<< 17-02-2015 >>

---

00:00:13	Jehan_	I do agree that in the long term, the compiler would benefit from fewer one-letter symbols.
00:01:42	gokr	Jehan_: Lots of Smalltalkers moved to Scala, but.. I am also allergic to the JVM eco system.
00:01:45	TEttinger	gokr, what is Urhonimo?
00:02:05	gokr	TEttinger: http://forum.nim-lang.org/t/870
00:02:30	Jehan_	gokr: It's not so much the ecosystem (though I think the JVM tooling is a ridiculous jungle), but the design decisions that enforced upon Scala.
00:02:44	*	johnsoft quit (Ping timeout: 245 seconds)
00:02:47	Jehan_	99% of Scala's problems, in my opinion, exists because of Java interop.
00:03:00	Jehan_	It's still a darn nice language.
00:03:17	*	yibter joined #nim
00:03:27	Araq	yeah but only because they copied a decent amount of Nim :P
00:03:28	gokr	The latest work on Cog (VM for Squeak/Pharo) may soonish propel it into the vicinity of Java performance.
00:03:30	*	johnsoft joined #nim
00:03:30	Araq	(just kidding)
00:04:27	gokr	Jehan_: Cog just got a whole new object memory/gc and they are good on their way to add adaptive optimizations which Eliot thinks should put them in the same ballpark as Hotspot.
00:04:58	gokr	But whatever ;)
00:05:16	TEttinger	gokr: clojure uses hotspot too and idiomatic clojure is dog slow
00:05:25	Jehan_	gokr: That's very nice. It just doesn't fix my one problem with Smalltalk.
00:05:38	gokr	Jehan_: Let me hear it! I can take it!
00:05:46	*	gokr grabs big fish...
00:05:50	Jehan_	Which is that image-based development is definitely a love-it or hate-it thing, and I'm pretty firmly in the hate-it camp.
00:05:56	gokr	Ha!
00:06:23	Jehan_	Otherwise, I have little issue with the language itself.
00:06:26	gokr	Well, then you should look at Pharo these days. Sure, it has the image - but it also has very nice source code management with git integration yaddayadda.
00:06:58	Jehan_	In fact, the major reason why I used to program in Ruby so much was that it was the closest thing to "Smalltalk for the commandline" that I could find (other than gst).
00:07:04	gokr	So the "dark side" of the image isn't really there anymore. People tend to rebuild them and throw them away quite often.
00:07:09	Jehan_	gokr: I have a copy of Pharo on my computer.
00:07:15	gokr	Which version?
00:07:29	gokr	Its evolving very fast these days.
00:07:31	Jehan_	gokr: 3.0?
00:07:51	gokr	Mmm, that's pretty new - although 4 is... around the corner I think.
00:08:03	Jehan_	What I want is basically the opposite of what Smalltalk is designed around.
00:08:19	gokr	What is that then?
00:08:22	Jehan_	I.e. a blank slate that I fill in.
00:08:49	gokr	I wonder if you would find pleasure reading about Ian Piumarta's COLA stuff.
00:09:18	Jehan_	The other thing is that I like to have my choice of editor.
00:09:23	dhasenan	gokr: COLA? Cache-oblivious lookahead array?
00:09:56	TEttinger	piumarta made a bunch of experimental languages, collectively the soda languages
00:10:10	gokr	TEttinger: Whoa! Impressed
00:10:40	gokr	Jehan_: He has more papers there, but this one I saw him present: http://piumarta.com/papers/DLS-2006-slides.pdf
00:10:51	TEttinger	he's been an inspiration to a lot of language writers, apparently
00:10:54	Jehan_	gokr: But yeah, oddly enough it's the Smalltalk environment that's my issue, not the language (and I do realize all the stuff that Smalltalk has done for IDEs).
00:10:54	gokr	I think it was one of the first presentations of those ideas.
00:11:07	gokr	Jehan_: The COLA stuff is not Smalltalk.
00:11:26	Jehan_	gokr: Yes, I was continuing my earlier train of thought.
00:11:31	gokr	Ah, ok.
00:11:58	gokr	Ian wrote the first port of the Squeak VM to Unix etc.
00:12:13	TEttinger	Maru is a recursive implementation of lisp that performs at something like 30% the speed of optimized C, which is pretty neat
00:12:18	gokr	He is ... a bit of a VM genius, but not a finisher ;)
00:12:37	gokr	Yeah, Maru is the latest incarnation of COLA.
00:12:43	gokr	AFAIK.
00:13:05	gokr	He wanted to get rid of the VM - make everything dynamic all the way down.
00:13:21	gokr	Oh well, its pretty cool stuff.
00:14:08	Jehan_	gokr: As far as speed is concerned, I generally need considerably less than, umm, "C hand-coded by an omniscient deity", so I'm pretty flexible as far as implementations go, and don't really need a hugely optimized VM. I frequently run Nim code with all checks turned on during production, too.
00:14:26	gokr	Those slides included the 400 line js implementation that was faster than Firefox at the time.
00:15:38	gokr	the fun part is that the language "creates itself" when you start it up. Almost like... you know Steele's legendary presentation?
00:15:53	gokr	growing a language I think it was called.
00:19:56	gokr	Araq: I pushed huge.nim to gitlab, it doesn't compile ;)
00:20:23	gokr	But its "ported" AFAICT, modulo the usual slab of bugs
00:26:53	*	Trustable quit (Remote host closed the connection)
00:37:33	*	infinity0 quit (Remote host closed the connection)
00:39:58	*	gsingh93 joined #nim
00:41:39	*	Sharcho joined #nim
00:47:38	Mat4	ciao
00:47:44	*	Mat4 quit (Quit: Verlassend)
00:50:17	*	17WAA8EJD joined #nim
00:53:06	*	Varriount joined #nim
00:55:00	*	Jehan_ quit (Quit: Leaving)
01:00:12	whitenoise	def-: so can Nim do bit fields?
01:01:31	def-	whitenoise: maybe someone wrote a module for them, i don't know
01:02:44	Varriount	whitenoise: Not out of the box, unfortunately. There's the 'packed' pragma, but that doesn't give much granularity.
01:03:21	fowl	theres a workaround
01:04:14	fowl	use importc/header or importc with emit
01:05:37	Araq	bye
01:07:16	*	filwit quit (Quit: Leaving)
01:12:53	*	17WAA8EJD is now known as fungos
01:19:06	*	smw_ quit (Read error: Connection reset by peer)
01:25:35	*	pogoplug joined #nim
01:25:59	*	reem quit (Remote host closed the connection)
01:27:39	pogoplug	I'm having problems bootstraping nim on Ansible linux. I get the following error: "nimcache/stdlib_osproc.c:1628:57: error: use of undeclared identifier 'POSIX_SPAWN_USEVFORK' LOC33 = posix_spawnattr_setflags(&attr, (NI32)((NI32)(POSIX_SPAWN_US..."
01:28:07	*	OnwardEuler joined #nim
01:28:33	def-	pogoplug: with which C compiler?
01:28:44	pogoplug	clang
01:28:51	pogoplug	and musl-libc
01:29:07	def-	ok, that explains it
01:29:16	pogoplug	not to me
01:30:42	def-	https://github.com/Araq/Nim/blob/devel/lib/posix/posix.nim#L1732-L1744
01:30:55	def-	I noticed the same issue with TCC and added this workaround
01:31:15	def-	I guess we need it in a more general way
01:34:32	*	reem joined #nim
01:35:42	def-	Made a PR to fix this, don't know if there's a better way: https://github.com/Araq/Nim/pull/2156
01:37:03	pogoplug	so its not a compiler issue, its a libc issue
01:37:03	*	koz_desktop quit (Remote host closed the connection)
01:37:58	pogoplug	def-: probably want to add that musl-libc doesn't specify VFORK in spawn.h
01:38:13	pogoplug	def-: I took your patch and add or defined(clang) and it built for me
01:38:24	pogoplug	of course that's not the right thing (its the libc not compiler)
01:38:28	pogoplug	but it moves me forwards
01:38:50	def-	pogoplug: yeah, i'll wait to see what Araq thinks about how to fix this properly
01:39:45	pogoplug	def-: pull in autotools of course!
01:39:57	def-	I doubt it
01:44:23	pogoplug	I need irc emoticons for sarcasm
01:50:09	*	dashed joined #nim
01:51:01	*	reem quit (Remote host closed the connection)
01:57:30	*	reem joined #nim
02:01:16	*	Sharcho left #nim ("Leaving")
02:05:04	*	JinShil joined #nim
02:06:56	*	pogoplug1 joined #nim
02:09:33	*	pogoplug quit (Quit: Page closed)
02:09:40	*	pogoplug1 quit (Client Quit)
02:11:27	*	chemist69_ joined #nim
02:11:35	*	pogoplug joined #nim
02:11:53	*	l04m33 joined #nim
02:14:40	*	chemist69 quit (Ping timeout: 255 seconds)
02:15:32	Varriount	pogoplug: We usually use ':3' or ';P' for sarcasm
02:17:14	*	JinShil quit (Quit: Konversation terminated!)
02:18:02	*	darkf joined #nim
02:18:44	*	JinShil joined #nim
02:23:30	flaviu	Varriount: I've never see ";P". Combining emoticons seems wrong :P
02:27:05	*	jholland quit (Quit: Connection closed for inactivity)
02:27:43	Varriount	flaviu: Ok, well at least ':3'
02:27:52	Varriount	I use that for when I'm joking.
02:30:07	l04m33	Hi folks, I'm tying out the asyncdispatch module, and my program here produced some `type mismatch` message: https://bpaste.net/show/19d03ec9303d
02:30:14	l04m33	Any ideas or pointers?
02:32:43	Varriount	l04m33: I don't think Asynccheck is a closure procedure.
02:32:43	*	pogoplug quit (Read error: Connection reset by peer)
02:33:08	Varriount	l04m33: You can wrap asynccheck in another procedure to get around the limitation
02:34:11	*	yibter quit (Quit: Textual IRC Client: www.textualapp.com)
02:34:23	*	johnsoft quit (Ping timeout: 246 seconds)
02:36:25	l04m33	Varriount: Thanks! Never read about `closure procedures`. Diving into that now....
02:36:48	*	johnsoft joined #nim
02:37:17	Varriount	l04m33: The rules about what can be implicitly converted to a closure and what can't are a bit complex, for reasons I can't recall at the moment.
02:37:24	Varriount	I do know that there are reasons though.
02:38:24	Varriount	l04m33: It's probably less of a headache to just for 'for f in y: asyncCheck(f)'
02:38:30	TEttinger	def-: I think it's cool what people are using TCC for these days
02:38:31	Varriount	*just do
02:39:55	l04m33	Varriount: I see. Maybe it had something to do with the translated C code.... I'm just trying out different ways of doing something :)
02:41:19	Varriount	l04m33: Closures have two pointers - one to the actual procedure, and one to the procedure environment
02:41:38	Varriount	It has to do with that, I think (though I'm not sure)
02:42:22	def-	you can use every proc as a closure inside its module. from another module they need the closure pragma for that, I think
02:43:08	Varriount	I think Araq said that it's like that because of 'default arguments', though I didn't quite understand why that was.
02:43:49	def-	"for f in fList: asyncCheck(f)" looks more nim-like to me anyway
02:44:05	def-	Varriount: right, there was something with that too, but I forgot as well. in this case I don't see any default arguments
02:45:16	def-	asyncCheck doesn't even return anything, so a map doesn't make sense
02:45:50	l04m33	def-: OK.... I came from the Python world you know :)
02:46:07	Varriount	l04m33: Ah, so you're like me.
02:46:38	Varriount	l04m33: If you use sublime text, onionhammer and I are the ones who develop the Nim extension plugin for it.
02:47:01	l04m33	So besides the manual, are there any other docs about closure usage?
02:47:28	Varriount	l04m33: Well, there are language-agnostic explanations of closures
02:50:02	l04m33	Varriount: I kinda understand what a `closure` is, but want to know more about how it works in Nim.
02:50:37	Varriount	l04m33: Well, in Nim, closures are usually inner functions
02:53:56	Varriount	l04m33: https://gist.github.com/anonymous/524404fe6fa42dc782ee
02:54:53	Varriount	l04m33: In that example, 'bar' is implicitly given 'closure' status
02:55:15	*	lavender joined #nim
02:55:55	Varriount	It grabs the outer variables ('c') and can use them past the scope of 'foo'
02:57:28	l04m33	Varriount: hmm.... That looks pretty like closures in Python to me
02:57:45	Varriount	l04m33: Yes. Their fairly similar
02:57:50	Varriount	*They're
02:59:23	Varriount	l04m33: The thing is, behind the scenes they're represented as type containing two pointers. One pointer points to the function location, the other points to some allocated memory that holds the 'grabbed' variables.
02:59:26	l04m33	I'm reading the manual, and I think we should have a section for closures in the tutorials....
02:59:44	Varriount	l04m33: If you want to write one up, that would be great.
03:00:09	Varriount	l04m33: There's also http://nim-by-example.github.io/ , which could use some love.
03:00:41	Varriount	(Not all of the information there is 100% accurate)
03:02:32	*	johnsoft quit (Ping timeout: 250 seconds)
03:03:05	onionhammer	Varriount hows that mystery branch coming
03:03:22	Varriount	onionhammer: I'm halfway done with nimble.py
03:04:05	Varriount	onionhammer: While it's in my mind, I'll push what I have so far to a new branch, that way you can see what I'm up too.
03:04:39	*	johnsoft joined #nim
03:08:52	*	koz_ quit (Ping timeout: 240 seconds)
03:08:58	Varriount	onionhammer: https://github.com/Varriount/NimLime/tree/refactor
03:10:06	Varriount	onionhammer: Beware, that branch is likely broken in a few places. I'm testing commands as I makes changes.
03:10:55	*	flaviu quit (Ping timeout: 255 seconds)
03:10:55	*	Mimbus quit (Ping timeout: 255 seconds)
03:12:11	DecoPerson	Strange error with closures and tuples: https://gist.github.com/Deco/c9d9741a1f67d7d5ca94
03:13:22	*	johnsoft quit (Ping timeout: 240 seconds)
03:13:35	Varriount	DecoPerson: Try renaming the inner closure. 'inc' is a built-in function (increments an ordinal)
03:14:06	DecoPerson	no difference
03:14:06	Varriount	Though, that should probably be reported as a bug, or at least a warning.
03:14:29	DecoPerson	oh wait, it's a different error now
03:15:16	DecoPerson	https://gist.github.com/Deco/c9d9741a1f67d7d5ca94
03:15:48	DecoPerson	Ah, annotating blah with {.closure.} solved it
03:15:49	def-	DecoPerson: proc blah {.closure.} =
03:16:22	Varriount	Hm, probably because blah() doesn't actually capture any outer variables.
03:16:41	def-	Varriount: exactly, if you make it capture something, it's implicitly a closure
03:16:59	Varriount	DecoPerson: Could you look and see if those bugs have been reported, and if not, report them?
03:17:10	Varriount	onionhammer: No comments?
03:18:21	DecoPerson	Varriount: sure
03:18:36	DecoPerson	Works perfectly, thanks! https://gist.github.com/Deco/c9d9741a1f67d7d5ca94
03:27:19	DecoPerson	Why doesn't my redefinition of inc shadow system.inc?
03:28:06	def-	DecoPerson: good question, i think it should
03:28:21	Varriount	DecoPerson: Bug in the compiler.
03:28:31	Varriount	Which is why it should be reported as a bug.
03:29:31	DecoPerson	I guess one is "proc inc() {.closure.}" and the other is "proc system.inc*(x: var T) {.stuff.}"
03:30:47	Varriount	DecoPerson: It's a magic proc
03:30:48	DecoPerson	Given that procedures aren't first-class values, like in Lua or Python (because overloading would be impossibly difficult)... how does it decide which overload to resolve to when treating them as such?
03:31:05	Varriount	(Literally, one of the pragmas is 'magic'
03:31:08	Varriount	)
03:31:32	Varriount	DecoPerson: What do you mean, first class values?
03:31:38	reactormonk	DecoPerson, you can make them first-class just fine
03:31:49	Varriount	DecoPerson: You can pass around procedures and such just fine.
03:32:12	reactormonk	DecoPerson, you could consider them polymorphic functions.
03:32:12	fowl	DecoPerson, ret.inc should resolve to the field if thats what you named it, it probably works with an object type
03:32:36	Varriount	DecoPerson: Most of the 'overloading' is done at compile-time. The compiler does it's best to infer the best matching procedure.
03:33:45	DecoPerson	This is not an issue with magic procs: https://gist.github.com/Deco/c9d9741a1f67d7d5ca94
03:35:34	Varriount	Hm. So the compiler is matching against the outer procedure first, not the inner.
03:35:39	Varriount	Definitely a bug.
03:35:42	DecoPerson	If there's any conflict at all, there seems to be an issue: https://gist.github.com/Deco/c9d9741a1f67d7d5ca94
03:36:08	DecoPerson	oh, wait... what the
03:36:26	fowl	DecoPerson, try result.incval = incval
03:36:28	DecoPerson	The IDE hints say there's an error, but it compiles fine with "nim c"
03:37:24	Varriount	DecoPerson: What IDE are you using?
03:37:40	DecoPerson	Oh, nvm, I forgot to save
03:37:42	DecoPerson	https://gist.github.com/Deco/c9d9741a1f67d7d5ca94
03:38:00	DecoPerson	Definately a bug
03:39:15	Varriount	DecoPerson: What IDE are you using?
03:39:33	DecoPerson	fowl: doing "result.incval = incval" worked
03:39:40	DecoPerson	and resolves to the inner proc
03:40:01	DecoPerson	Varriount: Sublime with NimLime, which runs "nim --verbosity:0 idetools --suggest" for hints
03:43:31	gmpreussner|work	quick question for wrapping a C library: i have a function that takes an array of SomeType, i.e. "SomeType* values". is it correct to convert this to "ref SomeTypeArray" with "SomeTypeArray* {.unchecked.} = array[1234, SomeType]" ?
03:45:41	gmpreussner|work	i thought i had read something about better C array support on the forums recently and recall this syntax, but i can't find it :/
03:46:22	*	JinShil quit (Quit: Konversation terminated!)
03:49:06	Varriount	gmpreussner|work: Huh?
03:49:41	gmpreussner|work	Varriount: C function looks like this: void Foo(SomeType* values)
03:49:51	gmpreussner|work	SomeType* means array of SomeType
03:50:01	gmpreussner|work	a pointer to a static array is expected
03:50:09	Varriount	gmpreussner|work: You can just pass it an array. It doesn't need to be unchecked.
03:50:50	Varriount	Isn't an array just a pointer to a block of memory? You don't need the extra ref/ptr, do you?
03:50:53	*	lavender_ joined #nim
03:50:53	gmpreussner|work	oh, ok. So it would be: proc Foo(values: array[123, SomeType]) ?
03:51:14	gmpreussner|work	in C and C++, the name of an array is a pointer to its first element
03:51:27	Varriount	Ah, sorry, yes, you need an unchecked array if you want to write it that way.
03:51:34	gmpreussner|work	i thought so :)
03:52:01	gmpreussner|work	so is it this then: proc Foo(ref SomeTypeArray) ?
03:52:09	Varriount	You should probably set the length as zero though. The unchecked pragma doesn't affect memory allocation for arrays, just bounds checking.
03:52:57	gmpreussner|work	i see. i saw a lot of examples that use 10_000 or 100_000_000 for unchecked arrays. 0 will work just as well then?
03:53:00	DecoPerson	Varriount, def-, fowl: https://github.com/Araq/Nim/issues/2157
03:53:34	*	lavender quit (Ping timeout: 245 seconds)
03:54:34	gmpreussner|work	Varriount: also, in general for non-array parameters, but pointers, is it safe to convert "SomeType* value" parameters to "value: ref SomeType" or should i use "value: ptr SomeType" ?
03:54:54	Varriount	gmpreussner|work: ptr
03:55:06	gmpreussner|work	ok thanks
03:55:11	Varriount	gmpreussner|work: 'ref' implies that the value is under the control of the garbage collector
03:55:20	gmpreussner|work	ah that's right
03:55:33	gmpreussner|work	so my array parameter should then be ptr as well, i suppose
03:55:55	gmpreussner|work	or is that a different situation?
03:56:38	gmpreussner|work	nevermind, i see that i have used ptr with arrays before, and it seems to work fine
03:58:10	Varriount	gmpreussner|work: Well, if my knowledge is correct, 'ptr array' is a pointer to a pointer to the first element of an array
04:00:40	*	johnsoft joined #nim
04:05:05	*	kapil___ joined #nim
04:09:45	*	zacts joined #nim
04:13:11	dhasenan	When I turned on the LineTooLong hint and compiled a hello world program, the compiler emitted some hints about the standard library (using a recent copy in the devel branch). Is there some way to constrain hints to only emit for my project rather than everything?
04:14:04	*	dashed quit (Quit: Connection closed for inactivity)
04:15:08	onionhammer	Varriount cool i'll check it out
04:18:42	reactormonk	dhasenan, uh-oh. File a bug.
04:19:38	onionhammer	Varriount you mind if i delete a couple of the old branches
04:22:26	onionhammer	Varriount it's all one commit :P
04:22:51	*	lavender_ is now known as lavender
04:24:47	DecoPerson	gmpreussner|work: here you go: https://gist.github.com/Deco/208ac923ecc66f42caf2
04:25:50	DecoPerson	Should make it easier to work with raw C pointers
04:26:24	*	reem quit (Remote host closed the connection)
04:28:28	gmpreussner|work	thanks DecoPerson
04:29:38	DecoPerson	It has mistakes... I should probably write some unit tests
04:32:13	*	koz_ joined #nim
04:32:28	*	zacts quit (Quit: ERC Version 5.3 (IRC client for Emacs))
04:35:37	DecoPerson	Isn't there a `..<` range operator?
04:36:09	TEttinger	holy crap, urho3d and urhonimo look awesome
04:36:41	TEttinger	I'm surprised it isn't mentioned on Urho3D's page
04:42:22	dhasenan	Is there a way to include an import trace when displaying errors? Like "Error: undeclared identifier: foo (in file imported at a.nim:12, imported by b.nim:47)"
04:42:26	dhasenan	GCC does this sort of thing, I think.
04:43:05	DecoPerson	I swear I saw someone do this in Nim: for i in 0..<len:
04:43:15	DecoPerson	Where the < means it will go up to len-1
04:46:59	whitenoise	is there a difference in memalloc and malloc
04:47:15	*	polde quit (Ping timeout: 272 seconds)
04:47:21	whitenoise	oh, that was a template you made, nvm
04:48:06	Triplefox	DecoPerson: yes, < is used there to mean "len-1"
04:48:26	DecoPerson	Triplefox: but whenever I try to use it, it says unrecognised operator
04:48:32	Triplefox	add a space
04:48:35	Triplefox	after ..
04:49:20	*	reem joined #nim
04:49:24	Triplefox	it's a bit dangerous when you use it with a variable that will count into 0, since a negative index into a seq will give you the whole seq
04:49:39	Triplefox	we have an issue talking about that problem
04:51:57	*	lavender_ joined #nim
04:52:56	whitenoise	types.nim(10, 18) Error: type expected
04:53:00	whitenoise	what does this mean
04:53:21	DecoPerson	Depends on your code, paste to gist.github.com
04:53:32	whitenoise	https://www.irccloud.com/pastebin/SDAUtUvq
04:54:23	DecoPerson	Line 10, I think you mean "student = Student(...)"
04:54:50	DecoPerson	var varName : TypeName says "varName is of type TypeName"
04:55:13	whitenoise	I see. it's expecting a type but I'm giving it basically an object
04:55:23	whitenoise	makes sense
04:55:28	Triplefox	the verbose student:Student = Student() also works i think
04:55:37	*	lavender quit (Ping timeout: 264 seconds)
04:55:39	DecoPerson	whereas var varName = TypeName(something: "something") says "varName is of type TypeName and construct it with these values"
04:55:50	whitenoise	Triplefox: that does work, i have done it with ints
04:55:53	whitenoise	x: int = 10
04:56:21	whitenoise	https://www.irccloud.com/pastebin/PbHvrlKR
04:56:21	Triplefox	yeah...stylistically it probably doesn't matter which you use, unless type inference is breaking for some reason
04:56:24	whitenoise	this works
04:56:52	DecoPerson	wow, just tried "var x: auto = 10" as a joke, and it broke the compiler
04:57:01	whitenoise	lol
04:58:47	DecoPerson	https://gist.github.com/Deco/53401151c669b944e554
04:59:06	Triplefox	hmm
04:59:26	DecoPerson	Not exactly high-priority :)
04:59:38	DecoPerson	Not exactly high-priority :)
04:59:44	DecoPerson	opps, wrong window
05:03:37	*	lavender_ quit (Ping timeout: 244 seconds)
05:06:43	Triplefox	aha, you can make it compile with "var x: auto int = 10"
05:06:56	whitenoise	what is the auto doing
05:07:05	DecoPerson	why would "auto int" even be a thing?
05:07:19	whitenoise	^
05:07:29	Triplefox	i noticed that auto is defined as a type class
05:08:36	Triplefox	i don't understand that in any detail but it would mean it's a kind of categorization, just not a useful one in that situation
05:08:59	Triplefox	> If a proc param doesn't have a type specified, Nim will use the "distinct auto" type class (also known as "any"):
05:10:00	whitenoise	ah, sort of like interface{} in Go
05:10:41	Triplefox	it looks like every time we use "ref" or "ptr" we're actually working within the typeclass system
05:12:48	Triplefox	the part i really don't get is user-defined type classes...it's probably really powerful once i understand it
05:14:05	DecoPerson	gmpreussner|work: Here's the fixed + unit-tested version: https://gist.github.com/Deco/208ac923ecc66f42caf2
05:16:47	*	polde joined #nim
05:19:14	gmpreussner|work	DecoPerson: looking nice. that will probably come in handy.
05:21:17	Triplefox	also i didn't notice that "unittest" existed so i guess i'll use that
05:26:17	*	reem quit (Remote host closed the connection)
05:26:59	*	reem joined #nim
05:39:01	ekarlso	∞is devel broken ?
05:43:13	*	lavender joined #nim
05:49:41	DecoPerson	ekarlso: Latest devel on Win8.1 64 is working for me
05:49:54	DecoPerson	What's your issue?
05:51:46	*	kashyap_ joined #nim
05:51:51	*	polde quit (Ping timeout: 272 seconds)
05:52:31	kashyap_	Some folks discussed my moving from Rust to Nim :) http://www.reddit.com/r/rust/comments/2vqy81/author_of_unix_in_rust_abandons_rust_in_favor_of/
05:53:50	TEttinger	hi kashyap_, yeah I saw it mentioned on ##fsharp, I looked closer and popped in here :)
05:54:55	kashyap_	:) ... I think I should write a more detailed article about my path to Nim
05:55:11	TEttinger	how did you get started with learning/writing Nim, kashyap_?
05:56:42	kashyap_	A lady called Maxime had tweeted about Nim a few weeks ago...That's when I looked at Nim ... when I saw Nim, I could see that it was wat I had been looking for!!!
05:57:07	TEttinger	neat
05:59:00	Triplefox	that makes it sound so serendipitous
06:01:02	ekarlso	anyone wanna volunteer to help build the package indx ?
06:01:32	Triplefox	build it how?
06:01:42	ekarlso	code it ;)
06:02:46	Triplefox	as in add more, or as in automate the indexing..?
06:03:16	ekarlso	Triplefox: I've actually started doing a big deal of it ..
06:03:22	ekarlso	but more help would be nice
06:04:33	DecoPerson	Something like npmjs.com, ekarlso?
06:04:39	ekarlso	DecoPerson: yeah
06:04:49	ekarlso	https://nim-pkg.svcs.io < initial stuffs
06:05:28	DecoPerson	Oh boy, here come the growing pains
06:05:34	Triplefox	firefox doesn't like that certificate :D
06:05:43	ekarlso	DecoPerson: what's wrong now ?
06:06:34	*	polde joined #nim
06:06:48	Triplefox	oh, the tags browser is cool
06:07:03	ekarlso	Triplefox: there's a branch for nimble also
06:07:34	ekarlso	allows you to "nimble register" in the current dir of a new package and it does it for you
06:07:36	DecoPerson	ekarlso: nothing wrong with your site. it's really well :)
06:07:49	DecoPerson	I was commenting on how Nim is building inertia, yet it is still immature
06:08:16	DecoPerson	What if heaps of people use the "do" notation in their packages, but the developers decide to change it completely...
06:08:27	DecoPerson	growing pains, as faced by every other language :)
06:08:58	Triplefox	i'm dodging the problem by building nim libraries that only use the subset least likely to change
06:09:41	ekarlso	nim libs needs more CI / Docs automation..
06:09:55	ekarlso	and not a static index :/
06:10:33	Triplefox	here is an idea, docs that build on unit test code
06:10:47	Triplefox	so the unit test can also be an example
06:11:31	DecoPerson	Triplefox: I don't agree with that idea. Take the unit tests I just wrote for my memory library: https://gist.github.com/Deco/208ac923ecc66f42caf2
06:11:58	DecoPerson	They're not very good examples, in fact effective coverage (the goal of unit testing) can be painful and result in ugly code
06:12:11	*	brson quit (Quit: leaving)
06:14:25	Triplefox	hide it because it is ugly? it still shows how code is used, and it doesn't have to be the central example
06:14:43	Triplefox	that goes for "this should fail" tests too
06:16:56	DecoPerson	Hmm, interesting. I suppose docs and tests could be meshed together. They could lead with introduction, rationale and introductory examples, and then have the typical type/proc documentation with expandable unit tests for each
06:17:14	DecoPerson	But you'd need a way to specify which unit test is for which proc(s)
06:19:06	Triplefox	i like it mostly because it would help keep code/tests/docs in sync, although it would force the workflow to be a certain way
06:19:15	*	bjz joined #nim
06:19:20	Triplefox	so there would certainly be a project at some point that couldn't handle it like that
06:28:48	*	bjz quit (Ping timeout: 250 seconds)
06:36:23	Varriount	onionhammer: You there?
06:42:49	*	TEttinger quit (Ping timeout: 244 seconds)
06:44:17	*	brson joined #nim
06:53:08	*	brson quit (Ping timeout: 264 seconds)
06:55:04	*	bjz joined #nim
07:04:07	*	kashyap_ quit (Ping timeout: 246 seconds)
07:12:08	*	kokozedman joined #nim
07:14:14	kokozedman	Hey guys, compared to other Hash map implementations (like STL, Judy, ulib, etc...), how good is Nim's Table?
07:18:41	Varriount	kokozedman: I don't think there have been any tests done. I know there has been talk about optimizing it though.
07:20:19	*	vendethiel- quit (Quit: q+)
07:26:39	DecoPerson	dhasenan: Regarding LineTooLong, you could try not enabling it and then wrap your code in {.push hint[LineTooLong]: on.} and {.pop.}
07:44:32	*	fizzbooze quit (Ping timeout: 246 seconds)
08:00:16	*	ChrisMAN quit (Ping timeout: 255 seconds)
08:01:10	*	HakanD___ joined #nim
08:33:23	*	lavender quit (Ping timeout: 256 seconds)
08:34:33	*	lavender joined #nim
08:36:01	*	Trustable joined #nim
08:36:15	*	Trustable quit (Remote host closed the connection)
08:36:59	*	jaapz joined #nim
08:37:08	*	Trustable joined #nim
08:37:13	jaapz	let's start a religious war, why would nim be better than rust?
08:37:42	vegai	hah
08:37:47	jaapz	other than that it's irc is on freenode like any other self respecting project
08:40:37	bjz	nim's name is shorter!
08:41:21	jaapz	bjz: you've convinced me
08:41:25	Triplefox	it comes first in the alphabet too
08:41:47	jaapz	another great reason
08:42:31	bjz	<insert syntactic preferences here>
08:42:50	jaapz	bjz: rust syntax is a bit weird
08:43:00	jaapz	and there's a lot of it
08:43:16	bjz	I will be the first to admit that
08:43:23	bjz	:)
08:44:00	bjz	nim has no annoying borrow checker
08:44:25	bjz	or ownership shenanigans
08:44:38	jaapz	some would say that's a strength in rust
08:44:49	bjz	some would (myself included)
08:44:57	bjz	but for some that is a plus
08:45:53	vegai	nim's syntax is also a bit weird when you're not used to it
08:45:56	bjz	nim's metaprogamming seems more powerful, and prettier
08:45:57	jaapz	does nim run on raspberry pi and other arm-type thingies?
08:46:04	bjz	vegai: indeed
08:46:21	jaapz	vegai: fair enough
08:46:40	vegai	only Forth's syntax isn't, really :P
08:46:49	gokr	It seems to me that there are some fairly deep differences that overrule things like "syntax" etc. Philosophy is very different. Memory handling very different. Error handling. And compilation targets.
08:47:02	bjz	gokr: indeed
08:47:17	bjz	jaapz: nim has exceptions, rust does not
08:47:22	gokr	Its kinda odd staring at syntax when those differences I mentioned are in comparison HUGE differences.
08:47:53	bjz	rust compiles to llvm IR, where as nim has multiple backends to intermediate language targets
08:48:19	jaapz	gokr: still, syntax is what people see and what can be incredibly annoying
08:48:31	bjz	jaapz: you get used to it
08:48:45	bjz	jaapz: so long as its not COBOL
08:49:06	jaapz	bjz: fair enough
08:49:08	gokr	Nim wants to be practical. Rust wants to satisfy a very specific use case (build a super multicore browser?) and sacrifices programmability to do it.
08:49:15	gokr	That's my perception at least.
08:49:27	jaapz	gokr: the browser thing is it's main use case indeed
08:49:38	vegai	I think it's trying generally to get into that place where C++ is
08:49:44	Varriount	Wait, rust doesn't have exceptions?
08:49:53	vegai	no, not really
08:49:55	Varriount	._.
08:49:57	gokr	Varriount: Ha! You kiddin?
08:49:58	Triplefox	just this evening i mentioned nim in some other channel and got a kneejerk "not case sensitive?? reject" response
08:50:02	vegai	Varriount: I like it, actually
08:50:13	bjz	Varriount: no, it has mondaic error handling, and task failure
08:50:17	*	dumdum joined #nim
08:50:48	Varriount	Hm. I think 'mondaic' should be a real word
08:50:51	bjz	Varriount: monadic as in, Option or Result
08:51:05	bjz	lol woops
08:51:17	gokr	It seems to me that serious use of error handling in Rust would simply blow the code into a mess.
08:51:36	bjz	gokr: I haven't found it to be a problem
08:51:43	gokr	But I may of course be utterly mistaken. But every call site along the chain would need to be instrumented.
08:51:45	Varriount	bjz: No, I spotted your error. I'm just saying that 'mondaic' sounds like it should be a real word.
08:51:54	bjz	Varriount: heh, yup
08:52:10	vegai	the Go folks seem to moderately succeed without exceptions too
08:52:22	jaapz	vegai: pls don't talk about Go error handling
08:52:24	vegai	but Rust does it a bit better for the reasons mentioned above
08:52:34	jaapz	vegai: if statements everywhere
08:52:36	*	davidhq joined #nim
08:52:56	vegai	jaapz: mm, sure.
08:53:06	vegai	but conceptually simple
08:53:13	vegai	which is sometimes good
08:53:18	jaapz	vegai: true enough but it clutters the code like a bitch
08:53:22	*	vegai nods
08:53:29	bjz	jaapz: Nim has a single BDFL who calls the shots, where as Rust has a team people
08:53:32	*	chemist69_ is now known as chemist69
08:53:39	jaapz	f, err := OpenSomething()
08:53:46	jaapz	if (err) { handle the error }
08:53:58	jaapz	not my cuppa
08:54:00	vegai	if you're crazy, you can inline that stuff inside the if clause, I think
08:54:06	DecoPerson	I'm of the school that if an exception has occured, the program should exit because it's an exceptional circumstance and the program could not possibly recover. If it's a situation the program has been designed to handle (disconnection, file not found, out of memory, etc), then it should be handle by traditional control flow... IMHO :)
08:54:08	*	davidhq quit (Max SendQ exceeded)
08:54:09	gokr	bjz: Do you use both btw?
08:54:31	jaapz	bjz: so kinda python-y
08:54:42	vegai	< bjz> Varriount: no, it has mondaic error handling, and task failure
08:54:45	bjz	gokr: mainly Rust - I follow nim though at a distance, and upvote all the posts on it :3
08:54:52	vegai	I think they abandonded the erlang-style task failure thing
08:55:00	vegai	didn't they?
08:55:17	*	davidhq joined #nim
08:55:22	bjz	vegai: there is still failure to the task boundary, but no *linked* task failure
08:55:27	vegai	ah, ok
08:55:48	jaapz	if it's really bad you can panic!() your way out of the program
08:55:53	bjz	vegai: which is sad, and means no supervisors, but Rust's goals have shifted a tad
08:56:22	bjz	vegai: towards 'pay only for what you use'
08:57:32	bjz	gokr: what kind of stuff are you making in nim right now?
08:58:22	gokr	bjz: I work at 3DICC and our system is a 3D virtual worlds kinda thing. 100% in Smalltalk for the moment, but we just hired Araq and are moving towards Nim.
08:58:36	gokr	Did you miss the announcement the other day of Urhonimo?
09:00:14	bjz	that's pretty neat
09:00:18	bjz	yeah, I did
09:00:44	gokr	There were basically 3 important things in that announcement IMHO: 1) Araq works at 3DICC 100% Nimming 2) Nim can now wrap C++ very well automatically and 3) Urhonimo is a Nim wrapper of Urho3D
09:00:52	bjz	how is smalltalk to use?
09:01:03	bjz	awesome!
09:01:06	gokr	Of those 3 I would say the first one is the most important part for Nim.
09:01:22	bjz	will be good for nim to have commercial backing
09:01:32	Varriount	gokr: Hm. Wrapping C++ means Qt wrappers, right?
09:01:36	bjz	gutsy for the company, but hats off to them
09:01:44	gokr	Araq's work to produce Urhonimo produced a huge improvement in c2nim/nim to wrap C++ fully automatically.
09:02:11	gokr	3DICC is small. Me and Araq is 50% of it ;)
09:02:34	bjz	ahh :D
09:02:35	Triplefox	it has an impressive name though
09:03:47	gokr	But yeah, having Araq be able to work 100% on/with Nim, from home, is quite a big deal for the Nim community IMHO.
09:03:52	bjz	gokr: it's also great that it is being used in production, especially by Araq
09:03:58	gokr	Definitely.
09:04:08	bjz	that is a big test for a language
09:04:22	gokr	Yeah, we will be phasing Nim in gradually.
09:04:53	bjz	it's one thing to bootstrap a compiler, and another to actually use it in a large project (that is the experience of Rust, at least)
09:05:04	gokr	Varriount: Yes, Qt should be very doable now.
09:05:18	bjz	compilers are different beasts to messy real world stuff
09:05:25	gokr	Definitely.
09:05:35	bjz	(even though compilers are also messy)
09:05:47	bjz	wait - is nim bootstrapped?
09:05:58	bjz	sorry - made that assumption
09:06:02	gokr	It is
09:06:07	gokr	Its all in Nim.
09:06:09	bjz	nice
09:06:19	gokr	I wrote an article on that process btw.
09:06:31	Triplefox	qt bindings will be nice, i know someone who might be interested in that
09:06:48	gokr	http://goran.krampe.se/category/nim/ - the "Bootstrapping" one at the far bottom
09:07:37	gokr	bjz: I also think this new C++ wrapping capability is a very strong feature of Nim.
09:07:50	bjz	yup, that is a big pluss
09:08:04	gokr	Urho3D is a beast, and we already have fairly advanced samples running quite fine.
09:08:45	bjz	I think nim would have an easier time wrapping C++ than Rust. It's overloading and template semantics would most likely match more closely
09:09:19	bjz	jaapz: there you go, it is easier to wrap C++ with nim than Rust
09:09:19	gokr	bjz: This is a small one I did yesterday: https://github.com/3dicc/Urhonimo/blob/master/examples/particle.nim
09:09:31	gokr	Its nice, because its fully clean - 100% Nim.
09:09:53	gokr	Araq's demo is much more advanced though: https://github.com/3dicc/Urhonimo/blob/master/examples/character.nim
09:10:13	bjz	great
09:10:21	gokr	It has an emit in the beginning, he will fix that, but otherwise also fully in Nim.
09:10:35	*	davidhq quit (Ping timeout: 246 seconds)
09:10:52	gokr	This is that demo: https://www.youtube.com/watch?v=-T0MrJkLB8Y
09:11:34	Triplefox	that sure does look like a 3d game
09:12:17	gokr	Its those 600 lines of Nim code I linked above.
09:13:25	gokr	Driving an animated avatar, Bullet physics etc etc. Tested on OSX, Windows and Linux (should run on iOS and Android too). :)
09:13:26	*	lavender quit (Ping timeout: 246 seconds)
09:15:17	Triplefox	what is your ultimate goal for wrapping this engine? or is that up for disclosure
09:15:52	gokr	Our current system is already used in production etc, we are not a VC company.
09:16:27	gokr	So we already have a 3D client with very advanced features. But we want to upgrade to the Urho3D engine.
09:17:41	gokr	So we are injecting Nim into our current client written in Smalltalk as a start.
09:18:10	Triplefox	sounds like a complex project, but also a good "real" test
09:18:27	gokr	Its a very complex project indeed.
09:18:36	gokr	And our system is HUGE.
09:18:59	gokr	So... its going to be a step-by-step thing.
09:20:08	Triplefox	when projects get bigger i always get frustrated that i can't really know it all very well anymore
09:21:29	*	davidhq joined #nim
09:21:44	gokr	One of our promos, 3D comes 48 seconds into it: https://www.youtube.com/watch?v=oK3xjPaKuSg
09:22:27	gokr	You can also search youtube for "Teleplace" and you will find lots more - since that was the name of the system earlier.
09:27:40	Triplefox	that pitch is getting me all fired up :3 like playing second life at the office
09:27:47	gokr	Yes it is.
09:28:20	gokr	And oh, its all in Smalltalk ;)
09:28:31	Triplefox	heh
09:28:48	gokr	But in a 2 years, perhaps all in Nim.
09:29:01	gokr	We will see.
09:30:06	Triplefox	i know a lot of people who are or used to work at IMVU, sort of similar domain but more consumer focused.
09:30:44	gokr	Yeah, we are mainly "big corp" focused at the moment - but future directions may include more consumer stuff.
09:31:09	Triplefox	their stack is some mashup of haskell and c++ and i think one other language
09:31:10	*	BlaXpirit joined #nim
09:32:15	gokr	On our server side we are basically all Smalltalk. On the client however we use several third party libraries in C and C++.
09:32:29	gokr	So C/C++ interop is important for us in Nim.
09:32:29	Triplefox	oh, it's actually way more than i thought, heh http://www.quora.com/What-is-IMVUs-technology-stack
09:33:24	gokr	Yeah, well, our full stack is of course also much more.
09:33:59	gokr	Btw, Araq's wife's birthday today - so he has a day off ;)
09:39:02	*	rinukkusu_ joined #nim
09:39:42	*	rinukkusu_ is now known as rinukkusu
09:42:57	*	pafmaf__ joined #nim
09:47:25	*	davidhq quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
09:48:09	ekarlso	actually gokr u only have to run build.sh
09:48:11	ekarlso	-,,-
09:48:15	dtscode	how is math.random seeded?
09:48:45	*	davidhq joined #nim
09:49:20	*	davidhq quit (Client Quit)
09:49:20	gokr	ekarlso: Ehm?
09:49:51	ekarlso	gokr: instead of doing the whole snippet you have in your blogpost ;)
09:50:37	gokr	ekarlso: I am not following.
09:57:45	BlaXpirit	dtscode, what do you mean "how"?
09:57:51	BlaXpirit	it's basically undefined
09:57:54	BlaXpirit	whatever C does
10:03:05	*	teapot joined #nim
10:05:27	*	gsingh93 quit (Quit: Connection closed for inactivity)
10:06:50	*	sitaram joined #nim
10:09:37	*	kashyap_ joined #nim
10:18:21	*	reem quit (Remote host closed the connection)
10:20:02	*	reem joined #nim
10:21:07	fowl	did you guys know theres a wikipedia page now http://en.wikipedia.org/wiki/Nim_(programming_language)
10:21:24	rinukkusu	sick
10:21:45	fowl	it needs more references
10:25:29	*	Varriount|Mobile joined #nim
10:26:00	*	Var|Mobile joined #nim
10:26:01	*	Var|Mobile quit (Read error: Connection reset by peer)
10:26:18	*	Var|Mobile joined #nim
10:28:26	*	Var|Mobile quit (Read error: Connection reset by peer)
10:29:33	federico3	nice
10:30:07	*	Varriount|Mobile quit (Ping timeout: 255 seconds)
10:35:15	*	Var|Mobile joined #nim
10:35:31	*	kuzy000_ joined #nim
10:44:04	*	Var|Mobile quit (Ping timeout: 255 seconds)
10:51:15	*	Varriount|Mobile joined #nim
10:52:15	*	Var|Mobile joined #nim
10:52:17	*	reem quit (Remote host closed the connection)
10:53:45	Var|Mobile	Hrm
10:54:36	Var|Mobile	¤Var¤
10:54:55	Var|Mobile	Interesting android keyboard
10:55:22	BlaXpirit	k
10:56:03	*	Varriount|Mobile quit (Ping timeout: 252 seconds)
11:04:53	*	HakanD____ joined #nim
11:05:58	*	TEttinger joined #nim
11:07:22	*	HakanD___ quit (Ping timeout: 245 seconds)
11:14:14	*	HakanD____ quit (Quit: Be back later ...)
11:17:13	rinukkusu	how does that make you guys feel? http://developers.slashdot.org/comments.pl?sid=6971335&cid=49059393
11:17:46	*	flaviu joined #nim
11:27:19	*	flaviu quit (Remote host closed the connection)
11:27:43	*	HakanD____ joined #nim
11:28:14	teapot	personally i don't think it's representative, all irc channels get heated now and then and tbh it's nothing to do with the language, just people being people
11:29:34	*	flaviu joined #nim
11:31:13	BlaXpirit	as much as I complain about Nim's module system, it's so impressive
11:31:45	BlaXpirit	just turned my library's structure inside out, but kept backwards compatibility
11:31:48	qwr	it's just different communication styles imho
11:32:52	BlaXpirit	you guys might wanna see how that conversation started
11:35:10	rinukkusu	the comment on stackoverflow got me here
11:35:40	rinukkusu	I know that irc communication and generally conversation on the internet differs from reall life
11:37:36	def-	rinukkusu: Hi. Usually the communication here is quite nice. That day was an exception, where it got really heated.
11:38:58	rinukkusu	no problem at all
11:39:05	def-	fowl: nice, can we copy the old one into it? https://en.wikipedia.org/wiki/Draft:Nim_(programming_language)
11:39:11	TEttinger	rinukkusu, yeah I came in #nim initially to prove to someone that the IRC channel probably wasn't crazy, and I got help on getting started in a complex setup very quickly from friendly, polite people
11:39:14	def-	(need to fix many links and stuff first)
11:43:29	teapot	yep I've always found this channel quick to help with my new-to-the-language problems. Often Araq has helped directly.
11:43:36	gokr	rinukkusu: Welcome. Yeah, I have been here daily since... october? But my perception is that the community is very friendly.
11:44:28	*	sepisoad joined #nim
11:44:54	sepisoad	is there a function to generate unique id (UID) in nim lang libraries
11:45:44	gokr	sepisoad: At the far bottom perhaps: http://nim-lang.org/lib.html
11:45:46	BlaXpirit	sepisoad, what I'd do in that case is search for "UID" at http://nim-lang.org/lib.html
11:46:30	gokr	Anyone who knows cgen inside and out?
11:46:38	BlaXpirit	looks like linux-only, that uuid thing
11:47:37	sepisoad	thanks @gokr & @BlaXpiri
11:47:54	qwr	some "oids" library is also there
11:48:29	qwr	http://nim-lang.org/oids.html
11:48:44	BlaXpirit	this looks more promising to me
11:49:33	BlaXpirit	it's based on C rand though :/
11:50:53	gokr	I just looked at the Cog VM and there is a UUID plugin there with specific code for mac, win, unix - using OS facilities. One could port that over.
11:51:32	BlaXpirit	I, as usual, turn to Python
11:51:45	gokr	Yeah, Python probably has something very similar.
11:52:43	gokr	Ha... the win code uses ole2. Talk about memory lane...
11:52:52	*	reem joined #nim
11:53:46	BlaXpirit	Python's implementation is extremely convoluted
11:54:20	gokr	It doesn't use OS facilities or?
11:54:43	BlaXpirit	it uses a ton of things.
11:55:29	BlaXpirit	https://github.com/python/cpython/blob/master/Lib/uuid.py if anyone is interested
11:55:37	BlaXpirit	I don't think it's worth copying
11:56:27	gokr	I could port over the one from Cog perhaps. Its very small.
11:57:11	*	reem quit (Ping timeout: 246 seconds)
11:57:41	gokr	I gather noone here except Araq (who is out having fun) knows cgen much? I stumbled onto a... bug.
11:58:42	*	untitaker joined #nim
12:00:03	*	jferg2010 joined #nim
12:02:00	*	reem joined #nim
12:02:29	Triplefox	There is a uuid generator in three.js, i didn't think i would have to port it but it's just a single function
12:02:53	*	reem quit (Remote host closed the connection)
12:06:08	*	HakanD____ quit (Quit: Be back later ...)
12:07:54	Triplefox	Also thinking this might be good to bind, or even port the minimal version http://www.pcg-random.org/
12:08:26	BlaXpirit	Triplefox, have you worked with that library?
12:08:34	BlaXpirit	is there any confirmation that it's good?
12:08:56	*	antrix joined #nim
12:09:24	Triplefox	There is more than one version of the library, and it's been around a little while. Maybe not good if you are leaning conservative
12:09:47	Triplefox	I haven't tried it myself yet but I've kept getting pointed towards it
12:09:51	BlaXpirit	it could be backdoored, who knows
12:10:34	Triplefox	It's not for secure numbers
12:11:29	BlaXpirit	then what's the problem with just using mersenne twister?
12:11:34	*	Gnist joined #nim
12:12:00	Triplefox	The minimal implementation is smaller than mersenne twister
12:12:27	BlaXpirit	that's not even an argument
12:12:46	*	sillesta joined #nim
12:12:49	*	teapot quit (Ping timeout: 246 seconds)
12:13:00	Triplefox	You're acting like this is important to you what random numbers people generate
12:13:25	BlaXpirit	I have raised concerns over this topic more than once, yes
12:14:07	Triplefox	You could implement what you want then, i want this one for myself
12:14:18	BlaXpirit	I have implemented
12:16:12	BlaXpirit	and it looks like Apache license will not allow me to use PCG generator in my library https://github.com/BlaXpirit/nim-random
12:16:19	*	OnwardEuler quit (Quit: Leaving)
12:19:12	*	Gnist quit (Quit: Leaving)
12:20:24	*	antrix quit (Quit: Leaving)
12:21:22	BlaXpirit	I have no idea how that would affect it. Apache license may be undesired.
12:25:20	*	Gnist joined #nim
12:31:00	Triplefox	Looks like it is only an issue with gpl1 and 2
12:31:47	Triplefox	Although dual licenses would be annoying
12:32:14	BlaXpirit	I don't think there is a way to add this without changing the license of my project
12:32:28	*	kapil___ quit (Quit: Connection closed for inactivity)
12:32:43	BlaXpirit	flaviu, maybe you have something to say. You have pointed me at PCG generators before, and I was about to add it, but ^
12:33:25	*	jferg2010 quit (Quit: Leaving)
12:33:25	*	teapot joined #nim
12:33:51	BlaXpirit	Xorshift, on the other hand, is in public domain, so just tell me which one(s) exactly should be added
12:34:03	def-	BlaXpirit: I think PCG would be nice to have
12:34:06	*	bjz quit (Ping timeout: 250 seconds)
12:34:13	*	teapot quit (Client Quit)
12:34:24	BlaXpirit	...
12:34:26	Triplefox	Also, derivative work may be relicensed so a native port will be fine
12:34:33	*	tapot joined #nim
12:34:49	BlaXpirit	where did you find that? -_-
12:34:57	Triplefox	Wiki
12:35:08	*	bjz joined #nim
12:35:30	BlaXpirit	hm
12:35:53	*	antrix joined #nim
12:37:46	*	flaviu quit (Ping timeout: 252 seconds)
12:39:55	*	flaviu joined #nim
12:40:05	flaviu	Sorry about that, my power went out.
12:40:31	flaviu	How about sending an email to Prof O'Neill? It's worth a try at least.
12:40:46	BlaXpirit	wiki might be lying though
12:40:59	BlaXpirit	I don't see anything about "derivative work may be relicensed"
12:42:24	flaviu	BlaXpirit: Section 9.
12:42:55	BlaXpirit	oh and 4 after the abcd list
12:43:19	BlaXpirit	mm not really...
12:44:06	BlaXpirit	flaviu, section 9 says about money
12:47:34	Triplefox	Oh well, I'm going to sleep now... have fun, don't stress too much just cause I raised the issue
12:47:46	*	rapind joined #nim
12:49:45	TEttinger	BlaXpirit: check out the xorshift PRNG shootout http://xorshift.di.unimi.it/
12:49:51	TEttinger	all of these are public domain
12:50:02	TEttinger	well all that he implemented
12:50:04	BlaXpirit	[:33:50] <BlaXpirit> Xorshift, on the other hand, is in public domain, so just tell me which one(s) exactly should be added
12:50:30	TEttinger	a ha
12:50:34	TEttinger	I mussed that
12:50:35	TEttinger	http://xorshift.di.unimi.it/xorshift128plus.c
12:50:44	TEttinger	that's probably the best right now
12:50:47	flaviu	PCG has some really cool ideas behind it btw, and the paper is easily accessible (if long)
12:52:14	flaviu	Yes, Xorshift* is very near ideal: http://i.imgur.com/9pAZMKE.png
12:53:07	BlaXpirit	flaviu, you say Xorshift*, TEttinger sent a link to 128+
12:54:13	flaviu	Ah, my bad.
12:54:17	TEttinger	yeah, on the shootout it talks about how xorshift+ combines properties of xorshift* with other things and it's a bit faster and has less flaws
12:55:21	flaviu	http://i.imgur.com/GdwPbHS.png
12:56:53	TEttinger	nice, how fast is pcg?
12:57:56	flaviu	http://i.imgur.com/Krh1ZFA.png
12:59:03	TEttinger	nice.
12:59:08	flaviu	I'm getting all this from the paper: http://www.pcg-random.org/pdf/toms-oneill-pcg-family-v1.02.pdf
12:59:21	TEttinger	ah ok
12:59:40	TEttinger	I'm curious about that apache license now
12:59:48	TEttinger	I thought it was GPL-compatible\
13:00:05	BlaXpirit	we're talking about MIT
13:00:10	BlaXpirit	GPL is completely irrelevant here
13:00:15	TEttinger	oh, ok
13:00:16	*	kapil___ joined #nim
13:00:23	*	Jehan_ joined #nim
13:00:34	BlaXpirit	this might be off-topic, dunno
13:01:25	flaviu	good idea.
13:02:40	BlaXpirit	oh hi, Jehan_. I finally gave http://forum.nim-lang.org/t/533/1#2886 a serious look, and yes, this is definitely worth implementing. the problem is providing a generic API for different RNGs
13:04:23	Jehan_	BlaXpirit: Yeah, I can relate. :)
13:04:51	BlaXpirit	may be worth changing the lowest common denominator from uint8 to uint32
13:05:40	BlaXpirit	but then again, there is at least one RNG that just provides bytes
13:07:11	flaviu	BlaXpirit: Which one?
13:07:24	BlaXpirit	SystemRandom/urandom/CryptGenRandom
13:07:46	flaviu	Concatenate the bytes together.
13:08:41	*	JinShil joined #nim
13:08:58	BlaXpirit	then it's gonna be a waste
13:09:28	flaviu	Who cares? If performance matters, people should use a fast PRNG
13:10:05	flaviu	That's a bit harsh, sorry.
13:10:10	BlaXpirit	I'm talking more about a waste of system's entropy
13:10:45	*	HakanD____ joined #nim
13:12:23	flaviu	Well, if you're just throwing away the bytes, entropy isn't used. The system's counter will think it's been used, but there are lots of other sources of error in that counter.
13:12:34	flaviu	And /dev/urandom won't ever block, so it's not a problem.
13:13:24	*	rapind quit (Remote host closed the connection)
13:14:14	BlaXpirit	I wish there were some kind of conditionals on whether a function exists or not
13:14:33	flaviu	when compiles(...): ...
13:14:57	BlaXpirit	flaviu, wait a second... I just realized that this actually may work in generics
13:15:15	BlaXpirit	gotta check this out
13:16:10	*	TEttinger quit (Ping timeout: 255 seconds)
13:16:49	BlaXpirit	omg it totally does work
13:17:10	BlaXpirit	https://bpaste.net/show/5f2866e08996
13:17:54	BlaXpirit	welp, I will go and rework the library now :|
13:19:33	*	reem joined #nim
13:19:33	*	reem quit (Remote host closed the connection)
13:21:44	*	tapot quit (Quit: Page closed)
13:22:55	*	teapot joined #nim
13:23:01	*	teapot left #nim (#nim)
13:30:39	*	untitaker quit (Ping timeout: 245 seconds)
13:30:45	*	HakanD_____ joined #nim
13:31:06	*	banister joined #nim
13:31:55	*	HakanD____ quit (Ping timeout: 255 seconds)
13:35:05	gokr	BlaXpirit: That's pretty slick :)
13:35:36	gokr	Lots of non Nimmers' eyes would roll back into the head though ;)
13:35:52	BlaXpirit	but this is exactly what I need
13:36:17	BlaXpirit	RNGs may provide any or all of randomByte, randomUint32, randomUint64
13:37:23	gokr	Just got another Urhonim example going - after hacking cgen.nim ;)
13:37:50	*	untitaker joined #nim
13:39:12	*	HakanD______ joined #nim
13:40:52	*	antrix quit (Quit: Leaving)
13:42:52	*	HakanD_____ quit (Ping timeout: 240 seconds)
13:44:24	*	HakanD______ quit (Ping timeout: 245 seconds)
13:45:12	*	teapot1 joined #nim
13:49:41	*	Gnist left #nim ("Leaving")
13:50:53	*	sepisoad2 joined #nim
13:52:01	teapot1	does anyone know if there are any plans to have features for safe sharing of memory between threads in nim?
13:53:03	*	kuzy000_ quit (Ping timeout: 265 seconds)
13:53:15	*	kuzy000 joined #nim
13:53:28	*	sepisoad quit (Quit: Leaving)
13:53:33	*	HakanD______ joined #nim
13:53:48	sepisoad2	sepisoad
13:58:23	*	sepisoad2 quit (Read error: Connection reset by peer)
14:01:33	*	BlaXpirit_ joined #nim
14:03:59	*	BlaXpirit quit (Ping timeout: 245 seconds)
14:05:41	*	rapind joined #nim
14:10:43	*	gsingh93 joined #nim
14:16:12	*	banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:20:23	*	reem joined #nim
14:21:53	*	jfchevrette joined #nim
14:24:09	*	Jehan_ quit (Quit: Leaving)
14:25:02	*	reem quit (Ping timeout: 250 seconds)
14:32:56	*	banister joined #nim
14:40:33	*	HakanD______ quit (Quit: Be back later ...)
14:45:21	*	wb joined #nim
14:49:45	ekarlso	if anyone is interested I got a vagrantfile that brings up fedora 21 with nim-vm and latest nim + nimble
14:49:52	ekarlso	fun for dev stuffs
14:51:41	jfchevrette	interesting. why not put it up on github and share the link?
14:51:51	jfchevrette	i'm interested :)
15:01:32	*	HakanD______ joined #nim
15:09:19	*	saml joined #nim
15:11:30	*	rapind quit (Remote host closed the connection)
15:20:52	*	JinShil quit (Quit: Konversation terminated!)
15:26:01	*	jholland joined #nim
15:32:31	*	saml quit (Quit: Leaving)
15:38:46	*	saml joined #nim
15:39:26	*	banister is now known as banisterfiend
15:40:11	shevy	ekarlso up with the link!
15:41:13	*	dumdum quit (Ping timeout: 256 seconds)
15:45:40	*	HakanD______ quit (Quit: Be back later ...)
15:48:36	*	rapind joined #nim
15:55:46	*	jfchevre_ joined #nim
15:56:10	*	jfchevrette quit (Ping timeout: 244 seconds)
15:57:44	*	kjo1 joined #nim
15:58:19	BlaXpirit_	flaviu, I'm pretty sure this is wrong https://github.com/flaviut/furry-happiness/blob/master/rand.nim#L25
15:58:56	BlaXpirit_	first of all, the conversion won't work if the number has a sign. second, it just truncates one bit regardless of sign
16:00:22	*	skroll3 quit (Ping timeout: 240 seconds)
16:02:10	*	skroll3 joined #nim
16:04:11	*	jfchevrette joined #nim
16:05:06	*	ChrisMAN joined #nim
16:06:07	*	nickles joined #nim
16:07:32	*	jfchevre_ quit (Ping timeout: 264 seconds)
16:07:52	*	nickles quit (Client Quit)
16:10:42	flaviu	BlaXpirit_: Well, it does work :P
16:10:43	*	Ven joined #nim
16:10:54	Ven	!kirbyrape
16:11:05	Ven	I'm sorry, I had to try.
16:11:31	*	kuzy000 quit (Read error: Connection reset by peer)
16:12:50	*	kuzy000 joined #nim
16:13:25	federico3	ekarlso: nim-vm?
16:13:32	flaviu	Or maybe it doesn't, investigating further.
16:15:53	reactormonk	How do I properly benchmark stuff? as in, circumvent dead code elimiation etc.?
16:16:48	flaviu	reactormonk: Typically, you set up an accumulator and do something non-optimizable with it like write it to stdout.
16:17:17	reactormonk	flaviu, yeah, but the problem with that it spends a non-trivial amount of time accumulating the results
16:17:34	reactormonk	e.g. converting to string.
16:18:36	flaviu	try grabbing the nth character of the string and adding it then.
16:18:54	reactormonk	nice idea
16:20:14	*	panzone joined #nim
16:24:19	*	panzone_ joined #nim
16:25:19	*	lavender joined #nim
16:26:17	*	banisterfiend quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
16:27:43	*	panzone quit (Ping timeout: 265 seconds)
16:29:26	reactormonk	flaviu, meh, can't reproduce the differences :-/
16:30:58	*	TEttinger joined #nim
16:33:28	*	panzone joined #nim
16:35:17	*	panzone quit (Client Quit)
16:35:20	*	Woflox joined #nim
16:35:31	*	gsingh93 quit (Quit: Connection closed for inactivity)
16:35:44	*	panzone_ quit (Ping timeout: 252 seconds)
16:44:31	*	irrequietus joined #nim
16:45:08	reactormonk	Araq, btw, any plans to move nim from Araq to nim-lang?
16:47:35	*	irrequietus quit (Client Quit)
16:47:51	*	HakanD______ joined #nim
16:48:33	ekarlso	federico3: nim version manager
16:50:03	TEttinger	but I like having a benevolent dictator for life
16:53:30	*	darkf quit (Quit: Leaving)
16:54:25	*	Jehan_ joined #nim
16:55:15	*	dumdum joined #nim
16:57:01	*	banister joined #nim
17:06:38	*	vbtt joined #nim
17:07:44	*	banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
17:09:37	*	dhasenan_ joined #nim
17:11:45	vbtt	hello
17:12:48	*	davidhq joined #nim
17:13:20	def-	hi vbtt
17:15:09	vbtt	pretty nice to see nim get good publicity in the last few months
17:15:34	vbtt	official support for araq is awesome
17:16:08	vbtt	although I don't care for c++, hopefully nim compiler issues will get fixed soonish now.
17:17:27	reactormonk	vbtt, been coding anything in nim?
17:17:32	Jehan_	vbtt: The importance of C++ support is to achieve a critical mass for external library support. Which is important for a lot of people.
17:17:54	Jehan_	It doesn't matter much to me, but that's because my work doesn't depend much on external libraries.
17:18:09	vbtt	Jehan_:I understand, I'm not against c++ support. doesn't excite me, that's all.
17:18:37	vbtt	reactormonk:not much really
17:18:38	*	Matthias247 joined #nim
17:19:07	vbtt	i wrote a blog about generics last year.
17:19:09	Jehan_	vbtt: As I said, it's not terribly important for me, but it makes Nim more viable for a number of people, which increases its userbase, which ultimately is good for Nim development.
17:19:15	vbtt	time to write another one about nim i think.
17:19:23	vbtt	Jehan_:agree
17:19:28	*	rapind quit (Remote host closed the connection)
17:19:59	reactormonk	vbtt, I can help, toss me something
17:20:22	vbtt	reactormonk:help with what?
17:21:02	reactormonk	vbtt, read-check etc. What do you want to write about?
17:21:28	vbtt	reactormonk:ah thanks.
17:21:36	vbtt	haven't decided yet. kinda short on time.
17:22:05	vbtt	i like server type code.
17:22:23	vbtt	maybe something about the concurrency models and async stuff.
17:24:17	vbtt	have to write more code in it first
17:25:21	*	rapind joined #nim
17:26:21	vbtt	btw, is there an stm library for nim?
17:28:27	reactormonk	What's stm?
17:28:41	Jehan_	software transactional memory.
17:28:43	vbtt	software transactional memory.
17:28:47	Jehan_	And not that I know of.
17:29:00	vbtt	so since heaps are isolated in nim, stm may be one good way to share data.
17:29:13	*	teapot1 quit (Ping timeout: 246 seconds)
17:29:26	Jehan_	Yeah, but … really hard to do without a GC for shared heaps.
17:29:48	Jehan_	There are also easier ways to do shared memory that don't involve STM.
17:30:22	vbtt	such as?
17:30:35	Jehan_	Also, transactional memory is still at the point where its performance is essentially too poor for practical purposes without hardware support.
17:30:47	vbtt	depends on what you use it for
17:30:48	Jehan_	Simple shared heaps that can be read-/write-locked.
17:32:03	Jehan_	vbtt: Well, pretty much anything? The problem is that the constant overhead for transactional reads/writes is so big that it tends to kill a lot of the gains you get from parallelization.
17:32:35	Jehan_	Furthermore, outside of specific data structures, there are easier approaches to get fine-grained parallelism working.
17:32:49	vbtt	Jehan_:depends on the implementation no? you could have read optimized stm, based on mvcc for e.g.
17:33:09	vbtt	it's mostly the writes that would be the problem
17:33:23	vbtt	anyway, does nim have shared heaps that can read/write locked?
17:33:38	Jehan_	vbtt: Of course it depends on the implementation, but I don't know of any that's both reasonably general and fixes the performance issues.
17:34:43	Jehan_	vbtt: No, but I've been advocating for it.
17:35:12	Jehan_	It's low-hanging fruit that would fit in nicely with the existing thread-local heaps.
17:35:41	vbtt	so these shared heaps would also be isolated? i.e. could objects in the shared heap point to other heaps?
17:37:38	Jehan_	vbtt: They'd be isolated. The point would be to have shared access to large data structures that you wouldn't have to move between threads.
17:38:14	vbtt	right, that would work too
17:38:14	Jehan_	Simple example: A large shared hashtable.
17:38:38	Jehan_	You'd still have to copy keys and values from the thread-local heap to the shared heap, but wouldn't have to move the entire table.
17:38:53	vbtt	i was thinking an stm that is isolated too.
17:39:04	Jehan_	Ah, gotcha.
17:39:13	vbtt	so,e g, use a cow btree for a hashtable
17:39:34	vbtt	then you have multiple parallel reads but single writer
17:39:44	Jehan_	But yeah, that would be tricky with GC again.
17:39:48	vbtt	very useful for caching type data that is common
17:39:53	vbtt	there would be no gc
17:39:59	vbtt	you'd have to remove the values yourself.
17:40:00	Jehan_	If you have one read/writelock per shared heap, GC works fine.
17:40:17	Jehan_	Since read access can't mutate the heap and with write access you have exclusive access and can GC at will.
17:40:38	vbtt	so all readers are locked out while the gc runs?
17:41:03	Jehan_	vbtt: Yup. Remember, though, that Nim's GC is soft RT.
17:41:33	Jehan_	And you actually have a lot of control over pause times.
17:41:44	vbtt	i see
17:42:19	Jehan_	This is why I said that it's low-hanging fruit that fits the existing GC model pretty well.
17:42:33	vbtt	yeah that would be pretty useful
17:42:43	Jehan_	It's not perfect, but it doesn't require a huge time investment, either.
17:44:27	vbtt	Jehan_:is it on the gsoc list?
17:45:30	vbtt	btw, is there is a way to deep copy an object in nim?
17:45:32	Jehan_	vbtt: The GSOC proposal is more ambitious. What I've been suggesting would be a bit too simple for GSOC.
17:46:02	Jehan_	vbtt: Yes, but I don't recall how you use it explicitly. It's used under the hood by spawn, I know that.
17:47:04	vbtt	also, is trunk stable enough mostly or should i use a published release?
17:48:00	*	rapind quit (Remote host closed the connection)
17:48:55	*	Ven quit (Remote host closed the connection)
17:50:49	reactormonk	vbtt, trunk is good
17:52:11	*	gsingh93 joined #nim
17:54:55	*	rapind joined #nim
18:04:20	*	banister joined #nim
18:13:33	*	lavender quit (Ping timeout: 250 seconds)
18:17:00	*	kjo1 left #nim (#nim)
18:38:10	*	UberLambda joined #nim
18:40:32	gokr	Araq: Here? Or still drinking champagne?
18:43:33	*	HakanD______ quit (Ping timeout: 264 seconds)
18:44:10	*	HakanD______ joined #nim
18:57:34	gmpreussner|work	is anyone using nim @ head on Windows today?
18:57:56	gmpreussner|work	i'm getting some odd segfaults at run-time in compiled programs. not sure what's going on.
18:58:12	gmpreussner|work	like local variable declarations can crash :/
18:58:20	*	singularity joined #nim
18:58:35	*	singularity is now known as singul42
18:58:43	singul42	hi there
19:00:01	singul42	I have just installed nimble and trying to install jester, but I get the following error:
19:00:06	singul42	Error: unhandled exception: Unable to query remote tags for git://github.com/dom96/jester/. Git returned: fatal: write error: Invalid argument [OSError]
19:00:25	def-	singul42: hi. do you have git installed?
19:00:46	singul42	yes, used a windows installer for it and it is also in the path
19:01:28	def-	On Linux it works for me
19:01:43	singul42	using "git" on the command line just gives me the git help
19:03:07	def-	as a workaround you can "git clone git://github.com/dom96/jester/" and run "nimble install" in the jester directory
19:03:14	def-	i don't know what causes the error
19:06:55	BlaXpirit_	Nim in Wine was such a great idea
19:07:21	BlaXpirit_	it allows me to test for: Windows, Nim 0.10.2, sizeof(int)==4 all at once
19:07:37	dom96	singul42: report this as a bug on Nimble's repo please
19:09:38	dom96	singul42: What happens when you execute: `git ls-remote --tags git://github.com/dom96/jester/` ?
19:09:47	*	dhasenan_ left #nim (#nim)
19:10:29	singul42	will try it in a second. just tried to clone it and also got an error. looks like git has problems on my machine with the git:// protocol
19:10:50	singul42	when using http:// I am able to clone jester
19:11:37	def-	that's an issue we've had a few times. sometimes we use git://, sometimes https://. should always be the same (probably https?)
19:12:00	singul42	using git ls-remote.... I get the error "fatal: write error: Invalid argument"
19:12:23	singul42	that's the same error I also get when trying to clone with git://
19:13:03	BlaXpirit_	looks like it's time to search for this particular problem
19:13:13	BlaXpirit_	this is not necessarily related to nimble
19:13:49	flaviu	Maybe it's related to corporate proxies?
19:14:22	singul42	at least I have no proxies running
19:14:29	dom96	Perhaps the ports which the git protocol uses is blocked?
19:14:32	singul42	also swicthed my firewall off for testing
19:15:40	singul42	well, I didn't restart windows after installing git .... hope that's not the problem
19:17:39	*	alexruf joined #nim
19:22:36	BlaXpirit_	made some changes to random library, if someone is interested https://github.com/BlaXpirit/nim-random/compare/602467577653aa5770208f49c98119903316bc96...master
19:24:06	BlaXpirit_	im away for now though
19:24:48	*	dom96 restored the draft of the Nim wikipedia article on https://en.wikipedia.org/wiki/Nim_%28programming_language%29
19:27:17	flaviu	BlaXpirit_: Added some comments on the commits
19:30:46	*	banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:31:10	BlaXpirit_	mostly unimportant things
19:32:03	ldlework	dom96: +1
19:33:03	flaviu	BlaXpirit_: Only the gitignore thing is unimportant IMO
19:33:50	ldlework	dom96: "Nim however differs greatly when it comes to semantics: for a start Nimr is statically typed."
19:34:19	dom96	ldlework: please fix it, I need to head out now.
19:35:15	*	koz_ quit (Quit: Leaving)
19:35:42	def-	ldlework: I'm also fixing many references and renames from nimrod to nim
19:35:53	ldlework	Well I just submitted the fix.
19:38:35	*	brson joined #nim
19:38:39	*	dumdum quit (Ping timeout: 256 seconds)
19:40:43	BlaXpirit_	flaviu, what's wrong with that cast?
19:41:07	flaviu	The cast is fine, the idea in general isn't.
19:41:26	BlaXpirit_	which idea
19:41:43	flaviu	Of falling back to seeding with the time if urandom is unusuable.
19:42:17	*	fizzbooze joined #nim
19:42:27	*	UberLambda quit (Quit: Leaving the Matrix)
19:43:17	*	alexruf quit (Quit: My Mac has gone to sleep. ZZZzzz…)
19:43:45	BlaXpirit_	what then
19:44:25	*	alexruf joined #nim
19:45:25	flaviu	If an attacker can cause that fallback to occur, you're pwned.
19:46:01	BlaXpirit_	people dont need to rely on that seed
19:47:17	*	singul42 quit (Quit: Page closed)
19:48:42	flaviu	But they will. It's the easiest way after all, and most people won't process the possible implications (or even be able to think of them!).
19:49:15	BlaXpirit_	I don't see how urandom can be unusable
19:49:51	flaviu	File descriptor exhaustion is one thing that comes off the top of my head.
19:50:50	flaviu	Or what about a chroot where no one has bothered to add /dev/urandom.
19:54:49	*	johnsoft quit (Ping timeout: 245 seconds)
19:55:30	*	johnsoft joined #nim
19:55:40	*	kuzy000 quit (Ping timeout: 250 seconds)
19:56:20	*	kuzy000 joined #nim
20:03:45	*	vendethiel joined #nim
20:06:32	*	dhasenan_ joined #nim
20:06:47	*	singul42 joined #nim
20:06:55	singul42	re
20:07:49	singul42	still having the problem with git:// protocol under windows, but as a workaround git config --global url."https://".insteadOf git:// helps
20:08:19	singul42	thanks for the help, hints and tips
20:11:22	TEttinger	singul42: woah that should help me too
20:11:25	TEttinger	thanks singul42
20:13:50	singul42	oh, I thought I am the only one on earth with the "fatal: write error: invalid argument" error when using nimble :-)
20:15:37	*	Mat4 joined #nim
20:15:41	Mat4	hello
20:18:10	singul42	hi Mat4
20:18:26	Mat4	hi singul42
20:21:29	*	Hakaslak joined #nim
20:26:48	*	Hakaslak quit (Ping timeout: 246 seconds)
20:30:48	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
20:32:37	ekarlso	https://github.com/ekarlso/nim-vm < cd vagrant; vagrant up < voila wokring rust devel
20:32:56	def-	rust devel?
20:33:04	ekarlso	ehm, nim devel
20:33:06	ekarlso	haha
20:33:08	ekarlso	:P
20:33:10	ekarlso	dunno where rust came into it
20:35:30	Mat4	from rust to dust
20:36:10	shevy	to lust
20:36:44	ekarlso	should it install nimble as well ?
20:41:51	vbtt	yes
20:47:47	*	UberLambda joined #nim
20:51:36	ekarlso	:p
20:56:42	*	HakanD______ quit (Quit: Be back later ...)
21:02:03	singul42	is there a nim library that offers high level procs to show a notification icon in windows' notification area? maybe also with some popup-menu functionality?
21:02:31	*	Jehan_ quit (Quit: Leaving)
21:03:56	BlaXpirit_	gtk?
21:04:04	BlaXpirit_	no idea
21:04:20	BlaXpirit_	Qt is my answer to everything, but it is not available for nim
21:04:52	flaviu	this is weird. If I replace one bit of code with another, it's 10x slower than the initial code. If I mark everything as inline, it's 2x faster than the initial code.
21:07:11	def-	flaviu: code pleae
21:07:22	singul42	well, looks like I have to dig into win32 api
21:07:28	def-	if it's short enough
21:08:37	*	TEttinger quit (Ping timeout: 250 seconds)
21:09:09	*	filwit joined #nim
21:09:15	flaviu	def-: I'm not really asking for help, but ok: https://github.com/flaviut/nim-random/compare/BlaXpirit:master...master
21:09:31	def-	flaviu: i'm just curious
21:11:34	*	nimthusiast joined #nim
21:14:54	*	alexruf quit (Quit: Textual IRC Client: www.textualapp.com)
21:16:12	filwit	BlaXpirit_: was reading logs, and noticed your pastbin, just wanted to point out a couple of alternatives in case you weren't aware of them: https://gist.github.com/PhilipWitte/a9ecf4d36b109ee179b7
21:17:51	BlaXpirit_	that's no good
21:18:31	filwit	I'm not sure of your use case. I just wanted to put it out there in-case you didn't know you could do those things.
21:19:37	filwit	why do I add a hyphen in the word "in-case"... silly brain...
21:20:51	filwit	BlaXpirit_: that said, the second example is basically exactly what your code is doing, but I find it a bit cleaner (I always find `compiles()` a bit too "hacky")
21:21:05	BlaXpirit_	but make it 4 classes
21:21:11	BlaXpirit_	across different modules
21:21:15	BlaXpirit_	it's oversimmplified
21:22:05	filwit	i'm not sure what you mean by "it makes 4 classes", sorry. Maybe I missed something in your original post.
21:22:35	BlaXpirit_	i mean do the same thing with 4 classes across different modules
21:24:07	filwit	that shouldn't be a problem. type classes apply to any other type (they're basically a way of wrapping up the `compiles()` statement into a clean package which only gets invoked once per unique-type)
21:24:27	filwit	but again, i don't know your use case
21:24:49	filwit	so i'm only showing you in case you didn't know about them
21:25:44	filwit	(ah, i think you're referring to the top example... yes that wouldn't work with 4 classes across modules, but the bottom example would)
21:28:13	*	Mat4 quit (Remote host closed the connection)
21:30:30	filwit	very cool to see the Nim wikipage back up!
21:30:55	filwit	guess the recent traction Nim has been getting did the trick :)
21:31:03	BlaXpirit_	ah i understand how your stuff works... yes, it'spretty much the same except you need to bother with typeclasses
21:31:03	gokr	131 users? yowza
21:31:36	Varriount	This comment might be interesting to those who wonder why Nim can't seem to get a Wikipedia article: https://news.ycombinator.com/item?id=9063800
21:31:54	Varriount	(or wasn't able to)
21:32:13	flaviu	filwit: I think it's up because they haven't noticed yet ;)
21:32:14	filwit	BlaXpirit_: yes, for a simple example the type-class might be overkill.. but type-classes can package a ton more than just checking for a single procedure (they can test anything), so it's very useful for this sort of thing.
21:33:31	flaviu	And don't typeclasses get lowered down to a series of when compiles(foo)?
21:36:23	*	singul42 quit (Quit: Page closed)
21:36:47	filwit	flaviu: that's my understanding, yes. But they clean up the code when you need to check multiple interfaces in multiple areas around your code. If you have an `Visual` type-class witch a few proc requirements (predraw, postdraw, draw, etc) and you need to work with "Visuals" in multiple places throughout your rendering pipeline, type-classes obviously make much more sense than doing `compiles(predraw(x)) and ...` all over the place
21:37:05	filwit	with** a few...
21:37:33	*	davidhq quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
21:37:53	flaviu	Yep, 20 lines once is better than 20 lines 20 times, especially for more complicated type classes
21:38:28	filwit	yep
21:39:24	filwit	https://www.youtube.com/watch?v=22-HSjMS3Ks
21:39:42	filwit	whoops... wrong video, sorry (nevermind that)
21:40:25	flaviu	haha
21:45:03	gokr	Tidbit: Just got a new demo running in Urhonimo (but it takes a hack in nim compiler) that is a stress test. It indicates that the Nim version is basically same speed as the C++ version.
21:45:04	*	nimthusiast quit (Ping timeout: 246 seconds)
21:46:29	*	reem joined #nim
21:46:38	filwit	I remember Araq (or someone) talking about using the Ref-Count GC as a leak-detector for non-GC'd Nim code before, but I forgot exactly how to do it (or what was possible).. Reading the recent discussion on the forums makes me think this would be a good thing to know (in order to help those who want to use Nim without a GC). Does anyone know more about this?
21:47:28	*	Jehan_ joined #nim
21:48:18	filwit	hey Jehan_, was just wanting to ask you about something (hopefully you know).. let me quote what I just said..
21:48:24	filwit	I remember Araq (or someone) talking about using the Ref-Count GC as a leak-detector for non-GC'd Nim code before, but I forgot exactly how to do it (or what was possible).. Reading the recent discussion on the forums makes me think this would be a good thing to know (in order to help those who want to use Nim without a GC). Does anyone know more about this?
21:48:47	filwit	Jehan_: Do you have any idea on how to use the GC in this way?
21:49:02	*	Boscop__ quit (Ping timeout: 245 seconds)
21:49:03	*	Woflox quit (Ping timeout: 244 seconds)
21:49:51	Jehan_	filwit: Not in the sense that I know how to actually implement it as part of the current codebase, but the idea is that a GC will detect unreachable objects. If that happens during manual memory management, that object is a leak.
21:50:02	Jehan_	The Boehm GC has explicit support for that, too.
21:50:33	Jehan_	That said, I'm not sure if the people who want a GC-less version of Nim know what they're in for.
21:51:04	filwit	I agree, but it would be a nice thing to point them towards.
21:51:22	Jehan_	Eh, that would only solve a fraction of their problems.
21:51:43	*	Boscop__ joined #nim
21:52:17	Jehan_	And let's not even get into the hard real-time stuff.
21:52:21	filwit	well didlybom had a good point in that he did not know what parts of the stdlib he could use without the GC.. I remember hearing that you could use the GC to detect this, but didn't know exactly how to direct him to it (which is why I ask)
21:52:41	Jehan_	At this point it may be easier to create a hard real-time GC or not do dynamic memory management at all.
21:53:03	Jehan_	Very little, really.
21:53:59	Jehan_	I mean, strings and seqs are the easy part, since they have copy semantics by default (like C++).
21:54:51	Jehan_	Basically, if I tried, I'd simply not use any ref types at all and do it like ParaSail.
21:55:08	*	jsudlow quit (Quit: Leaving)
21:55:30	Jehan_	Note that arguments and results are passed by reference by default even if not declared as var (unless they're very small).
21:55:31	*	reem quit (Remote host closed the connection)
21:55:36	Varriount	filwit: When the compiler has the GC switched off, it will warn when memory is being used that is ussually under GC control.
21:55:51	filwit	I imagine that would be best for embeded devices coding (areas you probably don't want the small overhead of the GC)
21:56:01	filwit	^ (directed at Jehan)
21:56:25	filwit	Varriount: Okay awesome. So Nim basically does this out of the box.
21:56:26	filwit	nice
21:56:56	Jehan_	Basically, what I'm saying is that you need to make do without GC, you probably want to make do without heap storage at all (or only under controlled circumstances).
21:57:05	filwit	hmm... I've tried with --gc:none before and didn't notice any messages though, maybe I wasn't looking hard enough..
21:57:39	Varriount	filwit: Odd. They should be there.
21:57:39	Jehan_	Nim is capable of working out of the box without ref types, which is the easiest way to get started.
21:57:46	*	davidhq joined #nim
21:57:50	filwit	Jehan_: yes i understood you, and I agree that's probably the best solution.
21:57:51	Jehan_	You'd still have to deal with string/seq-resizing.
21:58:33	ekarlso	nimvm now handles multiple versions and allows you to "nim-vm use v0.10.2"
21:58:44	def-	filwit: definitely there: xy.nim(4, 3) Warning: 'add(s, i)' uses GC'ed memory [GcMem]
21:58:46	Jehan_	I'd suggest to look at ParaSail and set Nim up to do it similarly.
21:59:24	filwit	Jehan_: that's an interesting idea actually... basically if you only used stack-allocation plus ptr types for referencing, you basically end up with something similar to Rust's memory management...
21:59:35	filwit	Jehan_: I feel like I'm really mistaken about that though...
21:59:37	Jehan_	Does nim-vm work like rvm? I haven't looked at that.
21:59:59	ekarlso	Jehan_: kinda yeah ;)
22:00:10	filwit	def-, Varriount: Yes thanks. I wasn't really looking for the messages when I tried awhile ago (was benchmarking something).. so it was just me.
22:00:19	filwit	def-, Varriount: thanks for the heads up.
22:00:21	Jehan_	filwit: ParaSail is sort of stack allocation + tree structures.
22:00:42	*	reem joined #nim
22:00:43	Jehan_	ekarlso: Hmm, not so exciting, unless it manages to not pollute shell profile files as much as rvm does?
22:00:52	Jehan_	I'm generally more of a fan of virtualenv.
22:01:23	Jehan_	filwit: Which works mostly like unique_ptr, minus the need to actually have pointers in a lot of cases.
22:02:35	filwit	I've never really taken a look at Parasail, but I remember watching a talk about it's memory-safety with a GC awhile back (didn't understand most of it at the time)
22:03:12	filwit	but now that I think about it, basically stack-allocating everything (except for seq/strings) would mostly be memory-safe in Nim with a GC
22:03:36	filwit	I don't know what virtualenv is, and my understanding of unique_ptr is limited too
22:03:55	filwit	it's been awhile since I've worked with C++ code heavily
22:04:08	filwit	and I never really used those smart pointer types when I did
22:04:13	*	pafmaf__ quit (Quit: This computer has gone to sleep)
22:04:26	BlaXpirit_	btw, when I did try to use typeclasses, the compiler failed big time
22:04:39	*	polde quit (Ping timeout: 272 seconds)
22:04:51	Jehan_	filwit: virtualenv was "as opposed to rvm". :)
22:04:52	filwit	BlaXpirit_: do you have an example?
22:04:56	ldlework	typeclasses compile for me fine, but the compiler doesn't enforce them when used as generic constraints
22:05:00	Jehan_	Two conversations at once, sorry.
22:05:17	*	davidhq quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
22:05:38	BlaXpirit_	filwit, quite likely that I don't
22:05:55	BlaXpirit_	not anymore
22:06:14	filwit	BlaXpirit_: i mean can you post a simple example online (of what is failing for you) so that we can help fix the issue?
22:06:44	BlaXpirit_	found it in logs, yay
22:06:45	filwit	the type-classes work fine in master and devel
22:06:58	ldlework	'work fine'
22:07:18	BlaXpirit_	https://bpaste.net/show/fddec6c851f2 on line 13 there are 2 generic arguments. if I make one of them normal, it works, just not with both
22:07:24	ldlework	filwit: they still don't work as constraints do they?
22:07:34	filwit	ldlework: yes my brain always types the wrong words that I'm thinking :S
22:07:54	ldlework	filwit: hehe
22:08:13	filwit	ldlework: if you mean `proc foo[T:MyTypeClass](x: T)` then yes, that works
22:08:40	filwit	BlaXpirit_: one sec, let me take a look
22:09:25	ekarlso	Jehan_: it just adds a single addition to PATH
22:09:27	ekarlso	nothing more
22:10:03	ldlework	filwit: as in it will only let me parameterize foo with a MyTypeClass? That's news to me.
22:10:21	ldlework	I'm pretty sure last time I checked user type-classes compiled but did not enforce against generics.
22:10:23	Jehan_	ekarlso: Yeah, still not keen on that, I'm afraid.
22:11:03	ekarlso	Jehan_: how else would you solve it ?
22:11:10	Jehan_	I mean, it's not as bad as rvm where I basically had to spend an hour ripping out everything it had put in there, but I still prefer it if my dot files are being left alone.
22:11:20	filwit	BlaXpirit_: to my knowledge you cannot use user-defined-TypeClasses like you're trying to (as regular type contstraints). Try: `proc draw*[D:Drawable](renderTarget: RenderTarget, obj: D)`
22:11:22	Jehan_	ekarlso: As I said, I'd prefer a virtualenv-like approach.
22:11:33	BlaXpirit_	filwit, maybe, maybe
22:11:40	ekarlso	Jehan_: hmmmms
22:11:43	BlaXpirit_	but that... is not very useful
22:12:00	filwit	BlaXpirit_: I may be wrong about that, let me check.
22:12:09	Jehan_	ekarlso: I mean, I'm not saying that it should be done this way. Just trying to explain my personal preferences. :)
22:12:23	BlaXpirit_	filwit, Drawable works just fine if the RenderTarget argument is not there
22:12:34	BlaXpirit_	RenderTarget works just fine if the Drawable is not there
22:12:44	BlaXpirit_	the combination of them breaks things
22:13:09	Jehan_	There used to be a problem with multiple generic arguments, that may be the same thing.
22:13:21	filwit	BlaXpirit_: your problem is with RenderTarget... define like this: type RenderTarget* = RenderTexture | RenderWindow
22:13:22	BlaXpirit_	your line of code doesn't work either
22:13:30	filwit	notice i removed the 'or'
22:13:39	filwit	i think that's the problem at least, (didn't test)
22:13:41	BlaXpirit_	doesn't help, filwit
22:13:44	filwit	crap
22:13:55	BlaXpirit_	no change at all
22:14:15	filwit	one sec, let me try to compile this...
22:17:03	*	BlaXpirit joined #nim
22:17:34	*	BlaXpirit_ quit (Quit: Quit Konversation)
22:18:05	*	kernasi joined #nim
22:18:07	kernasi	hola
22:18:22	filwit	BlaXpirit_: it looks like line 11 is causing the issue, but it's odd and I think it might be a bug
22:19:20	BlaXpirit	that's what everybody thought
22:20:31	filwit	BlaXpirit: actually I think it's because the semantics are wrong on that line (you can't pass 'RenderTarget' as a value to 'draw' cause it's a type not an instance) and just the error message is odd... one sec
22:21:24	ekarlso	Jehan_: how you think a venv approach could be done then ?
22:21:36	*	Cykacykacyka joined #nim
22:22:00	gokr	kernasi: welcome
22:22:05	Jehan_	ekarlso: Umm, I haven't looked at it, but I think it may require changes to nimble at least.
22:22:29	ekarlso	Jehan_: what about nim compiler versions then ?
22:22:30	*	rapind quit (Remote host closed the connection)
22:22:37	Cykacykacyka	hello, I came across this language and it seems really great from what I've seen
22:22:39	*	rapind joined #nim
22:22:40	*	rapind quit (Remote host closed the connection)
22:22:52	Cykacykacyka	does the YCM plugin for vim support this language ?
22:22:52	ldlework	Cykacykacyka: welcome
22:23:07	Cykacykacyka	and what could be a good non trivial project to try the language out ?
22:23:29	Cykacykacyka	and will you ever try and write a llvm frontend compiler for it ? :D
22:23:49	def-	Cykacykacyka: All I know for vim is this: https://github.com/zah/nimrod.vim/ (a bit out of date i think)
22:23:51	Jehan_	Different Nim versions can co-exist in different directories, the only issue there is the global config file.
22:24:41	def-	Cykacykacyka: someone was working on an llvm frontend, but it's been abandoned. you can compile through clang though, and that often gets nice speedups at compile and runtime
22:26:12	Cykacykacyka	yeah, but doesn't transpiling to c before bytecode mean that there are certain optimisations you simply can't do ?
22:26:47	def-	Cykacykacyka: a gameboy or nes emulator in Nim would be cool
22:27:51	Cykacykacyka	I haven't done any emulators yet, in fact, the only real life project I've made is an automisation framework for android in python, but hey, gameboy shouldn't be too hard, should it ?
22:28:40	BlaXpirit	nes emulator huh
22:28:57	BlaXpirit	isn't that too complicated
22:29:01	Cykacykacyka	and regarding the similarities to python, do you also do list comprehensions and the godly or operator in assignments ?
22:29:07	def-	at least those are fun projects i've been thinking about.
22:29:10	def-	BlaXpirit: he said non-trivial
22:29:26	BlaXpirit	Cykacykacyka, no
22:30:01	Cykacykacyka	BlaXpirit, although this is just speculation, I think that the cpu in the early gameboys are relatively well documented and are not brainwack craizy
22:30:20	def-	There seem to be quite good references on writing a gameboy emulator: https://github.com/mattrubin/Gambit
22:30:53	Cykacykacyka	and being a naive student with no understanding of anything, writing an emulator would give me good practical experience and force me to understand lower levels of hardware, shouldn't it ?
22:31:08	*	pregressive joined #nim
22:31:31	Araq	Cykacykacyka: the word "transpile" is verboten in this channel ... ;-)
22:31:43	def-	Welcome back, Araq
22:31:48	Cykacykacyka	Araq, noted.
22:32:17	Araq	(omg, seriously! if you don't know what you are talking about, stop acting like you do!)
22:32:33	def-	Araq: any plan to reupload the website soon? I like filwit's changes
22:32:54	Araq	that depends on dom96
22:32:57	def-	ok
22:33:00	Araq	I have it ready
22:33:06	Araq	for upload
22:33:45	Cykacykacyka	well, then elaborate, are you not compiling nim code to c code to binary ?
22:35:13	def-	Cykacykacyka: YCM looks similar to the nimsuggest we have
22:35:30	def-	should see how to get it integrated into vim
22:35:40	filwit	Araq: could you help me figure out why this type-class code isn't working: https://gist.github.com/PhilipWitte/8935a52bcf945e7a878b ?
22:36:07	Jehan_	Cykacykacyka: Transpiling generally means that the target language offers about the same level of abstraction as the source language and the translation process tries to preserve that; using C as a portable assembly language is not really the same thing.
22:36:18	Araq	filwit: no idea, sorry, I'm still busy getting static[T] into shape
22:36:27	filwit	Araq: np
22:36:43	filwit	I'm thinking it may be a bug in Type-classes anyways
22:37:19	Araq	but hey, at least I got their tests cleaned up
22:38:36	filwit	BlaXpirit: take a look at my gist: https://gist.github.com/PhilipWitte/8935a52bcf945e7a878b
22:38:53	filwit	BlaXpirit: it's not working, but the line i have commented out is along the lines of what you want to do
22:38:54	BlaXpirit	what am i supposed to see there
22:38:56	kernasi	how I can tell "nim" to use a different CC? in the output I see "CC" but setting CC doesn't have any effect; I'm used to this in C
22:39:03	kernasi	can I*
22:39:04	BlaXpirit	oh ok
22:39:06	kernasi	too much beer
22:39:13	filwit	BlaXpirit: i have `d is Sprite` just to make things compile
22:39:26	BlaXpirit	kernasi, --advanced shows some stuff
22:39:31	kernasi	I'm there :)
22:39:38	kernasi	but it still uses "gcc"
22:39:52	BlaXpirit	--gc:refc|v2|markAndSweep|boehm|none
22:39:59	*	UberLambda quit (Ping timeout: 256 seconds)
22:40:04	Jehan_	kernasi: --cc:clang
22:40:04	BlaXpirit	oh damn
22:40:07	kernasi	oh
22:40:15	BlaXpirit	CC sorry, I confused with GC
22:40:26	filwit	dom96: 127, new record :)
22:40:35	Jehan_	If you want to actually use a different version of gcc, that's slightly more complicated.
22:40:49	Jehan_	Since you then will have to specify the path.
22:40:51	BlaXpirit	wasn't it just 128 cuz UberLambda left
22:40:52	kernasi	no, just a wrapper :)
22:41:11	*	Demos joined #nim
22:41:12	filwit	BlaXpirit: oh i missed that, you're right
22:41:18	filwit	nevermind... 128 again!
22:41:22	BlaXpirit	:>
22:43:27	*	grumbly joined #nim
22:43:52	BlaXpirit	:>
22:44:45	grumbly	Does anyone have experience making nim use tinycc?
22:44:45	kernasi	I was actually looking for --passL:"-static" as I already replaced my libc :) thanks guys
22:44:59	gokr	kernasi: Bottom: http://nim-lang.org/question.html
22:45:09	kernasi	I wanted to replace $CC because I have a wrapper I use in C ("musl-gcc")
22:45:12	kernasi	but that it works as well
22:45:26	*	JinShil joined #nim
22:45:33	gokr	filwit: We were at 130 earlier
22:45:36	gokr	;)
22:45:46	def-	grumbly: yep, works perfectly, just need --cc:tcc
22:45:54	kernasi	gokr, nice, I change it in the config, perfect :)
22:46:06	kernasi	ya, sorry, I just downloaded nim and wanted to play around with it :)
22:46:47	def-	kernasi: --gcc.exe:musl-gcc should work
22:46:56	filwit	gokr: we're at that now too! quick, invite your friends!
22:47:04	grumbly	So I thought (--cc:tcc) however I get an error 'crt1.o not found' and 'crti.o not found'
22:47:13	filwit	:P
22:47:17	def-	filwit: https://gist.github.com/dom96/da2d4dead6fb00b71312
22:47:23	grumbly	I get farther if I force --cpu:i386
22:47:42	filwit	def-: er, okay then
22:47:50	grumbly	Both nim and tinycc work alone on hello world, but can't get them to work together.
22:47:59	filwit	this record goes up too fast these days I can't keep up!
22:48:39	kernasi	def-, yep :)
22:48:48	kernasi	seems flexible enough
22:49:00	def-	grumbly: maybe your tcc version is too old? i only tried with 0.9.26-r2
22:49:14	kernasi	tcc is not developed anymore, I wouldn't invest in that
22:49:19	kernasi	though it's nice
22:49:24	*	jfchevrette quit (Ping timeout: 245 seconds)
22:51:16	grumbly	def-, I'm using tcc v. 0.9.26 from here [http://download.savannah.gnu.org/releases/tinycc/]
22:51:43	grumbly	kernasi, Sounds like the right idea
22:52:05	Jehan_	kernasi: Set gcc.exe = "musl-gcc" in your config file.
22:52:13	def-	kernasi: grumbly if you google for it, there are some solutions: http://stackoverflow.com/questions/6329887/compiling-problems-cannot-find-crt1-o
22:52:25	Jehan_	The --cc argument can only be one of a number of predefined values.
22:52:29	kernasi	my god, the help in here is overwhelming
22:52:37	*	kuzy000 quit (Ping timeout: 264 seconds)
22:52:50	kernasi	Jehan_, ya, noticed that :)
22:53:08	grumbly	kernasi, Thank you; I hadn't found that one yet.
22:53:29	kernasi	redirecting your thanks to def-
22:53:30	Jehan_	For example, I have the following in ~/.config/nim.cfg
22:53:32	Jehan_	@if gcc48:
22:53:32	Jehan_	cc:gcc
22:53:32	Jehan_	gcc.exe = "gcc-4.8"
22:53:32	Jehan_	gcc.options.always = "-w -fpermissive"
22:53:32	Jehan_	@end
22:56:14	*	pregress_ joined #nim
22:56:54	grumbly	def-, Thank you. I'm sure if I spent a few more hours I could port that solution to my mac, but it's a headache. I'll just stick with Clang. Thanks for the help - this is a great IRC community!
22:56:59	*	pregressive quit (Ping timeout: 256 seconds)
22:57:33	def-	grumbly: oh right, i only tried on linux. clang is pretty fast at compiling already
22:58:34	*	grumbly quit (Quit: Page closed)
23:06:50	filwit	BlaXpirix: got it working: https://gist.github.com/PhilipWitte/8935a52bcf945e7a878b
23:08:10	filwit	BlaXpirit: ^; So basically you just need to prove that both the RenderTexture and RenderWindow can be called with `draw`
23:08:31	BlaXpirit	o.o
23:08:47	filwit	BlaXpirit: unfortunately that's not idea.. I'm still trying to figure out how to make that cleaner (using RenderTarget instead)
23:08:52	filwit	ideal*
23:09:10	BlaXpirit	ok thanks
23:09:26	fowl	i used generics like that in sfmls cpp wrapper
23:10:45	*	pregressive joined #nim
23:12:45	*	pregress_ quit (Ping timeout: 244 seconds)
23:17:25	*	BlaXpirit quit (Quit: Quit Konversation)
23:17:52	Jehan_	dom96: Where does nimble look for nim?
23:19:31	Cykacykacyka	has anyone used nim to produce binaries that run on android ? Not talking about using any of the android API's, basicall just using the c compilers provided by the android ndk ?
23:21:00	*	dom96_ joined #nim
23:21:36	dom96_	Jehan_: It looks in your PATH.
23:21:55	Jehan_	dom96_: Yeah, it wasn't finding it, but reinstalling from Git fixed that.
23:21:57	*	Var|Mobile quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
23:23:26	def-	Cykacykacyka: some old code here: https://github.com/gradha/nimrod-on-android
23:24:41	*	sillesta quit (Ping timeout: 250 seconds)
23:24:59	*	polde joined #nim
23:25:19	*	pregress_ joined #nim
23:27:28	kernasi	Jehan_, I had to add gcc.linkerexe = "/usr/local/musl/bin/musl-gcc as well; now it's perfect :)
23:27:37	kernasi	thanks for all your help
23:27:42	kernasi	great channel so far :P
23:28:02	Cykacykacyka	def-, this is not exactly what I need. I don't intend to app apps in nim, but rather 'traditional/linux-like' daemon that would do things with a socket and stuff.
23:28:09	*	pregressive quit (Ping timeout: 256 seconds)
23:28:22	*	jsudlow joined #nim
23:31:01	def-	Cykacykacyka: sorry, i misread. sounds like a case for cross compilation: http://nim-lang.org/nimc.html#cross-compilation
23:35:14	filwit	BlaXpirit: I found a different (probably better) way which uses RenderTarget inside the Drawable type-class: https://gist.github.com/PhilipWitte/8935a52bcf945e7a878b
23:36:42	filwit	BlaXpirit: however, that way is slightly different. It uses the type-classes slightly differently (in a way i'm not as familiar with) and avoids one of the procs completely.
23:37:16	filwit	BlaXpirit: not sure it will really be applicable to your real code
23:41:28	*	lavender joined #nim
23:42:45	*	pregress_ quit (Ping timeout: 252 seconds)
23:45:04	*	pregressive joined #nim
23:45:37	filwit	oh damn he left awhile ago...
23:47:38	*	rapind joined #nim
23:57:34	*	brson quit (Quit: Lost terminal)
23:57:44	*	brson joined #nim
23:59:46	Cykacykacyka	def-, actually, I shouldn't have jumped to conclusions, I don't have to use jni, but everything up to it is very well documented and fits my use case, thank you very much.