<< 20-11-2017 >>
00:02:51FromGitter<naelt> hello, im new to learn nim. i try some examples from nim in action on openbsd 6.2. i get a error message "Error: execution of an external compiler program 'gcc -c -w -pthread", when i try spawn a thread in a really simple example like let name = spawn stdin.readLine() . in the compilation step i set also --threads:on. i did not try it on other os, because openbsd is my target. it may be this a specific bsd error?
00:05:48*laas quit (Quit: WeeChat 1.9.1)
00:15:19*jinshil joined #nim
00:20:00*def-pri-pub joined #nim
00:27:13FromGitter<naelt> nim version 0.17.2
00:28:13FromGitter<techate> that's not a BSD-specific error message, it's not some kind of weird OpenBSD security f.e. that's just the error message provided when the command it calls to compile some code fails
00:28:37FromGitter<techate> it's probably that you need -pthreads instead of -pthread, or something. I remember that OpenBSD had special flags there.
00:30:45FromGitter<techate> with ktrace/kdump you might see the error provided by gcc itself, which might be more helpful: http://www.dotresults.com/2010/03/21/openbsd-strace-equivalent/
00:31:04FromGitter<naelt> thank you very much :)
00:35:15*endragor quit (Remote host closed the connection)
00:35:41*endragor joined #nim
00:38:46*astronavt joined #nim
00:39:21*endragor_ joined #nim
00:40:11*endragor quit (Ping timeout: 255 seconds)
00:43:45*endragor_ quit (Ping timeout: 248 seconds)
00:43:47*astronavt quit (Ping timeout: 255 seconds)
00:44:23*astronavt joined #nim
00:44:28*astronavt quit (Client Quit)
00:46:38*astronavt joined #nim
00:49:50*def-pri-pub quit (Quit: Leaving.)
00:59:02FromGitter<naelt> your answer lead me in the right direction, upgrading gcc solve my problem. thank you
01:08:33FromGitter<techate> cool. thanks for letting us know how you fixed it
01:08:33*yglukhov joined #nim
01:15:17*yglukhov quit ()
01:47:57*MJCaley joined #nim
01:59:05*vivus quit (Quit: Leaving)
02:10:21*rbrt quit (Quit: Oíche mhaith)
02:19:34*rbrt joined #nim
02:26:35*vlad1777d quit (Ping timeout: 240 seconds)
02:33:03*marenz__ quit (Ping timeout: 248 seconds)
02:55:38*chemist69 quit (Ping timeout: 255 seconds)
03:09:45*chemist69 joined #nim
03:21:37*arecaceae quit (Remote host closed the connection)
03:21:56*arecaceae joined #nim
03:25:59*def-pri-pub joined #nim
03:40:50*def-pri-pub quit (Quit: Leaving.)
03:56:54*rbrt quit (Quit: Oíche mhaith)
03:59:48*mr_yogurt joined #nim
04:02:08*rbrt joined #nim
04:05:07*astronavt quit (Disconnected by services)
04:05:37*astronavt joined #nim
04:06:25*astronavt quit (Disconnected by services)
04:07:56*astronav_ joined #nim
04:10:39*astronav_ quit (Remote host closed the connection)
04:11:59FromGitter<zacharycarter> I almost have my game graphics situation figured out
04:12:05FromGitter<zacharycarter> stumbling on the last bit
04:14:38*MJCaley quit (Quit: MJCaley)
04:18:05FromGitter<kdheepak> Does anyone know if I use Nim to create a touch bar application?
04:21:21*astronav_ joined #nim
04:34:16mr_yogurtis version 1.0 supposed to be 'soon
04:34:28mr_yogurt*'soon'?
04:34:33*rbrt quit (Quit: Oíche mhaith)
04:36:36*jinshil left #nim ("Good-bye!")
04:37:12FromGitter<techate> latest that I've seen: https://forum.nim-lang.org/t/3337#21029
04:38:57mr_yogurtalso, do plans exist for a major backwards-compatibility-breaking update sometime in the future? (is 1.0 supposed to be that update?)
04:44:44*onionhammer quit (Quit: WeeChat 1.9.1)
04:45:20*onionhammer joined #nim
04:46:24*SenasOzys quit (Remote host closed the connection)
04:48:00*rbrt joined #nim
04:49:22*SenasOzys joined #nim
05:07:06*rbrt quit (Quit: Oíche mhaith)
05:12:30*rbrt joined #nim
05:12:54*dddddd quit (Remote host closed the connection)
05:22:52*rbrt quit (Quit: Oíche mhaith)
05:26:04*rbrt joined #nim
05:40:53*skrylar joined #nim
05:47:14skrylarboop.
05:51:58*astronav_ quit (Remote host closed the connection)
06:03:16*Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
06:10:42*SenasOzys_ joined #nim
06:11:17*SenasOzys quit (Read error: Connection reset by peer)
06:25:24*def-pri-pub joined #nim
06:26:11*def-pri-pub left #nim (#nim)
06:30:44FromGitter<mratsim> @mr_yogurt 2.0 at the earliest
06:32:17FromGitter<mratsim> For now the only major change planned is destructors for seq/string but interface will stay the same
06:32:27*rbrt quit (Quit: Oíche mhaith)
06:35:52FromGitter<mratsim> @kdheepak you can probably use it, Nim compiles to Objective C and can import objective C libraries.
06:36:48*nsf joined #nim
06:45:00*endragor joined #nim
06:49:27*endragor quit (Ping timeout: 240 seconds)
07:01:05*endragor joined #nim
07:01:58*endragor quit (Remote host closed the connection)
07:02:46*endragor joined #nim
07:15:23*bluenote joined #nim
07:18:27*m712 quit (Ping timeout: 246 seconds)
07:21:25*m712 joined #nim
07:27:49*mr_yogurt quit (Ping timeout: 260 seconds)
07:33:51*gokr joined #nim
07:35:06*endragor quit (Remote host closed the connection)
07:46:19*endragor joined #nim
07:50:22*Ven joined #nim
07:50:46*Ven is now known as Guest85501
07:54:48*PMunch joined #nim
08:00:47*endragor quit (Remote host closed the connection)
08:04:11*Guest85501 quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
08:08:49*Ven joined #nim
08:08:57*Vladar joined #nim
08:09:12*Ven is now known as Guest409
08:10:25skrylarthinking about adopting debian's policy on what 'stable means'
08:10:34skrylarwell, mostly the "if there are no bugs in ten days" part
08:12:22*Guest409 quit (Client Quit)
08:13:28*Ven joined #nim
08:13:52*Ven is now known as Guest63647
08:16:42*miran joined #nim
08:18:43*vlad1777d joined #nim
08:19:50*xkapastel quit (Quit: Connection closed for inactivity)
08:21:16*endragor joined #nim
08:24:59*bluenote quit (Ping timeout: 260 seconds)
08:26:33*endragor quit (Ping timeout: 268 seconds)
08:31:17*claudiuinberlin joined #nim
08:31:30skrylarOn a downside, it seems ibus input modes don't work in EFL/FLTK
08:33:00*vlad1777d quit (Ping timeout: 246 seconds)
08:34:43*voidspacexyz joined #nim
08:38:20*sz0 joined #nim
08:38:39PMunchHmm, is there any functional difference between result = Something and just an implicit return of Something?
08:38:56*endragor joined #nim
08:39:34PMunchI think Araq mentioned something about result directly populating the memory of the return value, instead of first constructing an object and then returning it. Is this true for imlicit returns as well?
08:42:36*Arrrr joined #nim
08:42:36*Arrrr quit (Changing host)
08:42:36*Arrrr joined #nim
08:44:46*PMunch quit (Remote host closed the connection)
08:46:08*PMunch joined #nim
08:53:08skrylarPMunch, sounds about right
08:53:34skrylarassuming the return value is an object, and not a ref/ptr/something that has to be new/make'd anyway
09:02:39PMunchWhy wouldn't a ref/ptr work the same way?
09:03:02PMunchI mean putting the ref directly into the return position, obviously not the object itself
09:07:50skrylarbecause if result is a pointer, you have to assign it to an allocated object
09:08:19skrylarif result is just regular old data, then making a local one and copying it to result might save time (assuming it doesn't get optimized away)
09:08:29skrylar*avoiding
09:08:53skrylarref types might get auto allocated too, not sure.
09:22:58FromGitter<Varriount> Return values are usually inefficient and/or inflexible. "Out" parameters allow the caller to do things like reuse memory.
09:31:36skrylardon't out and var become pointers in c?
09:31:40skrylari know var does
09:32:32skrylari looked at efl again earlier. embedded systems people seem to like it for making like.. car UIs apparently
09:36:14FromGitter<Varriount> Var becomes a pointer to the stack location
09:39:28*Guest63647 quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
09:41:33*Ven joined #nim
09:41:57*Ven is now known as Guest85201
09:46:45*dexterk joined #nim
09:51:09PMunchVarriount, really?
09:51:31PMunchSo if I pass a var int to a procedure it will directly modify that position in my stack frame?
09:53:24FromGitter<andreaferretti> I hope so!
09:53:40FromGitter<Varriount> Yes.
09:54:26FromGitter<Varriount> That's the most efficient way to make the "assigning to the parameter changes the source value" work.
09:54:57PMunchHuh, neat!
09:57:00*Guest85201 quit (Read error: Connection reset by peer)
09:57:07skrylarout and var are synonymous afaik
09:57:16skrylari've read from 'out' parameters and it wasn't an error
09:57:32*Ven joined #nim
09:57:55*Ven is now known as Guest16662
09:58:56*Ven joined #nim
09:58:56*Guest16662 quit (Read error: Connection reset by peer)
09:59:19*Ven is now known as Guest4030
10:00:41*floppydh joined #nim
10:28:25*arnetheduck joined #nim
10:46:17*sz0 quit (Quit: Connection closed for inactivity)
10:54:45*Guest4030 quit (Read error: Connection reset by peer)
10:55:15*Ven joined #nim
10:55:39*Ven is now known as Guest21780
10:56:05*salewski joined #nim
10:58:10salewskiNote: gintro package is now using files from oldgtk3 instead repositories from ngtk3 for bootstrapping.
10:58:16*Guest21780 quit (Read error: Connection reset by peer)
10:58:32*Ven_ joined #nim
10:59:12salewskiThis gives fewer warnings during install, and I may be able to completely delete deprected ngtk3 soon.
10:59:46salewskiIn case it should not work for you, then let me know. Bye.
11:00:28*salewski quit (Client Quit)
11:04:19ArrrrHow do you turn {.experimental.} on in every nim file?
11:07:18*Senketsu quit (Remote host closed the connection)
11:13:08FromGitter<mratsim> I hate working with error code as return values :/ Give me a pretty exception instead of "CUDNN_STATUS_NOT_SUPPORTED" or "CUDNN_STATUS_BAD_PARAM" ...
11:16:41FromGitter<andreaferretti> yeah, it's really frustrating when some CUDA code gives you an opaque error like this
11:21:42*xkapastel joined #nim
11:27:32skrylarmratsim: that has been a topic of debate for the firebird wrapper
11:28:14skrylarthough exceptions are nightmares across language bounds
11:29:06skrylari recall one of the major reasons Go lacks them is because of the plumbing they require, and how despite java/c#/c++ having them for 20 years the plumbing isn't even standardized
11:29:07FromGitter<mratsim> The worse thing is I get cudnn_status_not_supported for a proc that shouldn't return that :/
11:29:45FromGitter<mratsim> Nice discussion on exception including for go: https://stackoverflow.com/questions/1736146/why-is-exception-handling-bad
11:30:37skrylarI rather like Lisp error handling admittedly
11:30:46skrylarexceptions are also one place that are simplified with a GC
11:31:09skrylaryou just longjmp to a handler and the gc sorts it later. ref systems have to inject book-keeping
11:31:59skrylarif some experimental ocaml programmers are to be believed, jamming the cpu / cache /etc all the time for that book keeping is *worse* than properly used GC
11:33:35skrylarI may have to open an issue on github and leave it for a few days, about if interbase/firebird users want exceptions. i can see the appeal, although people also don't tend to check every error code
11:33:46FromGitter<mratsim> By the way @andreaferretti are you still working on linear algebra/machine learning stuff?
11:39:38FromGitter<alehander42> the glang faq `[excptions] They are, by definition, exceptional `
11:40:22FromGitter<alehander42> I hate when people use semantic accidents to prove their point
11:41:08FromGitter<alehander42> exceptions are errors, when you e.g. have a parser for something, getting an error is extremely un-exceptional
11:41:21skrylarhttps://github.com/Skrylar/firebird/issues/6
11:41:46*MJCaley joined #nim
11:42:01skrylarwell, i find exceptions useful for fatal errors only
11:42:35skrylaralso writing tons of try/catch is quite annoying
11:42:47FromGitter<alehander42> you don't have to write a ton of try/catch-s
11:43:01FromGitter<alehander42> and this is very dependent on the usecase
11:43:11FromGitter<alehander42> in some cases they're very useful for non-fatal errors too
11:43:35FromGitter<alehander42> the ability to propagadate them to a single place of checking and continuing is very powerful
11:43:59skrylari dislike them for nonfatal errors because there is no adequate restart system
11:44:14FromGitter<alehander42> you write the restart system
11:44:26skrylarin the case of lisp (or some custom try/catches in C) there is a formal way to communicate nonfatal errors, i do like those
11:44:37FromGitter<alehander42> e.g. the parser example, you catch exceptions in your parseLine function, and you add them to error list, and continue with next parseLine
11:44:41skrylarotherwise its back to many try/catches :)
11:44:48FromGitter<alehander42> this is *one* try catch
11:45:06skrylarin that specific example it is
11:45:13FromGitter<alehander42> well there are many possible examples
11:45:18FromGitter<mratsim> Agree with alehander42, exceptions are just to handle error, say you pass a 3d Tensors to something that only accepts a 2d matrix? Exceptional? No, should be warned? Yes, should the proc automatically handle it? How it doesn't have the context ===> Exception. Obviously it should be done before any side effects happen.
11:45:35skrylarin the case of ex.dealing with a database, you need one for the overall connect/query/close and then another one for say if a single record update is being weird
11:45:53skrylarso you end up with try {stuff} try {update record} except {record is busy} | except {database is borked}
11:46:14FromGitter<alehander42> well that's the whole point, you have the choise
11:46:23FromGitter<alehander42> if you need granularity, you do several except-s
11:46:36FromGitter<alehander42> but often you just need to catch all db exceptions and print something generic
11:46:40FromGitter<mratsim> What is lisp way of handling? An "Either" type which returns (Result, Option[Exception]) ?
11:46:58FromGitter<alehander42> and you can do with one try catch and many excepts your example
11:47:01skrylarmratsim: they have exceptions like c++/nim and they also have restart handlers
11:47:39skrylarthe db code can issue a nonfatal exception "record update failed" and a handler can choose one of the restarts ex. "retry, abort, change query and run again" where the last one might be used in a REPL
11:48:07skrylaror in a system i worked on very briefly, if a texture had an issue the restarts could be "load a different one, load a smaller mipmap, load a generic placeholder"
11:48:11FromGitter<alehander42> well you can implement such a handler in several lines yourself
11:48:24skrylaryou can do everything yourself, always.
11:48:26*skrylar rolls eyes
11:48:30FromGitter<mratsim> Rust way is not bad too, it allows function chaining as well.
11:48:43FromGitter<alehander42> but it's so much easier to do it on top of exceptions than on top of just error codes :D
11:48:57FromGitter<alehander42> of course restart handlers are cool, but they're just extra
11:49:00FromGitter<mratsim> Depends on pattern matching I'd say.
11:49:11skrylarpattern matching is neat
11:49:18FromGitter<alehander42> yeah I like it too
11:49:32skrylaryou can pass in closures to functions to immitate restarts
11:49:41skrylarit's just that people aren't familiar with that pattern in the first place
11:50:11skrylarjust like people aren't generally aware of call/cc :p
11:52:26*MJCaley quit (Quit: MJCaley)
11:52:34skrylarbleh. embarcadero's site appears to be collapsing on itself
11:52:47skrylardoc wiki is down and funny sentences like this: "For upgrade and other InterBase licenses please reach out to your EMB_TEXT_SALES_REPRESENTATIVE or EMB_TEXT_LOCAL_EMBARCADERO_RESELLER partner for details."
11:55:10FromGitter<alehander42> wow embarcadero were the delphi ide guys right
11:55:18*Ven_ quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
11:55:34FromGitter<alehander42> distant memories of their delphi ide
11:55:35skrylaryea
11:55:43skrylari was using them to figure out how firebird worked
11:56:34*Ven joined #nim
11:56:59*Ven is now known as Guest96202
12:09:00AraqI don't like exceptions either :-) but my alternative, "better API design" doesn't convince anybody
12:09:25Araqand Rust and Golang are bad examples of how to not include exceptions
12:10:27Araqexceptions really feel like a dated solution though in a world where message passing is so important
12:11:26*Guest96202 quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
12:13:55FromGitter<mratsim> Well as long as you can give fine-grained error message like “Issue, shape %i from param %d is incompatible with shape %i from param %d” % (foo_shape, foo, bar_shape, bar) instead of “BAD_PARAMETER” I’m fine with any solution
12:15:07skrylarthere is the hell that interbase uses :)
12:15:23skrylara 20-word vector that is written to in magic pairs that have to be matched against documentation to figure out wat
12:17:52FromGitter<alehander42> I do like return-based errors with sum types and pattern matching, but it's good to have some `escalate` option
12:20:20*ttoe joined #nim
12:20:28FromGitter<alehander42> eg I do `a = weirdOperation() # must handle error case` ⏎ `a = weirdOperation?() # must be in Result<> function if error escalate it to the parent call otherwise unwrap`
12:23:52ttoeHi. I ran the parallel example from the manual: https://nim-lang.org/docs/manual.html#parallel-spawn-parallel-statement
12:24:56ttoeWeirdly I get the same result instantaneously whatever number i run the function "pi" with
12:25:20ttoeand i have no chance to observe if anything happened in parallel...
12:25:46Araqttoe it's just a toy examples, FP precision could prevent any real observations
12:26:16ttoeAraq: ah ok, thanks
12:27:14FromGitter<mratsim> This line will probably be slower than serial by the way due to false sharing: ⏎ ⏎ ```ch[k] = spawn term(float(k))``` [https://gitter.im/nim-lang/Nim?at=5a12ca22ba39a53f1ac130b5]
12:28:53FromGitter<andreaferretti> @mratsim Unfortunately my work on Nim in general has stalled due to other things at work :-(
12:29:07FromGitter<andreaferretti> Hopefully I will be able to contribute a little more after december
12:31:45FromGitter<mratsim> argh :/
12:32:14FromGitter<mratsim> well at least your workplace is supporting and using Nim :)
12:38:50*voidspace_xyz joined #nim
12:41:36*voidspacexyz quit (Ping timeout: 248 seconds)
12:58:18Araqinterestingly even Erlang has exceptions
12:58:55Araqwith its supervisor trees and tiny processes they still felt the need for exceptions
13:00:21skrylarmratsim: https://arxiv.org/abs/1711.02213
13:02:10FromGitter<alehander42> erlang's throw is perfect example of the `escalate` thing
13:02:32FromGitter<alehander42> it's basically a non-local return
13:02:45FromGitter<alehander42> and you're supposed to deal with it inside your module/process usually
13:05:10*voidspace_xyz quit (Ping timeout: 264 seconds)
13:05:33*voidspacexyz joined #nim
13:05:38Arrrris cast really caster than type() ?
13:05:41Arrrr*faster
13:09:40AraqArrrr, depends on the involved types.
13:10:03Araqbut it's not the same, there is no speed tradeoff here, they do different things?
13:10:41Arrrrsay a conversion between different integer types.
13:13:49skrylarAraq, it might be necessary to have a nim.cfg option to add things for staticexec's path
13:14:29skrylarwe're running in to a strange situation where ex. firebird on SuSE puts fb_config somewhere silly, which means PATH has to be edited to find it
13:14:40skrylaralthough in debian/arch its in a normal place, so it works fine with a nimble install
13:14:58Araqwho uses SuSE anyway :P
13:15:04skrylarqqtop does apparently
13:16:31Araqit's a valid point though
13:16:44Araqpath support staticExec sounds like a useful thing to have
13:21:09*dddddd joined #nim
13:23:29FromGitter<mratsim> @skrylar, funny someone sent me Flexnet PDF on Slack just 1 hour ago :p
13:24:08Araqhttp://www.erlang.se/euc/04/carlsson_slides.pdf claims 'throw/catch' is not widely used
13:25:58FromGitter<mratsim> For me exceptions are like GC, when Nim 1.0 will be posted, I bet that there will be a thread "Nim has GC so it's bad" and another "Nim has exceptions so it's bad".
13:26:11FromGitter<mratsim> On HN*
13:26:49skrylarmratsim: Import AI just mentioned it
13:27:01dom96kdheepak: You can most definitely create an app that makes use of the touch bar on the new MBPs using Nim
13:27:16skrylarhow can gc and exceptions be bad on hackernews. that place was founded by lispers and lisp does both of those things
13:28:07*d10n-work joined #nim
13:28:09FromGitter<mratsim> The bare-metal crowd, spear-headed by C/Rust and a bit less C++ people.
13:32:24FromGitter<mratsim> Mmm author of ZeroMQ says he should have use C + error codes instead of C++ + exceptions: http://250bpm.com/blog:4
13:36:12FromGitter<mratsim> (I'm not convinced)
13:36:23FromGitter<alehander42> Araq: but error/exit (the process exceptions) are more used ?
13:36:55*ttoe quit (Quit: Page closed)
13:37:45*Jesin quit (Quit: Leaving)
13:39:51*cspar_ joined #nim
13:42:28*cspar quit (Ping timeout: 240 seconds)
13:50:07AraqI don't know
13:54:32*sz0 joined #nim
13:55:06*SenasOzys_ quit (Remote host closed the connection)
13:55:31*SenasOzys_ joined #nim
13:59:46Araqerlang uses tail recursions instead of looping so what is 'break' in Nim becomes "ugh" in Erlang
14:04:09PMunchWouldn't break just be return with tail end recursion?
14:04:21PMunchSince you're reusing the stack frame
14:05:08Araqhow so? return would be a local return, you would need some propagation mechanism
14:09:09FromGitter<techate> since you're tail-recurring there's nothing to propagate to
14:09:48AraqI think it means your tail recursion becomes a real recursion
14:10:09Araqnot that "return" suddenly does the proper thing for you and leaves the loop
14:10:09FromGitter<techate> but it depends on how you do it. can't just reuse one of the existing tail-recursive functions in the language and expect your provided fun to break for you
14:11:15PMunchWait, what are you discussing?
14:11:56FromGitter<techate> Erlang not having break as a subtopic of exceptions vs. other error handling
14:12:44PMunchAh right
14:14:57*MJCaley joined #nim
14:15:02*LyndsySimon left #nim (#nim)
14:19:21*nsf quit (Quit: WeeChat 1.9.1)
14:22:49FromGitter<techate> https://gist.github.com/techate/4ff55eae9c7c4d40b102137bd5276fa3 f.e.
14:23:25FromGitter<techate> eh the 'echo n' should de-indented fully. copy&paste error.
14:23:55*fvs joined #nim
14:24:23FromGitter<techate> you don't need to do anything special to exit the loop. just stop looping.
14:28:47FromGitter<krux02> I saw the changes section for the 0.18 Version of Nim. Looks really good to me, I like it :D
14:31:47*crem quit (Ping timeout: 258 seconds)
14:33:10FromGitter<krux02> Nim is really developing Nicely into a good language
14:33:30*crem joined #nim
14:37:25PMunchkrux02, link?
14:39:42*ftsf quit (Ping timeout: 260 seconds)
14:44:09FromGitter<krux02> PMunch: https://github.com/nim-lang/Nim/blob/devel/changelog.md
14:44:34PMunchThanks :)
14:47:17PMunchNice, I like the changes to the JSON module
14:47:26*claudiuinberlin quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:47:31PMunchBut what's the difference between getBiggestInt and getInt?
14:50:48euantorBiggestInt = 64 bit always IIRC, int is platform dependent
14:50:52PMunchHaha GitHub markdown didn't like the ..code-block:: nim stuff :P
14:51:57FromGitter<mratsim> It's not Nim it's reStructured
14:52:12PMunchHmm
14:52:24*PMunch quit (Quit: Leaving)
14:52:48Araqyeah I'll never get used to markdown :P RST is just superior
14:53:36*claudiuinberlin joined #nim
14:54:20*MJCaley quit (Quit: MJCaley)
14:55:35*Jesin joined #nim
15:00:34FromGitter<krux02> what is so much better with RST compared to markdown?
15:00:43FromGitter<krux02> I really have no opinion, for me it is just differnet
15:01:19FromGitter<krux02> also was hast du so zu sagen?
15:01:35FromGitter<krux02> sorry wrong chat
15:02:09AraqRST has more features and trailing whitespace is not significant
15:03:23skrylarwell markdown was intended to be a way to format the way people wrote emails in to html
15:03:33*SenasOzys__ joined #nim
15:03:35skrylarit wasn't built to make docbooks
15:03:38Araqalso RST was carefully designed for documenting (Python) source code, markdown feels more like a hack for "I need a syntax for my forum software"
15:04:03skrylari sort of like texinfo. asciidoc is ok
15:04:15skrylarbut texinfo has roughtly 3 elements of syntax, and asciidoc has many
15:05:08skrylaronly used rst briefly with sphinx
15:05:17*ftsf joined #nim
15:05:26FromGitter<krux02> I us rst only in the Nim forum
15:05:39*SenasOzys_ quit (Ping timeout: 250 seconds)
15:06:31FromGitter<krux02> I read that rst is a remake of structured text, but with a specification first
15:06:59FromGitter<krux02> could be that Nim has that Future, too. Some reNim
15:07:05FromGitter<krux02> with specification first
15:07:28Araqplanned for Nim v2 :P
15:07:50Araqthough specs are hard to write and the ultimate form of a spec is an executable spec
15:07:56Araqalso known as "program"
15:08:23FromGitter<krux02> well yea, but a specification is kind of helpful
15:08:57FromGitter<krux02> and it helps to decide if something is a bug or a feature
15:12:57Araqspecs come after successes
15:13:19Araqit's hard to find many cases where the spec came first
15:15:28FromGitter<krux02> Algol 68
15:15:40FromGitter<krux02> But you can't really speak from success :P
15:15:57cremOften idea looks good until you try to use it, that's the problem with spec-first languages. Spec-last is much more live and malleable.
15:16:10FromGitter<krux02> And Plankalkül has the spec first
15:16:39cremC++ is kind of spec-first, especially in 90s.
15:16:53*nsf joined #nim
15:17:52FromGitter<krux02> Well, not really. There are a lot of c++ implementations all implementing non-spec features that then develop into the spec
15:17:55FromGitter<krux02> same with OpenGL
15:18:11FromGitter<krux02> (not a programming language, but same development process)
15:18:50AraqI don't really know, but I bet the first Ansi C++ standard came after "C with classes version 1"
15:20:43cremFor modern C++, usually features are tested before including into the standard. But in 90s, there were all those template exporting, weird throw definitions for functions, etc, which were not viable.
15:20:45FromGitter<krux02> Araq: I would really like to talk again about ``getImpl`` again and the non-symbol identifiers for the arguments
15:21:11Araqdon't talk about it, fix it
15:21:26FromGitter<krux02> well ok I plan to actually do it
15:21:28Araqbut document this breaking change
15:21:35FromGitter<krux02> but I don't know where the ast is created
15:21:45FromGitter<krux02> in changes.md?
15:21:48Araqsemtypes.nim
15:21:55Araqyes, in changes.md
15:22:02FromGitter<krux02> ok
15:22:29FromGitter<krux02> I take it as `green light` for the change
15:22:40*Senketsu joined #nim
15:24:43FromGitter<krux02> Araq: semtypes is long, can you tell me how and when in the compilation the implementation ast is constructed
15:25:09Araqwell it isn't, which is your problem
15:25:20Araqthe original AST is not touched in this case
15:25:33Araqinstead the type is constructed
15:26:11FromGitter<krux02> all I found out so far is, that when I call get Impl it returns the ast node that is stored in the symbol node
15:26:25FromGitter<krux02> but that means that the ast at that point in time is already existent
15:28:26FromGitter<krux02> I would like to know how and where that node with the ast is created
15:30:46*voidspace_xyz joined #nim
15:32:50Araqsemstmts.nim line 1465
15:32:54Araqs.ast = n
15:33:06Araqas well as in a few other places
15:33:23FromGitter<krux02> ok thanks I will look into that
15:33:31FromGitter<krux02> and probably have problems
15:33:55*skrylar mutters that everything that isn't the common lisp hyperspect is probably waste material
15:34:08*voidspacexyz quit (Ping timeout: 240 seconds)
15:34:10*SenasOzys__ quit (Remote host closed the connection)
15:34:30*SenasOzys__ joined #nim
15:34:31*fvs left #nim ("ERC (IRC client for Emacs 25.3.1)")
15:34:33FromGitter<krux02> well I don't know common lisp hyperspect
15:34:39skrylarfunny to think people used to deride common lisp for "having a hundred pages of spec", yet the C++ drinking game is every time something is "implementation specific" or "undefined"
15:34:49skrylarits called the hyperspec, i just mispelled it
15:35:08Araqskrylar common lisp misses the "linear Lisp" features though
15:35:26skrylarit also misses the compiler that lets you use it for games :|
15:35:27FromGitter<krux02> there is also unspecified
15:35:42FromGitter<krux02> for example the type for a lambday expression in c++ is unspecified
15:35:45skrylaryea but the thing is people were being rude towards lisp because "the c++ spec is shorter therefore c++ is good"
15:35:55skrylarBut the Common Lisp hyperspec leaves almost nothing undefined.
15:36:47FromGitter<krux02> there was a lisp assembler hybrid invented by naughty dog in the PS2 Aera
15:36:59skrylarit does seem to be best to have some amount of people working on random stuff, then coalesce the working parts of those random stuff in to a well specified core with extension points
15:37:33FromGitter<krux02> Emacs Lisp will rule all the Lisp implementations :P
15:37:52skrylari would have kept using lisp if it was viable :p
15:38:17FromGitter<krux02> I would have kept using C++ if it was viable
15:38:35skrylarits just that, for some reason, none of the lispers who do free things seem to care. even though they keep getting people saying they'd use it if they could run a game engine on it, they just get told "well the 80mb should be acceptable"
15:38:36FromGitter<krux02> or Go, or Scala
15:39:04FromGitter<krux02> none the less each of those languages left something in our Mind that we just liked
15:39:24skrylari don't know if smalltalk/lisp people are fundamentally derp or if they're just dead
15:39:30*fvs joined #nim
15:41:34FromGitter<krux02> well the one thing smalltalk people love is this change of the running program
15:41:58FromGitter<krux02> and I think the code exploration
15:42:16skrylarthere are commercial lisps that will treeshake and give you <10mb distributables
15:42:35skrylarthere used to be those for smalltalk but i dunno
15:42:49skrylarfor some reason the unix world decided to use the worst language they could find instead of the good ones
15:42:55skrylarand then everyone copied this decision
15:43:00FromGitter<krux02> well I honestly think that the future for commercial programming languaes is very bad
15:43:04AraqI don't use Lisp because I like static type checking
15:43:34Araqand the ((())) is just too optimized for parser simplicity for my taste
15:43:52skrylarallegedly there was going to be a real syntax but people got used to the parens before it was done
15:43:54FromGitter<krux02> I don't think C is the worst language
15:44:09FromGitter<alehander42> exactly, people love to boast how "you write directly ast", but that's actually tedious
15:44:25skrylari think someone said the syntax was going to be what haskell ended up using, or similar
15:44:29skrylarbut it just never happened
15:44:50*rbrt joined #nim
15:44:53FromGitter<krux02> my two cents, I think in the future we won't have syntax trees in text form anymore
15:44:59skrylarsmalltalk would probably be a fun scripting language in nim. if nothing else because ST-80 takes like no effort to teach noobs
15:45:12FromGitter<krux02> will take a long time until we are there, but text editors will just become less important
15:45:15FromGitter<alehander42> @krux02 I've played a lot with that idea
15:45:27skrylarsource-code-in-database is an old thing
15:45:37FromGitter<krux02> I don't mean in database
15:45:40AraqC is not the worst language but all its innovations turned out to be misguided
15:45:46skrylarthere's also job automation. lol
15:45:47FromGitter<alehander42> on one hackathon I implemented something like a central format with a bidirectional grammar and realtime translator between user definable syntaxes
15:46:03skrylarwell c is sold as a readable assembler
15:46:04Araqarrays decoy into pointers, type definitions reflect an expression's usage
15:46:08skrylarbut you don't write all your shit in assembly
15:46:37Araqassembly has far fewer corner cases and a dead simple syntax, not a syntax that is dynamite
15:47:01skrylarwould rather just do what red did
15:47:02Araqwhere you can leave out a * and introduce a backdoor
15:47:10FromGitter<krux02> I mean that instead of writing text and thinking an ast, and the computer then parses that ast hopefully in the same way as thought. I think the future will give us programming languages where the ast is presented to us in some form and we an operato on the ast, not text.
15:47:25skrylar"heres the lisp dialect for make run fast and the lisp dialect for i want to keep my hair attached"
15:47:49FromGitter<alehander42> the point was to have different "views" on the ast (even ui)
15:47:56FromGitter<krux02> there is scratch (programming for children) doing that, and there is a jetbrains project reseraching in that direction
15:47:59FromGitter<alehander42> and operate on them on need
15:48:05FromGitter<krux02> yes
15:49:07FromGitter<alehander42> isn't paredit something in that direction? basically you mostly need powerful text-like editing for trees
15:49:19FromGitter<alehander42> what's the jetbrains project
15:49:28Araqalehander42 sounds like the Nim design I started with
15:49:59*tefter quit (Remote host closed the connection)
15:51:06FromGitter<krux02> https://www.jetbrains.com/mps/?fromMenu
15:51:09skrylarwonder how the htmlparser package compares to ex gumbo
15:51:21FromGitter<alehander42> https://github.com/alehander42/hivemind the examples in the readme illustrate the idea (they define the syntaxes), but it's not easy to make something like that be robust and well defined in each case
15:51:24*claudiuinberlin quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
15:51:37FromGitter<alehander42> was that the idea behind filter Araq?
15:52:29skrylarthere was this http://ohayo.computer/landing.html but they seem to have rebranded to make dashboards instead of just being a tree ide
15:52:43*Jesin quit (Quit: Leaving)
15:53:51Araqalehander42: source code filters? I introduced them for code generation purposes like 'niminst' uses them
15:54:37AraqI wanted a templating system that is simple to understand and so I mostly switched the defaults around
15:55:07Araqinstead of quoting to get a verbatim string, you quote to get a non-verbatim expression
15:55:47Araqand since that's quite pervasive, it is done on the file level
15:56:17*sz0 quit (Quit: Connection closed for inactivity)
15:56:46Araqnot much to say about it really, if your templating solution doesn't support for loops or conditionals, rethink your solution
15:57:15cremAraq: any update on nim code formatter? After being spoiled with C++ and Python code formatted, I'm missing the same for nim very much. :)
15:57:34Araqbecause "separation of concerns" is never a good reason to cripple a notation.
15:58:19Araqcrem, I prefer to work on Nim version 1 instead or on --newruntime
15:58:48crem:(
15:58:52*Jesin joined #nim
15:59:17FromGitter<alehander42> I see, makes sense, I have to look at them
15:59:43FromGitter<alehander42> is the idea behind the code formatter to just pretty print PNode-s according to configurable style object ?
15:59:53FromGitter<alehander42> (I am trying to see how it is diff than repr
16:00:07Araqexcept for the "configurable" part, yes
16:00:24AraqI don't plan to support configurations, there is one style to rule them all
16:00:34FromGitter<alehander42> so how is it different than repr :D
16:00:50cremIf that default style will have 80 column code width limit, I'm fine. :)
16:01:08Araqcrem, it has, lol
16:01:34Araqalehander42: it isn't, repr uses compiler/renderer.nim and so does 'nimpretty' once it is finished
16:02:47FromGitter<alehander42> so probably what I don't get is
16:03:31FromGitter<alehander42> isn't the process just parse Nim -> prettypring it back, why can't it be done right now with parse + repr
16:04:35Araqbecause it eats your non-doc comments.
16:05:49*claudiuinberlin joined #nim
16:06:05FromGitter<alehander42> auch
16:09:11*claudiuinberlin quit (Client Quit)
16:10:17*claudiuinberlin joined #nim
16:13:55*claudiuinberlin quit (Client Quit)
16:15:12*claudiuinberlin joined #nim
16:16:01*rbrt quit (Quit: Oíche mhaith)
16:16:38*nsf quit (Quit: WeeChat 1.9.1)
16:17:30*claudiuinberlin quit (Client Quit)
16:18:10*claudiuinberlin joined #nim
16:19:37onionhammerin
16:21:19planetis[m]hey! in macros there is toNimIdent() newIdentNode() ident() (and formely ! operator) to construct and NimNode. what method is prefered?
16:21:35*claudiuinberlin quit (Client Quit)
16:21:50*floppydh quit (Quit: WeeChat 1.9.1)
16:22:11*floppydh joined #nim
16:22:50Araqident()
16:27:51*claudiuinberlin joined #nim
16:27:55*claudiuinberlin quit (Client Quit)
16:29:07FromGitter<krux02> Araq: well I kind of have no progress at all understanding that code in symtypes.nim
16:29:30FromGitter<krux02> can you explain me a bit of context?
16:29:59FromGitter<krux02> I tried to run that code with the debugger a bit and set breakpoints and display stuff
16:30:03FromGitter<krux02> just to understand that code
16:31:09FromGitter<krux02> what is ``semTypeNode`` supposed to do?
16:31:20FromGitter<krux02> what are it's arguments
16:31:28FromGitter<krux02> and what should it return?
16:32:24*Trustable joined #nim
16:32:25planetis[m]oh newIdentNode constructs a NimNode from a NimIdent the rest construct a NimIdent from a string
16:33:37*rbrt joined #nim
16:33:40planetis[m]and toNimIdent is not on the documentation in https://nim-lang.org/docs/macros.html was it added recently?
16:36:47*JappleAck quit (Ping timeout: 248 seconds)
16:36:50FromGitter<krux02> I recommend ``ident"someName"``
16:37:08FromGitter<krux02> creates an identifier node the shortest wad
16:37:09FromGitter<krux02> way
16:39:49planetis[m]thanks krux02
16:41:56*elrood joined #nim
16:48:49cremIn nimble, can i make project depend on another local package? I have a library i want two projects to depend on.
16:51:19Araqlocal package? it needs to have an URL otherwise it doesn't exist
16:51:28dom96that's false
16:51:58dom96You can depend on it. Just need to install each dependency separately.
16:52:07dom96(or better, use `develop`)
16:53:49*JappleAck joined #nim
16:57:09*JappleAck quit (Client Quit)
17:00:21FromGitter<krux02> Araq: I debugged the compiler for a while now. I came to the conclusion that the position semtypes.nim:1466 is not called at all from my testfile
17:00:57FromGitter<krux02> Araq: So where and how is getImpl implemented?
17:01:17Araqwell you keep changing the question
17:01:28FromGitter<krux02> no I don't
17:01:31*rbrt quit (Quit: Oíche mhaith)
17:01:38AraqgetImpl is different from your "parameters are not syms" problem
17:02:10FromGitter<krux02> well for me it is not different, because I don't know the implementation details and differences
17:02:28Araqproc getImpl*(s: NimSym): NimNode {.magic: "GetImpl", noSideEffect.}
17:02:50Araqgg mGetImpl compiler
17:02:50Araqcompiler\ast.nim
17:02:50Araq 639: mEqIdent, mEqNimrodNode, mSameNodeType, mGetImpl,
17:02:50Araqcompiler\vmgen.nim
17:02:51Araq 1025: of mGetImpl: genUnaryABC(c, n, dest, opcGetImpl)
17:02:55FromGitter<krux02> that get imple I alredy inspected
17:03:04FromGitter<krux02> it is a dead end
17:03:23FromGitter<krux02> it just returns an ast that is already constructed at the point when getImpl is called
17:03:36Araqyes, exactly
17:03:41FromGitter<krux02> so I need to change the code that constructs that AST the is returned from getImpl
17:03:42Araqso you're on the right track
17:03:58Araqhowever, you can also patch opcGetImpl
17:04:13Araqthat has less devastating effects
17:04:26FromGitter<krux02> well I can imagine that
17:04:45FromGitter<krux02> but I also need the symbols in the do expression we talked about
17:05:02FromGitter<krux02> weet to be honest the ``do`` expression is what is important to me
17:05:31FromGitter<krux02> I am working on a nim to glsl compiler and I write the parts that are glsl in a do block, so that I can inline them
17:07:39FromGitter<krux02> for example: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a130bdbf257ad9109bac1d8]
17:08:12FromGitter<krux02> this is then compiled into fragmentshader, vertexshader, varyings attributes, and uniforms
17:08:32FromGitter<krux02> ``mvp`` is a symbol that is used in that code block
17:09:07FromGitter<krux02> because it does not come from the arguments ``v`` or ``gl``, it has to be a uniform
17:10:11FromGitter<krux02> this means, that I have to check if any symbol that is used in that ast is actually one of the argument symbols or a locally defined symbol
17:10:49Araqkrux02: semProcTypeNode
17:11:19Araqsemtypes.nim, line 996 creates the parameter symbol
17:11:40FromGitter<krux02> ok, thank you
17:11:41Araqafter its contruction try
17:11:51Araq a.sons[j] = newSymNode(arg)
17:12:00Araqto patch the original AST
17:12:35FromGitter<krux02> ok I will try to understand that part of the compiler and do what you said
17:12:41AraqI don't know how debugging would have ever got you there, as I said you want to know why something is **not** changed
17:12:42*rbrt joined #nim
17:12:57Araqso my advice, nag me more instead of debug more
17:13:03*voidspace_xyz quit (Ping timeout: 248 seconds)
17:21:18FromGitter<krux02> ok
17:21:26FromGitter<krux02> but that nagging only works when you are there, too
17:21:40FromGitter<krux02> I mean you are there a lot, but no always
17:21:45FromGitter<krux02> the debugger is there always
17:23:43cremHmm. So I have a nim library which I don't want to publish (I'm fine with uploading it into private bitbucket repo if it helps), and want to use it in several of my toy projects, what's the best way to arrange that? Just have a symlink to a directory of that library from my project dirs?
17:23:57*voidspacexyz joined #nim
17:24:14cremAh, maybe nimble develop is indeed what I need.
17:24:33FromGitter<krux02> you can call nimble install
17:25:36cremI guess I don't fully understand what nimble does, will read the doc and come back. :)
17:25:59FromGitter<krux02> crem: nimble just puts the source of tha project into a common path
17:26:16FromGitter<krux02> well kind of
17:27:11cremIf I just do 'nimble init' for several packages, will they see each other?
17:27:49*SenasOzys_ joined #nim
17:28:24cremDoesn't seem like that..
17:28:33*SenasOzys__ quit (Ping timeout: 248 seconds)
17:29:31cremSo I have ~/dev/mylib and ~/dev/myproj, both started with `nimble init`, never uploaded to any external repo. What's the easiest way to make myproj see mylib?
17:33:23*floppydh quit (Quit: WeeChat 1.9.1)
17:33:24cremHm, probably I should not use nimble at all for personal projects..
17:37:14cremWhat's the most recent recommendation for project structure? Nimbook suggests src/, nimble reference suggests puttings sources to the root.
17:48:57*xkapastel quit (Quit: Connection closed for inactivity)
17:49:35*voidspacexyz quit (Ping timeout: 258 seconds)
17:52:08Araqcrem, check Nimble's documentation
17:52:53cremReading.. seems that I have to create a custom packages.json, and point nimble.ini's [PackageList] section to it. :-\
17:53:25FromGitter<Varriount> Wait what? Can't you just make a nimble file?
17:54:38cremTo make one non-public package depend on another non-public package.
17:55:14cremBoth of them do have .nimble files.
17:55:34cremBut without custome packages.json I don't see a way to make them depend on each other.
18:00:34*nhywyll joined #nim
18:02:57FromGitter<Varriount> Ah.
18:03:59FromGitter<Varriount> Well, I don't know of a package manager that has any easier mechanism for this kind of situation
18:04:13*claudiuinberlin joined #nim
18:04:32*sendell joined #nim
18:05:01FromGitter<Varriount> Python's pip requires a package index (which needs to be a webserver!)
18:06:46cremGo forced all packages to be in one source tree, so they see each other easily. I cannot say though that I like that.
18:07:08FromGitter<krux02> crem: I like that
18:08:05FromGitter<Varriount> What do you mean? They all have to be public?
18:09:06*bluenote joined #nim
18:11:56sendellis there already a nim library abstracting over WebGL and OpenGL (and possibly other graphics API) usable with the js target? Or is opengl+emscripten the way to go? :)
18:16:12FromGitter<krux02> webgl is opengl
18:16:44FromGitter<krux02> sendell, yes I think there is someone who did something like that
18:17:03FromGitter<krux02> but I am sorry I can't remember the name at the moment
18:18:16sendellno problem, maybe someone else will :) good to hear that it exists already
18:18:38FromGitter<krux02> but I think it works with emscripten
18:19:43sendellprobably. might help with js target restrictions too i guess
18:21:01dom96crem: Like I said, just use `nimble develop` on your local package's
18:21:08dom96add them normally as a dependency
18:21:16dom96You can just use their name's
18:21:22dom96requires "myPkg"
18:23:36cremdom96: indeed, that works! Thanks. Before that, I tried "nimble develop mylib" from mylib/.. and that didn't work.
18:24:43FromGitter<krux02> the experience "./koch temp c /tmp/scratch.nim" add debug statement "./koch temp c /tmp/scratch.nim" is realld not good
18:38:08Araqkrux02: hmm?
18:39:32*MJCaley joined #nim
18:48:21*ShalokShalom joined #nim
19:05:32bluenote@Araq: Yesterday I spend some time trying to convert a NimNode (from a getType expression) into the corresponding typedesc, and I couldn't find a solution.
19:05:58bluenoteDid I miss something obvious or would it make sense to have a feature request for a NimNode -> typedesc proc in macros?
19:06:44Araqbluenote, I think it's missing
19:11:11bluenoteI looked into how mNGetType is implemented (the inverse function), but couldn't really see what it would require to add it
19:12:33bluenoteand do you think the typedesc forwarding problem is hard to fix? https://github.com/nim-lang/Nim/issues/6784
19:15:18Araqbluenote, it's something I wanted to look into anyway
19:15:38Araqthere is a special casing in overloading resolution for templates and macros and that's not documented
19:15:46Araqand ideally can simply be removed
19:17:13bluenotecool, then I think the string interpolation idea from @dom96 could work nicely without giving up type safety
19:22:36*laas joined #nim
19:23:24laasWhat exactly does this error mean? Error: cannot use symbol of kind 'let' as a 'var'
19:23:46laasI'm trying to use var instead of let in a template, and it says this
19:26:33Araqlaas, you use macros.gensym incorrectly?
19:27:23*Sentreen quit (Ping timeout: 250 seconds)
19:30:18laasthis is actually in a template
19:30:24laasand I want it to have a specific name
19:32:57laasoh this is weird
19:33:05laasI only have this problem when I'm generating the template inside another template
19:40:32*Sentreen joined #nim
19:52:13*endragor quit (Remote host closed the connection)
19:54:13*endragor joined #nim
19:55:45FromGitter<mratsim> @sendell it’s zengine from zacharycarter
19:55:56FromGitter<mratsim> it uses emscripten
19:58:35*endragor quit (Ping timeout: 250 seconds)
19:58:47*kier quit (Ping timeout: 260 seconds)
20:01:21*Trustable quit (Remote host closed the connection)
20:06:13sendellthx mrasim i'll look at it :)
20:06:55*endragor joined #nim
20:08:08*fvs left #nim ("ERC (IRC client for Emacs 25.3.1)")
20:09:50*claudiuinberlin quit (Quit: Textual IRC Client: www.textualapp.com)
20:11:11*endragor quit (Ping timeout: 248 seconds)
20:14:33cremFor nimble package, it's not correct to put sources into src/ subdir, is it?
20:15:41laasas far as I understand, no
20:15:47laasit gave me warnings when I did
20:16:13laasabout my previous problem, here is the code: https://hastebin.com/icipunotug.cs
20:16:25laasthe first set of templates work, but the second set doesn't
20:16:56laasI get an error like "meta.nim(23, 17) Error: expression 'self128055' has no type (or is ambiguous)" in the first set of templates
20:17:13laasI really don't understand why, even when I make them completely equal
20:20:25*TjYoco joined #nim
20:21:39dom96srcDir = "src"
20:21:41dom96It is valid
20:21:46dom96And it is recommended
20:22:30cremThanks.
20:24:02*MJCaley quit (Quit: MJCaley)
20:24:10Arrrr!eval bool is enum
20:24:11NimBotCompile failed: ???(???, 0) Error: expression 'false' is of type 'bool' and has to be discarded
20:24:23Arrrr!eval echo(bool is enum)
20:24:25NimBotfalse
20:26:26FromGitter<alehander42> what is the "dom96 interpolation idea"
20:26:30*xkapastel joined #nim
20:27:08FromGitter<alehander42> string interpolation would be great
20:27:18dom96https://github.com/nim-lang/Nim/pull/6507#issuecomment-345071140
20:28:16*ofelas joined #nim
20:38:10*vlad1777d joined #nim
20:43:01*rbrt quit (Quit: Oíche mhaith)
20:43:15*TjYoco quit (Quit: Leaving)
20:44:10Arrrrlaas i was having your issue with another template.
20:54:40*SenasOzys_ quit (Read error: Connection reset by peer)
20:56:54*Arrrr quit (Read error: Connection reset by peer)
20:57:16*SenasOzys_ joined #nim
21:08:49*ofelas quit (Quit: shutdown -h now)
21:08:50*rbrt joined #nim
21:18:06skrylarwell, the cobweb experiment only has ~2 parts left
21:18:21skrylarwriting the adapter macro now :\
21:22:04*Vladar quit (Quit: Leaving)
21:24:10laasArrrr: so you don't know how to fix it? :/
21:31:39Araqlaas, gist the code and you'll be helped
21:34:02*SenasOzys__ joined #nim
21:36:31*SenasOzys_ quit (Ping timeout: 250 seconds)
21:37:23FromGitter<zetashift> https://www.reddit.com/r/nim/comments/7ebnr6/this_just_gave_me_the_flu/ whats the problem here?
21:37:36FromGitter<zetashift> my Nim-fu i still weak so i cant help him
21:37:44*d10n-work quit (Quit: Connection closed for inactivity)
21:50:47*nsf joined #nim
21:55:33*ShalokShalom quit (Remote host closed the connection)
21:58:02sendellhow can nim-glm require "nim >= 0.17.3" ? does that mean nightly?
22:01:28skrylarpossibly
22:01:29bluenoteyes, just checkout the devel branch and re-run ./koch -d:release boot
22:01:42skrylarglm is the math lib right?
22:01:57skrylarseems a little fey for a math lib to require nightly
22:03:06*nsf quit (Quit: WeeChat 1.9.1)
22:03:07bluenoteit's not unusual to depend on devel, Nim is moving quickly
22:03:19*skrylar targets the latest stable
22:05:27sendellthx bluenote
22:05:40sendellis nightly stable enough?
22:05:56skrylareeh
22:06:36skrylarI don't think Araq would tolerate a build being broken for too long, but the nature of nightlies is they are by definition not stable :)
22:08:06Araqnightly is stable enough but I will make it more stable
22:08:31Araqsomeone needs to give me some hook "CI green --> merge into master"
22:09:03Araqthen everybody can use the master branch, right now master reflects the releases which isn't all that useful since we have git tagging for that
22:10:25FromGitter<techate> @zetashift , he wants to know how to use cstrings safely, without them getting garbage-collected while they aren't garbage. the answer is probably A) use those functions, or B) keep them on the stack, or C) ensure they're always referenced by something the GC does care about. (it's also possible that he thinks that the term 'conservative GC' is spooky.)
22:11:17sendelli'll try installing it with choosenim :)
22:12:20skrylarAraq, debian testing bot? lol
22:14:56skrylarsomething like that can probably be done with buildbot
22:15:29*gokr quit (Remote host closed the connection)
22:21:50FromGitter<mratsim> I understand why GLM would want to depend on nightly / master
22:22:46sendellwhy?
22:22:58FromGitter<mratsim> I have this code to deal with the BackwardsIndex changes: ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ !eval echo “Please release 0.18 ASAP" [https://gitter.im/nim-lang/Nim?at=5a1355c2df09362e67326eb4]
22:23:47skrylarbleh. grsec has committed harikiri
22:23:53FromGitter<mratsim> every lib that do a[index] must introduce something similar if they want to be backward compatible
22:26:14*laas quit (Quit: WeeChat 1.9.1)
22:28:47Araqnew blog post has arrived: https://nim-lang.org/araq/karax.html
22:29:02skrylarwonder if anyone is running nim on fbsd
22:29:10skrylardon't see why it wouldn't work
22:29:52*ryanhowe joined #nim
22:30:26FromGitter<mratsim> @Araq: I pre-quote @dom96’s “I see what you did there” regarding "Unfortunately the nimble package does not install the examples so to tinker with; git clone is an alternative"
22:30:45skrylarAraq, i would probably break that up in to several pages but seems neat
22:31:15Araqskrylar, boy you have no idea, initially it was twice as long
22:31:17Araq:D
22:33:21skrylarhave contemplated if it would be worthwhile to try to process docbook with nim :p
22:33:29skrylarprobably easier to figure out how to get texinfo to use a stylesheet though
22:34:34AraqI use 'nim rst2hml x.rst' for my blog with a custom config
22:35:17skrylareh well.
22:35:29skrylaryou saw the docs i did for firebird. its more book than doc page. lol
22:36:05FromGitter<mratsim> that’s so alien to everything in the manual documentation "include karax / prelude”. I got stuck there, and then: ⏎ ⏎ ```nim c tools/karun ⏎ tools/karun -r helloworld.nim``` [https://gitter.im/nim-lang/Nim?at=5a1358d5540c78242d2a96ba]
22:36:19AraqI haven't, can you send me the link again, skrylar ?
22:36:46FromGitter<mratsim> tools/karun is?
22:37:15Araqa tool?
22:37:21Araqwhat do you mean?
22:37:27skrylarAraq, https://www.dropbox.com/s/p7vu32jdemfz2fy/firebird3.pdf?dl=0
22:38:02FromGitter<mratsim> the blog post doesn’t say that the tiny build tool is called karun.
22:39:37Araqnow it does
22:40:26FromGitter<mratsim> :D
22:41:15Araqskrylar, do you like LaTeX?
22:41:42skrylarAraq, i dunno, texinfo made that pdf
22:43:06skrylarits just hard to show the html version since i don't 'have a web host anymore
22:43:13skrylarcould have done the single page bulk export
22:45:03*nhywyll quit (Quit: nhywyll)
22:46:16skrylarFrom a wide search, it seems that the most reliable form of delivering help for desktop linux apps is.. a pdf.
22:47:15skrylarthough now that many distros have a "sensible-exec" type script, though theres maybe 7 different names for it, html inside /usr/share seems to be possible
22:49:56FromGitter<techate> last I heard xdg-open was the winner, adopted by BSD and Linux distro alike. There might also be some old names surviving alongside it
22:50:40skrylari saw a sensible-something name in xfce4 a few days ago
22:51:17FromGitter<techate> yep, I've got that as well. but I also have xdg-open
22:51:28skrylarBSD is just going to adopt whatever linux has. They're kinda non-factors in desktop space, but their work ethic is vastly more commendable than linus
22:52:20skrylaranyway. never actually used karax
22:52:36skrylardid a little with the JS backend and found the docs were quite lacking
22:52:54skrylarhad to find some weird chinese guy's gists to get basic content replacement down
22:54:44*elrood quit (Quit: Leaving)
22:55:19*marenz__ joined #nim
22:55:49skrylarafter cobweb is running, it might be amusing to .. i dunno, have some macro for the immgui freaks
22:56:02skrylarthat just secretly hooks up triggers to normal retained guis in the background
22:59:58*ketralnis joined #nim
23:04:50*rbrt quit (Quit: Oíche mhaith)
23:08:09*bluenote quit (Ping timeout: 260 seconds)
23:09:21Araqmratsim: instead of the version check you're supposed to check 'defined(nimNewRoof)'
23:10:20*onionhammer quit (Quit: WeeChat 1.9.1)
23:10:41skrylarand then in nim 2.6 there is nimNewRoofRly
23:10:59AraqnimNewNewRoof
23:11:05*onionhammer joined #nim
23:11:05AraqnimNewNewNewRoof
23:11:20Araqseems like a web scale solution to me
23:11:57skrylarnimQuiteTallCeiling was the best update. araq should have never iterated past that
23:12:18FromGitter<mratsim> well I don’t plan to keep backward compatibility for 10 years
23:13:48FromGitter<mratsim> I will probably move on to a new shiny language ParaSailSwiftlyWithRustyNim
23:15:02Araqwe all will but it will be called 'Nim'
23:15:20FromGitter<mratsim> With true polymorphism: syntax polymorphism, you can inherit from Rust BorrowChecker, Swift reference counting, Go GC :P
23:15:21Araqyou know the old joke about fortran, right?
23:15:40*skrylar rewrites mratsim in Dylan
23:15:43FromGitter<mratsim> I know the one about powerful languages reinventing Lisp ;)
23:16:11Araq"I don't know what the programming language of the year 2000 will look like, but I know it will be called FORTRAN"
23:18:21sendellSo... I bootstrapped 0.17.3 using koch to satisfy glm dependecy, but now zengine wont compile because of the indexing breaking change you were talking about, i guess: "Error: type mismatch: got (HSlice[system.float, system.int])". I guess i'm stuck here :p
23:18:49sendelllib and its dep wont compile using the same nim :/
23:19:25FromGitter<mratsim> open a ticket on zengine repo, zach is usually very fast to fix things
23:20:10dom96Araq: Can we prevent these breakages?
23:22:08AraqI doubt it
23:23:47Araqotherwise we would be version 1, right? you all keep voting for a better language
23:24:11Araqwe can make v0.18 be v1.0 though :P :P :P
23:26:03dom96:/
23:26:08dom96I can't say I voted for this change
23:26:26dom96Seems to me that it fixes some obscure issue at the expense of Nim's expressivity
23:27:12Araqbig words for a library solution that was replaced by some other library solution
23:27:45skrylaris the index operator borked or something
23:29:05*Jesin quit (Quit: Leaving)
23:29:33Araqnot borked, I rewrote the implementation in order to make it cleaner and fix some nasty bugs
23:30:09Araqbut I think this is about the changes I made to the '..' iterator
23:30:45Araqwhich btw never was consistent with 'countdown' or '..<' but consistency is only an argument when you feel like it
23:31:35*arnetheduck quit (Ping timeout: 240 seconds)
23:32:26skrylardidn't know ..< was a thing
23:32:33skrylari always did the annoying 0..(thing-1)
23:33:14sendelli guess learning nim on a nightly compiler was a bad idea anyway :p
23:33:43skrylarit kinda is
23:43:02*vlad1777d quit (Ping timeout: 255 seconds)