<< 23-02-2018 >>

---

00:04:19	*	miran_ quit (Ping timeout: 260 seconds)
00:22:43	FromGitter	<nitely> yes, like nimYAML
00:29:23	FromGitter	<nitely> nimYAML could have an API that require just a file name, and then the one that takes a stream for special cases I can't think off
00:30:12	*	MJCaley quit (Quit: MJCaley)
00:45:23	*	Jesin quit (Quit: Leaving)
00:59:22	*	MJCaley joined #nim
01:09:10	GitDisc	<pooboy> Hello guys !!
01:11:08	FromGitter	<zetashift> yello!
01:12:34	*	Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
01:14:32	GitDisc	<pooboy> nim is awesome !!!
01:14:53	GitDisc	<pooboy> How to use the modulus operator ?
01:17:11	FromGitter	<zetashift> I agree! And modulus is used like ```a mod b```
01:17:23	FromGitter	<zetashift> !eval 4 mod 2
01:17:24	NimBot	Compile failed: in.nim(1, 3) Error: expression '0' is of type 'int literal(0)' and has to be discarded
01:17:37	FromGitter	<zetashift> whoops
01:17:54	FromGitter	<zetashift> !eval echo(4 mod 2)
01:17:56	NimBot	0
01:18:19	FromGitter	<zetashift> this is what I get for learning multiple languages at the same time ><
01:21:05	GitDisc	<pooboy> Lol
01:21:16	GitDisc	<pooboy> Thanks !!
01:21:25	FromGitter	<zetashift> Are you coming from other languages @pooboy?
01:25:08	GitDisc	<pooboy> Yes..im coming from ruby,python , java ..im no programmer..im an enthusiast :)
01:25:46	FromGitter	<zetashift> @pooboy , this might help: https://github.com/nim-lang/Nim/wiki/Nim-for-Python-Programmers and http://blog.zdsmith.com/posts/nim-for-python-programmers.html
01:26:01	FromGitter	<zetashift> For quick small experiments you can use: https://play.nim-lang.org/
01:26:10	FromGitter	<zetashift> I'm actually in the same boat
01:27:07	GitDisc	<pooboy> 👌 👌 👌 👌 👌
01:27:33	GitDisc	<pooboy> Can i create my own package for nim ?
01:27:59	FromGitter	<zetashift> yes check out: https://github.com/nim-lang/nimble
01:28:40	GitDisc	<pooboy> Great !
01:53:56	*	yglukhov joined #nim
01:54:03	GitDisc	<pooboy> can we cross-compile binaries ?
01:55:11	FromGitter	<zetashift> @pooboy, yes see: https://nim-lang.org/docs/nimc.html#cross-compilation
01:55:35	FromGitter	<zetashift> I don't have any experience with it though but if you run into troubles just ask here or open a forum thread
01:58:05	*	yglukhov quit (Ping timeout: 240 seconds)
02:04:01	GitDisc	<pooboy> niceeeee
02:16:22	FromGitter	<zacharycarter> I guess I need to work on the playground this weekend :) will get me back into writing some Nim
02:18:47	GitDisc	<pooboy> 😇 😇 😇
02:19:52	GitDisc	<pooboy> how to write and call a procedure
02:21:13	FromGitter	<zacharycarter> pooboy: https://play.nim-lang.org/?gist=c2874ef069977ca708fd46ab6466cf6e
02:21:48	FromGitter	<zacharycarter> also please read the manual, answers to simple questions like yours are easily found there - https://nim-lang.org/docs/manual.html
02:24:06	GitDisc	<pooboy> Ohh cool
02:24:09	GitDisc	<pooboy> Thanks
02:24:49	FromGitter	<zacharycarter> sure thing
02:25:10	FromGitter	<zetashift> @pooboy also the https://nim-lang.org/docs/tut1.html#procedures official tutorial covers imho everything you need to make programs in Nim
02:25:18	FromGitter	<zetashift> @zacharycarter how's the gamedevving?
02:26:39	FromGitter	<zacharycarter> @zetashift - I haven't been doing much lately. I just started in the past few weeks to write the core of an engine, but I'm writing it in C++. I plan to add support for Nim as a language for gameplay code.
02:28:32	FromGitter	<zetashift> Have you seen http://kha.tech/ you might be able to draw some inspiration from that, IIRC it's core is written in C++
02:28:58	FromGitter	<zetashift> and with the wealth of information about C++ and gamedev I wouldn't blame you for writing it in C++
02:30:29	FromGitter	<zacharycarter> Ah I guess it's from the armory folks? Looks like they have a link to that project in the screenshot frame.
02:31:09	FromGitter	<zetashift> Armory is based on it yes but Kha is from a different person
02:31:19	FromGitter	<zacharycarter> gotcha
02:31:51	GitDisc	<pooboy> Noted
02:32:07	FromGitter	<zetashift> when I tried digging it into it they said don't think of it as an engine but as an alternative to SDL
02:34:30	FromGitter	<zacharycarter> I don't know - some of their features make no sense to me... like the ability to run on top of Unity. Why on earth would you want to run something like SDL on top of Unity?
02:36:09	FromGitter	<zetashift> I believe they added it so they could target consoles too as they didn't have the official SDKs
02:36:37	FromGitter	<zacharycarter> gotcha
02:42:21	*	poopBotTelefon quit (Ping timeout: 264 seconds)
02:50:23	*	MJCaley quit (Quit: MJCaley)
02:57:42	*	athenot_ joined #nim
02:58:12	*	athenot quit (Ping timeout: 260 seconds)
03:09:04	*	vlad1777d quit (Ping timeout: 268 seconds)
03:10:53	*	Faster-Fanboi_ quit (Quit: ZNC 1.7.x-git-876-42939c9 - https://znc.in)
03:12:36	*	Faster-Fanboi joined #nim
03:50:43	FromGitter	<zetashift> Can anyone help me why I get an out of bounds error? https://play.nim-lang.org?gist=8c1998c4d9932e0b643bd075fd806122
03:51:29	FromGitter	<zetashift> I'm following a data structures book with Nim and I've done this before in JS but that doesn't translate too well
03:52:46	FromGitter	<zetashift> doing ``` var existingNumbers = newSeqint (a.len) removes it but I don't get the correct answer then
03:53:05	FromGitter	<zetashift> whoops I mean ``` var existingNumbers = newSeqint (a.len) ```
03:56:25	FromGitter	<zetashift> it seems to always branch out to the true clause so I'm making a logic error but can't figure it quite out
04:05:47	FromGitter	<zetashift> Fixed the out of bounds error and the correct boolean operator is != and nim initializes int seqs with 0's
04:07:23	*	tefter quit (Remote host closed the connection)
04:15:36	FromGitter	<zetashift> and now it always prints false; Should I use arrays instead? https://play.nim-lang.org?gist=fb805d73085178df67094213cba633be
04:17:31	*	SenasOzys quit (Ping timeout: 268 seconds)
04:30:09	*	Sembei joined #nim
04:32:19	*	Pisuke quit (Ping timeout: 268 seconds)
04:46:11	*	Faster-Fanboi quit (Ping timeout: 256 seconds)
04:52:57	*	Faster-Fanboi joined #nim
05:07:01	*	dddddd quit (Remote host closed the connection)
05:07:19	*	pecanpie joined #nim
05:13:41	FromGitter	<Varriount> zetashift: I don't see how that algorithm is supposed to detect duplicate entries.
05:14:09	*	Faster-Fanboi quit (Ping timeout: 248 seconds)
05:14:51	*	pecanpie quit (Quit: Leaving)
05:15:03	FromGitter	<Varriount> You either need a second, nested `for` loop, or a `Set[int]` to store numbers that have been seen.
05:16:56	FromGitter	<Varriount> (Note that `Set[int]` is not the same as the built-in `set` type.)
05:20:52	*	Faster-Fanboi joined #nim
05:31:58	*	mal`` quit (Quit: Leaving)
05:41:42	*	mal`` joined #nim
05:46:24	*	r3d9u11 joined #nim
05:58:12	*	radagast_04 joined #nim
06:08:47	*	nsf joined #nim
06:25:50	*	r3d9u11 quit (Remote host closed the connection)
06:29:42	*	rockcavera quit (Remote host closed the connection)
06:38:54	GitDisc	<ethanpmorgan> Bonjour
06:38:58	GitDisc	<ethanpmorgan> My dudes
06:43:46	FromGitter	<zetashift> @Varriount I need a proc that checks for duplicates in a single loop; Everytime I encounter a new number I store it a 1 at the index of the corresponding number in existingNumbers
06:45:00	FromGitter	<Varriount> Yes, but
06:45:17	FromGitter	<zetashift> @ethanpmorgan hello! how's it going?
06:45:56	FromGitter	<Varriount> All the function will do is create an array filled with 1's
06:46:29	FromGitter	<epmor> Hey, gave up with Haskell, Nim seems like a great language to learn programming with, changed my previous project to from Haskell to Nim and Nim's quite rich in the stdlib compared to Haskell
06:46:31	FromGitter	<zetashift> here's the js: https://pastebin.com/9RTerCEP
06:47:44	FromGitter	<Varriount> You can't do it in a single loop, as you need to check whether the current number has been seen before
06:49:06	*	rockcavera joined #nim
06:49:11	FromGitter	<zetashift> mhhmmh, I guess I've just misinterpreted something wrong while reading; thanks for some clarity!
06:51:25	FromGitter	<zetashift> @epmor I had the same with Scala; still taught me alot about programming in general
06:53:28	FromGitter	<epmor> Yeah I've been jumping language to language to see what I like, but none of them seems "fun", so far Nim's simple and speedy so it's a win win. And the community is smallish which is nice
07:07:39	FromGitter	<tim-st> @zetashift something like `len(openArray) != len(set(openArray))` would solve your problem
07:07:59	FromGitter	<tim-st> (pseudo code)
07:09:54	*	r3d9u11 joined #nim
07:21:17	*	SenasOzys joined #nim
07:21:19	*	yglukhov joined #nim
07:25:48	*	yglukhov_ joined #nim
07:26:54	*	yglukhov_ quit (Client Quit)
07:27:57	*	yglukhov quit (Ping timeout: 264 seconds)
07:28:53	*	yglukhov joined #nim
07:29:04	*	miran_ joined #nim
07:30:48	*	rockcavera quit (*.net *.split)
07:30:48	*	avsej quit (*.net *.split)
07:30:48	*	Araq quit (*.net *.split)
07:30:48	*	derlafff quit (*.net *.split)
07:30:48	*	mgdelacroix[m] quit (*.net *.split)
07:30:48	*	pydsigner quit (*.net *.split)
07:31:20	*	joshbaptiste quit (*.net *.split)
07:31:20	*	subsetpark quit (*.net *.split)
07:31:20	*	hohlerde quit (*.net *.split)
07:31:20	*	jonafato quit (*.net *.split)
07:36:57	*	yglukhov quit (Read error: Connection reset by peer)
07:40:53	FromGitter	<Varriount> @tim-st Wouldn't it be better to just maintain a set of already seen integers, and check/add to it every loop?
07:41:23	FromGitter	<Varriount> That way you can bail out early if a duplicate is found.
07:48:40	FromGitter	<epmor> How do you use case with object types?
07:49:03	FromGitter	<epmor> I'll pastebin my code
07:49:23	FromGitter	<epmor> https://pastebin.com/wazpaKZp
07:50:11	FromGitter	<epmor> I haven't finished it, hoping to get it to be more usable, but just trying to get the case working with the object/struct
07:50:59	*	miran_ left #nim (#nim)
07:51:07	*	miran_ joined #nim
07:51:46	*	miran_ left #nim (#nim)
07:52:22	*	miran1 joined #nim
07:53:22	*	joshbaptiste joined #nim
07:53:22	*	subsetpark joined #nim
07:53:22	*	hohlerde joined #nim
07:53:22	*	jonafato joined #nim
07:56:07	*	pydsigner joined #nim
07:56:35	*	miran1 quit (Client Quit)
07:57:34	*	rockcavera joined #nim
07:57:34	*	avsej joined #nim
07:57:34	*	Araq joined #nim
07:57:34	*	derlafff joined #nim
07:57:34	*	mgdelacroix[m] joined #nim
07:58:31	*	miran joined #nim
07:59:49	*	mgdelacroix[m] quit (Ping timeout: 240 seconds)
08:00:02	*	byteflame quit (Ping timeout: 248 seconds)
08:00:02	*	survivorm[m] quit (Ping timeout: 248 seconds)
08:00:03	*	ehmry quit (Ping timeout: 248 seconds)
08:00:03	*	dyce[m] quit (Ping timeout: 248 seconds)
08:00:09	*	hohlerde quit (Ping timeout: 255 seconds)
08:00:13	*	notdekka[m] quit (Ping timeout: 240 seconds)
08:00:13	*	xincognito10[m] quit (Ping timeout: 240 seconds)
08:00:14	*	hiway quit (Ping timeout: 240 seconds)
08:00:14	*	unclechu quit (Ping timeout: 240 seconds)
08:00:18	*	planetis[m] quit (Ping timeout: 255 seconds)
08:00:21	*	Jipok[m] quit (Ping timeout: 252 seconds)
08:00:22	*	TheManiac quit (Ping timeout: 260 seconds)
08:00:22	*	Miguelngel[m] quit (Ping timeout: 260 seconds)
08:00:22	*	Yardanico quit (Ping timeout: 260 seconds)
08:00:22	*	Demos[m] quit (Ping timeout: 260 seconds)
08:00:50	*	macsek1911[m] quit (Ping timeout: 276 seconds)
08:00:51	*	shashlick quit (Ping timeout: 276 seconds)
08:01:12	*	tiorock joined #nim
08:01:12	*	tiorock quit (Changing host)
08:01:12	*	tiorock joined #nim
08:01:12	*	rockcavera quit (Killed (wolfe.freenode.net (Nickname regained by services)))
08:01:12	*	tiorock is now known as rockcavera
08:01:37	FromGitter	<epmor> Nevermind I got it
08:03:25	*	Ven`` joined #nim
08:04:40	FromGitter	<mratsim> @epmor, if you have questions on how to translate some Haskell constructs to Nim feel frree, several of us have Haskell or Scala background and added libraries for idiomatic (but fast) functional programming (chaining zip, map, fold, filter, etc)
08:07:08	FromGitter	<mratsim> @Lite5h4dow it’s me who told you you can use streams. It’s more of an advanced way to read anything (files, network, Gzip on the-fly uncompression). basically you open a handler to a “stream” (a file read line by line, a network stream, a file uncompressed on-the fly, a download ...), read some, process, read some, process on the fly, etc.
08:07:25	FromGitter	<epmor> Ah thanks @mratsim Just trying to implement Result into Nim
08:07:27	FromGitter	<mratsim> It can be a camera feed, a mic
08:09:12	FromGitter	<mratsim> @epmor, you can create a tuple: `type Result[T]: tuple[value: T, error: YourErrorSuccessEnumType]`
08:09:51	FromGitter	<mratsim> but in that case Nim uses exceptions rather than returning a result type like in Haskell Rust or Go.
08:10:10	FromGitter	<mratsim> I mean, it’s idiomatic to use exceptions
08:10:35	*	SenasOzys quit (Remote host closed the connection)
08:11:48	*	Ven`` quit (*.net *.split)
08:11:48	*	avsej quit (*.net *.split)
08:11:48	*	Araq quit (*.net *.split)
08:11:48	*	derlafff quit (*.net *.split)
08:11:50	*	joshbaptiste quit (*.net *.split)
08:11:50	*	subsetpark quit (*.net *.split)
08:11:50	*	jonafato quit (*.net *.split)
08:11:56	FromGitter	<epmor> Ah I see
08:12:00	FromGitter	<epmor> Cheers my dude
08:13:37	FromGitter	<epmor> Wanna make something that'll be used or useful
08:16:31	*	Ven`` joined #nim
08:16:37	*	Arrrr joined #nim
08:21:32	*	PMunch joined #nim
08:22:27	*	subsetpark joined #nim
08:22:33	*	jonafato joined #nim
08:22:38	*	joshbaptiste joined #nim
08:23:18	Arrrr	https://www.reddit.com/r/programming/comments/7zird1/nim_at_fosdem_frequently_asked_questions/
08:24:57	FromGitter	<mratsim> @epmor You can have a look at nimfp for example: https://github.com/vegansk/nimfp
08:25:12	FromGitter	<mratsim> For example the Either type: https://github.com/vegansk/nimfp/blob/master/src/fp/either.nim
08:26:40	FromGitter	<epmor> Ah sweet, thanks @mratsim
08:30:21	PMunch	Arrrr, did you like it?
08:31:56	Arrrr	Yes. But people on reddit will not know how does nim look like.
08:33:12	*	avsej joined #nim
08:33:12	*	avsej quit (Changing host)
08:33:13	*	avsej joined #nim
08:33:14	PMunch	Hmm, that's true, but it's a quick google search away :)
08:33:34	*	Araq joined #nim
08:34:01	Arrrr	Imagine people are lazy
08:34:06	*	derlafff joined #nim
08:34:45	*	xkapastel quit (Quit: Connection closed for inactivity)
08:35:06	FromGitter	<mratsim> You have 5 sec to capture someone attention on a webpage
08:35:42	FromGitter	<mratsim> at least that’s what growth hackers, designers and frontend dev work with
08:38:15	PMunch	Hmm, maybe I could sprinkly some code examples in there..
08:42:21	FromGitter	<mratsim> apparently last npm upgrade crashed all linux servers
08:43:20	PMunch	mratsim, yup pretty brutal :P
08:43:32	PMunch	It changed a bunch of permissions for system folders like /etc
08:44:10	FromGitter	<mratsim> I wonder if nimble did that how many people would hear of Nim :evilgrin:
08:45:15	PMunch	Haha, what kind of sysadmin would end up reinstalling over some broken file permissions?
08:46:20	FromGitter	<mratsim> Someone who didn’t have to deal with systemd or pulseaudio 10 years ago ;)
08:47:45	PMunch	Hmm, is GitHub down for you too?
08:48:00	FromGitter	<mratsim> I don’t even understand why you would deliver your production packages through npm without a docker or LXC or VM. It’s like Python, versioning/compat is too random
08:48:31	FromGitter	<mratsim> Works for me™
08:48:44	PMunch	Hmm, I can't pull repos or visit their main site..
08:51:22	FromGitter	<honewatson> Just wanted to say thanks to the Nim crew. You have done an amazing job! Its almost unbelievable what you have achieved!
08:52:28	PMunch	Agreed!
08:59:46	*	floppydh joined #nim
09:01:20	Arrrr	All the bugs that arise from naming both type and module the same ...
09:01:35	*	hiway joined #nim
09:03:56	Araq	Arrrr: filenames should be all lowercase anyway :P
09:04:14	Araq	honewatson: thank you. :-)
09:10:10	*	rokups joined #nim
09:14:54	*	dyce[m] joined #nim
09:14:54	*	Yardanico joined #nim
09:14:54	*	shashlick joined #nim
09:14:55	*	Miguelngel[m] joined #nim
09:14:55	*	unclechu joined #nim
09:14:55	*	mgdelacroix[m] joined #nim
09:14:55	*	Demos[m] joined #nim
09:14:55	*	xincognito10[m] joined #nim
09:15:01	*	byteflame joined #nim
09:15:02	*	macsek1911[m] joined #nim
09:15:02	*	TheManiac joined #nim
09:15:02	*	hohlerde joined #nim
09:15:02	*	notdekka[m] joined #nim
09:15:02	*	planetis[m] joined #nim
09:15:02	*	ehmry joined #nim
09:15:03	*	survivorm[m] joined #nim
09:15:04	*	Jipok[m] joined #nim
09:30:03	*	yglukhov joined #nim
09:34:22	*	couven92 joined #nim
09:41:14	FromGitter	<gogolxdong> @PMunch are you writing a protobuf library in Nim ?
09:44:05	*	Ven` joined #nim
09:44:06	*	Ven`` quit (Read error: Connection reset by peer)
09:45:18	PMunch	Yup
09:45:46	FromGitter	<gogolxdong> How is it going?
09:47:42	FromGitter	<gogolxdong> Where to apply ?
09:49:37	PMunch	Apply?
09:49:42	PMunch	It's going pretty well
09:50:18	PMunch	I'm developing a parser library as well to parse the file with. So currently I'm working on improving the error messages from the parser so it's easier to see what you did wrong
09:50:38	PMunch	But for parsing correct data it works pretty well
09:50:52	PMunch	And I've implemented the varargs and sig-sag signed ints
09:50:53	FromGitter	<gogolxdong> IPFS uses protobuf.
09:51:32	PMunch	So the only thing thats really left is to generate the types and the procs to read and write messages
09:52:11	PMunch	IPFS?
09:52:32	PMunch	https://github.com/PMunch/protobuf-nim
09:52:38	PMunch	https://github.com/PMunch/combparser
09:52:47	PMunch	That's the repos for both project
09:52:48	PMunch	s
09:52:50	FromGitter	<Vindaar> InterPlanetary FileSystem ;)
09:53:27	PMunch	Hmm, interesting
09:54:04	FromGitter	<gogolxdong> thanks . I will check and see whether can give a hand.
09:55:37	PMunch	That'd be great :)
09:57:36	PMunch	The protobuf-nim repository is kinda messy :P
09:57:55	PMunch	So don't hesitate to ask questions
09:59:38	*	yglukhov quit (Remote host closed the connection)
10:00:21	couven92	dom96, I didn't have time watching you live, but I am watching the YT now! Looking forward to it!!!! :)
10:02:07	couven92	tea actually sound real nice, hmm... git to make myself a cup first :P
10:03:07	*	Vladar joined #nim
10:05:34	FromGitter	<gogolxdong> Though we can't access into IPFS due to Data Security Law , it's a paradim-shifting practice which our government encourages.
10:07:18	FromGitter	<gogolxdong> What we can do is a interChinese FileSystem.
10:15:16	FromGitter	<narimiran> test 1 2
10:16:33	*	miran quit (Quit: Page closed)
10:18:36	FromGitter	<narimiran> ok, this works (i was having problem with IRC this morning). it's miran, btw, if it was not clear....
10:25:46	Araq	so ... again the question: how do we ensure imports from a nimble package?
10:25:52	Araq	import nimble / random ?
10:26:05	Araq	but nimble itself is a nimble package so that would be ambiguous
10:33:54	FromGitter	<andreaferretti> for what it's worth, there could be two nimble packages defining the same module
10:34:09	FromGitter	<andreaferretti> so it is not only an issue of nimble vs stdlib
10:34:49	FromGitter	<andreaferretti> path import are kind of a solution right now
10:35:07	FromGitter	<andreaferretti> import "nim-random/v0.6/random"
10:35:15	FromGitter	<andreaferretti> or something like that, I don't remember
10:35:29	FromGitter	<andreaferretti> but you have to know the actual path in the nimble directory
10:35:43	FromGitter	<andreaferretti> and it depends on the version number, which is unfortunat
10:35:44	FromGitter	<andreaferretti> e
10:36:39	FromGitter	<narimiran> doesn't sound like an elegant solution.... :(
10:37:06	FromGitter	<andreaferretti> it is not
10:37:15	FromGitter	<andreaferretti> just telling what works right now
10:38:18	FromGitter	<andreaferretti> there should be a way to define a namespace in your nimble package
10:38:24	FromGitter	<andreaferretti> so that one could do
10:38:31	FromGitter	<andreaferretti> import mynamespace/mymodule
10:38:44	FromGitter	<andreaferretti> or just import mymodule if it is not ambiguous
10:39:18	FromGitter	<narimiran> shouldn't nimble authors be the ones responsible to not use the same names as std lib?
10:39:49	FromGitter	<andreaferretti> yup
10:40:00	FromGitter	<andreaferretti> but modules can be added to the stdlib
10:40:02	couven92	narimiran: Yeah, how well is that going to go?
10:40:07	FromGitter	<narimiran> and if some conflict arises currently because of additions to std lib - well, nim is still in pre-1.0 phase and this should be expected
10:40:10	FromGitter	<andreaferretti> I think this is what happened with random
10:40:22	FromGitter	<narimiran> and you and your package need to find another name
10:40:43	FromGitter	<andreaferretti> the real problem is conflicts between modules in different nimble packages
10:41:27	Araq	most nimble packages do require import mynamespace/mymodule
10:42:52	FromGitter	<andreaferretti> yeah, but is there any check on uniqueness of namespaces?
10:43:44	FromGitter	<andreaferretti> I think something was done when nimble started adding warnings for modules named differently from the nimble package
10:49:59	Araq	the problem is that 'import jester/jester' or import random/random is not sexy and so Nimble allows for a different directory structure where 'import jester' works
10:50:37	Araq	which is fine for names like 'jester' that do not have technical names like 'random', I guess
10:55:10	*	SenasOzys joined #nim
11:03:09	dom96	andreaferretti: the package name **is** the namespace
11:03:51	dom96	And indeed, the warnings are there to ensure uniqueness
11:03:55	dom96	Eventually they will be errors
11:06:37	FromGitter	<andreaferretti> but can one do `import packagename/modulename`?
11:06:46	FromGitter	<andreaferretti> that is, is this supported
11:06:47	FromGitter	<andreaferretti> ?
11:07:08	FromGitter	<andreaferretti> I only remember `import "packagename/$version/modulename"`
11:07:44	dom96	yes
11:08:20	dom96	Where are you getting `import "packagename/$version/modulename"` from?
11:08:25	dom96	That was never possible
11:09:34	couven92	LOL, dom96 "nimisawesome" used 6 times! :D
11:09:43	dom96	8 times, no? :)
11:10:00	couven92	ah, low res on my small-on-the-side-screen
11:10:02	couven92	sry
11:10:12	couven92	yeah, 8 times! :)
11:10:19	FromGitter	<andreaferretti> @dom96 it is possible because it is the actual path on disk of the nim file
11:10:20	dom96	yeah, I wish I remembered to change the font size :(
11:10:38	dom96	andreaferretti: but it's not. The path is packagename-$version/modulename
11:10:44	couven92	and go was not awesome! :P
11:10:48	FromGitter	<andreaferretti> ah ok sorry
11:10:59	FromGitter	<andreaferretti> I got that wrong but it was what I meant
11:11:20	FromGitter	<andreaferretti> the question remains
11:11:22	dom96	also that requires your nim path to contain ~/.nimble/pkgs
11:11:26	dom96	does it by default?
11:11:39	FromGitter	<andreaferretti> whe. run by nimble apparently it does
11:12:07	couven92	dom96, though tbh, as a git teacher at UiT, I take issue at you running `git init` AFTER you were done coding the main parts! :P
11:12:23	FromGitter	<andreaferretti> the point is `import "packagename-$version/modulename"` is inconvenenient
11:12:29	couven92	AND calling your commit "Initial Commit"
11:12:39	FromGitter	<andreaferretti> does `import packagename/modulename` work?
11:12:55	dom96	couven92: oh? I was never taught git formally
11:12:58	FromGitter	<andreaferretti> otherwise I would not say that packagename is a namespace
11:12:59	FromGitter	<alehander42> I always call my first commit "First" :D
11:13:09	FromGitter	<andreaferretti> except in some philosophical sense
11:13:51	*	derlafff quit (Remote host closed the connection)
11:13:57	dom96	andreaferretti: it works for modules that the package puts in a 'packagename' folder
11:14:06	*	derlafff joined #nim
11:14:14	dom96	`import "packagename-$version/modulename"` shouldn't work
11:14:14	couven92	Yeah, I don't teach a subject with syllabus on Git, but I have been giving evening classes for three years now...
11:14:19	dom96	Amazed that it does
11:14:56	dom96	couven92: I can see why calling 'git init' earlier is a good idea, but what's wrong with my commit message?
11:15:16	couven92	dom96, it does not describe what you did in you commit at all!
11:15:21	FromGitter	<andreaferretti> @dom96 then it's just the path
11:15:40	dom96	couven92: I see lol
11:16:08	couven92	Normally I put the .gitignore and LICENSE in the Initial Commit
11:16:14	FromGitter	<andreaferretti> what I meant is
11:16:29	couven92	(in cases where those two need to be there, but do not really matter)
11:16:57	FromGitter	<andreaferretti> if I call my package `strutils` and it only consists of a file `strutils.nim`, I should be able of doing `import strutils/strutils`
11:17:18	FromGitter	<andreaferretti> this would distinguish it from the stdlib module
11:17:32	FromGitter	<andreaferretti> because the package name acts as namespace
11:17:42	dom96	andreaferretti: the only way to implement this would be to special case it in the compiler though
11:18:10	FromGitter	<andreaferretti> I know, that is why I was asking for the compiler to support a notion of namespace
11:18:24	FromGitter	<andreaferretti> which then would be the package name for nimble packages
11:19:18	FromGitter	<andreaferretti> an alternative that does not require changes in the compiler would be
11:19:19	couven92	dom96, you said you wanted to write your own Password manager, PMunch and I wanted to write a command-line integration for KeePass in Nim :)
11:19:39	FromGitter	<andreaferretti> for nimble to use the current directory structure
11:19:57	*	Ven` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
11:20:00	*	lesshaste joined #nim
11:20:11	FromGitter	<andreaferretti> but put in the path some directories named packagename, which are symlinks to packagename-$version
11:20:24	FromGitter	<andreaferretti> for the currently used version
11:21:03	FromGitter	<andreaferretti> (this breaks on windows and is probably unsafe to run two instances of nimble concurrently though)
11:21:55	PMunch	couven92, that is actually partially done :)
11:22:17	couven92	PMunch, get dom96 to join! :P
11:22:40	dom96	KeePass corrupted my password DB once
11:22:49	couven92	oooh...
11:22:51	dom96	So i'll be staying away from it forever now :)
11:23:52	*	vlad1777d joined #nim
11:24:21	PMunch	Ouch, that's not good
11:27:25	lesshaste	how does nim handle multiplication of complex numbers in terms of SIMD speedups?
11:28:00	*	lesshaste quit (Quit: Leaving)
11:28:38	*	lesshaste joined #nim
11:29:24	dom96	andreaferretti: yeah... so any other ideas? :)
11:29:54	*	r3d9u11 quit (Remote host closed the connection)
11:32:18	couven92	dom96, we saw that UNIX (or POSIX) has getpass for password, Windows has CredMan Win32 API functions to enter the password
11:32:31	couven92	create a password-terminal package? :O
11:32:42	dom96	Do it :P
11:32:45	couven92	:D
11:33:35	*	vlad1777d quit (Ping timeout: 240 seconds)
11:33:37	couven92	Though the Windows thingy is REALLY advanced... It will actually print the hostname and username questions on the command-line in your native locale
11:34:55	*	jackprob joined #nim
11:35:21	couven92	eating the chars for the password on the command-line is not a Linux thing, native Windows, also eats password chars!
11:36:13	*	MJCaley joined #nim
11:37:12	dom96	:o
11:37:15	dom96	But whyyy
11:37:16	FromGitter	<alehander42> o.O I didn't know that
11:37:57	couven92	because writing asterisks is a little unsafe, as it shows the length of your password in the cmd! :P
11:38:25	dom96	lol
11:38:35	dom96	Time to submit webkit/gecko bugs
11:38:40	couven92	(and yes every other GUI totally ignores that by printing password dots!)
11:40:35	couven92	dom96, you can totally have line-feeds in your password! You can even have emojis in your password (you can just not login if you do afterwards)
11:42:46	dom96	Hrm, I wonder if most websites just reject emojis and other unicode characters
11:44:48	couven92	at NDC London there was an Emoji talk, at the end of the talk the presenter said that you UNIX-password can contain any Unicode, but that the password input field on Mac and most linux fields don't allow you to enter any Unicode chars, meaning that you might not be able to enter your super-duper cool poop-emoji password! :P
11:46:51	*	MJCaley quit (Quit: MJCaley)
11:59:42	GitDisc	<Yardanico> 1231
11:59:49	*	GitDisc quit (Remote host closed the connection)
12:00:45	*	GitDisc joined #nim
12:02:20	*	Snircle joined #nim
12:06:00	*	yglukhov joined #nim
12:08:43	*	Yardanico is now known as Yardanico[m]
12:08:46	*	Yardanico_ joined #nim
12:11:53	*	Yardanico[m] left #nim ("User left")
12:14:04	*	dddddd joined #nim
12:14:07	Araq	well?
12:14:38	*	Yardanico_ quit (Quit: Leaving)
12:14:49	*	Yardanico joined #nim
12:17:25	dom96	Araq: Let's flip this around. If you were starting Nim from scratch, how would you solve this problem?
12:17:36	Araq	before dreaming up rules with symlinks and re-inventing how to resolve versions, can we decide on a name?
12:18:23	Araq	dom96: I would require 'import jester / jester' or maybe 'import jester / _'
12:18:57	dom96	A name for what?
12:19:23	*	Yardanico quit (Client Quit)
12:19:50	Araq	for "import from $nimblepath"
12:20:01	Araq	or some other way to solve the problem
12:20:03	*	Yardanico_ joined #nim
12:21:18	dom96	'import pkg/'?
12:21:30	dom96	We already have 'std', right? so it fits
12:22:59	FromGitter	<alehander42> @Araq j("\\\"") works, but j"\\\"" fails with " error, is that a known limitation of this syntax
12:23:34	FromGitter	<alehander42> `j("\\\"") works, but j"\\\""` (fixing paste)
12:23:39	dom96	alehander42: You escape " with a double "" in blah""
12:23:43	dom96	(raw string literal syntax)
12:23:48	Araq	j"" is a raw string literal, j("") is not
12:24:20	Araq	we don't have suffix operators so 'import pkg/' doesn't parse
12:24:29	FromGitter	<alehander42> I see, ok
12:25:07	Araq	besides, that does not fix the problem since 'import random/random' does not work due to its structure
12:25:09	dom96	Araq: huh?
12:25:17	dom96	We have 'import std/'
12:25:23	Araq	we don't.
12:25:26	dom96	so let's have 'import pkg/random'
12:25:32	Araq	we have import std / modulename
12:25:45	dom96	that's what I mean...
12:25:55	Araq	oh
12:25:56	dom96	What would 'import std/' even mean?
12:26:01	FromGitter	<alehander42> do we have import std/x :O
12:26:23	Araq	dunno, I thought you were proposing to import every file in 'std' with this syntax
12:26:29	Araq	;-)
12:26:31	dom96	lol why
12:26:39	dom96	That would be 'import std/*' ;)
12:26:42	Araq	ok, so 'pkg' is a keyword
12:26:47	Araq	?
12:26:53	dom96	But anyway, I think my solution is good
12:27:03	dom96	depends what you mean by "keyword"?
12:27:07	*	Yardanico joined #nim
12:27:26	Araq	well it's obviously not a directory, is it?
12:27:41	dom96	it's special-cased
12:27:54	dom96	It's not a keyword in the sense that you can't define a 'pkg' variable
12:27:59	dom96	(Same for 'std' right now, right?)
12:28:01	Araq	ok, so special cased, yes
12:28:09	*	Yardanico__ joined #nim
12:28:12	Araq	(yes, same for 'std')
12:28:18	dom96	Yeah
12:28:21	dom96	Sounds good to me
12:28:49	dom96	How does 'std' work right now?
12:28:55	dom96	How is 'import std/sha1' handled?
12:28:57	Araq	compiler voodoo.
12:29:02	dom96	Can I do 'import std/std/sha1'? :P
12:29:15	Araq	no.
12:29:32	dom96	why not?
12:29:33	Araq	(maybe if there is a bug, but not in principle)
12:29:40	dom96	okay
12:30:07	dom96	So before I go into arguing about "Sha1 shouldn't be in std/"
12:30:21	dom96	Are we happy with `import pkg/random` vs. `import std/random`?
12:30:51	Araq	std is like "lookup, if fails, try lib/pure/ etc", but 'std' itself is NOT in the search path
12:31:11	Araq	so if some module is physically in std/ then the std/ directory is required
12:31:22	dom96	yes, makes sense
12:31:44	dom96	but what does "etc" mean here?
12:31:58	Araq	lib/impure, lib/wrappers
12:32:04	*	Ven`` joined #nim
12:32:11	Araq	and some others that the stdlib happens to use
12:32:16	dom96	I think you should introduce a '--pkgPath'
12:32:27	dom96	and --stdPath?
12:32:35	Araq	I don't want to, we have --libpath and --nimblePath
12:32:41	*	Yardanico__ left #nim (#nim)
12:32:51	Araq	or maybe it's just called --lib, don't remember, I never use it
12:33:01	dom96	The problem is that Nimble has to pass --path explicitly when building
12:33:03	Araq	but it exists.
12:33:15	dom96	so that will then fail if somebody uses `import pkg/random`
12:33:22	dom96	since 'random' will be in the "path"
12:33:41	Araq	hmm not sure I'm following
12:33:57	Araq	would std / random (enforced) help with that?
12:34:03	dom96	nimble build -> nim c --path:~/.nimble/pkgs/random-0.1.0/ mybin.nim
12:34:29	dom96	The compiler won't know whether that's part of the stdlib or Nimble
12:34:39	dom96	because AFAIK all stdlib modules are defined via --path
12:35:13	dom96	Not sure what --libpath does
12:35:24	dom96	But --nimblePath should then be built on top of this new --pkgPath
12:35:38	dom96	You'll need a separate list in the compiler anyway I think
12:35:47	dom96	so I don't see why creating this distinction will hurt
12:35:57	dom96	it follows from this 'std'/'pkg' feature
12:36:10	Araq	I wanted to introduce import $package / foo so that you can do: --find:package=over/here
12:36:29	Araq	but you talked me out of $dollar because "too hard to teach"
12:36:34	dom96	yes
12:36:36	dom96	I did
12:36:54	dom96	because it doesn't work well with what he have currently
12:36:55	Araq	then --path would not be a stupid list but specialized by package
12:37:02	dom96	*we
12:37:21	dom96	everybody would need to change their imports to use this $dollar thing
12:37:43	dom96	It's an overly-generalised feature, I don't see anyone else using this
12:37:48	Araq	true but it's a clear distinction between path imports and package imports
12:38:05	Araq	well I don't see why 'std' and 'pkg' should be special
12:38:19	dom96	Yes, but I want the distinction only to be necessary for resolving conflicts
12:38:37	dom96	The current module system is good
12:38:43	dom96	and it works for 99% of cases
12:38:59	dom96	The remaining 1% need this `pkg`/`std` feature to disambiguate
12:39:57	FromGitter	<alehander42> why can't you import random from pkg vs import random from std ⏎ I think the keyword illustrates better the difference and reusing the namespace a / b syntax can be confusing (as they're not really namespaces)
12:40:50	dom96	That implies that 'pkg'/'std' are modules.
12:40:55	Araq	from strutils import toLowerAscii from std ?
12:41:34	Araq	can't see that one working out well, regardless if 'std' is a real keyword or not
12:42:05	Araq	from std / strutils import toLowerAscii # makes more sense IMO
12:42:26	dom96	agreed
12:42:50	FromGitter	<alehander42> ok, what happens if you have a directory called 'std'
12:42:55	FromGitter	<alehander42> or called 'pkg'
12:43:07	Araq	you really shouldn't and that's the end of the story. :-)
12:43:49	Araq	unless we come back to my $std proposal :P where we get these lovely dollars for "namespaces"
12:45:15	dom96	actually, that's a valid concern
12:46:08	dom96	There will undoubtedly be users who create 'std' or 'pkg' directories in their project
12:46:17	Araq	we can name them 'stdlib' and 'stdpkg' and then people who name their directory 'stdlib' need to be punished
12:46:31	dom96	stdpkg? lol
12:46:46	dom96	or
12:47:05	dom96	we also have 'cwd'
12:47:06	dom96	or 'dir'
12:47:08	dom96	or 'local'
12:47:17	dom96	wait
12:47:17	FromGitter	<tim-st> Is there something like a macro or template that can add one `of` line to a given case and the of statement should be passed?
12:47:20	dom96	Don't we have '.'?
12:47:27	dom96	'import ./pkg/foo'
12:47:31	dom96	Problem solved?
12:47:39	Araq	tim-st: a template cannot do that, but a macro surely can
12:47:40	FromGitter	<andreaferretti> if I understand correctly, as things are now, one can explicitly choose to opt in the stdlib prefixing std to the import path, but not opt in nimble packages
12:47:52	FromGitter	<tim-st> @Araq thanks!
12:47:57	FromGitter	<andreaferretti> so it makes sense that nimble packages are searched first
12:48:05	dom96	andreaferretti: we are changing that
12:48:19	dom96	You will be able to explicitly ask for a Nimble package
12:48:35	FromGitter	<andreaferretti> I see, I was trying to figure the smallest change that would work
12:48:50	Araq	dom96: 'import "./pkg/foo" ' works but I really think people shouldn't use 'pkg' or whatever we end up using
12:49:04	FromGitter	<andreaferretti> if nimble packages have the priority, there should be no need to explicitly choose them
12:49:28	dom96	Araq: I agree. But if there is some unforseen reason that they must use this directory name, it's good that there is a workaround
12:49:42	Araq	ok.
12:50:14	*	athenot_ quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
12:50:27	Araq	andreaferretti: problem is that nimble packages creep into my nimbledir and will break things
12:51:05	Araq	say for illustration that I install a 'strutils' package which can be imported as 'import strutils' (just an example)
12:51:33	Araq	now every Nim program doesn't compile anymore if nimble packages get preferred over the other stuff in my search path
12:52:22	Araq	I can patch my Nim code to say 'import std / strutils' everywhere
12:52:48	Araq	but I don't want to. :-)
12:53:01	FromGitter	<narimiran> nimble package owners should make sure their libraries have unique name. end of story.
12:53:27	*	couven92 quit (Ping timeout: 240 seconds)
12:53:28	Araq	'random' used to be unique before we added it to the stdlib
12:53:35	Araq	same for sha1.
12:53:44	Araq	and it will happen again, why wouldn't it?
12:53:49	dom96	yep
12:53:51	dom96	precisely
12:53:54	FromGitter	<narimiran> araq: i'll copy my comment from before:
12:53:58	Araq	we cannot dream up of strange stdlib module names just to prevent clashes.
12:54:11	FromGitter	<narimiran> "and if some conflict arises currently because of additions to std lib - well, nim is still in pre-1.0 phase and this should be expected"
12:54:25	FromGitter	<narimiran> "and you and your package need to find another name"
12:54:36	Araq	we want to be able to grow Nim past 1.0 is released
12:54:36	dom96	Nim won't be in a pre-1.0 phase forever
12:54:40	*	radagast_04 quit (Ping timeout: 240 seconds)
12:54:47	FromGitter	<narimiran> dom96 - prove that
12:54:53	FromGitter	<narimiran> :D
12:55:30	Araq	well that's a good joke but nothing more.
12:55:46	Araq	even if we never release 1.0 a working solution would be desirable.
12:55:52	FromGitter	<narimiran> agreed
12:56:58	FromGitter	<narimiran> imho, imports from stdlib should remain as they are, and be as simple as possible
12:58:13	FromGitter	<narimiran> imports from nimble should also be as simple as possible when there are no conflicts, and have some addition (details not yet known) when there are conflicts either with stdlib or another nimble package
12:59:53	FromGitter	<narimiran> and there is a possibility that "my code had worked in nim 1.1, but doesn't work in 1.2" because of same-name lib added to stdlib in v1.2 - and this code should be now corrected with more explicit import
13:00:52	*	BitPuffin joined #nim
13:01:24	Araq	yes and for this to work we need a way to disambiguate
13:01:25	*	r3d9u11 joined #nim
13:01:42	Araq	which we currently lack for nimble packages that do not use a "namespace"
13:02:20	Araq	or not, import "../package-1.1/foo" works I guess
13:02:36	FromGitter	<narimiran> first - is there a possibility for two nimble packages to have the same import?
13:03:05	Araq	not sure, seems unlikely
13:03:23	Araq	but it can clash with anything in your --path
13:03:29	FromGitter	<narimiran> thought so. then this makes the problem easier
13:03:42	FromGitter	<narimiran> damn
13:04:09	FromGitter	<narimiran> `nimble import random`?
13:04:31	*	couven92 joined #nim
13:04:50	FromGitter	<Vindaar> you could even nimble import -> nimport (only half joking here...)
13:05:08	dom96	Each Nimble package name has to be unique
13:06:31	dom96	I just remembered something though and I wonder how well it will interact with this disambiguation
13:06:56	dom96	It would be nice to scope Nimble packages
13:07:25	FromGitter	<andreaferretti> require each pckage author to use a the reverse of a url they own :-P
13:07:51	FromGitter	<andreaferretti> import com.apache.common.random :-D
13:08:02	dom96	npm does it like this: require('@myorg/mypackage')
13:08:32	dom96	import dom96/jester ? lol
13:09:27	FromGitter	<narimiran> @andreaferretti don't give them the ideas :D :D
13:09:31	dom96	I think it might be too late to enforce this
13:09:57	*	SenasOzys quit (Ping timeout: 264 seconds)
13:10:00	Araq	dom96: yglukhov argued against it before
13:10:12	Araq	too often does the "vendor" change for this to be viable
13:10:14	dom96	oh? What was his argument?
13:10:51	GitDisc	<ethanpmorgan>
13:10:51	GitDisc	<ethanpmorgan> https://cdn.discordapp.com/attachments/371759389889003532/416582680982061056/20180223_131037.png
13:10:53	FromGitter	<narimiran> if you have `uniqueName` package, you can import with either `import uniqueName` or `nimble import uniqueName` for the same effect
13:11:02	GitDisc	<ethanpmorgan> The deer is done. My first programming book
13:11:02	dom96	So rename and scope and an alias for the old name
13:11:19	yglukhov	hey guys whats up
13:11:19	GitDisc	<ethanpmorgan> The deed is done. My first programming book
13:11:22	Araq	what does "scope" mean here?
13:11:28	dom96	ethanpmorgan: yay :D
13:11:36	FromGitter	<narimiran> if you have `usedName` package, if you use `import usedName` it will import stdlib version. you need to import it with `nimble import usedName`
13:11:50	dom96	Araq: The username in "import dom96/jester"
13:12:02	Araq	ok so what I called "vendor"
13:12:16	dom96	vendoring is a different package manager concept
13:12:17	GitDisc	<ethanpmorgan> College has us doing VB.net I'll let Nim help fix my insanity
13:12:23	dom96	so "scope" is a better name
13:12:37	Araq	it really isn't. but "username" is ;-)
13:12:55	Araq	"scope" already has a different meaning in a PL context.
13:12:56	dom96	narimiran: no clue what you mean
13:13:17	Araq	he is proposing a new syntax, a 'nimble import' statement
13:13:24	Araq	crazy talk. ;-)
13:13:32	FromGitter	<narimiran> @dom96 that's my proposal for this import confusion
13:13:59	dom96	righttttt
13:14:26	Araq	maybe we should allow URL imports and let people deal with the mess
13:14:46	FromGitter	<narimiran> what if there is a great package from user `fuckthisshit`? i wouldn't want to see that in my code
13:15:01	dom96	Araq: We're not savages
13:15:24	Araq	then we only need a way to do string interpolation like import "$foo" so that $foo can become https://etcetc
13:15:38	dom96	narimiran: why not? Brainfuck is a thing :P
13:15:50	dom96	Hrm... could I create a poop emoji Nimble package?
13:16:00	dom96	I bet I could
13:16:07	Araq	yeah, what if the module name is 'fuckthisshit'?
13:16:14	FromGitter	<narimiran> tru dat
13:16:17	Araq	you fork and rename? there you go.
13:16:33	Araq	but real packages have real names, in general.
13:16:41	livcd	what if the code contains master/slave and all the SJWs get mad. 1st world problems.
13:16:49	Araq	LOL
13:16:53	couven92	naming is hard!
13:17:34	couven92	unfortunenately still one of the great unsolved problems... :P
13:17:54	livcd	kill me but i like go get :PPP
13:18:03	FromGitter	<narimiran> ok, @Araq, so you think `nimble install` is not a good/simple way to deal with this?
13:18:08	dom96	livcd: :'(
13:18:29	Araq	narimiran: you mean 'nimble import'?
13:19:03	FromGitter	<narimiran> yeah, sorry
13:19:08	dom96	what would that even do?
13:19:32	*	SenasOzys joined #nim
13:19:52	Araq	we don't need more keywords and the split between stdlib and nimble packages is questionable design to begin with IMO.
13:20:06	FromGitter	<narimiran> @dom96 `nimble import pkgname` would make sure you're importing nimble package, and not the stdlib one
13:20:29	Araq	I would like to have a solution that does more than just stdlib vs nimble package
13:20:47	Araq	but oh well, probably too late for that
13:22:20	livcd	I would prefer simple instead of smart
13:22:51	FromGitter	<narimiran> ok, you two (araq and dom) have mentioned you are afraid of post-1.0 time when something would be a nice addition to stdlib, and there is a nimble package of the same name - how often do you think something like that would happen?
13:23:08	Araq	again, it already happened twice.
13:23:23	FromGitter	<narimiran> nim is not in post-1.0 phase yet ;)
13:23:25	Araq	and in general stdlib names want to be "standard"
13:23:37	dom96	what it should have been from the beginning is 'std/' prefix for all stdlib packages
13:23:40	dom96	but it's too late forthat
13:24:12	FromGitter	<narimiran> @dom96 would that mean `import std/strutils, std/sequtils, std/math`?
13:24:17	Araq	http, url, uri, cgi, every protocol name makes a good "stdlib" name
13:24:30	dom96	narimiran: yes
13:24:38	Araq	or
13:24:44	FromGitter	<narimiran> let me find some vomiting smiley....
13:24:50	Araq	import std / [strutils, sequtils, math]
13:24:51	dom96	although 'import std/[strutils, sequtils, math]` would be possible
13:25:09	dom96	optional 'std' is a good compromise IMO
13:25:26	FromGitter	<narimiran> `import std: strutils, sequtils, math` to look more nim-like?
13:25:35	dom96	So let's get it in there and move on
13:25:43	Araq	yup.
13:25:55	Araq	ok, so we add 'pkg' too and call it a day
13:25:57	Araq	?
13:26:15	FromGitter	<krux02> I just joined.
13:26:20	FromGitter	<krux02> you want to add pkg to what?
13:26:33	dom96	yes
13:26:43	Araq	as the second pseudo directory, krux02
13:27:07	FromGitter	<krux02> the second?
13:27:15	dom96	Araq: But also --pkgPath (and possibly --stdPath)
13:27:19	Araq	the first one is 'std'
13:27:23	Araq	argh
13:27:34	FromGitter	<krux02> I know in linux there are a lot of folders with the suffix `.d` for directory
13:27:42	Araq	--pgkPath and --stdPath, why?
13:27:51	dom96	I already explained why :(
13:28:01	Araq	yeah I know
13:28:03	Araq	but I mean
13:28:11	Araq	"can't we do without?"
13:28:13	FromGitter	<narimiran> @Araq can you show us the example of the proposed import of stdlib package and nimble package?
13:28:24	dom96	no, we can't. Nimble needs a way to tell the compiler which is which
13:28:32	FromGitter	<narimiran> or even better - multiple packages
13:28:38	dom96	unless you have a better idea
13:28:52	dom96	narimiran: import pkg/random vs. import std/random
13:29:34	Araq	dom96: how does it even fix the problem? say I have 'import random' in my program.
13:29:54	Araq	then nobody knows if it's std/random or pkg/random
13:30:13	dom96	'import random' will behave as it does now, search through the stdlib paths and then the nimble paths
13:30:23	dom96	This is about 'import pkg/random'
13:30:29	FromGitter	<krux02> whell the compiler could complain and say it's not unique
13:30:41	dom96	If Nimble specifies --path:~/.nimble/pkgs/random-0.1.0/ then you'll get an error
13:30:48	FromGitter	<narimiran> if there is 'pkg' prefix, then `import random` should be only the stdlib one
13:30:49	Araq	why?
13:31:01	FromGitter	<narimiran> so there is no confusio
13:31:20	dom96	Araq: Because `import pkg/random` will search through the nimble paths
13:31:26	dom96	And only those
13:31:38	FromGitter	<narimiran> imports without prefix = always stdlib package, nothing else
13:31:48	Araq	oh I see your point. hm.
13:31:58	livcd	I would just force people to pick a different name than awesomepkg: random
13:32:00	FromGitter	<andreaferretti> @narimiran that shi has sailed
13:32:07	FromGitter	<krux02> well I see this as, when you use `import strutils` without the pkg prefix then installing a package called strutils could break coda that already works.
13:32:17	FromGitter	<andreaferretti> there is already a lot of code using import without prefix for nimble stuff
13:32:18	FromGitter	<krux02> well, not if you build with nimble
13:32:22	FromGitter	<krux02> I see
13:32:45	FromGitter	<narimiran> once again - nim is pre-1.0, everybody should expect breaking changes!
13:33:07	FromGitter	<andreaferretti> not a breaking change that breaks half of the existing ecosystem without a second thought
13:33:07	FromGitter	<narimiran> we all knew what we "signed for"
13:33:13	Araq	we don't break every Nim program out there without a deprecation path, that's crazy talk.
13:33:20	FromGitter	<narimiran> ok ok :)
13:33:40	Araq	btw we also must be able to bootstrap ;-)
13:34:03	FromGitter	<krux02> well, but I think that at some point things that are deprecated should be removed when they are deprecated for a long time, just to clean up the code.
13:34:28	FromGitter	<krux02> make it ten years or something, but give it a date when something can be removed
13:34:42	dom96	Araq: Only problem with my idea to use --pkgPath is that Nimble will stop working with older Nim compilers
13:35:11	Araq	dom96: it can query the compiler's version if we really care that much about it
13:35:18	dom96	true
13:36:00	Araq	ok, so --pkgPath is actually what I thought of, it's not a list of options, but a concrete mapping from package name to its location?
13:36:50	FromGitter	<andreaferretti> does nimble currently pass the path of the stdlib using --path?
13:36:54	dom96	Araq: no, there is no mapping
13:37:03	Araq	but there should.
13:37:04	dom96	Araq: It's just a single path
13:37:07	dom96	Why?
13:37:20	dom96	andreaferretti: no
13:37:25	Araq	because it's cleaner.
13:37:38	Araq	nimble tells the compiler how to resolve pkg/foo
13:37:40	dom96	Araq: what use is the package name?
13:37:56	FromGitter	<andreaferretti> @dom96 then why cannot --path be used where you think to use --pkgPath?
13:38:10	dom96	andreaferretti: because that will become "--stdPath"
13:38:31	dom96	andreaferretti: Nimble needs a way to distinguish
13:38:35	FromGitter	<andreaferretti> The compiler already knows where the stdlib lives, right?
13:38:40	FromGitter	<andreaferretti> It's not nimble telling
13:38:41	FromGitter	<krux02> I just realized I failed at contributing to the conversation :P just ignore me I am procrastinating
13:38:45	*	Arrrr quit (Ping timeout: 265 seconds)
13:38:51	dom96	andreaferretti: yes, but that's defined in Nim's config file... using --path
13:39:01	FromGitter	<andreaferretti> uh I see
13:39:06	Araq	actually, it's --lib
13:39:28	FromGitter	<andreaferretti> great, then there is no need to introduce two new names
13:39:35	FromGitter	<andreaferretti> --lib -> --stdPath
13:39:41	FromGitter	<andreaferretti> --path -> --pgkPath
13:39:43	Araq	which gives the location of system.nim. probably a bad thing to have both
13:39:48	dom96	https://github.com/nim-lang/Nim/blob/devel/config/nim.cfg#L34 ?
13:39:52	dom96	looks like --path to me
13:39:52	FromGitter	<andreaferretti> --lib and --path are good enough
13:40:02	Araq	especially since I'm not sure what --lib does...
13:40:26	Araq	feature creep.
13:41:00	dom96	why do you want a mapping?
13:41:30	Araq	to not further hardcode this nimble vs stdlib distinction
13:41:45	Araq	the stdlib should be a package like others
13:41:57	Araq	unrealistic at this point, I know, but still.
13:42:15	dom96	pkg and std will be hard coded
13:42:18	dom96	there is no avoiding it
13:42:47	dom96	--pkgPath/--stdPath (or whatever you want to call these, this isn't all that important right now) is the natural way to implement this
13:43:34	dom96	Unless you're thinking --map:pkg:path1 --map:pkg:path2
13:43:52	dom96	to allow more than just 'pkg' and 'std'
13:43:53	Araq	that is what I have in mind.
13:44:11	dom96	that's far more disruptive
13:44:43	dom96	We can make `--path` mean `--pkgPath` (add an alias so that it can be renamed in time)
13:44:55	dom96	and change the Nim config to use --lib for --stdPath
13:45:12	dom96	(and also create an alias for --lib, or just rename it since no outside tools use it)
13:45:16	Araq	I don't fully understand the difference between --path and --pkgPath
13:45:32	dom96	--path:foo -> --map:std:foo
13:45:40	dom96	--pkgPath:foo -> --map:pkg:foo
13:45:56	FromGitter	<andreaferretti> this would break everything
13:45:56	dom96	(--path = --stdPath in this case)
13:46:20	FromGitter	<andreaferretti> there are already projects using path, generally for things not in the stdlib
13:46:40	dom96	But perhaps you're wondering what the difference between --path and --pkgPath is as --path is currently implemented
13:46:44	Araq	yeah, my projects use it too
13:46:55	Araq	dom96: yes.
13:46:58	Araq	that too.
13:47:21	dom96	--path is used by ``import random``
13:47:25	dom96	to find 'random'
13:47:41	Araq	yeah ok
13:47:50	Araq	the other thing is more of a rewrite rule
13:48:08	dom96	In our new implementation `import random` will use the list of --stdPath followed by --pkgPath
13:48:11	Araq	rewrite pkg/random to bla-1.0/random
13:48:17	dom96	`import std/random` will only use --stdPath
13:48:22	dom96	`import pkg/random` will only use --pkgPath
13:48:33	*	athenot joined #nim
13:48:41	dom96	no
13:48:49	dom96	You don't need to worry about that
13:49:00	dom96	That is taken care of by --nimblePath
13:50:23	Araq	unless overriden by --pkgPath, yes
13:50:44	dom96	--nimblePath:~/.nimble/pkgs iterates through each directory in ~/.nimble/pkgs, for each package name it finds the latest version and adds it using --path (right now, but in the future it will use --pkgPath)
13:51:05	dom96	--pkgPath doesn't override anything
13:51:12	dom96	it just adds a new path to the list of paths
13:51:18	dom96	the same thing that --nimblePath does
13:51:39	Araq	you lost me.
13:51:56	Araq	so you mean
13:52:07	Araq	--pkgPath does work with --noNimblePath?
13:52:17	Araq	to resolve the issue?
13:52:38	dom96	I think so. Not 100% sure what you mean.
13:52:47	Araq	ok
13:52:49	dom96	--noNimblePath instructs the compiler to ignore --nimblePath
13:53:14	dom96	(needed because the Nim config file contains these --nimblePath flags)
13:53:47	dom96	Is everything clear?
13:53:53	Araq	yes
13:54:02	dom96	great
13:54:06	Araq	I don't like it, but now it is clear
13:54:40	dom96	why not?
13:57:06	Araq	because it's too obviously a patch.
13:57:28	FromGitter	<andreaferretti> what about
13:57:44	FromGitter	<andreaferretti> instruct the compiler to look into --path for import pkg/foo
13:57:57	FromGitter	<andreaferretti> change the compiler to use --lib instead ofp--path
13:58:09	FromGitter	<andreaferretti> that's it
13:58:25	FromGitter	<andreaferretti> --path stays exactly as it is today
13:58:38	FromGitter	<andreaferretti> nimble stays exactly as it is today
13:58:57	dom96	Araq: Not if it's named --pkgPath and --stdPath :)
13:59:04	FromGitter	<andreaferretti> the only thing to change is the logic for the imports in the compiler
13:59:04	FromGitter	<andreaferretti> to prefer --lib over --path
13:59:20	dom96	andreaferretti: that's exactly what I suggested above...
13:59:48	FromGitter	<tim-st> If someone has edit rights to an example section of nim macros, please add an good and easy example for beginners like this one I made:
13:59:49	FromGitter	<tim-st> macro echoX*(): typed = result = "echo x".parseStmt() ⏎ var x = 1 ⏎ echoX()
14:02:19	Araq	tim-st: that's wrong.
14:02:31	Araq	it should use 'quote do' or template+getAst
14:02:36	Araq	or the API construction.
14:03:02	Araq	string based AST manipulations are a last resort when everything else fails
14:03:37	FromGitter	<tim-st> I assumed this, but it's very easy to use^^
14:04:19	Araq	it's only easy when you don't do anything whatsoever with the string
14:04:58	FromGitter	<tim-st> For my case thing that's working
14:05:11	Araq	dom96: I think andreaferretti has the right idea here
14:05:21	Araq	tim-st: for your case a 'template' would do too.
14:05:42	FromGitter	<tim-st> ok, will have a look, thanks
14:06:43	Araq	the point is "if 'pkg' is a special case, we can make --path work with it too"
14:07:13	Araq	so --path:foo would effect 'std/foo'
14:07:26	Araq	er no
14:07:31	Araq	I mean 'pkg/foo'
14:08:51	dom96	I'm pretty sure that's precisely what I suggested above.
14:09:00	dom96	--path means --pkgPath
14:09:07	dom96	--lib means --stdPath
14:09:23	Araq	well my proposal leaves out the --lib part :P
14:09:48	dom96	so how do you specify stdlib paths to Nim?
14:10:08	Araq	you don't. Nim knows where the stdlib is
14:10:12	Araq	well
14:10:25	Araq	depending on how you look at it, it stays the same. the config sets --lib
14:10:35	Araq	but Nimble doesn't.
14:10:49	Araq	but I guess this is what you meant
14:11:00	Araq	and you only requested the rename for consistency.
14:11:33	Araq	--path is ambiguous then, influencing where to find nimble packages as well as being the general search path
14:11:36	dom96	yep
14:12:14	dom96	I don't really see a reason to argue over this
14:12:21	dom96	This is just a couple of extra flags and sequences in the compiler
14:12:36	Araq	it's already complex
14:12:39	Araq	and had bugs.
14:13:00	Araq	and it's misdesigned, I don't like --path nor --nimblePath
14:13:16	dom96	I love it
14:13:31	Araq	so asking me to add --pkgPath or changing how --path works
14:14:01	Araq	does not make me happy.
14:15:12	Araq	I like my import $variable / [stuff] idea. you specify what $variable resolves to elsewhere. problem solved.
14:16:13	Araq	and both --nimblePath and --path become unnecessary.
14:17:06	dom96	Please let's just get this pkg/std distinction
14:19:06	Araq	that's more work to implement and less flexible :-(
14:19:46	Araq	it's also as hard to learn as $variable, IMO.
14:20:21	dom96	$variable is too flexible
14:20:33	dom96	there is no need for this to be in the user's face
14:20:52	dom96	I don't understand how that can possibly be less work
14:21:15	Araq	I wrote the code for it and disabled it recently.
14:21:39	Araq	so ok, maybe that's unfair.
14:21:51	Araq	anyway, I like it much more.
14:22:26	dom96	and how do you define what $variable means?
14:22:34	Araq	if Nimble supports / enforces user names one day, Nim doesn't have to change, it was $jester all along
14:23:25	dom96	so your solution is to break all existing package imports?
14:23:30	Araq	via the config or command line, --var:jester=$nimble/jester
14:23:56	Araq	no, I don't break anything, I only add a feature
14:23:57	dom96	so I install jester using Nimble
14:24:09	dom96	I can't use the global installation anymore unless I specify --var:jester=...
14:24:44	dom96	I also personally dislike this `$` thing
14:24:51	Araq	yes.
14:25:59	dom96	So you're breaking an established workflow
14:26:48	dom96	This is a much more disruptive change and I don't envision many use cases for its flexibility
14:27:23	Araq	this whole problem is a problem of "Don't repeat yourself", I don't want to write 'import ../nimble/pkgs/jester-1.0/jester' because it's tedious and the relative addressing super fragile
14:27:55	Araq	so instead I want $nimble/jester or even short just 'jester'
14:28:01	Araq	*shorter
14:28:20	Araq	but it's too ambiguous so I propose $jester
14:29:23	Araq	which cannot be confused with a real path/filename
14:29:46	dom96	if you want we can ask the community what they think
14:32:32	Araq	well we're having this discussion in #nim
14:32:44	Araq	we had an RFC already
14:32:50	Araq	with plenty of opinions
14:32:56	Araq	and no real solution afaict
14:33:29	Araq	I thought my recent 'std' addition would be good enough but it's not and now I don't want to add even more patches on top of it
14:34:51	FromGitter	<data-man> @Araq: Can I to use --var:std=mySuperCoolStd?
14:35:08	Araq	data-man: in my proposal, yes.
14:35:18	dom96	If you implement this then that will be the end of it
14:35:35	dom96	I will double check your commit to make sure you've done it right :P
14:35:45	Araq	that would be ok if I was happy with --path and --nimblePath, but I'm not.
14:36:06	Araq	and Nimble is getting "multiple packages per git repo"
14:36:17	Araq	and maybe later "username support" in some form
14:36:32	FromGitter	<alehander42> hey, I have a global level var and nim js compiles a for label, value in <var>: to `for(.. in var_variable)` which is a weirdly obvious bug
14:36:34	Araq	so I don't think you can promise me it won't grow again.
14:36:43	FromGitter	<alehander42> does anybody remember such a bug
14:37:05	Araq	alehander42: nope. report it please
14:37:24	dom96	Araq: multiple packages per git repo won't be a problem
14:37:53	dom96	they'll just be installed separately
14:38:04	dom96	usernames may never happen
14:38:14	dom96	by the time they do 1.0 will be out
14:39:49	FromGitter	<data-man> I like --var:
14:41:08	dom96	data-man: why?
14:41:42	FromGitter	<data-man> E.g. for NewStd with new runtime.
14:45:07	Araq	dom96: what if we use --var without the $ as a general mechanism to override --path?
14:45:40	Araq	then Nimble can use --var:pkg/foo=path/foo-1.0/foo
14:46:17	FromGitter	<andreaferretti> wouldn't this break existing imports anyway?
14:46:40	FromGitter	<andreaferretti> or do you plan for this to also support just import foo when it is not ambiguous
14:47:28	dom96	Araq: That's better, but --nimblePath will also need to introduce the same mappings
14:48:17	dom96	I fear it's too much power though
14:48:45	Araq	username support is really important though and it's covered by it
14:49:07	Araq	just saying "nimble will never support usernames" doesn't seem wise.
14:49:25	dom96	I can achieve the same thing by getting Nimble to place the packages inside a username/ directory...
14:50:10	FromGitter	<data-man> And this will simplify the development of new functions in the stdlib. Or comparing performance with existing functions. Without using the git.
14:51:49	FromGitter	<krux02> Araq: I give up on the modulo discussion. `div` on integers should round towards 0 because that is the behaviour in assembler and in C and many other languages. I don't like it, but I can accept it. Same way should `mod` be the pair of `div` meaning (x div y) * y + (x mod y) should be zero
14:52:14	FromGitter	<krux02> and mod on floats should change it's behaviour to be the same as fmod in c++
14:52:34	*	sz0 joined #nim
14:52:42	Araq	data-man: that's already possible via a Nimscript config
14:53:00	Araq	krux02: what caused this change of mind?
14:54:22	FromGitter	<krux02> But I think a function `proc divmod(a,b: int): [a,b: int]` should exists, because both division and modulo division is calculated with the same asm instruction just with the result in a different register and integer division is the slowest math instruction even slower than floating point division.
14:55:07	Araq	dom96: what do you mean? "--nimblePath will also need to introduce the same mappings"?
14:55:25	Araq	krux02: agreed.
14:55:38	FromGitter	<krux02> Well I think it's important to be consistent and stick to standards. Certain things are just fighting agains windmills.
14:55:43	dom96	Araq: I mean it will also need to implicitly specify --var:pkg/foo=... for each package
14:56:01	dom96	otherwise I won't be able to write 'import pkg/random' and compile via Nim
14:56:36	Araq	that's conceptually what --nimblePath does, it's just that the implementation is different?
14:56:38	FromGitter	<krux02> It's like the guys who wants decimal floating point numbers because he things the float type is bad.
14:56:55	Araq	I prefer fractions fwiw
14:56:57	FromGitter	<krux02> Yes float is not perfect, but everything is float. Arguing against it is pointless.
14:57:15	Araq	and saturated arithmetic :-)
14:57:47	FromGitter	<krux02> And now I called `mod` from c++ glm `modulo` in nim-glm.
14:57:54	dom96	Araq: I'm just mentioning that you will still need --nimblePath
14:58:04	dom96	and that it will need to work with this new --var thing
14:58:14	Araq	yes, and add 'pkg' as a special name
14:58:14	dom96	I still prefer --stdPath/--pkgPath
14:58:15	FromGitter	<krux02> that modulo function has the behavior I want,
15:01:25	Araq	dom96: ok, but I am happy with the compromise.
15:01:41	FromGitter	<krux02> I solved the conflict by just choosing a different name.
15:02:25	dom96	Araq: I'm not :/
15:02:25	FromGitter	<krux02> Araq: should I make a pull request with the change to `mod` on float behavior?
15:02:56	Araq	krux02 was it wrong?
15:03:05	Araq	iirc is uses 'fmod' already
15:03:10	Araq	*it
15:03:36	FromGitter	<krux02> when it does use fmod, then it is good.
15:03:51	dom96	I feel like we already agreed on --stdPath/--pkgPath
15:04:01	dom96	and you keep pushing this --var stuff
15:04:03	FromGitter	<alehander42> @Araq https://github.com/nim-lang/Nim/issues/7249 that's the js problem, I can try to fix it quickly but I am not sure if the problem is in for generation or in what's assigned to a
15:04:35	FromGitter	<data-man> @krux02: Maybe this library will be useful for you http://www.jhauser.us/arithmetic/SoftFloat.html ⏎ zig language used it.
15:05:26	Araq	dom96: "I fear it's too much power though" it not a convincing argument for Nim
15:06:17	Araq	and all that --pkgPath does is to hardcode this pkg/ prefix
15:06:21	dom96	the fact is that it won't be obvious where this strange random special cased paths are coming from
15:06:38	dom96	'pkg' and 'std' is already bad enough
15:06:49	dom96	and you want to encourage people to add their own weird names
15:07:00	dom96	with nothing to distinguish them from normal paths
15:07:02	dom96	it's a bad idea
15:07:30	Araq	but $foo is ugly, you can't have it both ways
15:08:08	dom96	I don't want it at all
15:08:49	Araq	you are also overly concerned with a niche case, namely 'nimble install jester' and then writing in some nim file 'import jester' (not even 'import jester / _')
15:09:05	Araq	without any configuration change
15:09:05	*	Ven`` quit (Ping timeout: 240 seconds)
15:09:19	FromGitter	<andreaferretti> is it a niche case?
15:09:28	Araq	I always have a configuration file anyway.
15:09:30	FromGitter	<andreaferretti> I think most nim projects are written this way
15:09:33	dom96	that's literally what everyone does
15:10:05	Araq	well I don't.
15:10:29	dom96	You mostly work on a project that doesn't use Nimble
15:10:34	FromGitter	<krux02> Araq: nope mod does not use fmod: http://ix.io/Oft/nim
15:10:54	*	Jesin joined #nim
15:11:21	Araq	I work on quite some projects.
15:11:31	Araq	none that lacks a configuration file.
15:11:55	*	Ven`` joined #nim
15:12:05	*	pwntus_ is now known as pwntus
15:12:06	FromGitter	<andreaferretti> by a configuration file do you also mean a nimble file
15:12:07	FromGitter	<andreaferretti> ?
15:12:18	*	pwntus quit (Changing host)
15:12:18	*	pwntus joined #nim
15:12:21	FromGitter	<andreaferretti> because all my projects have nimble files, none of them has a .cfg file
15:12:27	Araq	a .nims or a .nim.cfg file
15:12:41	FromGitter	<andreaferretti> i don't have any of those
15:12:50	FromGitter	<andreaferretti> i have a nimscript file called foo.nimble
15:12:52	Araq	ok but you do have a .nimble file, so there :P
15:12:58	FromGitter	<andreaferretti> ah ok
15:13:00	FromGitter	<andreaferretti> that's good
15:15:14	dom96	Araq: ugh, why are you now arguing for the removal of --nimblePath?
15:15:28	Araq	I am not
15:15:32	dom96	You are
15:15:56	Araq	no, I want something like --var instead of --pkgPath because it's more flexible
15:16:15	dom96	I don't understand your latest argument then
15:16:27	dom96	But I'm still unconvinced
15:16:49	dom96	--pkgPath/--stdPath is the more strict implementation
15:16:56	dom96	if we want we can always make it more flexible in the future
15:17:02	dom96	we can't go the other way around
15:17:42	Araq	well I disagree. all we can do with the inflexible solution is to add more features
15:18:00	Araq	I want one "too powerful" feature and be done with it
15:18:30	Araq	and that is exactly what did happen in the past.
15:18:35	Araq	we had --path
15:18:35	dom96	That's my point, you want to create this flexible feature because of some imaginary problem in the future
15:18:40	Araq	then we needed --nimblePath
15:18:45	Araq	then --noNimblePath
15:18:49	Araq	and now --pkgPath
15:19:09	Araq	come on, at least my solution makes the old shit redundant in theory
15:19:20	dom96	it doesn't
15:20:08	dom96	by your logic it all is redundant, I mean how hard is it to open the configuration file and add --path:~/.nimble/pkgs/random-0.1.0 in there anyway?
15:20:34	dom96	the idea is to give tools the necesssary hooks to make it easy on the user
15:20:51	dom96	not to give the user more tools to make things less obvious
15:21:21	Araq	the original problem was that I am unable to say import pkg / random
15:21:32	Araq	I fail to see how a config file change deals with that
15:22:14	Araq	but ok, if I use the fact that --path is an ordered search
15:22:22	Araq	then it might work out indeed.
15:22:28	Araq	with 'import random'
15:22:32	dom96	Well, I was referring to your "old shit" comment
15:22:40	dom96	which I interpreted as referring to --nimblePath
15:22:56	Araq	no, it also refers to --path
15:24:29	Araq	maybe even moreso to --path. it's already subtle to see where things are coming from since the path is ordered and this order is used to prioritize the search
15:25:44	Araq	I fail to see how an explicit mapping makes things worse
15:27:07	dom96	because --path is already established
15:27:43	Araq	established but requires --pkgPath. --var is a single feature that covers both.
15:32:21	Araq	but how about this: we don't add 'pkg', import random means "import from stdlib", import jester means "import from nimble package" and the random nimble package is fixed
15:32:56	dom96	no, we will need a way to disambiguate between packages
15:33:04	dom96	I have always wanted this
15:33:10	dom96	I thought you were going to implement it this way
15:33:11	Araq	and in the future "no namespace nimble packages" are inspected to use sane names
15:33:18	dom96	but you implemented this weird $thing, then rolled it back
15:33:31	dom96	added 'std', called it a day and now you're complaining that you have to implement another thing
15:33:49	Araq	well the person who does the work is allowed to complain about the work
15:33:57	dom96	true
15:34:08	dom96	And fair enough
15:35:02	FromGitter	<GULPF> about the mod/div discussion - I already have a PR for changing `mod` for floats to use fmod
15:35:08	Araq	let's face it. 'import random' for a nimble package never was a good idea
15:35:10	FromGitter	<GULPF> https://github.com/nim-lang/Nim/pull/7118
15:35:38	dom96	So does that mean you'll remove 'std'?
15:35:48	Araq	when the established meaning of 'import foo' was local import or stdlib import.
15:36:00	Araq	(already ambiguous btw)
15:36:05	FromGitter	<krux02> Araq: how do I set a fucntion to just emit some code for the target language?
15:36:30	dom96	import pkg/random, std/random, ./random # is the perfect tri-fecta IMO
15:36:32	FromGitter	<krux02> @gulpf: ah yes sorry for that
15:37:53	dom96	We already have 'std'
15:38:02	dom96	and we should have had 'pkg'
15:38:35	dom96	And there should have been a way to tell the compiler "This is a 'std' path" and "this is a non-std path"
15:38:44	Araq	import dir/random # kaboom
15:38:56	dom96	?
15:38:58	Araq	'std' and 'pkg' are introduced as pseudo directories
15:39:11	Araq	it's not a perfect tri-fecta
15:39:21	Araq	it leaves out an important case.
15:39:29	dom96	which is?
15:39:38	Araq	import someOtherSubDir / random
15:39:55	dom96	wut
15:40:00	dom96	why wouldn't that work?
15:40:19	Araq	it would but it proves it's not all that natural
15:40:46	dom96	well, yeah... 'std' and 'pkg' are special-cases
15:40:48	dom96	'.' isn't
15:40:52	Araq	I like the "stlib is just another package" much more conceptually.
15:41:19	Araq	and then 'std' and 'pkg' is a weird beast.
15:41:24	dom96	okay, I was happy with that
15:41:28	dom96	and then you said "no" in that issue
15:41:57	dom96	But that would still leave us hanging on the 'pkg' disambiguation front
15:42:39	dom96	Hrm
15:42:41	Araq	the pkg disambiguation is solved by --var
15:42:55	Araq	--var:jester=dom96/jester
15:42:56	Araq	vs
15:43:00	Araq	--var:jester=araq/jester
15:44:09	dom96	that's a very obtuse way to solve it
15:44:15	Araq	or --var:jester=$nimble/araq/jester
15:44:35	Araq	depending on how Nimble's solution looks like.
15:45:34	Araq	krux02: proc foo() = {.emit: """code here""".}
15:45:47	Araq	but I think I misunderstand your question, krux02
15:47:58	dom96	I'd prefer to disambiguate in the code rather than via some flag
15:48:15	Araq	but then you're back at the original problem
15:48:20	*	athenot quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
15:48:20	Araq	violating DRY.
15:48:42	Araq	(which might be acceptable)
15:48:58	Araq	there is not a single point where you import dom/jester
15:49:02	dom96	also, what if I want to import std/random in one module, but pkg/random in another?
15:49:41	*	athenot joined #nim
15:50:00	Araq	you want to be able to say "my project imports jester. today it's dom's jester. tomorrow it's krux02's fork of it. then it's dom's again"
15:51:03	dom96	In fact
15:51:05	dom96	this cannot work
15:51:11	dom96	I'll tell you why
15:51:32	dom96	What if Jester (or some other package) wants 'pkg/random' but my program wants std/random?
15:51:38	dom96	That is sure to happen
15:51:49	dom96	library packages cannot have configuration files
15:52:00	Araq	then it's not pkg/random, but random/random
15:52:11	Araq	that Jester imports
15:52:20	dom96	?
15:52:34	dom96	What does 'random/random' mean?
15:52:45	Araq	well as I said
15:52:57	Araq	import random to mean "import nimble's random"
15:53:02	Araq	was a bad idea to begin with.
15:53:33	Araq	and no patch really works for it.
15:54:08	dom96	I feel like you're missing the point
15:54:14	dom96	This disambiguation needs to happen in the code
15:54:21	dom96	Otherwise libraries cannot disambiguate
15:54:52	dom96	You can't have `import random` mean two different things in two different modules
15:55:05	Araq	never argued otherwise
15:55:12	Araq	I am aware of this point.
15:55:54	Araq	but the Nimble package lists its dependencies
15:55:57	*	floppydh quit (Quit: WeeChat 2.0.1)
15:55:58	FromGitter	<krux02> Araq: thanks, it is solved. I think I was looking for importcpp "# % #". Gulpf already did that for `mod`.
15:56:26	Araq	and so 'import pkg / random' would usually work since the Nimble file contains the dependency to the random package
15:56:50	Araq	and the random package is unique, nimble enforces it.
15:57:00	FromGitter	<krux02> Gulpf: can you add an entry to the changelog.md in your pull request?
15:57:18	dom96	You keep forgetting that some users compile via nim
15:57:32	dom96	and that Nimble doesn't get rid of the stdlib 'random' from the path
15:57:37	Araq	no, I am not.
15:57:55	Araq	import pkg / random fails if there is no random nimble package.
15:58:17	Araq	well
15:58:22	*	solitudesf joined #nim
15:58:25	Araq	what are you arguing about?
15:58:28	Araq	import random
15:58:29	Araq	or
15:58:36	Araq	import pkg / random ?
15:59:08	dom96	You seem to be proposing that users will disambiguate by writing `import random` and --var:random=~/.nimble/...
15:59:30	dom96	right?
15:59:48	dom96	or are we back to the 'pkg'/'std' implemented via --var idea?
15:59:53	Araq	kind of, but no, I mean
16:00:11	Araq	yes, that would work but only as a workaround and not in a nimble package
16:00:39	Araq	usually you would resolve a package name, not a module name
16:00:47	Araq	via --var
16:01:09	dom96	You'd need this 'pkg'/'std' to be convention
16:01:19	dom96	so flexibility offered by --var is pointless
16:01:30	Araq	you can use it to make 'import random' mean 'import .nimble/.../random.nim'
16:01:42	dom96	but why would I?
16:01:48	Araq	but it's a bad idea because it should mean "import from stdlib"
16:01:54	dom96	yes
16:01:56	dom96	exactly
16:02:01	Araq	why would you? to make code compile quickly.
16:02:13	Araq	when you can't be bothered to fix it properly :P
16:02:27	Araq	(maybe random was recently added to the stdlib)
16:02:37	dom96	so you change it to 'pkg/random'
16:02:42	dom96	it's not hard to search and replace
16:02:46	Araq	the flexibility offered
16:02:49	dom96	and it's the correct fix
16:02:56	Araq	is mostly for username support.
16:03:13	Araq	IMO.
16:03:29	dom96	Like I said, it's not necessary
16:03:38	dom96	username support can be handled by Nimble
16:04:11	dom96	and we don't know if we truly want it yet
16:04:29	PMunch	Username support?
16:04:50	dom96	scoped packages
16:05:15	Araq	ok, how about this. we keep 'std' because it allows to say
16:05:30	Araq	"import from stdlib, not my local random.nim file"
16:05:41	Araq	and leave the rest as it is now?
16:05:46	*	Ven`` quit (Read error: No route to host)
16:05:50	Araq	and fix the random nimble package
16:05:55	dom96	No
16:05:58	*	Ven`` joined #nim
16:06:02	Araq	what problems would it cause?
16:06:04	dom96	You've already suggested this and I said 'no'
16:06:16	dom96	You yourself said that we will run into this problem in the future
16:06:35	Araq	hmm
16:06:45	Araq	well
16:06:48	PMunch	Mabye put up an RFC on GitHub? You've been discussing this all day :P
16:06:49	Araq	no.
16:07:07	dom96	lol
16:07:15	dom96	and RFC will take months
16:07:15	Araq	we won't run into this problem if Nimble gets stricter about "packages without namespace"
16:07:29	Araq	to which you never really replied.
16:07:39	dom96	But I don't want to write `import jester/jester` :P
16:07:39	FromGitter	<krux02> the easier it is to just use someone's code. The better it is.
16:08:29	Araq	yeah, ok. I'm not gonna add more features just because 'import jester / jester' looks a bit ugly
16:08:48	Araq	especially since the features are not wanted anyway.
16:09:14	Araq	IMO import jester / _ could be added to deal with that
16:09:25	Araq	but you don't like that either anyway, so why bother.
16:09:29	dom96	lol
16:09:38	dom96	another feature to make it slightly better but still ugly
16:09:54	dom96	when a package vs stdlib path distinction makes sense anyway
16:10:09	dom96	I am still wondering how 'std' works currently
16:10:10	FromGitter	<krux02> dom96 whas do you really want or need.
16:10:24	FromGitter	<krux02> you two are arguing for quite a while now and it doesn't seem to settle.
16:10:27	dom96	How does it know what is the stdlib and what isn't?
16:10:32	FromGitter	<krux02> sometimes it better to just sleep about it.
16:10:49	FromGitter	<krux02> Like I did and I changed my mind about the `mod` operator
16:11:08	PMunch	How about if nothing matches "jester" then "jester/jester" is attempted?
16:11:12	FromGitter	<krux02> I don't like the behavior but I agree that it should behave like it does.
16:11:33	*	Trustable joined #nim
16:11:41	*	Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
16:11:44	PMunch	Or to put it another way, if it's not found in the stdlib then look for a namespaced version in nimble
16:12:19	Araq	then jester still needs to be patched to allow import jester/jester
16:12:21	*	couven92 quit (Ping timeout: 264 seconds)
16:12:23	Araq	and so does random.
16:12:31	PMunch	Why?
16:12:39	PMunch	Both jester/jester and jester would work
16:12:39	dom96	Araq: How does 'std' distinguish stdlib paths and other paths?
16:12:56	Araq	because 'import jester' is a shortcut for 'import jester / jester' the long winded version needs to work too
16:13:11	Araq	and the short version of 'import random' is what got us into this trouble in the first place.
16:13:19	PMunch	Araq, not really
16:13:23	Araq	I can say 'import random/random'
16:13:33	Araq	yes, really.
16:13:53	PMunch	Oh shit, I've literally got to run
16:13:56	*	PMunch quit (Quit: Leaving)
16:14:00	dom96	PMunch: I would need to modify the jester repo and move ---
16:14:01	dom96	okay then
16:14:12	FromGitter	<krux02> I don't really understand the problem
16:14:34	dom96	krux02: there is no way for me to import the 'random' nimble package
16:14:42	Araq	dom96, https://github.com/nim-lang/Nim/blob/devel/compiler/modulepaths.nim#L150
16:15:24	dom96	Araq: hm
16:15:32	FromGitter	<krux02> how it looks like when I import things is really not the most important feature of the programming language. Because it is not what I spent most of my time with. And if I spend too much time to get my imports right and cross platform. Then I should start thinking if there is something wrong. But that is not the case.
16:15:50	Araq	nimble enforces a unique package name but then might not use it as the namespace
16:16:15	Araq	it's like nimble is fine with the 80% solution
16:16:29	dom96	that's not really fair, 'import random' is enforcing the namespace
16:17:00	dom96	the problem is that the stdlib is in the "global" namespace
16:17:41	Araq	yes but the stdlib came first and so took the most convenient syntax possible
16:17:56	Araq	and actually it wasn't a namespace either
16:17:59	dom96	yes, and I can't fault you for that
16:18:06	Araq	it's what --path allows
16:18:37	*	yglukhov quit (Remote host closed the connection)
16:19:01	dom96	This is such a simple change
16:19:09	dom96	from what I can tell
16:19:15	dom96	based on the code you've shown me
16:19:32	dom96	we can reuse --path for packages
16:19:54	dom96	and add --stdPath, change the config file, change the modulepaths code slightly, add another if for 'pkg', done.
16:20:25	FromGitter	<krux02> I really think this nimble thing needs very good Documentation. Not just about how to use it, but also on how it looks for something and the reasoning behind it. How it chooses a package on the path, how the path is structured, etc. And what the differce is when you build a file with "nim c" vs "nimple c".
16:20:53	dom96	krux02: Have you read the Nimble readme? Most of that is documented
16:21:11	FromGitter	<krux02> well I did, but it was quite a while ago.
16:23:51	*	fvs joined #nim
16:26:54	Araq	what if the Nim compiler read the .nimble file properly. would it make things simpler?
16:27:16	Araq	then nimble doesn't have to teach it how to resolve pkg / foo
16:27:35	FromGitter	<krux02> dom96: I just looked at the readme and it is missing the most important part, the introduction. It doesn't explain what problem it solves and how it tries to solve it. It just assumes you know what problem it solves and jumps right to requirements and installation.
16:28:01	Araq	krux02: well it's a package manager.
16:28:06	Araq	what else is there to say about it?
16:28:12	dom96	Araq: No
16:28:19	Araq	it's like a wife. can't live with one, can't live without one.
16:28:19	FromGitter	<krux02> It says it's a package manager. But I am on linux, I already have a package manager, why does Nim needs it.s own package manager?
16:28:23	*	Ven`` joined #nim
16:28:50	dom96	krux02: sure, that's a good question.
16:29:22	FromGitter	<krux02> And if you know what a package manager is, it doesn't explain what kind of package manager it is. Does it install handle binaries, does it handle src.
16:30:22	FromGitter	<krux02> Simply talk a little bit about the problem of managing dependencies in a software project and what main inspiration nimble has for it's structure.
16:31:22	*	couven92 joined #nim
16:32:20	FromGitter	<krux02> Simply answer the most important question: "Why do I need to care about nimble"
16:34:50	*	MJCaley joined #nim
16:35:32	Araq	IMO a real documentation page would be more important, I'm fed up with gitub pages
16:35:49	Araq	but whatever ... now what do we do with 'pkg'? nothing?
16:36:18	FromGitter	<krux02> I am OK with github pages. The content is what I care about.
16:36:21	*	Jesin quit (Quit: Leaving)
16:36:39	dom96	I guess we should sleep on it
16:36:49	dom96	I'm tired of discussing this
16:37:25	FromGitter	<krux02> yes, thumbs up for that idea
16:41:27	*	xkapastel joined #nim
16:47:55	*	miran joined #nim
16:56:07	*	solitudesf quit (Quit: solitudesf)
17:02:52	*	Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
17:05:48	FromGitter	<Vindaar> So I got the getPassword proc working. Implemented an unsetControlCHook analogously to setControlCHook. Just reverts c_signal back to the signalHandler defined in excpt.nim. Seems to work just fine!
17:07:18	FromGitter	<krux02> @Vindaar I have no idea what you are talking about.
17:09:05	FromGitter	<Vindaar> Haha, sorry. Mentioned it yesterday after Dom's stream. I tried to implement a proc to read a password from stdin, without showing characters using readLine instead of getch(). So that one can use Ctrl-C and have it quit, and also use backspace in the input
17:12:19	*	solitudesf joined #nim
17:13:15	FromGitter	<Vindaar> Here: https://gist.github.com/Vindaar/f5cb17a6c47f6d4cdf3a4995ce2bf3ae
17:25:18	shodan45	it's the anti-Nim! https://hackaday.com/2018/02/22/quickbasic-lives-on-with-qb64/
17:28:33	Araq	Vindaar: candidate for terminal.nim?
17:28:41	Araq	seems generally useful for many
17:30:14	*	yglukhov joined #nim
17:30:14	*	yglukhov quit (Read error: Connection reset by peer)
17:30:51	*	MJCaley quit (Quit: MJCaley)
17:37:56	*	miran quit (Quit: Konversation terminated!)
17:45:28	FromGitter	<Vindaar> Yeah, that was the intention. Should I make a PR?
18:02:36	*	Zevv joined #nim
18:03:19	Zevv	Hi, is there a reason that the nim source files do not have a newline at the end?
18:03:42	Zevv	This seems to confuse Git, it complains in 'git diff' telling me 'no newline at end of file'
18:03:57	*	SenasOzys quit (Ping timeout: 264 seconds)
18:05:06	*	sz0 quit (Quit: Connection closed for inactivity)
18:11:43	Araq	Zevv, usually they do
18:11:57	Zevv	yeah, once more I'm complaining too early without investigating properly
18:12:04	Zevv	that's because you guys are too helpful
18:12:11	Zevv	it's easier to ask then to figure it out myself :)
18:14:00	*	SenasOzys joined #nim
18:14:53	Zevv	my vim messes things up
18:14:54	Zevv	:set noendofline
18:14:57	Zevv	that's new to me
18:16:43	Zevv	sorry for the noise, once again
18:18:21	*	francisl joined #nim
18:23:34	*	natrys joined #nim
18:36:43	*	Syneh___ joined #nim
18:37:38	*	dashed_ joined #nim
18:38:45	*	francisl_ joined #nim
18:41:17	*	francisl_ quit (Client Quit)
18:42:40	*	francisl quit (Quit: francisl)
18:43:05	*	avsej_ joined #nim
18:43:06	*	avsej_ quit (Changing host)
18:43:06	*	avsej_ joined #nim
18:43:09	shashlick	what's the best way to get/send binaries with other devs? something that doesn't need an account ideally
18:43:36	*	huonw_ joined #nim
18:43:41	*	Cthalupa- joined #nim
18:43:44	*	francisl joined #nim
18:44:45	*	avsej quit (*.net *.split)
18:44:45	*	Cthalupa quit (*.net *.split)
18:44:46	*	huonw quit (*.net *.split)
18:44:46	*	Syneh__ quit (*.net *.split)
18:44:46	*	dashed quit (*.net *.split)
18:44:48	*	Syneh___ is now known as Syneh__
18:44:50	*	avsej_ is now known as avsej
18:44:51	dom96	shashlick: you mean share files?
18:45:02	*	dashed_ is now known as dashed
18:48:10	shashlick	yep, share a test build of an app
18:53:05	dom96	Maybe this? https://send.firefox.com/
19:06:54	shashlick	nice
19:17:49	*	francisl quit ()
19:21:34	*	francisl joined #nim
19:24:08	*	r3d9u11 quit (Remote host closed the connection)
19:24:38	*	r3d9u11 joined #nim
19:34:04	*	fvs left #nim ("ERC (IRC client for Emacs 25.3.1)")
19:41:53	*	r3d9u11 quit (Remote host closed the connection)
19:53:34	*	BitPuffin quit (Remote host closed the connection)
19:58:43	*	tongir joined #nim
20:05:53	*	tongir quit (Ping timeout: 248 seconds)
20:08:02	*	tongir joined #nim
20:19:34	Calinou	that one's pretty nice, but files don't last long
20:19:39	Calinou	they last 14 days on https://transfer.sh/
20:19:45	Calinou	(plus, it has a CLI available)
20:20:07	Calinou	shashlick: ^
20:23:51	*	yglukhov joined #nim
20:30:11	*	dddddd quit (Ping timeout: 276 seconds)
20:34:19	dom96	so I have a git repo URL such as https://github.com/nimble-test/multi.git
20:34:29	dom96	Is there a natural way to specify a subdirectory within that git repo inside this URL?
20:36:12	shashlick	Calinou: that's nice
20:37:06	*	Vladar quit (Quit: Leaving)
20:38:02	Calinou	dom96: you can't clone a subdirectory directly
20:38:16	dom96	yes, I know
20:38:19	Calinou	Git submodules allow referencing a Git repository commit inside another repository, and cloning it when passing --recursive
20:38:26	Calinou	(or `git submodule --update init`)
20:38:36	dom96	but is there a URL convention that would allow me to specify the subdirectory naturally?
20:38:45	dom96	GitHub kind of screws this up for HTTPS urls
20:38:46	dom96	hrm
20:38:52	Calinou	not that I know of
20:40:44	*	dddddd joined #nim
20:44:30	dom96	https://github.com/nimble-test/multi.git/alpha
20:44:48	dom96	Rule: look for *.git and use the rest as the subdirectory, too hackish?
20:45:35	Araq	Nimble already has two URLs in its JSON description
20:45:43	Araq	"url" and "web"
20:45:51	dom96	yeah, so I could add another field called "subdir" as well
20:45:55	Araq	make "url" the URL and "web" the starting point?
20:46:12	Araq	but then why does the "root" matter?
20:46:26	dom96	what?
20:46:33	Araq	can't I have
20:46:39	Araq	https://github.com/nimble-test/multi.git/alpha
20:46:40	Araq	and
20:46:43	Araq	https://github.com/nimble-test/multi.git/beta
20:46:47	Araq	as 2 Nimble packages?
20:46:54	dom96	yes
20:46:56	dom96	that's the idea
20:47:11	Araq	so ... you don't have to implement anything?
20:47:20	dom96	yes I do
20:47:25	Araq	and what?
20:47:26	dom96	You can't just ask git to clone that
20:47:57	Araq	ah ok
20:48:05	Araq	so yeah, do it as you described
20:48:24	dom96	it's hacky though
20:48:31	dom96	https://github.com/nimble-test/multi/alpha # won't work
20:48:31	Araq	why?
20:48:44	Araq	that's ok
20:49:10	Araq	I think it's more hacky if I have to look up what the root is just to be sure
20:49:35	Araq	and don't make Nimble ask about the root, it's already chatty and wants to know too much things
20:49:43	dom96	I think a lot of people would argue for adding a separate field to the JSON heh
20:49:47	Araq	*too many things
20:50:00	Araq	the JSON is not the problem
20:50:04	dom96	But I prefer this solution too
20:50:12	dom96	Because I can easily install on the command line
20:50:16	Araq	the problem is that you then need to ask the programmer about it eventually.
20:50:34	dom96	I don't think that's a problem
20:50:40	dom96	You won't ever need to ask the programmer that
20:50:54	Araq	how could you avoid it?
20:51:07	Araq	if the URL is allowed to be ambiguous?
20:51:07	dom96	I know that yglukhov would prefer requires("https://github.com/nimble-test/multi", subdir="alpha")
20:51:25	Araq	that is not a solution.
20:51:32	yglukhov	?
20:51:33	Araq	that is the REAL hack.
20:52:04	dom96	yglukhov: we're discussing how to handle nimble packages that are not at the root of a git repo
20:52:33	Araq	so ... how is the poor user of a package supposed to know about
20:52:40	Araq	requires("https://github.com/nimble-test/multi", subdir="alpha") ?
20:53:16	dom96	well, this only applies to packages not in the registry
20:53:29	yglukhov	ok, that sounds like something i would prefer. =) even though i can't see a single reason to do that.
20:53:30	dom96	You'd need to know the url needs to have '/alpha' appended
20:53:56	dom96	But I do too prefer https://github.com/nimble-test/multi/alpha
20:54:04	Araq	a URL is full of cruft already, the '.git' won't kill anybody
20:54:23	dom96	*https://github.com/nimble-test/multi.git/alpha
20:54:36	yglukhov	dom96: how do you know which part of url is git url or subpath
20:54:49	dom96	Everything after *.git/ is subpath
20:55:00	Araq	yeah, please go with that
20:55:06	dom96	hrm
20:55:10	dom96	but what about hg
20:55:19	Araq	hg is dead.
20:55:19	yglukhov	what if the package is served by git url which does not end in .git?
20:55:30	Araq	then make it end in .git.
20:55:45	yglukhov	=)
20:55:48	Araq	or maybe just require the dot in there
20:55:54	Araq	foo.bar/path/here
20:56:19	yglukhov	whats the point in a subpath package anyway?
20:56:25	yglukhov	=)
20:56:28	dom96	monorepos
20:56:43	yglukhov	like it should mean anything to me?
20:56:45	Araq	you don't have to have 5 github entries
20:57:01	dom96	Everything in a single massive repo
20:57:03	Araq	for a single larger package that consists of 5 different but related parts
20:57:38	yglukhov	why are those separate packages then? and not a single package?
20:58:20	dom96	https://github.com/zserge/webview/pull/111#issuecomment-367561320
20:58:29	*	rokups quit (Quit: Connection closed for inactivity)
20:59:09	dom96	(scroll up)
21:01:22	dom96	Maybe I should use // as a delimiter
21:01:30	dom96	https://github.com/nimble-test/multi//alpha
21:01:48	yglukhov	dom96: is it "requires" syntax you're talking about?
21:02:05	yglukhov	and nimble install syntax?
21:02:07	Araq	so ... because github ignores the '//'?
21:02:12	dom96	yes
21:02:16	Araq	and it's generally ignored in URLs?
21:02:37	dom96	Araq: Yeah, because in URLs it's generally ignored
21:02:43	Araq	I prefer the dot
21:02:46	*	el_tejon joined #nim
21:03:13	dom96	I'd use : but that is already used by git/hg
21:03:26	dom96	[email protected]:nimble-test/multi.git
21:03:43	Araq	well if you remove the special syntax in Nimble before you make the request
21:03:49	dom96	oh, it's only used by git: ssh://[email protected]/BitPuffin/linagl
21:04:09	dom96	yeah, I will of course remove it
21:04:20	Araq	then you can use whatever you want to.
21:04:55	dom96	https://github.com/nimble-test/multi/(/alpha)
21:04:56	dom96	:P
21:05:17	Araq	#/ perhaps
21:05:40	dom96	ahh yes
21:05:43	Araq	not sure if that can cause conflicts or not but #/ is for SPAs, right?
21:05:55	dom96	or...
21:05:59	dom96	?subdir=/alpha
21:06:14	Araq	can URLs have comments? :P
21:06:40	*	arthurz joined #nim
21:06:40	dom96	I think query key/values work perfectly here
21:06:47	dom96	and give us the basis for other parameters
21:08:10	dom96	wow, that confuses git
21:09:05	dom96	hg actually detects it and gives a nice error
21:09:13	dom96	abort: unsupported URL component: "subdir=/alpha"
21:09:52	dom96	git tries to clone into multi?subdir=/alpha
21:10:10	dom96	oh well, that does actually make sense
21:10:26	dom96	So anyway, any objections?
21:16:15	yglukhov	not that i care much, but yeah, i think it's not the best idea to invent a new url format, that is a superset of hg/git urls.
21:18:08	*	francisl quit (Remote host closed the connection)
21:20:42	dom96	What would you propose?
21:38:46	yglukhov	dom96: you have already said for myself =)
21:39:37	yglukhov	requires("https://github.com/nimble-test/multi", subdir="alpha")
21:40:58	yglukhov	another option is to forbid multi-package repos
21:41:13	*	Trustable quit (Remote host closed the connection)
21:42:37	yglukhov	tbh i still fail to understand the reason behind it. creating a git repo is not much harder than a new nimble file
21:46:26	yglukhov	but like i said, i'm not going to argue a lot about it. it's not the most critical question to me. maybe a hacky url extension is ok after all :)
21:46:55	*	yglukhov quit (Remote host closed the connection)
21:49:43	Araq	I would use #/ but ?subdir=/bla is ok too
22:03:48	*	nsf quit (Quit: WeeChat 2.0.1)
22:04:01	*	francisl joined #nim
22:18:50	*	athenot quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
22:45:40	dom96	Thought that would be easier to implement
22:46:44	FromGitter	<Vindaar> Well, here we go, opened a PR for getPassword: https://github.com/nim-lang/Nim/pull/7254
22:46:51	FromGitter	<Vindaar> First PR is scary :P
22:47:10	*	francisl quit (Remote host closed the connection)
22:49:52	*	Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
22:51:36	dom96	Nothing to be scared of :)
22:51:41	dom96	Thanks for the PR!
22:53:04	*	Snircle joined #nim
22:55:01	*	vivus joined #nim
22:55:33	*	solitudesf quit (Ping timeout: 264 seconds)
22:57:08	FromGitter	<Vindaar> Haha, no worries. :D I'm off to bed now. good night!
22:57:15	dom96	'night!
22:57:23	dom96	So I think it's about time I release a new Nimble version
22:58:06	FromGitter	<zacharycarter> :shipit:
23:19:23	dom96	https://github.com/nim-lang/nimble/blob/master/changelog.markdown#0810---23022018
23:23:39	*	Sentreen quit (Remote host closed the connection)
23:26:54	*	zolk3ri joined #nim
23:39:19	*	natrys quit (Quit: natrys)
23:50:56	*	xkapastel quit (Quit: Connection closed for inactivity)
23:52:46	GitDisc	<spaceghost> dom96: I was just thumbing through your nimkernel repository. I had to discover it by looking up what you're up to on github rather than see it weeks ago when I was looking at different young kernels on github. You should definitely tag it appropriatel.
23:53:36	GitDisc	<spaceghost> dom96: I was just thumbing through your nimkernel repository. I had to discover it by looking up what you're up to on github rather than see it weeks ago when I was looking at different young kernels on github. You should definitely tag it appropriately.
23:56:30	*	vlad1777d joined #nim