/pure/httpclient.nim`
05:19:36FromGitter<data-man> @gogolxdong: Sure.
05:21:52FromGitter<gogolxdong> which I have no idea where to check, does this mean I have a file named httpclient.nim?
05:22:55FromGitter<data-man> httpclient.nim is part of stdlib
05:23:50FromGitter<gogolxdong> yes ,does this mean I have a file named httpclient.nim clashes with stdlib.httpclient?
05:24:03FromGitter<data-man> Yes.
05:24:14FromGitter<gogolxdong> which I haven't
05:24:37FromGitter<data-man> locate "httpclient.nim"
05:32:54*nsf joined #nim
06:29:19*Guest46886 is now known as awal
06:34:59*couven92 joined #nim
06:36:28*PMunch joined #nim
06:36:46*sz0 quit (Quit: Connection closed for inactivity)
06:43:50FromGitter<gogolxdong> How to do with this ? `Error: 'spawn' takes a GC safe call expression`
06:43:54*gokr joined #nim
06:50:15FromGitter<gogolxdong> Can parseJson work at compile time?
06:51:42FromGitter<data-man> No.
06:55:54FromGitter<gogolxdong> I want to parse a json file.
06:56:29FromGitter<data-man> Why at compile time?
06:57:51FromGitter<gogolxdong> It's known at compile time, do I have to push it to runtime?
07:03:34FromGitter<data-man> How big is this file?
07:07:05*Electrux joined #nim
07:07:44Electruxhello is anyone here? i need a bit of help. i am trying to implement bubble sort but it is not sorting at all
07:07:45FromGitter<gogolxdong> It's mysql configuration , as mysql connection cannot be multiplex , I have to add parseFile() before everywhere opened a database connection instead of using const values.
07:08:16def-Electrux: hi. post the code on gist or pastebin
07:09:33Electruxdef-: https://pastebin.com/VB03PFuH
07:11:32Electruxthe values are never changed :/
07:11:42def-to me it looks like they're changing
07:11:48def-you just forgot the first value
07:11:57def-[5, 1, 2, 3, 6]
07:12:05FromGitter<data-man> @gogolxdong: Just load this file to const and parse this const at compile time.
07:13:21FromGitter<data-man> s/compile time/runtime :)
07:13:24Electruxoh... yes u are right! but the loops seem to be correct to me... why is first value not being modified?
07:15:14def-not sure, seems like you're not iterating over the correct elements
07:15:34def-this one should work, for reference: https://github.com/def-/nim-unsorted/blob/master/bubblesort.nim
07:15:35FromGitter<gogolxdong> It's global json variable which would be GC'ed.
07:15:48Electruxoh i see now... there definitely was wrong... the second loop must start from the minimum value, not from i
07:16:08Electruxthank u so much ^_^
07:17:04Electruxone more question... why can't i repr open array? it says: Error: interal error: genRepr()
07:22:28livcddom96: it's kinda tricky to comment on non-tech topics
07:23:39FromGitter<data-man> ```code paste, see link``` ⏎ ⏎ @Electrux: works for me on devel branch. [https://gitter.im/nim-lang/Nim?at=5b07b9fbba1a351a68c19b12]
07:25:30*yglukhov[i] quit (Remote host closed the connection)
07:26:05*yglukhov[i] joined #nim
07:27:01*leorize joined #nim
07:30:05*yglukhov[i] quit (Ping timeout: 240 seconds)
07:30:40hohlerdedom96: excellent job with the forum!
07:34:54*Vladar joined #nim
07:41:49Electruxso is stable branch not good?
07:42:10Electruxi mean is it missing something?
07:43:51PMunchWell, stable branch is stable
07:44:11PMunchBut there are always some bugs or missing features on it with the pace that Nim is developing
07:45:30*xkapastel quit (Quit: Connection closed for inactivity)
07:45:45PMunchWhen 1.0 drops bugs will start to be back-ported to that version. So if a bug is discovered and fixed it will also be backported into the stable version, but new features will wait until another minor or major version.
07:49:15FromGitter<data-man> @Electrux: A bugs exist always and everywhere. Even a gods are doing them. :-D
07:50:42*dddddd joined #nim
07:51:56PMunchdata-man, made another Nim org?
07:52:57FromGitter<data-man> Yeah. There are big plans. :)
07:53:52PMunchHaha :)
07:54:44PMunchAh, the way Nim wraps C just makes it so nice to use: http://ix.io/1bkM/Nim
07:54:50Electruxok it's ok... i was just wondering... i find nim to be a pretty cool language so would really like to use it as a primary language
07:55:22PMunchGot the roller-mouse I'm playing around with at work to scroll when I move the roller :)
07:55:28Electruxdata-man: thanks a lot :D
07:55:39Electruxshould i use the devel version?
07:55:47PMunchI think most people do
07:56:02PMunchI try to use stable when I'm writing a library though
07:56:30PMunchBut I like to push things to the limit so I often end up having to jump to devel anyways :P
07:56:37Electruxlol
07:56:39Electrux:D
07:56:45PMunchJust install choosenim, then you can jump between versions easily
07:57:03Electrux:O i used homebrew... i guess choosenim would actually be better
07:57:15PMunchAnd update to the newest stable/devel with ease as well
07:57:32PMunchUhm, what does your "nim --version" say?
07:57:53PMunchNot sure about homebrew, but many package managers out there have really outdated Nim packages..
07:57:54Electruxit is 0.18.0 stable
07:57:59PMunchAh okay :)
08:00:41Electruxi have been looking for an interesting but not too verbose programming language... i, for some reason, do not like "popular" languages... so i guess nim fits the bill perfectly... is it usable enough for daily use? and would nim with lua scripting be good? i m thinking of learning lua as scripting language
08:01:36PMunchI've used Nim daily for years now, perfectly fine
08:01:52PMunchYou will however find yourself without a library from time to time
08:02:46PMunchNever really got into Lua, so not sure how good it would be with Nim
08:03:03*floppydh joined #nim
08:03:56Electruxnim is able to use C libraries right?
08:04:02FromGitter<kayabaNerve> Yes
08:04:30PMunchYeah, Nim compiles down to C, so calling C is super easy
08:04:46Electruxthen i guess missing library should not be that big a problem... hopefully... i just started with nim yesterday haha :)
08:05:06PMunchIf you look at the snippet I just posted: http://ix.io/1bkM/Nim you can see those top XTest procedure definitions are actually imported from a C library
08:05:47PMunchThe x and xlib import all the way at the top are also just thin pre-made wrappers around their corresponding C library
08:06:09Electruxoh yes! that code looks pretty neat btw! :)
08:06:23hohlerdeElectrux: nim is perfectly fine for daily use, however, if you use the devel branch and build it on your own, you might encounter breaking changes sometimes. The impact of those are not that big, though.
08:06:27FromGitter<data-man> @Electrux: There is https://github.com/jangko/nimLUA
08:07:20PMunchElectrux, haha thanks. It's a bit heavy on the discards, should probably turn those into proper errors with some sort of handling, but for what I'm using it for it doesn't matter much if it doesn't work 100% of the time
08:08:20PMunchhohlerde, yeah I wouldn't worry too much about breaking changes. Even libraries I wrote a year or more ago only need a thing or two changed to work fine with latest Nim.
08:08:31Electruxhohlerde: oh... well i guess considering that this language is not that mature ( i guess? ) i can expect some major changes... thanks for warning me :)
08:08:34PMunchAnd mostly just minor stuff as well
08:09:13ElectruxPMunch: ooh... well what are u using it for?
08:09:33PMunchAll sorts of things really
08:10:25Electruxinteresting :D
08:10:30PMunchStarted doing university projects in it, then I went on to do my master thesis in it, and of course now I've written a couple small and large libraries, a game, and a utility program here and there
08:10:46PMunchJust the other day I wrote a generator for some binary test data for work
08:11:47Electruxok idk what is this about... https://imgur.com/a/cGgSIO7 i m now on development branch
08:12:00Electruxwow a game?!!
08:12:02hohlerdewrote an atm driver in nim using libusb and made heavy use of OOP. works really well.
08:12:10PMunchhttps://github.com/PMunch?tab=repositories
08:12:13Electruxmay i have a link to those things if they are open source please?
08:12:15Electruxoh thanks :D
08:12:51PMunchThe game was just a fairly short game-jam game
08:13:03PMunchAn atm driver?
08:13:35hohlerdeyep
08:13:45PMunchhttps://pmunch.itch.io/hold-the-line
08:13:49PMunchWhat's that?
08:14:21hohlerdeit communicates with the firmware of the atm abd offers a restful interface to issue commands such as withdrawal or bank account statement
08:14:31livcdhohlerde: is it on github ? :)
08:14:49PMunchWait, an actual ATM?
08:15:00hohlerdenope, it is not open source, as it is used by the company I work for
08:15:07hohlerdeyes, an actual ATM
08:15:13PMunchHuh, cool
08:15:24PMunchBetter hope you don't have some strange bug then :P
08:15:28Electruxoh wow! u wrote a driver in Nim at production environment? that's awesome! :D
08:15:38hohlerdethe ProCash 1500 you see in some bank branches
08:15:57hohlerdethe driver is used for showcases, not for production in the wild
08:15:57PMunchIf someone enters ARAQ as their PIN it empties the vault
08:16:12PMunchAh right, makes sense
08:16:16PMunchStill cool though
08:16:34Electruxbtw cool repositories PMunch :D
08:16:52hohlerdeI just wanted to try out if nim is capable of doing that
08:16:55PMunchElectrux, haha thanks :)
08:16:57livcdPMunch: i cant start your game :S
08:17:15PMunchlivcd, from the binary on itch or from source?
08:17:19PMunchWin or Linux?
08:17:22livcdfrom the binary on windows
08:17:22PMunchErrors?
08:17:45livcdhttps://glot.io/snippets/f1cyngmfw5
08:18:52hohlerdeand Electrux, the nim community is really nice and helpful, if you encounter problems or have questions
08:19:11PMunchHmm, strange..
08:19:44PMunchTBH I never tried the Windows version myself. But all of my teammates were on Windows so it should work..
08:20:00livcddoh :D
08:20:31hohlerdein my beginner days with nim had an debugging session with araq on a sunday afternoon. he solved my problem and I really appreciated that he took his precious time to help me.
08:23:02*yglukhov[i] joined #nim
08:23:40Electruxthat's great! hohlerde :D
08:24:04PMunchlivcd, what kind of machine are you running on. Win10 I assume?
08:24:22*Vladar quit (Remote host closed the connection)
08:24:26Electruxcan anyone please tell me what this means? https://imgur.com/a/cGgSIO7 has anyone encountered this? ( it's ok if it's a bug... i just want to know )
08:24:43*Vladar joined #nim
08:25:08ElectruxPMunch: is your game available for Mac?
08:25:13livcdPMunch: yes win10
08:25:38*sendell joined #nim
08:25:39livcdPMunch: tried with the compatibility setting but did not work
08:27:21PMunchElectrux, well it requires SDL
08:27:30PMunchBut if you have SDL then it should work
08:27:52ElectruxSDL is available on Mac OS
08:28:01hohlerdeElectrux, have not encountered that
08:28:36PMunchlivcd, might be some issue with the amount of channels or the size of the buffer..
08:30:39Electruxhohlerde: ok it occurs when u pass a var openarray to a function and use repr in that function
08:54:53*arecaceae quit (Remote host closed the connection)
08:55:17*arecaceae joined #nim
08:57:03FromGitter<matrixbot> `monday smalltalk` anyone seen this error before: nimcache\stdlib_system.c:13746:16: error: '_O_BINARY' undeclared
09:01:37FromGitter<7sDream> I meet it when I use msys2 instand of gcc as compiler in windows
09:01:54FromGitter<7sDream> instand of mingw
09:04:07*yglukhov[i] quit (Remote host closed the connection)
09:04:40*yglukhov[i] joined #nim
09:04:54FromGitter<matrixbot> `monday smalltalk` that makes sense actually, thanks
09:05:16*yglukhov[i] quit (Remote host closed the connection)
09:05:30*yglukhov[i] joined #nim
09:06:53*yglukhov[i] quit (Remote host closed the connection)
09:07:28*yglukhov[i] joined #nim
09:07:45*yglukhov[i] quit (Remote host closed the connection)
09:07:59*yglukhov[i] joined #nim
09:08:13*yglukhov[i] quit (Remote host closed the connection)
09:08:49FromGitter<matrixbot> `monday smalltalk` is it documented that msys/cygwin doesnt work? I get it all sorted now though
09:09:06*xet7 joined #nim
09:19:43FromGitter<7sDream> install guide say that mingw works, but say nothing about other ⏎ https://nim-lang.org/install_windows.html#notes-about-compiler-dependencies
09:34:03FromGitter<mratsim> @skrylar, I’ve fixed https://github.com/mratsim/Arraymancer/issues/240
09:42:11skrylarhooray.
09:47:14FromGitter<mratsim> I’m pretty sure I copy pasted like a dumb machine :P
09:47:45FromGitter<mratsim> I forgot to remove the return value and to add the var parameter and obvviously no test
10:28:43PMunchHmm, this article someone linked to on the forum in the topic "Perfecting Nim" (or something like that), is really interesting: http://joeduffyblog.com/2016/02/07/the-error-model/
10:28:49PMunchAbout error models
10:29:29PMunchEspecially the require/ensure stuff and the abandon/fail differentiation is really interesting
10:31:28FromGitter<mratsim> @PMunch we will provide and use tools to implement thmost of the suggestion in this article as it is a big inspiration for how we will handle errors at Status
10:33:25FromGitter<data-man> @PMunch: https://github.com/arnetheduck/nim-result/pull/1
10:34:23PMunchmratsim, ooh interesting
10:34:31PMunchAll done through macros I assume
10:34:45PMunchIs there a way in Nim to disable exceptions by the way?
10:35:05FromGitter<data-man> If you missed this in the chat. :)
10:35:16PMunchAnd have the compiler stop you from calling anything that can throw one
10:36:21PMunchdata-man, what is that?
10:38:12FromGitter<data-man> Just read https://github.com/arnetheduck/nim-result/blob/a41157eda91bd1fff589262e4d82cbd0d0a1e0f4/result.nim
10:39:18PMunchYeah I'm reading that rightnow
10:40:41PMunchNot a huge fan of the return type errors though. Too much strain the the success-path of your program with all those checks..
10:41:59Araqyeah, I'm writing a blog post on why exceptions are superior and should be used...
10:42:06PMunchWhat would be great in Nim, since we have the powerful compile-time stuff, would be to have handler code injected at the fault site, instead of adding an extra branch.
10:42:32Araqjust look at the result.nim implementation
10:42:55Araqso it lifts '+' from the ground types to Result[]
10:43:18Araqwhat next? lift -, div, mod, ==, <=, plus every other operation that can be lifted?
10:43:26Araqthat would be *every* operation.
10:43:29Araqimpossible.
10:43:49Araqcontrast this with exceptions: There is no code to write, exceptions propagate.
10:44:13Araqit's like having a dedicated side channel for the error path that simply does the right thing.
10:44:29PMunchDepends on what you mean by the right thing..
10:44:38FromGitter<mratsim> @dom96 no way to change our profile picture in the forum at the moment?
10:44:55Araqit unwinds the call stack, the *same* thing that Result tries to emulate.
10:46:01euantorThe fact that Rust already has two different approaches to error handling is quite telling: https://github.com/rust-lang-nursery/error-chain and https://github.com/rust-lang-nursery/failure
10:46:13euantorAnd both are from `rust-lang-nursery`
10:46:47FromGitter<mratsim> Haskell too has both exceptions and Result/Either
10:46:57FromGitter<mratsim> Rust is inspired by Haskell ;)
10:47:26Araqand the star of reliability, Erlang, has exceptions too.
10:47:27FromGitter<mratsim> C++ has both exceptions and error codes
10:48:34FromGitter<dom96> @mratsim settings
10:48:35euantorI do like Rust's `try!` macro and `?` for when calling a method that returns a `Result`
10:48:49FromGitter<mratsim> as long as the solution pass my design smell detector `c = a + b` I’m fine with it. `check foo_add(c, a, b)` ist verboten
10:49:05FromGitter<dom96> @mratsim in your profile page. Why haven't you looked there?
10:49:39FromGitter<mratsim> (https://files.gitter.im/nim-lang/Nim/Rywd/2018-05-25_12-49-13.png)
10:49:52FromGitter<mratsim> tried clicking on the photo as well but nothing happens
10:50:10PMunchYeah, there's no change picture button there
10:50:12FromGitter<mratsim> aah it’s gravatar
10:50:47FromGitter<mratsim> ugh :/ that’s a pain.
10:53:13Araqthe only reason why exceptions are bad is because you can write open(); write(); close() solution? pack 'close' in a destructor (C++) or invent a {.finally.} pragma so that close can only be called in a 'finally' part
10:57:08Araqbut no, we don't need no stinkin' decent problem analysis, we can just hype some inferior FP inspired solution instead...
10:57:34PMunchA finally pragma would actually be kinda neat. But it hides the bigger problem that unchecked exceptions are just that, unchecked. If I have to go dig up some documentation or do some other magic to check if a procedure might randomly terminate my program I think that's room for improvement.
10:57:56Araqand solve the problem with MOAR boilerplate and brute force, that has always served us well
10:57:57PMunchWith return values and discard at least it's painfully obvious that I'm not doing something I possibly should do
10:59:17PMunchI agree that the way exceptions are handled is cleaner than return values, but I still don't feel like it's a very robust solution..
10:59:34Araqfeelings are irrelevant.
11:00:03PMunchs/feel/think
11:00:41PMunchOr rather s/feel like/think that/,
11:01:33FromGitter<alehander42> so why can't checked exceptions do the job ? people don't like the writing code overhead, but it might be less than error types
11:02:06FromGitter<alehander42> (I still think error results with some sugar for easily bubbling up / marking which calls can error can be also good, haven't looked much at nim-result)
11:02:52FromGitter<brechtm> Hi all
11:03:42PMunchAs he mentions in the article here: http://joeduffyblog.com/2016/02/07/the-error-model/#preconditions-and-postconditions by having ensures and requires they greatly reduces the number of exceptions
11:03:57PMunchI think that if that was the case then checked exceptions would be nice
11:04:00Electruxm not even sure if m qualified to say this... and no offense... it would be very nice if exceptions are used at a minimum... :/
11:04:07FromGitter<brechtm> Are nakefiles still Good Practise or should I use Nimble instead?
11:04:36PMunchHaving to catch an ArgumentException that you know can't happen because you are passing in a static integer is just annoying
11:05:11PMunchbrechtm, I think nimble replaced nake
11:05:32FromGitter<alehander42> ahh, so smarter inference of which exception paths are impossible for a callsite is needed
11:05:33FromGitter<data-man> @Araq @dom96 ⏎ Returning to the night question. ⏎ Is https://t.me/nim_lang the official channel? If so, then we need to add it to the site.
11:05:47FromGitter<alehander42> that might be hard
11:05:58zahary_At Status, we had some very long discussions about error handling after the "Perfecting Nim" post. We have an internal proposal for how to error handling in our own projects that is not so secret anymore. You can take a look here:
11:05:58zahary_https://gist.github.com/zah/d2d729b39d95a1dfedf8183ca35043b3
11:06:24zahary_I think it solves many of the raised concerns from both sides of the argument
11:06:25PMunchdata-man, it is the "official" yes. It was created before FOSDEM so we had a way to more easily co-ordinate while there
11:06:40FromGitter<brechtm> @PMunch I'm using this nakefile (https://github.com/yglukhov/nim-sdl-template/blob/master/nakefile.nim) that handles multi-platform (macOS, iOS, Android) builds. Should it be possible to do the same with Nimble?
11:06:48PMunchAs IRC isn't really all that great on mobile :P
11:07:11Araqzahary_, I'm completely sold on exceptions now fwiw. array's [] index operator doesn't return an Either
11:07:24Araqinteger operations don't produce an Either.
11:07:46Araqso you have to deal with these failures anyway
11:07:53Araqvia exception handling.
11:08:10PMunchI again kinda agree with the article again on that one, those should be hard errors, not something you can recover from..
11:08:34dom96huh, you've changed your mind about exceptions Araq?
11:08:41Araqnope, wrong, there are no hard unrecoverable errors.
11:08:58Araqthat's Erlang's key point.
11:09:12Araqyou let the *subsystem* fail, not the full program.
11:09:19PMunchHow often do you see: while true: try: echo myArray[i]; i+=1; except OutOfBoundsError: break
11:09:23Araqyou can map subsystems to processes
11:09:44Araqbut lacking lightweight processes, mapping them to proc calls is a good alternative
11:09:48FromGitter<brechtm> @PMunch: I've encountered such situations in Python
11:10:11Araqand why should subsystem arrangements depend on concurrency primitives anyway
11:10:12zahary_well, I think I'm covering these subleties in my gist (when a non-recoverable error can be recovered, why error code doesn't work well for this type of errors, etc)
11:10:19FromGitter<brechtm> (except StopIteration)
11:10:25Araqdom96, yeah...
11:10:28PMunchAraq, that's fair. Nim lacks the granularity that Erlang has in that regard, so failing the entire program might be a bit too harsh.
11:10:44dom96I'm a big proponent of exceptions too
11:11:09dom96As far as I'm concerned, Nim's exceptions with exception tracking are perfect.
11:11:24dom96The common downside quoted about exceptions is that they are not part of the function type
11:11:33dom96Well, exception tracking makes it part of the function type
11:11:54PMunchOh yeah, Nim exceptions are better than many other kinds of exceptions, that's for sure
11:12:38PMunchbrechtm, what kind of horrible Python code have you been reading?
11:12:49zahary_there are some flaws. a proper exception tracking system should still discrimate between recoverable and non-recoverable errors
11:12:57dom96zahary_: I think your proposal can be handled via exception tracking too.
11:13:01FromGitter<brechtm> @PMunch: writing!
11:13:06zahary_the reasoning is explained in my gist: https://gist.github.com/zah/d2d729b39d95a1dfedf8183ca35043b3
11:13:26zahary_dom, my proposal is based on the exception tracking mechanism with some features on top of it
11:14:13FromGitter<brechtm> @PMunch: though that was with iterators, not lists
11:17:40PMunchbrechtm, with iterators? Do they throw an exception when they are exhausted in Python?
11:18:11FromGitter<brechtm> @PMunch: yes, next(exhausted_iterator) throws StopIteration
11:18:13PMunchzahary_, don't have time to read your proposal atm, but I'm looking forward to it :)
11:18:27dom96At least Lobste.rs likes my post https://lobste.rs/s/6tougq/nimforum_lightweight_alternative
11:20:34FromGitter<alehander42> huh which community is lobste.rs :O looks interesting
11:21:57*gokr quit (Ping timeout: 240 seconds)
11:22:05FromGitter<alehander42> @zahary_ would the handleErrors be ergonomic enough to make e.g. try "useless" (work well even for simple try usecases)? If a user decides to adopt an error-handling strategy, he'd want to use it everywhere
11:22:50zahary_well, it will support as-expression and as-statement syntax, so indeed one can use it everywhere instead of try
11:24:06zahary_handleErrors is a codename chosen for clarity btw. so far 'rescue' (stolen from Ruby) is the favorite for the final name. other suggestions are welcome
11:25:28FromGitter<alehander42> isn't this a misleading name? in ruby one uses `begin rescue `, so I always translated `rescue` to `catch/except` in my head
11:30:47Araqzahary_, 'handleErrors' assumes that there are unrecoverable errors that are mapped to 'raise' anyway
11:31:13AraqI would argue that the true unrecoverable errors are mapped to quit()
11:31:23Araqand not of anyone's concern.
11:32:00Araqhowever, things we are currently unsure about like IndexError or OverflowError should be a properly recoverable error
11:32:05zahary_well, as you said, something non-recoverable errors can be handled at the subsystem level
11:32:11zahary_sometimes ..
11:32:37AraqIndexError and OverflowError turn Nim procs into total functions
11:32:49Araq(ignoring endless loops)
11:33:20Araqand that's a key ingredient of why this "in practice, I'm catching IndexError on this buggy subsystem" works so well
11:34:13zahary_Otherwise, in an RFC I would actually recommend having control about how certain errors are handled at the raise site by introducing a compile-time define for this. You say that certain exceptions should be handled with quiting
11:34:37Araqand with a safety zone in the allocator even "out of memory" situations can be handled
11:34:46zahary_The rationalle for this is enabling some optimizations in code that can assume no stack unwidning will be ever happening
11:35:20Araqthis is essentially what an OS does too. it quits a subsystem on OOM and yet can continue to work
11:36:35*gokr joined #nim
11:38:09Araqassuming no OS, "assume no stack unwinding" is pretty much impossible.
11:38:32Araqconceptually "quit" does unwind the stack too.
11:38:53zahary_sure, although I've discussed some alternative approaches here:
11:38:53zahary_https://gist.github.com/zah/d2d729b39d95a1dfedf8183ca35043b3#appendix-aborting-tasks-and-the-need-for-a-resource-supervisor
11:38:53zahary_You can have your own OS-like resources supervisor
11:39:42zahary_stack unwinding is one way to release resources, a supervisor is the other. The OS is a kind of supervisor
11:41:37*gokr quit (Ping timeout: 252 seconds)
11:42:13*Snircle joined #nim
11:42:53*natrys joined #nim
11:43:38Araqok, well, maybe outline how a supervisor would work in Nim user-land
11:44:21AraqI see that you cannot use 'quit' and are forced to use 'raise'
11:44:41FromGitter<brechtm> Does Nimble somehow support building for iOS?
11:44:52AraqI don't see why these unhandled exceptions that propagate into the type signature would hurt us
11:45:12zahary_I've mentioned one way - relying on the GC heap, which is walkable. Another way is to put a Context structure in a thread-local var and register all the resources there (as some kind of collection that can be walked when things go awry)
11:46:39Yardanico@brechtm what do you mean by this?
11:46:42Yardaniconimble is a package manager
11:48:04FromGitter<brechtm> @Yardanico I understand it is supposed to replace Nake. And I want to convert this nakefile (https://github.com/yglukhov/nim-sdl-template/blob/master/nakefile.nim) that can build my project for macOS, iOS and Android.
11:48:26zahary_because it's unreasonable to handle them - they are not expected to ever happen, so the software could not have a reasonable policy about how to handle them. the callers on the every level of the call graph would be annoyed for having to spend time thinking about these non-recoverable errors and figuring out that they should be propagated up
11:48:27FromGitter<brechtm> I suppose I could define a task that builds for IOS
11:48:36Yardanico@brechtm Nimble supports nimscript
11:48:40FromGitter<brechtm> But perhaps there are better ways
11:48:57shashlickdata-man you can test the nim-0.16.0.tar.gz posted in the download directory on the website
11:49:26Yardaniconimscript is basically a interpreter (well, virtual machine) used for running compile-time stuff in Nim compiler
11:49:39Yardanicobut it can be used separately (for nimble)
11:50:25Yardanico@brechtm most of the code in this nakefile can be run by nimble+nimscript easily without changes
11:51:01FromGitter<brechtm> @Yardanico can I call Nimble's build functionality in these tasks. There seems to be little documentation about these Nimble tasks.
11:51:20PMunchzahary_, yeah if we have checked exceptions we should also have require/ensure and a group of possibly unchecked exceptions that map to errors like div by zero, OOM, stack-overflow, etc.
11:52:20PMunchbrechtm, my idea was to use the nimble tasks and basically do what you now do through the nakefile
11:53:22FromGitter<brechtm> So that's pretty much a 1-to-1 conversion. Nimble doesn't really provide anything extra wrt to management of the iOS versions of the dependencies?
11:54:30shashlickdata-man: what's Nim compression
11:54:58PMunchbrechtm, doubtful
11:55:42FromGitter<data-man> @shashlick: New Nim's org for compression/decopression stuff. The big plans. :)
11:55:57*gokr joined #nim
11:58:46FromGitter<data-man> All interested persons are welcomed!
12:02:00*gokr quit (Ping timeout: 260 seconds)
12:04:28FromGitter<data-man> @shashlick: This archive contains the sources only?
12:04:57Araqzahary_: ok, so you argue you need to write 'raise' which should not result in a .raises annotation?
12:08:06zahary_I imagine two nice ways for this to work: 1) We can copy Java and introduce a base type for all non-recoverable errors. This exception may be implied by default from the higher-level `errors:` pragma in my proposal. 2) We may allow the user to specify the non-recoverable errors list per scope - errors can expand to something including the list specified in the current scope - system.nim provides a default
12:08:46*CodeVance joined #nim
12:10:54Araqproc hiddenRaise(e: ref Exception) {.raises: [], noreturn.} =
12:11:10Araq proc inner(e: ref Exception) = raise e
12:11:38Araq cast[proc (e: ref Exception) {.raises: [].}](inner)(e)
12:12:05Araqsolves the problem?
12:12:21Araqif you don't want to track unrecoverable errors, don't track them
12:13:00Araqit's pretty much what we do for system.`+`
12:13:01*elrood joined #nim
12:15:42zahary_well, we don't have control over built-in exceptions such as 'IndexError', 'OverflowError', etc. My reasoning is that they should be included in the list. We'll strive to use asserts for pretty much everything else
12:17:39*gokr joined #nim
12:18:56Electruxi literally cannot understand half of the second part of the nim guide.. :( ( the guide is good... just my fault probably )
12:19:16PMunchWhich guide?
12:19:18PMunchThe manual?
12:20:20Electruxhttps://nim-lang.org/docs/tut2.html
12:20:26Electruxthis one
12:20:49FromGitter<mratsim> same question about NimCompression :P
12:20:55shashlickdata-man: yes Nim sources and csources combined
12:22:50PMunchElectrux, maybe just need the manual?
12:22:55FromGitter<data-man> @shashlick: Just use zip files from github releases.
12:23:15PMunchBut to be fair, those are the advanced concepts. Might be good to master the basics first :)
12:23:18shashlickdata-man: is there a write up on the goals of Nim compression
12:24:06shashlickI've added code to download from GitHub if archive isn't on Nim Lang, but haven't pushed yet
12:24:26shashlickAlso, dom96 wanted to keep those archives as is for choosenim
12:26:56AraqElectrux: ignore the harder parts until you are ready for them. In fact, that's what most Nim starters do.
12:27:01FromGitter<data-man> @shashlick: Goals: wrappers for libarchive, zstd, lz4, etc. Common interface for these wrappers. Port some C libs to Nim. So far, that's enough.
12:27:05ElectruxPMunch yea, i guess i will practice the language before understanding all of the 2nd tutorial page :)
12:27:57ElectruxAraq: thanks... i will do that...
12:28:31Electruxcan someone give me suggestions as to what should i program with the language? practice problems perhaps... or something? ( hopefully enjoyable )
12:28:46Electruxthank you very much for all your help u guys :)
12:28:56FromGitter<mratsim> I find that the only way to learn is do something you want to do
12:29:09FromGitter<mratsim> otherwise it’s easy tog et discouraged.
12:29:28FromGitter<mratsim> say, you’re tired of the music players that exist, try doing one yourself.
12:29:58FromGitter<mratsim> or you want to automate stuff with your Raspberry Pi, well just do that
12:30:10Electruxi like making terminal based stuff... :D
12:30:19Electruxand perhaps delve into some GUI
12:30:20PMunchOther options include simple programming challenges such as advent of code
12:30:29PMunchAt least just to get your feet wet
12:30:35FromDiscord<r00ster> how could I raise an error myself with a custom message? because my library can be only used on certain OSes so I raise when its the wrong OS
12:30:42FromGitter<mratsim> To get my feet wet I use Project Euker
12:30:50FromGitter<mratsim> Euler*
12:31:02PMunchr00ster, raise newException(
12:31:11PMunchCrap, didn't mean to hit enter
12:31:30FromGitter<mratsim> when defined(windows): {.error: “Unsupported OS”.} —> error is compile-time
12:31:34PMunchraise newException(<your error type>, "This is the message")
12:31:39Electruxok then, i will try to make terminal based stuff and do some programming challenges... thanks a lot :D
12:31:51PMunchOh yeah, that's a much better solution :)
12:32:07FromDiscord<r00ster> oh ok thank you
12:32:10PMunchElectrux, no problem. And be sure to ask any questions you might have :)
12:32:33Electruxsure thing! thanks again ^_^
12:34:09shashlickdata-man ya sure I can help with wrappers
12:40:29FromGitter<data-man> @shashlick: Join to us! :)
12:50:34dom96data-man: Awesome organisation :)
12:50:56dom96Electrux: I always start by making IRC bots :)
12:53:47FromGitter<data-man> @dom96: Why you haven't joined? :(
12:54:13FromGitter<data-man> And Araq :(
12:54:32dom96because I don't check emails as soon as I wake up
12:54:35*gokr quit (Ping timeout: 252 seconds)
12:55:09FromGitter<data-man> Hmm, good morning? :)
12:58:00*couven92 quit (Read error: Connection reset by peer)
12:58:04dom96I wake up late too :P
12:58:11*pwntus quit (Remote host closed the connection)
13:00:38FromGitter<data-man> Move untar to the new org. ;)
13:01:51Yardanico@data-man I got an invite too :D
13:02:11FromGitter<data-man> Of course! :)
13:04:33FromGitter<data-man> More Nim-orgs, more globalization. This is the trend. :-D
13:05:11shashlickI like the idea
13:05:31shashlickThat way more people can check in code without having to wait for PR
13:05:45FromGitter<data-man> Yes.
13:06:24FromGitter<mratsim> You sound like a man with a plan
13:07:06FromGitter<mratsim> https://memegenerator.net/img/instances/68923504.jpg
13:08:41FromGitter<data-man> Also planned NimDB, NimGit, NimNet, NimAlgorithms, NimDataStructures, NimGraphics, NimSystem, NimMath, NimTinyTools, etc. :)
13:09:48Yardanicohmm, well, that's maybe useful, but still
13:10:21Yardanicoyou'll always need to be online to invite people to these orgs (so they can create/transfer their repos)
13:10:22dom96I don't think we need to do that...
13:10:25Yardanicoyeah
13:10:29FromGitter<brechtm> Is it possible to specify compiler options in a .nimble file?
13:11:56dom96use a .nims or .nim.cfg file for that
13:12:32dom96Here is an example: https://github.com/nim-lang/nimforum/blob/master/src/nim.cfg
13:12:54FromGitter<brechtm> thanks
13:12:55dom96you can also name it mainFile.nim.cfg and it will only apply when compiling `mainFile.nim`
13:19:15FromGitter<brechtm> @dom96 Does that allow for the options to be dependent on the task being executed?
13:19:39PMunchbrecdtm, if you are still looking at tasks the nimble file for that project has quite a lot of them: https://github.com/nim-lang/nimforum/blob/master/nimforum.nimble
13:20:12FromGitter<brechtm> @PMunch yeah... it's still all a little overwhelming :-)
13:20:53FromGitter<brechtm> I'm getting the feeling Nimble is not flexibly enough to the task of elegantly replacing my nakefile. At least if I want to use "nimble build" for creating the macOS binary (and tasks for iOS and Android targets).
13:22:08FromGitter<brechtm> Granted, that is a fairly non-standard task though.
13:22:26dom96nimble tasks can do everything that a nakefile can
13:22:45livcdi always read nakefile like nakedfile
13:23:40FromGitter<brechtm> @dom96 yes, but how to pass one set of arguments to nim c on build, and another set for a task?
13:23:56*yglukhov[i] joined #nim
13:24:29FromGitter<brechtm> "nimble c" calls "nim c", after resolving dependencies, right?
13:25:37dom96I'm not sure what you mean
13:25:52dom96Maybe you could show me your current nakefile?
13:26:45FromGitter<brechtm> @dom96 https://github.com/yglukhov/nim-sdl-template/blob/master/nakefile.nim
13:27:49FromGitter<brechtm> Perhaps I should not use separate tasks for each target (macOS, iOS, Android), but instead set a target variable or something?
13:28:11dom96s/runNim/exec "nim ..."/ ?
13:28:51dom96The only thing Nimble tasks need is access to any parameters passed to a task
13:28:57dom96But I don't think that nakefile is using this
13:29:51FromGitter<brechtm> @dom96 but how do I map "nimble build" do what's defined in defaultTask?
13:30:38dom96`nimble c`?
13:31:23FromGitter<brechtm> @dom96 I'm sorry, I'm not following. nimble c where?
13:32:01dom96I'm not sure what you mean to be honest
13:32:06dom96What's there to map?
13:32:22dom96You can literally execute the exact same thing this nake file is executing
13:33:12FromGitter<brechtm> By copying these tasks to the .nimble file. That, I understand. But how do I make "nimble build" execute defaultTask?
13:33:26dom96oh, that's what you want
13:33:28dom96You can't
13:33:31dom96You need to define your own task
13:33:39skrylarmratsim: so this facebook paper.. amusingly, they aren't even using RNNs at all. it's just linear layers and a gaussian attention network
13:34:01FromGitter<brechtm> @dom96 aha, that's what I thought :-/
13:34:33FromGitter<brechtm> @dom96 but nimble will still resolve any dependencies before a task is executed, I suppose?
13:34:46dom96brechtm: yes, but you should execute `nimble c` for that
13:35:05FromGitter<brechtm> @dom96 yup, I figured so much. Thanks!
13:35:21dom96brechtm: and not being able to override 'nimble build' is a Nimble issue, you can't do it with nakefiles either
13:36:14*Guest52326 quit (Quit: Discord > IRC)
13:36:26*SunDwarf joined #nim
13:37:19FromGitter<brechtm> @dom96 override 'nimble build' in a nakefile? Or something equivalent in nake?
13:38:16FromGitter<brechtm> Nake does have the notion of a default task which it executes on invoking nake without any arguments, no?
13:38:52dom96brechtm: but then how will Nimble users use it?
13:39:38dom96if you want a clever hack, you could create a Nim file that executed your default task and specify it in Nimble's `bin` (bin = @["mydefaulttask"]) :)
13:40:04dom96`nimble build` will build that binary, but not run it
13:40:17dom96you could however execute your default task at compile time with `staticExe`
13:40:20dom96*staticExec
13:40:23dom96but that might be going too far lol
13:40:55*CodeVance quit (Quit: Leaving.)
13:41:10*enthus1ast quit (Ping timeout: 264 seconds)
13:41:21*CodeVance joined #nim
13:41:45*xkapastel joined #nim
13:42:27FromGitter<brechtm> @dom96 It's a game, so it's not really of use to Nimble users. I only wanted to use Nimble for dependency management.
13:45:04dom96I would just create an `android`/`ios` task and instruct your users to run `nimble android` etc.
13:46:29PMunchBy the way brechtm, have you seen my SDLgamelib? It was born from the same repo as your nakefile
13:46:57FromGitter<brechtm> @PMunch no, but I will now :-)
13:47:44FromGitter<brechtm> @PMunch did you keep the multi-platform support?
13:50:01PMunchThink so
13:50:03FromGitter<Varriount> @data-man Love all the work you've been doing recently. I was especially amused by the recent "Use binary search everywhere".
13:50:44PMunchDon't see why it should be broken
13:50:56PMunchLast time I checked it was cross platform still :)
13:51:20FromGitter<brechtm> @PMunch I mean the building for iOS and Android, but since it's a Nimble package, I don't suppose so?
13:51:27shashlickdom96: you good with the choosenim and untar PRs?
13:52:37PMunchbrechtm, it's just a library to more easily use SDL
13:53:08PMunchIt's meant to be used _with_ nim-sdl-template, not _instead of_
13:54:25FromGitter<data-man> @Varriount: Big thanks! More amusing things are planned! :)
13:56:27Electruxhow would u reverse a string? does nim provide a built in function in some module?
13:58:17PMunchvar s = newString(oldString.len); for i in 0 .. oldString.high: newString[oldString.high - i] = oldString[i]
13:58:22Araqunicode.reverse
13:58:30PMunchYeah, or that
13:58:47PMunchThere's one it strutils too for pure ASCII no?
14:00:56shashlickdom96: I've implemented binary build installs, 64bit and direct download from GitHub and intend adding some test cases for choosenim before pushing
14:01:12Electruxoh... thanks :) i was gonna go for strutils.split, algorithm.reverse, and perhaps create a new string from that... but this makes it much easier... thanks again
14:02:04dom96shashlick: Untar - No, you need to investigate why it fails in 0.18.1 and not 0.18.0. Adding a try around this is just ignoring the problem.
14:02:09Araqand you all know what follows now...
14:02:15Araqhttps://nim-lang.org/docs/theindex.html search for 'reverse'
14:02:29Araqand you notice Araq is wrong and it's unicode.reversed()
14:02:50Electruxwow that's massive :O
14:03:01dom96shashlick: You're doing similar things in choosenim
14:04:55shashlickWell if readstr is returning a blank string, you have to handle it in choosenim
14:05:15FromGitter<krux02> yay, I have my first example of compiling Nim to opengl instractions plus glsl running.
14:05:36dom96shashlick: Added more comments on the PR
14:05:41FromGitter<krux02> meaning the first step is done so that Nim bocomes a hybrid GPU+CPU programming language
14:08:03Electruxbtw, do we not need null terminator in string?
14:09:16shashlickOk cool
14:09:56FromGitter<krux02> Electrux: The null terminator for strings is there just for C interop, nothing else
14:10:09Electruxoh ok
14:10:27Electruxso i don't need it if i don't explicitly interface with C code
14:11:06FromGitter<krux02> A lot of C APIs require strings to be null terminated. If Nim did not have the terminator, temporary string copies would be inevitable on each call
14:11:15FromGitter<krux02> yes
14:12:39FromGitter<krux02> It is not like Nim has no bugs. It can in fact happen that there is a nim library that relies on the null terminator, or you see the null terminator. But this should not be of interest to use until you actually face that bug
14:16:37*endragor quit (Remote host closed the connection)
14:17:03YardanicoElectrux in fact in devel branch of Nim you can't access null terminator anymore
14:17:35Yardanicowell, you can if you convert it to cstring (maybe not to cstring)
14:19:41FromGitter<data-man> Discovered bug (by scanf) in ```parseUntil*(s: string, token: var string, until: string; start = 0): int``` when ```until``` == ```nil``` ⏎ What this proc should return in this case? Currently fixed with result == 0. ⏎ There are other opinions?
14:20:58FromGitter<arnetheduck> @PMunch most languages distinguish between recoverable and non-recoverable errors, since ages past.. c with `abort`, c++ with `std::terminate`, java with `RuntimeException`, rust with `panic` etc.. the concept is hardly new, it's merely a matter of where you put the bar / what you make convenient for developers to use
14:21:17*nsf quit (Quit: WeeChat 2.1)
14:21:53PMunchHow do you mean RuntimeException relates to that?
14:26:37FromGitter<arnetheduck> `RuntimeException` was meant for unrecoverable stuff - the things that you would not include in your function signature (and therefore wasn't expected to handle - out of sight, out of mind).. only problem is that java checked exceptions suck, so everyone started using them for everything
14:27:01FromGitter<krux02> anyoen here doing game programming
14:27:19FromGitter<krux02> I have something to show
14:28:17FromGitter<krux02> http://ix.io/1bmy/nim
14:35:21*yglukhov[i] quit (Remote host closed the connection)
14:36:16FromGitter<krux02> that code I posted, I can compile it, and with macro magic the part that is written in the block under `framebuffer.renderDebug ...` is compiled to glsl including all the bridges to variables like M, V, P.
14:37:05FromGitter<arnetheduck> another way to look at it is that the fact that you *can* catch `RuntimeExceptions` de facto destroys the distinction between recoverable and unrecoverable errors - thus it's plausible to think that anything that opens a loophole for them in nim will be "abused", leading us back to where we are today, ie where function signatures don't teach you what can go wrong
14:37:33jxyIs there a way to know the details of the current memory pool? Like which file/line/statement allocated how much in the current pool at run time?
14:39:04Araqarnetheduck: yeah and my solution is to accept errors are recoverable.
14:39:04jxyor which object is holding refereces that cannot be GCed?
14:39:09euantorjxy: Try `dumpNumberOfInstances` and compile with `-d:nimTypeNames`
14:39:13euantorhttps://nim-lang.org/docs/gc.html
14:39:41Araqand unrecoverable errors call 'quit'
14:40:01Araqand 'quit' should never be used in a library...
14:40:06*endragor joined #nim
14:40:33jxyeuantor: thanks
14:41:09*yglukhov[i] joined #nim
14:43:32PMunchWoo, just checked my first Nim code into a work repository. Vive la revolution!
14:44:33*endragor quit (Ping timeout: 256 seconds)
14:48:36*PMunch quit (Quit: Leaving)
14:49:34jxyHow come '[GC] occupied memory' shows a much smaller number than '[Heap] total number of bytes'?
14:49:51*gokr joined #nim
14:50:58skrylarAraq, tell that to miguel de icaza
14:51:30skrylar"i can't find a file, instead of asking the process embedding me where it could possibly be i will just call abort()" -- Mono
14:51:54skrylarand this is the embedded api so there isn't the excuse of "you're being weird"
14:54:17*mostly-harmless joined #nim
14:59:01Araqjxy: the GC freed things?
14:59:19Araqusually the allocator doesn't return memory back to the OS.
15:05:22jxyI'm calling GC_fullcollect and echo GC_getStatistics() and dumpNumberOfInstances, in that sequence
15:05:26*BitPuffin joined #nim
15:05:53jxyshouldn't these two numbers agree? What's the difference between these two numbers?
15:06:09FromGitter<brechtm> I remember there being an online service like http://repl.it that also supported Nim, but I can't find it. Anyone?
15:11:26Yardanicoplay.nim-lang.org
15:11:31Yardanicoah, like repl.it ?
15:11:41Yardanicoyeah, I remember one which supported a lot of langs including Nim
15:12:29FromGitter<brechtm> Got it: https://glot.io/new/nim
15:12:36*yglukhov[i] quit (Remote host closed the connection)
15:13:00FromGitter<brechtm> Ah, but I didn't know about play.nim-lang.org
15:13:25shashlickdom96: can you also please review https://github.com/nim-lang/nimble/pull/464 if you have a chance?
15:14:56FromGitter<brechtm> play.nim-lang.org doesn't seem to be linked from nim-lang.org?
15:15:19*DarkArctic joined #nim
15:19:58Araqjxy: well one is about the heap's size, the other about the used memory
15:20:23Araqthere is allocated memory from the OS and then this is either "occupied" or "free"
15:22:44jxyok
15:23:33jxydoes the nim runtime uses anything like '[HEAP] ref cArray[system.char]'?
15:24:26Araq'cArray' ?
15:24:31Araqnever heard of it
15:25:54jxyright, that's my code doing
15:26:26jxyI need to figure out why they don't get GCed
15:40:29FromDiscord<treeform> Hey dom96, why in the untar package you have a zlib wrapper but statically link with zlib (on windows), while in the zip package you dynamically link with zlib?
15:40:32FromDiscord<treeform> https://github.com/dom96/untar/blob/master/src/untar/gzip.nim#L3
15:40:43*CodeVance quit (Read error: Connection reset by peer)
15:40:48FromDiscord<treeform> https://github.com/nim-lang/zip/blob/master/zip/zlib.nim#L5
15:41:20FromDiscord<treeform> I can't find a dll that works with nim on windows64. Why not always statically link on windows?
15:41:26FromDiscord<treeform> Why not always statically link?
15:42:53shashlickI think we need a static zlib wrapper which both libs can depend on
15:43:18shashlickcan go in data-man's NimCompression org
15:43:20FromDiscord<treeform> I would go as far as to say zlib needs to be part of standard library. So many things depend on it.
15:44:09FromDiscord<treeform> It only has snappy? https://github.com/NimCompression
15:46:04shashlickso far
15:47:16FromDiscord<r00ster> it is right that I dont actually need to use a `cint` for my C binding right? because it works surprisingly when im just declaring the argument as `int` instead of `cint`.
15:47:50*SenasOzys joined #nim
15:49:39FromDiscord<treeform> you might have issues with 32bit/64bit
15:50:07FromDiscord<treeform> are you on 32 bit platform?
15:51:01FromDiscord<r00ster> im on 64 bit
15:51:29FromDiscord<r00ster> will 32 bit users have problems? im not sure now if i should also change all the other ints to cints. In C the functions have the type `int`
15:51:51FromDiscord<r00ster> i mean the arguments*
15:54:07FromGitter<Varriount> r00ster: For Nim, `int` is 64 bits on a 64 bit system, and 32 bits on a 32 bit system. For C, `int` is 32 bits on both 32 and 64 bit systems.
15:54:53FromDiscord<treeform> So he should have issues now because he is on 64 bit? It should just work on 32 bit system?
15:56:09*Trustable joined #nim
15:56:10*Trustable_2 joined #nim
15:56:22*Trustable_2 quit (Remote host closed the connection)
15:56:35FromDiscord<r00ster> well I dont have issues currently when im using the functions where i changed cint to int. Works perfectly fine. The cint functions too. I just there need to cast to .cint when i want to use an int variable there for example
15:58:52*yglukhov[i] joined #nim
16:00:00*skrylar quit (Remote host closed the connection)
16:04:20FromGitter<Vindaar> I need to give a short talk on neural networks on Saturday. Thought I'd be a good idea to bring something like a live demo. So I wrote some code to train an MLP on MNIST w/ arraymancer. Then have that send test samples during training with the predictions and current accuracy to another program built with nim-plotly on the JS backend. That way one can watch the training on the fly. ⏎ There's a gif on the repo here:
16:04:20FromGitter... https://github.com/Vindaar/NeuralNetworkLiveDemo ⏎ everything's a little rushed right now, but it's still cool I think :)
16:05:20FromDiscord<treeform> that looks neat
16:07:09FromGitter<Vindaar> thanks!
16:08:16Electruxwhy did the creators of this language opt for camelCase function names? i am curious :)
16:09:14FromDiscord<treeform> probably because of pascal
16:10:11Electruxsorry i do not understand...? :/
16:11:25FromDiscord<r00ster> pascal is a very old programming language. it used camelCase and yea it looks like Nim got inspired by Pascal
16:12:04FromDiscord<r00ster> pascal is an old programming language. it used camelCase and yea it looks like Nim got inspired by Pascal
16:13:27*floppydh quit (Quit: WeeChat 2.1)
16:14:22*yglukhov[i] quit (Remote host closed the connection)
16:14:52Electruxoh
16:15:27Electruxok thanks for telling :)
16:16:09FromGitter<data-man> @Electrux: Pascal, evolution - https://en.wikipedia.org/wiki/Delphi_(IDE)
16:23:51Electruxhow do u make a 2d array from an already provided matrix of numbers?
16:24:16*yglukhov[i] joined #nim
16:24:39FromGitter<Vindaar> when you say "provided matrix of numbers", what do you mean exactly?
16:24:41Electruxis it just [ [ ... ], [ ... ], [ ... ] ... ]
16:25:19Electruxi mean i have hard coded matrix of numbers for which i would like to create a 2D array
16:25:54Electruxso would i just have to do something like: arr = [ [ 1, 2, 3 ], [ 4, 5, 6 ], [ 7, 8, 9 ] ]?
16:26:03FromGitter<Vindaar> yeah, you can do that
16:26:04*gokr left #nim (#nim)
16:26:23Electruxare there other ways too?
16:26:48Araqmacros
16:27:21FromGitter<brentp> I am getting weird errors with --threads:on and using spawn, flowvars. The errors are saying that things are nil that are not. When I use `gc:boehm`, these problems go away. This happens with devel. How could I debug this?
16:28:42*yglukhov[i] quit (Ping timeout: 252 seconds)
16:29:51Electruxoh... well i guess i am too beginner right now to start delving into macros
16:30:14*yglukhov[i] joined #nim
16:32:05*BitPuffin quit (Ping timeout: 260 seconds)
16:34:36*yglukhov[i] quit (Remote host closed the connection)
16:34:56jxyIs there a way to make Nim's memory management more verbose? I want to make it scream at me when it allocates and frees something.
16:44:32*yglukhov[i] joined #nim
16:46:01Electruxi can do this right? var x, y = getx(), gety()
16:46:20*sendell quit (Remote host closed the connection)
16:49:01*yglukhov[i] quit (Ping timeout: 245 seconds)
16:49:21YardanicoElectrux var (x, y) = (getx(), gety())
16:49:24Yardanicotry it this way
16:53:43Electruxthe first version is incorrect? or inconsistent? Yardanico
16:53:48Yardanicoincorrect
16:58:29Electruxoh... ok thanks a lot :d
16:58:31Electrux:d
16:58:33Electrux:D
17:14:21*yglukhov[i] joined #nim
17:18:27*yglukhov[i] quit (Ping timeout: 240 seconds)
17:19:11Araqbrentp: do you 'cast' thread local pointers?
17:19:34Araqjxy: compile with --gc:none and then the compiler reports allocations
17:21:22Electruxbtw, that neural network thing is so awesome!!!! @Vindaar :O
17:21:49FromGitter<brentp> @Araq I am doing something like this: https://forum.nim-lang.org/t/3854#23978 which I see someone has just warned me is unsafe.
17:22:13FromGitter<brentp> @Vindaar indeed very cool! glad to see nim-plotly getting used. :)
17:23:11FromGitter<Systra72_twitter> Hi! Trying to install Arraymancer on Windows using Msys2. Missing the libcrypto-1_1-x64.dll. Anyone know what to do?
17:24:25Araqbrentp: this will never work reliably with the default GC, stick to boehm for now :-)
17:24:55FromGitter<brentp> @Araq boehm is very slow for my workload.
17:25:27Araqwell then you need to redesign your program
17:25:44*athenot quit (Remote host closed the connection)
17:26:28*athenot joined #nim
17:27:01FromGitter<brentp> ok. how? this is my first try and parallel stuff in nim and I'm not sure what are the best things to try.
17:30:14FromGitter<brentp> I need the large seq[int8] but can modify the program structure
17:32:26*natrys quit (Quit: natrys)
17:34:26*yglukhov[i] joined #nim
17:38:44*yglukhov[i] quit (Ping timeout: 252 seconds)
17:48:17FromGitter<Vindaar> @brentp thank *you* for starting nim-plotly :)
17:51:14Araqbrentp: allocate the buffers in the main thread, don't realloc/add to them
17:51:24Araqand let the worker threads fill them.
18:01:37FromGitter<brentp> @Araq thanks. that's getting me further.
18:02:27FromGitter<data-man> @Araq: May I to add ```maxLen``` to parseOct and parseBin procs in parseutils? Tiny refactoring strutils.parseXXXint.
18:03:52Araqno.
18:04:18FromGitter<data-man> parseHex has maxLen.
18:04:54FromGitter<data-man> https://gist.github.com/data-man/f750df0593c71daacca1e5cb5c5b5573
18:04:56AraqI'm tired of this bullshit, just a couple more parameters here, just another default parameter there (splitLines with keepLineEndings because we cannot ever learn how parsers work)
18:05:39FromGitter<data-man> Just parsing duplicated in strutils.
18:06:11FromGitter<Quelklef> @data-man careful adding extra parameters, you may end up with something like clisp's stdlib :-)
18:06:43AraqparseOct should never be used anyway because octal numbers are an abomination.
18:07:02FromGitter<data-man> untar use it.
18:07:18Araquntar should not be used because it uses fucking octals.
18:11:49FromGitter<data-man> This is the standard. We did not invent it.
18:11:52*SenasOzys quit (Ping timeout: 245 seconds)
18:16:50Araqwe don't have to create ripple effects either
18:17:08Araqfixing Nim's untar package does not mean to add new features to the stdlib
18:18:29FromGitter<data-man> For consistency with parseHex. And maxLen is very useful.
18:20:44Notkeais nimble.directory broken?
18:21:33Araqlet me tell you a secret.
18:21:58FromGitter<data-man> When I added these proc for scanf, I missed maxLen. :(
18:24:01Araqconsistency is best way to bloat everything. so what if parseOct is not "consistent" with parseHex. one is called parseHex, the other is called parseOct.
18:24:15AraqparseJson doesn't have 'maxLen' either.
18:25:00Araqor parsecfg. and omg, parseCfg takes a Parser parameter and parseOct doesn't
18:25:20Araqsimilarities do not make laws.
18:27:08FromGitter<data-man> You too exaggerate. :)
18:28:08Araqyes, in the hope to stop this madness.
18:28:25YardanicoNotkea what do you mean by that?
18:28:33YardanicoNotkea website doesn't open for you?
18:28:51Notkea502 Bad Gateway for me
18:29:38YardanicoNotkea oh, strange, federico3 is the creator of this website, so I pinged him :)
18:30:01federico3looking...
18:30:48AraqI have patched over 60_000 loc the last weeks. and I noticed much in Nim's stdlib and in the compiler's featureset is was added to please some maniac's vision of "consistency"
18:31:41FromGitter<data-man> Did consistency binarySearch with other procs from algorithm is not useful? :)
18:33:25Araqstrutils contains isUpper and tons of stuff for consistency with Python ffs. Python needs these because it's slow and lacks set[char].
18:34:02AraqNim is fast and has set[char], it's a different language, we don't have to copy Python.
18:36:05Araqand it's just wrong to assume that we help Python newcomers and it causes no harm otherwise. it does harm. the code needs to be maintained. these procs end up in the documentation making the useful things harder to find.
18:36:36Araqtests for these run whenever somebody pushes a PR.
18:40:21FromGitter<data-man> Duplicating a number's parsing code in strutils is also not very good.
18:40:42Araqthat's indeed another problem.
18:41:07Araqusually the algorithm can be put into a template and then you write 1-5 different proc headers around it.
18:41:13*kobi joined #nim
18:41:26*kobi is now known as kobi7
18:41:34kobi7hi! congrats for the new forum
18:42:34FromGitter<data-man> One tiny PR. It's all. Please. :)
18:42:42AraqNo.
18:43:07AraqI'm still waiting for the unicode update btw.
18:44:08FromGitter<data-man> Soon™. On the this week.
18:44:16*SenasOzys joined #nim
18:46:52FromGitter<data-man> Today is Monday, right? :)
18:50:17dom96kobi7: thanks :)
18:50:27kobi7guys, how do I make an object field readonly?
18:51:05kobi7dom96, do you know apache has "ab" apache bench - which you can try for your beast
18:51:31dom96kobi7: yes, wrk is better
18:51:45kobi7wrk?
18:51:55dom96https://github.com/wg/wrk
18:52:10kobi7oh
18:52:27kobi7cool
18:52:53kobi7dom96: I have a field in an object, that I want to allow read access but not write access. how do I do that?
18:53:17dom96proc field*(x: Type): RetType
18:53:29kobi7* is rw, though
18:53:48kobi7sorry, it's a proc then?
18:53:49dom96Not if you define a proc
18:54:00dom96procedures can be called without () in most instances
18:54:21kobi7ok, good enough for me
18:54:37kobi7just for the record, there's no shortcut in the syntax for that
18:55:09Yardanicoyou can make a template :D
18:55:17Yardanicoor maybe a macro
18:56:59kobi7if it's in the same file, I can read the type's private fields? I mean the module is the file, correct?
18:58:06kobi7Yardanico: cool to try.
18:58:20Yardanicokobi7 yes, you can
18:59:27*yglukhov[i] joined #nim
19:00:35kobi7so, template get ( name, type : untyped) ... then take name, prepend '_', create field, create proc with name that returns type
19:01:01Yardanicoyou want to have fields as _field ?
19:05:55shashlickdom96: simpler PR for untar - https://github.com/dom96/untar/pull/4
19:07:48dom96shashlick: Tested it?
19:08:03kobi7Yardanico: we tinkered with the idea of readonly fields, so the _field is the private one, and proc field returns it. (theoretically)
19:08:23kobi7I guess a template is enoough to do this?
19:08:43Yardanicowell yeah, but you'll have to define your type with _field yourself
19:09:11*miran joined #nim
19:09:15*jhorwitz joined #nim
19:09:30jhorwitzHi everyone, been a while but was thinking of getting back into Nim
19:09:39kobi7right, but the definition can be the new keyword, say readonly field : string, which then creates it like that
19:10:23dom96jhorwitz: great :)
19:11:31FromGitter<krux02> Araq: I like consistency. But copying the API from other languages doesn't make Nim as a language on it's own consistent it is the exact opposite.
19:12:22FromGitter<krux02> it basically is a poptential split in possible styles of the language for no benefit.
19:12:48shashlickdom96: yep
19:12:55shashlicklet me run on Linux as well
19:14:31jhorwitzWhat is most needed as far as contributions currently?
19:17:34shashlickdom96: yep passes on linux as well
19:18:41*Guest79630 quit (Quit: Exiting)
19:21:06dom96ok thanks, merged
19:21:47shashlickcool
19:22:25shashlickany other concerns on the choosenim PR? once that's in, I can start testing and deliver the new features
19:22:51*jhorwitz quit (Quit: leaving)
19:23:35*jhorwitz joined #nim
19:28:20*kobi7 quit (Ping timeout: 256 seconds)
19:28:20*yglukhov[i] quit (Read error: Connection reset by peer)
19:28:41*yglukhov[i] joined #nim
19:29:50Araqjhorwitz: https://github.com/nim-lang/Nim/issues/7870 is rather easy to start with
19:30:08Araqor pick any issue tagged with "easy".
19:30:24Araqor look at the "needed libs" github repo
19:30:44dom96or fix things that affect your own projects
19:31:47shashlicklooking into wrapping 7z to extract archives, has this already been done before?
19:32:06Araqkrux02: consistency causes bloat though, it needs to be a careful tradeoff
19:32:39*mwbrown joined #nim
19:33:04Araqwhen everybody uses 10% of what is in strutils but pays the cost for compiling it completely we either should trim it down or at least get incremental compilations to work
19:33:19Araqpreferably both.
19:33:29FromGitter<krux02> yes
19:34:29dom96A project that I would really like to see made is a langserver (https://langserver.org/) implementation for Nim.
19:35:24dom96It should be a separate tool that performs its own analysis and relegates to nimsuggest (where appropriate)
19:35:43dom96That's what I would be writing right now if I didn't have thousands of other projects :)
19:35:45FromGitter<data-man> That's just the point, that everyone has his own 10%, often not overlapping with others.
19:37:49jhorwitzI can take that issue Araq
19:38:47shashlickdom96: replied to your comments
19:39:09shashlicki'm trying to wrap 7z so that we can use the same mingw.7z in choosenim and finish
19:39:21shashlickthis installer project is taking its own sweet time
19:39:50dom96shashlick: I'd prefer xz
19:40:57Araqno, 7z is superior
19:41:14dom96why?
19:41:19FromGitter<data-man> 7z supports xz
19:41:27shashlickwell, that's what's on both nim-lang.org and sourceforge, if I can convince araq to link to the official mingw :D
19:41:46Araqxz doesn't work, it requires 'tar' as interim step
19:41:57shashlickwell, so I'm only wrapping 7z extraction, xz is already covered in untar
19:42:13dom96yeah, and we've already got a native untar library in Nim
19:42:45dom96shashlick: Not on Windows
19:42:55dom96and only if the OS has `unxz` in its PATH
19:43:02jhorwitzWhat is involved in running the tests in C++ mode?
19:43:05Araq7z is one of my favourite tools and this is where the discussion ends for me.
19:43:11jhorwitzforgive my dumb question
19:44:21shashlickdom96: good point
19:44:49shashlickwe can look into xz for windows as part of data-man's nimcompression effort
19:45:03FromGitter<data-man> No, ```7z t anyfile.xz``` works without tar. :)
19:46:06FromGitter<data-man> @shashlick: I prefer libarchive. Incredible library!
19:46:23Araqjhorwitz: it's not dumb at all. Nim's tester is called "testament" and there is logic that supports targets: "C JS Cpp" or similar
19:46:52Araqeasiest is to change this logic so that "C" implies "also rerun as C++"
19:47:45shashlick7z is lgpl though
19:50:36Araqouch but at least lgpl doesn't affect our MIT license when we use it as a library
19:51:03federico3that's the point of the linking exception
19:51:22dom96sadly xz also appears to be lgpl
19:51:48dom96if you want extra challenge: create a native 7z/xz library in Nim and license it under MIT :D
19:52:33FromGitter<data-man> zstd, brotli, lz4, lizard, etc.
19:57:55FromDiscord<awr> does the {.nimcall.} pragma implies that a procedure is not a closure
19:57:58FromDiscord<awr> *imply
19:59:15dom96yep
20:02:09*miran quit (Ping timeout: 264 seconds)
20:09:13*DarkArctic_ joined #nim
20:10:04shashlickwell, was thinking of compiling in 7z extraction into choosenim from source
20:10:15*jhorwitz quit (Ping timeout: 245 seconds)
20:12:49*DarkArctic quit (Ping timeout: 248 seconds)
20:13:45*DarkArctic_ quit (Ping timeout: 260 seconds)
20:13:52shashlicklibarchive seems cool
20:15:46*yglukhov[i] quit (Remote host closed the connection)
20:16:21*yglukhov[i] joined #nim
20:21:02*yglukhov[i] quit (Ping timeout: 260 seconds)
20:26:33dom96ooh google search console uncovers some interesting bugs in the forum
20:26:54FromGitter<Varriount> Like?
20:28:19federico3marshal.store is able to generate invalid JSON that marshal.load cannot open :-/
20:34:38*nsf joined #nim
20:35:17*Cthalupa- joined #nim
20:38:23*Cthalupa quit (Remote host closed the connection)
20:39:22*Cthalupa- is now known as Cthalupa
20:40:31Araqfederico3, bug report please
20:45:15*Electrux quit (Ping timeout: 260 seconds)
20:45:16dom96Varriount: UTF-8 profile name handling problems
20:56:20*Electrux joined #nim
21:00:07*max3 joined #nim
21:01:00*Electrux quit (Ping timeout: 260 seconds)
21:06:41*gb00s quit (Quit: The Lounge - https://thelounge.github.io)
21:14:41*pydsigner quit (Quit: I'll Be Back!)
21:14:51*pydsigner joined #nim
21:15:13*elrood quit (Quit: Leaving)
21:16:20dom96Fixed :)
21:16:38dom96Didn't realise we had profiles like this lol https://forum.nim-lang.org/profile/%F0%9F%98%80
21:26:57Yardanicodom96 lol
21:41:12*sz0 joined #nim
21:45:49*gb00s joined #nim
21:46:42*CodeVance joined #nim
21:55:19*Electrux joined #nim
21:56:17*Trustable quit (Remote host closed the connection)
21:59:55*Electrux quit (Ping timeout: 260 seconds)
22:01:12*xkapastel quit (Quit: Connection closed for inactivity)
22:11:23dom96In case anyone's curious, here is what happens when someone deletes their account https://forum.nim-lang.org/t/3800
22:16:20*Electrux joined #nim
22:17:48FromGitter<kayabaNerve> How is everyone?
22:18:10FromGitter<kayabaNerve> dom96: Why would you want to leave?
22:18:32dom96No idea. Some people apparently do.
22:18:58FromGitter<kayabaNerve> :O They must have been brainwashed by the Pythonites. Such a horrible gang...
22:19:06FromGitter<kayabaNerve> Or traumatized by seg faults.
22:19:37*Vladar quit (Quit: Leaving)
22:20:27*Electrux quit (Ping timeout: 240 seconds)
22:24:03arecaceaehow do I offset a pointer?
22:35:02arecaceaenvm, I found it
22:35:49arecaceaehow do I get a pointer to a string, even if its allocated as static-memory? - addr(s[0]) only seems to work if its a var string - I just want to read it
22:38:16AraqunsafeAddr
22:39:11arecaceaeAraq: thanks
22:44:00*nsf quit (Quit: WeeChat 2.1)
22:44:47FromGitter<gogolxdong> got this on devel ,`Error: system module needs: nimGCunref`
22:45:07FromGitter<kayabaNerve> @gogolxdong What were you trying to do?
22:50:11*mostly-harmless quit (Ping timeout: 256 seconds)
22:50:32FromGitter<gogolxdong> using --gc:regions
22:51:53*mostly-harmless joined #nim
22:55:02FromGitter<gogolxdong> never mind , have made it right, it's container enviroment.
23:03:15*mostly-harmless quit (Ping timeout: 252 seconds)
23:03:33*leorize quit (Quit: WeeChat 2.1)
23:04:56*mostly-harmless joined #nim
23:24:00*xkapastel joined #nim
23:36:00*Electrux joined #nim
23:40:34*Electrux quit (Ping timeout: 264 seconds)
23:40:41*CodeVance1 joined #nim
23:43:27*CodeVance quit (Ping timeout: 240 seconds)
23:45:16*Guest71713 joined #nim
23:47:50Guest71713Minere BTC enquanto navega e trabalha sem deixar o pc lentão. https://getcryptotab.com/718967 https://www.youtube.com/watch?v=luzqQN3kL4g&t=166s