<< 08-03-2018 >>

---

00:00:04	FromGitter	<ZarsBranchkin> Oops yeah, nim-mode
00:01:17	FromGitter	<ZarsBranchkin> https://i.imgur.com/sqUF8Bq.png
00:01:30	FromGitter	<krux02> yes I just tested it
00:01:35	*	elrood quit (Quit: Leaving)
00:01:35	FromGitter	<krux02> yes that keyword is missing
00:01:37	Temperance[m]	how do I define a function pointer with void return?
00:01:51	FromGitter	<krux02> I found out I can actually use func to define functions
00:02:12	FromGitter	<krux02> either I never tried it or last time I did it didn't work
00:02:29	FromGitter	<ZarsBranchkin> I remember seeing it somewhere it in docs, apparently deprecated
00:02:32	FromGitter	<krux02> it's just emacs that has problems highlighting and indenting the line when I use func
00:02:49	FromGitter	<krux02> yea min should release a few keywords
00:03:47	FromGitter	<ZarsBranchkin> Oooh, it seems like some keyword for future use https://nim-lang.org/docs/manual.html#pragmas-nosideeffect-pragma
00:05:31	Temperance[m]	I think I understand, nim uses 2 spaces for indentation, doesn't it?
00:06:59	FromGitter	<ZarsBranchkin> I think that's the official convention, but should work with different amount of spaces too. Haven't tested myself
00:09:08	Temperance[m]	does it work with tabulations?
00:09:53	Temperance[m]	NICE
00:10:11	Temperance[m]	took an example with glfw and nuklear
00:10:15	Temperance[m]	and managed to wrap some initialization code inside a proc and expose it with*
00:10:15	Temperance[m]	I'm liking this stuff a lot.
00:10:23	FromGitter	<ZarsBranchkin> No, tabs aren't allowed, only spaces
00:10:34	FromGitter	<ZarsBranchkin> Yeah, I find it real fun interfacing with C functions
00:11:02	Temperance[m]	I'm having many problems understanding why this kind of mechanisms aren't there for plain C
00:11:34	Temperance[m]	>No, tabs aren't allowed, only spaces
00:11:48	Temperance[m]	is this something that can be configured?
00:11:55	Temperance[m]	or is it kinda, enforced
00:12:35	FromGitter	<ZarsBranchkin> I think it's enforced
00:12:39	FromGitter	<ZarsBranchkin> Yeah, here https://nim-lang.org/docs/nep1.html#introduction-spacing-and-whitespace-conventions
00:19:23	*	MJCaley joined #nim
00:25:28	*	dddddd joined #nim
00:32:59	*	xet7 quit (Quit: Leaving)
00:38:58	*	xet7 joined #nim
00:45:27	Temperance[m]	I'm having a little bit of a problem understanding what's happening here
00:45:28	Temperance[m]	type fptr = (proc(x: int):int)
00:47:22	Temperance[m]	does this declare a type called "fptr" that is a proc that receives a parameter x which is an int, and should return an int?
00:47:22	Temperance[m]	am I getting this correctly?
00:47:52	FromGitter	<ZarsBranchkin> Yeah, something like that. Here is more information on that https://nim-lang.org/docs/manual.html#types-procedural-type
00:49:59	*	geocar joined #nim
00:56:37	*	rbrt quit (Quit: Oíche mhaith)
00:58:08	*	rbrt joined #nim
00:58:38	*	rbrt quit (Client Quit)
01:00:42	*	vlad1777d joined #nim
01:05:01	Temperance[m]	well, insofar I'm kinda liking nim a lot
01:06:35	Temperance[m]	The fact I was able to import some package and have a window opened by glfw in one day seems like a whole success for a day
01:45:46	*	rockcavera quit (Remote host closed the connection)
01:47:19	*	rockcavera joined #nim
01:48:22	*	zarthur joined #nim
01:49:36	*	rockcavera quit (Remote host closed the connection)
01:51:44	*	arthurz quit (Ping timeout: 255 seconds)
01:56:51	*	rockcavera joined #nim
02:07:34	*	S1t1Schu joined #nim
02:11:05	*	S1tiSchu quit (Ping timeout: 240 seconds)
02:12:42	*	rockcavera quit (Ping timeout: 245 seconds)
02:19:33	*	rockcavera joined #nim
02:48:06	*	MJCaley quit (Quit: MJCaley)
02:59:31	*	geocar quit (Quit: Connection closed for inactivity)
03:18:45	*	Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
03:35:50	FromGitter	<zacharycarter> is there a way to create multiple modules for the VM in memory?
03:36:14	FromGitter	<zacharycarter> or should I just register everything in the globalCtx?
03:36:29	FromGitter	<zacharycarter> I was hoping I could make several modules and compose them into a modulegraph and use that when compiling my nimscript
03:36:43	FromGitter	<zacharycarter> but I'm not seeing how to do that clearly - I've created several modules, but don't see a way to make them depend on one another
03:46:50	*	SenasOzys quit (Ping timeout: 268 seconds)
04:09:07	*	triple_a joined #nim
04:10:17	*	triple_a left #nim (#nim)
04:19:13	FromGitter	<krux02> to be honest i dont know what you mean with modules in memory.
04:40:10	*	dbatbold joined #nim
05:03:38	*	yglukhov joined #nim
05:14:04	*	yglukhov quit (Remote host closed the connection)
05:27:55	*	sz0 joined #nim
05:35:02	*	yglukhov joined #nim
05:36:42	*	nsf joined #nim
05:39:13	*	yglukhov quit (Ping timeout: 240 seconds)
05:55:39	*	yglukhov joined #nim
06:00:27	*	yglukhov quit (Ping timeout: 265 seconds)
06:02:59	*	smt` joined #nim
06:06:43	*	smt quit (Ping timeout: 252 seconds)
06:29:14	*	craigger quit (Quit: bye)
06:31:08	*	craigger joined #nim
06:36:54	*	yglukhov joined #nim
06:41:30	*	yglukhov quit (Ping timeout: 256 seconds)
06:58:27	*	yglukhov joined #nim
07:18:28	*	Yardanico joined #nim
07:23:43	*	nohusuro joined #nim
07:27:02	*	dddddd quit (Remote host closed the connection)
07:33:57	*	zama quit (Ping timeout: 245 seconds)
07:37:03	*	couven92 joined #nim
07:40:34	*	zama joined #nim
07:46:26	*	rokups joined #nim
07:49:39	*	yglukhov quit (Remote host closed the connection)
07:51:57	*	yglukhov joined #nim
08:01:27	*	yglukhov quit (Remote host closed the connection)
08:06:57	*	sz0 quit (Quit: Connection closed for inactivity)
08:07:52	FromGitter	<narimiran> wait what? nim has REPL?! (cc @krux02)
08:10:08	*	yglukhov joined #nim
08:15:11	*	yglukhov quit (Remote host closed the connection)
08:18:42	*	yglukhov joined #nim
08:20:05	FromGitter	<andreaferretti> yes, just one that does not really work well with the C FFI
08:20:18	FromGitter	<andreaferretti> hence with a not small part of the stdlib
08:20:26	FromGitter	<andreaferretti> so it's hidden by default
08:26:34	*	PMunch joined #nim
08:36:46	*	sendell joined #nim
08:36:50	*	dbatbold quit (Ping timeout: 260 seconds)
08:49:08	FromGitter	<jaco60> Hi, i need to reverse a string. Is there already a `reverse` or `reversed` proc for strings in the std library (i’ve no found…) or conversions methods from string to seq allowing to use the ones from `algorithm`), or should i write my own proc (not a big deal, though) ?
08:49:43	*	Arrrr joined #nim
08:52:27	PMunch	jaco60, https://rosettacode.org/wiki/Reverse_a_string#Nim
08:52:50	FromGitter	<jaco60> Arf
08:53:26	PMunch	Seeing the complexity of the unicode variants I really feel like this should be put into a library. Possibly strutils
08:55:46	*	floppydh joined #nim
08:58:15	FromGitter	<jaco60> +1
08:58:52	euantor	you mean like `unicode.reversed`? https://nim-lang.org/docs/unicode.html#reversed,string
08:58:54	euantor	;)
08:59:37	PMunch	Haha, yeah I guess like that
09:00:12	FromGitter	<jaco60> Arf (bis repetita)
09:00:18	euantor	It is quite well hidden, and it doesn't seem like there's a version that doesn't handle UTF8
09:01:07	PMunch	Maybe rosettacode should be updated to just show this instead?
09:01:33	PMunch	euantor, well UTF-8 first is generally a good approach
09:01:55	PMunch	If you put another reverse in strutils everyone would just use that and mess up with unicode stuff
09:02:16	euantor	true. It would definitely make sense to update rosettacode. Do they accept contributions like that - I've never looked
09:02:26	PMunch	I think so
09:02:54	PMunch	And besides, writing a proper ASCII string reverser is a lot easier than writing a UTF-8 one
09:03:14	PMunch	So if you need for for just ASCII, then it's not too hard to write it
09:04:17	euantor	`algorithms.reversed` also works: https://nim-lang.org/docs/algorithm.html#reversed,openArray[T]
09:04:35	euantor	THough you get a `seq[char]` out of it
09:05:04	FromGitter	<jaco60> I have an account on Rosetta, i may add the unicode.reversed solution
09:05:23	euantor	awesome, that would be helpful :)
09:06:36	PMunch	Maybe we should go through some of these as well: http://rosettacode.org/wiki/Reports:Tasks_not_implemented_in_Nim
09:06:47	PMunch	Might be a good idea for a live-stream :)
09:07:17	FromGitter	<jaco60> i don’t understand why the existing solution import unicode and rewrite reversed
09:07:28	euantor	yeah, there are probably a lot that could be improved a bunch
09:08:34	PMunch	jaco60, if you look at the example it's the exact same as the one in the manual. I wouldn't be suprised if the same person who wrote that rosettacode snippet was also the one who added it to the unicode module
09:09:36	euantor	This one is super easy: http://rosettacode.org/wiki/DNS_query
09:09:51	euantor	https://nim-lang.org/docs/nativesockets.html#getHostByName,string
09:10:20	Araq	PMunch: that is exactly what happened
09:10:36	Araq	first it was on rosettacode and then it was added to the stdlib
09:11:42	PMunch	euantor, well it says both IPv4 and IPv6. Not sure if you can do that with getHostByName
09:11:51	euantor	it looks like it
09:11:57	euantor	https://github.com/nim-lang/Nim/blob/master/lib/pure/nativesockets.nim#L413
09:12:20	PMunch	That is either
09:12:31	PMunch	Depends on what you got from the lower procs
09:12:48	PMunch	You can't specify "I want this as IPv6"
09:12:52	euantor	Still might be worth having it there though
09:13:37	euantor	This is also an easy one: http://rosettacode.org/wiki/URL_parser
09:19:07	*	Vladar joined #nim
09:20:31	PMunch	Does Nim have a URL parser built in somewhere?
09:20:42	euantor	yep, in the `uri` module
09:20:48	PMunch	Nice
09:24:04	FromGitter	<jaco60> Added the unicode.reversed in Rosetta
09:25:05	FromGitter	<jaco60> I could be better to know the nim version where this proc was added
09:28:41	FromGitter	<jaco60> s/I/It
09:36:28	*	Arrrr quit (Ping timeout: 256 seconds)
09:38:45	Araq	probably exists since at least 0.16
09:39:19	FromGitter	<jaco60> Ok, i’ll add this
09:44:40	*	Arrrr joined #nim
09:44:41	*	Arrrr quit (Changing host)
09:44:41	*	Arrrr joined #nim
09:50:48	*	yglukhov quit (Remote host closed the connection)
10:00:28	PMunch	You could always check the git commit that introduced it as well jaco60
10:01:18	FromGitter	<jaco60> Wow… https://forum.nim-lang.org/t/3540 : that’s what we have to do to extract a readable IP Address from a Hostent ?
10:05:43	Araq	that code only deals with IP v4
10:06:16	PMunch	jaco60, Araq, apparently reversed in the unicode module was added for 0.11.0
10:11:14	*	yglukhov joined #nim
10:15:55	*	couven92 quit (Read error: Connection reset by peer)
10:16:47	PMunch	Hmm, is there a way to borrow all for a distinct type?
10:18:54	*	Arrrr quit (Quit: Leaving.)
10:19:53	Araq	no and if you need that, why do you have a distinct type?
10:35:05	*	smt` quit (Ping timeout: 256 seconds)
10:44:43	*	yglukhov quit (Remote host closed the connection)
10:46:09	*	smt joined #nim
10:50:10	PMunch	Overloading
10:50:41	PMunch	I want to have two types, which are both ints, behave exactly like ints, but serialize differently when written to a stream
10:51:29	PMunch	Currently I have them as distinct ints and have an overloaded write proc for each of them
10:51:52	PMunch	But that means you have to convert back and fro between them, which is annoying
10:56:50	PMunch	Hmm, with a X to seq converter it won't use high as it techincally wants an openarray, is this a bug?
10:57:21	*	yglukhov joined #nim
10:58:10	PMunch	Hmm, same for []=
10:58:16	PMunch	Well that's annoying
10:59:41	PMunch	Especially since I can't return an openArray from a converter..
11:03:00	*	Ahmedkh joined #nim
11:03:11	*	Ahmedkh left #nim ("ERC (IRC client for Emacs 25.3.1)")
11:05:31	Araq	"I want to have two types, which are both ints, behave exactly like ints, but serialize differently when written to a stream" seems confusing
11:08:42	PMunch	Well, that's how protobuf works.. You can have an int64, a sint64, a fixed64, or a sfixed64. All of those are 64bit integers, fixed64 is unsigned. The two fixed variants serialize as a regular 64bit int. The normal int64 serializes as a Varint, but always uses 10 bytes to store a negative number. sint64 uses zig-zag encoding which is more efficient for negative numbers.
11:09:13	PMunch	So I want to have a protoWrite procedure, which takes either of these and writes them as per the spec
11:10:10	PMunch	But internally in your program it doesn't really make sense to differentiate between them
11:12:53	dom96	hrm, couldn't you use a converter?
11:13:24	dom96	to implicitly convert back to the int but still able to write procs on the distinct type
11:16:54	Araq	you should use the new pragma annotations
11:17:08	Araq	field* {.sfixed64.}: int64
11:17:23	Araq	it's a property of the field, not of the type
11:17:49	FromGitter	<andreaferretti> +1
11:17:57	FromGitter	<andreaferretti> it is a perfect use case
11:19:03	Araq	dom96: did we move fsmonitor to a nimble package? https://forum.nim-lang.org/t/3617
11:20:07	dom96	apparently it's here? https://github.com/nim-lang/Nim/blob/devel/lib/packages/fsmonitor.nim
11:20:40	PMunch	dom96, yeah I have converters now. But that doesn't work for seq
11:20:45	PMunch	Not that that's a huge issue
11:21:06	PMunch	Hmm, how does that work Araq?
11:21:10	PMunch	The properties I mean
11:21:20	*	xkapastel quit (Quit: Connection closed for inactivity)
11:21:31	Araq	read the manual
11:21:56	*	yglukhov quit (Remote host closed the connection)
11:26:44	PMunch	Hmm, yes I remember vagualy someone talked about this when it was introduced
11:27:08	*	regtools quit (Quit: leaving)
11:27:39	FromGitter	<jaco60> Any advice before publish it in Rosetta ? (i don’t know if it’s idiomatic nim as i’m a beginner) https://gist.github.com/jaco60/bb2fbc4d79b2f64adaa6f8a0d9d7a679
11:27:54	PMunch	So convert my writer procs to a single macro, then annotate the fields and have the macro read the custom pragma to decide serialization?
11:29:37	PMunch	jaco60, looks good to me. Could use an array instead of a seq since you're not changing the length but that's a minor thing
11:35:41	dom96	jaco60: why not just echo(url)?
11:36:46	FromGitter	<jaco60> Because i wanted to follow other solutions, like Go
11:37:15	FromGitter	<jaco60> echo url would be rather « raw », compared to the other solutions
11:37:32	dom96	You'll get the same output, but with much less code
11:39:04	FromGitter	<jaco60> i don’t get the same output at all
11:39:08	dom96	well, similar output
11:39:28	FromGitter	<jaco60> No… I just get the original url
11:39:39	dom96	echo(parseUri(url))
11:39:42	dom96	that's what I meant
11:39:51	*	yglukhov joined #nim
11:40:08	FromGitter	<jaco60> yes, but no… Try `echo res`, you’ll see
11:40:13	*	yglukhov quit (Remote host closed the connection)
11:41:11	dom96	oh right
11:41:17	dom96	echo system.`$`(parseUri(url))
11:41:20	PMunch	echo res just outputs the same URI that you passed in
11:43:51	PMunch	Hmm, same info. But the formatting isn't as nice
11:44:24	FromGitter	<ZarsBranchkin> Uhm, why is the std/ part nescessary? I can't import the modules when it's there
11:44:41	FromGitter	<ZarsBranchkin> Error: cannot open 'std/uri'
11:45:10	dom96	Are you using Nim 0.17.2?
11:45:15	PMunch	jaco60, speaking of formatting: https://gist.github.com/3352f266c96db3cef74627feedf7cba6
11:45:25	FromGitter	<ZarsBranchkin> Ah, probably still have 0.17 on this computer
11:45:43	dom96	but actually it's more idiomatic not to use that
11:45:52	dom96	'std/' is only necessary for disambiguation
11:46:25	FromGitter	<ZarsBranchkin> Also backwards compatible if that is left out, so probably should remove that
11:47:28	dom96	yeah, and that's another reason to just write `echo system.`$`(parseUri(url))` ;)
11:48:03	*	regtools joined #nim
11:49:20	FromGitter	<ZarsBranchkin> System is necessary because uri overrides default $?
11:49:48	dom96	yep
11:50:11	dom96	system has a generic `$` for all objects
11:51:38	Araq	well the idioms still evolve
11:51:46	Araq	I like 'std'
11:52:22	Araq	speaking of which, can we do without 'pkg' please :P
11:53:06	Araq	ah let me read the RFC again, forgot the problems already
11:53:17	dom96	no, we can't
11:54:09	*	SenasOzys joined #nim
11:58:07	Araq	import random # uses std lib random
11:58:07	Araq	import random/random # uses nimble's random
11:58:42	Araq	still is the best solution. fix the nimble packages that don't play nice
11:59:01	PMunch	That looks super confusing :P
11:59:25	Araq	package management is all about enforcing a style/directory layout anyway
11:59:38	FromGitter	<jaco60> Posted...
12:01:09	*	yglukhov joined #nim
12:02:30	*	Snircle joined #nim
12:05:21	*	yglukhov quit (Ping timeout: 248 seconds)
12:08:26	Araq	what if we make 'pkg' mean "not in std" but make it use the --path/nimblepath ?
12:08:41	Araq	then it would just work with today's nimble, right?
12:16:48	GitDisc	<mraxilus> hey can anyone here provide some help with reassigning pointers in nim
12:17:35	GitDisc	<mraxilus> I'm trying to implement a stack to learn about pointer/memory mangement in nim -> https://pastebin.com/zk4BtZtd
12:18:14	GitDisc	<mraxilus> and on line 43 I get an error about not being able to assign a new address to one of my ptrs
12:21:49	*	yglukhov joined #nim
12:22:59	*	vlad1777d quit (Ping timeout: 268 seconds)
12:25:33	PMunch	mraxilus, try changing your input type to a "var Stack[T]
12:25:49	GitDisc	<mraxilus> what's the difference?
12:25:57	*	yglukhov quit (Ping timeout: 240 seconds)
12:26:14	PMunch	Normally you can't modify your input parameters in Nim
12:26:24	FromGitter	<ZarsBranchkin> Oh true, that solves it. It tells compiler that argument may get modified by the procedure
12:26:37	PMunch	So you must specify that your input parameter is var so it can be modified
12:26:42	PMunch	Correct
12:26:47	GitDisc	<mraxilus> Ah, I understand now, thanks
12:27:16	PMunch	That way you won't be able to pass it e.g. a stack that's declared with let (not that it makes much sense in this case)
12:42:10	FromGitter	<dom96> mraxilus: PMunch: A better idea would be to use a 'ref' type in that case
12:42:19	FromGitter	<dom96> You won't have to write 'var Stack' everywhere then
12:45:04	Araq	that's not better, just different
12:45:19	FromGitter	<andreaferretti> I prefer var stack
12:45:26	FromGitter	<andreaferretti> Stack stays a value type
12:45:37	FromGitter	<andreaferretti> and functions that mutate the stack say so in their signature
12:46:07	GitDisc	<mraxilus> yeah, I'm usually a fan of explicitly stating where mutations happen
12:46:22	GitDisc	<mraxilus> think I'll stick with ptr
12:46:56	Yardanico	why ptr?
12:46:57	*	floppydh quit (Remote host closed the connection)
12:47:25	PMunch	Isn't that the opposite of what you just said?
12:47:31	PMunch	Or did you mean var instead of ptr?
12:47:45	*	floppydh joined #nim
12:47:50	GitDisc	<mraxilus> var yeah my bad
12:49:43	PMunch	dom96, I tried to compile a local version of Nim after I made some changes. But to avoid having to build from C sources I just used the nim version I had from choosenim, but I get this error: http://ix.io/U34/
12:49:46	PMunch	Is that normal?
12:50:50	Yardanico	PMunch, https://github.com/dom96/choosenim/issues/45
12:52:45	PMunch	Aha
12:52:50	PMunch	What even is nim0?
12:53:21	Yardanico	probably executable koch creates while bootstrapping
12:53:25	Yardanico	I don't really know
12:53:42	Araq	nim0 is the 0th step in the bootstrapping process
12:54:01	Araq	which is just a copy of nim iirc
12:55:23	PMunch	Hmm, building from csources didn't get me much farther either..
12:55:25	PMunch	http://ix.io/U3c/
12:55:46	Yardanico	PMunch, ugh, are these csources 0.18.0?
12:55:54	PMunch	Yup
12:56:01	PMunch	According to bin/nim --version at least
12:57:28	PMunch	Oh wait
13:03:03	*	yglukhov joined #nim
13:07:38	*	yglukhov quit (Ping timeout: 276 seconds)
13:23:42	*	yglukhov joined #nim
13:25:37	FromGitter	<narimiran> @dom96 "A better idea would be to use a 'ref' type in that case" - did i understand this correctly: if you have 'ref object' there's no need to have 'var ' parameter in your proc, you can still modify it?
13:26:50	FromGitter	<narimiran> i thought using 'var' is a must if you wanted to modify something, didn't know about this other possibility (if i didn't misunderstand what you just said). any pros/cons of each method?
13:28:21	*	yglukhov quit (Ping timeout: 268 seconds)
13:29:13	euantor	yeah, you don't need `var` for a `ref` object
13:31:23	PMunch	What you are passing in with a ref object is not the object, it's the ref. So you're not allowed to change the ref, but you can change what it references (that is just treated as memory). If you used var with a ref object that would imply that you changed the reference, similar to double pointers in C.
13:32:06	Yardanico	ptr var ref object :D
13:32:22	Yardanico	this is valid 0_0
13:32:38	PMunch	Hmm, what would that be
13:32:47	PMunch	I guess var implicitly makes a ref
13:32:55	PMunch	So it would be a pointer to a ref to a ref object?
13:33:23	FromGitter	<narimiran> thanks guys for the clarification!
13:37:07	FromGitter	<narimiran> btw, yesterday i had a situation where i had a seq consisting of arrays. and i want to send one array to another proc without copying it (`let a = seq[index]` would have copied it, from what i understand). i did is this way, tell me if there is a better way:
13:40:11	FromGitter	<narimiran> `let a = addr seq[index]`, and then in that proc, i had a parameter which said: `p: ptr array[length, type]`
13:44:21	*	yglukhov joined #nim
13:46:30	PMunch	narimiran, you should probably use var in that case as well: https://play.nim-lang.org/?gist=8dd766c0da08dbb5115dba9426cc3a90
13:46:51	PMunch	That means you are not copying the data you are sending in
13:48:32	*	yglukhov quit (Ping timeout: 245 seconds)
13:49:05	*	SenasOzys quit (Ping timeout: 240 seconds)
13:49:10	FromGitter	<narimiran> @PMunch: here's that part of the code (it won't work, just to show what i'm doing): https://play.nim-lang.org/?gist=ceed5afa714abc90ad7bee280ed42680
13:50:55	FromGitter	<narimiran> i thought if on line 8 i don't use `addr` that would copy Ts[t], no?
13:51:18	PMunch	Not if T is marked as var as well in the 'f' proc
13:55:13	FromGitter	<narimiran> @PMunch: here's your example modified: https://play.nim-lang.org/?gist=43d821c97ef03c2afd6343dc45e7b53b to include new `var z = x[1]` - the values in `x` haven't changed! (in contrast to when you have called `test(x[1])`)
13:56:07	PMunch	Ah yes, the "var z = x[1]" copies the array I think
13:56:43	PMunch	So in your case addr is the way to go
13:56:48	*	SenasOzys joined #nim
13:57:55	FromGitter	<narimiran> ok, so here is the wanted version: https://play.nim-lang.org/?gist=c2c11da40fc0c032dfa11dcd922910b7 using `addr` and `ptr`
13:58:44	PMunch	Yes
13:58:59	PMunch	Note that you could also have z declared using a let statement here
13:59:14	PMunch	As it is the reference, not the data
13:59:16	FromGitter	<narimiran> indeed!
14:00:15	FromGitter	<narimiran> ok, this was the first time i have used `addr` and `ptr` so i wanted to check if that's the way to go, or there is some better/idiomatic way to achieve the same thing.
14:00:38	PMunch	For performance reasons (which I guess is what you are going for here) you might also want to throw on a {.inline.} pragma
14:00:46	PMunch	Or even just declare the whole thing as a template
14:04:49	FromGitter	<narimiran> you have guessed correctly about performance :) {.inline.} goes in the definition of my `f` proc, right? what would be advantages of using template here?
14:04:57	*	yglukhov joined #nim
14:05:48	PMunch	Yes
14:06:05	PMunch	Well, maybe not that great of a benefit
14:06:20	*	vlad1777d joined #nim
14:06:24	PMunch	template and {.inline.} would probably have similar performance benefits
14:07:09	PMunch	{.inline.} is just a flag to the compiler to make it inline the body of the procedure. templates does more or less the same but in Nim
14:09:29	*	yglukhov quit (Ping timeout: 260 seconds)
14:11:08	FromGitter	<narimiran> just tried both versions (inline and template) and there is no noticeable difference in the end regarding running times
14:12:21	PMunch	Did you notice a speedup between not using them?
14:12:26	Araq	template currently has the edge because the compiler's handling of .inline is rather stupid
14:12:48	Araq	the compiler delegates .inline to the C compiler
14:13:19	Araq	template performs the inlining in the frontend, this can elide more copies, for example
14:15:10	PMunch	Yeah, and .inline. is just a hint
14:15:18	PMunch	The compiler might ignore it completely
14:16:33	FromGitter	<narimiran> PMunch: haven't noticed a significant difference (it is smaller than difference between two consecutive runs)
14:17:34	Araq	PMunch: that's usually not a problem.
14:17:44	FromGitter	<narimiran> Araq: so you would recommend to use templates in cases like this one? its performance will be at least as good as proc, and it might be better?
14:18:05	Araq	no, our guidelines say to prefer inline procs
14:18:32	Araq	as they are easier to reason about (no control flow effects)
14:19:29	Araq	that they are currently slower in some circumstance is just an unfortunate implementation detail. you can always optimize and use 'template' later
14:23:31	FromGitter	<narimiran> i have run every version five times, and these are average times:
14:23:52	FromGitter	<narimiran> proc: 3.65 sec; inline proc: 3.65 sec; template: 3.80 sec
14:25:34	*	yglukhov joined #nim
14:29:17	FromGitter	<narimiran> and all these are very good times for such a huge domain i'm calculating. (for comparison, julia runs this in 4.74 seconds, python+numpy+numba in ~13 sec)
14:29:35	*	yglukhov quit (Ping timeout: 240 seconds)
14:30:01	FromGitter	<jaco60> Well, iI give up trying to figure out how to use `nativesockets.getHostByName` to get IP addresses. It’s beyond my skills :(. If someone is able to explain me how to get *readable* strings from the `Hostent.addrList` field, i’ll be grateful.
14:33:27	*	smt quit (Ping timeout: 252 seconds)
14:35:44	dom96	http://sogrady-media.redmonk.com/sogrady/files/2018/03/lang.rank_.118.png
14:35:54	dom96	Time to create more projects guys!
14:38:35	dom96	(From https://redmonk.com/sogrady/2018/03/07/language-rankings-1-18/)
14:39:03	federico3	we need better dissemination
14:39:39	zarthur	Indeed @dom96, and my 2 cents, if #nim could get as good Webframework as Rails it could take it much higher
14:40:04	*	PMunch shudders while thinking about Ruby on Rails
14:41:21	dom96	zarthur: Perhaps you could that web framework :)
14:41:24	dom96	*could write
14:42:01	federico3	zarthur: the web "development" world is already really saturated with languages and frameworks
14:42:12	zarthur	@PMunch it is all about how you want to position #nim. Ruby is high because of Rails, C# is for all/everything summed up, R for the wide adoption by para-programmers in application to data science, CSS, well not a language per-se but drives quite a bit of websites
14:43:25	zarthur	yeah writing a successful Web Framework is not a trivial undertaking, but today the World needs better web experience. Ruby is slow and illogical, rip it apart with #nim :-)
14:45:18	*	douglascorrea joined #nim
14:45:28	zarthur	Well, @pmunch, with all the fairness, but when a Web Dev (not me) reading "Nim in Action" needs Nginx reverse proxy mapped to Jester 0.0.1 to power a website one sets oneself on fire
14:45:52	federico3	zarthur: want to write detailed requirement on https://github.com/nim-lang/needed-libraries/issues ?
14:47:05	federico3	zarthur: what's the issue? 0.0.1?
14:47:40	PMunch	zarthur, well that's fair. But Jester really isn't that hard to set up and it's super easy to use. At least for simple sites which is all I have really used it for
14:47:46	FromGitter	<andreaferretti> rosencrantz is versio 0.3.3 - clearly more advanced :-P
14:48:37	dom96	https://www.reddit.com/r/programming/comments/82wpiw/the_redmonk_programming_language_rankings_january/dvdks2l/
14:48:54	zarthur	Exactly, aim, high @pmunch, nim needs a pro web framework with ORM (database objects mapper)
14:49:39	dom96	zarthur: yeah, strange how most of the web runs on PHP behind nginx, apache, lightttpd huh?
14:49:42	federico3	zarthur: again, want to write detailed requirement on https://github.com/nim-lang/needed-libraries/issues ?
14:50:17	dom96	I bet most web frameworks recommend you run them behind nginx or another http server
14:50:21	zarthur	@dom96 PHP has good web frameworks
14:50:36	PMunch	zarthur, yeah I was using MySQL for my simple site. But now we have Ormin: https://github.com/Araq/ormin
14:50:57	zarthur	All the web servers are Web Framework agnostic
14:52:58	FromGitter	<krux02> zarthur: then built a web framework.
14:53:07	FromGitter	<andreaferretti> zarthur: not exactly true
14:53:20	FromGitter	<andreaferretti> many web frameworks are meant to be exposed directly
14:53:29	FromGitter	<andreaferretti> without a reverse proxy
14:53:39	FromGitter	<andreaferretti> hence they are frameworks and servers in one
14:53:47	federico3	that's usually not a good idea
14:54:38	zarthur	Does ormin natively integrate into Jester?
14:54:44	livcd	zarthur: Ruby is loosing its momentum because its "usefulness/popularity" is driven mainly by an opinionated web framework.
14:55:20	zarthur	@andreaferretti exactly my point, but in the book... ^
14:55:43	federico3	yep - it's better not to force ORMs and other stuff on users
14:56:08	PMunch	zarthur, what do you mean by natively integrate into Jester?
14:56:23	zarthur	@livcd, this is why I said above rip it apart with #nim
14:56:53	dom96	Does the book say you "need" it?
14:56:58	zarthur	@federico3 that's what I am conveying in reference to the book
14:58:15	PMunch	Hmm, I should really do that documentation server. If for nothing else than to show how to combine Jester+Ormin+Karax to create a simple website
14:58:22	federico3	I'd rather have a tool to create a skeleton for web apps and pull in only the libraries you really need and generate the glue code
14:58:42	livcd	federico3: like with js projects and npm ? :D
14:58:45	PMunch	Or Jester+Ormin+<Some templating engine (ideas?)>
14:58:49	federico3	[spam] PMunch: and httpauth
14:59:03	federico3	livcd: no, like turbogears was doing
14:59:23	PMunch	Oooh nice federico3, didn't know about that
15:00:23	federico3	PMunch: Nim already has templating albeit it does not integrate well in jester
15:00:32	Temperance[m]	I suppose that given that nim compiles to C, using nim with emscripten must not be too hard, right?
15:00:49	FromGitter	<andreaferretti> for templates there is https://nim-lang.org/docs/filters.html
15:01:04	FromGitter	<andreaferretti> I cannot see why it would not integrate nicely in jester
15:01:09	FromGitter	<andreaferretti> or anything else really
15:01:15	FromGitter	<andreaferretti> they are just functions
15:01:25	federico3	andreaferretti: that's why
15:01:57	PMunch	Oh yeah, those strange things. Haven't really tried them
15:02:00	PMunch	But I like the idea
15:02:08	dom96	IMO those filters just need a nicer syntax
15:02:17	dom96	and a way to inline them into procedures
15:02:37	Yardanico	Temperance[m], it's already possible
15:03:04	federico3	andreaferretti: many frameworks discover automatically the template location and "apply" them to the routes you want
15:03:06	Yardanico	Temperance[m], for example see https://github.com/Jipok/Nim-SDL2-and-Emscripten
15:03:10	dom96	Temperance[m]: It's already used in production
15:03:21	Yardanico	dom96, really?
15:04:03	dom96	Reel Valley uses emscripten IIRC
15:04:20	Yardanico	dom96, I think it doesn't, hmm
15:04:44	Yardanico	dom96, ah, yes, it does
15:05:00	PMunch	federico3, I guess one could do something like that with a macro
15:05:06	FromGitter	<andreaferretti> @federico3 I don't like autodiscovery, I think that explicitly using the functions you want is preferable
15:05:15	PMunch	Search through a folder and include whatever is in there and create routes for it
15:05:17	Yardanico	Temperance[m], check out Reel Valley on facebook (it's a game), it uses nim+emscripten
15:05:26	Yardanico	it's also available on mobile platforms, but it uses native code there
15:05:50	federico3	PMunch: the routing has to be explicit because you might want to apply the same template to many routes
15:05:58	*	r3d9u11 joined #nim
15:07:06	livcd	so then an opinionated game framework in Nim ? :-X modeled after Reel Valley ?
15:07:23	Yardanico	livcd, Reel valley uses NimX
15:07:34	Yardanico	for gui :)
15:07:45	livcd	oh
15:08:05	Yardanico	livcd, a lot of parts of Reel Valley are open-sources
15:08:07	*	sendell quit (Remote host closed the connection)
15:08:07	Yardanico	open-sourced*
15:08:13	PMunch	federico3, wait. What did you want it to do then?
15:08:34	Yardanico	livcd, check "Technical Insights of the Game" here - https://yglukhov.github.io/Making-ReelValley-Overview/
15:09:01	Yardanico	livcd, and yeah, their game engine is rod - https://github.com/yglukhov/rod
15:10:23	zarthur	@livcd, I am not a proponent of Ruby on Rails, but "loosing momentum" feels like fake news, and it does not mean much sorry if assaulting but how much momentum, it is very high on the chart
15:11:08	zarthur	in m view the Ruby on Rails community just wanted to whine a bit to entice the maintainers to keep going delivering good stuff
15:11:50	*	yglukhov joined #nim
15:12:11	FromGitter	<Vindaar> So I recompiled a small program, which just extracts some time information from some data on the current 0.18.1 devel branch. Previously compiled on a few months ago on the then devel branch. I noticed that the duration I output suddenly shows 0 days, 0 hours and 0 minutes. Well, code boils down to this example from the docs: ⏎ https://play.nim-lang.org/?gist=ebc0c0a8f45eb2a31e61ce54262afb9a ⏎ I know there were
15:12:11	FromGitter	... quite a few changes to the times module, but is that expected behavior now? Meaning initInterval only adds the given field (in this case seconds) to the TimeInterval instead of correctly populating the different fields.
15:12:43	federico3	PMunch: you "apply" a template to a route with a simple syntax and the framework will scan ./templates for it and give a meaningful error if there's any issue. Currently templates are imported globally and you have to type the tpl name in 3 places
15:12:55	FromGitter	<Vindaar> Looking at the implementation of initInterval shows that indeed, the div and carrying to populate all fields is gone
15:15:30	PMunch	federico3, I would love to see a mock up of how this would work in Jester. Not runnable of course
15:16:34	*	yglukhov quit (Ping timeout: 260 seconds)
15:18:32	dom96	Vindaar: yeah, there was quite a few changes to this. Do open an issue if the behaviour is odd now.
15:18:46	federico3	PMunch: will do
15:18:57	FromGitter	<Vindaar> @dom96 ok, will do
15:33:21	*	rinzai joined #nim
15:38:06	*	yglukhov joined #nim
15:40:24	*	douglascorrea quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
15:41:13	*	dddddd joined #nim
15:43:07	FromGitter	<Vindaar> done: https://github.com/nim-lang/Nim/issues/7315
15:44:35	*	miran joined #nim
15:47:34	Temperance[m]	if I end up liking nim I might try to do a toy game engine for fun, specially since I'll want to be able to test glsl and stuff with it.
16:03:26	*	floppydh quit (Quit: WeeChat 2.0.1)
16:05:29	*	PMunch quit (Quit: Leaving)
16:06:12	*	douglascorrea joined #nim
16:09:55	FromGitter	<krux02> :D I have there something for you
16:10:11	FromGitter	<krux02> but only if you are not on a mac
16:10:37	FromGitter	<krux02> @Temperance githib.com/krux02/fancygl
16:11:10	FromGitter	<krux02> but if you try it out, tell me about it, it's not in a final state it is more a research project
16:11:32	miran	krux02: you have (at least) one typo in your link ;)
16:11:43	FromGitter	<krux02> I actually write a master thesis about it and it will have eventually shaders that are written in it
16:11:46	FromGitter	<krux02> ah sorry
16:11:53	FromGitter	<krux02> github.com/krux02/opengl-sandbox
16:12:23	*	noonien joined #nim
16:12:27	noonien	hello folks!
16:13:07	miran	hello noonien!
16:13:08	FromGitter	<Vindaar> hey!
16:15:45	FromGitter	<krux02> @Temperance: the video is also linked on the github page, but it is a good introduction: https://www.youtube.com/watch?v=JO0iqGDgFqA
16:16:08	noonien	i'm trying to see if i can compile a simple blinky program using nim, for a cortex m0 mcu. from what i understand, this should be as easy as compiling with `nim c --cpu:arm --gc:none --os:standalone`, correct?
16:20:11	*	r3d9u11 quit (Ping timeout: 256 seconds)
16:28:00	Temperance[m]	can't open fancy gl github
16:28:25	miran	Temperance[m]: see the other link few messages below
16:28:27	dom96	noonien: hrm, maybe federico3 can help out here
16:28:38	noonien	cool
16:29:03	noonien	is there any way of passing default build flags using nimble?
16:29:24	noonien	or should i just create build-debug and build-release tasks in the .nimble file?
16:29:34	Temperance[m]	will look into it tonight
16:30:15	dom96	noonien: create a .nim.cfg file and add your flags there
16:30:30	dom96	name it module.nim.cfg
16:30:45	dom96	or just nim.cfg to apply to your whole project IIRC
16:30:59	noonien	i've got a nimble package created with "nim init"
16:31:10	noonien	should i put that file in the src folder?
16:31:25	dom96	yeah
16:31:31	noonien	cool
16:50:10	noonien	does nim have a runtime?
16:52:10	noonien	can i compile nim without the stdlib, or, can i choose which functions from the stdlib i use, and not compile the ones i don't use?
16:55:40	FromGitter	<zacharycarter> is it possible to compile multiple modules for the VM at runtime?
16:55:44	FromGitter	<zacharycarter> using a modulegraph or something?
16:55:59	FromGitter	<zacharycarter> or should I just register all my vm callbacks in the main context?
17:01:02	Araq	hmm, not sure
17:01:47	Araq	noonien: with --deadCodeElim:on only what you use ends up in the binary
17:03:06	noonien	i see, i'm already using that, i'm trying to create an empty program and am getting: `/build/newlib-OTNHfJ/newlib-2.4.0.20160527/build/arm-none-eabi/newlib/libc/stdlib/../../../../../newlib/libc/stdlib/exit.c:70: undefined reference to `_exit'`
17:03:14	noonien	i guess exit is a special requirement
17:05:26	Araq	exit.c is not produced by Nim
17:07:09	FromGitter	<zacharycarter> Araq: any idea who might know more about the VM stuff? I've written a macro which stores callbacks I want to register with a VM module context in a table. At runtime I register these callbacks with the VM context - but it'd be neat if I could build a module dependency hierarchy and separate all my callbacks into separate modules, instead of just dumping them all into the main module.
17:07:14	Yardanico	noonien, arm-none-eabi means what you're compiling for barebones?
17:07:28	Yardanico	I mean not for an OS?
17:07:52	noonien	Yardanico: yes
17:08:05	Yardanico	well then you can't use some modules from stdlib
17:08:12	Yardanico	and GC
17:08:14	noonien	that's not a problem
17:08:34	Yardanico	(well you can port GC, but I don't know how :D)
17:11:45	FromGitter	<tim-st> Not directly nim related, but mayber someone knows if there's a better way to use a sequence of uint as strKey than like `0|1|2|4|1`?
17:12:04	FromGitter	<tim-st> (enums)
17:15:41	Araq	https://hal.inria.fr/hal-01724997/document
17:15:53	Araq	zacharycarter: I think you're doing it wrong ;-)
17:16:08	Araq	check out how putEnv was added to Nim
17:17:36	FromGitter	<zacharycarter> okay thanks I will
17:19:54	FromGitter	<zacharycarter> ah I think I get it now :) thanks Araq!
17:20:36	*	xkapastel joined #nim
17:27:46	noonien	Yardanico: I'm trying to compile an empty program
17:27:54	noonien	just an empty main function
17:28:28	Yardanico	noonien, you still need to disable gc, maybe it will help
17:28:29	Yardanico	but IDK
17:29:36	*	MJCaley joined #nim
17:31:07	*	rbrt joined #nim
17:39:42	dom96	noonien: looking at nimkernel might help out too: https://github.com/dom96/nimkernel
17:40:28	noonien	awesome, i'll give it a look, thanks!
17:42:57	*	yglukhov quit (Remote host closed the connection)
17:46:05	federico3	noonien: an example of nim.cfg for avr in query
17:46:31	noonien	just a second, i'll put up a gist of what i've got so far
17:54:08	noonien	this is what i've got so far: https://github.com/noonien/nrf51288_blinky_nim
17:54:16	noonien	and the error i'm getting: http://vpaste.net/FlOf9
17:55:04	noonien	i don't know how hard it'll be to initialize the hardware in nim, as a temporary workaround i was thinking of compiling the nim code as a library and initializing the hardware in C code
17:55:31	noonien	the problem with nimkernel is that it calls the linker manually
17:55:33	*	Chiel joined #nim
17:55:48	noonien	and one of the reasons i'm trying to move to nim is so i can simplify the build process
17:56:53	Chiel	Hello Nim people. We are considering using Nim on an embedded linux system with pretty tight memory constraints. Would it be possbile to compile (part of) the Nim standard libs as shared libraries to reduce the size of the individual binaries?
17:57:32	Chiel	I can only get this to work using the C api, but I'm not sure how to call pure nim code from an Nim lib compiled with --app:lib
17:58:01	*	SenasOzys quit (Remote host closed the connection)
17:58:18	*	SenasOzys joined #nim
18:03:04	federico3	Chiel: like in https://forum.nim-lang.org/t/2051 ? --app:lib gives you a C api
18:04:07	federico3	(besides: Nim includes only the the needed code in a binary, not the whole stdlib: are you sure this optimization would be useful?(
18:04:13	federico3	))
18:04:31	Chiel	Probably, I have a number of tools all doing asynchronous networking, which adds up
18:04:39	Chiel	Just considering the options
18:05:45	federico3	can you share some numbers?
18:07:47	Chiel	I have a single binary which is now approx 500K, of which only 40K is main application symbols, the rest all comes out the stdlib
18:08:22	Chiel	Adding one and one, I suppose I should be able to have two of such programs runninng at a total of (40+40+460) = 540K
18:08:46	Chiel	The other option would be to make one single binary and fork() early
18:09:24	Chiel	or one binary which multiple apps built in, and simply choose an other path to run at startup
18:09:33	Chiel	basically, just linking it all into a single blob
18:11:20	federico3	like https://en.wikipedia.org/wiki/BusyBox
18:11:25	Chiel	just like that
18:12:03	Chiel	busybox originally solved it like that to be able to run small on uclinux, which has no MMU, and no shared libs
18:12:29	federico3	do you also want to do execute-in-place?
18:12:43	Chiel	No, we run from cramfs, so that is not an option at this time
18:13:07	Chiel	I have approx 4mb of flash, but loads of RAM available
18:19:08	federico3	Chiel: assuming you are using -d:release --opt:size --deadCodeElim=on , have you tried compressing the binaries?
18:19:28	Chiel	Well, in the end they are compressed by the cramfs, which does a lot
18:19:37	federico3	(i mean apart from cramfs)
18:20:35	Chiel	No, not yet. That does not really help in the end because it will end up in the cramfs anyway
18:20:53	federico3	cramfs uses zlib - perhaps xz does better
18:20:59	Chiel	Cramfs does as well
18:21:26	Chiel	squashfs that is
18:21:27	Chiel	I'm sorry
18:22:15	Chiel	anyway, I'm still fine for now, just looking for options
18:24:09	noonien	Oops, my mistake, nim was not at fault, i just had to pass --specs=nosys-specs to the linker
18:24:19	noonien	According to: https://stackoverflow.com/questions/19419782/exit-c-text0x18-undefined-reference-to-exit-when-using-arm-none-eabi-gcc
18:24:49	noonien	And https://launchpadlibrarian.net/170926122/readme.txt
18:25:00	noonien	Perhaps nimc should do that by itself?
18:25:18	Araq	noonien: there is a subset of Nim that maps to C 1 to 1, there is nothing you need to use C for
18:25:53	Araq	Chiel: the library can be built as a DLL but it's usually not done because it's inconvenient
18:26:22	Chiel	How does one import and export from a shared lib with mangling intact?
18:27:25	Araq	by telling the compiler how to mangle
18:31:20	Araq	proc foo*(x, y: int) {.exportc: "prefix_$1_intint".}
18:32:24	Chiel	Ah, that makes sense, thank you
18:36:25	Araq	noonien: Nim can't guess your configuration. the embedded stuff is full of special cases and setups
18:37:10	noonien	That makes sense, i just figured arm-none might be a common setup in the embedded world
18:38:11	dom96	it might make sense if this option works for all arm-none projects
18:38:23	dom96	and if Nim can easily know that that is what you're using
18:40:09	*	Trustable joined #nim
18:41:48	*	jxy quit (Quit: leaving)
18:45:50	*	jxy joined #nim
18:59:23	FromGitter	<samdmarshall> heya, I had a quick question while trying to work on a small project. I have some sequences that i'm filtering down. the resulting sequences should have a single elements, how would I go about pulling an element from the sequence without hard-coding it directly to a specific index? I thought there was an `any` or similar operator that would return an element (doesn't necessarily have to be the first element) from
18:59:23	FromGitter	... the sequence, but I cannot seem to find anything when I checked the docs
19:01:50	FromGitter	<Vindaar> Not quite sure if that's helping you, but you're aware of the sequtils module, right? https://nim-lang.org/docs/sequtils.html
19:02:56	Araq	I think we only have https://nim-lang.org/docs/sequtils.html#any,openArray%5BT%5D,proc%28T%29 and siblings returning 'bool' not the element that matched
19:03:04	FromGitter	<samdmarshall> yeah I looked at that but didn't see anything that would return a singular element from the sequence for me
19:04:47	FromGitter	<tim-st> Is there a way to ensure my enum type has every ord in range 0..255 ?
19:05:24	FromGitter	<samdmarshall> normally i'd be fine with just using an iterator for such results, but i'm doing this in a `cmp` method so I need to unbox the element from the seq
19:09:17	*	athenot_ quit (Remote host closed the connection)
19:09:28	FromGitter	<cabhishek> Hi folks, I am getting `ProtocolError: Connection was closed before full request has been made` error when trying to load test my server. Any ideas how I should I handle this run time error? I am using `AsyncHttpClient`.
19:09:58	*	athenot joined #nim
19:16:59	*	SenasOzys quit (Ping timeout: 256 seconds)
19:21:21	FromGitter	<nitely> @samdmarshall you are lookig for a `reduce` function
19:22:48	FromGitter	<nitely> I does not seem Nim has one, at least not one with that name in the stdlib
19:23:12	FromGitter	<jaco60> there is a foldr proc, right?
19:24:37	FromGitter	<ZarsBranchkin> Yeah, seems you're right https://nim-lang.org/docs/sequtils.html#foldr.t,untyped,untyped
19:29:28	FromGitter	<nitely> ah, right. That one should work
19:30:07	FromGitter	<samdmarshall> how would I use that to return a singular element of the sequence?
19:30:52	miran	samdmarshall: can you be more detailed about what exactly you're trying to do?
19:31:49	FromGitter	<nitely> @samdmarshall you want the first one that matches your cmp?
19:32:27	FromGitter	<samdmarshall> i'm not sure how more detailed I can be: I have a seq that has probably one, but maybe more, element in it. I don't care which element of the sequence I am given, I just want one of them
19:33:25	*	SenasOzys joined #nim
19:34:06	miran	so basically you want a random element from that seq?
19:34:17	FromGitter	<samdmarshall> yes
19:34:27	FromGitter	<data-man> LLVM/Clang 6.0.0 released. http://releases.llvm.org/6.0.0/docs/ReleaseNotes.html
19:35:09	miran	https://nim-lang.org/docs/random.html#rand,int
19:35:24	*	Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
19:35:38	miran	seq[rand(seq.len-1)] ?
19:35:46	FromGitter	<samdmarshall> that would look something like `my_seq[rand(len(my_seq)]` yes?
19:36:30	FromGitter	<samdmarshall> yeah that is kinda horrible and was hoping for a simple operator, I might as well just hardcode the first element if that is the case
19:36:54	FromGitter	<ZarsBranchkin> Can always write a template for that, haha
19:37:35	FromGitter	<samdmarshall> returning the first element isn't that big of a deal, I was hoping there was a language construct for this
19:37:49	miran	take a note that max in `rand` proc is included in the range, so your example would sometimes fail. but you can use seq.high instead of seq.len-1
19:38:24	FromGitter	<ZarsBranchkin> Language construct is kind of a sketchy statement in Nim, since the standard libraries are just pure Nim as well. There is no difference if they implement it, or you write a template/procedure for that
19:38:59	FromGitter	<ZarsBranchkin> So if you would write `any` template, that picks random element like that, you could as well call it a new sequence construct
19:39:02	FromGitter	<samdmarshall> ok, let me rephrase, I was hoping there was something in the stdlib that already did this
19:39:20	*	Snircle joined #nim
19:39:45	FromGitter	<samdmarshall> randomizing the element selection is going to over-complicate this if I have to implement that myself
19:40:05	dom96	what's wrong with: seq[rand(seq.len-1)]?
19:40:47	FromGitter	<ZarsBranchkin> ```template any[T](list: OpenArray[T]): T = ⏎ list[rand(list.high)]``` ⏎ ⏎ And that's about as legit as what stdlib could provide you [https://gitter.im/nim-lang/Nim?at=5aa191bf888332ee3a1567e7]
19:40:53	FromGitter	<samdmarshall> it is just adding complexity to the code for the sake of it when I already know it will have at least one element
19:41:22	FromGitter	<ZarsBranchkin> Now with that you can just do: ⏎ ⏎ ```var someElement = mySeq.any``` [https://gitter.im/nim-lang/Nim?at=5aa191e1458cbde557102355]
19:42:02	FromGitter	<ZarsBranchkin> Which would be as efficient/fast as just writing out `mySeq[rand(mySeq.high)]`
19:42:06	dom96	Not that I want to bikeshed, but `pick` is a better name for this function
19:42:24	FromGitter	<samdmarshall> @ZarsBranchkin yeah, i'm aware I can do that; my question was about existing functionality and if I was missing something because I was trying to look it up using the wrong terminology -- not a question of "can I actually do this"
19:43:01	FromGitter	<samdmarshall> @dom96 yeah, I agree, I gave `any` in my example as that is what it has been called in other languages I know
19:43:08	FromGitter	<ZarsBranchkin> Now that you were interested in it, why not make a pull request and suggest such addition to sequtils? Now it will be part of stdlib
19:43:12	miran	samdmarshall: well, in python you would have `choice`, but it is not that much simpler than the nim code presented before
19:43:47	FromGitter	<ZarsBranchkin> The language grows from such additions anyway, that's why I was slightly confused about what's the problem here
19:44:04	FromGitter	<samdmarshall> ¯\_(ツ)_/¯ all I was trying to figure out if there was a convenience method that exists or not, if not then i'll just hardcode and move on in my work
19:44:39	miran	hardcode means you will have `seq[0]`?
19:45:12	FromGitter	<ZarsBranchkin> what about `seq.pop`, if you don't care about which element you get?
19:45:21	FromGitter	<ZarsBranchkin> That's part of stdlib already
19:45:54	FromGitter	<ZarsBranchkin> https://nim-lang.org/docs/system.html#pop,seq[T]
19:46:33	*	Chiel quit (Quit: Lost terminal)
19:47:17	*	Ven`` joined #nim
19:51:04	FromGitter	<samdmarshall> @ZarsBranchkin because that requires adding more mutability to my code
19:51:58	FromGitter	<samdmarshall> @miran if it means I don't have to write the implementation and is documented as part of the language then the complexity it has doesn't matter, as I don't have to maintain it
19:52:07	FromGitter	<ZarsBranchkin> Then probably index would be as simple as it gets
19:52:13	FromGitter	<samdmarshall> whereas, I do need to maintain my own code
19:53:44	miran	well, i still don't see the problem with `seq[rand(seq.high)]`, because it is quite self-explanatory in what it does, but that's just me.... :)
19:54:14	FromGitter	<samdmarshall> it takes infinitely longer to parse and comprehend than seq[0] does, so
19:54:56	miran	that kind of thinking won't get us much further than some addition of integers :D
19:55:07	*	douglascorrea quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:55:56	FromGitter	<ZarsBranchkin> Not sure if you should be worrying about parsing time, Nim expects you to meta-program it, that's what it was made for
19:57:04	*	douglascorrea joined #nim
19:57:11	*	rbrt quit (Quit: Oíche mhaith)
19:58:24	FromGitter	<samdmarshall> i'm talking about humans
20:01:12	FromGitter	<ZarsBranchkin> If I were you, I'd just write that template and forget about it. Don't see how it's any different from chunking up your code via funtions, except here you abstract away tougher to read code, to deal with what you just mentioned
20:01:41	FromGitter	<ZarsBranchkin> Remember that templates are compile time thing and are there for exactly that - make code more readable to human
20:03:38	FromGitter	<nitely> well, I mentioned `reduce/fold` cuz it can be use as an `any` that returns a seq value instead of boolean, which came up at some point in the conversation. But it does looks like you would better off just doing `my_seq[0]` to get the first element or `my_seq[my_seq.high]` to get the last one
20:05:35	shashlick	dom96: looks like nimble runs "before install" before installing dependencies so I cannot use dependencies in the before install section
20:06:06	FromGitter	<ZarsBranchkin> Yeah, fastest performing and prettiest looking would be template `first` that simply equals to `my_seq[my_seq.low
20:06:12	dom96	shashlick: what are you planning to use?
20:07:34	shashlick	for my nimgen dependent packages, nimgen needs to be installed before you can generate the nim files
20:07:45	shashlick	so steps are 1. install nimgen, c2nim dependencies 2. before install, generate nimgen artifacts 3. install to pkgs dir
20:08:08	shashlick	alternative is to run nimgen in the "after install" step but I need to be in the right pkgs directory and not quite sure how to do thata
20:09:21	shashlick	i can find the install path with "nimble path packagename" but then how do I "cd" into that in a .nimble script
20:11:01	FromGitter	<nitely> @ZarsBranchkin I believe ruby is one of the few languages that has stuff like `myseq.first` and `myseq.last`, but seriously, anyone who has program before knows what `my_seq[0]` means
20:12:34	FromGitter	<jaco60> To mimic a lot of FP languages, we could have myseq.head and myseq.tail (or myseq.car and myseq.cdr for lispians) :)
20:13:19	miran	b-but why? what's wrong with s[0]?
20:13:45	miran	well ok, s[1..^1] is not as pretty as pythons s[1:]
20:14:17	*	PMunch joined #nim
20:16:14	FromGitter	<samdmarshall> my 2-cents: index by zero as a concept is silly
20:16:20	FromGitter	<samdmarshall> so I prefer `.first()`
20:16:47	miran	let us not go there. also no vim vs emacs, and no spaces vs tabs :D :D
20:17:11	FromGitter	<ZarsBranchkin> Yeah, that's why i'd use seq.low or seq.high as indices, when implementing first and last
20:17:47	FromGitter	<ZarsBranchkin> And now i'm feeling like just making a PR with this for system.nim
20:17:58	*	Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
20:28:28	FromGitter	<krux02> @samdmarshall what is so bad with index by zero
20:29:38	FromGitter	<samdmarshall> I don't like hardcoding values if I can avoid it
20:29:42	FromGitter	<krux02> miran: I could imagin that nim would automatically apply the this pragma in `[]`
20:30:12	FromGitter	<krux02> meaning `a[0 .. len-1]` would work, because it can resolve len as `a.len`.
20:30:23	FromGitter	<krux02> and then all other symbols would work, too
20:30:30	FromGitter	<krux02> a[low .. high]
20:30:40	PMunch	Oooh, that would be fancy
20:30:59	FromGitter	<krux02> and if you define them also your own symbols `a[low .. middle]`
20:31:12	PMunch	Man, that would be really neat
20:31:21	PMunch	But I'm not sure if the this pragma would work
20:31:32	PMunch	Or wait, would it?
20:31:37	PMunch	For let's say a seq
20:31:54	FromGitter	<krux02> I can try a prototype with macros
20:31:57	FromGitter	<krux02> maybe
20:32:24	FromGitter	<krux02> I'll be back until then you get some waiting music: https://danieldeluxe.bandcamp.com/track/soul-siphon
20:32:43	PMunch	Would these be procs though, and not fields
20:32:48	PMunch	I don't think this works with procs
20:34:05	FromGitter	<ZarsBranchkin> Seems neat, but how would it react to variables named that way? I guess use those, if there sre none, assume a.low etc
20:41:35	*	rokups quit (Quit: Connection closed for inactivity)
20:42:01	dom96	shashlick: You can probably use this: https://nim-lang.org/docs/nimscript.html#cd,string
20:42:21	dom96	shashlick: Also, this might be a valid alternative to executing `nimble path ...`: https://github.com/nim-lang/nimble/blob/master/src/nimblepkg/nimscriptapi.nim#L44
20:42:32	shashlick	dom96: looks like my fix for #280 is wrong - https://github.com/nim-lang/nimble/pull/448/commits/4770556939a7dfd53c608926a6ccf0320dd5e77f
20:42:33	dom96	Definitely plan to expand this API in the future :)
20:43:19	FromGitter	<krux02> @ZarsBranchkin yes, but unfortunately I don't see an easy way to just prototype that functionality right now.
20:43:23	shashlick	the pre-install hook needs to be moved here: https://github.com/nim-lang/nimble/blob/master/src/nimble.nim#L349
20:43:37	dom96	shashlick: :/
20:43:51	dom96	It's time to add a test for this
20:44:57	FromGitter	<krux02> as far as I know macros don't allow to control the rules for symbol resolution
20:47:05	shashlick	getPkgDir() returns blank
20:51:19	dom96	hm, odd
20:54:06	*	rbrt joined #nim
20:56:12	*	nsf quit (Quit: WeeChat 2.0.1)
21:00:11	miran	https://www.tiobe.com/tiobe-index/ nim is not even in the first 100 languages? or it isn't even included?
21:00:28	miran	(yeah yeah, i know, tiobe index is shit, but just saw a link on HN)
21:02:25	*	miran quit (Quit: Konversation terminated!)
21:02:46	FromGitter	<ZarsBranchkin> miran: Apparently they do track Nim https://www.tiobe.com/tiobe-index/programming-languages-definition/#instances
21:03:01	FromGitter	<krux02> tiobe index is not shit, but it is just a metric they use, nobody can tell you how popular a language really is
21:03:26	FromGitter	<krux02> but to be honest how big can a language be when the entire community of that language shares a single chat room?
21:03:55	FromGitter	<jaco60> and we don’t need Tiobe to know that nim is not a mainstream programming language… yet
21:04:46	PMunch	krux02, what do you mean. We have plenty of chat rooms!
21:04:55	PMunch	They're all linked together though..
21:05:09	PMunch	Except from #nim-nologs and #nim-offtopic I guess
21:05:14	PMunch	And the Telegram channel
21:05:52	Araq	number of github projects seems to me a pretty reasonable indicator
21:06:14	dom96	Not everybody uses GitHub :)
21:06:23	Araq	no need for Tiobe's weirdness. do they still think Delphi and Pascal are different languages?
21:06:37	dom96	They are different languages though :P
21:06:55	FromGitter	<krux02> PMunch that is what I mean. All chat rooms are linked together, and it is still not overcrowded
21:07:16	PMunch	krux02, yeah I was just making a joke :P
21:07:58	FromGitter	<krux02> dom96: that is true, but it is just meant as a metric
21:08:35	dom96	I think measuring user count and message frequencies might be a good way to measure popularity
21:08:58	FromGitter	<krux02> I never programed in Delphi or Pascal, but I though Delphi to Pascal is like C++ to C.
21:09:05	dom96	or at least it gives you another metric to look at
21:09:08	FromGitter	<krux02> But it is just a complete random guess
21:09:38	dom96	Creating something that measures this is on my todo list in fact
21:09:44	FromGitter	<krux02> dom96: message frequency is good, but then you can only compare nim to nim, not to other languages
21:10:07	FromGitter	<krux02> dom96: at least there is still the chat log
21:10:12	FromGitter	<ZarsBranchkin> Heh, recently had to start using pascal(or delphi, not even sure). Was pleased to see some similarities to Nim, like variable type declerations at least
21:10:12	Araq	oh they fixed it. now only Delphi/Object Pascal is listed
21:10:22	FromGitter	<krux02> so you are able to track back to when the first chat was logged
21:10:26	*	rbrt quit (Quit: Slán agat)
21:10:37	PMunch	Yeah I was thinking about parsing the irclogs for new nicks and such to get some cool metrics
21:11:14	FromGitter	<krux02> yes, but then you need to parse the user from the bridges as well correctly
21:11:33	FromGitter	<krux02> some regular expressions to the win
21:11:57	dom96	krux02: Why only Nim to Nim?
21:12:09	Araq	krux02: "Delphi to Pascal is like C++ to C" that is correct but everybody made the switch
21:12:19	dom96	I could join channels dedicated to other languages and start collecting metrics
21:12:46	FromGitter	<krux02> yes that is true, but not all other languages have "official" channels
21:13:13	FromGitter	<krux02> the imporance of the channel in other languagas is just much different
21:13:44	FromGitter	<krux02> so I don't think you can compare it
21:14:24	Araq	the truth is: our docs and the search feature are so good that most don't use google or stack overflow nor #nim. and they all develop commercial closed source products
21:14:52	FromGitter	<jaco60> :)
21:15:06	Araq	so we think nobody uses it when in reality entire civilizations now depend on it ...
21:15:15	GitDisc	<treeform> Araq, is that true?
21:15:20	dom96	krux02: I think there is plenty of official channels to compare to: #rust on Mozilla's IRC network and #crystal-lang as an example.
21:15:36	GitDisc	<treeform> because that is what I am doing...
21:15:37	dom96	#d seems pretty official too
21:15:55	FromGitter	<jaco60> just compare SO channels, too
21:15:56	Araq	treeform: nah, I think I live in my own universe
21:16:04	PMunch	We should create a IRC -> Stack Overflow bot to get some more traffic going :P
21:16:33	PMunch	Any message by a new user followed by a question mark automatically gets added to Stack Overflow with the #nim tag
21:16:34	GitDisc	<treeform> I think stackoverflow as a good place before, now its too full of crap.
21:16:48	*	athenot quit (Ping timeout: 255 seconds)
21:16:54	PMunch	I still use it when I google for things. Lot's of good answers on there
21:17:04	PMunch	But I almost never ask questions there
21:17:10	GitDisc	<treeform> I end up there less and less.
21:17:15	*	PMunch quit (Quit: leaving)
21:17:29	GitDisc	<treeform> and when I do end up its not very useful anymore.
21:17:42	FromGitter	<ZarsBranchkin> Same, I have never asked question on stack overflow yet, but it has given me answers most of the time
21:18:09	*	athenot joined #nim
21:20:26	FromGitter	<jaco60> (https://files.gitter.im/nim-lang/Nim/Ryao/b_1_q_0_p_0.jpg)
21:20:41	FromGitter	<jaco60> The new bible of IT students
21:21:17	GitDisc	<treeform> Araq, do you think nimscript can be used as a scripting language sort of like lua or more like AngelScript? I did not find a way to included as a library or eval strings or files? Or expose compiled things to it?
21:22:03	dom96	treeform: nimble does this, and there is even a package that exposes a nice API for it IIRC
21:22:28	GitDisc	<treeform> I did not find it, I was poking around vm.nim it looks like thats where it lives.
21:23:16	GitDisc	<treeform> is this it? https://github.com/Serenitor/embeddedNimScript
21:23:26	GitDisc	<treeform> I guess I was looking for some thing inside the compiler or stdlib.
21:23:30	GitDisc	<treeform> but if this works it works
21:23:48	dom96	yeah, I think this is it
21:25:06	FromGitter	<krux02> I recently got over 500 points on stack overflow for a single answer I gave.
21:25:10	Araq	https://github.com/Araq/nimedit/blob/master/api.nim#L131
21:25:18	Araq	^ treeform
21:25:24	FromGitter	<krux02> But I too think stack overflow became worse
21:26:03	FromGitter	<krux02> it was a nice place to get answers, and often it still is, but when you ask a question, be warned it will be closed down in no time and tagged as not constructive.
21:26:23	FromGitter	<krux02> and the people who tag you are also pround and get achievements for doing that
21:27:04	Araq	btw somebody could work on NimEdit... ;-)
21:27:20	FromGitter	<krux02> Worst thing I had was, I was writing an answer to someone and while I wrote it the question was closed because it is "not constructive" and I was not able to give this person the answer.
21:27:52	FromGitter	<krux02> Araq: well I like nim, but to be honest, not everything needs to be written in Nim
21:27:57	GitDisc	<treeform> Araq, I written my own editor about 4 years ago, I don't need another one.
21:28:20	Araq	krux02, blasphemy :P
21:28:22	GitDisc	<treeform> Araq, it looks like that will get me started on embedding it.
21:28:24	FromGitter	<krux02> treeform: what does it do?
21:28:58	GitDisc	<treeform> krux02, the only really good property it has, is I know how it works. I don't recommend it.
21:29:00	dom96	I wrote my own editor too :P
21:29:04	Araq	well I'll extract some useful libraries from NimEdit, maybe.
21:29:12	dom96	Araq: Yeah, you should.
21:29:21	dom96	or improve embeddedNimScript
21:29:30	dom96	treeform: do try out embeddedNimScript if you can
21:29:47	dom96	It's good to use these libraries instead of copying and pasting custom code all the time
21:30:42	FromGitter	<krux02> is embeddedNimScript like python?
21:30:54	GitDisc	<treeform> All I really want is: load scripts, run scripts, expose compiled code to them, have a repl, and maybe break on errors and set breakpoints.
21:31:17	GitDisc	<treeform> embeddedNimScript does not do break on errors and set breakpoints, but maybe that can be hacked in.
21:31:41	Araq	needs some VM additions
21:31:46	Araq	but can't be hard
21:33:14	GitDisc	<treeform> I like the AngelScript model where you write what kind of looks like C++, so if you decided to port it to compiled code it is easy.
21:33:26	GitDisc	<treeform> Another idea is just to ship the game with a nim+c compiler as well?
21:33:31	shashlick	dom96: is it possible to get the output of an exec command in nimscript
21:34:12	FromGitter	<krux02> isn't that the return value?
21:34:21	dom96	Doesn't look like it https://nim-lang.org/docs/nimscript.html#exec,string
21:34:43	Araq	treeform: let make this clear: NimScript is Nim. modulo the VM's restrictions. it uses the same syntax, the same static type system, the same macro system, etc etc
21:34:46	dom96	I'm still a proponent of removing this nimscript thing and making the ordinary procedures work in nimscript
21:34:56	dom96	s/nimscript thing/nimscript module/
21:35:09	Araq	hmm?
21:35:51	dom96	just define osproc.execCmdEx for the VM
21:37:40	Araq	yeah, another thing we should have done for 0.18 :P
21:37:47	Araq	oh wait
21:37:53	Araq	we have staticExec()
21:38:19	dom96	huh? So you agree now that we shouldn't have a specific module dedicated to nimscript?
21:38:27	Araq	no.
21:38:40	Araq	I agree we need something like execCmdEx
21:38:50	Araq	but it turns out we already have.
21:39:18	shashlick	does staticExec work in nimscript
21:40:38	Araq	try it :-)
21:42:07	Araq	"If you are reading this you are missing nimblepkglist.js or have javascript disabled in your browser."
21:42:12	Araq	nope...
21:43:26	dom96	It's time to just remove this and link to https://nimble.directory
21:43:37	shashlick	how do you get the package name within nimble
21:44:01	dom96	`packageName`
21:44:04	dom96	possibly
21:44:06	Araq	packages?callback=gotPackageList:1 Uncaught ReferenceError: gotPackageList is not defined
21:44:06	Araq	at packages?callback=gotPackageList:1
21:44:23	Araq	why not fix that "callback"?
21:44:33	Araq	why is it broken anyway?
21:44:44	dom96	because it has been freezing my browser since forever
21:44:47	dom96	and many other's too
21:45:11	shashlick	dom96: packageName is blank in the after install
21:45:51	dom96	shashlick: :/
21:46:03	shashlick	i'm installing from local dir, not sure if that matters
21:47:20	dom96	No idea
21:47:38	*	Vladar quit (Quit: Leaving)
21:50:46	Araq	dom96: so ... you intentionally disabled it?
21:50:56	dom96	Araq: no?
21:51:01	dom96	No idea why it's broken
21:52:26	dom96	Does c2nim support objective C?
21:54:16	shashlick	got it working - staticExec() works thankfully, but have to hardcode packageName in .nimble file
21:55:47	dom96	oh cool, Nim is #2 on this: https://www.slant.co/topics/6032/~systems-programming-languages
21:57:09	shashlick	ugh, but now I am upsetting nimble - Cannot completely remove C:\Users\gt.nimble\pkgs\nimkerberos-0.1.0. Files not installed by nimble are present.
21:57:28	Araq	dom96: no to Objective C.
21:58:10	dom96	Araq: :'(
21:58:24	shashlick	so I have to use before install, which means nimble needs to be fixed
22:20:24	dom96	Maybe it's time to add libclang support to c2nim? :)
22:23:04	*	Ven`` joined #nim
22:25:44	shashlick	dom96: so if I push a fix to nimble, does it get pulled by koch tools?
22:32:01	*	smt joined #nim
22:44:00	FromGitter	<cabhishek> Hi everyone, I think this is a bug in `AsyncHttpClient`: ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ Running above code results in `Exception message: Connection was closed before full request has been made` error. Just to compare, equivalent async python code runs fine. [https://gitter.im/nim-lang/Nim?at=5aa1bcaf6f8b4b9946059324]
22:52:39	*	douglascorrea quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
22:56:57	shashlick	dom96: running nimble test doesn't do anything - I have a tests\testkb.nim and it doesn't run anything
23:02:32	*	Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
23:04:51	FromGitter	<zacharycarter> dom96, araq: I see why you guys use nimscript as configuration now :D I'm re-writing FRAG and going to write an editor this time - I'm using nimscript as configuration for things like shaders etc
23:10:03	FromGitter	<zacharycarter> With the macro I can write code like - ⏎ ⏎ ```add("bar")``` ⏎ ⏎ :D [https://gitter.im/nim-lang/Nim?at=5aa1c2cbf3f6d24c6852a9a7]
23:11:41	Araq	cabhishek report it on github please
23:12:09	*	Trustable quit (Remote host closed the connection)
23:12:17	Araq	shashlick: koch tools uses the latest nimble if you have a .git dir, aka have a Nim from git
23:12:43	*	MJCaley quit (Quit: MJCaley)
23:15:49	*	douglascorrea joined #nim
23:28:05	noonien	how can i see the called compiler/linker commands using nimble?
23:30:07	noonien	also, does nim have a proc main?
23:30:08	FromGitter	<zacharycarter> not sure that you can
23:30:25	FromGitter	<zacharycarter> no
23:30:45	FromGitter	<zacharycarter> isMainModule will be set to true when the module being invoked is the main module
23:31:04	FromGitter	<zacharycarter> so you can write
23:31:29	FromGitter	<zacharycarter> ```proc main() = ⏎ echo "foo" ⏎ ⏎ when isMainModule: ⏎ main()``` [https://gitter.im/nim-lang/Nim?at=5aa1c7d1c3c5f8b90d4d62b6]
23:31:46	FromGitter	<zacharycarter> or just call `main()` without the check if you don't care about it
23:32:22	FromGitter	<zacharycarter> regarding the nimble question - ask dom96
23:32:38	noonien	i'm just wondering if there's a reason to create a main proc in nim
23:32:42	dom96	noonien: try --verbose or --debug
23:32:55	FromGitter	<zacharycarter> yes there is - there's a performance benefit apparently
23:33:00	FromGitter	<zacharycarter> to having a main function
23:33:18	FromGitter	<zacharycarter> there's also a performance benefit to using the inbuilt result variable, rather than declaring and returning your own
23:33:34	noonien	dom96: i tried both, none output the compiler/linker command
23:34:33	noonien	i'm guessing not declaring a main function makes all declared variables global
23:34:47	*	littleli joined #nim
23:34:51	dom96	noonien: hrm, try `nimble c --verbosity:3 file.nim`
23:34:55	FromGitter	<zacharycarter> dom96: I'm thinking that knowledge about the VM needs to more prominent - game devs love this kind of stuff
23:35:15	FromGitter	<zacharycarter> the fact that I can execute code inside a configuration file is huge
23:35:41	noonien	dom96: nimble or nim?
23:36:11	noonien	nim shows me calls to gcc, so that appears to work
23:36:48	dom96	noonien: both should work
23:37:08	noonien	oh, my bad, nimble c works as well
23:37:32	Temperance[m]	I've been trying with a friend, to write a game engine in c++, the guy is relatively good with shaders but he isn't in c++
23:37:56	FromGitter	<zacharycarter> using what rendering API?
23:38:06	Temperance[m]	I expected him to design the shaders and all the stuff, and I told him I would provide him the rest for wathever visuals he dreamed to make.
23:38:21	FromGitter	<zacharycarter> I've written a couple of game libraries in Nim - but I'm re-writing the first one now.
23:38:44	FromGitter	<zacharycarter> http://fragworks.io/ and https://github.com/zacharycarter/zengine
23:38:55	Temperance[m]	But we have had lots of problems, he has problem with c++
23:38:56	Temperance[m]	I wanted to go with opengl
23:39:01	Temperance[m]	Vulkan is still too green and too difficult for me to consider it.
23:39:16	noonien	dom96: hmm, from what i can tell, only the linker call is shown
23:39:20	Temperance[m]	I'm thinking that with how easy is to build nim projects, It might be a good idea to do the c++ parts on separate files and just use nim as glue.
23:39:28	FromGitter	<zacharycarter> that's why I'm going to be using bgfx again
23:39:38	FromGitter	<zacharycarter> why do you need C++?
23:39:38	dom96	noonien: try removing nimcache :)
23:39:41	Temperance[m]	I have to study nim more, but if I can take out cmake out from the equation It will be a definitive win.
23:39:49	noonien	i did, no change
23:40:04	noonien	the cache and the binary
23:40:28	FromGitter	<zacharycarter> Temperance[m]: you don't need c++ :D just bind to it
23:40:59	dom96	noonien: in that case I'm out of ideas, Araq?
23:41:18	Temperance[m]	Knowing how to be efficient is something that is learned. You have to have powerful tools, you have to be able to relate asm to code easily, you have to be able to use intrinsics.
23:41:25	Temperance[m]	even if nim provided all that stuff, (and I have some problems debuggining yesterday) those are things that would probably take me a little bit of too much time relearning to do in nim.
23:41:28	noonien	from what i can tell, the commands are comments in the generated .c files
23:42:02	FromGitter	<zacharycarter> ah I see what you mean by glue code now - yeah I mean most frameworks I've written have leveraged several C/C++ libraries
23:42:29	FromGitter	<zacharycarter> for instance - http://fragworks.io/ - uses https://github.com/bkaradzic/bgfx for the rendering API
23:42:36	FromGitter	<zacharycarter> and the new iteration of frag will too
23:43:45	Temperance[m]	Still, several things in nim seemed really attractive, and the fact that the garbage collector can be deactivated, and yes, several c++ libraries could be reused, makes me think nim unlike other modern languages, has a little bit of tought probably put into making people like me not so unhappy.
23:43:48	noonien	Temperance[m]: hehe, sucky build systems is one of the reason i'm trying to switch to nim as well
23:44:08	noonien	that and the '70s preprocessor
23:44:43	Temperance[m]	It took me half a year to learn cmake, and I still am not sure I fully grasped it. I can't pretend to work with other people in c++ if they don't know how to deal with a tool that is HARDER to use than the language it is supposed to build.
23:45:01	FromGitter	<zacharycarter> I prefer premake
23:45:14	noonien	yeah, all build systems seem to invent a new DSL
23:45:23	noonien	when the real problem, imho, is within the language
23:45:24	Temperance[m]	I tried it, seemed simpler, but didn't like it.
23:45:26	FromGitter	<zacharycarter> and generally use - https://github.com/bkaradzic/GENie - if I"m using C++
23:46:36	FromGitter	<zacharycarter> https://github.com/bkaradzic - has kind of tamed the C++ world for me, via studying his code
23:47:06	FromGitter	<zacharycarter> at least in terms of producing portable and easy-to-read C++ code
23:47:35	Temperance[m]	-> this was written by him, and it kinda is my bible https://gist.github.com/bkaradzic/2e39896bc7d8c34e042b
23:47:48	FromGitter	<zacharycarter> yeah! that is a great read :D
23:47:54	noonien	sure, the problem isn't that you can't write good code. the problem is that c/c++ preprocessors are so bad it's really hard to trust anyone not to take "shortcuts"
23:48:27	Temperance[m]	To be honest, I only use c++ for 3 or for features. However, Most likely nim already covers said problems.
23:49:01	noonien	a good module system would probably not cause c/c++ to need such convuluted build setups
23:49:19	Temperance[m]	I would code plain C if it weren't for how hard is to search for "plain C" results, thanks to c++ having the same name, and for namespaces. I really don't like endless names you get in C
23:50:23	Temperance[m]	that being said. I think .h files have extra uses, It's nice to have already done a list of the functions in some part of your code.
23:50:36	Temperance[m]	The last part I mean it about modules and stuff.
23:54:19	noonien	do you think they're worth all the hassle you have to put up with to setup a complex enough build system to build your, and others' code?
23:58:20	*	MJCaley joined #nim
23:58:33	Temperance[m]	I don't think at all they're worth the hassle we have to put with. That being said I'm not sure headers had to be completely related to all those hassles either. C/C++ compilation workflow is a complete mess, and while yes, headers feel redundant, even if you didn't need headers, C/C++ would still be a pia to compile.