<< 03-06-2020 >>

---

00:00:02	*	ftsf joined #nim
00:08:07	FromDiscord	<Yardanico> @Avatarfighter the best one is probably getFileInfo
00:08:22	FromDiscord	<Yardanico> https://nim-lang.org/docs/os.html#getFileInfo,string
00:13:41	*	krux02_ quit (Remote host closed the connection)
00:21:25	*	ljoonal quit (Quit: ljoonal.xyz)
00:22:12	*	Guest21672 quit (Ping timeout: 256 seconds)
00:25:06	*	ljoonal joined #nim
00:25:44	*	adnan338 joined #nim
00:26:29	*	gangstacat quit (Quit: Ĝis!)
00:28:08	*	dadada__ joined #nim
00:29:54	*	gangstacat joined #nim
00:31:05	*	zacharycarter quit (Ping timeout: 272 seconds)
00:40:29	*	nekits quit (Quit: The Lounge - https://thelounge.chat)
00:42:02	*	nekits joined #nim
00:51:31	*	dadada__ quit (Ping timeout: 246 seconds)
00:58:10	*	dadada joined #nim
00:58:34	*	dadada is now known as Guest74631
01:05:12	*	rockcavera quit (Remote host closed the connection)
01:05:31	*	rockcavera joined #nim
01:13:18	*	chemist69 quit (Ping timeout: 260 seconds)
01:14:32	*	Tlongir joined #nim
01:15:21	*	chemist69 joined #nim
01:17:26	disruptek	zedeus: it's broken for me today, i dunno. maybe i was stoned when i pushed it.
01:18:11	disruptek	zedeus: i know i can do what i want with a stream. i like the idea of bundling a channel api. not sure about anything else, but if the design is right, future abstraction is pretty much unlimited.
01:19:23	*	konvertex quit (Ping timeout: 256 seconds)
01:21:54	*	Guest74631 quit (Ping timeout: 240 seconds)
01:24:56	zedeus	it's not a must, but writing (or thawing) straight to the socket would be nice to avoid a copy
01:26:23	*	opal quit (Ping timeout: 240 seconds)
01:26:27	disruptek	just make a stream api for your thing, i guess.
01:27:15	zedeus	sockets have `write` and `read`, do you need anything else?
01:27:44	disruptek	i mean, if i cannot setPosition/getPosition then that changes things.
01:27:54	zedeus	ah.
01:27:55	disruptek	i don't exploit this now, but i was planning on it.
01:28:11	*	dadada__ joined #nim
01:28:14	disruptek	but it's `whatever`. we need to make this animal in any event.
01:33:26	*	dgb quit (Quit: WeeChat 2.2)
01:33:37	disruptek	and i need to solve recursion differently.
01:34:05	disruptek	well, not recursion, but stack overflow.
01:43:45	*	clemens3 joined #nim
01:46:14	*	dddddd quit (Ping timeout: 256 seconds)
01:52:20	*	dadada__ quit (Ping timeout: 258 seconds)
01:53:06	*	rockcavera quit (Remote host closed the connection)
01:55:05	*	opal joined #nim
01:58:06	*	dadada joined #nim
01:58:12	*	rockcavera joined #nim
01:58:31	*	dadada is now known as Guest72120
02:19:55	*	sendell sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/wYkROJIraZreFWlUuHVuzSMK >
02:20:14	sendell	isnt this supposed to emit a performance warning ?
02:20:38	sendell	"passing a copy to a sink parameter"
02:22:25	*	Guest72120 quit (Ping timeout: 264 seconds)
02:22:27	*	muffindrake quit (Ping timeout: 246 seconds)
02:24:47	*	muffindrake joined #nim
02:28:12	*	dadada__ joined #nim
02:39:59	*	Senketsu quit (Ping timeout: 252 seconds)
02:51:54	*	dadada__ quit (Ping timeout: 240 seconds)
02:56:09	*	ljoonal quit (Quit: ljoonal.xyz)
02:56:28	*	ljoonal joined #nim
02:58:11	*	dadada joined #nim
02:58:33	*	dadada is now known as Guest60363
03:02:41	FromDiscord	<Elegant Beef> Im currently attempting wrap `pty.h` and when i attempt to compile this it says "Undefined reference to `openpty`"↵https://play.nim-lang.org/#ix=2o91
03:02:59	FromDiscord	<Elegant Beef> What am i missing? 😄
03:09:06	*	rockcavera quit (Remote host closed the connection)
03:19:46	*	Tongir joined #nim
03:22:00	*	Guest60363 quit (Ping timeout: 256 seconds)
03:22:42	*	Tlongir quit (Ping timeout: 272 seconds)
03:28:04	*	dadada__ joined #nim
03:28:38	*	lritter joined #nim
03:31:08	*	B4s1l3 is now known as opDispatch
03:36:23	*	opal quit (Ping timeout: 240 seconds)
03:39:34	*	opal joined #nim
03:42:27	*	opal quit (Excess Flood)
03:44:38	*	opal joined #nim
03:46:27	*	zacharycarter joined #nim
03:47:10	*	Tongir quit (Ping timeout: 265 seconds)
03:52:02	*	dadada__ quit (Ping timeout: 260 seconds)
03:55:02	*	waleee-cl quit (Quit: Connection closed for inactivity)
03:58:08	*	dadada joined #nim
03:58:32	*	dadada is now known as Guest58692
04:06:02	*	supakeen quit (Quit: WeeChat 2.8)
04:06:39	*	supakeen joined #nim
04:08:32	*	Tlongir joined #nim
04:11:04	*	zacharycarter quit (Ping timeout: 256 seconds)
04:22:13	*	Guest58692 quit (Ping timeout: 258 seconds)
04:23:16	*	Tongir joined #nim
04:25:52	*	Tlongir quit (Ping timeout: 256 seconds)
04:28:07	*	dadada__ joined #nim
04:50:06	*	narimiran joined #nim
04:52:06	*	dadada__ quit (Ping timeout: 256 seconds)
04:58:09	*	dadada joined #nim
04:58:33	*	dadada is now known as Guest19028
05:02:05	*	opDispatch quit (Quit: Konversation terminated!)
05:06:12	disruptek	zedeus: added socket support... maybe. untested. see if it works or hack it until it does?
05:06:16	disruptek	!repo frosty
05:06:17	disbot	https://github.com/disruptek/frosty -- 9frosty: 11marshal native Nim objects via streams 15 6⭐ 0🍴
05:06:43	zedeus	oh great, thank you. will test when I wake up
05:07:00	disruptek	cool.
05:07:11	FromDiscord	<Elegant Beef> Disruptek save me, before i go crazy!
05:07:23	disruptek	nope, i'm out.
05:07:32	zedeus	after a few hours of painful attempts at memory optimizations, I tried simply making it not display images, and the memory usage decreased to negligible levels
05:07:34	FromDiscord	<Elegant Beef> I mean the appropriate response was "Arent you already"
05:07:45	zedeus	now, to make a zero-copy request proxy..
05:07:59	disruptek	memory copies aren't /that/ expensive.
05:08:10	disruptek	it's allocs that suck.
05:08:34	zedeus	sure, but it gets expensive when you're serving hundreds of images per second
05:08:35	disruptek	beef: what's your problem?
05:09:06	disruptek	zedeus: you make the job harder on yourself, so it's hard to feel compassionate.
05:09:36	FromDiscord	<Elegant Beef> I cant get pty.h functions to be included properly, seems im doing something slightly wrong
05:09:55	disruptek	which pty functions?
05:09:57	FromDiscord	<Elegant Beef> Since complains about undefined reference to the openpty function
05:10:05	FromDiscord	<Elegant Beef> openpty and forkpty
05:10:13	FromDiscord	<Elegant Beef> The only two in the pty.h file 😄
05:10:37	disruptek	where is the code?
05:11:02	FromDiscord	<Elegant Beef> https://play.nim-lang.org/#ix=2o9b
05:11:07	disruptek	hmm.
05:11:31	FromDiscord	<Elegant Beef> I did look at the nim termios implementation to see how that's done
05:12:17	FromDiscord	<Elegant Beef> No clue why it compiles on the playground
05:12:21	leorize[m]	fun fact, that example compiles on the playground
05:12:31	disruptek	yeah, i was just about to say.
05:12:45	disruptek	what's your output?
05:13:03	FromDiscord	<Elegant Beef> https://hatebin.com/haoqepytnl
05:13:24	leorize[m]	well you shouldn't use them anyway
05:13:50	disruptek	link with -lutil
05:13:53	leorize[m]	switch to `posix_openpt`
05:14:25	leorize[m]	go posix-compatible
05:20:08	FromDiscord	<Elegant Beef> Well thanks disruptek, and uhh leorize that just seems like more aimless work for me:D
05:21:08	leorize[m]	well you want to create a badass terminal emulator
05:21:17	leorize[m]	it should be proper badass and work cross-platform
05:21:34	*	Guest19028 quit (Ping timeout: 240 seconds)
05:21:57	FromDiscord	<Elegant Beef> Im not disputing that, but i couldnt even get pty.h to work 😄
05:22:37	disruptek	-lutil doesn't work?
05:22:44	FromDiscord	<Elegant Beef> No it did
05:23:06	disruptek	cool.
05:23:29	disruptek	well, i have a fuckton of porn to catch up on.
05:23:35	FromDiscord	<Elegant Beef> lol
05:23:44	FromDiscord	<Elegant Beef> They make so much daily, i dont know how you do it
05:23:55	disruptek	it's my cross to bear.
05:23:57	disruptek	bare.
05:23:59	disruptek	whatever.
05:24:00	leorize[m]	well posix_openpt is a syscall and it's implemented directly in libc, so you don't need `-lutil`
05:24:03	FromDiscord	<Elegant Beef> Both work in this case
05:24:34	FromDiscord	<Elegant Beef> Well ill look into it, but i have absolutely 0 idea where to
05:25:05	FromDiscord	<Elegant Beef> I know the man page exists
05:26:41	FromDiscord	<Elegant Beef> Do you know the posix equivlent of forkpty?
05:26:49	*	kenran joined #nim
05:27:04	silvernode	So I just found this language https://min-lang.org which is apparently written in Nim
05:27:28	silvernode	Looks like Lisp style syntax
05:27:56	leorize[m]	@Beef forkpty is a convenience wrapper on top of openpty
05:28:12	*	dadada__ joined #nim
05:28:12	leorize[m]	which is a convenience wrapper on top of `posix_openpt` :P
05:28:52	FromDiscord	<Elegant Beef> I mean my understanding was you open a pty then split that for you clients and give them a slave which does magic and works somehow
05:28:56	FromDiscord	<Elegant Beef> I really have 0 idea how pty works
05:30:04	leorize[m]	read posix_openpt man page, it covers how to create a slave device
05:30:26	FromDiscord	<Elegant Beef> I mean im starring at it
05:36:16	FromDiscord	<Elegant Beef> Do you know of documentation that gives more detail on how to use PTYs?
05:36:34	FromDiscord	<Elegant Beef> Like even if i wrap these i have less than 0 idea what to do with the file handles to get function with these
05:37:15	FromDiscord	<Elegant Beef> Nevermind a search got that
05:43:29	leorize[m]	lol
05:43:54	FromDiscord	<Elegant Beef> After reading that it's still a very blurry image
05:44:00	Prestige	Is there a way to return early from a proc at any point? E.g. https://play.nim-lang.org/#ix=2o9i
05:44:24	*	endragor quit (Remote host closed the connection)
05:44:38	leorize[m]	uhmm just use return?
05:44:56	leorize[m]	and maybe not using that `if false` to block your `return`?
05:45:09	Prestige	wow i am tired
05:51:23	leorize[m]	@Beef what are you trying to do?
05:51:47	FromDiscord	<Elegant Beef> Get CLI application support
05:51:55	*	dadada__ quit (Ping timeout: 258 seconds)
05:52:24	leorize[m]	you're basically trying to spawn an application attached to that terminal?
05:52:30	FromDiscord	<Elegant Beef> Yea
05:52:51	leorize[m]	so first do all that pts setup magic
05:53:55	leorize[m]	then `fork()`
05:54:36	leorize[m]	in the child process, close fds 0, 1, 2, or just close stdout, stdin, stderr
05:56:12	leorize[m]	then call `setsid()`
05:57:35	leorize[m]	now `open()` the slave terminal, then `dup2()` the resulting `fd` to 0, 1 and 2
05:57:40	leorize[m]	setup completed
05:58:03	leorize[m]	now `execve()` to a new session leader
05:58:10	*	dadada joined #nim
05:58:34	*	dadada is now known as Guest24122
06:02:50	FromDiscord	<Elegant Beef> I do have to ask, how the fuck did you learn this
06:05:32	FromDiscord	<Elegant Beef> Seems setsid already forks
06:05:42	Zevv	APUE
06:06:12	Zevv	http://www.apuebook.com/
06:06:50	Zevv	ptys and session handling are everybody's nightmares
06:07:05	FromDiscord	<Elegant Beef> I mean it's vastly under documented online
06:07:09	FromDiscord	<Elegant Beef> Same with Xlib imo
06:07:53	FromDiscord	<Elegant Beef> It's been long enough i'd expect a well written explanation of the steps required to support CLI applications in a terminal emulator
06:07:56	Zevv	it's considered so common basic knowledge among unix beards that no one bothers writing about it I guess
06:08:03	FromDiscord	<Elegant Beef> Apparently so
06:08:13	FromDiscord	<Elegant Beef> If i get a working terminal im certainly going to
06:08:30	Zevv	just like the beginners frustration "How the H*LL doe I read a single key from a terminal?!"
06:08:50	FromDiscord	<Elegant Beef> non blockingly do it*
06:08:54	Zevv	well, says the beard, you just have to set your terminal to non-canonical mode
06:09:15	Zevv	with a dazzling sequence of tcsetattr arguments
06:09:54	FromDiscord	<Elegant Beef> The fuck does `*const` mean in C land?
06:10:10	Zevv	well, non-blockingly is something everyone runs into at one point in her career, and it's just complicated on all os's
06:10:31	FromDiscord	<Elegant Beef> Hey i tried to manually read the inputs from the keyboard
06:10:31	Zevv	https://cdecl.org/
06:10:39	FromDiscord	<Elegant Beef> Which ended aswell as you'd expect
06:11:45	FromDiscord	<Elegant Beef> > declare args as array of const pointer to char↵Ah yes that doesnt hurt to read
06:12:13	FromDiscord	<Elegant Beef> So it's an array cstrings?
06:12:39	Zevv	what's your whole line? need more context for a sensible answer :)
06:12:50	FromDiscord	<Elegant Beef> `char *const args[]`
06:13:17	FromDiscord	<Elegant Beef> Or is it an array of pointers to the first element in the string
06:13:46	leorize[m]	@Beef how did I learn that? I read the man pages and guess :P
06:14:06	FromDiscord	<Elegant Beef> See i like guessing but not guessing aimlessly
06:14:12	FromDiscord	<Elegant Beef> some intuition behind the guesses is nice
06:14:21	Zevv	args[] is an array of pointers to char, which are typically used as strings in C
06:14:54	Zevv	well, if you guess often enough, you learn to steady your aim
06:15:26	FromDiscord	<Elegant Beef> You dont get better at something practising it wrong 😄
06:15:51	*	endragor joined #nim
06:17:07	FromDiscord	<Elegant Beef> Well the first two functions work atleast↵https://play.nim-lang.org/#ix=2o9p
06:17:11	FromDiscord	<Elegant Beef> So im on some path
06:18:50	Zevv	you *do* get better, as long as you are effective at observing your results.
06:19:33	FromDiscord	<Elegant Beef> Some programmers write shitty code an never improve so, yea let's hope im in colum b 😛
06:20:03	*	opal quit (Ping timeout: 240 seconds)
06:20:24	Zevv	I also know lousy programmers who write teriffic code, and terrific programmers who write lousy code. It's a difficult subject.
06:22:00	FromGitter	<waghanza> @dom96 I have update results https://github.com/the-benchmarker/web-frameworks/tree/cloudify#results
06:22:02	*	Guest24122 quit (Ping timeout: 256 seconds)
06:22:25	leorize[m]	@Beef: please import them as their real name :P
06:22:45	leorize[m]	pty is not the same as pts (i know, i know)
06:23:59	FromDiscord	<Elegant Beef> you dont mean 1:1 just without the y right?
06:27:54	leorize[m]	yea
06:28:08	FromDiscord	<Elegant Beef> pedant
06:28:09	*	dadada__ joined #nim
06:28:45	leorize[m]	@waghanza once Nim 1.2.2 is out can we see an another update? there are some hot optimization coming with that one :D
06:29:50	FromGitter	<waghanza> @loerize probably not ... but I hope so
06:30:30	FromGitter	<waghanza> tbh, the infrascture is fully sponsored (by digitalocean) and I have first to publicize results
06:30:34	Avatarfighter	wow I did not expect to see nim so high on the list
06:30:55	FromGitter	<waghanza> to thanks them about the awesome stuff I can do since I have a free cloud
06:30:59	FromDiscord	<Elegant Beef> I mean it's a mid level language 😄
06:31:22	FromGitter	<waghanza> @Elegant beef what do you mean ?
06:32:08	FromDiscord	<Elegant Beef> Nim is a mid or low level language depending on you view, so if it lost to interpreted languages it'd be quite disconcerning
06:32:14	FromDiscord	<Elegant Beef> Unless i'm missing something
06:33:49	FromDiscord	<Elegant Beef> I didnt even notice C/CPP o nhere
06:33:52	FromDiscord	<Elegant Beef> on here*
06:33:54	FromDiscord	<Elegant Beef> Lol
06:34:02	FromGitter	<waghanza> I think I understand you, but I don't know, tbh, if it's lower or mid level, since I do not code in `nim` ⏎ but seems to be quite the same level as `crystal`
06:35:36	FromDiscord	<Elegant Beef> It's at the same level as C/C++/Rust AFAIK
06:37:41	*	PMunch joined #nim
06:40:26	*	fredrikhr joined #nim
06:41:27	FromDiscord	<Elegant Beef> And yea leorize you're now going to get bugged by me whenever i actually go to attempt using the pt
06:47:00	leorize[m]	lol
06:51:49	*	dadada__ quit (Ping timeout: 264 seconds)
06:58:15	*	dadada joined #nim
06:58:38	*	dadada is now known as Guest27436
07:14:19	*	opal joined #nim
07:14:53	voltist	So I'm writing a proc to read PNG image data, and I'm getting this really weird problem. When I read a 4-byte unsigned int with streams.readUint32, it gives a number that is deffo wrong (in billions not 13), despite seeming to read the correct amount of data. Any ideas?
07:15:26	voltist	Reading from the right place of course
07:15:27	supakeen	Endianness?
07:16:21	voltist	supakeen: Can that be changed with the streams library?
07:16:36	supakeen	I don't know, but it might be the problem here, see if you get \x13\x00\x00
07:16:39	supakeen	Or the reverse.
07:17:40	supakeen	Oh there is endians.swapEndian32 here: https://nim-lang.org/docs/endians.html#swapEndian32%2Cpointer%2Cpointer
07:17:57	bung	!repo buffer
07:17:58	disbot	https://github.com/oswjk/nimpb -- 9nimpb: 11Protocol Buffers for Nim 15 21⭐ 4🍴 7& 7 more...
07:18:04	bung	wrong
07:19:14	bung	try this https://github.com/bung87/buffer
07:19:42	*	adnan338 quit (Ping timeout: 258 seconds)
07:19:48	bung	also you can take https://github.com/bung87/icon as example
07:20:05	voltist	supakeen "see if you get \x13\x00\x00" How could I see a hex version?
07:20:15	voltist	bung: I'll have a look at that, thanks
07:20:22	bung	repr youvar
07:20:48	voltist	Oh I could just read a single byte and check if its 13
07:22:00	*	Guest27436 quit (Ping timeout: 258 seconds)
07:22:03	supakeen	voltist: I can also ask you, did you get the number `318767104`?
07:22:19	voltist	Yeah I've solved it now
07:22:29	voltist	It is indeed endianness
07:22:33	supakeen	Nice ok :)
07:22:33	voltist	And yes that is the number
07:22:40	supakeen	Then my crystal ball worked well today!
07:22:48	voltist	Indeed
07:23:11	voltist	I considered calling this project crystal actually, but settled on ezra :)
07:23:47	PMunch	(by the way binaryparse was written with the recent knowledge of parsing PNGs, so it should be capable of doing them)
07:25:18	voltist	PMunch: What do you mean 'doing them'? Decoding, reading the chunks, endianness, etc?
07:25:27	PMunch	Yes
07:25:37	PMunch	Well not the endianness part IIRC
07:25:53	PMunch	But it should be the right endianness for PNGs I think
07:26:07	FromGitter	<sheerluck> Nim ball is crystal ball implemented in Nim
07:26:19	PMunch	Haha :P
07:27:31	Prestige	that's hilarious
07:28:08	*	dadada__ joined #nim
07:28:27	*	ljoonal quit (Quit: ljoonal.xyz)
07:28:46	*	ljoonal joined #nim
07:28:53	supakeen	Perhaps the readInt* in streams could take an endianness flag (or it could be set on the stream?)
07:29:01	*	supakeen thinks of a weekend thing to do
07:29:32	*	supakeen edits nimdeas.txt
07:30:31	voltist	PMunch: Could you provide an example of binaryparse's PNG powers?
07:31:08	*	konvertex joined #nim
07:31:23	PMunch	Not really, the most complex thing I did with it was for my old work..
07:31:25	PMunch	https://github.com/PMunch/binaryparse/blob/master/binaryparse.nim#L687
07:31:36	PMunch	That is the most complex example in the wild
07:32:06	voltist	Ok
07:32:10	voltist	I'll have a look at it
07:32:40	voltist	Oooo looks whats coming https://github.com/nim-lang/RFCs/issues/27
07:32:41	disbot	➥ [RFC] Option for endianness in streams
07:33:37	FromDiscord	<Elegant Beef> I think it's clear if you detect big endianness just crash and say "try again later, with different hardware"
07:33:45	PMunch	For PNG you'd need to use the call procedure syntax and write a procedure that can tell which block is next and trigger that parser.
07:34:59	*	Senketsu joined #nim
07:37:17	voltist	How is this swapEndian32 meant to work in https://nim-lang.org/docs/endians.html ? I feel like I'm missing something because of its strange argument order
07:45:16	voltist	Oh got it now
07:46:51	*	Vladar joined #nim
07:47:53	voltist	WTF PNG why is the image a different endianness to everything else
07:48:15	PMunch	Haha, because reasons
07:48:31	voltist	It wishes to hurt me
07:50:20	PMunch	Isn't the image data just in bytes?
07:51:44	*	dadada__ quit (Ping timeout: 256 seconds)
07:52:35	voltist	I meant image size
07:54:02	PMunch	Oh, that is indeed a bit odd
07:54:29	PMunch	Sure it's not just the magic at the beginning that is a certain endianness?
07:55:59	voltist	Oh actually it could be because the rest of the IHDR header is single bytes (so I wouldn't notice)
07:56:22	voltist	Or something like that
07:56:24	voltist	It works anyway
07:58:20	*	dadada__ joined #nim
08:00:39	*	ljoonal quit (Quit: ljoonal.xyz)
08:01:41	*	ljoonal joined #nim
08:22:11	*	dadada__ quit (Ping timeout: 265 seconds)
08:28:08	*	dadada joined #nim
08:28:17	*	ftsf quit (Ping timeout: 246 seconds)
08:28:33	*	dadada is now known as Guest14069
09:20:35	voltist	Anyone here worked with zip/zlib much?
09:22:21	FromDiscord	<Firefell> sent a long message, see http://ix.io/2o9X
09:22:51	FromDiscord	<Firefell> (edit) 'http://ix.io/2o9X' => 'http://ix.io/2o9Z'
09:23:54	PMunch	voltist, not much, but I have tried it
09:24:18	voltist	PMunch: Did you have any trouble with Z_VERSION_ERROR?
09:24:30	voltist	I can't seem to track down why it is happening
09:24:52	PMunch	Yup
09:25:39	PMunch	Just a sec
09:25:46	voltist	All good
09:26:48	PMunch	Change this line to be a uint instead of a uint32: https://github.com/nim-lang/zip/blob/master/zip/zlib.nim#L14
09:26:51	PMunch	That fixes it
09:27:25	voltist	Ah ok
09:27:29	voltist	I'll try it out
09:27:49	voltist	Have you submitted the fix as a pull request?
09:28:33	PMunch	Well... https://github.com/nim-lang/zip/issues/39
09:28:35	disbot	➥ ZlibStreamError due to size mismatch of data types (Linux) ; snippet at 12https://play.nim-lang.org/#ix=2iJQ
09:29:44	voltist	Huh OK
09:30:04	FromDiscord	<willyboar> PMunch: I just rode this https://peterme.net/scraps/web-frameworks-in-nim.html
09:30:37	voltist	Honestly I don't see why it can't be used as a temporary fix until the 'proper' solution is
09:31:08	PMunch	@willyboar, any input?
09:31:19	PMunch	voltist, well, ask Araq :P
09:31:35	PMunch	It's been a year and that library is still broken on Linux which is a shame..
09:31:43	voltist	Might do
09:32:04	PMunch	I wouldn't even mind a when defined(linux) and defined(amd64): uint else: uint32 or something like that
09:32:17	PMunch	Just someone please make it work..
09:33:48	voltist	Yeah an ever so slightly unsightly fix is better than a library that doesnt work IMO
09:34:58	PMunch	Agreed
09:35:21	FromDiscord	<willyboar> I like the doc part
09:37:39	PMunch	Yeah I've been toying some more with that idea
09:37:57	FromDiscord	<willyboar> in the explanations repo?
09:39:09	PMunch	Nah, just on paper so far
09:40:22	FromDiscord	<willyboar> ## comments really don't work w/ jester?
09:41:02	PMunch	Well, they don't break anything
09:41:18	PMunch	But they can't be used to create API documentation AFAIK
09:42:07	PMunch	The idea here was to be able to add them to the routes and then have something similar to "nim doc" that would generate API documentation like swagger.
09:43:58	FromDiscord	<willyboar> something like this?
09:43:59	FromDiscord	<willyboar> http://redocly.github.io/redoc/
09:45:03	FromDiscord	<willyboar> This looks very cool
09:48:31	PMunch	Yes, although they are more typically rendered like this: http://petstore.swagger.io/
09:48:59	*	chemist69 quit (Ping timeout: 252 seconds)
09:49:49	FromDiscord	<willyboar> great tool
09:50:12	*	chemist69 joined #nim
09:50:31	FromDiscord	<willyboar> Also I would like something similar to rails error pages
09:51:17	FromDiscord	<willyboar> This things makes the developers life easy but they are beautiful the same time
09:51:19	FromDiscord	<willyboar> 😛
09:51:39	PMunch	Example?
09:52:02	FromDiscord	<willyboar> wait i think crystal have something similar
09:52:31	FromDiscord	<willyboar> https://github.com/crystal-loot/exception_page
09:55:52	PMunch	Huh, those are actually pretty neat!
09:56:11	FromDiscord	<willyboar> this is inspired by rails
09:57:22	FromDiscord	<willyboar> also the rails routes command
09:59:33	FromDiscord	<willyboar> https://blog.codemy.net/rails-routes-http-methods-and-sql-queries/
09:59:36	*	Vladar quit (Quit: Leaving)
10:01:36	PMunch	I have PTSD from using Rails..
10:01:57	*	dddddd joined #nim
10:02:43	FromDiscord	<willyboar> Rails is the best tool to create fast a slow app
10:02:52	PMunch	Haha :P
10:02:59	PMunch	I spent two days trying to change a route
10:03:20	FromDiscord	<willyboar> But it has some wonderful tools
10:03:23	PMunch	Changed every instance I could possibly think of, but some magic somewhere had it fucked..
10:04:01	PMunch	That's one of the things I've written down, everything should just be "normal" Nim code. You should be able to manually do everything and override everything
10:05:22	FromDiscord	<willyboar> I agree.
10:05:31	*	Guest14069 quit (Ping timeout: 260 seconds)
10:05:38	FromDiscord	<willyboar> Also i would love a nice plugin system
10:06:22	PMunch	That is another thing I've written down :P
10:07:01	PMunch	Basically an easy way to add things like authentication and such
10:09:25	FromDiscord	<willyboar> someone create a PR to add to jester this functionality
10:10:38	PMunch	I think it'd be easier to rewrite the whole thing
10:10:45	PMunch	I mean you could add some of it
10:10:57	PMunch	But the type-safety would be a big breaking change
10:12:03	FromDiscord	<willyboar> it is crucial to decide the path of the framework from the beginning
10:12:34	FromDiscord	<willyboar> jesters path was to be simple
10:13:13	*	dadada joined #nim
10:13:20	FromDiscord	<willyboar> This PR adding plugins it is a huge change
10:13:37	*	dadada is now known as Guest21617
10:14:51	*	gamut joined #nim
10:16:19	gamut	Hi can I get the const name as a string when I only have its value (37) ? const ERROR_CODE* = 37
10:23:00	*	Senketsu quit (Ping timeout: 246 seconds)
10:36:14	*	Guest21617 quit (Ping timeout: 240 seconds)
10:40:56	*	gamut quit (Remote host closed the connection)
10:43:07	Zevv	you need an enum
10:43:38	*	dzamo[m] joined #nim
10:44:36	Zevv	https://play.nim-lang.org/#ix=2oab
10:44:44	Zevv	Idiom in Nim is not to use CAPITAL_CONSTS, though
10:49:15	*	naught-fowl quit (Remote host closed the connection)
10:54:21	*	Vladar joined #nim
11:08:35	silvernode	Man I wish I could read programming documents without loosing attention and forgetting everything in 5 minutes.
11:08:50	FromDiscord	<Rika> add?
11:08:59	FromDiscord	<Rika> or adhd i dont know the difference
11:09:53	silvernode	Yeah I have one of those. It seems to be more common now than ever before. Every new people I meet these days warns me they have A.D.H.D.
11:10:14	silvernode	every new person*
11:11:43	FromDiscord	<j$> it's really easy to not focus, compared to the past, no practice
11:12:26	FromDiscord	<Rika> there are times where i'm laser focused on programming for hours and times where i cant even write a single line in an hour so i dont know if i have it or not
11:15:50	FromDiscord	<willyboar> Ι τηινκ τηισ ηαππενσ το εωερυονε
11:16:07	FromDiscord	<willyboar> (edit) 'Ι τηινκ τηισ ηαππενσ το εωερυονε' => 'I think this happens to everyone'
11:16:10	FromDiscord	<willyboar> sorry
11:17:22	FromDiscord	<Rika> sorry for what
11:17:39	FromDiscord	<Rika> its not exactly a great thing to have adhd is it
11:17:48	FromDiscord	<willyboar> Gitter users will understand 😛
11:17:52	FromDiscord	<Rika> oh
11:17:54	FromDiscord	<Rika> edits lmao
11:18:06	FromDiscord	<willyboar> exactly...
11:20:46	FromDiscord	<willyboar> PMunch this is great too :
11:20:47	FromDiscord	<willyboar> https://github.com/samueleaton/sentry
11:22:29	FromDiscord	<Rika> sounds like something thats kinda easy to recreate at a glance no?
11:23:51	FromDiscord	<willyboar> I don't have an idea
11:24:15	FromDiscord	<willyboar> I didn't try :p
11:26:39	FromDiscord	<KrispPurg> Does rlocks requires threads to be on? And is it better to handle 429s on a http request, using rlocks or locks?
11:26:46	FromDiscord	<KrispPurg> (edit) 'requires' => 'require'
11:29:26	*	NimBot joined #nim
11:29:26	*	FromDiscord joined #nim
11:34:17	PMunch	@Willyboar, that solves a similar goal to what I did in webexperiments
11:34:26	PMunch	Although that was more geared towards production
11:34:33	PMunch	And the entire program didn't have to be re-run
11:39:47	FromDiscord	<willyboar> I was talking mostly for development env
11:40:45	*	theelous3 joined #nim
11:44:19	silvernode	The other day things were just so clear when I was trying to design a program for the first time. Any other time I just stare at one line trying to think of what the next design concept should be
11:44:47	silvernode	Then 3 hours have passed and I am still in one spot, with 2 lines of code now instead of one.
11:44:55	PMunch	@willyboar, yeah for that a system like sentry is nice. Should be easy to throw together in Nim though
11:45:39	PMunch	silvernode, that's why I like to type up some stuff on paper first
11:45:48	PMunch	And explain them to myself as I take a walk
11:46:06	PMunch	That way I get a better picture of what I'm actually trying to do, and what my problems will be
11:47:15	silvernode	I think I need to try having all of us get together in one editor in real time, then press the record button, get on voice chat, and call it a Nimcast. The end result of every Nimcast episode is a program written on the fly
11:47:54	silvernode	Someone shouts out an idea and that's what we make
11:48:50	silvernode	Anyone from the IRC or Forum can join and the main goal is to laugh, have fun and make something cool.
11:49:34	silvernode	I learn better that way I think
11:50:46	silvernode	Someone can be appointed as the director so people can be given a feature to work on.
11:51:17	silvernode	All of that can be planned before each episode.
11:51:22	FromDiscord	<willyboar> silvernode: you can do something similar to disruptek
11:51:35	FromDiscord	<Rika> omg i want to join that shit
11:51:37	silvernode	Yeah he sits on twitch and writes code
11:51:58	FromDiscord	<willyboar> I mean the tools
11:52:00	FromDiscord	<willyboar> 😛
11:52:10	*	krux02 joined #nim
11:52:27	FromDiscord	<Recruit_main707> silvernode: when you talk about all, how many people is that?
11:52:58	FromDiscord	<Rika> prolly 8 is around max
11:53:10	FromDiscord	<Rika> any more than 8 people editing code in real time is prolly @_@
11:53:19	FromDiscord	<Rika> even 8 is prolly @_@ level
11:54:06	FromDiscord	<Recruit_main707> and what code would they/we write?
11:54:08	silvernode	idk, maybe 5 people per episode? A forum post will created prior to each episode and tell people they can reserve their slot. I think of it in terms of a RPG party in a game during a dungeon run.
11:54:37	FromDiscord	<Recruit_main707> that would be cool
11:54:40	silvernode	The forum post will also have a vote on the next topic for the program
11:55:02	FromDiscord	<Recruit_main707> i might try to extrapolate it to another servers (if you dont mind)
11:55:05	FromDiscord	<willyboar> you can do it with a separate nimcasts site
11:55:10	silvernode	if not a vote, then just a bunch of ideas that get mixed up and randomly selected
11:55:30	FromDiscord	<willyboar> and add votelink, twitch link, etc in forum
11:56:06	silvernode	The rules would be that topic submissions have to be very focused so that it can't be a crazy program and can be written in a couple of hours.
11:56:32	silvernode	yeah I would make a nimcast site
11:56:35	FromDiscord	<willyboar> Maximum 2-3 epidodes is better
11:57:10	FromDiscord	<willyboar> it is nice to wait for Fancy Library in Nim episode 2 😛
11:57:20	FromDiscord	<Rika> ~~add custom editor support~~
11:57:27	silvernode	yeah maybe an episode would need to be spit up into parts. Those can be bigger programs occasionally. Then we can tell people to turn in next week to see the conclusion.
11:57:33	FromDiscord	<willyboar> but more than 2-3 epidodes is too much
11:57:48	silvernode	tune*
11:58:34	silvernode	I think most programs would be small enough to fit in one show.
11:58:46	silvernode	or at least should be
12:00:11	*	rockcavera joined #nim
12:00:50	*	lritter quit (Quit: Leaving)
12:00:55	silvernode	I am thinking that all 5 slots will be laid out in the forums post. Each slot will be a role and responsibility , so like slot 1 is the party leader, slot 2 is designer, slot 3 is some other role role
12:01:27	silvernode	Each episode can follow that format
12:01:50	silvernode	Each episode will have different people filling those roles.
12:03:29	silvernode	You will get noobs to fill a role sometimes too which is a great chance to teach and learn in a fun way. Also a good way to learn teamwork skills and fast adaption skills
12:04:30	FromDiscord	<willyboar> well different people in each episode it is difficult
12:04:44	FromDiscord	<willyboar> at least in the beginning
12:04:47	silvernode	Maybe one slot can always be reserved for a programming noob who is new to the community
12:05:11	silvernode	that way you don't have the leader slot potentially be a noob.
12:05:37	silvernode	the leader slot should always be someone who is a seasoned developer.
12:05:56	silvernode	or at least experienced to some degree
12:06:02	*	supakeen quit (Quit: WeeChat 2.8)
12:06:41	*	supakeen joined #nim
12:06:51	silvernode	Lots of potential here but I would like for all of us to contribute ideas to form a solid format that works like a machine. Somethiing streamlined.
12:07:31	silvernode	The number one goal is that it has to be really fun.
12:07:46	silvernode	for the participants and the viewers.
12:08:07	silvernode	Which ultimately means the project topic has to be fun
12:08:25	silvernode	Examle: write a joke generator.
12:09:07	FromDiscord	<Recruit_main707> can projects be somewhat "serious", more than a joke generator for instance?
12:09:28	FromDiscord	<Zed> for all the zeds in this channel i have found our editor!
12:09:30	FromDiscord	<Zed> https://github.com/zedapp/zed
12:09:30	silvernode	Sure, but they have to be fun in some way
12:09:35	FromDiscord	<Recruit_main707> complex rather than serious
12:10:43	FromGitter	<Vindaar> @voltist: sorry, wasn't around lately. Should be here again regularly now!
12:10:57	silvernode	Zed, how come the main website is a whois notice?
12:10:58	PMunch	@Zed, "Welcome to Zed, a code editor built using web technologies" welp that's a shame
12:11:09	FromDiscord	<Zed> lol yeah
12:11:47	FromDiscord	<Zed> but hey it's named after me so im using it
12:11:48	silvernode	The only web technologies editor I am remotely willing to use is VSCode
12:12:37	silvernode	I can admit that VSCode does a great job at fooling me that it's not web browser code.
12:12:45	FromDiscord	<Zed> silvernode: not sure about the website, i didn't even notice it
12:12:47	PMunch	Hmm, I should really give Kakoune another go
12:13:12	FromDiscord	<Recruit_main707> vscode supports multi client edition
12:13:17	livcd	How do you pronounce `Kakoune` ?
12:13:36	FromDiscord	<willyboar> I am using moe just to try it
12:13:52	PMunch	I pronounce it like cocoon, but with an a instead of the first o and a slightly harder k sound
12:14:00	FromDiscord	<willyboar> https://github.com/fox0430/moe
12:14:07	PMunch	No idea if that's how you're supposed to pronounce it though :P
12:14:13	silvernode	I quite like Micro, it's one of those editors where you try a hotkey that makes sense and it just works.
12:14:26	PMunch	moe? micro?
12:14:36	silvernode	it also has linting out of the box
12:14:46	silvernode	micro
12:15:33	FromDiscord	<willyboar> PMunch there are so many...
12:15:38	silvernode	I was writing a quick nim function for testing the other day with Micro and when I saved it, Micro underlined lines of code that had errors.
12:15:56	FromDiscord	<Zed> did they?
12:15:57	PMunch	willyboar, yeah..
12:16:15	PMunch	I mean Vim³ is of course the best one
12:16:43	PMunch	(That's a small 3 for those who don't have a big enough font :P)
12:17:02	FromDiscord	<Zed> how many vims are there?
12:17:19	FromDiscord	<willyboar> original or with the clones?
12:17:31	PMunch	https://github.com/oakes/vim_cubed
12:18:05	FromDiscord	<Zed> i love it
12:18:28	FromDiscord	<willyboar> I have seen it. The #2 Nim library in github stars
12:18:35	PMunch	Haha :P
12:18:39	PMunch	#2?
12:18:47	PMunch	What's #1?
12:19:02	FromDiscord	<willyboar> Nim itself
12:19:04	FromDiscord	<willyboar> 😛
12:20:20	PMunch	Ah, makes sense
12:20:25	PMunch	What's #3?
12:21:19	FromDiscord	<willyboar> nitter
12:22:08	PMunch	Oh really?
12:22:10	PMunch	Huh
12:22:28	PMunch	Oh right, that thing
12:22:42	PMunch	I thought it was like a Twitter clone in karax or something :P
12:22:44	FromDiscord	<willyboar> https://github.com/search?q=stars%3A%3E%3D500+language%3Anim
12:33:00	FromDiscord	<willyboar> PMunch i suggest to create a new repo or even a new org to discuss this framework ideas in the issues
12:33:17	PMunch	The web framework thing?
12:33:25	FromDiscord	<willyboar> yes
12:33:39	FromDiscord	<willyboar> because we will lose them in this chat 😛
12:33:57	PMunch	I guess that might be a good idea, but it could easily devolve into design by comity..
12:34:39	PMunch	I was thinking about discussing it first with alehander92 and potentially others that have shown a particular interest. Then maybe create an MVP before getting feedback
12:35:18	FromDiscord	<willyboar> There is not better place to discuss it
12:36:35	FromDiscord	<willyboar> this is a great way to get some feedback before the MVP
12:37:21	FromDiscord	<flywind> I think we need a better async ecosystem first. Even a robust async database package is rare.😀
12:39:22	*	adnan338 joined #nim
12:41:20	FromDiscord	<willyboar> @waghanza it would be nice to the framework results to add in the end of the line (or somewhere else) an icon with what kind of tool is. Something similar to techempower.
12:42:46	*	lritter joined #nim
12:43:15	FromDiscord	<willyboar> @flywind i think nim could be a leading language in web dev
12:48:39	FromDiscord	<flywind> I agree.
12:50:54	*	adnan338 quit (Ping timeout: 258 seconds)
12:53:18	*	adnan338 joined #nim
13:18:39	silvernode	nitter is pretty cool
13:18:48	silvernode	never came across it until now
13:19:40	silvernode	Nim is so simple to compile and get an efficient web server going
13:22:48	*	waleee-cl joined #nim
13:26:25	PMunch	Hmm, kakoune still seems to lack a feature or two that I'd like..
13:27:33	*	Vladar quit (Quit: Leaving)
13:28:00	*	lritter quit (Quit: Leaving)
13:47:50	alehander92	hm
13:50:11	alehander92	we can do that with willbyoar and PMunch
13:51:02	alehander92	the web framework repo, but i think it might be best to just have a simple chat first, e.g. twitch/mumble or zoon
13:51:04	alehander92	zoom*
13:52:26	PMunch	Sounds like a good idea
13:52:35	alehander92	also, it's good to see what https://github.com/dom96/jester/issues/226
13:52:37	alehander92	is about
13:52:38	disbot	➥ Adding support for "plugins" to Jester ; snippet at 12https://play.nim-lang.org/#ix=2obd
13:53:37	alehander92	ok, when would be a good time for you guys
13:56:22	alehander92	willyboar i know the guy who did the rails error button thing
13:57:40	alehander92	actionable errors*
13:57:53	alehander92	not sure if you were talking about them when you said error pages
14:04:43	FromDiscord	<Zachary Carter> Aren't "plugins" just middleware?
14:04:53	PMunch	Yes
14:05:02	FromDiscord	<Zachary Carter> most web application server frameworks have this concept
14:05:07	FromDiscord	<Zachary Carter> maybe we should call them that then
14:05:53	PMunch	That's what I was expecting to call them in docs and such :)
14:06:02	FromDiscord	<Zachary Carter> because I think people in the web dev world are more familiar with the term middleware
14:06:03	FromDiscord	<Zachary Carter> gotcha
14:08:11	*	zacharycarter joined #nim
14:11:03	PMunch	alehander92, hadn't seen that issue before. That is similar to what I was thinking of
14:16:19	*	Trustable joined #nim
14:16:28	disruptek	IT'S ALL ABOUT THE BENJAMINS
14:16:28	alehander92	one thing i want is
14:16:46	alehander92	to use stuff like nim-json-serialization or similar light-er json impl-s
14:16:56	alehander92	without intermediate json nodes
14:17:14	alehander92	for more effectiveness + type safety
14:18:02	FromDiscord	<Rika> why would intermediate json nodes be less type safe
14:18:05	FromDiscord	<Rika> or less effective
14:18:07	alehander92	and to also to somehow reuse how much we can from type info for validations + api generation + model access
14:18:37	alehander92	Rika because you access dynamic-like values which take more space and have all the same JsonNode type
14:18:53	alehander92	and imho very often one has a clear schema mappable to a type
14:19:05	alehander92	of course, when one doesn't, then JsonNode is very useful
14:19:20	disruptek	zedeus: i just realized i dunno if your socket handles endian for you.
14:19:39	alehander92	but it's a very ruby/python way to do things, not very nim-ish to me when you can just deal with typed values
14:20:02	*	Vladar joined #nim
14:20:04	alehander92	otherwise the JohnAD plugin stuff looks superb
14:20:33	FromDiscord	<Rika> hm makes sense
14:21:27	disruptek	alehander92: you are all over it, agree.
14:21:45	disruptek	just proper types.
14:23:48	disruptek	i might just write the type system, honestly.
14:24:11	disruptek	i have so much experience with this crap and i hate it so much.
14:24:20	disruptek	maybe that's how we figure out our expertise.
14:24:28	disruptek	mine is package management, because i hate it so much.
14:24:33	disruptek	and json, because i hate it so much.
14:24:51	alehander92	what do you mean, write the type system
14:25:14	disruptek	a new json that is actually nimish.
14:25:28	disruptek	idiomatic.
14:25:32	alehander92	but why , if you can just auto-typify it
14:25:38	alehander92	btw!
14:25:47	disruptek	because the semantic is determined by json.
14:25:48	alehander92	that's a great tool. take some json, infer schema/types
14:25:52	alehander92	probably somebody did it
14:25:53	alehander92	tho
14:25:57	disruptek	!repo openapi
14:25:58	disbot	https://github.com/disruptek/openapi -- 9openapi: 11OpenAPI Code Generator for Nim 15 27⭐ 3🍴 7& 4 more...
14:26:02	disruptek	yeah, me.
14:26:08	alehander92	awesome :D
14:26:20	alehander92	but you base it on some existing schemas?
14:26:32	disruptek	you give me the schema, i give you the nim api.
14:26:52	alehander92	but what i am saying is, one can do it without preexisting schema
14:26:56	alehander92	you give tool the json
14:26:59	alehander92	it gives you a schema
14:27:07	disruptek	you give me the json, i give you the nim api.
14:27:17	disruptek	you give me the yaml, i give you the nim api.
14:27:25	alehander92	something like that yeah
14:27:27	disruptek	you give me the aws gateway, i give you the nim api.
14:27:35	disruptek	this is already a thing.
14:27:36	alehander92	i give you a cake?
14:27:42	disruptek	bentley eats cake.
14:27:48	disruptek	it's about his only trick.
14:27:50	alehander92	isn't this no ok for dogs
14:27:57	disruptek	he can make a cheesecake disapear.
14:28:05	alehander92	ah, no chocolate
14:28:06	alehander92	smart
14:28:33	disruptek	he's a dog with spots.
14:28:37	disruptek	~bentley
14:28:38	disbot	bentley: 11https://imgur.com/gallery/yEXiWWG -- disruptek
14:28:38	disbot	bentley: 11a good boy
14:28:53	disruptek	he's a very handsome dog.
14:28:54	alehander92	good stuff
14:29:08	alehander92	i was in the mountains with a dog yesterday
14:29:14	disruptek	yeah?
14:29:18	alehander92	its relationship with his owner
14:29:23	disruptek	which mountains?
14:29:27	alehander92	pushed me to a lot of happy places
14:29:35	alehander92	the .. local mountains
14:29:41	alehander92	not totally sure which one
14:29:46	alehander92	i am very bad at this stuff
14:29:53	alehander92	it had trees
14:30:00	disruptek	this dog is ridiculous with me. i never had a dog that was so loyal.
14:30:13	alehander92	it feels good
14:30:24	disruptek	it sucks for him, because when i'm coding, he isn't playing outside.
14:30:37	disruptek	but he will never leave my side.
14:30:47	alehander92	you just
14:30:54	alehander92	throw exceptions
14:30:58	alehander92	and yell "catch"
14:31:08	disruptek	jesus.
14:31:19	alehander92	sorry
14:31:34	disruptek	there are states in the union where you can be shot for such an offense.
14:31:45	alehander92	i am in the european union
14:31:48	disruptek	let us not speak of it.
14:31:53	alehander92	we dont get offended so easily
14:31:55	FromDiscord	<Rika> ._.
14:32:30	alehander92	cool dogs
14:32:32	FromDiscord	<Recruit_main707> disruptek, about openapi, we would need a flatbuffers parser
14:32:36	alehander92	ok guys, bb
14:32:42	FromDiscord	<Recruit_main707> see ya
14:32:59	disruptek	what's your application?
14:33:23	*	naught-fowl joined #nim
14:33:24	FromDiscord	<Recruit_main707> wdym, what im gonna use it for?
14:33:29	disruptek	yeah.
14:33:54	FromDiscord	<Recruit_main707> to add support for a nim framework
14:34:16	disruptek	what does this have to do with openapi?
14:34:41	FromDiscord	<Recruit_main707> flatbuffers are also based on schema files
14:34:50	FromDiscord	<Rika> ??
14:35:17	FromDiscord	<Recruit_main707> the api part is already done
14:35:40	disruptek	flatbuffers looks awesome.
14:35:43	disruptek	do you have an impl?
14:35:55	FromDiscord	<Recruit_main707> ideally, it would be done using cpp and the things they have, i have personally started porting the go generator, but its though
14:36:00	PMunch	Hahaha, that catch joke alehander92 :P
14:36:00	FromDiscord	<Recruit_main707> > do you have an impl?↵yes
14:36:11	disruptek	where is it?
14:36:17	disruptek	!repo flatbuffers
14:36:18	disbot	https://github.com/Skrylar/skflatbuffers -- 9skflatbuffers: 11FlatBuffers implementation in pure Nim. 15 3⭐ 0🍴 7& 1 more...
14:36:27	FromDiscord	<Recruit_main707> those are the other ones
14:36:29	FromDiscord	<Recruit_main707> https://github.com/Albus70007/NimFlatbuffers
14:36:40	FromDiscord	<Recruit_main707> i have to commit the latest changes
14:38:40	FromDiscord	<Recruit_main707> if i finish the codegen (in cpp) i could try to add them to the official repo
14:38:47	*	Vladar quit (Quit: Leaving)
14:38:50	disruptek	fuck cpp.
14:38:55	FromDiscord	<Recruit_main707> i know...
14:39:15	zacharycarter	why do you need to use C++?
14:39:25	zacharycarter	and why are there multiple flatbuffers repos now?
14:39:37	zacharycarter	what's the point in having two implementations?
14:39:40	FromDiscord	<Recruit_main707> those of google are stubborn
14:39:55	zacharycarter	but surely there are other implementations impelmented in other languages?
14:40:29	FromDiscord	<Recruit_main707> becuase i dont know how skylar's one works, and my idea was to be able to add them to their repo, so i based mine on the other existing implementations of the other languages
14:40:47	zacharycarter	hmm okay
14:40:55	zacharycarter	I still don't get the C++ part
14:41:09	FromDiscord	<Recruit_main707> its a requirement
14:41:15	disruptek	why aren't you using skrylar's impl?
14:41:33	FromDiscord	<Recruit_main707> > becuase i dont know how skylar's one works
14:41:40	disruptek	does it work?
14:41:45	FromDiscord	<Recruit_main707> idk
14:42:07	FromDiscord	<Recruit_main707> i also asked if it was being maintained and he didnt answer
14:42:07	zacharycarter	https://github.com/google/flatbuffers/tree/master/rust/flatbuffers/src
14:42:12	zacharycarter	I don't see any C++ there
14:42:25	disruptek	you don't think you have anything to learn from existing code but you think you'll learn something from a from-scratch impl?
14:42:35	FromDiscord	<Recruit_main707> the implementation its done in each language, the code generator in cpp
14:42:36	disruptek	after you impl it, won't it be existing code?
14:43:05	disruptek	what makes you think you know more about this than skrylar?
14:43:13	disruptek	he actually successfully implement it, right?
14:43:24	disruptek	do you think he learned anything?
14:43:45	zacharycarter	okay now i get the C++ part
14:43:48	FromDiscord	<Recruit_main707> i dont get what you are going at, im sure skrylar knew what he was doing, probably even better than me
14:44:03	zacharycarter	until the codegen part is done though - this is all pretty useless right?
14:44:08	*	BehrangDadsetan joined #nim
14:44:24	*	BehrangDadsetan is now known as BenIsCoding
14:44:30	FromDiscord	<Recruit_main707> ye, essentially, bigger projects need codegen
14:44:48	alehander92	recruit_main707 i think he is trying to say that it's wiser to try to learn/reuse a lib first
14:44:57	alehander92	before redoing it from scratch
14:45:13	FromGitter	<kaushalmodi> > i think he is trying to say that it's wiser to try to learn/reuse a lib first ⏎ ⏎ +1
14:45:17	alehander92	but we often do the second thing because it sounds more fun
14:45:39	FromDiscord	<Rika> i dont know, making a library from scratch doesnt sound so fun to me
14:45:47	disruptek	look, i'm guilty of it, too.
14:45:53	alehander92	well, its' just NIH
14:46:17	FromDiscord	<Recruit_main707> flatbuffers are not that complex, and how they were implemented i doubt google would let us add them to their repo, and if i was writing a codegen, i wouldnt know how skylars implementation managed things
14:46:20	disruptek	frosty is probably better impl in skflatbuffers.
14:46:50	disruptek	who fucking cares about google.
14:47:09	FromDiscord	<Recruit_main707> well, 1st its free propaganda
14:47:19	disruptek	doesn't sound that cheap to me.
14:47:29	FromDiscord	<Rika> arent there better ways to do propaganda
14:47:36	disruptek	you have to replace a good library with another one which may not be as good and doesn't even fucking exist.
14:47:43	FromDiscord	<Recruit_main707> 2nd writing a parser and codegen from scratch is the unfun part of reimplementing flatbuffers
14:47:59	disruptek	tell a friend about nim.
14:48:01	FromDiscord	<Rika> why do we need a codegen
14:48:02	BenIsCoding	Hey everyone! Are Nimmers primarily Windows/Linux users? Does anyone else have a bad time with MacOS in an environment with brew/xcode mix? Any advice for me?
14:48:08	FromDiscord	<Rika> dont we have macros for that or something
14:48:14	FromDiscord	<Rika> there are some people who use macos here
14:48:17	FromDiscord	<Rika> a good amount actually
14:48:20	disruptek	BenIsCoding: it does seem to be mostly windows/linux.
14:48:35	alehander92	it should be usable with osx tho
14:48:35	zacharycarter	I don't have any issues with osx and Nim
14:48:38	disruptek	you won't have a problem with osx.
14:48:41	alehander92	if it has objc codegen
14:48:43	PMunch	Yeah brew puts Nim files in weird places which breakssome tools
14:48:44	BenIsCoding	So maybe doing my work mostly in a docker container might make sense?
14:48:45	FromDiscord	<Rika> dont use brew to install nim is all i know
14:48:56	FromDiscord	<Recruit_main707> Rika, i know macros would be very well fitted for that, but the parser part...
14:48:58	zacharycarter	why are you using xcode with Nim?
14:49:03	disruptek	just download it and build it. you'll be fine.
14:49:05	zacharycarter	or do you just mean the xcode developer tools?
14:49:08	FromDiscord	<Rika> what about the parser part @Recruit_main707
14:49:09	PMunch	BenIsCoding, just install Nim through choosenim and you should be fine
14:49:28	zacharycarter	also remember that on osx gcc == clang due to aliaising
14:49:29	BenIsCoding	I tried to download it and just build it. 1.2.0 and 1.0.6 won't build outside clean docker containers.
14:49:33	disruptek	choosenim is pretty iffy on osx.
14:49:44	zacharycarter	they build fine for me
14:49:45	disruptek	what is the error?
14:49:45	BenIsCoding	choosenim 1.2.0 won't work for me
14:49:45	FromDiscord	<Rika> so is brew isnt it
14:49:56	zacharycarter	I don't use choosenim
14:50:09	disruptek	neither is ben.
14:50:09	FromDiscord	<Recruit_main707> Rika, well, you are criticising that i reimplement an already existing implementation of flatbuffers, but want me to create a .fbs parser and codegen by hand also?
14:50:13	BenIsCoding	Things like docker run -ti zuazo/irssi \ -c irc. freenode.net -p 6667 \
14:50:20	BenIsCoding	oops :)
14:50:37	BenIsCoding	I get errors like use of undeclared identifier '__stdinp'
14:50:38	disruptek	no ssl?
14:50:43	disruptek	use ssl, dude.
14:50:58	BenIsCoding	wait.. is the ssl comment for me?
14:51:02	disruptek	yes.
14:51:04	BenIsCoding	is that in relation to irssi?
14:51:06	zacharycarter	do you have xcode developer tools installed?
14:51:11	disruptek	yes.
14:51:17	BenIsCoding	I had no clue that IRC entered the SSL world :)
14:51:24	disruptek	yes.
14:51:25	FromDiscord	<Rika> @Recruit_main707 i fail to understand what you mean
14:51:28	BenIsCoding	Yes, I have XCOde and everything installed
14:51:36	zacharycarter	not just xcode
14:51:39	zacharycarter	the developer tools
14:51:45	disruptek	command-line tools
14:51:48	zacharycarter	and what compiler are you using
14:51:57	zacharycarter	clang or gcc installed via brew?
14:52:07	disruptek	but it makes him install that when he runs clang from cli.
14:52:10	zacharycarter	it sounds to me like you're missing system header files
14:52:16	BenIsCoding	I followed Arak's advice and tried to manually patch lib/system/io.nim so that it does not use macros but apparently there is stuff in c_code that still is failing even after my manual patching
14:52:29	disruptek	why would you need to do any of this?
14:52:36	disruptek	run clang.
14:52:38	disruptek	what does it say?
14:52:52	zacharycarter	yeah you shouldn't have to do any of this
14:52:58	zacharycarter	Nim should just work on macOS
14:52:58	BenIsCoding	(base) Bent:nim-1.2.0 ben$ clang -v
14:52:58	BenIsCoding	Apple clang version 11.0.3 (clang-1103.0.32.62)
14:52:58	BenIsCoding	Target: x86_64-apple-darwin19.5.0
14:52:58	BenIsCoding	Thread model: posix
14:52:58	BenIsCoding	InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin
14:52:58	disruptek	don't patch anything.
14:53:18	BenIsCoding	I was also surprised that a newbie is asked to manually patch core files :)
14:53:21	disruptek	and show me the output from a clear sh build_all.sh
14:53:28	disruptek	s/clear/clean/
14:53:28	BenIsCoding	But I figured.. it's Arak who suggested it
14:53:36	disruptek	that's nuts.
14:53:40	alehander92	but it wasn't araq man
14:53:40	zacharycarter	well he uses windows so he doesn't know
14:53:42	*	kenran quit (Quit: leaving)
14:53:46	alehander92	just kidding
14:53:47	disruptek	i'm sure he was confused if so.
14:53:49	zacharycarter	he's probably just trying to get it working for you
14:54:01	disruptek	~paste
14:54:01	disbot	paste: 11a frowned-upon behavior in chat; please use a service such as https://play.nim-lang.org/ or http://ix.io/ or https://gist.github.com/ and supply us a URL instead. -- disruptek
14:54:16	BenIsCoding	I was about to ask where to best paste :)
14:54:39	disruptek	if you use ix.io, the bot will move the url into playground.
14:54:47	disruptek	and broadcast it here:
14:54:49	disruptek	~news
14:54:49	disbot	news: 11the #nim-news channel has a Nim news feed of updates to pull requests, issues, and packages. -- disruptek
14:55:22	FromDiscord	<Recruit_main707> @Rika, they said that why would i implement flatbuffers for nim if it was already done, i said it was because in order to create a codegen, it would be easier if it was based on the "official" implementations and if i actually knew how things where being done (unliky with skylar's impl)
14:55:32	BenIsCoding	Here is my gist https://gist.github.com/bdadsetan/7d9ef3677afa1968336e965e8619ec55
14:55:52	FromDiscord	<Recruit_main707> and as we said, flatbuffers without a codegen, are not useful
14:55:56	disruptek	~news is the #nim-news channel has a Nim news feed of updates to pull requests, issues, and packages. The bot also broadcasts ix.io -> playground URLs here automatically.
14:55:57	disbot	news: 11the #nim-news channel has a Nim news feed of updates to pull requests, issues, and packages. The bot also broadcasts ix.io -> playground URLs here automatically.
14:55:57	BenIsCoding	wow.. ix.io looks fun!
14:56:15	disruptek	there's a nim ix paster in here:
14:56:20	disruptek	!repo xs
14:56:22	disbot	https://github.com/disruptek/xs -- 9xs: 11xstreamstartup.com 15 2⭐ 0🍴
14:56:32	zacharycarter	something is very screwed up with your environment
14:56:45	zacharycarter	but I'm not sure what you've done
14:56:59	zacharycarter	did you install clang or gcc or something with brew?
14:57:08	zacharycarter	and how are you clonining / building Nim?
14:57:20	alehander92	recruit_main ok, it was more about the fact
14:57:28	alehander92	you can try to ask skylar
14:57:33	BenIsCoding	I do want to learn nim, I love how expressive it is. I love the idea of having a much better front-end to clang than C++, I love the idea of meta-programming with nim style templates and macros.
14:57:45	BenIsCoding	But this is a tough start.
14:57:47	alehander92	to see if one can reuse his impl, that's all
14:57:50	BenIsCoding	I downloaded the .xz file
14:57:57	zacharycarter	well this isn't Nim's fault...
14:57:59	BenIsCoding	did not clone from git
14:57:59	disruptek	can you build c programs?
14:58:00	FromDiscord	<Recruit_main707> and they also say i shouldnt be using cpp despite having half of the job done, and that i should use macros because they are (obviously) better fitter
14:58:01	alehander92	if it can't be adapted easily, yeah
14:58:10	zacharycarter	you've done something to your environment
14:58:21	zacharycarter	plenty of people install Nim on macOS and don't have issues
14:58:30	BenIsCoding	I was using C and C++ to compile stuff for my mega2560 microcontroller and wanted to use Nim instead
14:58:31	zacharycarter	including myself
14:58:51	disruptek	which osx is this?
14:58:52	BenIsCoding	So compiling C/C++ generally works, although it partially uses avr compile tools
14:58:59	FromDiscord	<Recruit_main707> alehander92 i tried to ask him if it was being maintained before jumping into doing it myself :P
14:59:03	BenIsCoding	Catalina 10.5.5
14:59:11	disruptek	is that current?
14:59:20	BenIsCoding	I believe it's the most current
14:59:24	disruptek	okay, cool.
14:59:36	disruptek	i just want to know if this might be the first of many such queries.
14:59:52	disruptek	but i wouldn't expect a patch release to break us.
14:59:57	FromDiscord	<Recruit_main707> and ive just remembered, skylars impl is not even complete (it doesnt have a builder)
14:59:58	disruptek	but on the other hand, it's clang 11.
15:00:05	disruptek	it's not like clang 6 or clang 8.
15:00:20	BenIsCoding	I believe I recently upgraded my XCode and the Build tools
15:00:23	alehander92	well targetting a microcontroller
15:00:29	alehander92	and using custom tools/targets
15:00:35	alehander92	sounds very specific to me
15:00:43	disruptek	it's xcode clang, though.
15:00:46	alehander92	as often one might use crosscompilers/different headers etc
15:00:46	disruptek	c'mon.
15:00:56	FromDiscord	<willyboar> alehander92 i meant exception pages
15:01:11	FromDiscord	<willyboar> error pages is 404,500 etc
15:01:16	BenIsCoding	I am not even close to using nim to compile for my microcontrollers yet. I am just trying to build nim without using brew on macosx
15:01:20	alehander92	so one should make sure super-normal compilation just targetting osx works normally
15:01:35	BenIsCoding	ok, let me try that then
15:01:39	BenIsCoding	will do a hello world
15:01:56	*	aeverr joined #nim
15:02:00	zacharycarter	brew shouldn't even be in the picture
15:02:12	zacharycarter	I think your problem is you're compiler isn't finding the correct system headers
15:02:15	alehander92	keep in mind `gcc.exe` and `linker.exe` if one has more than one versions
15:02:21	alehander92	willbyoar ah ok
15:02:29	FromDiscord	<willyboar> PMunch i found another crystal reloader
15:02:31	alehander92	recruit ok, i thought you havent talked, sorry
15:02:31	FromDiscord	<willyboar> https://github.com/f/guardian
15:02:33	disruptek	i think it's 64 versus 32.
15:02:52	alehander92	recruit if you've done your research, then it sounds ok, sorry
15:02:52	disruptek	you're pointing at 32-bit headers somewhere.
15:03:36	BenIsCoding	Ok, plain C works. At least my 4 liner hello world example does.
15:03:55	disruptek	what file type is the output?
15:04:09	zacharycarter	how are you compiling the plain C file?
15:04:12	zacharycarter	with what command?
15:04:32	BenIsCoding	$ file a.out > a.out: Mach-O 64-bit executable x86_64
15:04:48	BenIsCoding	I used gcc and clang
15:05:00	zacharycarter	what do you mean gcc and clang?
15:05:11	zacharycarter	you were able to build it with both?
15:05:17	disruptek	tell me something more about what "arak" told you to do.
15:05:19	zacharycarter	keep in mind, on macOS gcc is aliased to clang
15:05:31	BenIsCoding	Yes. I built once with gcc and once with clang. Both times successfully.
15:05:38	zacharycarter	if you type in `gcc -v` you'll see
15:05:39	disruptek	we already know he typed clang.
15:05:45	zacharycarter	okay
15:06:13	disruptek	i wish i was on my osx right now.
15:06:19	BenIsCoding	> **<Araq>** patch io.nim and set stdioUsesMacros to false
15:06:34	zacharycarter	you definitely don't need to do any of that
15:06:42	disruptek	but what was he thinking?
15:07:02	zacharycarter	no clue
15:07:06	disruptek	!last Araq
15:07:06	disbot	Araq spoke in 12#nim 25 hours ago 12https://irclogs.nim-lang.org/02-06-2020.html#13:41:16
15:08:21	PMunch	@willyboar, writing those shouldn't be hard..
15:08:24	BenIsCoding	The conversation was https://irclogs.nim-lang.org/01-06-2020.html#15:23:22
15:08:44	PMunch	!last Zevv
15:08:44	disruptek	cool
15:08:44	disbot	Zevv spoke in 12#nim 4 hours ago 12https://irclogs.nim-lang.org/03-06-2020.html#10:44:44
15:09:18	disruptek	dude.
15:09:24	PMunch	Zevv, I was toying around with my idea of containing all of async in the returned Future: http://ix.io/2oby
15:09:32	disruptek	is it brew clang or not?
15:10:13	BenIsCoding	No, it's XCode clang.
15:10:57	BenIsCoding	All I am trying to do is to create a simple UI with nim that shows what comes through bluetooth from my microcontroller that I was hoping to program in nim. Nim is a good candidate for these tasks right?
15:10:59	disruptek	why would xcode clang load headers from /usr/local?
15:11:12	zacharycarter	they shouldn't
15:11:16	zacharycarter	did you install clang with brew?
15:11:20	BenIsCoding	It might not have. It was my assumption because of the error messages.
15:11:21	disruptek	yes, nim is good for that.
15:11:27	Zevv	disruptek: features and closures are no-no, says mr A.
15:11:27	disruptek	stick with it; we'll sort you out.
15:11:45	PMunch	Zevv, basically the Future you get back has a pointer to an iterator. You should set this pointer to a place where you want to get any new iterators. Then the caller is responsible for checking if it has updated after every call and clear it, and also run all the iterators passed on there to completion.
15:12:18	disruptek	Zevv: what is this cryptic shit?
15:12:25	Zevv	not you, I was talking to PMunch!
15:12:28	PMunch	disruptek, I believe that was directer at me :P
15:12:29	BenIsCoding	brew list and brew cask list do not show any clang
15:12:35	disruptek	lol
15:12:47	Zevv	PMunch: let me grok that
15:12:57	disruptek	you understand how confused i am, right?
15:13:05	Zevv	aren't we all
15:13:10	PMunch	I've gotta walk the dog, so I'll be AFK for a while
15:13:11	disruptek	so this is the link you gave us:
15:13:14	disruptek	"@Araq, currently when I run choosenim 1.2.0 it won't compile because it is looking for the __stderr symbol which is not defined in my /usr/local/include because it assumes I am using the XCode compiler and the XCode system includes
15:13:35	disruptek	"my /usr/local/include and clang are from brew and apparently behave differently than the XCode system includes such as stdio.h
15:13:41	Zevv	PMunch: that makes sense indeed. Problem is that I kind of gave up on iterators and am into this continuation passing thing up to my elbows
15:13:49	BenIsCoding	I might have jumped the shark there. I noticed they are different and it would fit my error.
15:14:07	PMunch	Zevv, what is the problem with iterators?
15:14:09	disruptek	Zevv: i wanna work on it.
15:14:28	Zevv	disruptek: please please do. I can't do it. I started six times, but I can't seem to get there
15:14:44	disruptek	so whatever is head is current?
15:14:46	Zevv	PMunch: well, it's all part of a bigger story. Did you follow any of this CPS stuff?
15:14:47	*	D_ quit (Ping timeout: 260 seconds)
15:15:11	PMunch	I did have a look at it, but not much more than that
15:15:37	Zevv	I was whining here about wanting to have Lua coroutines, and did this mockup coro thing with ucontext. Then you came with your coros and we all had fun. Then Ar4q reached out and told me he was looking into CPS - which disruptek also happened to mention a few weeks ago.
15:16:05	Zevv	I kind of knew about that but never understood it fully. So I got some papers from ar4q, and I feel that CPS might be _the_ way to go.
15:16:20	Zevv	Is sense that ar4q also thinks that, given he calls it "the big game changer"
15:16:41	Zevv	properly implemented it will allow for true coroutines, async, and can also be used to properly implement interators
15:16:56	FromDiscord	<mratsim> if we can fix iterator chaining and make closures more convenient I'm all for it
15:16:58	disruptek	it's a paradigm changer, though.
15:17:20	disruptek	it enables a style of concurrent programming that is unbelievably natural and ergonomic.
15:17:27	Zevv	anyay, I do now grok what it does, I did some baby steps to do a poc in macros, and it has the potential to be pretty great
15:17:31	Zevv	It also unifies async and threads
15:17:34	*	tane joined #nim
15:17:50	disruptek	and memory graph movement among threads.
15:17:52	Zevv	disruptek is rambling about styles of programming, but that's just not the point - we *hide* that style of programming
15:17:54	*	D_ joined #nim
15:18:02	disruptek	which means you get free arc in threaded compute.
15:18:53	Zevv	there's only the problem that someone will need to come up with a proper algorithm to do a transformation from the code you type in and chop it up into a number of separate functions which all do tail calls
15:19:12	PMunch	Hmm, I see
15:19:13	Zevv	the effect is that you get code that runs without a stack - and that's the gist of it.
15:19:25	Zevv	All "memory" is no longer in the stack, but in tiny objects living on the heap.
15:19:35	PMunch	What? It would still have a stack wouldn't it?
15:19:39	disruptek	nope.
15:19:44	Zevv	You can store this object in your fridge, take it out next week and continue where it left off
15:19:48	Zevv	no stack.
15:19:54	disruptek	just data in memory.
15:19:59	FromGitter	<waghanza> yeah, we have plenty of features to add ❤️
15:20:05	PMunch	Ah right, but that's just because you're storing its stack frame on the heap
15:20:06	disruptek	as fast as you can move the data, you can move compute.
15:20:16	Zevv	and that was the insighed I gained from playing with this. You lose the stack, and so you are free to jump wherever you want, whenever you want
15:20:21	FromGitter	<waghanza> however, we are focusing on producing results now
15:20:22	Zevv	You can make any type of control flow *in coed*
15:20:24	Zevv	in code
15:20:38	PMunch	So to "thaw" it you just copy it back onto the stack and execute it
15:20:44	Zevv	so iterators, exceptions, coroutines, it can all be expressed in normal code. No magic from languages or compiles.
15:20:48	disruptek	you could say that.
15:20:52	FromGitter	<waghanza> but a nice `UI` with plenty of documentation
15:21:03	FromDiscord	<mratsim> you don't need ARC for multithreaded tasks
15:21:09	FromDiscord	<mratsim> there is a single owner
15:21:11	Zevv	PMunch: basically, what you get from a function is a "continuation" - that's nothing more than a proc pointer with some data to pass it.
15:21:20	Zevv	So you do `cont->fn(cont->data)`
15:21:24	PMunch	I guess that is similar to what I'm doing in my thing
15:21:30	PMunch	But it's the revere
15:21:30	Zevv	This function has a contract to return a new continuation.
15:21:39	PMunch	reverse*
15:21:40	Zevv	So `while cont: cont = cont->fn(cont)`
15:22:01	PMunch	It returns a structure with a place to register continuations
15:22:20	Zevv	it takes continuation and returns a continuation.
15:22:21	FromDiscord	<mratsim> https://github.com/mratsim/weave/blob/master/weave/workers.nim#L41
15:22:26	*	BenIsCoding quit (Quit: leaving)
15:22:34	FromGitter	<waghanza> for example, some php frameworks use nginx, and other use swoole, I think it deserve a documentation
15:22:34	Zevv	You can also have an event looop with a special function: it takes a continuation, and returns nothing.
15:22:50	Zevv	But secretly it stored your continuation somewhere, and calls it when a timer expires
15:22:56	Zevv	So there you have event handling -> async
15:23:01	FromDiscord	<mratsim> continuations are basically generalized closures: https://github.com/mratsim/weave/blob/master/weave/datatypes/sync_types.nim#L32
15:23:06	Zevv	you can also have a pool of continuations, and call them in round robin order
15:23:15	Zevv	So you get fibers
15:23:23	*	BenIsCoding joined #nim
15:23:31	disruptek	all of computation is available.
15:23:34	Zevv	you can also make a proc that passes two continuations back and forth. Thats producer-consumer coroutines
15:23:40	disruptek	but first we have to make it work.
15:23:46	Zevv	right
15:23:49	*	fredrikhr quit (Read error: Connection reset by peer)
15:23:58	Zevv	that's just details :/
15:25:44	Zevv	anyway, it would allow a lot of stuff that otherwise would need help from the language implementation.
15:25:55	PMunch	This does indeed seem very similar to what I'm doing. Just that I'm not calling a continuation, I'm registering it so that my caller can call it.
15:26:34	*	fredrikhr joined #nim
15:27:21	Zevv	I'm first cooking dinner and all, but I will take a proper look at your snippet. In the mean time if you are bored: https://github.com/zevv/nimcsp
15:27:22	PMunch	Hmm, well not quite..
15:27:52	Zevv	The most important thing of it all is https://github.com/zevv/nimcsp/blob/master/eventqueue.nim#L30
15:29:53	disruptek	are you sure you cannot represent goto?
15:29:58	disruptek	i thought we talked about this.
15:30:19	PMunch	Oh well, I really have to walk the dog and get some dinner now
15:30:51	Zevv	disruptek: well, I can represent it by a stub function
15:31:02	Zevv	that was try #3 and #4
15:31:13	Zevv	and the target I tried to represent with a block:
15:31:35	Zevv	problem is that i also seem to need 2 nested macros
15:31:40	Zevv	one untyped, and one typed
15:31:58	Zevv	(I hate these names, but i need 1 before semcheck, and one post-semcheck)
15:32:13	Zevv	so the before semcheck gets code thats not valid nim - it needs to transform it to thaat
15:32:24	Zevv	it can do the basic splitting and all
15:32:48	disruptek	fuck dude, we'll just do it in the compiler.
15:32:54	Zevv	but then comes the lambda lifting. Cant do that afer semheck because it is not valid with free variables
15:33:16	Zevv	so need to do the splitting after that and abuse the scope of nested functions as intermediate step
15:34:41	Zevv	so my current path is: untyped macro does the splitting with temporary virtual gotos. The result is the function properly split into tail calls, but these are still nested so that all locals are valid through the nested scopes
15:35:17	Zevv	untyped macro does not have enough info to do lambda lifting though - everything is just and "ident" still - so it is passed out and eaten by a typed macro
15:35:58	Zevv	that will figure out that it is actually symbols and not idents, and that might be able to lift the lambdas. FOr each nested proc it will create a Cont derived object with members that correspond to the locals in that proc
15:36:22	Zevv	it will then take the local out and add it to the preamble like "var foo = cont.MyCont.foo"
15:36:32	Zevv	so the function body never has to know this happened
15:36:46	Zevv	if that is done it can take the proc out of it parent scope and make it top level.
15:37:01	Zevv	does any of that make sense?
15:37:51	disruptek	yes, just digesting the last part.
15:38:01	*	narimiran quit (Ping timeout: 264 seconds)
15:38:37	Zevv	oh PMunch I see now, thats pretty slick, why didnt I think of that
15:46:39	*	Trustable quit (Remote host closed the connection)
15:53:19	*	dadada joined #nim
15:53:22	dadada	hello
15:53:49	dadada	why is it legal to overwrite a parameter with let parameterName = something?
15:53:57	disruptek	seems to me this code works.
15:54:08	disruptek	what don't you like about it?
15:55:24	dadada	for one thing, the inconsistency that normally you can define a identifier only once, for example, let a = 1; let a = 1; would not be legal
15:55:55	disruptek	oh, i was talking to zevv. sorry.
15:56:12	disruptek	do you have a playground?
15:57:22	disruptek	it's legal for technical reasons. it turns out it's handy because of the `var a = a` idiom to give you a mutable version of an immutable parameter.
15:57:35	FromDiscord	<Generic> that's one reason
15:57:45	FromDiscord	<Generic> shadowing can be useful so you don't have to come up with akward names after casting
15:57:59	disruptek	but this is a bad idea.
15:58:09	FromDiscord	<Generic> why is it a bad idea?
15:58:11	disruptek	we talked about making it an error/warning.
15:58:28	FromDiscord	<Generic> that would break a ton of code
15:58:28	disruptek	because at two points in the same scope, the symbol had different types.
15:58:36	disruptek	of course.
15:58:45	disruptek	that doesn't mean that code isn't already broken.
15:59:25	disruptek	it's magic that is invisible to me as code reviewer.
15:59:42	disruptek	i may have no idea what types are arriving from different calls.
15:59:52	FromDiscord	<Generic> ah yeah, having in the same scope multiple variables with the same name is confusing
16:00:03	FromDiscord	<Generic> though I would allow it to override variables from a previous scope
16:00:03	disruptek	this is valid code? var s = 3.float; var s = 3.int; var s = "3"
16:00:23	disruptek	of course.
16:00:37	disruptek	but that's not the same for more important reasons.
16:00:56	FromDiscord	<Generic> that doesn't seem to be valid code
16:01:21	disruptek	this is valid code? var s = 3.mayProduceFloat; var s = 3.mayProduceInt; var s = "3".mayProduceStr
16:01:35	FromDiscord	<Generic> it seems to be incidently to be the behaviour I described
16:01:46	FromDiscord	<Generic> so you can only describe variables from a previous scope
16:01:59	disruptek	what?
16:01:59	FromDiscord	<Generic> *override
16:02:10	FromDiscord	<Generic> https://play.nim-lang.org/#ix=2obU
16:02:14	FromDiscord	<Generic> try removing the block statements
16:04:11	disruptek	we're talking about changing type of a symbol introduced in a proc parameter within the body of the proc, right?
16:04:22	disruptek	i'm saying it should not be allowed.
16:04:27	disruptek	you're saying it should, right?
16:05:06	FromDiscord	<Generic> ah, I thought we we're only talking about local variables shadowing other local variables
16:05:25	disruptek	i'm saying i can pass you x today and i have no idea what i'm getting back. the return time is not on my screen due to type inference.
16:05:34	disruptek	tomorrow you might change the type.
16:05:49	disruptek	is that okay?
16:06:44	disruptek	does the consumer choose how to use the data or does the producer?
16:06:59	disruptek	imagine that your code follows your types, not the other way around.
16:07:13	FromDiscord	<Generic> that's a general problem with type interference
16:07:58	FromDiscord	<Generic> that if I change the return type of a proc, it might not immediately produces an error everywhere where changes would be necessary
16:08:04	disruptek	maybe, but i don't think it's too much to ask that we have this small limitation in symbol lookup in order to prevent bugs.
16:08:14	disruptek	it could always be relaxed later.
16:08:34	disruptek	it costs you almost nothing to permit it and it saves bugs.
16:08:49	disruptek	what do you want to do? quibble over a character substitution?
16:09:45	disruptek	procs follow types. fite me.
16:11:23	FromDiscord	<Generic> I think there are legit uses for shadowing parameters
16:11:34	FromDiscord	<Generic> say you cast the parameter so some other type
16:12:05	FromDiscord	<Generic> if some rest of the proc still depends on the type of the parameter
16:12:21	*	Senketsu joined #nim
16:12:32	FromDiscord	<Generic> the difference between those two types will be great enough that it would generate an error
16:14:17	Zevv	disruptek: "seems to me this code works", referring to PMunch snippet?
16:16:18	*	BenIsCoding quit (Quit: leaving)
16:17:16	*	disruptek quit (Ping timeout: 272 seconds)
16:17:36	*	disbot quit (Ping timeout: 256 seconds)
16:18:36	*	Senketsu quit (Ping timeout: 246 seconds)
16:23:16	*	disruptek joined #nim
16:23:44	*	disrupte1 joined #nim
16:23:47	disrupte1	ugh.
16:24:07	disrupte1	prolly gonna lose electricity again, but...
16:24:14	*	disbot joined #nim
16:24:40	disrupte1	my point was, we don't know when we might have to destroy the parameter symbol. it's not something you can predict.
16:24:58	disrupte1	rather, we want the liberty to choose how to specify its behavior.
16:28:20	*	disrupte1 quit (Client Quit)
16:48:17	*	bebarker joined #nim
16:51:08	*	bbarker quit (Ping timeout: 246 seconds)
17:00:30	*	FromDiscord quit (Remote host closed the connection)
17:00:44	*	FromDiscord joined #nim
17:07:19	*	FromDiscord quit (Remote host closed the connection)
17:07:32	*	FromDiscord joined #nim
17:13:48	FromDiscord	<Shucks> Using the winim library. How could I convert a wchar array to a nim string? Pretty old same question here: https://forum.nim-lang.org/t/580
17:16:33	disruptek	!last zevv
17:16:34	disbot	Zevv spoke in 12#nim 26 hours ago 12https://irclogs.nim-lang.org/02-06-2020.html#14:34:29
17:18:06	disruptek	Zevv: no, your code works. do you just want to remove hacks?
17:18:25	FromDiscord	<willyboar> disruptek do you know C#?
17:18:35	disruptek	not really.
17:18:47	disruptek	i only hack it.
17:19:11	FromDiscord	<willyboar> a streamer has a problem with
17:19:15	FromDiscord	<willyboar> https://www.twitch.tv/lana_lux
17:19:54	disruptek	they probably know more about it than i do.
17:20:09	disruptek	i only hacked in ffi cache between python and c#.
17:23:21	FromDiscord	<willyboar> by the way i hope you return to streams soon
17:23:47	disruptek	i just can't afford it, honestly.
17:24:14	FromDiscord	<willyboar> I totally understand
17:24:30	Prestige	time cost?
17:24:41	disruptek	opportunity cost.
17:24:47	Prestige	makes sense
17:27:52	*	sagax quit (Ping timeout: 256 seconds)
17:33:06	Zevv	disruptek: it works, but indeed, the "hacks" need to go. And it works only for trivial cases
17:33:32	Zevv	but this code skips the "goto" part, and I can think things with an 'if' and a 'break' where it is just wrong
17:33:50	Zevv	so I think it is safer to first do as the paper says, including the intermediate (virtual) goto step
17:34:10	Zevv	when that's done in the untyped macro, the typed macro picks it up and figures out which locals to lift and put in the derived Cont
17:34:18	disruptek	okay.
17:34:27	Zevv	and when it is all kind of working, we throw it away and someone does it properly in the compiler
17:34:34	Zevv	because I'm 100% sure this will not fly with a macro only
17:34:40	disruptek	really?
17:34:43	Zevv	really!
17:34:58	Zevv	like, using a variable injecting template - no way to tell
17:35:19	Zevv	and how the hell do we split up a non-closure iterator 'for' loop?
17:36:01	Zevv	so this is strictly proof-of-concept only
17:36:23	Zevv	but it should be sufficient to write a few simple networking apps, some coro filters, an alternative iterator implementation and things like that
17:36:27	disruptek	it's a limited subset of what you want, maybe, but you start somewhere.
17:37:25	Zevv	right. And as a playground it's way friendlier then compiler hacking. The basic algo should be the same in NimNodes or PNodes, I guess
17:37:39	disruptek	why do you say you cannot tell with a template injection?
17:37:51	disruptek	isn't the input typed?
17:38:11	Zevv	i mean; if I have a piece of code that calls an unhygienic template, that template will inject locals into scope
17:38:23	Zevv	but I can't see that in a macro - I can only see a template invocation
17:39:06	Zevv	I'm not even sure lifting locals is universally possible in a macro
17:39:19	disruptek	why is the first pass untyped?
17:39:31	Zevv	btw, your "!last" is broken.
17:39:39	disruptek	well, i lost power there.
17:39:40	Zevv	It says 26 houres, but it was like, 01:02
17:39:44	Zevv	ah right :)
17:39:59	disruptek	i doesn't write at every message.
17:44:26	*	disbot quit (Ping timeout: 246 seconds)
17:44:34	*	disruptek quit (Ping timeout: 260 seconds)
17:48:05	*	disruptek joined #nim
17:48:10	disruptek	damnit.
17:48:33	FromDiscord	<Zachary Carter> huh.... how is this possible? https://github.com/septag/rizz/blob/8d3d4e11fee51c2153c970485a2289218272b6a0/src/rizz/graphics.c#L2267 & https://github.com/septag/rizz/blob/8d3d4e11fee51c2153c970485a2289218272b6a0/3rdparty/sokol/sokol_gfx.h#L12582
17:48:52	FromDiscord	<Zachary Carter> `_sg_lookup_pipeline` is static but it's accessed in another c file...
17:49:03	*	disbot joined #nim
17:49:57	disruptek	btrfs can fsck right the fsck off.
17:58:18	*	dcmertens joined #nim
17:59:19	Zevv	disruptek: I can think of more complications. What about `if a() and b(): ...`. If a() is false, b will not be evaluated - effectively control flow needing goto's
18:00:31	disruptek	but so what?
18:00:44	disruptek	first of all, we can put gotos in if we want.
18:01:06	Zevv	I guess so.
18:01:27	disruptek	why not?
18:01:53	Zevv	not sure. it still doesn't click for me somehow
18:02:16	Zevv	I can do it by hand, but I can't seem to write an algo to do it for me
18:10:14	*	Trustable joined #nim
18:10:27	disruptek	why doesn't Cont_tocker_2 inherit from Cont_tocker_1?
18:10:56	Zevv	hm I guess it could, and probably should, as they come from nested scopes
18:10:59	Zevv	good point
18:11:29	disruptek	we know who is in which scope by looking at the parent type.
18:11:46	Zevv	true
18:12:17	disruptek	so that lets us write the prelude, right?
18:12:50	Zevv	that's the idea - if we have a good way to make out the locals.
18:13:54	Zevv	also, this first try starts by transforming a proc. I have another start where I start from a block, this might be easier because it doesn't need a special case for handling the proc itself: https://github.com/zevv/nimcsp/blob/ico/xfrm2.nim#L108
18:14:04	Zevv	'spwanPre' is a bad name, but that's where the transormation starts.
18:14:46	Zevv	not sure if it is better or worse in usability, but it makes the xfrm proc a bit simpler as it does not have to care about nnkProcDef
18:16:11	disruptek	of course, if we change the type then we may not be able to fold identical continuations.
18:16:20	dcmertens	I am writing a function that takes to arrays that should be the same length. If they are not the same length, which kind of exception should I raise?
18:16:32	disruptek	value error
18:16:33	dcmertens	I saw a list here: https://github.com/nim-lang/Nim/issues/8363
18:16:35	disbot	➥ Rework Nim's exception handling ; snippet at 12https://play.nim-lang.org/#ix=27OQ
18:16:50	dcmertens	disruptek, cool, thanks
18:19:04	disruptek	Zevv: makes sense.
18:21:19	disruptek	Zevv: really, we only care about symbols that get used, right?
18:22:14	disruptek	we can just drop everything else off the stack.
18:22:28	*	natrys joined #nim
18:23:31	disruptek	some of it we can probably even lower.
18:28:41	Zevv	well, you *only* want to pick up what is used in that scope
18:28:59	Zevv	you don't want to move around stuff that's not needed
18:30:58	disruptek	right, so we drop stuff and lower stuff.
18:31:18	disruptek	basically, we cut the scope in half. some stays, some goes.
18:31:42	disruptek	always the parent goes. so now it's /some/ of what remains between the continuation and the start of scope.
18:31:58	disruptek	we could start by shipping everything.
18:32:10	disruptek	what's the worst that could happen?
18:32:25	disruptek	if you use what we give you, then it's in use.
18:32:33	Zevv	I think if we can do this in one recursive pass it will happen by magic. You descend the AST depth first, until you come to an inner block, goto-ify as needed and lift what you find. Being done the recursion goes up one level, but it will only find a nested proc there, which it can effectively discard as it is guarenteed not to use anything from its outer scope
18:32:35	disruptek	else, it's just a pointer that is needlessly copied.
18:32:53	Zevv	the question is still: why does the author take the 'goto' route - is there a reason this can not be done in one pass?
18:33:14	Zevv	I think we're kind of on the same track here
18:33:15	disruptek	i was thinking we'd start from the bottom, but now i'm not so sure.
18:33:33	Zevv	well, the bottom of the AST tree I guess
18:34:34	Zevv	it's not linear. It's blocks. I think it's only block: while: and for: that create a true block.
18:35:13	Zevv	I'm just not sure of the exact algorith. Every time I do it by hand it's kind of fiddling about, intuitin. I don't have a strict set of rules yet, and the papers are too vague about that step
18:36:03	disruptek	why is it guaranteed not to use anything from its outer scope?
18:36:23	Zevv	because all its locals are lifted at that time.
18:36:49	Zevv	well, not only locals, all free variables also
18:37:02	disruptek	oh, so you just start from a scope that you know holds only used variables.
18:37:10	disruptek	then you work backwards.
18:37:25	Zevv	depth first
18:37:47	Zevv	but then again - if the author has a good reason to do the goto's first, that probably means there's no straight resursive way
18:38:18	disruptek	i cannot see why, though. to me, it's way better.
18:38:39	disruptek	when i jump to the next cont, i can also free any portion of my env that i don't need.
18:38:57	disruptek	i get it for free, basically.
18:39:01	Zevv	sure, the only thing that counts it what the next cont wants
18:39:05	Zevv	the rest is discardable
18:39:25	Zevv	I have problems writing this code. "kont" means "ass" in dutch.
18:39:46	disruptek	perfect.
18:39:59	Zevv	I do like the inheritence tree of continuation types tho - makes perfect sense
18:40:22	disruptek	well, not any more.
18:40:54	disruptek	it might end up being more efficient, though.
18:40:55	Zevv	we also need a postamble by the way, to put locals back into the contination of the next call
18:41:08	Zevv	postlude
18:41:13	Zevv	what the hell do you call that
18:41:26	disruptek	you call it the ass, of course.
18:41:35	Zevv	haha, that might stick :)
18:41:41	Zevv	"epilogue"
18:42:25	Zevv	anway, these are the ContXXX constructors in the 'returns'
18:42:41	disruptek	i guess technically we could repack the env each call.
18:42:51	Zevv	it changes over calls.
18:43:06	disruptek	i know. maybe it makes more sense to let the c compiler optimize it.
18:43:21	Zevv	Oh right, I assume the packing/unpacking itself will get mostly optimized away as well
18:43:37	disruptek	we just pass a link to the parent.
18:43:48	disruptek	everything's a pointer.
18:44:23	Zevv	hmm not sure about that
18:44:36	Zevv	nothing is a pointer, it's only data
18:44:46	Zevv	please copy away and let the compiler do the smarts here
18:44:53	Zevv	waht if you want to send a closure to a different thread
18:45:08	disruptek	what if you have a large object in scope?
18:45:16	disruptek	there's no stack.
18:45:22	Zevv	copy it into the contiuation
18:45:27	Zevv	it's the only way
18:45:36	disruptek	only when you need to move the graph.
18:45:40	disruptek	else, why copy it?
18:45:48	Zevv	if 'next' needs it, you copy it
18:46:02	disruptek	why copy, though?
18:46:05	Zevv	threads
18:46:30	disruptek	but this only makes sense if you are actually moving.
18:46:37	disruptek	that can be done outside of this.
18:46:40	Zevv	hm ok, fair enough
18:46:50	Zevv	but still, what do you pointerize and what do you copy?
18:46:59	Zevv	I'd rather start with copy only
18:47:22	disruptek	look, i already know what's in scope. everything else, i delegate to the parent.
18:47:36	disruptek	i do my stuff, you do yours.
18:47:54	Zevv	doesn't feel good.
18:48:07	disruptek	you can always copy it if you want.
18:48:47	Zevv	don't go there.
18:48:48	disruptek	if you do, you're just copying a chunk of memory. not bad. you have the type so you know what everything is; it's a singly alloc.
18:48:53	Zevv	Stuff that needs to be reffed is already a ref, probably
18:49:08	Avatarfighter	What would be a good way of reducing the memory footprint of gigabytes of strings in a program? I'm thinking of implementing a radix tree to reduce my programs memory but I'd like to know what you guys do to optimize memory lol
18:49:09	Zevv	copying is virtually free in 2020
18:49:22	disruptek	but not memory consumption.
18:49:38	disruptek	Avatarfighter: why keep them in memory?
18:49:39	*	natrys quit (Quit: natrys)
18:49:40	Zevv	disruptek: it's no differetn from a stack. There's always only 1 contination 'live' holding the data
18:49:59	disruptek	then there's no reason to copy anything.
18:50:25	Zevv	There is. You ref something and you store your continuation away. Someone else goes messing with your reffed data.
18:50:28	Zevv	that was not in the contract
18:50:34	Avatarfighter	disruptek: I'm currently keeping urls to be scraped when my workers are available in memory in a sort of queue atm but I ran into a problem with a certain site where gigabytes of urls are in memory forcing my kernal to reap my process
18:50:47	disruptek	Zevv: use the heavier version of cps, then.
18:50:54	disruptek	who gets to choose the semantic?
18:51:11	Zevv	So, does reffing make it more simple or more complicted to make a poc?
18:51:22	Zevv	because you don't want to ref a bool
18:51:31	Zevv	so you need 2 ways to do stuff - copying and reffin
18:51:35	Zevv	2 ways is more work then 1 way
18:51:53	Zevv	Avatarfighter: same question again: why keep them in memory?
18:52:27	disruptek	2 ways is twice as many ways.
18:52:37	Zevv	right
18:52:37	Avatarfighter	Zevv: I dont really know where else to put them tbh
18:52:42	Avatarfighter	I'm looking for ideas lol .-.
18:52:59	disruptek	so you copy your refs. where is your god, now?
18:53:17	disruptek	Avatarfighter: files are a good place to keep data.
18:54:15	Avatarfighter	they are, aren't they :P
18:54:44	Zevv	Right. And if you think files are for grandpa's only, you can whip yourself up some nice database
18:54:57	Zevv	or an apacha kafka instance might fit well, since you're queuing
18:55:23	disruptek	amazon sqs and just run 1000 scrapers at once.
18:55:30	*	fredrikhr quit (Ping timeout: 260 seconds)
18:55:30	Zevv	You can also encapsulate the url's in ICMP ping packets and send them on a round trip to new zealand. By the time they come back they have the same payloud and you can handle them again.
18:55:40	Avatarfighter	To be honest I was keeping stuff in memory because the sites I usually scraped weren't too big so I didn't get too much memory usage lol
18:55:46	Zevv	effectively you're storing your data in tubes only.
18:56:09	disruptek	you're storing it in other hosts.
18:56:21	disruptek	may as well use email; it generally has a 4-day retention.
18:56:26	Zevv	no, you're storing it in the cables
18:56:40	disruptek	it spends most of its time in hosts.
18:56:52	disruptek	the speed of light is pretty fast these days.
18:56:53	Avatarfighter	Store your data in ping requests and fold your router with gigabytes worth of pings
18:57:01	Avatarfighter	flood*
18:57:13	Zevv	I made this project once that did that - it was a fuse block device on top of raid, and the underlying blocks were managed by a little program juggling packets to far away hosts.
18:57:23	Zevv	I was able to store a few kb of that on that, but it was not on my machine.
18:57:41	Zevv	If you want to read a file you need to wait for the proper pings to come back. And every ping that comes back in is sent out again.
18:57:53	Avatarfighter	yeah that seems really interesting actually lol
18:57:58	disruptek	lose a packet, lose a file.
18:58:13	Avatarfighter	spend the file 3 times for that redundancy
18:58:14	Zevv	^^ "bevice on top of raid"
18:58:15	Avatarfighter	send*
18:58:36	disruptek	i need something just a hair more reliable than this to replace btrfs.
18:58:44	Zevv	just a *hair* right
18:58:53	disruptek	yeah, i'm not aiming high here.
18:58:56	Avatarfighter	Dang I might have to start hoarding data via packets
18:58:57	disruptek	i'm aiming hair.
19:00:04	disruptek	gonna try bcachefs i thinks.
19:10:28	*	FromDiscord quit (Remote host closed the connection)
19:10:45	*	FromDiscord joined #nim
19:20:58	*	narimiran joined #nim
19:23:17	*	Senketsu joined #nim
19:24:07	FromDiscord	<willyboar> !last yardanico
19:24:08	disbot	Yardanico spoke in 12#nim 28 hours ago 12https://irclogs.nim-lang.org/02-06-2020.html#14:26:03
19:33:24	*	terps joined #nim
19:46:47	*	dcmertens quit (Quit: Leaving)
19:47:35	*	FromDiscord quit (Remote host closed the connection)
19:47:50	*	FromDiscord joined #nim
19:48:20	Zevv	disruptek: still here?
19:52:20	*	NimBot joined #nim
20:10:11	Zevv	in case you're reading up: I need to do this more structurally. I'll try to write up collection of patterns and their transformation equivalents, and see how they compose. With luck this will work recursively directly transforming to procs. If not, it will be harder to traverse the AST since parts of the tree will move about.
20:10:33	Zevv	I can write a bunch of minimal transforms like this: http://ix.io/2odg and use those as unit tests as we go
20:11:03	disruptek	makes sense.
20:11:19	Zevv	and maybe we can even use this to drive the macro by using pattern substitutions instead of manual AST handling code
20:11:24	Zevv	but don't count on it :)
20:12:26	Zevv	also, I don't think there is a way to iterate non-closure iterators without 'for', so I'm not sure if for loops are transformable at this time. Probably not.
20:12:51	Zevv	these iterators are inlined, but we can't see that
20:12:55	disruptek	one thing at a time.
20:13:13	disruptek	but i'm not sure what pattern subs would buy us.
20:14:31	disruptek	seems like not having for loops is ideal.
20:17:12	*	terps quit (Ping timeout: 260 seconds)
20:17:35	PMunch	Zevv, you liked my approach?
20:19:32	*	dgb joined #nim
20:23:38	*	silvernode quit (Ping timeout: 272 seconds)
20:23:43	Zevv	yeah, altough there is still the nested iterator, but i liked it!
20:24:10	disruptek	using procs probably makes the most sense so far.
20:25:04	Zevv	not using for loops or injecting templates, but both should be just fine
20:25:17	*	qwertfisch quit (Read error: Connection reset by peer)
20:25:32	Zevv	anyway, doing the napnap now
20:25:35	Zevv	later!
20:25:38	disruptek	peace
20:25:42	Zevv	peace
20:26:52	*	qwertfisch joined #nim
20:27:26	*	narimiran quit (Ping timeout: 272 seconds)
20:44:58	*	gangstacat quit (Quit: Ĝis!)
20:47:42	*	gangstacat joined #nim
20:53:44	*	terps joined #nim
20:55:05	disruptek	skrylar[m]: what's your skflatbuffers license?
20:57:35	*	qwertfisch quit (Remote host closed the connection)
20:58:01	*	qwertfisch joined #nim
21:17:36	FromDiscord	<KrispPurg> Quick question: Can you jump a nimble pkg version to a much bigger version number like from v0.0.9 to v1.0.0?
21:18:14	FromDiscord	<Yardanico> Yeah sure why not
21:18:27	FromDiscord	<Yardanico> Just don't forget to tag a release
21:18:44	FromDiscord	<KrispPurg> what happens if you do?
21:19:02	FromDiscord	<Yardanico> If you don't have releases at all, it'll install latest commit
21:19:12	FromDiscord	<Yardanico> If you have some release tag, it'll install the tag
21:19:34	FromDiscord	<KrispPurg> ah
21:19:45	FromDiscord	<Yardanico> Nimble by default install latest commit if there are no tags, otherwise last release tag
21:20:31	FromDiscord	<KrispPurg> alright
21:24:22	*	PMunch quit (Quit: leaving)
21:31:31	*	tiorock joined #nim
21:31:31	*	tiorock quit (Changing host)
21:31:31	*	tiorock joined #nim
21:31:31	*	rockcavera is now known as Guest75063
21:31:31	*	Guest75063 quit (Killed (barjavel.freenode.net (Nickname regained by services)))
21:31:31	*	tiorock is now known as rockcavera
21:35:11	*	tiorock joined #nim
21:35:11	*	tiorock quit (Changing host)
21:35:11	*	tiorock joined #nim
21:35:11	*	rockcavera is now known as Guest69753
21:35:11	*	Guest69753 quit (Killed (orwell.freenode.net (Nickname regained by services)))
21:35:11	*	tiorock is now known as rockcavera
21:43:51	*	tane quit (Quit: Leaving)
21:45:05	FromDiscord	<Shucks> Is there something built in which could convert a byte array/seq to its numeric value? In my case the bytes represent an integer.
21:45:54	FromDiscord	<Rika> seq[byte] to seq[int]?
21:47:55	FromDiscord	<Shucks> seq[byte] to int
21:48:06	FromDiscord	<Shucks> with 4 stored bytes
21:49:41	*	exelotl quit (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.)
21:50:22	FromDiscord	<Elegant Beef> This is a silly way but it works :D↵https://play.nim-lang.org/#ix=2odG
21:53:06	FromDiscord	<Shucks> indeed oO
21:53:48	FromDiscord	<Shucks> was hoping there is some more "api" like way. Since I want to convert byte arrays into different types aswell.
21:53:53	*	terps quit (Ping timeout: 252 seconds)
21:54:30	FromDiscord	<Rika> maybe converting that seq[byte] to a string then using a StringStream or something
22:05:52	*	rockcavera quit (Ping timeout: 246 seconds)
22:06:09	*	rockcavera joined #nim
22:11:00	FromDiscord	<Elegant Beef> More silliness from me
22:11:00	FromDiscord	<Elegant Beef> https://play.nim-lang.org/#ix=2odK
22:11:06	FromDiscord	<Elegant Beef> Making silly code is fun
22:17:28	leorize[m]	Can someone with the latest macOS run this command and give me the output? `openssl s_client -showcerts -servername incomplete-chain.badssl.com -verify 10 -connect incomplete-chain.badssl.com:443 `
22:20:32	*	fredrikhr joined #nim
22:28:22	FromDiscord	<Shucks> > More silliness from me↵No clue how your magic works. But it does
22:51:35	*	brainproxy joined #nim
22:52:21	*	Trustable quit (Remote host closed the connection)
22:59:03	*	couven92 joined #nim
23:01:56	*	fredrikhr quit (Ping timeout: 258 seconds)
23:05:22	*	naught-fowl quit (Remote host closed the connection)
23:09:47	FromDiscord	<Elegant Beef> @Shucks it basically makes an int64 with the bit stream and then returns it, it will only work for less than 8 byte types
23:10:12	*	krux02_ joined #nim
23:10:44	FromDiscord	<Elegant Beef> it abuses the fact you can use bitwise on int64 and the cast
23:10:50	FromDiscord	<Elegant Beef> Cast converts the bit stream from A to B
23:11:01	FromDiscord	<Elegant Beef> Without any validational logic
23:12:47	*	krux02 quit (Ping timeout: 260 seconds)
23:13:22	*	tiorock joined #nim
23:13:22	*	rockcavera is now known as Guest60163
23:13:22	*	Guest60163 quit (Killed (tolkien.freenode.net (Nickname regained by services)))
23:13:23	*	tiorock is now known as rockcavera
23:14:04	*	tiorock joined #nim
23:14:04	*	rockcavera is now known as Guest33821
23:14:04	*	Guest33821 quit (Killed (card.freenode.net (Nickname regained by services)))
23:14:04	*	tiorock is now known as rockcavera
23:16:23	*	Senketsu quit (Ping timeout: 252 seconds)
23:17:29	*	beatmox quit (Read error: Connection reset by peer)
23:19:14	*	beatmox joined #nim
23:49:09	*	sagax joined #nim
23:51:46	*	beatmox quit (Quit: ZNC 1.8.0 - https://znc.in)
23:51:59	*	beatmox joined #nim
23:53:26	Prestige	Looks like we don't have bindings for xft, how would I got about creating it?
23:56:01	FromDiscord	<willyboar> What is xft?
23:56:30	Prestige	X FreeType lib: https://www.x.org/releases/X11R7.6/doc/man/man3/Xft.3.xhtml
23:56:42	Prestige	Wanting to use it to write a status bar for linux
23:57:06	FromDiscord	<Elegant Beef> You'd start making a wrapper 😛
23:57:32	FromDiscord	<Elegant Beef> Look at the importc and header pragmas