<< 01-03-2014 >>

---

00:03:51	Araq	good night
00:05:52	*	tinAndi quit (Quit: ChatZilla 0.9.90.1 [Firefox 27.0.1/20140212131424])
00:06:15	*	Puffin quit (Quit: WeeChat 0.4.3)
00:06:34	*	BitPuffin joined #nimrod
00:12:34	*	darkf joined #nimrod
00:19:19	*	EXetoC1 joined #nimrod
00:21:14	*	EXetoC quit (Ping timeout: 264 seconds)
00:25:11	*	nequitansMobile joined #nimrod
00:31:56	*	skyfex quit (Quit: Computer has gone to sleep.)
00:43:36	*	nequitansMobile quit (Ping timeout: 240 seconds)
00:49:12	*	EXetoC1 is now known as EXetoC
00:49:13	*	icebattle quit (Quit: leaving)
01:14:52	*	io2 quit ()
01:26:31	*	nequitans_ joined #nimrod
01:27:14	nequitans_	Ah, Araq, I fixed the indexing issue (thanks), but that wasn't the cause of the slowdown. Your guess about the random number generator appears to be the cause
01:42:48	*	Matthias247_ quit (Read error: Connection reset by peer)
01:46:21	*	DAddYE quit (Remote host closed the connection)
01:46:37	*	zielmicha quit (Quit: Connection closed for inactivity)
02:32:55	*	XAMPP-8 joined #nimrod
02:34:25	*	XAMPP-8 quit (Client Quit)
02:38:52	*	nande quit (Remote host closed the connection)
03:00:16	*	Tyler__ quit (Quit: Leaving)
03:08:52	*	xtagon joined #nimrod
03:16:33	*	DAddYE joined #nimrod
04:09:57	*	q66 joined #nimrod
04:19:05	*	skrylar joined #nimrod
04:19:06	*	skrylar peeks in for a moment
04:19:57	skrylar	Araq: so what was the final verdict on --cs:partial; are we losing the neat style-insensitivity or has forcing a naming style become a failed experiment?
04:21:55	*	comex_ is now known as comex
04:23:12	Demos_	I think araq is asleep
04:26:44	skrylar	i'll just have to send dream ghosts to haunt
04:27:05	*	filwit joined #nimrod
04:27:54	filwit	skrylar: i think removing the T/P prefix would be best for adoption
04:28:01	filwit	(just read the logs)
04:28:38	filwit	besides, local variables should not be named ambiguous things like "file" anyways.
04:29:14	Discoloda	that is what originally made me ignore Nimrod, case insensitive identifiers (why, i imagine, T/P became convention)
04:29:15	filwit	so: var prefsFile = File.open("..") shouldn't cause naming conflicts, case-sensitive or not
04:30:02	filwit	(not really valid code, but you get the idea)
04:30:56	filwit	Discolada, yeah the Hungarian notation was always a sore spot for me, but i really like the case-insensitivity (though it probably does cause some conflits
04:31:53	filwit	still, the compiler should be able to figure this out: var file: File; open(file, "...") # because this overload doesn't take a typedesc, it implicitly accepts the local var
04:32:02	Discoloda	i guess my C code is worse, i dont use typedefs for structs so i end up with "struct foo" everywhere
04:32:35	filwit	Discolada, I've seen a lot of major C project do the same though (wayland for instance), not sure why
04:33:14	Discoloda	i feel its more explicit and uses the struct/union namespace for those types
04:35:17	filwit	brb
04:38:50	filwit	Ultimately I never really saw the point of T/P prefixes to begin with. It seems to be a left-over from when Nimrod didn't have 'type T = ref object' and you had to make a seperate alias (PFoo = ref TFoo) for everything
04:39:09	skrylar	Discoloda: what made you eventually decide to un-ignore nimrod?
04:39:32	skrylar	filwit: i actually like pascal type names, but it's not something i'm going to riot over or anything.
04:39:34	Discoloda	c2nim
04:40:12	skrylar	there's a lot of C/C++ code that has "CThisClass", enums have hungarian notation "e.g. ThisEnum = teDoAThing" so its not exactly unheard of
04:40:34	filwit	skrylar: yeah i saw your forum post
04:40:49	skrylar	Qt names everything QBlah and also notates enum flags
04:41:12	filwit	skrylar: i don't see the point of a prefix at all unless you have a global (and i have a better idea for that anyways)
04:41:26	skrylar	Discoloda: sounds like nimrod is full of those weird things that look like dealbreakers at first and then you come back two weeks later and realize they aren't, lol. For me it was the C compilation step
04:41:35	skrylar	filwit: for enums or for type names?
04:41:35	filwit	skrylar: "global" means "singleton"
04:41:45	filwit	skrylar: both?
04:41:51	Discoloda	the c compilation is also what brought me back
04:42:16	skrylar	i thought the point of prefixing enums was so you could write taJustified instead of JustifiedTextAlignment or TextAlignment.Justified everywhere
04:42:39	filwit	i prefer the latter really
04:42:53	skrylar	Haskell's algebraic types lead to that awful JustifiedTextAlignment style of naming too
04:43:32	filwit	are you really typing TextAlignment.Justify so often that it becomes a chore?
04:43:33	skrylar	but i'm an old pascal/delphi derp, so i tend not to mind pascalisms
04:43:53	skrylar	filwit: depends on if i'm using an awesome text editor or not
04:44:05	filwit	plus, this problem disapeers when you have intellisense/code-complete
04:44:13	skrylar	I didn't mind as much with Sublime's fuzzy text complete, but its a chore in everything else (e.g. vim)
04:45:04	filwit	meh, I use Kate at the moment but am used to Visual Studios. Both basically means i start typing three characters then hit tab most of the time
04:45:07	skrylar	It also looks stupid when you have to type ThisEnum.DoAThing | ThisEnum.DoAnotherThing when you're composing a set of obviously the same enum type
04:45:37	skrylar	eh, i use Ctrl-P in vim which does that but sublime does one better
04:45:48	filwit	well that's why you can just use DoAThing | DoAnotherThing unless there's a conflict
04:46:07	skrylar	it uses a fuzzy match instead of a prefix match, so you can type "sfm<tab>" and it will go "Oh, SuperFlippingMachine"
04:46:57	filwit	honestly, what the compiler should do as far as naming resolution is this: if the function takes an enum, you don't need to prefix the enum name and it doesn't conflict with local names when they're of non-acceptable types
04:47:20	skrylar	doesn't that add extra symbol table and scoping rules though?
04:47:42	skrylar	one of the pascalisms is to avoid symbol table lookups because they exponentially slow the compilation of huge codebases
04:47:48	filwit	so if you have: var justify = 10; foo(Justify) # take TextAlign enum - no conflict since local is type int
04:48:26	filwit	skrylar: you could be right, i'm not sure. I haven't really thought about the performance implications.. but honestly i would accept a slower compiler for this kind of thing
04:48:38	skrylar	slower compilers have to be accepted very carefully
04:49:00	filwit	true, but i doubt it would be a huge deal unless you had a shit-ton of name conflicts?
04:49:07	skrylar	specifically what we all *don't* want is a "well lets put this in for that feature, now lets tie everything in to it since we already have it" and then the compiler becomes C++0x
04:49:35	skrylar	i'm not objecting to it intelligently solving the name conflicts btw
04:49:59	filwit	plus, i really don't see how that would slow things down... it just looks at the proc overloads and says: if this takes an enum, resolve to that enum type
04:50:17	filwit	otherwise it looks for locals
04:50:32	skrylar	i have no idea how the AST passes run though i should probably look some time
04:50:42	skrylar	i know macros don't get access to certain type information from parameters
04:51:37	skrylar	having an enum value the same identifier as a structure or function name could confuse what "let x = Foobar" means
04:52:05	skrylar	(admittetly other languages already solve that by just saying its ambiguous and listing off the potential matches)
04:52:22	filwit	the only area i ever need to use 'TFoo' is if i have a type that's design to be a singleton.. and that could be resolved if Nimrod eventually adds "type Foo {.global.} # creates singleton and all access is to the global var not typename"
04:52:49	skrylar	I don't think thats something you would need compiler help for
04:53:00	skrylar	that sounds like something a regular macro could do
04:53:07	filwit	yeah, good point
04:53:22	Demos_	I like T/P prefixes
04:53:25	skrylar	if a regular macro can't do that, it *should* be able to
04:53:34	filwit	plus it's not a huge deal to just make "TFoo; var Foo*: TFoo"
04:53:34	Demos_	although I don't see what is so bad about ref TBlah
04:53:51	Demos_	also I dont like the ref object syntax, since it limits what people can do with the object
04:54:05	filwit	Demos_, how does it limit?
04:54:19	skrylar	Demos_: that goes back to a pascal-ism where pointer types had to be typedef'd because it allowed the compiler to sweep everything in one-pass with simpler grammar
04:54:19	Demos_	it enforces refernce semantics
04:54:21	filwit	i'm pretty sure you can get the non-ref type
04:54:26	filwit	i just can't remember how
04:54:53	skrylar	you can do a much simpler grammar if you only need to accept an identifier
04:55:08	skrylar	which is the legacy of PFoo
04:55:12	Demos_	still, I really dont mind the prefixes
04:55:30	skrylar	I don't either, but its not really up to me. lol
04:55:40	filwit	Demos_: the biggest problem with the T/P thing is adoption concern. I mean most people can "live" with it, but they're not going to move away from "language X" for something they feel is straight out of the 80s (not my words)
04:55:46	Demos_	I dont understand why people have their panties in a bunch
04:55:56	Demos_	oh dear god
04:55:58	skrylar	filwit: if thats the only reason, i don't think its legitimate
04:56:10	Demos_	"the syntax is not C so it must be slow"
04:56:21	skrylar	Once you're so far outside of C syntax, changing one tiny thing isn't going to make them go "oh well i've changed my mind then!"
04:56:38	filwit	skrylar: i think it's very ligament if Nimrod hopes to gain popularity and eventually be funded somehow
04:56:47	Demos_	these are the people who are going to bitch about every way in which nimrod is different from java
04:57:02	skrylar	filwit: but D is a C-like and doesn't have commercial support either
04:57:16	Demos_	skrylar, well...erm...facebook
04:57:20	Demos_	a little
04:57:26	filwit	skrylar: D actually offers commercial support now
04:57:37	skrylar	outside of digital mars?
04:58:04	filwit	not sure if anyone pays for it or not, but Remedy games is hireing D devs.. so..
04:58:25	Demos_	interestingly I found nimrod really easy to pick up dispite never having used another language with a similar syntax
04:58:40	skrylar	I'm not an expert on language marketing, but I'm still fairly sure people make their decisions on tooling support more than whether there is a T prefix
04:58:46	filwit	skrylar: i'm not sure what you mean by "outside digital mars" you mean DMD?
04:59:05	skrylar	i remember araq posting about Delphi still being marked as the 17-th most used language or something
04:59:43	Demos_	well delphi was kinda killed by having too much commercial support right?
04:59:47	filwit	skrylar: of course there are many reasons people use one language over the next. The point is what is statistically the best option.
04:59:54	skrylar	Demos_: Delphi compilers are bumfooz expensive
05:00:18	skrylar	even back in delphi 5 days they wanted a few thousand for a single developer
05:00:23	filwit	skrylar: I'm very unfamiliar with the Delphi/Pascal world, so i can't really comment
05:01:14	skrylar	filwit: 'statistically the best' is how you get to live with C++ :P
05:01:40	filwit	all I'm saying is this: statistically speaking, we would *probably* get more interest if we removed the T/P thing, right? And then you folks could still use it if you wanted too.. so it's a win-win
05:01:43	Demos_	christ over $4,000 for delphi
05:02:10	skrylar	filwit: we really couldn't; it would schism from the standard library
05:02:23	Demos_	we probably will not get more intrest by doing that but whatever, Also, I want to point out that it is useful in documentation browsing
05:02:28	filwit	skrylar: no, i don't just mean pure democracy. I'm not saying lets just take a poll and do whatever that leads too. I'm saying it's relevant tho
05:03:03	skrylar	filwit: yeah but that same argument could say why are 2x spaces the compilers indent marker?
05:03:10	skrylar	we could get more interest if we supported tabulators
05:03:11	skrylar	etc
05:03:46	filwit	skrylar: accept that has demonstrably negative implications, while removing T/P just means more adoption
05:03:53	Demos_	I thought it was just alignment, if you want to use 4x spaces you can afaik
05:04:11	skrylar	as I said the other night, you need a 'killer app' and then it doesn't matter what decisions you made along the way
05:04:26	skrylar	Ruby is dog slow and always has been, kind of quirky in many ways, and 37signals made everyone stop caring
05:04:26	filwit	skrylar: completely agree with that
05:04:26	Demos_	right, except that T/P helps when you are style-insensitive, which I personally like
05:04:49	Demos_	what was c++'s killer app?
05:04:55	skrylar	the thing with Matz is he never sat and said "What would make my language more appealing to C coders?", he's said up front "I wrote this to make it fun to use"
05:05:07	filwit	Demos_, it doesn't actually help with case-insensitive though
05:05:10	skrylar	Demos_: C++'s benefit was that it stacked on to C
05:05:31	skrylar	the alternative was Objective-C, but the two guys who made Obj-C wanted extortionate fees
05:06:35	skrylar	(I don't actually have a super-strong lamentation about dropping T/P, I just don't think that people are choosing against Nimrod because of syntax. I mean people are choosing Rust and the syntax is godawful for that.)
05:06:58	skrylar	They didn't even consider what looked like code, they just slapped a parser over text and went "welp it works lets keep it"
05:07:42	filwit	skrylar: i almost passed up Nimrod because of this sort of thing, and I've met a bunch of people who have complained about it a ton. The only people who seem to like it don't really care either way, so we should just remove it and everyone is pretty happy :P
05:07:57	skrylar	filwit: as long as i get to keep my style insensitive parser
05:08:26	filwit	skrylar: agree there. I always liked the case-insensitive thing Nimrod's got going on
05:08:42	filwit	it means everyone can agree-to-disagree
05:08:47	Demos_	I will probably keep using T/P prefixes
05:08:48	filwit	:P
05:08:50	xtagon	What's the benefit of style insensitive parsing anyways? Do people commonly typo their capitalization?
05:08:57	Demos_	also, lets remember that C coders are used to this sort of thing
05:09:00	skrylar	xtagon: it removes the argument over style
05:09:01	Demos_	namespaces and whatnot
05:09:27	filwit	xtagon: if i really like camelCase but you really like snake_case you can use my lib how you like
05:09:30	Demos_	and I find it nice to be able to just type the funciton name without looking up capitalization in the docs
05:09:36	xtagon	Ah
05:09:47	skrylar	xtagon: basically, people can scream blue in the face over "official styles" but nobody cares; thats a "me vs you" opinion argument. If the compiler is style insensitive, then it simply *does not matter* and so we can do work instead.
05:10:17	filwit	^ exactly
05:10:23	skrylar	And yes, there is also the benefit that six different libraries will look the way *you* want to use them, instead of the miniature hell of "SDL_SemiCaps(ftdothings())"
05:10:24	Demos_	I really just want generics to work correctly
05:10:40	Demos_	or rather I want a definition of "working correctly" and then I want that
05:10:50	skrylar	lol
05:10:51	filwit	plus it makes things more care-free in terms of compiler mistakes (accept that Nimrod also it very touchy in other areas like comment indentation)
05:11:06	skrylar	there is at least a good reason for that; it lets comments be used for things
05:11:37	skrylar	all though elastic tabstop would be pretty cool too
05:11:43	filwit	skrylar: i think it should be optional, and off by default
05:11:49	Demos_	essentially I think that nimrod does things so differently (dare I say correctly) that people who are really used to the way C/C++ does things will be spooked
05:11:59	skrylar	filwit: off by default just brings back the arguments
05:12:03	xtagon	I'd love to be able to use tabulators for indentation. But I doubt anybody cares or feels the same way
05:12:15	Demos_	there is a very good reason for disallowing tabs
05:12:28	Demos_	if you accedently mix tabs and spaces it creates hell
05:12:47	filwit	skrylar: well the biggest (and only) problem is IDE adoption. Most text-editors aren't going to know exactly where to place the '#' unless they understand the AST
05:12:55	skrylar	Demos_: moonscript solved that by defining what a tab means to the compiler and documenting it; but that's just a style&tool argument again
05:13:02	xtagon	I've never accidentally mixed tabs and spaces. Makefiles use tabs for indentation and I've never had an issue with them
05:13:02	filwit	skrylar: i use Kate, but can't use block-comment/uncomment because of this
05:13:02	skrylar	Demos_: but thats not what an elastic tabstop does anyway
05:13:34	skrylar	elastic tabstop is a neat concept that basically uses a character as an indendation marker, then it automatically aligns those markers up in the renderer
05:13:51	filwit	xtagon: mixing tabs and spaces is a very bad idea with white-space-significant code
05:13:52	skrylar	so you get the option to set indent level without borking text alignment
05:14:00	skrylar	but this requires tool support
05:14:03	Demos_	liek '&' in LaTeX's align envirenment
05:14:42	Demos_	I am impressed with nimlime and VisualNimrod is coming along again
05:15:07	filwit	Demos_: there's a VS plugin in the works?
05:15:32	skrylar	I think a decent nimrod IDE that supported elastic tabstop, the style-insensitive parser and integrating nimrod pretty would lead to a pretty argument-free environment
05:15:42	filwit	Demos_, if so, that's awesome (though i would much prefer to see a cross-platform IDE like MonoDev supported over VS)
05:15:43	skrylar	that was basically what the SCID concept was meant to do
05:16:06	Demos_	yeah it supports highlighting, building, and project navigation
05:16:19	filwit	Demos_, that said, VS support would probably go much further than MonoDev support in terms of gaining attention
05:16:21	Demos_	should have completion by the end of the week (I hope)
05:16:38	filwit	Demos_, nice man, good work!
05:17:01	*	r0b1 quit (Ping timeout: 264 seconds)
05:17:09	skrylar	anyhows, i should get back to reinstalling windows on this machine; ubuntu has given me too many migraines lol
05:17:27	filwit	that's cause Ubuntu :P
05:17:33	filwit	use Manjaro
05:17:35	skrylar	yeah i used to use mint
05:17:41	xtagon	When GNU/Linux gives you migrains, just try another distro ;)
05:17:45	Demos_	yeah I started it a few months ago and it has been kinda on hold with classes. The VS plugin API makes me want to stab myself with a rusty spork
05:17:50	skrylar	But the fact that you have to play whack-a-distro is one of the primary failures of linux
05:17:52	filwit	Mint is not much better, unless you use Mint Debian
05:17:59	filwit	i can't stand non-rolling distros
05:18:12	xtagon	I consider the fact that I get to play whack-a-distro until I find one I like a feature of Linux
05:18:20	Demos_	yeah I installed ubuntu and it was crashing and panicing everywhere, went back to arch
05:18:32	filwit	skrylar: yeah the many distro game is a problem, but it's getting better really
05:18:45	Demos_	you could use a *BSD
05:18:47	skrylar	I kinda like how the BSDs take responsibility over their system and you can safely say "I use FreeBSD" or "FreeBSD has this problem" without people getting in to the whole "NO NO THIS IS JUST THE KERNEL" b.s.
05:18:51	filwit	skrylar: but if you like Windows, by all means use it. I would just dual-boot though for testing
05:19:21	skrylar	filwit: i actually don't like windows, but the tools i need just aren't here on nix
05:19:27	Demos_	I /really/ want windows to support booting off something like btrfs or zfs. Never gunna happen thouhg
05:19:41	filwit	skrylar: accept that BSD is quickly becoming the same thing as Linux in that regard. It's all about popularity and open-source that causes "fragmentation"
05:19:53	skrylar	I mentioned doing audio work a while back, and linux is about on par with studio software 20 years ago. lol
05:20:16	filwit	skrylar: interesting. all my best tools are Linux things these days
05:20:39	skrylar	filwit: i haven't found a good non-orgmode outliner, and most of the audio softsynths i ran in to were poorly built
05:20:43	filwit	skrylar: but mostly because i chose to migrate to linux awhile ago and haven't used Windows in awhile
05:20:58	filwit	(beyond VS and Unity3D that is)
05:21:17	skrylar	i actually still miss my OSX; but eh.
05:21:27	filwit	skrylar: i don't even know what a non-orgmode outler or audio softsynths is.. lol
05:21:44	skrylar	filwit: outliners are for organizing / heavy writing
05:22:05	filwit	skrylar: have you used KDE and KAuthor?
05:22:18	skrylar	filwit: and a softsynth is kind of like a computer version of a guitar/piano
05:23:18	skrylar	filwit: KDE yes, KAuthor no
05:23:53	filwit	skrylar: my main tools, besides coding, are art tools, and I switched to FOSS awhile ago and haven't needed to look back. Blender/Gimp/Inkscape/Krita all run much better on Linux than Windows, so I stay in Linux most of the time I can
05:24:05	skrylar	Krita is nice, I will admit
05:24:48	filwit	skrylar: yeah, KAuthor is for heavy authoring, not sure if it's what your looking for but it's pretty nice in terms of usability (granted i don't have much to compare it too)
05:25:01	filwit	skrylar: what tools do you use on Windows for these things?
05:25:02	skrylar	i can't even find kauthor with targeted searching
05:25:24	skrylar	filwit: well on my mac, i used notebook http://www.circusponies.com/ and tinderbox
05:25:28	skrylar	well, wanted to use tinderbox
05:26:06	skrylar	Notebook is vaguely Onenote with bulleted lists on steroids and a neat app skin to look like a book
05:26:09	filwit	wait.. so you want a fancy notepad program?
05:26:30	filwit	why not use Google docs and stuff?
05:26:34	Demos_	my advice is to get a paper notepad
05:26:40	filwit	^ lol
05:26:40	Demos_	nothing like actually writing stuff down
05:26:46	Demos_	seriously
05:26:48	skrylar	bleh, i hate handwriting :)
05:27:36	filwit	i can understand switch OSes for the sound tools, but a notepad program? there's got to be something on Linux that is basically good enough unless you're really picky about features
05:27:57	filwit	what about the Benjin program or whatever it's called?
05:28:13	skrylar	never heard of it
05:28:13	filwit	I'm on Windows right now, or i would look through my KDE programs
05:28:21	filwit	let me look it up
05:28:41	skrylar	But yeah, 300$ of audio software is my present motivation to switch the computer back over
05:28:58	skrylar	especially after spending a good six months demoing linux options
05:29:26	skrylar	with great hits like qtractor deciding i don't actually need those MIDI notes to get recorded when i ask it to. haha
05:29:29	filwit	skrylar: http://worldofgnome.org/bijiben-3-11-5-and-owncloud/
05:29:54	Demos_	YES! I just figured how the fuck to get a reference to the active project in VS
05:30:03	Demos_	jesus christ that was harder than it should have been
05:30:15	Demos_	two casts and the use of an internal object
05:30:23	filwit	skrylar: yeah if you have a lot of audio windows-only tools you should probably use that. Still i would recommend dual-booting
05:30:28	filwit	Demos_: nice!
05:30:47	skrylar	i might set up a dual boot again some time later
05:30:52	filwit	Demos_: ps. no idea what you just figured out, but good job!
05:30:54	filwit	lol
05:30:57	Demos_	skrylar, you could run windows in KVM
05:31:01	skrylar	what i've been doing so far is one computer was dedicated ubuntu and the other is dedicated windows
05:31:06	*	q66 quit (Quit: Leaving)
05:31:26	filwit	skrylar: dual-boot is really good for benchmarking different OSes
05:31:51	skrylar	i did actually write a book in linux using wine for quite a while
05:32:01	filwit	skrylar: you wrote a book?
05:32:10	skrylar	i did
05:32:13	skrylar	nanowrimo 2013
05:32:13	filwit	skrylar: i'm a bit book-worm :)
05:32:17	filwit	big*
05:32:28	skrylar	i was one of those scrivener derps for a while
05:32:37	Demos_	skrylar, you should write a book on nimrod once 1.0 comes out!
05:33:09	filwit	skrylar: mind if i ask what book you wrote?
05:33:20	skrylar	filwit: it was never published
05:33:26	filwit	ah, too bad :(
05:33:30	skrylar	i should still have it around to clean up as a PDF
05:33:45	filwit	what genre ?
05:33:53	filwit	i'm a fantasy/scifi guy
05:33:55	skrylar	fantasy/fiction
05:33:58	filwit	nice
05:34:11	filwit	i've always wanted to write a book.. but never have the time
05:34:39	skrylar	It really sucks when you decide to do it in a month and then realize the only thing you have to go on is the vague idea of something you did before
05:35:30	filwit	yeah you probably have to be pretty dedicated. I really love Brandon Sandersons work, and he's a machine with writing (like 5+ books a year)
05:35:47	skrylar	Then you find out, most people actually spend all year developing characters and outlining the story instead of just going "I WILL DO IT ALL NOW" xD
05:36:14	*	Demos_ quit (Read error: Connection reset by peer)
05:37:05	filwit	yeah. i've read a lot of novel (100+) and can narrow down my favorites to a consistent story-arch graph almost, haha
05:37:25	skrylar	theres a website about the snowflake method which goes over the 'correct formula' for fiction
05:37:44	filwit	i've often thought that if I ever wrote a book, i would just plot the story events along that graph and i'd have a best-seller haha
05:38:10	skrylar	there's cool stuff like Aeon Timeline that helps with that; it makes these pretty timelines and tracks events you throw in it
05:38:14	skrylar	I ended up not using it at all though
05:39:51	filwit	novels are probably like indi-games.. lot of competition and hard to actually get reconized
05:40:32	skrylar	just have to finish them, really; then go bother tons of agents until you get one that wants it
05:40:40	filwit	though idk, I've gotten some of those free-amazon books for kicks-n-giggles before, and they're always pretty lame (accept Emperors Edge, that was excellent)
05:40:49	skrylar	but i don't want to be a professional author, and they recommend writing several books in a row because it will take years to get the first one published
05:41:09	filwit	you can self-publish on Amazon now
05:41:17	skrylar	just because you can doesn't mean you should, lol
05:41:35	*	Demos joined #nimrod
05:41:37	Varriount|Mobile	I dream of writing a fiction novel one day...
05:41:37	skrylar	i remember seeing this one horrible book sample on amazon that was like
05:41:56	filwit	yeah seriously... some of those books i got off self-published amazon i couldn't make it a 5th way through they where so damn lame
05:42:09	skrylar	"due to political reasons we are unable to reproduce so i have come back from the future and we have decided to sex everyone up here for great glory" or some equally cringe-worthy "... wut?"
05:42:18	skrylar	yay crap-fi \o/
05:42:34	filwit	lol
05:43:09	skrylar	someone was trying to get me to write this one dream out as a book, which involved the planet's orbit stopping and so they built a huge axle to make it spin manually, until some cult decided to terrorize the world axle
05:43:26	filwit	i found this really good book series on there once (Emperors Edge) and ever since them I try the highest rated freebies out now and then... i don't even know why though.. they're all horrible..
05:44:15	filwit	skrylar: lol wut.. that sounds either very un-scientific or a very cool alien twist
05:44:28	Varriount|Mobile	The hope for finding the diamond in the dross
05:44:42	filwit	dross?
05:45:05	Varriount|Mobile	Term for junk found during mining.
05:45:35	Varriount|Mobile	Specifically metallic ores, iirc
05:45:36	filwit	(ps, we're completely spaming #nimrod right now... Araq's gunna be anoyed when he ways up trying to read the back-log, lol, sorry man ;P )
05:45:37	skrylar	filwit: steampunk physics in dreams \o/
05:45:57	skrylar	filwit: just git-rebase it all in to the offtopic channel that was never sanctioned lol
05:46:22	fowl	whatsw steampunk physics
05:47:00	Varriount|Mobile	I deal with sketchy physics in fantasy novels by just imagining that the entire book takes place in a universe with alternate physics.
05:48:27	Discoloda	#nimrod-offtopic exists
05:48:38	filwit	i really don't like books where things are completely undefined. I always hated things like LOTR where Gandolf can randomly shoot lightning out his ass and save the day in a way you never saw coming..
05:48:54	skrylar	asspull@tvtropes
05:49:11	filwit	Discolada: oh rly? thanks we should probably move this convo over there guys, really
05:49:34	skrylar	thats sort of a sad trope because the way you avoid an ass pull is to telegraph and that can get equally boring when you start calling al of the plot twists because of what they set up in advance
05:50:04	filwit	yeah, but if it's done cleverly it's a lot better, IMO
05:50:22	filwit	i really like Mistborn and Way of Kings cause they do it right
05:50:53	filwit	(ps, i can't figure out how to join a different channel on this crappy web IRC..)
05:51:18	Discoloda	. /join doesnt work?
05:51:35	filwit	i'm a IRC newb, didn't know you could do that
05:51:53	filwit	worked, thanks
06:08:45	Demos	wow pipelight is using 300% of my CPU
06:09:57	filwit	cause Wine
06:10:14	filwit	wait.. 300%.. how?
06:10:20	filwit	statistically impossible..
06:11:36	Demos	cause wine
06:16:49	Demos	I think it does CPU decodeing of the video
06:18:15	xtagon	Is there a builtin proc to unshift or shift a seq? All I see in the docs is pop
06:19:57	Demos	I doubt it, shifting is slow
06:23:45	*	xtagon quit (Quit: sleeeep)
06:24:16	filwit	Pipelight works pretty much perfect on my machine.. but then my machine is good and i'm using Arch
06:24:25	filwit	Never looked at the CPU usage though
06:24:35	Demos	it is smooth but the cpu usage is hight
06:28:28	filwit	The CPU usage is tooo damn high!
06:30:26	Varriount|Mobile	She can't take much more o' this capt'n!
06:31:06	Demos	I have been trying to get primary VGA passthrough working so I can just use netflix in KVM but primary VGA passthrough is hard to do since you need to disconnect the VGA from the host before connecting it to the guest
06:32:01	*	BitPuffin quit (Ping timeout: 264 seconds)
06:32:34	Demos	turned quality to low and CPU usage went down quite a bit
06:34:47	filwit	how do you adjust the quality?
06:34:59	filwit	right-click?
06:35:06	filwit	(not on linux or i'd try)
06:35:19	Demos	in account options
06:35:51	filwit	k i'll look next time i reboot
06:44:52	Araq	seriously guys
06:45:23	Araq	T/P are about value vs reference based semantics. That has always been the reason.
06:45:24	Discoloda	Araq?
06:45:46	Araq	I *copied* Pascal often and changed it into something meaningful
06:46:18	Araq	Delphi uses 'T' for classes which have reference semantics and so Delphi's convention makes no sense.
06:47:20	Araq	Likewise the need for symbol forwarding stems from nimrod's macro system.
06:47:47	Araq	"having a fast 1 pass compiler" was not the reason to do it this way and never has been.
06:48:04	Araq	Nimrod uses 3-5 passes anyway.
06:50:34	Araq	I don't disagree it looks like straight from the 80ies in places, but you should know better in #nimrod. ;-)
06:57:55	Varriount|Mobile	Hi Araq!
07:11:29	*	DAddYE quit ()
07:12:01	skrylar	i already removed the T/P from the skutf code on github; all though the styleinsensitivity is something i'm attached to
07:12:32	*	Varriount_ is now known as Varriount
07:12:49	Varriount	My goodness Earthbound has some wierd enemies.
07:13:36	skrylar	Earthbound (NES)?
07:13:42	skrylar	well, snes
07:13:58	Varriount	Yep. (On my phone, using an emulator0
07:14:43	Varriount	Earthbound and Mother 3 are the best rpg games I've ever played.
07:14:58	skrylar	earthbound is such a strange game overall
07:15:41	Varriount	"Demonic Petunia became Tame" <- WTF?
07:16:24	Varriount	It's the only game I've played where you have to face melting clocks and cups of coffee as enemies.
07:17:22	Demos	Sounds like "The Void"
07:17:50	Varriount	"The Void"?
07:18:39	Demos	Russian game
07:18:43	Demos	very strange
07:20:16	skrylar	that one sounded interesting but when i talked to people who played it, they said it was 'russian hard'
07:22:34	Varriount	Oooh, another strange game -> http://jayisgames.com/archives/2013/06/undertale_demo.php
07:25:26	Demos	I also recomend "The Precursors"
07:25:31	Demos	very good russian game
07:25:49	Demos	and with that I am oging to go to sleep
07:30:01	*	Demos quit (Ping timeout: 240 seconds)
07:59:15	Araq	skrylar: yeah I'm collecting data about the style-insensitivity now and will blog about it
07:59:53	Araq	title of the blog post is most likely "consistency is a poor man's approach to correctness"
08:02:08	Araq	zahary1: have a look at bug #966
08:02:20	Araq	seeems to be caused by the new delegators
08:55:01	Varriount	Araq: Is the style sensitivity option something that can be altered by the push and pop pragmas?
09:03:21	skrylar	Varriount: how would that work? if one module has strict mode on and another doesn't, what does strict mode have to do to use the nonstrict code?
09:04:11	skrylar	right now its a global compiler flag IIRC because it turns off normalizing identifiers
09:11:52	filwit	hey Araq
09:12:26	*	Endy joined #nimrod
09:17:12	Araq	hi filwit
09:17:19	Araq	you're doing a great job on reddit :-)
09:28:53	filwit	hey, glad you approve man
09:29:43	filwit	i realized after i came here that i didn't truly know what the final direction was for the T/P thing but my second post kinda hinted at being removed
09:30:00	filwit	i don't mean to advertise something that you're against
09:31:59	filwit	anyways, i saw that guys post about a "potential future language" and though it would be a perfect opportunity to plug Nimrod a little :)
09:32:36	filwit	seems like it was approved by the larger reddit community which is a good sign :)
09:33:59	filwit	i really wounder if that's because it's new to those people or because it's becoming familiar. I would guess that later
09:35:22	Araq	I'm playing with the idea to disallow upper cased letters ... lol
09:35:40	Araq	so there is only foo_bar_baz possible and everybody uses that
09:36:19	Araq	and ultimately we'll get rid of fonts that make _ look like a thing of the 70ies
09:36:27	filwit	i really think that would be "bad for business" and i personally wouldn't like it
09:37:12	filwit	wait get rid of fonts? are you just joking.. i'm a bit confused
09:37:27	Araq	I prefer fooBar too, but thinking about it, it makes perfect sense
09:37:44	Araq	well I'm not serioulsy suggesting this change, of couse
09:37:56	filwit	oh, i thought you where..
09:38:00	Araq	but it does make perfect sense...
09:38:21	filwit	idk really, what's so much better about it than have just case-insensitivity?
09:38:25	filwit	what does it fix?
09:38:48	Araq	why do we have Z and z in the first place? it just creates another option to misspell things
09:39:05	Araq	Chinese does fine without upper cased letters
09:39:39	filwit	yeah i actually thought once you could make a editor which would auto-inject the '_' mark
09:39:48	filwit	(when you press space)
09:40:03	filwit	so typing code would feel more like writing language
09:40:14	Araq	why do we have 2 ways to write the same letter? why not 3 or 5?
09:40:35	Araq	it's just some crappy cultural artifact
09:40:52	Varriount|Mobile	German has ß
09:40:55	filwit	haha, araq is getting deep
09:40:58	filwit	but it's true
09:41:14	filwit	it's pretty pointless if you think about it
09:41:41	rixx	Varriount|Mobile: yes, and russian has a different alphabet, and romanian has some letters from the cyrillic alphabet
09:41:49	filwit	plus theses day's it's just a psuedo-class thing. Oh, you didn't use correct grammer, you must be dumb
09:42:01	filwit	when no one texts with correct grammar
09:42:07	Varriount|Mobile	Pointless it may be, but you will never get me to willingly, purposefully use "i" instead of "I"
09:42:13	filwit	kinda a pointless formality
09:43:07	Araq	happiness comes from eliminating choices
09:43:31	Araq	I take away the upper letters and make everybody happy
09:43:58	Araq	Varriount|Mobile: ß has an uppercased representation now. but nobody noticed.
09:44:20	Varriount|Mobile	What is programming, but a series of choices to bring about a desired solution?
09:44:24	*	Araq bets less than 1% of all germans know about it
09:44:25	filwit	haha damn, i'm sure there's some "german nazi" joke in there somewhere, i just can't find it :P
09:47:14	filwit	so wait German doesn't use upper-case either?
09:47:27	Araq	it does
09:47:41	Araq	ß used to have no uppercased representation as it never starts a word
09:47:42	filwit	i didn't even think about asian-languages capitalization before you mentioned it
09:47:53	filwit	ahh, i see
09:48:12	filwit	but they made it so folks could YELL IN CAPS?
09:48:23	Araq	yes, exactly
09:48:35	Araq	ALLCAPS is even on the passports here, lol
09:49:16	filwit	i haven't seen a passport in awhile, couldn't tell you what ours look like
09:49:16	*	Matthias247 joined #nimrod
09:49:20	Araq	meaning if your name is Gruß it became GRUSS
09:49:36	Araq	and if your name is Gruss it became GRUSS too lol
09:49:54	Araq	on the passports
09:50:01	filwit	is Gruss a common German name?
09:50:08	Araq	nah, I made it up
09:50:08	filwit	never heard that one before
09:50:15	filwit	oh, i see
09:58:10	rixx	Araq: nah, I think more people know about ẞ
09:58:47	rixx	I think it was even mentioned in elementary school once
09:59:18	rixx	and I think I've seen it in advertising, too
09:59:25	filwit	i wounder if kids still get taught cursive in elementary school here
09:59:31	filwit	y'all have cursive in german?
09:59:44	rixx	yup
10:00:06	Araq	I'm not sure it's still taught
10:00:09	rixx	but they teach it later nowadays, I think
10:00:37	rixx	it is still taught, seen my siblings use it
10:00:48	rixx	and they are 10 and 8
10:01:02	filwit	that sounds about right
10:01:15	rixx	but they started pretty late
10:01:18	Araq	elementary school is mostly to teach children that one can cook his own lunches and that one can move backwards
10:01:27	rixx	:D
10:01:57	rixx	Araq: and how to deal with other children and being a little silent in class ;)
10:02:04	rixx	well. not too loud, maybe
10:02:07	filwit	i don't get the expression :|
10:02:32	filwit	do you cook your own lunches in german elementary school or something?
10:02:48	filwit	or is that just a phrase i'm unfamiliar with?
10:03:08	rixx	would probably be better than the stuff they serve there
10:03:17	filwit	i watched this german movie called "the wave" once which was pretty good
10:04:04	filwit	yeah our cafeteria food was gross too. public schools don't exactly feed the students brain-food
10:04:06	rixx	ah, well. not exactly representative of our educational system though
10:04:39	filwit	they do alright though i suppose, it's just they give you pizza & cheeseburgers as options... what kid is going to pass that up?
10:04:56	*	Varriount|Mobile misses German bakeries.
10:05:49	filwit	oregon has some really good food that i miss
10:06:12	filwit	the only place you can eat vegetarian and actually have a good meal
10:06:24	filwit	at least in my experience in the US
10:06:40	filwit	well no, California had souplantation which was good too
10:06:50	rixx	filwit: ah, no, they serve real food here, but it's … think mashed potatoes made from instant powder >.<
10:07:14	*	rixx stopped eating that stuff after fourth grade and just brought her own lunches
10:07:40	filwit	yeah I have a Hungarian neighbor, and he's made some really good Hungarian and German dishes for us before.
10:07:51	filwit	but they where both very rich
10:08:49	rixx	well, really good German dishes don't require you to be rich
10:08:57	rixx	Hungarian ones neither, actually
10:09:07	rixx	Hungarian ones even less
10:09:22	rixx	damn, now I wanna make Hungarian Goulash
10:09:50	filwit	well he's not the healthiest person alive, and i'm sure there are unhealthy people everywhere. Still, his dishes where tasty :)
10:10:01	filwit	yeah it was Hungarian Goulash that he made too
10:10:18	filwit	with just chopped potatoes and chopped pork or something
10:10:27	filwit	and some sort of sauce
10:11:29	filwit	good combo really. I like the "classic american" cookout food too (probably not from america): corn, bakedbeens, burger/hotdogs
10:12:09	filwit	but if i had to pick a perfect meal (since' i'm making myself hungry right now) i would probably choose steak and potatoes and broccoli
10:12:38	filwit	with A1 steaksauce and a beer (only time i actually like beer is with that combo)
10:13:22	filwit	probably should stop spamming #nimrod with food posts
10:16:59	Araq	nah it's fine
10:18:32	rixx	weeell, now I have to choose: Goulash for the whole hackerspace ooor studying for my exam
10:18:56	rixx	(exam is about formal languages/automata and compilers, so there's that)
10:20:19	filwit	i haven't seen you around before rixx, are you new?
10:20:33	rixx	filwit: yup, just getting started with nimrod
10:20:46	filwit	cool, and welcome
10:21:07	rixx	thanks :)
10:21:28	filwit	if you're researching languages then you're in the right place ;)
10:21:53	rixx	yep, a friend of mine pointed me this way when I was complaining about Python's lack of static typing (Python being my language of choice for all private projects, because it's so damn convenient)
10:22:26	rixx	well, he first tried to get me to use Ada, as always, but then he told me about Nimrod and I chose to give it a try
10:22:47	filwit	never really used Python besides hacking on Blender UI, but I'm kinda apposed to dynamically-typed languages in general
10:22:56	rixx	still gonna do my next project in python, though, because I wanna get through with it fast ^^
10:23:00	filwit	we where just talking about that the other day
10:24:10	skrylar	Ada is neat but the tooling is eeeh unless you pay
10:25:08	Araq	Ada is simply not the right tool unless you care about bounded memory usage by construction
10:25:22	rixx	Ada sounded pretty neat after everything my friend told me. I was basically left with saying "Yeah, but I don't wanna program tanks"
10:25:35	Araq	it sucks when it comes to resizable arrays and strings
10:25:48	Varriount|Mobile	rixx: We have a user in here who is working on some neat CPython interfacing lib
10:25:49	Araq	been there, done that, decided nimrod is necessary
10:26:56	Varriount|Mobile	Program tanks?
10:27:20	rixx	afaik that's one of the things Ada is actually used for
10:28:03	filwit	i've heard it was developed by some military, but that's about the extent of my experience with Ada
10:28:12	filwit	looked at it once and didn't like the syntaxd
10:28:26	Varriount|Mobile	rixx: https://github.com/micklat/NimBorg
10:29:18	rixx	nah, Ada is pretty nice, but I'm so not into Pascal-like begin/end syntax
10:29:30	Varriount|Mobile	If you want to use python from Nimrod . If you want to use nimrod from python, I'd suggest just using cffi and exporting C procedures.
10:30:02	rixx	I don't really plan on using the two together
10:30:50	filwit	i have always been more of a C-style person (just what i am used to), but Nimrod's made me appreciate indent-style
10:31:10	filwit	it can be easier to shift blocks of code around when you don't have curly-braces
10:31:59	filwit	btw, Araq, you mentioned removing comments from the AST, you still plan on that or did you change your mind?
10:32:28	filwit	i brought it up last night and someone said they like it being part of the AST, but i forget who and why
10:33:38	filwit	i just think it would make support Nimrod on simple text editors a bit easier, but other than that i could care less really
10:34:32	Araq	yeah that's still in the pipeline
10:34:43	Araq	I think only doc comments should be part of the AST
10:35:00	Araq	you're talking about alignment of comments though, editors don't care about ASTs
10:35:16	Varriount|Mobile	If you want regular comments in the ast, make them doc comments
10:36:05	filwit	Araq: right
10:36:40	Araq	filwit: so apparently we need more bikeshedding before the people can accept nimrod's syntax
10:37:03	filwit	like i've been saying man :P
10:37:27	filwit	actually some things about Nimrod i've really started to appreciate in terms of style
10:37:44	Araq	well the *real* reason is simply a lot of inherent disrespect for my work as I didn't invent unix
10:37:46	filwit	but i still think there's some inconsistencies that could help
10:38:06	rixx	oh, damn, Araq, you didn't? what am I doing here then?
10:38:06	Varriount|Mobile	"Nimrod: Just like that deadly space fungus, it grows on you"
10:38:15	Araq	but so be it
10:38:24	filwit	Araq: yeah, unfortunately people are pretty picky, it's true
10:38:27	rixx	Araq: I thought you were the secret child of K&R or something
10:39:16	*	EXetoC quit (Quit: WeeChat 0.4.3)
10:39:28	filwit	Araq: i'm sure no-one disrespects your work, it's just not everyone needs the performance of C, so that leaves them with other GC languages they might prefer the syntax of.. so syntax is important
10:40:24	filwit	still, there are a lot of people from Python/Ruby that seem attracted to Nimrod's syntax as-is
10:41:22	filwit	it's the C++/C#/Java crowd that seem to have very rigid ideas about what's right/wrong with syntax
10:42:34	filwit	that said, I kinda actually agree with many of C#'s designs.. like "str + str" for concatenation
10:42:38	Matthias247	filwit: actually I though the same when I first looked at Nimrod. But it's simply the unfamilarity
10:44:04	filwit	Matthias247: yes but i hear that a lot. and how many people look at it and simply turn away? Some we will never get because they're too set in there ways, but other things could probably help people accept an unfamiliar thing if it has good reason too it
10:44:41	filwit	it's true it's rather arbitrary when it comes to syntax, but some things are also a bit universal
10:44:50	filwit	math, strings, lists, ect
10:46:54	*	r0b1 joined #nimrod
10:46:58	*	r0b1 quit (Changing host)
10:46:58	*	r0b1 joined #nimrod
10:46:59	filwit	those things have a lot of cross-over, and if people can just look at something and think 'that make sense' (like "hello " + p.name + ".") it could help grow the language. at least, I think it would.
10:47:30	filwit	ultimately though, these things could be also changed after Nimrod 1.0, and it might be a better plan to do so
10:47:36	filwit	(if they get changed at all)
10:47:39	dom96	gutten morgen
10:47:59	filwit	morning dom
10:48:38	Varriount|Mobile	I think the dynamic language crowd is attracted to nimrod because nimrod gives many of the benefits that a dynamic language normally has.
10:48:49	dom96	My worry is that we will become too much like C if we fold to people's demands too much.
10:49:33	filwit	nah, the language already is leaps and bounds better than C because there's no ending marks and no headers
10:49:54	Varriount|Mobile	Let's add braces. And mandatory semicolon delimiters >:)
10:50:06	filwit	plus the ref/ptr/addr syntax is way better (because they're descriptive)
10:50:10	Araq	"+" for concatenation is wrong
10:50:17	*	io2 joined #nimrod
10:50:21	filwit	Araq: i disagree
10:50:29	filwit	Araq: cause of "commonality"
10:50:32	Araq	@[1,2] + @[2,3] # what does it do?
10:50:48	filwit	that's [1,2,1,2]
10:50:56	Araq	in nimrod it's not ambiguous
10:51:12	Araq	and it's @[3,5]of course ;-)
10:51:15	filwit	if you did this: @[1,2][] + @[2,3][] it shoudl be addition
10:51:21	rixx	wat
10:51:34	rixx	Araq's version is much more intuitive
10:51:45	filwit	Araq: i know what it is in Nimrod, i just meant what i think it *should* be
10:52:08	filwit	and i mispelled, i meant [1,2,2,3]
10:52:32	Araq	well lifting via `[]` is awful though
10:52:49	filwit	rixx: i dissagree. i think if you took a poll most people would understand: mylist + otherlist as concatenation not addition
10:53:10	filwit	the reason is because often list's are not just numbers
10:53:36	rixx	filwit: maybe I'm thinking too mathematically
10:53:45	filwit	so it's a wider application to use the most common symbol of addition '+' for the most common _application_ of lists (joining)
10:54:02	rixx	doesn't make it right, though
10:54:05	filwit	mylist + other list (join); mylist += node (add)
10:54:16	Matthias247	for lists I find + to e ambigous and would prefer named operators (join, add)
10:54:23	rixx	but I like Araq's versin better, and I agree that appending [] to change the meaning is pretty horrible
10:54:32	filwit	rixx: that's why i was saying do this for add: mylist[] + otherlist[] (addition)
10:54:34	Matthias247	however for strings I find + for concetanation naturally
10:54:58	Araq	could easily be done, Matthias247
10:55:00	filwit	rixx, Araq: then you can also do things like this: mylist[0..5] + otherlist[5..10]
10:55:15	rixx	I don't. I got used to + concating strings, but I don't think it's nice
10:56:14	Araq	Lua uses '..' fwiw, Perl uses '.'
10:56:22	Matthias247	And Qt uses %
10:56:33	Matthias247	which is totally strange
10:57:26	filwit	Araq: neither of those 'imply' addition, which is why they're worse than '+'
10:57:42	Araq	what? o.O
10:58:15	Araq	concatenation has nothing to do with addition except for the fact that afterwards result.len == a.len + b.len
10:58:33	filwit	./../% don't imply addition, while '+' is taught to us to mean 'add' and when you apply 'add' to lists you normally mean 'add together'
10:59:47	filwit	that's pretty much as simplified as i can make my argument
11:00:41	filwit	so if you can find fault in that, sure i'll take it. Ultimately you'd have to take a study about what "most people" associate when you "imply" addition of strings (math or join), but I'm assuming that later
11:01:20	*	[1]Endy joined #nimrod
11:02:42	Araq	for english speaking people, perhaps you're right
11:03:20	filwit	that's what i mean, "most common" unless something else becomes the world trading language
11:03:40	Araq	in German it never occurred to me to conflate "hinzufügen" and "addieren"
11:04:24	*	Matthias247 quit (Read error: Connection reset by peer)
11:05:01	*	Endy quit (Ping timeout: 264 seconds)
11:05:02	*	[1]Endy is now known as Endy
11:05:20	filwit	Araq: interesting you guys have two words for 'add'? i mean i guess english does as well, but the common 'math' term is always 'addition' which is part of the crux of my argument
11:06:33	filwit	Araq: based off google-translate, it looks like the first word is more like 'append' and the second is more like 'addition'
11:06:34	dom96	wow. Mozilla spend 115K per month for the CI
11:07:05	filwit	CI?
11:07:33	dom96	continuous integration
11:07:37	dom96	http://taras.glek.net/blog/2014/02/06/cost-efficient-continious-integration/
11:09:12	Araq	filwit: yeah
11:09:18	*	ponce__ joined #nimrod
11:09:26	*	brson quit (Quit: leaving)
11:09:30	Araq	hi ponce__ wb
11:09:34	ponce__	hi!
11:09:47	ponce__	I hate to make language suggestions but
11:10:24	Araq	no worries, go on
11:10:28	ponce__	I thought about just "public" and "private" sections as a * replacement
11:11:01	Araq	brb
11:14:20	filwit	pounce__ i suggested something similar on the forums already: http://forum.nimrod-lang.org/t/381/3
11:15:33	filwit	pounce__ ultimately changes like that will probably take a lot of discussion before it would be changed (if it changes in the near future at all) because it would probably require a lot of work and there are somewhat bigger issues to address (like IDE tooling and things)
11:16:36	ponce__	I only say that because I see people complaining, I suspect that the current syntax is more readable
11:17:09	ponce__	especially if couloured differently in IDE etc
11:17:46	*	EXetoC joined #nimrod
11:18:34	filwit	ponce__ i agree with you though, i think there could be a better solution to the '*' too, but ultimately that's probably the least needed syntax change in Nimrod ATM
11:20:25	filwit	ponce__ though it is brought up a lot, probably because it's confused with C's semantics
11:20:30	dom96	filwit: I would still love an Aporia logo if possible.
11:21:23	filwit	dom96: okay i'll get around to making you one. Just last few times I tried i could tell you didn't really like the designs.
11:21:45	ponce__	compiling HEAD babel with HEAD nimrod compiler makes: "babelpkg\packageinfo.nim(4, 14) Error: cannot open 'version'"
11:22:04	ponce__	what did I do wrong?
11:22:23	dom96	filwit: Yeah, sorry. :\
11:22:28	filwit	dom96: i've been meaning to sit down and do a bit of design touch-up/artwork/mascot changes to Nimrod for awhile and layout a visual design for the docs
11:22:53	dom96	Better design for docs is more important.
11:22:56	dom96	Do that first if you can
11:22:59	filwit	dom96: no problem man, i'm not asking you to like all my design :P just that means it takes a little longer
11:23:22	filwit	ponce__ use devel
11:23:49	ponce__	ok
11:25:15	filwit	dom96: i will and maybe we can work together on the HTML or something (where I do the front end and you modify the docgen?) I'm hoping we can add in some JS to make seaching for procs by type-association very convenient
11:25:39	dom96	filwit: sure
11:25:59	filwit	dom96: that would probably only require a hidden tag to be exported in the docgen along with every proc so i could search through them in JS
11:26:32	*	Varriount|Mobile quit (Remote host closed the connection)
11:26:46	*	Varriount|Mobile joined #nimrod
11:27:10	dom96	filwit: We need to be careful not to bloat the html too much though
11:28:32	filwit	dom96: it could be in the anchor tag, so <a name="jmpname" data-param="int float int")>
11:28:41	filwit	data-params*
11:29:03	dom96	That's abusing html a bit heh
11:29:13	filwit	how so?
11:29:21	*	eigenlicht joined #nimrod
11:29:24	filwit	data-* attributes are part of HTML5
11:29:24	dom96	Specifying a non-standard argument
11:29:29	dom96	oh
11:29:34	dom96	Didn't know that lol
11:29:35	filwit	no, it's a standard now
11:29:37	filwit	yeah
11:29:55	dom96	Thought you just made up an attribute and added it in there.
11:30:02	filwit	lol, no :P
11:30:10	filwit	it will validate
11:30:32	filwit	as long as you use <!doctype html>
11:30:44	dom96	good
11:31:10	filwit	yeah, it makes doing this sort of thing much easier
11:31:29	filwit	was really happy when they i found out about that addition to HTML5
11:32:34	filwit	of course, before html5 there was xhtml which basically allowed the same type of things if i recall
11:32:53	dom96	ahh, xhtml. Good times.
11:33:06	filwit	yep, haha
11:33:37	filwit	i still kinda prefer xhtml to html5 since it was consistent with it's </> instead of having some takes not require the /
11:34:03	filwit	but it's been discontinued, so what are you gonna do
11:34:08	dom96	I like html5 simply because I can finally remember the doctype.
11:34:17	filwit	lol, good point
11:51:42	*	ddl_smurf joined #nimrod
11:54:16	Araq	ponce__: tbh I understand how the "trailing" '=' for procs looks weird at first, how pragmas look strange, how type sections can feel 'off' but that the export marker continues to be brought up suprises me
11:54:40	skrylar	i sometimes forget to type it
11:54:54	skrylar	filwit: xhtml is still valid
11:55:18	skrylar	just because some /'s are optional in html5 doesn't mean you can't still just export perfect XML all the same
11:55:18	filwit	skrylar: thought 2.0 was officially dropped
11:55:53	skrylar	filwit: as a standard it might be abandoned (which is kinda dumb) but since it was largely just "proper html", you can still do that for all intents and purposes
11:56:12	Araq	it works nicely and has been designed after people noticed the interface/implemetation split of pascal doesn't work well. And it's not my invention, I took it from Oberon.
11:56:35	skrylar	(nb: i don't have a problem with the export marker, i just forget to type it a lot)
11:57:15	skrylar	i should write a "reasons why nimrod seems derpy at first glance (but it really isn't)"
11:57:34	filwit	skrylar: yeah i know it still is accepted in browsers. i only meant it's not being promoted and supported to the level HTML5 now is. I'm sure there's some odd tablet browser that doesn't render correctly with it by now or something
11:58:14	skrylar	i always had trouble with browsers accepting xhtml when passed as xhtml :\
11:58:21	Araq	python uglifies private names with a leading '_'. how is that better? and more importantly: why isn't that constantly brought up against python? I find the other __builtin__ stuff equally ugly
11:58:54	skrylar	because python is used by NASA and dissing python = dissing NASA which is uncool
11:58:55	skrylar	:P
11:59:06	Varriount|Mobile	Araq: Private names are rarely used in small projects
11:59:17	skrylar	(then again LISP is also used by NASA but nobody cares about that lol)
11:59:33	filwit	Araq: because it's a common style practice in other langauges like C whereas '*' mostly alien (although i agree with what you're saying)
11:59:36	Varriount|Mobile	Most libraries use the __all__ variable to selectively export things
11:59:49	filwit	Araq: because _** is a common....
12:00:09	skrylar	amusingly people sometimes use _foo to mean private in C/C++ but this is in error
12:00:18	skrylar	the spec says that anything starting with underscore is reserved for the compiler
12:00:41	filwit	i thought it was two underscores
12:00:43	skrylar	but people wholesale ignore that and so compilers ended up having to use double underscore instead
12:01:03	filwit	at least it's not Dart where '_' is protected and '__' is private
12:01:18	EXetoC	:E
12:01:41	skrylar	because 'private blah blah' is too stupid
12:02:20	skrylar	i think D allowed both C++ and Java styles for visibility, which IMO are more directly explicit than underscore soup
12:02:50	filwit	D reserves __ for the compiler, technically
12:03:01	filwit	__traits(..)
12:03:10	filwit	__XMM(..) // simd
12:03:21	ponce__	Araq: I guess it's frequently discussed because a line like "ip*: string" create confusion with the C pointer declaration
12:03:31	skrylar	according to IBM C its two underscores or one underscore and a capital letter+end of identifier
12:03:41	skrylar	so _fooBar is allowed but _FOOBY isn't
12:04:02	filwit	...
12:04:18	Araq	skrylar: that is correct
12:04:30	Varriount|Mobile	ponce__: But after you found out that "*" was a visibility modifier, were you still confused?
12:04:47	ponce__	I think the * for public is a good idea
12:05:08	ponce__	that probably takes a bit of time to sink in
12:05:21	Araq	I have nothing against ! for public, it's like 1 line in the compiler to support that too
12:05:22	ponce__	once used to the C family
12:05:49	skrylar	i'm not sure making concessions to C syntax is a great idea
12:05:53	Araq	var s!: string # hmm
12:05:58	skrylar	there's a certain point that if you're different enough you might as well do what you want
12:06:18	filwit	Araq: use ! for public and ? for package?
12:06:27	*	askatasuna quit (Ping timeout: 244 seconds)
12:06:28	filwit	Araq: i almost like the sound of that..
12:06:28	skrylar	kind of like how asian grammar is completely different to european grammars so its a full "well, i guess i just need to learn" trigger instead
12:06:40	filwit	Araq: then again, * and + make just as much sense really
12:06:43	Araq	filwit: meh, don't start with "package"
12:06:47	skrylar	filwit: protected isn't needed?
12:06:53	skrylar	private members are already visible to that module
12:07:03	skrylar	(admittedly people complained about this in D1)
12:07:10	filwit	Araq: :P just planning ahead for when it's so requested you have no choice but to add it.
12:07:21	ponce__	"package" visibility is seldom used in D
12:07:31	Araq	I don't even think it can work, filwit :P
12:07:56	filwit	why? it just exposes a symbol to only the files within the same folder?
12:08:21	filwit	i could see why that could require more work that it's worth right now, but impossible?
12:08:27	ponce__	http://pastebin.com/A0ihp2zQ < talking about visibility, I can't help but notice that "publicness" get repeated many times
12:08:51	ponce__	so I guess people talk about it because until our eyes are accustomed, it jumps out
12:09:04	filwit	Araq: then again, there might be a better solution with submodule & includes, so idk
12:09:23	Araq	well it requires a notion of "package friends", filwit
12:09:31	dom96	I don't get this package visibility you guys are talking about.
12:09:41	Araq	currently there is an implicit notion of a "friend" module
12:09:43	dom96	Wouldn't the compiler need to be aware of babel even more then?
12:10:00	dom96	Also, if you change the export marker to ! I will actually leave and never come back :P
12:10:14	filwit	Araq: all files in the same folder would be friends like in D
12:10:19	skrylar	dom96: also make sure that public members are in allcaps
12:10:31	skrylar	proc NEWTHING!(...)
12:10:33	ponce__	like the template notation in D (!), it's too much expressive ;)
12:10:42	skrylar	just to complete the notion of shouting at users
12:10:54	*	psquid quit (Quit: work)
12:11:06	dom96	skrylar: perfect
12:11:14	Araq	proc foo+(a, b: int) # what about + then?
12:11:35	skrylar	proc `+`+(a, b: int) # not confusing at all
12:11:39	filwit	Araq: why change the '*' ?
12:12:00	filwit	skrylar: no different than proc `*`*(a, b:int)
12:12:01	dom96	skrylar: well to be fair: proc `*`*(a, b: int) exists.
12:12:08	filwit	jinks
12:12:12	ponce__	that would make http://pastebin.com/bPd3z9RT
12:12:23	EXetoC	dom96: no, package as in a directory in said project
12:12:37	dom96	EXetoC: So then we have two different definitions of "package"?
12:12:42	skrylar	dom96: i will now hide under a mushroom
12:12:44	dom96	That's completely sane.
12:13:05	EXetoC	the programming community?
12:13:08	filwit	wait, there's a current definition of package?
12:13:23	Varriount|Mobile	I thought package private/public was only needed for silly languages that use the concept of packages
12:13:30	dom96	filwit: A babel package.
12:13:32	filwit	i thought it was basically just a path thing at the moment, anthing before the '.' in an import was a folder
12:13:38	Varriount|Mobile	Eg: java
12:13:42	filwit	filwit: babel is part of the compiler?
12:13:51	filwit	dom96 ^
12:13:53	Araq	filwit: because it looks like a pointer declaration and + only looks like UML
12:13:54	dom96	no.
12:14:01	dom96	well, it is a bit.
12:14:46	ponce__	how about public: blocks http://pastebin.com/RPhkdjCU
12:14:50	dom96	Araq: Seriously. If you want to appeal to everyone then you might as well just make a new D and call it "a better C".
12:15:24	filwit	don't listen to him Araq, CHANGE ALL THE THINGS!!!!
12:15:45	ponce__	dom96: they would still differ in many ways
12:15:54	Araq	now who is the angel and who the demon ...
12:16:06	Varriount|Mobile	ponce__: Can be done with macroz
12:16:10	dom96	ponce__: That will result in too much indentation.
12:16:39	dom96	I was under the impression that human beings are intelligent.
12:16:43	dom96	And capable of learning new conventions.
12:16:53	Araq	dom96: you're wrong though
12:16:54	EXetoC	only some are
12:16:54	filwit	Araq: seriously though, i don't think changing * to + is a valuable change really
12:17:01	dom96	Is learning that * means "export" so hard?
12:17:04	filwit	Araq: i agree with dom96 on this one
12:17:13	Araq	filwit: yeah I'm not serious
12:17:21	ponce__	dom96: no, but is "export" so valuable that it has to be on every line?
12:17:41	dom96	Araq: Gah. Stop joking around.
12:17:48	filwit	Araq: i can never tell
12:18:01	dom96	Araq: That's how you get quoted infavourably.
12:18:24	Araq	at least I get quoted. that's already better than getting ignored.
12:18:26	filwit	lol.. he's trolling his own IRC
12:18:52	skrylar	Araq: speaking of demons, when is the compiler going to support futhark
12:18:54	dom96	Araq: Be careful, you may regret it later :P
12:18:55	filwit	because everyone always talks about petty changes, he suggests the pettiest of them just to watch the reactions..
12:19:08	skrylar	[actually it probably does considering those are unicode alphas]
12:19:31	EXetoC	no but srsly, we need to extend the concept further, but allowing more than just optional begin/end and {} etc, if we do indeed get that
12:19:32	dom96	ponce__: That's true. For objects it is likely that you want to export every field. For that case it may make sense to add something like that, but I think a pragma may be nicer.
12:19:54	skrylar	isn't that something a macro could already do (or should be able to) :3
12:19:58	EXetoC	everybody wins. "def p(x): ..."
12:20:07	ponce__	if public: private: blocks are possible with macros, might as well keep the current syntax and eventually wrap it when needed
12:20:19	skrylar	well i meant a "make all fields public" macro
12:20:27	dom96	EXetoC: Funny. I was actually making a macro which allowed you to do that a while back :P
12:20:36	skrylar	all though yeah a certain template syntax would probably work too
12:21:02	dom96	ponce__: I think it would make sense to finally settle on an OOP macro, and let the OOP macro provide that.
12:21:20	Araq	oh yeah. dom96 is right
12:21:21	EXetoC	"go do_stuff()" -> ???
12:21:36	filwit	Araq: just realized i can make my macros use '+' instead of {.get.} to generate getters...
12:21:56	dom96	Araq: Indeed. I think we should add that for 0.9.4 together with a closure macro.
12:21:58	Araq	filwit: so '+' for readonly?
12:22:06	filwit	yeah basically
12:23:15	dom96	Oh and in regards to allowing pragmas to be used using the {pragma} syntax, I am fine with that.
12:23:31	EXetoC	dom96: well, if you could just push/pop inside a type
12:23:38	dom96	Aporia highlights pragmas anyway, so it should look pretty much the same to me anyway.
12:24:49	filwit	dom96: that means pragmas and sets will be highlighted the same unless you have more advanced semantic checking going on
12:24:52	EXetoC	and then you could group anything by visibility
12:25:03	dom96	filwit: oh damn. I forgot about sets.
12:25:18	dom96	In that case I change my opinion.
12:25:26	dom96	Pragmas should stay as they are.
12:25:51	EXetoC	I think .pragma. was proposed
12:26:34	filwit	yeah i really have no problem with the existing */{..} syntax.. it's only the T/P thing and a select few type names (seq) and there operators that i have any problem with at all
12:26:53	*	nequitans_ quit (Ping timeout: 252 seconds)
12:27:08	Araq	filwit: aliasing & to + for strings is *very* easy to do
12:27:23	filwit	oh, and the inconsistency with type/proc's use of :/= too.. but that needs more discussion
12:28:09	Araq	actually maybe we should do this instead, filwit :
12:28:10	filwit	Araq: i know, but people will always do what the docs say, and it's not like i don't know what '&' does.. i'm just saying for adoption purposes only. it makes things simpler, IMO
12:28:23	Araq	proc foo(a, b: int) is
12:28:26	Araq	echo a, b
12:28:38	*	askatasuna joined #nimrod
12:28:48	filwit	proc foo(..): int is ... ?
12:28:52	Araq	yes
12:29:06	filwit	not bad. but what about types then?
12:29:20	filwit	type Foo is ...; type Bar: Foo is ... ?
12:29:24	Araq	type foo = int # stays?
12:29:38	Araq	it's much more of an equals in this place anyway
12:29:47	Araq	const foo = 3+4 # = makes sense
12:29:55	Araq	var foo = 4 # = makes sense
12:30:20	Araq	proc foo: int = 4 + 5 # = also makes sense :P
12:30:26	filwit	yes but those are fundamentally different (to the coder) than defining types/procs (even though const is somewhat both)
12:30:55	filwit	Araq: i don't dislike the '=' for procs at all.. i just dislike how type/proc makes different use of them
12:31:10	filwit	Araq: in fact i would probably prefer = to 'is'
12:31:35	Araq	filwit: well what are you complaining about then?
12:31:45	filwit	Araq: but I still think types should be this: type Foo: int; type Bar: Foo = ...
12:32:01	filwit	because it matches what procs are like
12:32:13	filwit	and that's the closest idiom, not 'const'
12:32:20	Araq	'type' introduces a section, 'proc' does not
12:32:32	Araq	'type' is much more like 'const' already
12:32:39	filwit	only cause you made it that way... proc could introduce a section as well
12:32:49	filwit	in fact i could probably add that in now that i know more about the AST
12:32:53	Araq	':' means "has type" '=' means "aliases type"
12:33:21	filwit	: doesn't mean "has type" for procs
12:33:30	filwit	it means "returns type"
12:33:56	Varriount|Mobile	Which is the single exception.
12:34:13	Araq	filwit: yes, but you're not arguing for '->' are you?
12:34:21	Araq	proc foo -> int
12:34:37	filwit	i did actually, i just called it 'of'.. see my forum post
12:35:00	Araq	'of' is commonly spelled 'as' in Basic iirc
12:35:16	filwit	i would be fine with either if that's a concern
12:36:33	filwit	Araq: here's what I don't like about the differences between the '=' with type/proc: type T = U <body>; proc F:U = <body>
12:37:07	filwit	Araq: the stmt body follows the = for one but not the other
12:37:26	filwit	Araq: moreover, it's also different for if/else/etc statements
12:38:03	filwit	Araq: if we're talking about 'syntax cleanup' i would argue that any change which made things more consistent on all fronts would be the best direction
12:38:34	filwit	Araq: may not be possible, but then no change at that level would be an easy one, so probably shouldn't be jumped to anyways
12:39:00	Araq	again, it's not U <body>, it's <body>
12:39:22	Araq	object of foo\n a, b: int # is the <body>
12:39:36	filwit	but i sure looks like 'U' body the way it's written, especially since you cant do object[...]
12:40:21	Araq	that's a different concern then
12:40:44	filwit	you can't do: type Foo =
12:40:48	filwit	object of Foo
12:40:51	filwit	a: int
12:40:52	filwit	either
12:41:17	filwit	wait.. can you?
12:41:19	filwit	one sec
12:41:33	Araq	lol I dunno either
12:42:14	Araq	but either way, 'type' is not 'class' and so 'type' doesn't resemble 'proc'. I guess that's what it comes down to, you want them to be as similar as possible
12:42:27	dom96	For my closure macro I was considering also implementing some proc type sugar:
12:42:35	dom96	(int, string) -> int
12:42:54	dom96	But then that's a problem if you don't want a return type
12:43:04	dom96	You could just write void but that's not really in the spirit of Nimrod.
12:43:05	EXetoC	-> void?
12:43:08	filwit	Araq: didn't know you could put 'object' on the same line as the recList
12:43:09	EXetoC	right
12:43:15	dom96	Araq: What do you think?
12:46:53	filwit	Araq: it's odd how 'object of T' can be on the same line as the recList that follows it, or it's recList can be indented..
12:47:08	Araq	dom96: -> is to appeal to the math/FP people and so '-> void' is fine here, I think
12:47:57	filwit	Araq: personally I would love to see ':' replace '=' everwhere so that if/else/custom-macro was consistent with type/proc.. if it really came to it i would probably even prefer "const foo: 10"
12:48:13	dom96	Araq: Alright. I was also considering: ~(int, string) or something.
12:48:38	filwit	Araq: sorry, i meant "i think** i would love..."
12:49:28	*	Varriount|Mobile would rather not have any hasty language changing decisions made.
12:49:48	dom96	Let's just s/:/=/ :D
12:50:19	Varriount|Mobile	filwit: You scare me. Is this revenge for me dropping that medical link earlier?
12:50:44	filwit	Varriount|Mobile: I don't think anything being talked about is actually being seriously talked about as a change that will happen anytime soon
12:52:03	filwit	Varriount|Mobile: lol, no, but that's alot for that :P
12:52:15	filwit	dom96: btw, did you see the PM i sent you earlier?
12:53:10	dom96	filwit: nope?
12:54:07	filwit	dom96: well just /join me an i'll ask again
12:54:26	dom96	What? That's not quite how IRC works lol
12:54:33	dom96	Just /msg me
12:54:40	filwit	oh, that's probably why it didn't work...
12:54:42	filwit	:P
12:55:32	filwit	g'damn.. can you just open a chat.. im a newb at IRC and this webIRC isn't helping..
12:56:27	filwit	i just wanted to ask you about the super secrete google-analytic anyways..
12:58:25	filwit	dom96: i tried to /msg you three times but it didn't work..
12:59:06	EXetoC	"/query dom96" or "/msg dom96 ohai"
12:59:20	dom96	filwit: I PM'ed you
12:59:30	filwit	k
13:25:08	filwit	g2g, bbl
13:25:14	*	filwit quit (Quit: Page closed)
13:49:22	*	nequitans_ joined #nimrod
13:57:38	nequitans_	Hello!
13:57:45	dom96	hey
13:58:30	nequitans_	Thanks for your guys' assistance yesterday -- Araq's suggestion about by slowdown being related to random() was right
14:01:11	nequitans_	Once I fixed that issue (quickly wrote my own generator in nimrod), I decided that I wanted to do some metaprogramming to make doing this again simpler
14:01:17	Varriount|Mobile	Hah, PHPFail -> http://codegolf.stackexchange.com/questions/21114/weirdest-way-to-produce-a-stack-overflow/21160#21160
14:01:19	nequitans_	ala: https://gist.github.com/geetduggal/9290100
14:02:29	nequitans_	However, I get this error at compile time: Error: internal error: param not set: :env
14:03:34	nequitans_	ideally, it would be generecized too
14:03:38	Varriount|Mobile	nequitans: Sounds like something wrong with the compiler
14:04:04	Varriount|Mobile	The env parameter is the hidden param passed to iterators
14:04:30	*	darkf quit (Quit: Leaving)
14:05:18	nequitans_	VarriountIMobile, I see -- does the template itself look legit ('cause it's probably likely that I'm doing something wrong there :))
14:06:22	Varriount|Mobile	nequitans: What line/col is the error reporting?
14:07:19	nequitans_	line 6 right after the proc
14:07:38	Varriount|Mobile	Oh! Its the closure proc too, I think
14:08:06	Varriount|Mobile	env parameters are also passed with them, to capture outer variables
14:09:01	Araq	nequitans_: I think this bug has already beeen reported
14:11:12	nequitans_	Araq, noted. Is there a hack to get around it in the interim?
14:12:10	Araq	try without the 'block' which is kind of deprecated anyway
14:14:10	Varriount|Mobile	How is 'block' depracated?
14:14:21	nequitans_	Araq, it worked!
14:14:57	*	brihat left #nimrod (#nimrod)
14:15:18	Araq	Varriount|Mobile: 'block' in templates is a hacky way to introduce a fresh scope. With hygienic templates that is not necessary anymore.
14:17:11	Varriount|Mobile	nequitans: What are you working on? Some sort of simple threading thing?
14:20:23	nequitans_	VarriountIMobile, well after I became interested in nimrod and super excited after seeing its metaprogramming capabilities, I wanted to use it for my next project (I'm towards the end of my phd). One aspect requires creating a lot of random matrices (up to tens of thousands of rows and columns) and then doing an optimal selection amongst them. Threads makes it faster :)
14:21:20	Varriount|Mobile	Ooh, nifty
14:23:28	nequitans_	I really like how it is so easy to write things like 'withFile', 'debug', 'parallelFor' in a syntax similar to nimrod itself without modifying the language (e.g. i assume you have to do this in python, for instance). i haven't looked into rust and julia's metaprogramming capabilities, but i wonder if this is an advantage of nimrod?
14:24:56	profmakx	is there a way to get hold of .h #defined constants in nimrod code?
14:29:04	Varriount|Mobile	profmakx: Use c2nim to extract them from the header, or use some varient of the cimport pragma
14:29:33	profmakx	{.importc, noDecl.}: cint
14:29:35	profmakx	like that?
14:29:55	dom96	also {.header: ...}
14:30:00	profmakx	AF_INET6 = 23 ## for network protocol IPv6. this will almost certainly not work on anything but the system that it was taken from (I'll assume linux)
14:44:17	Araq	profmakx: you can use var AF_INET6 {.importc, header: "<foo.h>".}: cint to get "portable constants"
14:44:33	Araq	but this has its disadvantages of course
14:44:45	profmakx	does that work in an enum?
14:44:51	profmakx	maybe i should just try
14:46:22	Araq	no
14:47:00	nequitans_	Varriount|Mobile, only one part of the project is in impure nimrod now (calls previously written C++ code via a compiled library) ended up writing some basic graph algs on my own). Would be cool to make it all pure Nimrod just to show how easy it is to do without sacrificing speed
14:47:51	Varriount|Mobile	nequitans: Even if part of it is in nimrod, its still neat!
14:50:01	Varriount|Mobile	Hm. What part of nimrod uses "::"?
14:53:22	*	dom96 wonders if GTK has a resize end signal
15:05:11	Araq	Varriount|Mobile: nothing but it's a valid operator symbol
15:23:22	EXetoC	Araq: so you think that the rule that states that you shouldn't nest more than about 3-4 levels doesn't really help when it comes to readability?
15:24:25	Araq	the rule *at least* needs 1 exception
15:26:49	EXetoC	you even said that you liked deep nestings, when we asked you about the two-space indent convention, so that's why I'm asking
15:34:46	Araq	iterating over an N dimensional space needs N loops
15:35:10	Araq	it's foolish to split that up into multiple procs to avoid the deep nesting
15:42:41	*	BitPuffin joined #nimrod
15:43:04	profmakx	having a nice iterator there would sometimes help more fwiw
15:46:54	*	nequitans_ quit (Remote host closed the connection)
15:49:01	EXetoC	surely the complexity varies from case to case
15:50:42	*	BitPuffin quit (Quit: WeeChat 0.4.3)
15:51:14	renesac_	@[1,2] + @[2,3] indeed looks like @[3,5] for me
15:51:51	EXetoC	what else if it does indeed work? & is for concatenation
15:52:08	renesac_	Araq, what about the possibility of specifying the operator precedence via a pragma or something?
15:53:40	Araq	renesac_: we have "strong spaces" planned to affect operator precedence which is IMO a superior solution to this problem
15:53:55	Araq	profmakx: yes but then you have that nesting in the iterator
15:54:32	renesac_	well, as I said, 'shl' and friends don't mix well with 'strong spaces'
15:56:02	renesac_	and that 'first character' rule probably leads to ankward operators... and I don't really think it simplifies anything for the user
15:56:27	renesac_	as there are many operators to remember...
15:56:39	renesac_	*characters
15:59:25	Araq	well fair enough with strong spaces we can have <<
15:59:34	EXetoC	Araq: and that's essentially why we got this two-space convention?
16:00:02	Araq	EXetoC: no, I'm used to 2 spaces, that's why. No deep thoughts behind it really.
16:00:12	Araq	feel free to use 4 spaces
16:00:23	Araq	the compiler doesn't mind, nor do I
16:02:56	Araq	renesac_: it's not only about "simplifying" things for users
16:03:19	Araq	which operator overloading and the macro system symbol lookup can happen lately
16:03:31	Araq	and so precedence must not be bounds to symbols
16:04:53	renesac_	macros can't lookup the precedence information someway?
16:11:21	Araq	well surely it could be done
16:11:30	Araq	but what's the point really?
16:11:59	Araq	x<<1 & 5 should be parsed as I wrote it, not as you defined &'s precedence
16:13:36	renesac_	well, I'm not sure I'm really into this strong spaces bandwagon, as I like to have spaces between operators and values
16:14:24	renesac_	and if you don't use this feature, the << operator will have lousy precedence...
16:16:16	renesac_	and having both 'shl' and '<<' is bad
16:17:39	renesac_	though we likely would still need some non-special signal operators...
16:19:58	renesac_	and or not xor shl shr div mod in notin is isnot of -> & (confusion with concatenation), | , ! , ^ (no conflict?), <<, >>, // (python-like), % (confusion with format), the others may be left as is....
16:20:32	Araq	can you wait until we implemented it and you tried it in practice before you complain?
16:20:42	renesac_	well, and I haven't looked up the precedences
16:21:05	renesac_	ok...
16:21:29	renesac_	but then we would have two versions for those operators?
16:21:58	Araq	I don't think so, but you can always alias yourself
16:23:55	Araq	however =<< and =>> are cool and 'shr' doesn't easily enable it
16:24:18	Araq	so perhaps we should copy C more here
16:25:00	Araq	but then 'shr' and 'shl' should be deprecated and ultimately removed though most people are not opposed to having aliases to make programmers feel right at home
16:25:20	Araq	so I dunno and it's not important anyway
16:25:39	Araq	see you later
16:38:40	*	nequitans_ joined #nimrod
17:03:34	*	Matthias247 joined #nimrod
17:21:08	*	Mat3 joined #nimrod
17:21:12	Mat3	hi all
17:22:51	*	brihat joined #nimrod
17:22:51	*	brihat1 joined #nimrod
17:23:15	*	brihat left #nimrod (#nimrod)
17:23:19	*	brihat1 quit (Client Quit)
17:23:47	*	brihat joined #nimrod
17:39:08	nequitans_	i have a simple mersenne twister random number generator in pure nimrod.. dunno if it's in a package already (first glance says, no, but i'm not sure)
17:41:58	dom96	Not that I know of.
17:44:28	renesac_	I was going to implement some xorshift random number generators
17:45:43	renesac_	well, I implemented a simple one, but was going to make it more generic and this xorshift1024* looked interesting also:
17:45:49	nequitans_	dom96, noted. renesac_ -- i implemented xorshift last night to test out whether random nums were slowing me down. decided to use mersenne today since it is basically a standard, proven generator. i could provide both since already implemented. do you know how to seed a xorshift generator?
17:45:59	renesac_	http://xorshift.di.unimi.it/
17:46:28	nequitans_	renesac_, interesting
17:47:37	renesac_	he don't show the seeding procedure, but I don't think you need the warm up that mersenne needs
17:49:36	nequitans_	ic
17:50:02	renesac_	I'm far from expert in those issues though
17:50:20	renesac_	and none of those are cryptographic secure anyway
17:50:36	renesac_	so speed seems the most important thing
17:51:31	nequitans_	yea, i think the idea behind them is application to simulations and statistics more than crypto stuff
17:52:34	renesac_	well, it is true that peer review can get picky about the quality of the RNG you used for simulations...
17:57:59	renesac_	anyway, adding that multiplication in the end of xorshift is like making a lot more shifts, and is cheap, at least on current x86
17:58:37	renesac_	that is why murmurhash is also so good in that table
17:58:37	nequitans_	yea, it's pretty interesting that the xorshift under certain measures of quality performs so well
17:59:08	renesac_	one thing that bothers me though is the lack of 0 as possible output
17:59:09	renesac_	:P
17:59:43	*	nequitans_ quit (Read error: Connection reset by peer)
18:00:59	*	nequitans_ joined #nimrod
18:01:06	nequitans_	ack, computer froze!
18:01:41	renesac_	I think that a murmurhash-like algorithm with a bigger state would be a killer rng... but I don't have the number theory knowledge to build it
18:01:46	renesac_	[14:59:08] <renesac_> one thing that bothers me though is the lack of 0 as possible output
18:02:26	nequitans_	interesting
18:05:22	renesac_	does an addition/subtraction step fix that lack of 0 thing?
18:05:30	renesac_	w/o introducing the lack of another number?
18:05:42	renesac_	well, away
18:05:49	*	renesac_ is now known as renesac|away
18:07:53	nequitans_	yea, that's an interesting question
19:11:34	*	jbe joined #nimrod
19:13:31	jbe	does anyone know if there is a way to yield process control in nimrod without specifying a wait time - similar to sleep(0) on windows or sched_yield on unix?
19:17:10	jbe	never mind, i can just use the posix library
19:18:12	dom96	yeah, I think that may be the only way
19:23:34	Mat3	use the Posix wrapper
19:30:45	jbe	yep that works fine, thanks
19:33:18	*	noam_ quit (Ping timeout: 240 seconds)
19:34:37	*	noam joined #nimrod
19:39:52	Mat3	hello Araq: Have you found a strategy for limiting the type zoo yet ?
19:40:59	Araq	Mat3: well I have 2 "narrowing" instructions but I have no idea whether that's correct for every edge case
19:41:15	Araq	people will report bugs though, so I'm not too worried
19:45:49	*	noam quit (Ping timeout: 264 seconds)
19:45:51	Mat3	I think that says, you extend all types from a basic one ?
19:46:30	Araq	yes, well now that you say it, it can't work for uint64 :-)
19:47:21	renesac|away	:/
19:47:24	Araq	it uses int64 operations for everything and then narrowSigned and narrowUnsigned ops bitmask the result
19:47:36	*	renesac|away is now known as renesac
19:47:57	*	noam joined #nimrod
19:48:37	renesac	don't this introduce some overhead?
19:48:51	renesac	or from the C code the compiler optimizes it out?
19:52:06	Araq	renesac: I'm talking about the VM used for compile time ops
19:52:16	Araq	not about the other code generators
19:52:36	renesac	oh, ok
19:52:49	Araq	and currently we treat uint64 as a second class citizen already so it's fine for now
19:52:55	renesac	then it makes sense
19:53:30	renesac	and nimrod don't exposes a uint128 right? (I remember seeing something like that somewhere in the compiler though)
19:53:50	renesac	for SIMD
19:54:24	*	noam quit (Ping timeout: 252 seconds)
19:54:32	Araq	well actually we need to rework all this for int128
19:54:47	Araq	we'll have a more generic intXX type internally then
19:54:55	Mat3	I am a bit confused; Integer and unsigned integers can be both represented as ... integers (2 complement make it possible)
19:55:06	Mat3	so what's the problem ?
19:55:13	Araq	Mat3: that is correct but div for instance differs
19:55:27	Araq	so you need sdiv and udiv
19:55:36	profmakx	Mat3 you are confused in the wrong way
19:55:48	profmakx	you want to think about how you n bits are interpreted
19:56:09	Araq	for addition/substraction it doesn't matter, for multiplication it's dfiferent again iirc
19:56:38	Araq	and it's all different when you consider overflows
19:56:39	profmakx	primitive types are a pita
19:58:02	renesac	well, in my code I aways end up with uints and bitwise operations, so good nimrod support for them is important for me...
19:58:15	Araq	profmakx: that's why I keep saying I'm glad unsigned floating point numbers don't exist
19:58:33	Mat3	not really, All needed ist masking of a single bit for integer divisions. There are plenty processors avariable which only support unsigned division (if at all)
19:58:52	Mat3	^is
19:59:59	Araq	Mat3: how does that work in detail?
20:03:16	Araq	renesac: I know, don't worry
20:03:21	Mat3	simply mask out the sign bit and adjust negative numbers
20:04:22	Mat3	example:
20:05:19	Mat3	MOV RBX; RAX ; RAX : RBX
20:05:34	Mat3	TEST RBX, RBX
20:05:43	*	Demos joined #nimrod
20:06:13	Mat3	sorry
20:06:24	Mat3	MOV RDX, RAX
20:06:35	Mat3	TEXT RAX, RAX
20:06:39	Mat3	^test
20:09:05	Mat3	then, xor rdx, rax set the high bit of rdx if both have a different sign
20:10:28	Mat3	because rdx can be used as mask for 'NOT' it is now possible to:
20:10:35	Mat3	MUL RDX
20:11:11	Mat3	regardless of the sign of both values, because
20:11:19	Mat3	NOT EAX
20:11:23	Mat3	NOT EDX
20:11:28	Mat3	ADD EAX, EDX
20:12:20	reloc0	i'm currently trying to understand nimrod better. and writing some code with it, trying to learn it.
20:12:28	Mat3	(please replace RDX with RBX)
20:12:34	Mat3	so:
20:12:46	reloc0	every now and then i have some problems with the documentation.
20:13:23	reloc0	can i help to improve documentation or give example programs?
20:13:33	reloc0	and if so - what's the best way?
20:15:07	Mat3	mov rbx, rax; test rax, rax; xor rbx, rdx; test rdx, rdx; mul rdx; test rbx, rbx; not rax; not rdx; add rax, 1
20:15:39	Mat3	that code should divide signed and unsigned integers
20:16:30	Araq	reloc0: the docs are in RST in $nimrod/doc, just edit the .txt files and make a pull request
20:16:44	Araq	nimrod rst2html file.txt
20:16:49	Araq	produces the html version
20:18:55	reloc0	Araq: cool. thanks
20:18:56	Mat3	Araq: You will find such tricks in every common Z80 assembly book for example
20:19:16	Araq	Mat3: yeah but I'm lazy so I am asking you ;-)
20:19:36	Araq	also a sequence of instructions is bad for a VM
20:19:52	Araq	where CISC is the name of the game
20:22:23	Mat3	well, just implement an unified division instruction (your Out-Of-Order cpu will have no problem with it)
20:24:07	Mat3	I just found out the Small C compiler compiles this way, check it out
20:24:18	Mat3	ciao
20:24:30	*	Mat3 quit (Quit: Verlassend)
20:35:17	Varriount	reloc0: The library documentation is mostly in the docstrings of types, procedures, and other code in the library source files
20:40:22	Varriount	OrionPK: ping
20:47:13	*	Demos quit (Ping timeout: 240 seconds)
20:55:01	*	q66 joined #nimrod
20:57:01	Varriount	Araq: Do iterators have result variables?
20:57:21	Araq	no
21:00:35	renesac	Varriount, why the bug I filled about the implicit initialization of the result variable, even when it isn't assigned to, was tagged as 'feature'?
21:01:29	renesac	Araq, right, good to know
21:01:57	renesac	https://github.com/Araq/Nimrod/issues/868 <-- this one
21:02:07	Varriount	renesac: I don't know. I don't control the issue tracker.
21:02:36	renesac	ah, as you were the only one to answer I though it was you
21:03:20	renesac	maybe I should rename it, reflecting the edit
21:04:05	Varriount	renesac: I stated that it was arguable whether this was a bug or a feature.
21:04:59	renesac	how "proc foo():int = var bar = 42" returning 0 is a feature?
21:05:36	renesac	(in my edit I discovered the 'result' identifier wasn't the problem
21:05:44	Araq	it *is* a feature
21:05:56	Varriount	The result value is always initialized to 0
21:06:11	Varriount	Or rather, the memory that it uses is zeroed
21:06:16	renesac	yes, but this hides programmer errors
21:06:28	renesac	if you return w/o even touch the return value
21:06:46	renesac	I'm not against it being initialized to 0
21:07:28	renesac	but the compiler should complain when a proc is supposed to return something, and you don't use either 'result' or 'return' to set this return value
21:07:28	Araq	well we can easily produce a warning
21:07:33	fowl	<renesac> if you return w/o even touch the return value -- what if this what you meant to do, knowing the semantics, then its not an error
21:07:57	renesac	'proc foo():int' will complain of missing implementation
21:08:16	Araq	we recently got the necessary control flow analysis
21:08:39	renesac	right
21:09:44	*	Demos joined #nimrod
21:09:45	renesac	well, if it has at least one return somewhere, returning 0 by default w/o touching may be desired
21:10:09	Araq	indeed the compiler uses this idiom heavily
21:10:32	renesac	right
21:11:00	Demos	you can say proc foo(): int = 4 though right?
21:11:02	renesac	I wonder if it is easy do differentiate the two cases
21:11:06	Demos	and it will return 4
21:11:09	renesac	Demos, yes
21:11:16	Araq	IME it doesnt hide bugs, it hides 'result = 0 | false'
21:11:36	Demos	it hides boneheaded errors
21:11:54	Demos	imo it should be a warning
21:11:57	Araq	I said IME
21:11:57	renesac	yeah
21:12:00	Demos	like it is in C
21:12:11	Araq	so ymmv
21:12:34	renesac	and allowing or not to redeclare a 'var result' is another issue
21:13:00	renesac	that shadows the original one, but isn't used by default for returning
21:13:59	renesac	well, any experienced nimrod coder wouldn't make this error anyway... just me when messing around the first time...
21:14:24	renesac	that isn't high priority though..
21:14:34	Varriount	Araq: How exactly do I produce a new test result for the babel package build test?
21:15:02	Varriount	Unlike the other tests, this one doesn't use test files with specs
21:16:28	renesac	Araq, how fast would be to querry the memory allocator for the capacity of an array, for example?
21:16:52	renesac	I find it kinda a waste to duplicate this information...
21:18:39	renesac	only for the few uses where you allocate a seq but don't want it's length to follow the capacity...
21:20:20	renesac	you would need to query the memory alocator, maybe in one step via a realloc with special flags, once every array duplication (assuming normal dinamic array growth here)
21:21:02	Varriount	renesac: sizeof()?
21:21:39	renesac	I think if you ask sizeof() a pointer it will return 4 or 8...
21:23:07	renesac	and for a struct, it will return the baseline size, not the size of the slot the memory allocator decided to insert it
21:23:46	Araq	Varriount the tester has lots of special stuff in categories.nim
21:24:35	Araq	renesac: it would be very fast
21:25:10	renesac	so, isn't this idea interesting?
21:26:38	Varriount	renesac: But what could it be used for?
21:26:44	dom96	I spent an entire day trying to get scrolling past bottom working in Aporia.
21:26:45	renesac	you ask for a newSeq(cap=10000), and then internally it calls a 'realloc(shrink=false)' every doubling
21:26:57	renesac	(not using the current API)
21:27:17	renesac	Varriount, arrays and seqs with 4/8 bytes less overhead
21:27:25	*	dom96 </3 GTK
21:27:41	renesac	that could translate to even bigger savings
21:28:14	renesac	if your seq type is a tuple 'len, pointer'
21:28:33	Araq	renesac: it's in my todo for years now
21:28:46	renesac	if you add another field, it would not be a power of two anymore, and would likely be allocated in a bigger slot
21:28:53	Araq	so sorry it's not interesting ;-)
21:29:00	renesac	hehe
21:29:04	renesac	*heh
21:30:13	renesac	btw, is it safe to store information in lower bits of pointers?
21:30:17	renesac	or it confuses the GC?
21:33:45	dom96	Who wants $10K? If somebody uses Nimrod for this it would give us some really good publicity: http://www.raspberrypi.org/archives/6299
21:35:18	renesac	* Okay, so sue us**
21:35:20	renesac	** Please don’t sue us.
21:35:24	renesac	lol
21:35:56	Araq	renesac: sorry not safe and would have desastrous results
21:36:19	renesac	:(
21:36:42	renesac	can't the GC mask those bits, or it already use them?
21:36:51	renesac	*uses
21:38:31	jbe	how should i go about iterating through some bytes that i have a pointer to? (sdl surface pixels actually)
21:40:11	dom96	Araq: Wow, that reddit post of yours is at ~50 upvotes. There are people out there that respect you :)
21:42:32	dom96	jbe: Something like this perhaps? for x in 0 .. <width: for y in 0 .. <height: pixels[x + y]
21:44:11	Varriount	dom96: Link to the post?
21:44:35	dom96	http://www.reddit.com/r/programming/comments/1z6j37/on_garbage_collection_in_games/cfrb3jy
21:45:00	dom96	Might as well look at all the comments though to get the context: http://www.reddit.com/r/programming/comments/1z6j37/on_garbage_collection_in_games/
21:45:31	Demos	idetools seems to have trouble with spaces in paths
21:45:40	Araq	renesac: it is very easy to unmask these bits but the consequences are awful for performance
21:46:14	*	q66 quit (Ping timeout: 252 seconds)
21:46:18	Varriount	Demos: escapeShell?
21:46:28	Varriount	Or just put quotes around the path
21:46:35	fowl	jbe, check out the source for the graphics library
21:46:39	renesac	really? I would think that masking was a cheap operation
21:46:41	renesac	:P
21:46:43	Demos	I tried both
21:46:49	jbe	dom96: like, cast the pointer to a ptr byte, and then use thepointer[i]?
21:47:19	dom96	jbe: If you're using sdl then you might as well use the graphics module.
21:47:35	Demos	maybe not the spaces
21:47:53	dom96	jbe: Cast it to an array[0..1000_000-1, int32]
21:48:04	renesac	too bad then... another place that might hide a bit would be the sign bit of a seq lenght, but it isn't nearly as widely useful...
21:48:52	Araq	the masking itself is not measurable of course :P
21:48:53	renesac	but at least I could def a bigint as a 'distinct seq'
21:49:35	renesac	or 'seqWithHiddenBit'
21:50:05	Araq	there are other consequences that make it slow
21:50:09	Demos	OK idetools actually has trouble with the dirty file being in my temp folder
21:50:23	renesac	seqs are magic, and you pointed me where they were defined sometime ago, but I lost it
21:50:46	renesac	Araq, like what, now I'm curious (not criticizing)
21:50:47	jbe	dom96: thanks. well, it's just to load files into a texture packager. i'll see how it works
21:50:53	renesac	?
21:54:00	fowl	dom96, can i put sdl2 on nimrod-code
21:55:01	dom96	fowl: sure. The wrapper is complete right?
21:55:55	Varriount	dom96: The comments in that thread... they bring tears of pride to my eyes. Go filwit! Go Araq!
21:56:37	dom96	Varriount: hehe
21:56:43	fowl	dom96, net and audio needs work
21:57:13	*	nequitans_ quit (Ping timeout: 240 seconds)
21:57:20	dom96	fowl: it should be fine.
21:57:35	*	brson joined #nimrod
21:57:37	fowl	ok
21:57:49	jbe	dom96: thanks. well, it's just to load files into a texture packager. i'll see how it works
21:57:56	jbe	oops
22:06:19	fowl	dom96, do we have something that launches a platform-independant open file dialog
22:06:36	fowl	gtk on linux, win gui on windows
22:06:38	dom96	fowl: dialogs module
22:09:23	Demos	oh it fails if run in COMMAND.COM
22:09:50	Varriount	Command.com doesn't exist in windows anymore.
22:09:56	Demos	whatever
22:10:02	Demos	the old style command shell
22:10:07	Varriount	\nitpick
22:10:22	Demos	fails in powershell as well, does it assume some sh stuff?
22:12:02	fowl	dom96, how about naming, is import sdl2/ttf ok? :D
22:13:02	dom96	fowl: yeah
22:15:59	Varriount	Araq: Almost have testament/babel integration done. I'm testing the output right now.
22:17:13	*	OrionPK quit (Ping timeout: 240 seconds)
22:21:02	Araq	excellent
22:21:14	skrylar	meepity meep
22:22:19	Varriount	Hi skrylar!
22:22:32	Varriount	Did you see Araq & filwit's post on reddit?
22:35:01	*	Endy quit (Ping timeout: 264 seconds)
22:37:29	Demos	WoooWWWWWWW VS completion
22:39:39	profmakx	when people say "syntax is odd" they mean "it doesn't look like C" right?
22:39:42	profmakx	or
22:39:56	profmakx	"it looks like C, but works differently, I want it to work like C"
22:42:07	Varriount	profmakx: That's my guess.
22:42:35	Demos	yeah, also nimrod ignores pseudo software engeneering
22:42:54	Varriount	Demos: Huh? "Pseudo software engineering"?
22:43:42	Varriount	Araq: What should the tester do if it can't find babel?
22:43:44	Demos	stuff like strict scope control (beyound sane modules) and extremely explicit everything.
22:44:01	Demos	and sticking everything to objects
22:44:15	Varriount	Demos: You mean, Java?
22:44:42	*	Varriount is now known as Java
22:44:42	Demos	yeah, and c++ a little
22:44:50	Java	MOOOAR OBJECTSSS
22:45:06	Demos	some of these things seem to make sense when you are coming from C/C++ and have very little control over scope
22:45:11	*	Java noms all the memory
22:45:11	Demos	imo
22:45:17	*	Java is now known as Varriount
22:47:52	dom96	oh look. A new programming language: http://wasyl.eu/programming/2014/03/01/introducing-september.html
22:49:32	profmakx	reinventing lisp macros now with curly braces>
22:49:53	profmakx	i really hate syntax with {} for blocks and ; as terminator
22:49:55	Varriount	To be fair, he does say that it's a hobby
22:50:07	Varriount	profmakx: So do I! :D
22:50:11	profmakx	of course
22:50:19	profmakx	everyone should have implemented at least one language
22:50:20	dom96	It seems that 99% here do :P
22:50:33	*	profmakx calls bias
22:51:25	*	psquid joined #nimrod
22:51:45	Varriount	Hi psquid
22:52:36	dom96	It seems like a perfect place to mention Nimrod
22:52:55	Varriount	dom96: That could be seen as inflammatory
22:53:40	dom96	True.
22:55:07	renesac	only mention nimrod if you have something to add
22:55:18	renesac	and well, rust was already mentioned in the reddit thread
22:55:27	dom96	of course it was...
22:55:33	Varriount	>:/
22:55:56	Varriount	Throw the infidels into the pits of DOOOOM!
22:56:14	*	Varriount gets the torches and pitchforks
22:56:25	profmakx	i should put out a disclaimer of the form
22:56:28	renesac	you can answer that rust poster, saying that nimrod metaprogramming is much more powerful, and also works on the AST
22:56:30	profmakx	"spare me the rust talk"
22:56:51	dom96	I could mention Araq's strong spaces idea here: http://www.reddit.com/r/programming/comments/1z9ppe/introducing_september_working_on_an/cfruobb
22:56:56	renesac	but compare with september too, or it will be purely a flame
22:57:37	dom96	or I could show an example of 'until' here: http://www.reddit.com/r/programming/comments/1z9ppe/introducing_september_working_on_an/cfrusbg
22:59:37	fowl	dom96, i guess ihave to set up a window though
22:59:59	dom96	fowl: hrm?
23:00:04	EXetoC	profmakx: that syntax should be illegal
23:00:20	fowl	dom96, sry, for dialogs.nim, btw this module should be moved to babel
23:00:30	fowl	it uses glib/gtk2
23:00:53	dom96	fowl: hrm, yeah.
23:01:18	Varriount	My very fuzzy havanese is sitting in front of a space heater. Why?
23:01:23	dom96	I guess it may as well be a part of the gtk package?
23:03:15	fowl	i would put it in a separate package, possibly with browsers.nim
23:03:36	dom96	I think the tester uses browsers.nim
23:04:56	*	Varriount|Mobile quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com ))
23:26:29	*	jbe quit (Quit: Leaving)
23:36:06	fowl	yea it does
23:37:38	*	runvnc joined #nimrod
23:37:42	runvnc	hello
23:37:52	fowl	hey
23:37:52	runvnc	is there an easy way to convert from a tuple to json
23:38:20	Araq	marshal.nim uses a json representation
23:38:36	runvnc	oh ok thanks a lot even better
23:50:55	*	ddl_smurf quit (Quit: ddl_smurf)
23:52:16	*	ddl_smurf joined #nimrod
23:58:02	NimBot	nimrod-code/Aporia master b150d1c Dominik Picheta [+0 ±3 -0]: Added first implementation of scrolling past bottom.
23:58:28	NimBot	nimrod-code/packages master 71ec516 Billingsly Wetherfordshire [+0 ±1 -0]: added sdl2