<< 10-02-2014 >>

---

00:00:29	filwit	Araq: okay great. so my thinking was pretty accurate. I can hack fix it myself, so it's not a big deal.
00:01:12	Araq	I will tell you the invariant in the vm when you document it afterwards. deal?
00:02:11	filwit	i'm not sure exactly what you mean, but I don't mine documenting things as I learn them
00:02:28	Araq	good, will tell you tomorrow, I need to sleep now
00:02:41	filwit	okay, night
00:07:36	*	darkf joined #nimrod
00:25:47	*	xenagi joined #nimrod
00:42:40	*	ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
01:01:25	Demos	BitPuffin, ping
01:14:06	BitPuffin	Demos: pong?
01:14:24	Demos	linagl is your library right?
01:14:30	BitPuffin	hell ye
01:14:47	BitPuffin	missing somoething?
01:14:52	Demos	I am trying to write a linear algebra library (because why not) and am running into issues with static[T] not existing
01:15:21	BitPuffin	Demos: why not? Because linagl :P
01:15:23	Demos	well I want to have a type param that would make the matrix dynamic sized, but I am not sure how to do that without compile time constant values
01:16:00	BitPuffin	Demos: right now unless I have missed something you can only pass ranges as parameters to generics
01:16:15	BitPuffin	so that's why linagl atm does that
01:17:02	BitPuffin	so you have to do type TMat5 = TMatrix[float32, range[0..4], range[0..4] instead of TMatrix[float32, 5, 5]
01:17:37	BitPuffin	Demos: but you should be using linagl instead :) it is in the public domain and everything
01:20:41	BitPuffin	Demos: so you could even use it for your secret illegal drug dealer projects
01:25:12	*	Demos_ joined #nimrod
01:25:18	Demos_	sorry, I got disconnected
01:25:47	Demos_	so it is just really stupid that you can only pass ranges as generic params
01:26:35	BitPuffin	Demos_: check the logs if you missed what I said
01:27:17	Demos_	yeah you can use ranges, but static[int] is a documented feature that does not work...
01:27:36	BitPuffin	Demos_: I don't even know what it is supposed to do
01:27:57	Demos_	you can have template params that are compile-time constant values
01:28:06	*	Demos quit (Ping timeout: 245 seconds)
01:28:13	BitPuffin	hmm
01:29:23	BitPuffin	Demos_: docs link?
01:29:35	Demos_	http://build.nimrod-lang.org/docs/manual.html#special-types
01:29:40	Demos_	note that none of that code works
01:33:29	*	ics joined #nimrod
01:46:41	Demos_	using ranges is a fairly bad solution. I also can not get code like type Foo[T] = object, proc DoIt[Foo](): int = stuff with Foo.T
01:53:32	Demos_	BitPuffin, Also why are vector and matrix seperate types?
02:08:03	*	EXetoC quit (Quit: WeeChat 0.4.2)
02:25:35	*	askatasuna joined #nimrod
02:32:51	Discoloda	what arguments can a generic take?
02:33:13	*	askatasuna quit (Ping timeout: 272 seconds)
02:33:58	*	dmac quit (Ping timeout: 246 seconds)
02:40:16	*	BitPuffin quit (Ping timeout: 246 seconds)
03:12:16	Demos_	types that do not cause the compiler to break when you try and pass them to a generic
03:13:16	*	BitPuffin joined #nimrod
03:34:14	*	aftersha_ joined #nimrod
03:38:14	*	BitPuffin quit (Read error: Connection reset by peer)
03:38:52	*	BitPuffin joined #nimrod
03:44:27	*	carum joined #nimrod
03:45:22	*	brson quit (Ping timeout: 246 seconds)
03:47:20	Demos_	ping BitPuffin
03:47:20	*	brson joined #nimrod
03:48:50	Demos_	linagl does not work very well... or at all
03:51:13	*	aftersha_ quit (Quit: Computer has gone to sleep.)
03:59:21	*	BitPuffin quit (Ping timeout: 245 seconds)
04:01:09	*	xenagi quit (Quit: Leaving)
04:14:52	*	carum quit (Remote host closed the connection)
04:20:13	*	BitPuffin joined #nimrod
04:22:10	*	ics quit (Ping timeout: 260 seconds)
04:24:49	*	ics joined #nimrod
04:32:32	*	BitPuffin quit (Ping timeout: 253 seconds)
04:41:48	Demos_	cd ..
04:42:43	*	carum joined #nimrod
04:43:43	*	carum quit (Remote host closed the connection)
05:08:54	*	micklat joined #nimrod
05:10:49	*	micklat quit (Remote host closed the connection)
05:13:18	*	dmac joined #nimrod
05:28:17	*	brson quit (Ping timeout: 246 seconds)
05:28:23	*	brson joined #nimrod
05:33:47	*	Demos_ quit (Quit: Leaving)
05:36:30	*	isenmann joined #nimrod
06:08:21	*	micklat joined #nimrod
06:15:00	*	ddl_smurf quit (Quit: ddl_smurf)
06:15:18	*	DAddYE joined #nimrod
06:17:57	*	brson quit (Read error: Operation timed out)
06:20:41	*	ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
06:25:37	*	ics joined #nimrod
06:42:36	*	demizer joined #nimrod
06:46:04	*	demizer quit (Quit: WeeChat 0.4.2)
06:46:51	*	demizer joined #nimrod
07:06:43	*	aftersha_ joined #nimrod
07:09:01	*	filwit quit (Quit: Leaving)
07:24:04	micklat	I want to submit a pull request. How do I run the test suite?
07:34:26	*	athaudia quit (Ping timeout: 265 seconds)
07:41:12	*	aftersha_ quit (Quit: Textual IRC Client: www.textualapp.com)
07:41:29	*	aftersha_ joined #nimrod
07:41:55	*	aftersha_ quit (Client Quit)
07:56:51	*	psquid quit (Ping timeout: 272 seconds)
07:58:00	Araq	micklat: "koch tests" does the trick
07:58:24	Araq	but currently many tests fail, I'm working hard to make them green or disable them
07:59:03	Araq	bbl
08:39:54	*	aftersha_ joined #nimrod
08:47:11	*	aftersha_ quit (Quit: Computer has gone to sleep.)
09:19:07	*	skrylar joined #nimrod
09:19:23	skrylar	meep
09:30:13	*	athaudia joined #nimrod
09:30:48	*	DAddYE quit (Remote host closed the connection)
09:32:05	*	CarpNet joined #nimrod
09:37:13	skrylar	i will say that when i saw nimrod compiled to C originally i passed over it for Rust, since haXe's C++ target was dog slow. After seeing how fast it actually runs despite using GCC i am impressed
09:46:25	*	dmac quit (Ping timeout: 272 seconds)
10:00:11	*	micklat quit (Remote host closed the connection)
10:03:34	*	radsoc joined #nimrod
10:29:57	*	BitPuffin joined #nimrod
10:30:05	BitPuffin	pong demizer
10:30:09	BitPuffin	Demos_
10:30:15	BitPuffin	seems like he isn't here though
10:30:19	BitPuffin	would like to know what is wrong
10:30:28	skrylar	its a pretty quiet night
10:30:30	BitPuffin	linagl has probably been affected by regressions or something
10:30:49	BitPuffin	I noticed that some bug I had posted was re-opened on github
10:30:55	skrylar	do you know the last version your code worked on?
10:31:14	BitPuffin	skrylar: well no since it was a git version
10:31:26	BitPuffin	0.9.3 doensn't really help because there is a million 0.9.3's :P
10:31:32	skrylar	hrm
10:31:48	skrylar	yeah, i should probably poke about that. Rust encodes the git revision in to -v
10:32:10	skrylar	I was going to suggest using bisect to find the regression, but that requires you know what hash was good
10:32:21	BitPuffin	skrylar: I have never seen you here before btw, so hey! :)
10:32:29	skrylar	BitPuffin: 'elloes
10:32:33	BitPuffin	anyway
10:32:39	BitPuffin	He just said that linagl doesn't work
10:32:43	BitPuffin	he didn't say what didn't
10:32:48	BitPuffin	kind of unhelpful
10:32:50	skrylar	:/
10:32:53	*	ics quit (Ping timeout: 252 seconds)
10:32:55	BitPuffin	I haven't touched it in a while though
10:33:04	BitPuffin	But I could do some work on it after work today
10:33:24	BitPuffin	let's see if he created a bug report
10:34:41	BitPuffin	nope no bug report
10:34:50	BitPuffin	god damn it Demos!
10:35:25	skrylar	no bug report, no bug :o
10:35:32	*	ics joined #nimrod
10:36:13	BitPuffin	and I just tried compiling and running the tests
10:36:15	BitPuffin	no errors
10:36:21	BitPuffin	so kind of hard to know what he's talking about :(
10:36:34	BitPuffin	unless he's like using 0.9.2 or something
10:36:38	BitPuffin	because that is not supported
10:36:53	skrylar	yay releases-that-youre-not-supposed-to-use synrome \o/
10:37:03	BitPuffin	\o/
10:37:13	BitPuffin	yeah hopefully Araq releases 0.9.4 soon
10:37:15	skrylar	rust does that too. i think its really silly
10:37:31	BitPuffin	although if generics aren't fixed by then I'm gonna have to require 0.9.5 pretty quickly to use linagl
10:37:45	skrylar	whats broken about generics right now?
10:38:25	BitPuffin	uh well for example, you can't pass anything but types and ranges as parameters to them right now
10:38:56	BitPuffin	and when they aren't types you can't use them in the return type
10:39:19	skrylar	i thought thats what the generic system was for
10:39:32	skrylar	i thought if it wasn't a type, you were supposed to use a template
10:39:52	BitPuffin	skrylar: njaeeaeahe
10:39:59	BitPuffin	anyways take this as an example
10:40:04	BitPuffin	this is a workaround that uses ranges
10:40:10	BitPuffin	and it still doesn't work
10:40:12	BitPuffin	proc sub*[T; R, C: range](m: TMatrix[T, R, C], therow, thecol: TInteger): TMatrix[T, R.low..R.high-2, C.low..C.high-2] {.noSideEffect.} =
10:40:38	BitPuffin	this is for getting a submatrix of a matrix of R*C dimensions
10:42:13	BitPuffin	skrylar: so as you can see there is no way for me to say that "this returns a matrix that is one R-1*C-1 dimensions
10:42:35	*	skrylar doesn't know math ._.
10:42:56	BitPuffin	skrylar: okay well a two dimensional array then
10:43:07	skrylar	i'm looking at the line
10:43:54	skrylar	I think the range part in your return definition is doing it
10:44:11	skrylar	is one allowed to do expressions when declaring a generic type?
10:44:23	BitPuffin	skrylar: basically what it is trying to say is that if you have an array (just gonna use C ish syntax for fun) that is [4][4] that you pass in, you will get a [3][3] array back
10:44:41	BitPuffin	skrylar: you are supposed to be allowed to do it but it doesn't work right now
10:44:57	skrylar	I'd probably templatize it then
10:45:15	skrylar	I have one section of code that uses templates and 'emit' just because the pragma syntax is derpy
10:45:31	skrylar	e.g. proc foo(...) # works
10:45:45	skrylar	proc foo(...) {.importc: bar.} # nope
10:47:05	BitPuffin	odd
10:47:25	skrylar	i also had bad issues where if i try to use 'macro' at all, the compiler segfaults
10:47:59	BitPuffin	skrylar: have you talked to Araq about it?
10:48:10	BitPuffin	and checked if it was reported and if not reported it? :P
10:48:12	skrylar	BitPuffin: this is the first time i've been to the IRC, tonight
10:48:21	BitPuffin	ah I see
10:48:31	skrylar	i poked in to nitpick about unicode
10:48:35	BitPuffin	well Araq is our leader so it's him you should pester about these things
10:49:15	skrylar	I noticed 'char' was 8-bit and that the 'unicode' module operates on runes, which is incorrect; i'll have to poke/patch that at some point
10:49:37	BitPuffin	incorrect?
10:49:39	skrylar	its a common mistake, but 'graphemes' are the correct unit to iterate by when dealing with a unicode string, not runes
10:50:04	BitPuffin	ah
10:50:13	BitPuffin	well I don't know super much about encodings
10:50:16	BitPuffin	just a little
10:50:32	skrylar	BitPuffin: basically according to the TRs, a 'code point' is the actual numeric value for a single unicode thingy; though a lot of people call those 'runes'
10:51:05	BitPuffin	TR?
10:51:06	skrylar	but say, an umlauted A
10:51:15	BitPuffin	umlauted? :P
10:51:17	skrylar	BitPuffin: unicode technical report
10:51:20	BitPuffin	ah
10:51:44	skrylar	if you put an A with an umlaut in the stream, it can be a special "A-with-umlaut" code point or "A" followed by "combining diacritic: umlaut"
10:52:04	skrylar	these are both valid unicode
10:52:50	skrylar	the pain in the ass is that there is no 'actual' limit to how many diacritic marks can be stuffed on a single rune, which means you can have a single letter and 5,000 symbols stuck to it; though the stream safe subset only allows 30
10:53:41	BitPuffin	lol
10:54:08	BitPuffin	ah umlaut is like åäö?
10:54:14	skrylar	yeah
10:54:27	BitPuffin	or maybe without å
10:54:36	skrylar	I wrote a unicode lib in C99 last year, it was painful.
10:55:06	BitPuffin	well not surprising you know so much about that then :P
10:55:14	skrylar	heh
10:55:26	skrylar	Its amazing how much of a pain it truly is to get text on the screen :)
10:57:08	BitPuffin	hehe
10:57:11	*	askatasuna joined #nimrod
10:57:16	BitPuffin	That's what happens
10:57:58	skrylar	i'm still surprised by the binary size for nimrod so far
10:57:59	BitPuffin	it's funny how you think those things are easy
10:58:14	BitPuffin	but it's because it is solved
10:58:22	skrylar	its easy until you want something special
10:58:35	skrylar	i think thats why there are no good outliners for windows, really
10:59:07	skrylar	on the mac you can just ask the OS to do most of this for you; on Windows the instant you want more than a stock control you're back to learning about font metrics and gap buffers
11:01:20	skrylar	there's also horrible things like what Unity or Love2D do
11:01:46	BitPuffin	what do they do?
11:01:52	skrylar	Love2D just goes through the string every frame and asks SDL to render it by character; Unity creates and GC's every widget on the screen per frame
11:02:10	BitPuffin	HAHA
11:02:12	BitPuffin	what
11:02:13	skrylar	its no wonder that on a 1000$ engine, people are still selling GUI replacements
11:02:21	skrylar	because the GUI is just godawful
11:03:36	skrylar	Unity is really no better than OGRE with a pre-made level editor and Mono/C# bindings ^^;
11:04:02	BitPuffin	so you are saying use NeoAxis?
11:04:20	BitPuffin	It is "better" though
11:04:26	BitPuffin	since ogre is a rendering engine
11:04:27	skrylar	I wouldn't actually pay for an engine o.o
11:04:32	BitPuffin	unity also comes with physics and shit
11:04:37	BitPuffin	and I know
11:04:39	BitPuffin	there are plugins
11:04:45	BitPuffin	anyways I hate premade engines anyway
11:04:58	skrylar	There is one and only one pre-made engine I respect, and that is the Unreal engine
11:05:11	skrylar	Because people can punch that thing in to doing pretty much anything they want it to
11:05:17	BitPuffin	lol
11:05:27	BitPuffin	do you know what a god awful codebase UE 3 has?
11:05:33	BitPuffin	I've heard terrible horror stories
11:05:38	BitPuffin	but basically to sum it up
11:05:52	BitPuffin	the UE 3 games you see on the market, such as Bioshock Infinite etc
11:05:58	skrylar	Torque is.. meh, I don't like it. I actually bought a license before they fully opened it but you have to patch the C++ side to do anything interesting, and TorqueScript is pretty hilarious
11:05:58	BitPuffin	are NOT even running UE 3
11:06:06	skrylar	yeah everyone hates 3
11:06:13	skrylar	ubisoft still uses 2.5
11:06:40	BitPuffin	it's like, something that has been patched and orked around and all that stuff until it barely works but is yet better than the original
11:06:50	BitPuffin	oh so you meant UE 2.x?
11:06:57	skrylar	sounds like gamebryo lol
11:07:05	skrylar	gamebryo is just terrible
11:07:16	skrylar	it sounds like a great idea and you see Civ + TES games use it
11:07:38	skrylar	then you find out that Civ had to bugfix the hell out of the engine to fix the terrible performance leaks
11:07:55	skrylar	and Bethesda has been monkeypatching it until Skyrim where they basically rewrote everything
11:08:30	skrylar	all though i still find Source to be the most embarassing engine IMO
11:08:38	BitPuffin	Why?
11:08:47	BitPuffin	I'd say that source is probably one of the most robust on the market
11:08:47	skrylar	Its a jank Quake 2 engine
11:09:09	skrylar	No scripting layer, only supports BSP maps, does not handle large maps with any grace
11:09:29	BitPuffin	it has lua and squirrel scripting?
11:09:46	skrylar	it didn't the last time i knew people who tried to use it
11:09:54	BitPuffin	I guess it's not built to be a silver bullet (nothing can be)
11:10:04	BitPuffin	which I think is fine
11:10:17	BitPuffin	in fact I think that's the whole problem with premade engines
11:10:22	BitPuffin	they are built to solve everything
11:10:25	skrylar	https://developer.valvesoftware.com/wiki/Source_Engine_Features no scripting layer
11:10:28	BitPuffin	which makes them a mediocre solution for everything
11:11:10	skrylar	well that was why the unreal engine used to be good
11:11:12	BitPuffin	https://developer.valvesoftware.com/wiki/VScript
11:11:22	skrylar	it was a chunk of modules glued together by unrealscript
11:12:37	skrylar	heh squirrel and gamemonkey. i remember those
11:13:24	skrylar	Sadly LuaJit is pretty much the reason non-lua scripting languages don't matter for games; at least in my opinion
11:14:38	skrylar	the amount of speed that thing punches out of lua is absurd
11:15:41	BitPuffin	well
11:16:39	BitPuffin	I might argue that you shouldn't use a scripting lang for the stuff that needs absurd perf
11:17:10	skrylar	maybe
11:17:19	skrylar	though i don't like regular code being slow for no reason
11:17:37	BitPuffin	*cough* nimrod
11:17:52	skrylar	i haven't had speed issues yet ._.
11:18:02	BitPuffin	that's not what I meant
11:18:21	BitPuffin	I meant that nimrod is pretty much as pretty as a scripting language but still fast as a systems programming language :
11:18:23	BitPuffin	:)
11:18:38	skrylar	well, i didn't use it at first because GC and compiles to C
11:18:53	BitPuffin	you don't have to use CG
11:18:55	BitPuffin	GC*
11:18:57	skrylar	mandatory GC makes me 'ew' and i used haxe which compiles to C++ and is dog slow
11:19:23	skrylar	a simple roguelike in haxe took like 2-15 seconds to build
11:19:23	BitPuffin	and compiles to C is a nice feature because it means you can run on places that don't support nimrod (ie, every console on earth)
11:19:40	skrylar	it helps that its *straight* C
11:19:59	skrylar	regular C compiles pretty fast if you aren't a derp with 50mb of headers to trawl
11:19:59	*	aftersha_ joined #nimrod
11:20:02	BitPuffin	skrylar: anyways, we have optional GC (I also hate mandatory GC with a passion) and the compilation speed is pretty decent!
11:20:31	skrylar	BitPuffin: yeah, i saw that the GC was competent
11:20:34	BitPuffin	skrylar: yes and I believe nimrod also does some voodoo with compiling it on multiple threads etc
11:20:53	BitPuffin	or like
11:21:06	BitPuffin	It compiles the C while it generates the next C for it to compile
11:21:10	skrylar	nimrod doesn't emit headers for things that don't need it, so GCC isn't trawling megabytes of useless data
11:21:13	BitPuffin	so it's utilizing resources well
11:21:30	skrylar	I used to use a program 'makeheaders' for my straight C which did the same thing
11:21:49	skrylar	scanned your C files and made minimized headers, so 1) no manual maintenence and 2) much faster compiling even without parallel builds
11:22:01	BitPuffin	skrylar: anyway, there is no reason why we couldn't compile straight to machine code (you are in fact welome to implement it if you so please :P)
11:22:14	skrylar	BitPuffin: nah, i don't mind since its very quick
11:22:17	BitPuffin	we just use C because we can utilize optimizers and shit
11:22:30	skrylar	the full build of the compiler only took ~10 seconds IIRC
11:22:40	BitPuffin	yup
11:22:50	BitPuffin	and that's the biggest nimrod source there is more or less
11:23:02	skrylar	i don't mind 10 seconds for a massive code base, especially considering you don't usually *need* to full build all the time
11:23:06	skrylar	:)
11:23:45	BitPuffin	yep
11:23:47	BitPuffin	build in caching
11:23:49	BitPuffin	luvit
11:24:01	skrylar	ccache + lazy compilation :3
11:24:08	BitPuffin	we also *can* compile to C++, objective-c, js
11:24:11	skrylar	though i don't think nimrod is lazy compiling on my machine right now
11:24:18	*	aftersha_ quit (Client Quit)
11:24:35	skrylar	it might be, but it hasn't been slow enough for me to get concerned yet
11:24:46	skrylar	0.2s on average to rebuild my test code so i'm not worried, lol
11:25:15	BitPuffin	:)
11:25:47	skrylar	i might be interested in trying to figure out how to shield Ogre builds in nimrod at some point
11:26:01	skrylar	i think using the dynlib will keep link times down
11:26:55	skrylar	my least favorite part about ogre is that it brings in a few hundred megabytes of templatized C++ sadness which means simple code will take ridiculous amounts of time to compile
11:27:06	skrylar	which doesn't exist in say, Horde3D, because all of that has been shielded
11:28:35	skrylar	BitPuffin: though i do like pascal types :)
11:37:43	skrylar	BitPuffin: huh. i just wondered if templates and tags could be used to generate shaders
11:38:24	skrylar	e.g. use tags to limit it so only 'shader-able' methods from a module could be called in a block, and then a macro that rewrote the actual code to GLSL
11:42:52	*	EXetoC joined #nimrod
11:54:27	BitPuffin	skrylar: of course it could
11:54:48	BitPuffin	in fact I am planning to do something like that
11:55:13	*	aftersha_ joined #nimrod
11:56:14	EXetoC	wot
11:58:10	BitPuffin	EXetoC: sup
12:13:37	skrylar	generics are <3
12:14:09	skrylar	i rewrote my rectangle math code over from Rust, and the total LOC is way down mostly because the generics are less derpy
12:16:47	skrylar	BitPuffin: is there a nimrod equivalent to "maybe returns" already?
12:19:34	BitPuffin	skrylar: maybe returns?
12:22:56	skrylar	BitPuffin: in rust they have a generic maybe type, somewhat haskell-ish but its basically a variant type with "Some X" and "None"
12:23:07	skrylar	So you can return Maybe<Rectangle>
12:24:16	EXetoC	it returns either way in normal circumstances, but anyway, there's variants
12:25:01	skrylar	EXetoC: the function always returns a value; its just a special variant that replaces 'nil' in case the encased type is a non-nillable pointer
12:25:28	skrylar	easy enough to write a Nimrod version of the same, I just wasn't sure if that was already a thing
12:25:51	EXetoC	I was just nitpicking
12:25:53	EXetoC	https://github.com/fowlmouth/nimlibs/blob/master/fowltek/maybe_t.nim
12:26:03	EXetoC	http://build.nimrod-lang.org/docs/manual.html#object-variants
12:26:19	skrylar	alright
12:26:33	skrylar	i was mostly just curious if there was already a convention for that, before i started doing anything with it
12:26:36	skrylar	i knew how to write it
12:26:40	EXetoC	it's slightly less useful, or maybe it's just a little clunkier, but we might get something similar soon
12:27:06	EXetoC	ok
12:28:02	skrylar	thats a weird API for it (i just read the source)
12:29:50	EXetoC	yeah?
12:31:36	skrylar	EXetoC: I guess it can't be helped all that much, though it would look better if you could do
12:31:54	skrylar	case x; of Something: blah, Nothing: blah
12:32:32	skrylar	"if X" is strange because it could be an uncertain boolean, in which case it now reads as though you were checking the boolean instead of merely the *existence* of the boolean
12:33:36	skrylar	but thats a whole hive of bees that probably doesn't need to be worried about at the moment
12:33:39	*	carum joined #nimrod
12:33:59	EXetoC	for TMaybe? if so, see the 'converter'
12:34:16	*	carum quit (Remote host closed the connection)
12:35:40	*	carum joined #nimrod
12:37:15	*	aftersha_ quit (Quit: Computer has gone to sleep.)
12:38:05	EXetoC	have a look at this if it's unclear http://build.nimrod-lang.org/docs/manual.html#convertible-relation
12:38:36	skrylar	I understand how the code works and I was mostly just nitpicking the syntax for using that TMaybe
12:39:04	skrylar	I do appreciate the help though
12:39:47	EXetoC	right. well, I wouldn't mind referencing 'has' explicitly tbh
12:40:49	skrylar	i suppose the good thing about nimrod is that with templates there can be a few dozen ways to use something like a TMaybe and they won't bloat the actual binaries like C++ would
12:41:17	*	aftersha_ joined #nimrod
12:49:54	BitPuffin	I should probably read through the build version of tfm
12:50:01	BitPuffin	so that I am up to date with latest features
12:50:26	EXetoC	ya
12:52:44	BitPuffin	would be cool if nimrod got some of the nice sugar from ruby I think
12:53:05	BitPuffin	so that you can do `a = 5 if a == 2` or whatever
12:54:09	EXetoC	if (let x=p();x==1): ?
12:54:54	skrylar	tfm?
12:55:35	skrylar	EXetoC: that incurs the need for indentation
12:56:01	EXetoC	I don't think so
12:56:08	skrylar	BitPuffin: i think that would pretty much be served by an 'iif' method/macro i think
12:56:24	EXetoC	if (let x=p();x==1): echo "yup"
12:56:38	skrylar	i think VB and some Pascals had it so it was like that; x = iif(a == 2, 5, 0)
12:56:39	BitPuffin	skrylar: tfm = the fucking manual
12:56:44	skrylar	BitPuffin: oh. :P
12:56:52	BitPuffin	is in rtfm :P
12:56:57	BitPuffin	as
12:58:21	EXetoC	let x = (discard 3; 4); assert x == 4
12:58:28	skrylar	i opened the manual page for the gtk2 unit; it made my scrollbar sad
12:58:32	EXetoC	except you need to replace the ; with a newline
12:58:54	BitPuffin	I don't see how what you are typing is addressing what I said lol
12:58:55	EXetoC	it'd be nice if that could be avoided, for the sake of IRC mainly :>
12:59:12	EXetoC	BitPuffin: then what does that do?
12:59:23	BitPuffin	oh you just mean what it would do?
12:59:56	BitPuffin	well I gues syou would do `a = 5 if a == 4 else 3`
12:59:59	BitPuffin	or something like that
13:00:05	BitPuffin	can't remember exactly how ruby does it
13:00:13	BitPuffin	or maybe that it just keeps its value
13:00:15	BitPuffin	yeah probably that
13:00:41	EXetoC	let x = if y: z else: w
13:00:51	EXetoC	no?
13:01:10	EXetoC	if not, then you'll have to translate to Nimrod, because I don't know Ruby c(:)
13:01:50	BitPuffin	EXetoC: yeah i know that you can do that in fact I believe I do it quite a few times
13:02:00	BitPuffin	it's just harder to follow in certain cases
13:02:13	BitPuffin	can't remember where it was annoying though
13:03:22	skrylar	heh, i didn't know you could
13:03:34	skrylar	I just re-did code earlier to use abs/min/max instead
13:03:36	EXetoC	it's pretty clear when you break it into multiple lines. I usually expand it to four, but I suppose three lines is pretty readable too
13:04:24	BitPuffin	EXetoC: yeah but the ruby way is readable on one line :P
13:08:16	EXetoC	oh well
13:19:23	*	carum quit (Remote host closed the connection)
13:25:55	*	aftersha_ quit (Quit: Computer has gone to sleep.)
13:34:01	*	darkf quit (Quit: Leaving)
13:55:22	*	Araq_ joined #nimrod
13:55:50	Araq_	hi skrylar welcome
13:56:07	Araq_	hi demizer welcome
14:03:55	*	radsoc quit (Quit: Leaving)
14:07:09	*	radsoc joined #nimrod
14:12:28	*	[1]Endy joined #nimrod
14:13:59	BitPuffin	Araq_: you don't by any chance have any idea of what demos meant when he said that linagl doesn't seem to work at all?
14:14:19	BitPuffin	because it works here
14:20:59	Araq_	BitPuffin: no but even I would, I wouldn't tell you
14:21:19	Araq_	because you don't like nimrod's if-expressions :P
14:22:42	OrionPK	oh dear
14:22:51	BitPuffin	Araq_: oh snap diddely
14:23:11	BitPuffin	Araq_: I do like them, I just think the different variations are more readable in certain situation
14:23:13	EXetoC	he said it wasn't ideal, because of the limitations that you too had run into
14:23:13	BitPuffin	s
14:23:26	EXetoC	did he also say that it didn't work at all?
14:24:40	Araq_	BitPuffin: can't agree, sorry, if then else is the standard order
14:24:52	Araq_	and python's is weird
14:25:11	*	OrionPK quit (Remote host closed the connection)
14:25:23	EXetoC	yes, it's consistent
14:25:25	*	OrionPK joined #nimrod
14:25:30	Discoloda	b?t:f is the only logical pick expression
14:25:34	BitPuffin	EXetoC: well those limitations will be removed when the compiler is fixed, so it's not really an issue with linagl
14:25:34	EXetoC	weird indeed
14:25:38	BitPuffin	EXetoC: and yes he did
14:26:00	BitPuffin	Demos_ | linagl does not work very well... or at all
14:28:29	BitPuffin	I am a bit confused why he doesn't understand why matrices and vectors are different types, one is a 2d array and one isn't
14:28:43	EXetoC	vector.nim doesn't work for me: Error: internal error: typeRel: tyNon
14:29:09	Araq_	that's a regression that I cause in an attempt to make the compiler more stable ...
14:29:10	EXetoC	*tyNone
14:29:15	Araq_	*that I caused
14:29:25	EXetoC	fair enough
14:29:32	BitPuffin	https://gist.github.com/BitPuffin/8916869
14:29:38	BitPuffin	works here
14:29:43	BitPuffin	I updated the compiler yesterday
14:30:36	EXetoC	I updated 1-2 days ago
14:30:38	BitPuffin	EXetoC: do you have the latest version of the compajler
14:32:28	EXetoC	I don't know if I remembered to pull, but I'll try again tonight
14:32:49	BitPuffin	EXetoC: could possibly be that the version that babel pulls isn't the head version
14:33:09	EXetoC	I cloned the repository
14:34:26	BitPuffin	well then there we go
14:34:29	BitPuffin	should work
14:36:43	EXetoC	we'll see
14:37:18	BitPuffin	EXetoC: why can't you try now? :P
14:42:25	*	Araq_ quit (Quit: ChatZilla 0.9.90.1 [Firefox 27.0/20140127194636])
14:48:59	*	ddl_smurf joined #nimrod
14:54:03	*	aftersha_ joined #nimrod
14:59:51	*	aftersha_ quit (Quit: Computer has gone to sleep.)
15:09:33	skrylar	hmm..
15:09:43	skrylar	Fossil seems neat, but I like my git-rebase.
15:12:44	*	aftersha_ joined #nimrod
15:14:52	*	Demos joined #nimrod
15:16:28	Varriount	Hello honey badgers!
15:16:39	Demos	hello Varriount
15:16:50	Demos	fixed your sleep scedule I see
15:17:36	BitPuffin	Demos: whatchu talkin bout linagl not working
15:18:25	Demos	BitPuffin: I just sent you a memo, initially it was because babel had the wrong version, then there are issues with the side effect checker
15:18:52	BitPuffin	Demos: I haven't seen any issues with the tests I have written at least
15:18:53	BitPuffin	hrm
15:18:55	Demos	also [0,0,0] and [0.0,0.0,0.0] do not seem to convert to array[float32, 0..2], which is kinda stupid
15:19:17	Demos	I was also having the mag function fail the side effects check
15:19:21	Demos	I am on 0.9.3 master
15:19:58	BitPuffin	Demos: Err I don't think you should mix int vectors with float vectors really
15:20:07	BitPuffin	Unless you do so explicitly
15:20:51	Demos	BitPuffin: right. But when initializeing stuff having that conversion is nice. It is not something you can fix, would probably need to be special cased in the language (do not want conversions happening everywhere, only on initialization)
15:20:53	BitPuffin	Demos: right it fails because I have to use C sqrt
15:21:07	Demos	does C sqrt have side effects?
15:21:22	BitPuffin	Demos: no but nimrod can't verify that calling C doesn't mean side effects
15:21:40	BitPuffin	so yeah that's a minor issue
15:21:57	BitPuffin	Demos: anyway most of the warts (such as temporarily having to use ranges) is limitations of the compiler atm
15:22:03	BitPuffin	that's why I have pre-defined types
15:22:03	Demos	well it makes normalize and mag just not work. Should remove the noSideEffect pragma on them until it does
15:22:27	BitPuffin	so that it won't break peoples code who doesn't do anything out of the ordinary
15:22:34	BitPuffin	Demos: yeah absolutely, that's just a miss on my end
15:24:48	BitPuffin	Demos: mag is currently a much bigger pain in the ass than in should be because we only have sqrt for floats
15:25:08	BitPuffin	but it's not a biggie
15:25:20	BitPuffin	It gives you what you expect now I guess :P
15:25:49	Demos	frankly I think there should be a "noSideEffectsJustTrustMe" pragma. I think haskell has something like that for C functions
15:26:14	BitPuffin	oh and it only works for "float" and not "float32" because I can't do TReal stuff on it
15:26:17	BitPuffin	Demos: agreed
15:26:33	Varriount	noReallyThisHasNoSideEffectsEvenThoughItSeemsLikeItDoesNoReally
15:27:07	Demos	{.noSideEffectsErrornosCanDieInAFire.}
15:28:39	Demos	does float accept floatXXs as well, I seem to recall some issue relateing to that when I used linagl on the devel branch
15:29:01	Demos	I thought "oh float must be a typeclass, oh wait that is dumb"
15:32:32	BitPuffin	Demos: there was some issues with typeclasses
15:32:38	BitPuffin	and no float is float
15:32:46	BitPuffin	TReal includes float float32 float64 etc
15:32:59	Varriount	Anyone know if the semantics for "instanciating" a new closure iterator has changed at all?
15:34:48	BitPuffin	Demos: anyways mag is fixed now
15:35:27	BitPuffin	Demos: anyways it is really useful if you use linagl, because it helps me find bugs and also makes it so that there is one more person who can complain about bugs holding it back in the compiler :)
15:36:06	Varriount	Also, any ideas on how to iterate through a closure iterator that has args, when given as an argument to another closure iterator?
15:38:19	Demos	BitPuffin: allright, I may submit a pr for some accessor functions (although having a function called "x" is bound to clash with /something/
15:40:32	BitPuffin	Demos: accessor functions?
15:40:38	BitPuffin	Demos: you mean like vec.x
15:40:41	BitPuffin	don't do that
15:40:46	BitPuffin	swizzling is already implemented
15:40:53	BitPuffin	but it's currently broken
15:40:57	BitPuffin	because bugs in the compiler
15:41:27	BitPuffin	Demos: but in theory you should be able to do vec.xyxyzyzyxgzyzxyzgrgbgbrgbrbgrbaststspqpqqsptsq
15:42:41	Demos	what I saw was vec.{"xyzzzwagggg"}
15:43:11	Demos	I just saw zah say he was fixing deligators tonight, so hope remains
15:43:17	*	aftersha_ quit (Quit: Computer has gone to sleep.)
15:45:40	Varriount	Hm. I wonder if there's any way we could get funding for nimrod.
15:46:09	BitPuffin	Demos: yeah initially in was that but then when zahary added delegators I made it into v.ecxhrexhoa
15:46:34	BitPuffin	Demos: may have removed that though
15:46:45	BitPuffin	it's really just a matter of changing the name etc
15:47:12	BitPuffin	Demos: as you can see the {"xyxyxyxy"} thing is deprecated
15:47:18	Demos	what is a matter of changeing the name? deligators or getting funding?
15:47:30	BitPuffin	because it is preferred to not use it since it will be removed in favor of x.xz
15:47:41	BitPuffin	Demos: delegators
15:47:57	Demos	yeah I really like the .xxxyzwang thing, much better than the way glm does it (macros for each combination)
15:48:05	BitPuffin	Demos: yup
15:48:11	BitPuffin	bbl
15:53:22	*	[2]Endy joined #nimrod
15:56:44	*	[1]Endy quit (Ping timeout: 245 seconds)
16:09:06	skrylar	:|
16:10:05	skrylar	well that was a subtle bug: forgot the type specifier to the last parameter in a function, Nimrod didn't find this suspicious at all (i wonder what type it thought it was? or did it just use the same type as the parameter before it?) and so the reason my code wasn't drawing was ... because it wasn't actually calling the right method, not the cliprect being broken. *derps self*
16:11:14	Varriount	skrylar: If procedure parameters aren't given a type, they default to the "expr" type
16:12:38	Demos	Varriount: really, auto would be more sane
16:12:49	EXetoC	"auto* = expr"
16:12:59	Demos	really now.?...?
16:13:00	Varriount	Demos: The 'auto' type is the 'expr' type
16:13:05	*	Discoloda left #nimrod (#nimrod)
16:13:05	Demos	neat
16:13:16	*	skrylar wouldn't mind a compiler hint about that
16:13:21	skrylar	if nothing else it lets you know it did it
16:13:28	*	Varriount agrees with skrylar
16:13:32	Demos	should be a warning
16:13:54	skrylar	such would have saved me about 45 minutes of going "wut?"
16:13:57	Varriount	Actually, such a feature might not be too hard to implement.
16:14:01	EXetoC	you can already do p(x, y: T) so it's a source of confusion
16:14:06	skrylar	Fun times with ncurses: Can't use debugEcho because curses is eating the buffer
16:14:14	Varriount	:3
16:14:31	EXetoC	p(x, y: T) -> p(x: T, y: T)
16:14:49	Varriount	I have no sympathy, ncurses doesn't work on windows without cygwin
16:15:00	Demos	I hear ncurses is a big ball of wtf
16:15:06	skrylar	Varriount: http://pdcurses.sourceforge.net/ *whistles*
16:15:08	Varriount	^ That too
16:15:21	Varriount	:O
16:15:27	EXetoC	termbox?
16:15:34	skrylar	termbox is nix only
16:15:52	Varriount	skrylar: Why aren't programs like this more well-known?
16:16:06	skrylar	Varriount: i donno. I found out about it because ToadyOne used it for Liberal Crime Squad
16:16:07	Demos	does COMMAND.COM even allow stuff like curses?
16:16:12	skrylar	Demos: yes
16:16:25	EXetoC	they don't make that obvious at all
16:16:30	skrylar	The guy who wrote termbox even mentioned it
16:17:04	skrylar	well its.. done through weird APIs, for windows
16:17:33	Demos	welcome to windows EnjoyYourStayEx
16:17:39	skrylar	And full of stupidity, like you can ask Windows for the "size" of your console but then it tells you the virtual size, and then a window size, but only kind of, so you have to calculate your *actual* buffer size
16:18:02	OrionPK	time for someone to wrap this game engine i guess https://github.com/okamstudio/godot
16:18:37	*	skrylar peeks at it
16:18:50	Demos	looks like reddit hugged that website to death
16:18:52	skrylar	the .org url is taking a small forever to open
16:19:18	OrionPK	ya
16:19:32	OrionPK	it loaded up eventually
16:20:05	OrionPK	a pure nimrod engine would be better yet, of course :P
16:20:06	*	carum joined #nimrod
16:20:20	skrylar	Port it :P
16:20:48	skrylar	maybe i'll look at this when the website is functionally usable later
16:22:32	*	skrylar is still trying to get a morphic-inspired GUI built :(
16:23:50	Varriount	In its defence, some parts of the winapi are wierd due to maintaining compatability with things like MS-Dos
16:24:14	*	carum quit (Ping timeout: 245 seconds)
16:24:37	*	Demos quit (Ping timeout: 248 seconds)
16:26:18	Varriount	Demos: I don't think COMMAND.COM actually exists anymore. It's just cmd.exe now. \nitpick
16:29:52	skrylar	Varriount: some of it is also the painful legacy of C
16:30:33	skrylar	even stuff like this http://libagar.org/ gets really derpy to use, because you have to fake all of the OO
16:31:09	EXetoC	we have a couple of upcoming engines, but it takes ages as always
16:31:38	Varriount	Hm. Interesting. Does libagar use native api's, like wxwidgets, or something like java's gui toolkits?
16:32:13	skrylar	it uses SDL
16:32:15	EXetoC	BitPuffin: have you used PA yet?
16:33:06	BitPuffin	EXetoC: PortAudio?
16:33:07	Varriount	skrylar: Araq has a repo named "Claro", which he had envisioned to be a nimrod gui toolkit
16:33:17	EXetoC	BitPuffin: yes
16:33:24	BitPuffin	EXetoC: nein
16:33:28	Varriount	However it's not currently developed, and quite.. lacking
16:33:42	EXetoC	BitPuffin: ok
16:33:47	BitPuffin	EXetoC: have you?
16:33:48	*	Demos joined #nimrod
16:33:52	skrylar	Varriount: i've been designing one based on Morphic (smalltalk) and the DOM (html)
16:34:04	Varriount	skrylar: Sounds interesting.
16:34:11	skrylar	Morphics.. basically just a scene graph really
16:34:27	skrylar	there is a 'morph' class and every widget is just a combination of those
16:34:31	Varriount	So, by DOM, you mean, seperate the data from the layout?
16:34:33	EXetoC	BitPuffin: are you on devel? I just built the latest Nimrod source, and vector.nim still generates an ICE
16:34:40	EXetoC	BitPuffin: nope
16:34:43	BitPuffin	EXetoC: I am not on devel
16:34:52	skrylar	e.g. a button is a morph with a background on push/release, a checkbox is just a text morph stuck to a button morph, that kind of thing
16:34:58	EXetoC	that ought to explain it then
16:35:13	BitPuffin	EXetoC: indeed
16:35:14	EXetoC	Demos: are you on the devel branch?
16:35:24	BitPuffin	I'm on master
16:35:24	Varriount	EXetoC: Any ideas on how generic closure iterators could be properly passed in around in nimrod?
16:35:55	BitPuffin	God damn it must the repository name be capitalized
16:36:03	BitPuffin	makes it a pain to cd into xD
16:36:08	Varriount	So that chaining things iterators like filter and map would work?
16:36:21	BitPuffin	from now on I shall change the dir name
16:36:24	EXetoC	indeed. speaking of which, I should tweak my shell a little
16:36:31	Varriount	BitPuffin: Not if you're on a case-insensitive file system. :3
16:36:34	skrylar	Varriount: kinda. I took inspiration for the actual layouting and event from the CSS box model and the event system from how the DOM passes them to JavaScript. Mostly because I started the project in Rust and it was a lesser evil to put every common message through a channel type than it was to inherit things. Dynamic dispatch over there is.. kind of a massive PITA right now.
16:36:39	BitPuffin	Varriount: fuck that shit
16:37:11	EXetoC	Varriount: I don't know any details
16:37:23	Varriount	skrylar: Pfft, dynamic dispatch. Nimrod has advanced static dispatch trees.
16:37:26	Demos	BitPuffin: merge meh pr to linagl
16:38:10	Demos	and EXetoC I switch around often, usually I am on devel or master, rarely stable
16:38:25	Demos	with meh ssd it takes <1m to reboot the whole compiler
16:38:36	skrylar	Varriount: my partial ideas are that making controls should be easy and you shouldn't have to disassemble *everything* to make a new part, the GUI should be loadable from resource files (e.g. same concept as Apple's nibs), and ideally I'd like it to be action-driven. Qt/Delphi has a thing where you can make 'actions' that are attached to toolbars and menu buttons, so those have the same button text and enable state, but then for some
16:39:27	BitPuffin	Demos: what is it about
16:39:40	Demos	I removed {.nosideeffect.} on normalize
16:39:52	skrylar	end-user customization of the GUI is really easy if you've already split off what each button should do in to little units :)
16:39:57	Varriount	skrylar: And what about native look and feel?
16:40:01	Demos	since normalized calls mag and that causes the side effect checker to reject it
16:40:06	skrylar	Varriount: i don't really care about that so much.
16:40:14	BitPuffin	yeah
16:40:42	Demos	Varriount: I think native look/feel is a lost cause. I mean Qt does it by essentially "brute artistic force" and imo tends to look more native than wxWidgets
16:40:49	skrylar	Varriount: Drawing controls isn't integral to my class tree, so if you want that you *can* do it
16:40:56	BitPuffin	Demos: I guess that change doesn't really merit that you need to say that you dedicate the work to the public domain or whatever
16:41:00	BitPuffin	a comment or something would be nice
16:41:02	BitPuffin	though
16:41:11	skrylar	Demos: it calls the windows style APIs nowadays
16:41:13	BitPuffin	I should really get like a waiver file or something
16:41:34	skrylar	there are ways of asking XP and beyond to draw the modern theming on to a GDI canvas
16:41:37	EXetoC	Demos: 65s here (SATA). not too bad
16:41:43	Demos	BitPuffin: what... I change one line. And I do not intend to keep my own fork and not send prs
16:41:49	BitPuffin	:P
16:42:07	BitPuffin	That's why I said it doesn't really merit forcing you to say that
16:42:15	BitPuffin	it is merged
16:42:24	Demos	also, afaik CC0 is not copyleft :D
16:42:32	skrylar	all though for what I want to do, I don't really gain a benefit from native L&F; I actually *like* the weird notebook skin that Notebook had on the mac, and the subdivision UI of Blender
16:42:46	BitPuffin	Demos: didn't say it was
16:42:49	BitPuffin	Demos: it is public domain
16:42:50	skrylar	but you can certainly shovel it under what i have written, since its not tightly coupled
16:43:18	EXetoC	Demos: 49s when upping the CPU freq, but <1m isn't too specific :>
16:43:32	skrylar	Varriount: hilariously the current revision draws out to {n,pd}curses, because who doesn't want an outliner that runs in PuTTY? :D
16:44:43	BitPuffin	pfff, copyleft
16:44:47	*	BitPuffin hates the gpl
16:45:16	*	Demos does not mind the gpl, would like authors to licence under LGPL with a static linking exception
16:45:23	skrylar	^
16:45:39	Demos	although one can static link and offer object files
16:45:53	skrylar	SDL moved off of LGPL lately
16:46:07	BitPuffin	no
16:46:15	BitPuffin	fuck both GPL and LGPL and everything of the like
16:46:34	BitPuffin	but yeah licensing under that is definitely preferred
16:46:40	BitPuffin	over regular pure bs GPL
16:46:49	BitPuffin	skrylar: wasn't that a long long time ago? :P
16:46:58	skrylar	i don't really mind the "if you fiddle with our lib we want the changes" licenses; you *are* kinda getting big libraries for free in those cases
16:47:10	skrylar	BitPuffin: i donno, it looks like nimrod is using bindings for 1.2
16:47:53	skrylar	all though with big libs like OGRE I don't really *want* to fiddle with them
16:49:07	skrylar	I kinda wonder about overengineering sickness in the C++ world; it seems like C++ people love 500 template pileups, 9 hour compile times and half a million lines of code to print hello world
16:49:17	Varriount	^
16:49:39	Varriount	skrylar: Don't forget the restrictive mindset that the Java world is stuck in
16:49:54	Demos	I agree, although there do exist good c++ template libraries (Eigen!)
16:49:59	BitPuffin	:P
16:50:15	skrylar	Demos: C++'s native architecture makes *any* template library bad
16:50:30	Demos	there is a c++ lib called libcinder designed for "creative coding", but compile times for a trivial (<1000 line) project get to like 6mins
16:50:49	skrylar	Now all the code has to go in headers which aren't compiled, so now GCC has to churn over 60mb of code per file
16:51:11	skrylar	C++'s grammar is horrendous to begin with, let alone megabytes of data to run through it
16:51:26	Demos	skrylar: sure, Eigen does pay consider compile time though. And afaik nimrod's import statement works the same way just without the stupid scope issues
16:51:28	EXetoC	Compiling C++ code on my 1x2GHz CPU was too damn slow
16:51:37	*	XAMPP quit (Ping timeout: 260 seconds)
16:51:42	skrylar	Demos: import is symbolic
16:51:44	Demos	but yeah, having every single header bring in /all/ of boost sucks
16:52:09	skrylar	IIRC nimrod supports symbolic modules, which means they get compiled once and not again unless you touch it
16:52:26	Demos	skrylar: right, I think I misspoke.
16:52:39	skrylar	its okay :)
16:52:50	Demos	nimrod has modules, that is the important part
16:53:02	Demos	the whole way #include is transitive is imo a bigger problem than compile times
16:53:04	skrylar	I can get over the need for forward declarations (ick) considering its a really minor amount of derp for what you get out of it
16:53:19	Demos	what in nimrod?
16:53:23	skrylar	yeah
16:53:53	Demos	yeah forward decls are kinda meh. Not ideal but it can be nice having the whole public interface at the top of the file
16:54:00	Varriount	skrylar: Are you using SDL or Cairo?
16:54:26	skrylar	Demos: you only need forward decls for recursive definitions
16:54:57	skrylar	Demos: and its for the pascal reason (compiler gets to skip an extra step which means 500k SLOC codebases compile exponentially faster than C++)
16:55:20	skrylar	Varriount: neither, ncurses at the moment
16:55:37	skrylar	Varriount: i do a lot of switching due to an uncomfortable desk chair, so i've been working on it through ssh most of the time.
16:55:42	Varriount	A console GUI?
16:55:57	skrylar	yes, but the design is the same otherwise
16:55:58	Demos	skrylar: but don't you need forward decls for recursive definitions in c++ as well?
16:56:21	skrylar	Demos: no, a C++ class can call its own methods without them IIRC
16:57:49	skrylar	Varriount: i'm trying to keep the module set up in such a way that the only thing that changes between targets are the implementations of the Draw function; I don't imagine it would be too hard to even put those in a different file and just use 'include' with some 'when' statements
16:58:19	skrylar	but its easier to get a console behaving, and i can actually use the program that i need to write at least; doing it in a full GUI takes a lot more time
16:59:04	skrylar	Vim has a CLI and GUI version so maybe its not a bad idea to have both for an app :)
17:00:32	BitPuffin	yeah I was actually thinking about that earlier
17:00:52	BitPuffin	maybe it's better to forward declare everything at the top of the nimrod file like a pseudo header because it makes it quicker to skim
17:01:18	Demos	frankly I think it does not matter that much. If you need to skim it than point docgen at it
17:01:20	EXetoC	I use the terminal version only for the shell features
17:01:24	BitPuffin	as a guy who prefers to just open a source file to loook what is available I like it
17:01:32	Demos	but yeah
17:01:47	BitPuffin	Demos: yeah I know but I don't wanna have to leave by text editor and open like firefox to view it :P
17:01:56	Demos	the argument that headers are somehow good because they seperate interface and implementation is not a very stong argument
17:02:36	BitPuffin	It is a fair point I think
17:02:44	BitPuffin	but headers have really bad consequences
17:02:59	EXetoC	I hate repeating myself in that way
17:03:05	BitPuffin	I wouldn't actually mind it if nimrod had headers, just as long as import would work the same way
17:03:09	BitPuffin	EXetoC: yaeh I guess
17:05:18	Demos	well I would not mind if nimrod's module API definition (for binary modules) had the same format as a "header" but I don't want to have to write it myself
17:05:59	EXetoC	that could be automated of course
17:06:08	BitPuffin	D has the .di modules iirc
17:06:22	BitPuffin	yeah automatic generation would be sweet
17:07:30	*	DAddYE joined #nimrod
17:10:19	dom96	hello people
17:10:35	dom96	welcome skrylar!
17:11:45	BitPuffin	oi dom96
17:12:01	dom96	Demos: Have you tried using the {.noSideEffects.} pragma on wrapped C functions?
17:12:04	*	DAddYE quit (Ping timeout: 265 seconds)
17:12:15	dom96	BitPuffin: oiiii
17:13:19	dom96	hrm, the jester example doesn't compiler.
17:13:23	dom96	*compile
17:13:26	dom96	How did this happen!?
17:13:32	EXetoC	dom96: how's the current socket interface? for a game for example.
17:13:45	dom96	EXetoC: How is it for a game?
17:13:48	dom96	EXetoC: Good?
17:14:03	Demos	dom96: I think I did
17:14:06	EXetoC	I suppose I should wait a month or two if I'm not in a hurry
17:14:15	*	aftersha_ joined #nimrod
17:14:20	*	DAddYE joined #nimrod
17:14:21	Demos	dom96: do you mean on the declerations in math.nim or on the calls?
17:14:25	EXetoC	dom96: just so you know what the use case is
17:14:28	dom96	Demos: Declarations.
17:14:39	dom96	EXetoC: It should be good for everything.
17:14:39	Demos	yeha i tried that last night
17:14:48	EXetoC	dom96: the one we have now? ok
17:15:07	dom96	EXetoC: Sure. But the new one will replace it.
17:15:19	dom96	May not bother with a deprecation path.
17:18:48	*	DAddYE quit (Ping timeout: 265 seconds)
17:19:50	dom96	Araq: Looks like you broke my dear jester.
17:24:39	*	aftersha_ quit (Quit: Computer has gone to sleep.)
17:24:40	skrylar	dom96: helloes
17:25:05	skrylar	EXetoC: you could always wrap enet :)
17:25:42	dom96	That is in fact already wrapped :P
17:25:58	skrylar	well there you go
17:26:21	*	skrylar snickers that rust was at 0.9 and TCP/UDP sockets were completely broken >_>
17:26:30	Demos	there is also a zmq wrapper, I hear people like zmq
17:27:04	skrylar	I've never actuallyed used zmq, other than reading the pages about it a long time ago. I think thats more of an internal messenging setup than anything you want public-facing
17:27:06	*	dom96 looks depressingly at the amount of people using Rust.
17:27:44	dom96	I'm the same in regards to zmq. Haven't read much about it and currently regard it as an unnecessary dependency.
17:28:10	skrylar	yeah.. i finally left #rust a few days ago. ran in to a showstopping bug with the stdlib (they fixed it now i think) but there's just too much derp
17:28:45	OrionPK	bleh, why doesnt the dom module have a console/console.log
17:29:02	skrylar	like public release versions that contain broken sadness, syntax which is finger painful, and the worst part: 'oh well some layer of LLVM will optimize this really slow thing away at some point' </rant>
17:29:22	*	XAMPP joined #nimrod
17:29:26	dom96	skrylar: You should write a blog post titled "Why I left Rust for Nimrod" ;)
17:29:41	skrylar	dom96: i had started to write one about tips for using rust
17:29:45	skrylar	i think now my tip would just be "don't"
17:30:17	skrylar	the memory safety rig is neat until you need to actually do something
17:30:37	skrylar	i was reading that the mozilla guys were having trouble getting the DOM modelled under it, and thats what the language was designed *for*
17:31:08	dom96	You are literally the first person I came across to say something negative about Rust.
17:31:19	skrylar	lol
17:32:35	EXetoC	skrylar: was that recently?
17:32:35	dom96	I'm mostly too busy working on Nimrod to judge it for myself. I once decided to try it, after waiting 45 minutes for the compiler to bootstrap I decided to give up.
17:32:42	skrylar	EXetoC: yes
17:32:47	Demos	skrylar: I have never used rust but that is the impression I got from reading about the type system
17:33:00	skrylar	well the borrowing mechanics are really neat
17:33:22	skrylar	it will get mad if you try to shove a pointer somewhere and it can't prove that pointer is going to live long enough not to dangle
17:33:29	Demos	I /know/ that c++'s unique_ptr<T>s are only really useful when you get non-owning raw pointers from them, and restricting that to some subset where the language can prove valid access would be hard
17:33:51	Varriount	dom96: Did you push your iocp stuff yet?
17:33:55	dom96	Varriount: yeah
17:33:58	Varriount	:D
17:34:05	*	XAMPP quit (Ping timeout: 260 seconds)
17:34:07	skrylar	the derpiness comes in that they removed the GC heap, which means you can't really do the dynamic dispatch stuff
17:34:12	dom96	Varriount: It's in devel actually. Because Araq accidentally merged it in there ...
17:34:18	EXetoC	skrylar: I thought it was alright, apart from the implementation limitations at that time, but it's too bad even the language devs are struggling
17:34:23	dom96	But I am still working on the newasync branch.
17:34:25	Varriount	dom96: Why wouldn't it be in devel?
17:34:31	skrylar	EXetoC: well they won't rewrite anything :|
17:34:39	dom96	Varriount: Because it's not ready.
17:34:59	skrylar	EXetoC: there is so much about Rust now that is WONTFIX because "well the AST isn't built for that" e.g. macros are crippled because they have a specific pass that has access to basically nothing
17:35:46	dom96	Nimrod really shines when it comes to macros.
17:36:01	Demos	dom96: well once you get passed the rather terse error messages and bugs
17:36:03	skrylar	their compiler runs in passes, which means errors come in all jigsaw puzzled; fix all the errors? get new errors
17:36:35	dom96	skrylar: Oh yeah. Speaking of Maybe: all you need is a pattern matching macro and you're set I think.
17:36:42	skrylar	oh yeah, the 45 minute bootstrap is because they forked LLVM
17:37:14	skrylar	lisp macros + pascal types <3
17:37:18	EXetoC	dom96: what part of it?
17:37:34	skrylar	dom96: i do wish that nimrod wouldn't monomize generics for distinct pointers
17:37:35	dom96	EXetoC: Linux support
17:38:10	skrylar	maybe its being nitpicky but pointer types do not need a struct for each pointer, they can just use a void* and typecast like a C programmer would
17:38:17	dom96	Demos: True... but the macros are still IMO pretty sexy.
17:38:51	dom96	skrylar: No idea what "monomize generics" means heh
17:39:05	EXetoC	dom96: the pattern matching. the problem I had was the verbosity involved when constructing Nothing
17:39:19	Varriount	"ref object of PObject" - What is that, a ref of a ref of an object?
17:39:38	dom96	Varriount: No. A ref object inheriting from a 'ref TObject'
17:39:48	dom96	Varriount: It's not a double ref.
17:40:13	skrylar	dom96: say you have an object with a single T field, now say you use that object three times with different pointer types.
17:40:17	Demos	I agree, once the polish is applied they will be amazing. But quite frankly all of nimrod's metaprogramming features need refinement. I ran into a lot of trouble trying to do some type-level arithmatic yesterday
17:40:37	skrylar	dom96: if its a distinct pointer, nimrod will generate 3 structures in the C file which each have 'void*' as the type for the distinct pointer but are identical structs
17:41:09	skrylar	e.g. it doesn't know that it can share the generic because they are all void* types
17:41:50	dom96	skrylar: oh, well i'm sure that can be fixed?
17:42:03	skrylar	probably; i wouldn't know how
17:42:17	skrylar	I know Ada does some things so that generics don't bloat up the code base very much
17:42:32	dom96	Wouldn't the C compiler take care of the bloat in this case though?
17:43:05	skrylar	I don't know if GCC or LLVM will de-dup everything
17:43:08	Varriount	skrylar: Generics are hard.
17:43:24	skrylar	supposedly LLVM has a mechanism for shared generics but I don't know what it is
17:43:31	*	[2]Endy quit (Read error: Operation timed out)
17:45:21	*	XAMPP joined #nimrod
17:45:25	dom96	EXetoC: What verbosity is that? The need to instantiate the generic?
17:45:43	EXetoC	dom96: yes
17:45:56	dom96	EXetoC: Yeah, that's a bit of a PITA
17:47:40	Demos	hm conversions from (int, int) to (float, float) are invalid even with a cast, how annoying
17:47:58	Demos	not a cast[]() style cast ofc
17:48:42	dom96	I must admit, int/float types in Nimrod are a tad annoying.
17:50:19	Demos	yeah, all those implicit conversions in C are annoying but not having them makes you realize why they are there
17:51:21	EXetoC	dom96: that has more to do with tuples
17:51:33	*	radsoc quit (Ping timeout: 272 seconds)
17:51:50	Demos	EXetoC: yeah, this is true
17:55:03	*	DAddYE joined #nimrod
17:55:11	EXetoC	I don't know if it should be implicit for anything other than ranges
17:56:53	Demos	yeah, and makeing implicit conversions for tuples is probably more trouble than it is worth. Even explicit conversions could be hard. I am all for implicit conversions for initialization (when a numeric literal appears in code)
17:58:05	EXetoC	sure. in many cases it is
18:01:16	*	Demos quit (Ping timeout: 252 seconds)
18:06:55	*	brson joined #nimrod
18:17:24	BitPuffin	you know what, I'm gonna work a bit on my multiplayer thing
18:18:20	OrionPK	what network lib are u using
18:19:09	BitPuffin	right now just internal nimrod shits
18:23:33	skrylar	blegh. multiplayer code.
18:23:47	skrylar	scenegraph sync hell
18:26:48	BitPuffin	:P
18:26:58	BitPuffin	it's a bit more difficult than that
18:27:25	BitPuffin	at least for this project
18:35:01	reactormonk	skrylar, o/
18:35:12	reactormonk	How do I get map(x.inputs, proc(i: TItem): TTypeID = i.typeID).toSet.incl(y.output.typeID) running?
18:35:36	reactormonk	(TSet[TItem], proc (TItem): TTypeID)
18:35:41	reactormonk	aka mapping over a set
18:36:52	Araq	OrionPK: I can't reproduce your leak
18:37:05	Araq	it stays below 300K forever
18:37:11	OrionPK	look at task manager or the console?
18:37:14	OrionPK	looking*
18:37:23	Araq	console
18:37:28	OrionPK	look at task manager
18:37:36	Araq	humm
18:37:39	EXetoC	reactormonk: where does it fail? does incl work like that when it takes a var as the first arg?
18:38:27	Araq	OrionPK: that means nimrod's reporting is wrong?
18:39:20	OrionPK	araq maybe.. or it's leaking memory outside of nimrod's GC?
18:40:36	reactormonk	EXetoC, it's the map itself.
18:40:43	reactormonk	just defined it in sets
18:41:10	Araq	OrionPK: ok, will look at it later
18:41:48	*	aftersha_ joined #nimrod
18:42:57	Araq	skrylar: we require the "no forwarding rule for the macro system. pascal has nothing to do with it
18:43:12	*	filwit joined #nimrod
18:43:37	EXetoC	reactormonk: you haven't gotten it to work? have you looked at system.map?
18:44:31	filwit	Araq: what symbol in vm did you want me to document?
18:44:41	Araq	lol
18:45:05	Araq	make a top level comment how it works
18:45:30	filwit	yeah a doc comment, got it
18:45:37	filwit	but what part?
18:45:59	Araq	the calling convention
18:46:11	Araq	the registers
18:46:25	filwit	okay, so just everything i've been going over for awhile
18:46:38	filwit	okay
18:46:46	filwit	^doubt okay
18:46:49	filwit	double*
18:46:50	Araq	the bizzare invaraints wrt PNode
18:46:54	NimBot	Araq/Nimrod devel ba8e74d Simon Hafner [+1 ±1 -0]: added `map` to sets
18:46:54	NimBot	Araq/Nimrod devel 59dd15c Simon Hafner [+1 ±0 -0]: added tests for set
18:47:15	reactormonk	Araq, request to rename set to bitset
18:47:56	Araq	reactormonk: why no make a map for everything that has items?
18:48:53	reactormonk	Araq, because initSet
18:49:09	reactormonk	Araq, I need items() and the constructor.
18:49:30	Araq	oh so your map returns a set, i see
18:49:35	reactormonk	yup.
18:49:41	*	BitPuffin quit (Ping timeout: 260 seconds)
18:49:46	Araq	very well then
18:49:54	reactormonk	That's why I pressed for a null-element by type.
18:50:11	filwit	rename: set=>bitset, seq=>list, list=>chain :P
18:50:29	filwit	(just kidding, though they are my favorite names)
18:50:36	reactormonk	filwit, ;-)
18:51:00	reactormonk	Araq, issue for mapping type => proc that generates a nulle-element of that type
18:51:15	reactormonk	?
18:51:22	Araq	no
18:51:44	Araq	fix some bug instead
18:51:51	Araq	there are over 230 issues
18:52:10	reactormonk	that's a feature request. 1minute of work.
18:52:31	filwit	those are bugs caused by you though Araq (since you wrote the compiler). It's only fare you have to fix them all.
18:52:43	filwit	makes perfect sense
18:53:14	*	Varriount slaps filwit around a bit with a average-sized meteor.
18:53:29	Araq	reactormonk: nope
18:53:31	Araq	wrong
18:53:52	reactormonk	Araq, go on...
18:53:54	Araq	I recently realized it's not enough to write the code and test it
18:54:04	Araq	you have to maintain it ...
18:54:05	reactormonk	you also have to fix it?
18:54:43	reactormonk	Araq, do you give someone, say dom96, the authority to act on behalf of the nimrod project for GSoC?
18:55:19	Varriount	Can dom96 even do that? Don't you have to be 18 or older, or something?
18:55:27	Araq	hey
18:55:34	Varriount	Hi
18:55:38	Araq	dom96 is not *that* young
18:56:32	Araq	filwit: actually most bugs are zahary's :P
18:56:33	dom96	But I am but a toddler.
18:56:59	reactormonk	Araq, I could write liftedMap[A,B](zero: proc(): A, add: proc(var x: A, B)): proc map*[A, B](data: TSet[A], op: proc (x: A): B {.closure.})
18:57:21	reactormonk	well, the generics in the first part would need to be different from the second part
18:57:30	OrionPK	dom96 how many years have you been programming?
18:57:40	*	dom96 counts
18:57:54	reactormonk	I think I'm at 6
18:57:57	dom96	~6 ish
18:58:04	OrionPK	how old are you again? 19?
18:58:07	dom96	18
18:58:14	reactormonk	not too shabby
18:58:16	OrionPK	since you were 12?
18:58:19	dom96	yeah
18:58:25	OrionPK	I was doing HTML when I was 12 or 13 iirc
18:58:32	reactormonk	I have written my first code at 12, but that was VB ^^
18:58:32	dom96	Give or take 1 or two years
18:58:36	OrionPK	but no actual programming until I was maybe 16 or 17
18:58:44	reactormonk	^ same
18:58:55	dom96	Yeah, I was doing VB.NET back then
18:59:03	reactormonk	dom96, well, do you want to do the GSoC?
18:59:12	OrionPK	I did some VBA when I was 15 or something
18:59:19	*	DAddYE quit ()
18:59:33	dom96	reactormonk: I don't think a single person can do it.
18:59:52	reactormonk	dom96, I'm pretty sure as well. Are you in for kicking other people to do it?
19:00:02	dom96	Ahh... VB.NET. Those were the days.
19:00:26	reactormonk	dom96, if you wanna have the American speech - it's gonna look good on your resume
19:00:30	OrionPK	.NET wasnt around, unfortunately
19:00:38	dom96	reactormonk: I wouldn't mind doing it, but after reading the FAQ it looks like a lot of papers need to be filled out. Including tax forms and what not, which I have no experience with.
19:01:02	reactormonk	dom96, I'm sure there's a mentor channel somewhere where you can get help for that
19:01:03	dom96	reactormonk: hrm, that's true.
19:01:21	dom96	reactormonk: Do you want to work on Nimrod over the summer?
19:01:27	filwit	beyond my initial programming in BASIC (on an hand-me-down Apple IIc), I learned programming from writing Macromedia Director Lingo scripts
19:01:30	reactormonk	dom96, nope, need to finish my masters
19:01:43	Varriount	I do! I want to work on nimrod over the summer!
19:01:50	filwit	i looked up Director, apparently it's still alive and making Shockwave games today..
19:02:00	dom96	Varriount: Loving the excitement!
19:02:11	*	DAddYE joined #nimrod
19:02:43	dom96	hrm, actually. I started with C not VB.NET hah
19:02:54	reactormonk	Araq, do you approve of dom96, so he can go ahead if he decides to?
19:02:55	filwit	so basically, soon we loose dom96 to college life?
19:03:02	dom96	Was disappointed I could only create console apps though.
19:03:02	Varriount	I started with Python :3
19:03:09	filwit	:|
19:03:39	dom96	(The book I got only gave examples of how to print things in a console. I wanted GUI apps!)
19:04:16	dom96	filwit: I'll be working on Nimrod while at lectures, so you won't lose me :P
19:04:24	Varriount	^ Like me
19:04:37	NimBot	Araq/Nimrod devel e534ad4 Simon Hafner [+0 ±2 -0]: fixed a bug in `map` for sets
19:04:47	dom96	I'm really worried I will be extremely bored in the first year.
19:05:35	Araq	reactormonk: you know... having me review your code has its advantages
19:05:39	dom96	If we spend a month reviewing "Hello World" I will just tell them to give me the end of year exams now and put me in the second year :P
19:05:48	Araq	reactormonk: yes I approve dom96 for everything
19:06:04	dom96	:)
19:06:05	reactormonk	Araq, yup. But it also costs you time. PRs it is?
19:06:24	reactormonk	your choice.
19:07:11	Araq	PRs please
19:07:34	Araq	we have quality issues and dom96 said it's because I do too few code reviews
19:07:34	reactormonk	kk
19:07:49	Varriount	dom96: Did I tell you why I think stdout and stderr are getting mixed up?
19:07:56	Varriount	In osproc.nim
19:07:59	dom96	Varriount: nope. Do tell.
19:07:59	reactormonk	dom96, so go ahead with GSoC
19:08:14	dom96	reactormonk: How about you help me register?
19:08:24	reactormonk	dom96, sure
19:08:38	dom96	Right, let me finish watching the walking dead first.
19:08:45	reactormonk	http://www.google-melange.com/gsoc/homepage/google/gsoc2014
19:08:51	Varriount	dom96: In short, what I believe is happening is that multiple writes are happening at once, getting the output stream mixed up
19:09:00	reactormonk	dom96, I'll have to do some homework in about 1h though
19:09:05	EXetoC	Araq: what are the endian conversions in oids for?
19:09:06	reactormonk	and from what I see, it's gonna take a while
19:09:13	Varriount	The equivalent of two programs trying to write to the same file at the same time.
19:09:53	Araq	EXetoC: it's always big endian or something. fyi I ported this from the C client library
19:10:04	Araq	it's not like I came up with this myself
19:10:28	EXetoC	odd, but ok got it
19:11:18	reactormonk	Araq, any plans to harden repr against nil?
19:11:45	dom96	reactormonk: 10 mins left
19:12:06	dom96	Varriount: Interesting, any ideas how to fix it?
19:12:07	reactormonk	dom96, kk
19:12:28	Araq	bah, not 'repr' again
19:12:39	Araq	I wonder why it's so loved
19:12:51	Araq	it's hack and doesn't work that well
19:12:52	filwit	what's the alternative to repr?
19:12:59	Varriount	dom96: I'll have to look at the windows internals of osproc.nim
19:14:42	dom96	Araq: because it's useful for debugging.
19:15:07	dom96	people don't want to write a function which goes through each field manually and echos it
19:15:44	reactormonk	Araq, I'd love to deprecate repr in favour of a $ that works on everything
19:16:38	Araq	reactormonk: $ for everything is not good
19:16:54	reactormonk	Araq, so repr for everything then.
19:17:17	Araq	meh fine, so make it work with 'nil'
19:17:35	Araq	it works with nil afaict
19:17:50	Araq	but I never use repr so I might be wrong
19:18:12	reactormonk	I get a few nil access errors with repr
19:18:14	*	zahary joined #nimrod
19:20:22	Araq	hi zahary
19:21:33	*	Mat3 joined #nimrod
19:21:39	Mat3	hi all
19:23:39	Araq	hi Mat3
19:23:40	zahary	hi
19:23:58	Araq	zahary: I changed how typedesc works in the compiler but now stuff broke ...
19:24:25	dom96	ooh, House of Cards Against Humanity
19:24:29	*	Mordecai joined #nimrod
19:24:50	zahary	what did you change?
19:25:17	Mat3	hi Araq
19:25:30	Mat3	and zahary
19:27:00	reactormonk	http://pastebin.com/2XkfVm5m
19:27:18	Araq	typedesc[T] stays as it is, typedesc becomes internally tyTypeDesc[tyNone]
19:27:19	reactormonk	http://pastebin.com/2XkfVm5m
19:27:21	reactormonk	oops
19:27:38	Araq	so that skipTypes doesn't crash when you skip tyTypedesc
19:27:52	Mat3	exist there a time-frame for stable Nimrod versions ?
19:28:04	Araq	but I'm puzzled by the logic in sigmatch.nim wrt tyTypeDesc
19:28:10	Araq	Mat3: december 2014
19:28:32	zahary	tyTypeDesc[tyNone] would be void in the current design
19:28:35	dom96	reactormonk: ok, time to do this.
19:28:45	zahary	still different than just typedesc
19:28:58	Araq	zahary: I don't think so
19:29:09	Araq	tyEmpty is void, tyNone is ... none
19:29:17	zahary	ah, you could be right
19:30:28	reactormonk	dom96, just found a bug in sqlite fastrows I believe
19:30:30	zahary	so, what's puzzling in sigmatch?
19:30:31	reactormonk	but go for it
19:30:57	Varriount	dom96: Preliminary research on stdout and stderr pipes suggest that the problem *might* be due to buffering
19:31:34	reactormonk	dom96, http://www.google-melange.com/gsoc/homepage/google/gsoc2014
19:31:40	dom96	"Mentor Summit at Google: Representatives from each successfully participating organization are invited to Google to greet, collaborate and code"
19:31:41	dom96	0_o
19:31:58	dom96	So if my application is successful I will get to visit Google?
19:32:10	reactormonk	dom96, looks like it
19:32:18	dom96	hrm, why not.
19:32:43	zahary	dom96: have you read what does it take to become a mentoring organisation?
19:32:52	reactormonk	dom96, I suppose name is just nimrod
19:33:03	reactormonk	any official nimrod logo?
19:33:14	Varriount	reactormonk: The crown
19:33:25	reactormonk	dom96, ^
19:33:33	Varriount	The unofficial logo is a honey badger
19:33:51	dom96	zahary: I'm reading the FAQ now.
19:33:58	dom96	reactormonk: Let me read the FAQ first.
19:34:07	reactormonk	dom96, go fir it
19:34:18	dom96	zahary: Would you have time to mentor?
19:34:23	*	xtagon joined #nimrod
19:34:27	Varriount	Hi xtagon
19:34:36	xtagon	Hey Varriount
19:34:46	reactormonk	o/ xtagon
19:34:58	xtagon	What's new?
19:35:08	*	Mat3 is now known as Mat3-work
19:35:30	dom96	This looks pretty in-depth: http://en.flossmanuals.net/GSoCMentoring/what-is-gsoc/
19:38:18	zahary	dom96: I might be surprised by how much time is needed, but I could certainly try (mentoring)
19:39:12	dom96	zahary: According to the FAQ "Five hours per student per week is a reasonable estimate."
19:39:16	dom96	http://www.google-melange.com/gsoc/document/show/gsoc_program/google/gsoc2014/help_page#howmuch
19:39:41	zahary	that sounds quite reasonable
19:40:14	dom96	But make sure to keep yourself free because if you start disappearing students may get really discouraged.
19:40:57	dom96	Araq: You might as well be a mentor too.
19:41:21	zahary	I'm rarely without access to email, live chat as IRC is another story
19:41:49	dom96	zahary: I see. Can I have your email?
19:42:23	dom96	I think as long as you have some sort of way to communicate with the student it will be fine.
19:42:29	Varriount	dom96: Is there any reliable way to test osproc getting stdout and stderr mixed up?
19:43:43	zahary	I have one of yours (morfeusz), so I'll send you a ping
19:43:54	dom96	zahary: ok. Thanks.
19:44:08	dom96	Varriount: hrm, not sure.
19:44:28	dom96	Varriount: Perhaps create some app which writes to the stdout and stderr
19:44:37	xtagon	Nimrod is going to participate in GSoC? :D
19:44:44	reactormonk	http://dpaste.com/1602184/ <- that's the data structure that kicks repr
19:44:47	dom96	xtagon: Maybe, interested?
19:45:08	reactormonk	Varriount, 2> /dev/null to test for stdout
19:45:22	Varriount	Windows, remember?
19:45:32	xtagon	dom96, I'm not enrolled in an academic institution right now
19:46:02	dom96	xtagon: ahh, I wish I could participate as a student but I have the same problem heh.
19:46:33	reactormonk	Varriount, oh sap
19:46:45	Varriount	Hrm. http://stackoverflow.com/questions/1545619/can-i-capture-stdout-stderr-separately-and-maintain-original-order
19:46:49	*	Mordecai quit (Quit: bbiab)
19:47:31	dom96	Varriount: Perhaps there is a different way to merge the two streams.
19:49:54	*	radsoc joined #nimrod
19:50:32	xtagon	No, you NEVER cross the streams!
19:50:52	xtagon	http://www.youtube.com/watch?v=jyaLZHiJJnE
19:52:36	*	carum joined #nimrod
19:57:41	dom96	Araq: What do we actually want students to be working on?
19:57:44	Araq	zahary: sigmatch line 884 following
19:59:09	Araq	you check the bindings before even checking whether it's typedesc or typeDesc[T]
19:59:31	Araq	which means (a, b: typedesc) only matches (int, int) but not (int, float)
20:01:26	zahary	that was true at some point, then we decided that typedesc is not bindOnce
20:01:44	zahary	this is currently handled earlier while the proc implicit generic params are collected
20:01:52	*	JStoker quit (*.net *.split)
20:01:53	*	joelmo quit (*.net *.split)
20:01:53	*	gsingh93_ quit (*.net *.split)
20:02:05	zahary	(a, b: typedesc) will be translated to (a: typedesc1, b: typedesc2)
20:02:15	*	JStoker joined #nimrod
20:02:28	zahary	now, I'm not sure if the code in sigmatch still needs to have its current structure
20:02:48	Araq	also this part
20:02:50	Araq	internalAssert prev.sonsLen == 1
20:02:52	Araq	let toMatch = if tfUnresolved in f.flags: a
20:02:53	Araq	else: a.sons[0]
20:02:55	Araq	result = typeRel(c, prev.sons[0], toMatch)
20:02:57	*	tdc joined #nimrod
20:03:34	Araq	you don't even check what 'a' is, you simply access a.sons[0]
20:04:14	*	gsingh93_ joined #nimrod
20:04:26	*	joelmo joined #nimrod
20:04:46	zahary	this is prev; it can't be something different than a typedesc
20:05:01	zahary	ah, nevermind
20:05:37	tdc	What is the Nimrod way to split a module into several files?
20:06:13	Araq	tdc: either use different modules with symbol "forwarding" via 'export' or use simply use 'include' instead of 'import'
20:07:37	tdc	excellent! with include the doc parsing doesn't work - right? So maybe "forwarding" is cleanest?
20:07:46	*	vbtt joined #nimrod
20:07:59	Araq	tdc: doc2 can see through "include"
20:08:19	tdc	ok. Thanks!
20:09:47	dom96	It seems we need two org admins anyway.
20:11:52	dom96	Araq: If we do this I am going to need some feedback here.
20:12:05	Araq	dom96: ok, I'm number two
20:12:15	Araq	there are lots of things one could do
20:12:34	dom96	Araq: Firstly, we need concrete ideas. Because we need to submit a link to the ideas list in the proposal.
20:12:48	dom96	Araq: Secondly, how many students do we want to take?
20:12:59	reactormonk	dom96, one or two.
20:13:01	Araq	implement true coroutines that can capture the full stack
20:13:14	dom96	reactormonk: Yeah, definitely not more than 2.
20:13:21	vbtt	+1 to the coroutines idea!
20:13:22	dom96	Varriount already wants to do it so that's 1.
20:13:57	zahary	I propose making nimrod a platform for GC research - this means implementing precise stack marking and allowing GC implementation files to pick their code generation mode
20:14:08	dom96	Araq: Create a gist with ideas.
20:14:45	Varriount	dom96: Hm?
20:14:58	dom96	Varriount: You want to participate in GSoC
20:15:05	Araq	also we can have go-like light-weight concurrency with excellent C interop with some new technique that I invented
20:15:23	Varriount	dom96: Yes...
20:15:25	Araq	well likely re-invented :P
20:15:50	dom96	Araq: All that sounds very difficult :P
20:15:51	Varriount	dom96: There is likely no sure way to untangle stdout and stderr
20:16:00	vbtt	Araq:don't true coroutines give us light-weight concurrency already?
20:16:01	Varriount	Not even the command prompt can do it.
20:16:16	reactormonk	Varriount, muh.
20:16:24	Araq	vbtt: sure but without multi-core support
20:16:47	*	aftersha_ quit (Quit: Computer has gone to sleep.)
20:16:59	zahary	so Araq, you are right, I managed to crash it with some carefully crafted code
20:17:01	dom96	Varriount: What would you like to do?
20:17:09	vbtt	Araq: as i see - you want to multiplex coroutines on multiple system threads.
20:17:13	dom96	(For GSoC)
20:17:16	Varriount	dom96: For what? GSoC, or osproc.nim?
20:17:21	dom96	^
20:17:32	Varriount	Anything that you think I can handle, that is needed.
20:17:39	Araq	vbtt: no *I* don't want that at all
20:17:56	Varriount	If you think that coroutines are possible for someone of my skill set, I will try my best.
20:18:07	Araq	so every "cheap" coroutine can race, no thanks
20:18:18	*	shodan45 joined #nimrod
20:18:44	*	BitPuffin joined #nimrod
20:18:51	*	Varriount|Mobile joined #nimrod
20:19:00	Araq	dom96: 64 new record?
20:19:00	dom96	Araq: How are you planning on mentoring people about implementing coroutines?
20:19:01	vbtt	Araq: fwiw, i think true coroutines, single threaded are good enough.
20:19:06	dom96	Araq: no
20:19:22	Varriount|Mobile	Meep
20:19:46	vbtt	Araq:also just coroutines may be the right sized project complexity-wise.
20:19:46	Araq	dom96: well I can give detailed instructions of what needs to be done to implement them
20:19:52	*	shodan45 is 65 \o/
20:20:23	*	shodan45 notes that 65 is not his age :P
20:21:01	dom96	Araq: ok, anything you can think of that I could mentor?
20:22:04	Varriount|Mobile	asyncio development
20:22:53	dom96	That will be done by the time summer arrives :P
20:23:00	vbtt	actually single threaded coroutines are more desirable imo. you know exactly one of them is running and you know the switch points. they also use a shared heap.
20:23:10	Araq	vbtt: exactly
20:24:24	vbtt	and the programmer can manually spin up multiple threads if they desire.
20:25:46	vbtt	ok. yeah i definitely support the single threaded coroutines idea! and if it eventually works well with the async io, that would be ideal.
20:26:01	vbtt	(basically i want a gevent like system in nimrod)
20:26:39	Araq	vbtt: dom96 has a working prototype of asyncio with nimrod's "yield"
20:27:47	Araq	zahary: please have a look at why tests/macros/tmacrogenerics fails
20:28:01	Araq	I changed its supposed output, but it fails again
20:28:19	Araq	and I'm not sure I should have changed the supposed output
20:28:45	zahary	have you tried to fix the handling in typedesc values in the new VM?
20:28:58	vbtt	Araq:ok. dom96: is your stuff checked in?
20:29:16	dom96	vbtt: newasync branch
20:29:22	dom96	vbtt: It's not finished yet
20:29:33	dom96	vbtt: And I have some code which is still not pushed.
20:30:01	*	psquid joined #nimrod
20:30:08	*	psquid quit (Changing host)
20:30:08	*	psquid joined #nimrod
20:30:42	*	carum quit (Remote host closed the connection)
20:30:44	Araq	zahary: yes the codegen of the vm is aware of tyTypeDesc. But it still gets it wrong
20:30:49	vbtt	dom96:thanks. do you have examples i can look at? or docs?
20:31:05	dom96	vbtt: Not just yet.
20:31:12	*	[1]Endy joined #nimrod
20:31:47	Varriount|Mobile	Araq: If you ever get the urge to work on iterator and generic bugs, I have an entire file full of them
20:32:05	dom96	Araq: Here is KDE's idea list: http://community.kde.org/GSoC/2011/Ideas
20:32:19	dom96	The idea list is the most important in the application.
20:33:20	EXetoC	should categories.nim be extended to include "tests/stdlib" when specifying the stdlib category?
20:34:05	EXetoC	rather than just lib/pure/*.nim and the highlite module
20:34:20	vbtt	Varriount:are you familiar with how coroutines are used?
20:34:52	demizer	Hello, everyone. Is there a way to read the html documentation for the master branch?
20:35:22	Varriount|Mobile	vbtt: Vaguely
20:35:35	*	Mat3-work is now known as Mat3
20:36:15	Varriount|Mobile	demizer: Run the documentation generator on a clone of the master branch
20:36:16	EXetoC	demizer: build.nimrod-lang.org. a hyperlink to build.nimrod-lang.org/docs/manual.html is missing
20:36:23	vbtt	Varriount:you could start by looking at Python greenlets, gevent(which uses greenlets) and lua coroutines.
20:36:28	demizer	sweet thank you
20:36:31	EXetoC	or that
20:36:34	vbtt	Varriount:are you familiar with python generators?
20:36:49	Varriount|Mobile	vbtt: Quite familiar with those
20:37:17	dom96	Araq: Is `is` in a `when` broken?
20:38:26	EXetoC	not for me
20:38:47	EXetoC	are the operands types?
20:38:49	*	micklat joined #nimrod
20:38:57	*	[1]Endy quit (Read error: No route to host)
20:39:18	Varriount|Mobile	vbtt: But iirc, Python generators are not coroutines
20:39:34	*	[1]Endy joined #nimrod
20:39:37	dom96	hrm, I think I see the issue.
20:39:50	vbtt	Varriount: simplifying here but goroutines are the next level of generators.
20:40:17	Varriount|Mobile	"Next level" in what way?
20:40:19	dom96	What's the behaviour of 'is' when comparing a variable to a type?
20:40:34	EXetoC	can individual output tests be executed?
20:40:34	vbtt	Varriount: generators can have one suspended function. coroutines are a stack of suspended function calls.
20:40:51	vbtt	Varriount:sorry meant coroutines (not goroutines)
20:43:02	NimBot	dom96/jester master 1527e0d Dominik Picheta [+0 ±1 -0]: Fixes #12
20:43:34	vbtt	Varriount: so a function f() calls start_coroutine g(). then g() calls h(). so the stack looks like this: f() -> [ g() -> h() ]
20:43:58	vbtt	the square brackets represent the coroutine
20:44:18	vbtt	now h() could 'suspend' and control returns to f().
20:44:42	Varriount|Mobile	So, like Python's 'yield from'?
20:44:45	vbtt	now f() calls f2() and the stack looks like this. f() -> f2()
20:45:23	vbtt	however the coroutine [g() -> h()] is still around in a suspended state and can be 'resumed'.
20:45:49	vbtt	Varriount:yes like 'yield from'. the difference is that you don't have to write 'yield from' in each function that can be suspended.
20:46:24	vbtt	you just write normal code and any function can call 'suspend()' which suspends the current coroutine.
20:46:56	Araq	dom96: I wouldn't say "broken" but maybe the new vm still gets it wrong occasionally
20:47:05	dom96	Araq: nah, it's fine.
20:47:07	vbtt	(i'm describing asymmetric coroutines, btw, there are also symmetric ones where the idea is similar, i.e. suspended function stack, but the semantics are different)
20:48:20	dom96	Araq: Add ideas here please: https://github.com/Araq/Nimrod/wiki/GSoC-2014-Ideas
20:48:51	vbtt	Varriount:also, different from 'yield from' in that you can start a new coroutine and call any function (i.e. the called function doesn't have to be written with 'yield' or 'yield from')
20:49:14	vbtt	Varriount:i.e. the semantics are more like threads.
20:49:15	Varriount|Mobile	The question is, how difficult would it be to implement coroutines? I would probably need help, I don't think I could implement them on my own.
20:49:58	vbtt	Varriount:with Araq's help, it shouldn't be too hard. Are you familiar with how function calls work w.r.t the stack?
20:50:32	*	psquid quit (Quit: oop)
20:51:06	*	psquid joined #nimrod
20:51:06	*	psquid quit (Changing host)
20:51:06	*	psquid joined #nimrod
20:52:28	Varriount|Mobile	vbtt: Again, I have some knowledge. Bits and pieces that I've picked up from reading
20:53:29	Varriount|Mobile	Things like, different kinds of function calls modify certain registers to reflect return addresses and values and arguments
20:53:39	Mat3	Varriount|Mobile: You will need some knowledge of how to avoid call-frames (or better about misusing them for coroutine sheduling)
20:53:40	dom96	Is there much information about how coroutines are implemented out on the internet?
20:53:40	vbtt	Varriount: that's probably something you should read up on. also, i recommend using coroutines to get an idea of the semantics. (e.g. lua, greenlets, stackless etc.)
20:54:50	vbtt	dom96:yes but i haven't found a good resource that covers it from the ground up. also it depends on your function call implementation, and coroutines are much easier in interpreted languages which store the function frames as heap objects.
20:55:12	Varriount|Mobile	vbtt: Good thing I like reading
20:56:00	Varriount|Mobile	I've been reading a boom on the internals of Windows NT to get to bed
20:56:09	Varriount|Mobile	*book
20:57:54	vbtt	Varriount:w.r.t the stack i meant that the function locals are stored on the stack (c semantics and also nimrod). if you want to suspend functions, naturally you need a strategy to either copy the locals or have some kind or segmented stack or store the locals in the heap, not stack. so you have to make these decisions that araq can help you with.
20:58:40	vbtt	anyway, g2g. bbl
21:00:30	dom96	wow, my cahbot still compiles.
21:00:53	*	[1]Endy quit (Ping timeout: 248 seconds)
21:01:45	dom96	bbl
21:05:52	vbtt	Varriount:feel free to ping me on irc with questions if i'm around.
21:06:53	Mat3	coroutines: http://home.hccnet.nl/a.w.m.van.der.horst/forthlecture6.html
21:08:35	*	carum joined #nimrod
21:10:47	vbtt	Mat3:wikipedia is probably a better introduction :)
21:13:53	Mat3	well, this coroutine implementation is only one single definition (CO) which should be easily understandable in my opinion
21:16:44	vbtt	perhaps, but only to those comfortable with forth.
21:20:24	*	Varriount|Mobile quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
21:23:27	OrionPK	in a for x in.. how come the `x` variable is `res` and not `x` in endb
21:23:41	OrionPK	(when i look at locals)
21:27:05	Araq	OrionPK: inlining of iterators is to blame, I guess
21:28:11	*	micklat quit (Remote host closed the connection)
21:28:12	OrionPK	:\
21:30:57	*	tdc quit (Quit: Leaving)
21:33:12	*	carum quit (Remote host closed the connection)
21:34:21	*	carum joined #nimrod
21:45:24	*	carum quit (Remote host closed the connection)
21:52:46	Varriount	Araq: Any plans on passing closure iterators with arbitrary parameters as arguments to another iterator?
21:53:35	Araq	no sure what you mean
21:54:02	Araq	closure iterators got more stable recently and can capture from an outer proc
21:54:13	vbtt	nice ^
21:54:35	Varriount	Araq: Yes.. but that only works if the iterators are expressly designed that way.
21:55:30	vbtt	btw, would be nice to have the syntax allow easy definition of iterators that just capture from the outer proc. this is the more common case, imo.
21:56:09	vbtt	it is less common to pass parameters into the iterator with each step. however the syntax makes defining those easy.
21:56:48	Varriount	vbtt: I'm thinking about iterator chains, like map(filter(...), ...)
21:58:56	Varriount	vbtt, Araq: See https://gist.github.com/Varriount/8925071
21:58:58	zahary	we have talked about this - I'd like to see it with inline iterators actually (closure iterators are slow)
21:59:50	Varriount	zahary: nd what was the outcome of the talk?
21:59:56	Varriount	*And
22:00:28	dom96	back
22:00:50	dom96	Ideas guys, ideas: https://github.com/Araq/Nimrod/wiki/GSoC-2014-Ideas
22:01:49	zahary	well, they are not trivial to implement, but the outcome of such discussions is very easily influenced by working code :)
22:02:08	vbtt	Varriount: cant you chain iterators already? by wrapping it in iter() or something?
22:03:44	vbtt	I mean - something like map() should accept an iterable, right?
22:04:58	*	Demos joined #nimrod
22:05:39	Varriount	vbtt: There is not iter() procedure or iterator, as far as I'm aware.
22:06:20	Varriount	vbtt: Yes, but the iterator map() accepts has a known number and types of arguments
22:06:56	Varriount	One argument, that accepts item in a sequence given to map()
22:07:37	Varriount	In order to chain iterators, you would have to go through each one-by-one, accumulating the result of each and passing sequences.
22:08:35	Araq	Varriount: oh that's what you mean
22:09:51	Demos	can we steal things from Haskell's various iteration libraries or is that too much type wankery (I know little about how haskell does this, only that their implementation is very general and composable, and probably slow)
22:10:13	dom96	OT: Anyone here watch House of Cards?
22:10:42	zahary	ideally expressions such as map(foo, …).filter(…).group_by(…) should be lazy and shouldn't produce any allocations
22:11:03	dom96	Varriount: Thanks for the edits :)
22:11:06	EXetoC	enter type class interfaces?
22:11:22	Varriount	dom96: Feel free to strike down any that are two ambiguous
22:11:34	dom96	We will need to make those ideas more in-depth though
22:11:46	dom96	or else we risk getting rejected as an org
22:12:59	Varriount	How in-depth?
22:13:14	dom96	http://community.kde.org/GSoC/2011/Ideas
22:13:59	Varriount	dom96: You want the ideas formatted like that?
22:15:09	dom96	Perhaps not exactly like that, but a explanation of the project or idea would be nice.
22:15:23	dom96	"Fix bugs with iterators/generics" is a bit vague.
22:15:41	dom96	It needs to be concrete, something we can measure.
22:16:06	vbtt	"fix bugs" is not going to attract many takers.
22:16:18	dom96	indeed.
22:16:45	vbtt	new, cool sounding features will.
22:16:47	Demos	well you can word it differently, it /is/ the most pressing issue
22:18:02	dom96	hrm, i'm not sure how this is going to work. What if someone completes the ideas listed on that page before GSoC starts?
22:18:51	Mat3	ciao
22:20:55	*	radsoc quit (Ping timeout: 260 seconds)
22:21:06	BitPuffin	zahary: are you fixing generics?
22:22:04	Demos	well I don't think that is going to happen. And there are other related things such as "add support for spiffy generic features", "improve babel", "write a better random number generation library", "improve the time/date library", "improve c2nim", "improve the importcpp pragma (having type Foo[A, B] = object {.importcpp:"SomeGeneric<A,B>".} would be grand)"
22:22:09	*	Mat3 quit ()
22:26:05	Varriount	Demos: CryptGenRandom ftw
22:26:21	Varriount	dom96: Updated again, what do you think?
22:26:38	Demos	Varriount, is that a thing?
22:27:04	Varriount	Demos: There's a easily-wrapped random number generator in the windows api
22:27:24	Varriount	Or rather, crytographic random data generator
22:27:54	Demos	right, the nimrod API needs to be good. Also we want more than a crypto secure generator.
22:28:20	Demos	having access to a good PRNG is no good if everyone is running it once and modding the output
22:29:18	Araq	Demos: what do you mean "improve c2nim"? it's perfect. you might however simply wrap and use clang's parser instead
22:29:42	BitPuffin	"it's perfect" - biggest lie 2014
22:30:03	Araq	this is the much better solution when you try to translate stdio.h
22:30:21	Demos	yeah that was what I was thinking keep c2nim as a lightweight converter and have like clang2nim
22:30:35	Demos	anyway my point was that nimrod has no shortage of projects to get done
22:31:11	Araq	dom96: can you mentor a parser generator?
22:31:23	Araq	in fact, I want re2c for nim
22:32:16	reactormonk	Araq, re is what?
22:33:40	Araq	and a nim backend for Ragel
22:33:49	Araq	reactormonk: "regular expression"
22:34:07	reactormonk	Araq, ah, a code generator for regexp
22:37:47	Varriount	Demos: How does the page look now?
22:38:52	Demos	standard library is OK
22:39:32	Demos	what about making the docgen use huristics and comments to group "methods" together?
22:39:42	dom96	Araq: I think so
22:39:53	Demos	by huristics I mean "is the first param named self or this"
22:40:03	filwit	well, huh.. Godot engine looks pretty cool. Even the scripting is python-like.
22:40:10	dom96	Varriount: Get rid of those bullet points in front of headers.
22:40:14	dom96	*headings
22:40:41	EXetoC	Demos: that's so verbose. how about just 'o' :-)
22:41:39	Demos	EXetoC, well false positives. The idea would be to have some kind of annotation in the doc comments
22:42:38	EXetoC	I think just a list of all variations would be fine, but maybe I'm too optimistic
22:42:45	dom96	Varriount: That's better
22:43:16	EXetoC	and a slightly more flexible output
22:43:28	shodan45	filwit: I tried to look at the code for some of their examples, but only found xml files with vaguely code-like stuff in them
22:44:02	reactormonk	Demos, I'd prefer an index that gives you every proc that takes a type
22:44:14	Varriount	Will no-one else add ideas?
22:44:26	reactormonk	Varriount, documentation improvements all over.
22:44:36	reactormonk	or adapting yard or something similar
22:44:38	Araq	Varriount: I missed something here
22:44:38	Demos	right, but there is the issue of procs where the first param is in no way special
22:44:44	Demos	like pow(x,y) or whatever
22:44:59	Araq	what is the site you're all looking at?
22:45:05	filwit	Godot is like exactly what I'm writing for Nimord, even the scripts look familiar to my goals.
22:45:07	filwit	I can build games on Linux and publish to Windows, Mac, Android, etc.. that's cool
22:45:16	BitPuffin	what's the way to deal with parameters sent to a program
22:45:19	Varriount	Araq: Github
22:45:20	BitPuffin	like ./mysoftware bla
22:45:20	dom96	Araq: https://github.com/Araq/Nimrod/wiki/GSoC-2014-Ideas ?
22:45:23	BitPuffin	how do I get the bla part
22:45:27	filwit	shadon45: the examples are in the main git
22:45:39	filwit	shodan45* ^
22:46:13	Varriount	dom96: I can't fill out the "implement coroutines" idea, since I have no idea how to go about implementing it. :/
22:46:16	filwit	called 'demos', then the scripts are .gd
22:46:56	shodan45	filwit: I'll look again... I swear some of them only had images + an xml file
22:47:15	filwit	shodan45: https://github.com/okamstudio/godot/blob/master/demos/3d/platformer/enemy.gd
22:50:02	BitPuffin	.gd?
22:50:12	filwit	Godot engine
22:50:16	filwit	makes sense
22:50:19	BitPuffin	oh
22:50:21	BitPuffin	yeah
22:50:23	BitPuffin	I saw that
22:51:32	BitPuffin	is it getOpt?
22:52:36	Varriount	Ooh, I have a project idea. It was originally gradhas. Allow nimrod to include the standard library into the nimrod executable at bootstrap, and use those modules (unless a new one can be found)
22:54:52	Demos	Oh, fixing the way module loading works would be good as well
22:55:34	BitPuffin	WHAT
22:55:36	BitPuffin	parseopt
22:55:38	BitPuffin	is deprecated
22:55:44	dom96	parseopt2
22:55:57	dom96	also
22:56:10	dom96	There is a nice library for this stuff in babel
22:56:24	BitPuffin	parseopt2 lol
22:56:28	BitPuffin	dom96: what's the name
22:56:35	dom96	github.com/fenekku/commandeer
22:57:17	BitPuffin	dom96: I think babel should do a nimrod doc automatically or something
22:57:19	BitPuffin	would be sweet
22:58:21	dom96	perhaps
23:00:53	BitPuffin	dom96: then you could just create a symlink to the docs directory and it would get updated as pkgs were added :D
23:01:27	Araq	"Implement support for the JavaScript backend in the tester" wtf?
23:01:38	Araq	the tester supports that already
23:01:52	EXetoC	that was easy
23:01:54	EXetoC	next
23:02:08	dom96	Araq: It does?
23:02:18	Araq	yes
23:02:27	dom96	Are there any tests for the JS backend?
23:02:40	Araq	there is even support to run the *same* test for multiple targets
23:02:50	Araq	including the JS target
23:02:50	EXetoC	oh yeah
23:03:09	Araq	dom96: yes there are
23:03:15	dom96	hrm, ok.
23:04:14	dom96	Araq: I don't think you should be writing how these things should be done.
23:04:22	dom96	Araq: You should be writing what the end result should be.
23:04:39	Araq	nah, poor students will get it all wrong
23:04:59	dom96	That's why you mentor them...
23:05:35	dom96	You know, students pick where they go.
23:05:48	Araq	well I'll keep these notes for myself then, fine
23:06:00	dom96	Step-by-step instructions will probably discourage them
23:06:21	Varriount	Or make them underestimate the challenge of the task
23:07:04	Araq	hey, it's simple :P
23:07:09	Araq	hardly a challenge
23:07:19	EXetoC	1. add features 2. fix bugs 3. release 1.0
23:08:16	dom96	oh well, just dump your ideas in there for now.
23:08:32	dom96	We can refine things later.
23:08:37	dom96	And i'm too tired to do it now.
23:08:52	Varriount	dom96: What about the rest of the application? Do you need help with it?
23:09:04	*	Varriount *really* wants this to happen
23:09:36	dom96	I'm worried about the W8-BEN IRS form I will have to fill out.
23:09:43	dom96	I think that's after we get accepted though.
23:11:27	Varriount	If I get to work on nimrod as part of the GSoC, it means that I won't have to get a job over the summer, and lose (even more) time to nimrod.
23:12:08	dom96	Yeah, the great thing is that you already know Nimrod. If we get another student they will need to spend a fair share of their time learning Nimrod.
23:12:25	dom96	Perhaps we should just request 1.
23:12:32	dom96	But then that feels a bit... unfair. lol
23:12:32	Varriount	I can help teach them
23:12:45	Araq	pfff nimrod is dead simple to learn
23:12:47	Varriount	Nimrod is not that hard to learn.
23:13:11	Varriount	No harder than Java is, and much less likely to alter your brain
23:13:38	dom96	I really want this to happen too. Not only because of the t-shirt but it means I will probably have an easier time getting in to the GSoC as a student next year.
23:14:06	Varriount	It seems to me that using Java tends to alter (and restrict) the mind of a programmer when programming, which is a shame.
23:15:11	Varriount	I imagine that to a hard-core java user, nimrod is quite scary - no classes, seperation of 'methods' and objects, etc.
23:15:16	NimBot	Araq/Nimrod devel a600ce5 Zahary Karadjov [+1 ±4 -0]: fixes #797; generic procs can be used in places expecting matching concrete proc types
23:17:07	EXetoC	my mind does indeed seem very unaltered atm
23:17:19	Demos	I am also a student. Although I have far less experience in nimrod than Varriount
23:17:45	Varriount	Demos: Not that much less experience.
23:18:09	Demos	you came from python right?
23:18:13	Varriount	Yeah.
23:19:16	Araq	zahary: var prc = if arg.kind in nkLambdaKinds: arg[0].sym else: arg.sym
23:19:46	Araq	is slightly wrong as nkIteratorDef can be used anonymously
23:20:25	dom96	Demos: You're welcome to participate too if you'd like.
23:20:29	Araq	in fact, I want to get rid of nkLambda, nkProcDef suffices, if n[namePos] is empty, it's anonymous
23:21:12	Demos	well apperently students apply next month. But yeah GSoC sounds great
23:21:20	Araq	which at that point shouldn't be the case anymore
23:21:58	*	shodan45 quit (Quit: Konversation terminated!)
23:21:58	dom96	If we get accepted it'll be fun :)
23:24:02	Araq	good night
23:29:53	BitPuffin	uh
23:29:55	BitPuffin	what the fuck
23:30:04	BitPuffin	how do I make an optional argument with commandeer
23:30:12	Varriount	commandeer?
23:30:15	dom96	aww yeah. Finished Physics homework in 5 mins.
23:30:16	BitPuffin	like I can run the command like "bla" or "bla server"
23:30:20	BitPuffin	dom96: noob
23:30:34	dom96	BitPuffin: u wut m8
23:31:45	*	xenagi joined #nimrod
23:34:43	dom96	Reddit always makes AMAs funny. Gotta love the Bill Gates AMA.
23:35:43	xenagi	why is that one funny?
23:36:40	BitPuffin	fucking commandeer
23:36:45	dom96	People make lots of funny jokes.
23:36:46	BitPuffin	let me make an optional argument
23:39:17	dom96	This is pure gold: http://www.reddit.com/r/IAmA/comments/1xj56q/hello_reddit_im_bill_gates_cochair_of_the_bill/cfbsntu
23:40:23	BitPuffin	/Users/isak/src/nim/combat-cravings/src/combat
23:40:24	BitPuffin	Not enough command-line arguments
23:40:26	BitPuffin	Error: execution of an external program failed
23:40:28	BitPuffin	._.
23:40:47	dom96	BitPuffin: Fork that shit.
23:40:56	BitPuffin	dom96: fuck that shit
23:43:01	dom96	BitPuffin: noob
23:45:35	EXetoC	parseopt/parseopt2? just migrate later
23:46:16	BitPuffin	dom96: no u
23:47:14	BitPuffin	https://gist.github.com/BitPuffin/888fe85faab4438925b6
23:47:21	BitPuffin	get a read from nil ._.
23:48:26	BitPuffin	wait now it works
23:58:55	*	darkf joined #nimrod
23:59:19	EXetoC	just use discard, but nil works for now