<< 03-03-2014 >>

---

00:00:17	dom96	Varriount: Isn't that a bug?
00:00:45	NimBot	Araq/Nimrod devel 5f7ac15 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Adds posix.timegm(), brother of posix.mktime().
00:00:45	NimBot	Araq/Nimrod devel 3857496 Andreas Rumpf [+0 ±1 -0]: Merge pull request #927 from gradha/pr_adds_posix_timegm... 2 more lines
00:01:22	*	Matthias247 quit (Read error: Connection reset by peer)
00:07:12	Araq	Kelet: marshal uses nimrod's typeinfo stuff. It is reasonably easy to adapt marshal to produce a binary reader/writer
00:07:21	Araq	(welcome btw)
00:08:57	filwit	http://forum.nimrod-lang.org/t/386
00:09:21	filwit	^ okay folks, made a forum post about a new syntax idea i had last night
00:10:55	filwit	i think it address everyone's concern (just about) and has a (theoretically) realistic path implementation (changes can be done in steps, and things like {} sytle is more of an extension once certain symbols are rearranged)
00:11:06	filwit	path too**
00:11:11	filwit	to**
00:11:35	fowl	var mySet = @{ A, B, C }
00:11:41	fowl	what is {a,b,c} then?
00:11:52	filwit	please read the whole post
00:12:08	fowl	yeah
00:12:16	fowl	it appears you want it to be a statement list
00:12:17	filwit	{} would be reporposed for white-space insignificant "body"
00:12:50	fowl	you know @ is a unary function on openarray that creates seqs right
00:13:22	filwit	fowl: didn't know that actually
00:14:19	*	q66 quit (Quit: Leaving)
00:14:25	filwit	fowl: but i'm sure that problem could be addressed, or perhaps a better syntax for set/seq could be found (though i think the @[]/@{} is just fine)
00:15:38	fowl	{} is also used for dictionary
00:15:42	EXetoC	why should it look more like seq construction rather than array construction?
00:15:43	renesac	filwit, not having the program behave like it looks like (it's indentation) is bad in my opinion
00:16:08	fowl	@{"hello":"there"} == @[ ["hello","there"] ]
00:16:12	renesac	and one could use #{ and #} for redundancy in defining blocks
00:16:18	renesac	for e-mails, etc
00:16:25	renesac	no need to repurpose {}
00:16:49	filwit	renesac: the point of use {} for brackets is purely for popularity
00:17:02	renesac	it will lose popularity among python users
00:17:03	renesac	:P
00:17:03	filwit	renesac: and honestly, set's don't really need them
00:17:14	renesac	not only sets are defined by them
00:17:17	fowl	this is true
00:17:19	renesac	there are {}
00:17:22	renesac	*{:}
00:17:25	renesac	tables
00:17:26	fowl	i think it would be better to just use set(x,y,z)
00:17:39	filwit	fowl: i agree actually
00:17:48	Araq	there is also a{i}
00:18:00	Araq	(similar to a[i] )
00:18:07	fowl	filwit, i dislike your proc pragma change, if you want it to be more like types, it should be proc foo* {.pragma.} (args)
00:18:40	filwit	fowl: i'm not against that actually, though I'm not sure i like it better
00:18:42	Araq	filwit: also you can already do: type Foo = object {.pragmas.}
00:18:54	fowl	and swapping around : and = in the proc def will break tons of code, a better idea would be to make proc declarations the same as do() ie proc foo*(x:int) -> string:
00:19:11	filwit	Araq: i'm not sure what you're comparing that {.pragmas.} too
00:19:44	filwit	fowl: how would that break any less code? and this isn't like a "we need it NOW" kind of thing
00:19:50	Araq	it comes late then, similar to how it's done for procs
00:20:25	Araq	but I'm sleeping anyway, bye
00:20:28	filwit	Araq: i know you can do that, but isn't that fundamentally different?
00:20:29	fowl	filwit, because when the proc hits the "=" in the definition, it knows it needs a body, you allow "= returnType : body"
00:20:36	filwit	Araq: okay, will talk later
00:21:08	filwit	fowl: we're talking about parser changes man. '=' would now mean 'is type' and ':' would invoke the body
00:21:38	fowl	to me it looks like proc () = float: foo is a void function calling a template float with body code foo
00:21:54	fowl	and thats what i expect others who know nimrod to read that as
00:22:04	renesac	fowl, a template in this case would need 'do'
00:22:29	renesac	because some other ambiguity
00:23:08	filwit	one thing is that ':' really should be use for "start body" because of everything else that uses it if/else/custom-macros etc
00:23:24	fowl	yea
00:23:32	renesac	on the other hand, ':' is also used to specify type
00:23:42	renesac	I'm not sure what is more consistent
00:23:46	fowl	but now we have do -> ty : body, proc = ty: body, proc : ty = body,
00:24:27	renesac	yeah, it would be good to unify 'do' and normal procs
00:24:31	filwit	fowl: the last would be removed
00:24:54	filwit	fowl: and the first is only to pass things around and probably could have a better solution?
00:25:24	Araq	I hate to say it, but if we remove ':' for types things might get easier: proc foo(a, b int) int: body
00:25:26	EXetoC	death to excessive use of brackets - optional or not
00:25:30	dom96	filwit: I don't think allowing users to choose between brackets/indentation will result in a popularity increase.
00:25:59	renesac	^
00:26:02	fowl	filwit, remove the last option and you've rendered 100% of nimrod code uncompilable
00:26:19	Araq	but I think these go-like declarations really look ugly
00:26:32	filwit	dom96: completely disagree really. it might not "make a splash" but based on complaints i've heard it would help adoption
00:26:45	filwit	Araq: they could be on multiple lines
00:26:56	filwit	Araq: eve the ':' could be on a new line
00:26:56	Araq	filwit: that's awful
00:27:04	filwit	Araq: it's choice
00:27:12	EXetoC	everything is being criticized all the time...
00:27:13	renesac	filwit, it would just split the language, and make people read those ugly code
00:27:34	Araq	I'm sleeping anyway
00:27:35	dom96	filwit: You want to increase consistency right? Well, this would make things less consistent.
00:27:37	filwit	renesacs: kinda like how case-sensitivity doesn't do?
00:27:38	filwit	:P
00:28:13	filwit	dom96: i want to increase consistent with symbol usage. I always liked Nimrod's "you choose" attitude
00:28:24	renesac	well, most languages don't really force some case standard
00:28:39	filwit	yes, the lib-writers doo
00:28:45	renesac	and with case-insensivity you can call the way you want
00:28:56	EXetoC	you still need prefixes
00:28:58	filwit	which means you have to do: SDL_CreateWindow(getNameFrom(something_else_entirely()))
00:29:12	filwit	but that's nicer in Nimrod because there's choice
00:29:29	dom96	I don't get why you want to change the syntax of type definitions though.
00:29:40	filwit	this is basically the same idea, only with statment structure
00:29:49	dom96	I can understand why you may want: proc x(y: int) -> int: though.
00:29:55	filwit	dom96: i don't, types are pretty much left how they are
00:30:15	filwit	dom96: i even question needing to enfore ':' use at the end of the post
00:30:52	fowl	type Foo: bar: int
00:30:52	fowl	<- what is this
00:31:05	dom96	filwit: I think the : is required in if statements at least.
00:31:11	filwit	Araq: i actually agree that removing the ':' might make things easier for procs
00:31:36	filwit	dom96: yeah i agree mostly, which is why i was promoting enforcing it :P
00:31:44	dom96	writing proc x = int doesn't make sense to me though.
00:32:02	dom96	"procedure x equals type int"?
00:32:08	renesac	still makes a severe compatibility break
00:32:24	renesac	dom96, I agree, -> is much better, and already used in 'do'
00:32:45	filwit	dom96, renesac: no what Araq said was better.. just remove ':' altogether
00:32:51	filwit	proc x int
00:32:55	filwit	proc x int: ...
00:33:02	filwit	proc x int
00:33:03	filwit	..
00:33:04	fowl	filwit, the only change listen here that i can agree wiht is allowing : in type defs ie type A : object
00:33:05	filwit	?
00:33:26	fowl	listend*
00:33:27	renesac	filwit, he didn't really said, it was better, he said ' I hate to say it'
00:33:29	renesac	XD
00:33:29	fowl	listed
00:33:38	renesac	but yeah, it is an option
00:33:47	EXetoC	renesac: though it's feasible to automate migrations for grammars that are parseable by normal human beings
00:34:08	dom96	I wonder if he was kidding.
00:34:19	dom96	because that looks absurd to me
00:34:20	filwit	i don't think so, just pointing something out
00:34:24	filwit	lol
00:34:43	renesac	anyway, I'm oposed to identation insensitity, and I don't see that much gain for such a big change
00:34:44	filwit	we had talked about the merits of ':' in the past for var declarations
00:34:55	renesac	actually, it is pure bikeshedding
00:35:34	filwit	renesac: i don't see for what reason you are apposed to case-insensitivity
00:35:52	filwit	renesac: plus this is of course a bikeshedding issue.. but one that gets brought up a lot
00:36:07	fowl	i wish someone would write a c-like syntax
00:36:08	renesac	we just need a good FAQ entry on that
00:36:11	renesac	I may provide this
00:36:31	filwit	renesac: this isn't a "let get started guys" proposal, but a "future directions to make Nimrod inviting and consistent" proposal
00:37:09	renesac	brackets aren't inviting for me... else I would be using/trying D or Rust
00:37:12	filwit	renesac: do you just "not like" the idea or others using brackets? I don't really find that a valid complaint
00:37:30	filwit	the whole point of this is that many feel the exact oposite
00:37:38	filwit	opposite*
00:39:06	dom96	https://gist.github.com/dom96/959f27018ea3247a4498
00:39:18	dom96	To me it's perfectly consistent.
00:39:31	dom96	The comments are how I would read the code in my mind.
00:39:47	EXetoC	Chances are I'll be reading such code written by others, and that's why it matters
00:39:54	renesac	from __future__ import braces
00:40:22	filwit	EXetoC: the same rules of case-inconsistency apply to this
00:41:16	filwit	dom96: it's not consistent. because Bar's reclist just follows the 'object' part
00:41:39	filwit	dom96: only you have to have a \n to invoke that, since not even [,] work for objects (only tuples)
00:41:46	dom96	'object' starts a new section
00:41:53	filwit	i know
00:41:53	dom96	similar to the way 'type' does
00:41:57	filwit	i know..
00:42:05	dom96	So it is consistent.
00:42:27	dom96	It's only slightly different because it comes after an '='
00:42:30	filwit	only it's a completely different sort of symbol. 'type' is a keyword, 'object' is not
00:42:39	dom96	object is a keyword
00:42:48	renesac	'if' only has a ':' because it needs that when you do it in one line
00:42:53	filwit	it doesn't matter, it's not used the same, dom96
00:42:57	fowl	filwit, are yous ure object[a:int] doesnt work? i know tuples work both ways now
00:43:02	renesac	let, type, object don't need ':' for one-liners
00:43:26	renesac	so they don't have it either when you do in multiple lines
00:43:30	dom96	let, var, const, type, object, tuple -- all consistent.
00:43:31	filwit	yes, in that i agree.. i've agreed many times that the ':' could probably be optional
00:43:46	dom96	oh, tuple and object are good examples of consistency
00:44:05	filwit	type Foo = object[a, b: int] # doesn't work still right?
00:44:13	dom96	Why should it?
00:44:32	dom96	What's the difference between that and s/object/tuple/ ?
00:44:36	EXetoC	filwit: well, I don't feel too strongly about that
00:44:56	filwit	dom96: i'm not following you
00:45:13	dom96	My point is that you may as well just use a tuple in that case anyway.
00:45:16	EXetoC	dom96: the construction and the significance of the names
00:45:40	filwit	dom96: that has no relevance on weather or not the syntax is consistent
00:46:45	dom96	filwit: Alright. Fair enough. I suppose consistency doesn't necessarily matter if there is another way of doing the same thing.
00:47:39	filwit	dom96: i don't even fully understand the difference between object/tuple in Nimrod, beyond exposure semantics
00:48:06	dom96	Well AFAIK. With object you can inherit from different objects
00:48:17	dom96	With tuples you can access them by index
00:48:19	fowl	object can have invariants and use the when stmt inside them
00:48:30	dom96	yeah, that too.
00:48:35	dom96	tuples are much simpler
00:48:44	dom96	which is why it makes sense to define them using that syntax
00:48:55	dom96	and why it makes sense to define objects with that long syntax
00:49:18	filwit	okay that makes sense
00:49:19	dom96	But in the name of consistency we can ignore that I guess.
00:49:42	*	darkf joined #nimrod
00:49:59	renesac	filwit, changing subject, you should make a tutorial on how to use intrinsics in nimrod
00:50:13	renesac	do you have some wrapper already made?
00:50:37	EXetoC	what's that? compiler pragmas?
00:50:48	filwit	renesac: yeah, it's somewhat primitive atm. But one sec
00:50:51	EXetoC	nvm. wrapper?
00:51:11	EXetoC	unless that was an unrelated question
00:51:14	filwit	EXetoC: he means simd stuff
00:51:35	EXetoC	ok I didn't know about that
00:52:43	renesac	yeah, <intrins.h> and friends
00:53:04	dom96	filwit: But seriously. Please stop worrying about these things. Just accept the way things are for now and work on more important things, like making the docs nicer, that should give us a better popularity boost than some consistency adjustments.
00:53:08	filwit	renesac: here's a basic lib with a Vec4: https://gist.github.com/PhilipWitte/9316566
00:53:29	filwit	renesac: just wrap more of xmmintrin.h to get more functions
00:54:40	renesac	cool
00:55:24	filwit	dom96: i'm not "worrying" about them at all anymore. I had an idea last night about it and posted to Araq personally before posting here. He suggested making the forum post and talking about it with you all. I will being making a parser which works like this on my own if you folks really don't like it (to experiment), but that will be far away from now.
00:59:22	dom96	filwit: I will take a more thorough look at your proposal tomorrow because it is quite late here and I am very tired. But for now I will stick to what I said, I think the language is perfectly consistent and I still don't understand why you think otherwise.
01:04:16	filwit	dom96: can you do "type X = var\n " ? no.. there's only one place in the compiler where a body directly follows a keyword and gets assigned to a unique symbol. However, there's many places where a stmt body is passed to something, and even if/proc work differently here. It's true that proc is somewhat consistent with how 'var' works, but then, these are sorta fundamentally different beasts. Definiting a proc is more like definging a type than a var really.
01:06:33	dom96	I disagree. Definining a proc is exactly like defining a var.
01:06:50	dom96	var x: int = 5
01:06:56	dom96	proc x: int = 5
01:07:11	renesac	filwit, the '=' use in procs is consistent with variable declaration:
01:07:12	renesac	https://gist.github.com/ReneSac/02076a8ed3cfc4d559c3
01:07:34	filwit	which is why i always wanted to make types like this: "type Foo: Bar = <body>"
01:08:15	filwit	but then this is completely inconsistent with if/else/macros and I started to see how type wasn't the problem
01:08:41	filwit	ultimately, it would be nice if all "body's of executing code" followed a consistent symbol.. ':'
01:09:34	filwit	and then, you could apply the same to types as well to be consistent, which also opens the door to allow optional {} style, which wins all arguments in terms of adoption (honestly, why NOT do it if you have the option?)
01:09:42	dom96	Sometimes you just have to accept that not everything can be consistent. Because the two things are different enough that they just cannot be.
01:10:46	*	ddl_smurf joined #nimrod
01:10:55	filwit	sorry, but i don't accept arguments of complacency :P i'm not saying this is "so important Nimrod will never get uber-popular without it".. i'm saying "i think there's a better way, here's why"
01:11:26	dom96	Araq was going to implement the {} and begin/end style
01:11:30	dom96	using filters
01:11:31	renesac	filwit, this is also about convenient line-breaking
01:11:53	dom96	I'm not sure why he changed his mind.
01:11:54	renesac	'let x: int = 1 + \n 2' is also permited
01:12:09	renesac	'=' isn't special case
01:12:31	dom96	Yeah, that's true.
01:12:54	filwit	renesac: yet, but this fundamentally feels different than 'object\n <recList>'
01:13:09	filwit	and in the case of types, the ':' doesn't apply
01:13:23	dom96	The problem I can see with allowing {} is that you may get code which doesn't care about indentation which will happen, and then you are confused if you are used to the indentation style.
01:13:39	dom96	You scroll up and down "ugh, this code uses braces"
01:14:02	renesac	besides haskell, I don't know any language that allows both styles
01:14:05	dom96	Araq's original plan was to allow the IDE to render the Nimrod AST whichever way you preferred.
01:14:20	dom96	Which would solve that problem.
01:14:23	renesac	neither side seems open to add support for other block definition style
01:14:39	filwit	dom96: that is a valid concern. but I'm not sure how much that will happen in practice. Usually people will just chose one or the other, and you could also have compiler-option to enforce specific ones if you really wanted
01:15:05	dom96	I'm talking about other people's code here./
01:15:12	dom96	Not my own code which uses different styles.
01:15:23	dom96	Say I want to fix some bugs in other people's code.
01:15:37	renesac	I'm more ok if {} is used only for redundancy, w/o letting people disregard proper identation
01:15:48	EXetoC	I want consistency without having to rely on tools
01:15:56	filwit	dom96: IDE rendering is a good point too..
01:16:00	renesac	but even then, I don't think this cruft is necessary in the language
01:16:20	dom96	filwit: However, there is no reason not to get the IDE to translate to the indentation style
01:16:35	dom96	You can still get this feature, even if the language doesn't support {}
01:16:52	renesac	and for redundancy, one could easily implement '#{' '#}' w/o breaking any backwards compatibility
01:16:53	filwit	that's very true
01:17:15	filwit	renesac: it's not for redundancy. that would be pointless
01:17:43	renesac	so for what?
01:18:01	renesac	most of the arguments against identation for defining blocks that I hear are:
01:18:05	filwit	i meant, no one is arguing for {} as just a redundant wrapper for code
01:18:08	renesac	they are lost in e-mails and blog comments
01:18:27	renesac	the IDE might use them to re-ident code cop
01:18:31	renesac	ied from somewhere else
01:18:46	renesac	to make block start and finish more visible
01:18:58	renesac	all of those are served by redundancy
01:19:05	renesac	actually, ask for that
01:20:08	filwit	well i don't mind the object definition so much when it's like this:
01:20:10	filwit	type Foo =
01:20:13	filwit	ref object
01:20:16	filwit	a, b: int
01:20:28	filwit	because it's clear what is happening
01:20:33	filwit	but still, i think:
01:20:37	filwit	type Foo: ref object =
01:20:40	dom96	The advantage of the type *section*
01:20:42	filwit	is best
01:20:44	dom96	is that you can write 'type' once
01:20:51	dom96	and define multiple types
01:20:53	renesac	think of 'object' like a unary operator, but that can receive a block of code
01:20:58	filwit	i don't see how this effects that, dom96
01:21:01	dom96	Which is the same advantage that let/var etc has
01:21:07	renesac	yeah, it is kinda inconsistent, but is convenient
01:21:20	dom96	Well, I find it on the same line ugly :P
01:21:21	fowl	renesac, how is that at all true
01:21:44	renesac	maybe it was a bad analogy..
01:21:50	fowl	lol
01:22:02	renesac	as it isn't a operator, is part of the syntax that can't be emulated, it seems
01:22:18	fowl	i meant the receiving a block of code part
01:22:48	renesac	the fields definition can span multiple lines
01:22:54	renesac	I would call that a block, even w/o the ':'
01:23:36	EXetoC	bacon
01:23:42	dom96	It's not a block per se
01:23:55	filwit	yes there are only two places where something consumes a non-comma-separated "body".. that's 'types' and procs
01:23:56	dom96	But I should really sleep
01:24:03	dom96	Good night guys
01:24:07	filwit	later
01:24:26	fowl	in c the stuff inside struct {} is code
01:24:30	filwit	sorry, not only two places, if/else/etc do it too
01:24:30	renesac	filwit, and template and macro definitions
01:24:36	renesac	and that
01:24:54	filwit	fowl, renesac: i agree with dom96 actually, {} should just be an IDE thing
01:25:02	fowl	its not a "body" at all, its called ident defs
01:25:02	filwit	it just makes the most sense
01:25:28	fowl	you can also find ident defs in var decls and function parameters
01:25:42	renesac	fowl, right, I was really imprecise, it was only an analogy, but didn't help
01:25:51	filwit	fowl: i know it's not a StmtBody (it's a RecList), but it's still not comma-separated
01:26:15	filwit	fowl: so it "feels" like a "body of code" especially since it can have 'when' statements and 'case' too
01:27:46	filwit	i think this is best solution, honestly..
01:27:51	filwit	type Foo Bar: <body>
01:27:58	filwit	proc foo int: <body>
01:28:20	filwit	var foo: int = 0 # '=' is assignment and unique to var/let/const
01:28:48	filwit	then have {} be IDE feature if it's really wanted, etc
01:29:28	filwit	or i guess you could do..
01:29:34	filwit	var foo int: 234
01:29:45	filwit	for complete consistency, but whatever
01:30:37	renesac	actually, dont the (;) syntax already enables C/java like blocks of code?
01:30:58	filwit	renesac: not really, it's just for combining lines
01:31:18	renesac	what would't work?
01:31:38	filwit	doesn't change how white-space is used
01:31:51	filwit	it just allows.. var a = 10; echo a
01:36:36	renesac	ok, it seems that I can't nest (:)
01:36:48	renesac	* (;)
01:37:32	renesac	and yes, identation rules still are followed
01:41:31	Varriount	filwit: I don't mind others using brackets and semicolons, as long as *I* don't have to.
01:42:01	Varriount	(I'm reading logs while taking a break from frantic essay writing)
01:42:52	filwit	Varriount: dom96's points about this where valid. It could cause people to mix-match braces and cause bugs because they're expecting one and get the other. Plus, IDE rendering solves all concerns about Brackets or styling in general
01:43:08	BitPuffi1	Araq: wtf, pow _alias_?? So un-nimrodic
01:43:27	filwit	he's alseep
01:43:31	filwit	asleep*
01:43:45	BitPuffi1	yeah
01:44:32	Varriount	filwit: I am open to the idea about adding @ to set literals ('open' meaning I'm neither for, nor against the idea)
01:46:37	filwit	Varriount: no point really, that was only to free up {} for curly-brace style.. but I (think) i agree with dom96 that issue is better addressed with IDE rendering
01:47:26	EXetoC	I think that was the original proposal
01:47:29	filwit	Varriount: my main reason for that wasn't that I wanted curly-braces, it was that there's a lot of people that do. But this issue is a "meh, once we have a good IDE no one will complain" thing i think.
01:47:36	Varriount	filwit: Hm. Then can you edit your post to reflect your new thoughts?
01:47:45	filwit	EXetoC: yeah, Araq originally planned on having both styles
01:47:53	renesac	filwit, maybe I should't be saying it, but in 10+ years of python I don't remember any IDE that renders python with braces
01:48:05	renesac	well, there is the issue of {} having another meaning, like nimrod
01:48:05	filwit	Varrioiunt: i'be been thinking about deleting it entirely
01:49:30	filwit	renesac: yes it's true that, in terms of {} effecting adoption, it might still be a good idea to add to the language.. but somehow i doubt it since this is ALL just future speculation anyways
01:49:46	renesac	it ends up being a non-issue once the person has to use a language
01:49:48	filwit	renesac: might as well spend time writing the IDE rather than adding changing the language
01:51:01	renesac	and yeah, you might make Araq's syntax-agnostic IDE dream true
01:51:02	filwit	renesac: well my idea wasn't even initially about {} braces.. i just realized that if the code worked like that it would be very easy to market to C#/Java/JS/C++ folks because it looks like a great blend of elements and is consistent that way
01:52:02	Varriount	Nimrod will never be palatable to Java people. It's not constrained enough nor OO enough.
01:52:19	EXetoC	yeah can't discriminate against narrow-minded people :-)
01:52:35	renesac	yeah, and it don't have semi-colons at the end of every line
01:52:35	filwit	Varriount: that's not true once we have OOP macros
01:53:03	renesac	well, I'm going to write that FAQ entry latter, please remind me
01:53:17	filwit	Varriount: which is part of the main drive to these things.. write now i'm making my own OOP macros.. a lot of my code looks like this: "part Foo: var x: int; ..."
01:53:55	filwit	Varriount: see, the body follows the ':' symbol, unlike procs
01:54:25	filwit	Varriount: and the "body" of types is also different, but in an entirely different way
01:54:47	renesac	'part' is supposed to produce a proc, or is opening a new block?
01:55:16	renesac	like a macro or block statement, that also uses ':' , so it is consistent?
01:55:26	*	r0b2 joined #nimrod
01:55:42	filwit	it's produces a type, and edits the incoming procs (adds a 'this' param, handles special pragmas, etc).. it also injects code based on an editor-generated json file ;)
01:55:56	renesac	hum, right
01:56:40	filwit	'part' stands for 'application part' because most of "what" a part actually is is defined in the editor
01:58:02	filwit	this is much of the reason all this is considered non-important to me actually.. see, I really like C#/Java OOP idioms, at least in defining things
01:58:04	Varriount	filwit: Is this also related to your game engine?
01:58:43	renesac	seems good (I already saw your prototype image), but I'm afraid of your sense of style
01:58:47	renesac	^^'
01:58:47	filwit	Varriount: yes
01:59:38	Varriount	renesac: It occurs to me that a lot of objections programmers make against trying new languages are very similar to the objections that 6 year olds make against trying new food.
02:00:16	filwit	Varriount: only because there's a lot of time it takes to switch languages
02:00:34	filwit	Varriount: which removes some of the wit from your jest, good sir
02:01:16	renesac	well, one of the reasons I decided to try nimrod over rust was the python like syntax
02:01:45	renesac	and because it seemed a little more stable
02:01:55	filwit	renesac: yes, and much like your attraction to style, others are attracted to C-like style and OOP. Now lets compare markets...
02:01:55	Varriount	filwit: But it does not remove the fact that programmers will still completely reject things out of hand.
02:02:10	Varriount	filwit: Also http://www.marco.org/2012/06/29/php-addiction
02:02:29	filwit	renesac: the point is, doing things for demonstrable benefit, even if it's only nit-picking, is the _direction_ we should take
02:03:15	filwit	Varriount: this is why it's important to have attractive features to "the most" people if your goal is to gain in popularity
02:03:22	renesac	I don't really think style is the main thing holding back nimrod
02:03:39	filwit	renesac: oh, deer-god, neither do I
02:03:44	filwit	dear** god
02:03:53	filwit	not.. a deer god.. you know what i mean..
02:03:53	renesac	I also programmed in D for a time, and still program in C, even if I don't like this syntax
02:03:58	renesac	lol
02:04:27	renesac	and people jumped in python, ruby, etc, even w/o the braces
02:05:20	filwit	i don't really mind syntax oddities that much, but when things are inconsistent, it makes it harder to _learn_ and that's what I care about (because a) i need scripts easy-to-learn-as-possible, and b) adoption potentially means funding)
02:05:51	renesac	nimrod syntax is different from it's main contenders (D, rust, c++11, etc)
02:06:03	renesac	yeah, I agree about not liking inconsistencies
02:07:24	renesac	and wanting a simple and small language, that is still succinct and powerful, but that is a though balance
02:10:51	fowl	filwit, hows your game engine thing coming
02:10:56	Demos	I found nimrod quite easy to get into, I dont know it as well as I know c++ but still. And I had no experience with anything that had a non-c-ish syntax
02:11:18	filwit	i deleted my forum post, but i posted a git of it as a backup: https://gist.github.com/PhilipWitte/9317301
02:11:31	filwit	in case anyone still wants to leave a comment, do so there
02:12:25	filwit	fowl: it's coming along, i still have some major piece to finish. I want to release code when it can rendering imported models (at least) from Blender and the editor can-drop things onto a Stage
02:13:34	filwit	fowl: i'm still working on the Blender exported and much of the rendering pipeline, plus i would also like it to animate from Blender, and I haven't even started on that yet (though it shouldn't take more than a couple of solid moths).
02:14:35	filwit	fowl: for the last little while i've been setting up how the editor talks to the scripts macro (including trying to fix bugs in my way), and now things are going pretty smoothly there
02:15:43	filwit	fowl: i'm thinking at first I may try and make a GUI system or Art program as a way to build-up on of the core features (since the GUI will be built just like game-object prefabs)
02:15:45	fowl	what do you use for importing assets
02:15:54	filwit	json
02:15:56	filwit	atm
02:16:02	filwit	will be binary eventually
02:16:02	fowl	i mean importing models
02:16:10	filwit	my own file format
02:16:17	filwit	still under dev
02:16:23	fowl	why your own
02:16:27	filwit	it's just a python blender script
02:16:39	filwit	cause the other's have bugs and don't support everything
02:16:41	fowl	use assimp it can handle most formets
02:17:43	filwit	FBX isn't supported perfectly on BLender due to licensing issues (they write their own), though i hear this is getting better. Collada doesn't support some basic things either, so just to be safe I'm making my own so i can grab any data i need.
02:17:54	filwit	fowl: assimp?
02:17:58	filwit	let me look into it
02:19:12	filwit	Last time i checked, Collada blender exports didn't support multiple UV laters, and FBX-SDK is iffy with open-source software
02:19:24	filwit	but i only really heard that about FBX and didn't look into it
02:19:31	*	xenagi joined #nimrod
02:19:35	filwit	still, writing and model exporter really isn't hard
02:19:57	Demos	I would not use Collada, I have heard horror stories
02:20:11	filwit	UV layers**
02:20:36	fowl	filwit, http://assimp.sourceforge.net/main_features_formats.html
02:21:10	filwit	Oh AC3D, how i miss you..
02:21:42	filwit	fowl: so does assimp convert .blend files to .dae files correctly?
02:22:10	filwit	fowl: cause it's possible i could just use openCollada, but I don't really trust that based on past experience
02:22:12	fowl	i dont know
02:22:40	fowl	its well-tested though, been around forever, and i have a wrapper for it :p
02:23:05	filwit	fowl: okay i'll look into it more then
02:29:55	filwit	fowl: this looks promissing.. dunno how i've never heard of this before
02:30:03	filwit	promising*
02:31:12	filwit	fowl: looks like it support opening up .blend files directly and reading their bone-data, and supports multiple UVs and stuff.. nice. thanks for the tip. this could save me some time
02:31:33	filwit	fowl: especially if you already have a wrapper for it :D
02:31:37	fowl	for sure
02:32:22	filwit	how completely would you say your wrapper is?
02:32:52	filwit	(i'm about to try it, just asking
02:34:23	fowl	i thought it was complete, looking at it now it only has 5 functions.. i'll check it again
02:34:37	fowl	i have an example usage in fowltek/musings/nim3d
02:34:43	filwit	k
02:35:37	Varriount	filwit: I do agree that colons for type blocks and other block-like constructs should be acceptable (possibly optional)
02:39:10	*	amitp joined #nimrod
02:40:05	Varriount	filwit: How do you plan for your engine to cover 2d stuff?
02:40:14	filwit	Varriount: i think you're mostly right about programmers not wanting to learn new languages, but this is exactly why i think these inconsistencies are important for adoption. The easier things for the most people to learn, but less friction your going to have in building a fanbase
02:40:16	renesac	Varriount, why?
02:40:21	fowl	Varriount, 2d is peasant dimension
02:40:54	Varriount	renesac: fits in with other block statements (block, if, when, etc)
02:41:11	renesac	I already explained that things that have colons when introducing blocks, also have colons when typed in a single line
02:41:12	filwit	Varriount: with 2D transform constructs rendering in an orthographic camera
02:41:20	renesac	type, var, etc, don't follow this rule
02:41:59	fowl	we need a lispy syntax anyways
02:42:25	fowl	(let (s "hello")) (echo s)
02:42:26	renesac	of course, the ':' is a bit redundant when you have blocks defined by identation after them
02:42:28	filwit	renesac: so possibly the best solution is to allow all ':' be optional
02:42:35	Varriount	Here's a crazy idea that might make for a neat macro - use type declaration syntax to create proc types
02:42:56	renesac	I'm not sure if it is the best, as a bit of redundancy sometimes help
02:43:10	renesac	that is why python kept the ':' after 'if'
02:43:12	renesac	and etc
02:43:21	filwit	Varriount: that's kinda my attempt with the unify : thing..
02:43:25	renesac	there is a guido blog post about that
02:44:05	filwit	renesac: i personally think my above code (type T object: BODY) is the best so far
02:45:26	filwit	renesac: but here you have to enforce the use of ':', mainly because of procs
02:46:45	Varriount	https://gist.github.com/Varriount/9317670
02:46:46	renesac	you remove the '=' and put a ':'?
02:47:22	filwit	Varriount: what's even the point of the 'type' if everything's gunna be under it accept global vars :P
02:48:11	filwit	renesac: it's exactly like my earlier proposed solution only it removes the '=' alltogether (was brought up by Araq)
02:48:19	Varriount	filwit: I dunno. But it's not like that syntax is all that different from the alternative
02:49:03	filwit	oh wait, dur guys.. this is way the best syntax.. you don't even need {.pragma.} brackets anymore
02:49:07	filwit	just...
02:49:11	filwit	type Foo
02:49:14	filwit	final
02:49:17	filwit	object:
02:49:21	filwit	a, b: int
02:49:32	filwit	or..
02:49:36	renesac	too many lines and indentation
02:49:39	filwit	type Foo final object: ..
02:49:57	filwit	renesac: all the returns are optional
02:50:14	filwit	renesac: just goes till it hits the _required_ :
02:50:15	renesac	there you also added a block w/o :
02:50:32	renesac	after 'Foo'
02:50:52	Varriount	filwit: Updated the gist.
02:51:18	Varriount	Maybe we should just have a community Gist to put random syntactical musings on.
02:51:23	filwit	renesac: not a block.. just all whitespace is ignored before the : and only symbols are considered
02:51:38	filwit	Varriount, on sec
02:51:45	renesac	blocks w/o ':' are already allowed when it don't cause ambiguity...
02:52:03	renesac	and that is what you are trying to remove with that syntax..
02:52:21	renesac	*when it don't causes ambiguity in one line
02:52:38	Varriount	*doesn't cause ambiguity
02:52:44	Varriount	:3
02:52:48	*	ddl_smurf quit (Quit: ddl_smurf)
02:52:50	renesac	^^"
02:53:27	renesac	I often forget to use " doesn't " , sorry
02:53:46	Varriount	renesac: I'm just nitpickin' for fun
02:53:57	filwit	renesac: no i'm trying to make thing consistent and simpler ultimately. If forcing the : mean we can remove the {. .} syntax altogether, i thinks it's a valid exchange in the name of consistency, especially when it also unifies the type/proc/if/else/custom-macro/ets
02:55:05	*	amitp left #nimrod (#nimrod)
02:55:18	Varriount	filwit: One thing you need to keep in mind is how much stress the compiler and it's associated parts would have to pick up, handling these syntactic changes.
02:56:05	filwit	Varriount: i was never considering doing this in the short term
02:56:40	filwit	Varriount: the syntax works as-is, and there's more important things to fix for adoption than syntax.
02:56:56	filwit	Varriount: i'm just trying to influence the future mate :)
02:57:04	Skrylar	does anyone happen to have an opinion on TDM-GCC or MinGW
02:57:09	Varriount	filwit: I know.
02:57:20	Varriount	Skrylar: Aren't they the same?
02:57:23	renesac	'type Foo final, noinit ref object of something:' is difficult to parse visually, IMHO
02:57:26	renesac	w/o separators
02:57:26	filwit	Skrylar: what about them?
02:57:35	Skrylar	Varriount: TDM-GCC used to actually support win64 where mingw didn't
02:58:02	Varriount	Skrylar: http://mingw-w64.sourceforge.net/
02:58:04	renesac	and araq don't like this go style type declarations
02:58:53	Varriount	renesac: I can't say I blame Araq, Go type declarations make my eyes water. I'm sure I'd get used to it if I used Go often however...
02:58:59	filwit	renesac: not when there's color coding, which is easy to setup to distinguish when there's consistent rules
02:59:18	fowl	lol
02:59:40	renesac	and pragmas also use ':'
02:59:47	filwit	renesac: you can highlight the first word after 'proc' as bright red (like github) and then each word until : like and attribute
03:00:03	fowl	unsigned int x = 0; x -= 1; // x is now 4294967295
03:00:04	fowl	c is silly
03:00:29	filwit	integer math is sill
03:00:31	filwit	silly
03:00:39	renesac	fowl, it's how your hardware works
03:00:57	Varriount	filwit: If there is one thing I think detracts from nimrod, it's the unsigned int stuff.
03:00:58	filwit	fowl: nevermind i thought you where saying something else
03:01:40	filwit	Varriount: there's good argument for that though, so that should stay
03:02:00	Varriount	I know that Araq doesn't like it (and personally, I don't see the need for it unless interfacing with C stuff) but I've seen many people stumble over it.
03:02:12	Skrylar	yay integer math
03:02:17	Varriount	filwit: I'm not saying it should be removed, quite the opposite.
03:02:20	Skrylar	i was learning how to do fixed-precision math earlier this week
03:03:27	Skrylar	filwit: i have not actually seen very many good arguments for why unsigned is bad; when i asked earlier, i was given references to how C++ handled it poorly but not how unsigned is the devil
03:04:00	Skrylar	I think Araq's mentioning subtraction with length values was one of the only concrete reasons, because blah.len()-5 could cause integer wrapping... but using a signed type doesn't actually eliminate the threat of wrapping, it just delays it to a further edge case
03:04:33	fowl	ah the unsigned stuff happens in nimrod too
03:04:43	renesac	well, it is said to be used unnecessarily everywhere when people want positive stuff
03:04:55	Varriount	Skrylar: Mainly because it's used by many for the sake of representing positive/natural numbers, which causes people to miss the fact that it wraps
03:05:11	filwit	Skrylar: i've heard it can have a negative impact on performance if people are constantly casting from uint->int, but i'm not entirely sure though. Probably shouldn't say 'demostrable' without knowing myself.. but araq always seemed to make good arguments about it. Just can't remember them.
03:05:13	Skrylar	Varriount: that just sounds like an education problem to me :/
03:05:32	fowl	Skrylar, nimrod's `..` iterator guards against that though, iterating from 0 .. -1 will not run
03:06:01	Skrylar	casting between floats and ints is bad for performance too, but you don't see people banning floats
03:06:08	renesac	Skrylar, I'm not sure if there exists such an edge case, as MAX_INT < abs(MIN_INT)
03:06:17	*	almo140 joined #nimrod
03:06:19	renesac	so a single subtraction will still be negative
03:06:41	Skrylar	what does a 'negative length' actually mean?
03:06:52	renesac	an error
03:06:52	Skrylar	i usually tended to use unsigned when a negative value makes no actual sense
03:07:17	Skrylar	especially if a language has try/catch/raise, then i doubly don't use magic return numbers
03:07:19	filwit	Skrylar: you don't normally cast from float to int, unless you really have too. but might use uint more? possibly..
03:07:30	Demos	compilers are allowed to assume that signed arithmatic does not overflow, not so with unsigned. You usually do not want mod2 arithmatic
03:07:45	filwit	Skrylar: yes, i have the same habit with unit in other languages
03:07:46	renesac	Skrylar, that is exactly what Araq is against
03:07:55	fowl	well.. if you have the code "for x in 0 .. len(something)-5" and len returned unsigned and it wrapped you would have a bad time, since len returns a signed int the index can go negative and keep the loop from running
03:08:13	renesac	uint should be used when you are not really doing normal arithmetic with the numbers
03:08:16	Skrylar	renesac: using exceptions or magic return codes?
03:08:24	renesac	no
03:08:30	renesac	[00:06:52] <Skrylar> i usually tended to use unsigned when a negative value makes no actual sense
03:08:31	Demos	uint is for when you want bits, not numbers
03:08:32	filwit	Skrylar: but Nimrod does have uint, I wouldn't even mind having all uint stuff in 'unsigned' module either really
03:08:40	fowl	^ Demos
03:08:58	Skrylar	i still use it that way, because it makes more sense to me :\
03:09:21	renesac	well, nimrod has 'Natural' and 'Positive' types
03:09:34	renesac	just int ranges
03:09:36	Skrylar	at that point you should just rename the type to bits
03:09:37	Skrylar	:P
03:09:39	Demos	you never need the "extra numbers", if the biggest signed type is too small you probably need a Arbitrary-precision lib anyway
03:09:57	renesac	Skrylar, nimrod has a type called 'byte' as alias to 'uint8'
03:10:05	Skrylar	Demos: disagree; netcode uses uints, as do hardware interfaces and the like
03:10:17	Demos	well that is bytes right?
03:10:37	Demos	or if you might have like 1s complement numbers or something
03:10:38	Skrylar	Demos: "nobody needs the extra numbers" is a statement born out of the "i have 12gb ram so what is wasting a few bytes" mentality
03:10:49	Skrylar	stuff like MIDI channels *do* want the extra numbers
03:10:50	filwit	bbl folks
03:11:11	renesac	Demos, the 'extra bit' is very valuable when you have 8 bits per value
03:11:52	Demos	well this is an exception, and they exist. But defaulting to a signed type is a good idea
03:12:04	renesac	and even with 32 bits, it is still a useful difference between 2GB and 4GB w/o doubling the space you are using to store the number
03:12:11	Skrylar	i agree and accept that there are problems from implicit and careless conversion between signed/unsigned, but i don't agree that unsigned is 'evil'
03:12:20	Demos	and if you know you are storing gazillions of these things than go banannas with uint64s or whatever
03:12:26	Skrylar	but as i said a week ago, implicit conversion issues are not unique to uints either
03:12:45	Skrylar	see TCL which calls it the "shimmering" problem where strings/ints get typecast back and forth
03:13:22	renesac	a typecast from int to uint should have no performance impact
03:13:34	fowl	Skrylar, but tcl is gross and everything is a string
03:13:45	renesac	it should be only a thing to keep the compiler type system happy and blissfully ignorant
03:14:00	Skrylar	renesac: there was a link posted here in the log (i don't have it on hand; just did a reinstall) that went in to detail about how C++'s idea of implicit conversion from uint to int causes hidden issues
03:14:37	fowl	its in the wiki
03:14:49	renesac	well, implicit conversion is bad in many cases
03:14:52	Skrylar	though i'm guessing the int/uint problem is similar to the gc/nogc problem; in that there are people who absolutely need real-time code and they will always disagree with GCs
03:15:00	renesac	but is safe in a few ones
03:15:36	renesac	I wanted nimrod to support those few ones
03:15:42	renesac	like uint8 to int
03:15:53	Skrylar	you can jank it by writing your own converter
03:16:09	renesac	I may try that
03:16:16	Skrylar	and as a bonus the converter is scoped, so you can import it selectively without exposing it to external code
03:16:26	renesac	I tried to fix it in sigmach.nim
03:16:33	Skrylar	I use this in my UTF code a few times, because a codepoint is a uint32
03:16:55	Skrylar	(NB: a utf codepoint *is* actually specced as an unsigned 32-bit value; there is no 'negative codepoint' in 6.2)
03:16:58	renesac	but then there was ambiguity between a proc(x:int64) and proc(x:uint64) when given a uint8
03:17:38	Skrylar	all though i use a distinct uint32, which allows me to shield a codepoint from regular math so people don't do stupid things unintentionally
03:18:04	Skrylar	"a + b" means nothing to a unicode codepoint afterall :)
03:18:06	renesac	I have to try if using a converter has the same problem
03:18:23	renesac	Skrylar, yeah, I'm also using a distinct uint32 for my bignum module
03:18:45	renesac	for many reasons
03:18:54	Skrylar	i like distinct types as a concept
03:19:07	renesac	if I get intrinsics working, I may upgrade it to uint64
03:19:08	Skrylar	they work especially well when you're dealing with opaque pointers
03:19:22	renesac	and no change in the code that depends on it
03:19:38	Skrylar	i do however wish that two distinct pointers didn't cause generic code to get expanded twice
03:19:56	Skrylar	e.g. a generic linked list will compile a separate implementation for each distinct pointer even if the C code for both is identical
03:20:04	renesac	hum
03:20:30	renesac	yeah, this is a problem with the current compiler
03:20:46	renesac	hopefully it will be optimized in a not too far future
03:20:56	renesac	at least it isn't a inherent language problem
03:21:57	renesac	and for hashing, rngs, etc you also want unsigned ints
03:22:13	renesac	and with compression, you are dealing with bytes, not numbers, etc
03:22:13	Skrylar	I don't think they're going away
03:22:17	renesac	yeah
03:22:19	Skrylar	they're just shoved in a separate module
03:23:12	renesac	not just that, but the compiler still has poor support for them
03:23:32	renesac	https://github.com/Araq/Nimrod/issues/936
03:24:25	Skrylar	renesac: Araq mentioned a solution for problems like that in the IRC a while back
03:24:35	renesac	(yeah, I know how to work around of all them)
03:24:38	Skrylar	it had to do with making a converter for the lit(T) type IIRC
03:24:56	renesac	yeah, I did it in my bigint module
03:25:01	Skrylar	i think mostly, the unsigned module just needs to get polished up
03:25:42	Skrylar	time to wait for gcc to install \o/
03:25:50	renesac	and error messages should point begginers to import the unsigned module
03:27:48	Skrylar	probably some spooky message that says to only do that if they really need unsigned types
03:32:47	*	xenagi quit (Quit: Leaving)
03:36:50	Varriount	renesac: Intrinsics?
03:38:07	Varriount	Skrylar: "Warning: Unecessary use of the unsigned module will cause you to be eaten by a grue."
03:41:11	Demos	Varriount, that "quote" reminds me of http://eigen.pri.ee/shooter/
03:43:33	Varriount	Demos: Do you think it's sppoky enough?
03:43:42	Demos	yes
03:46:18	Varriount	Demos: I hate being assigned a 2 1/2 page paper on the weekend, when it's due on Monday
03:46:45	Demos	yeah last semester I had programming assignments due on sunday
03:47:01	renesac	Varriount, I would need 128bit multiplication
03:47:55	renesac	https://gist.github.com/PhilipWitte/9316566 <-- this should be useful as a start point
03:47:58	renesac	for me
04:05:04	*	BitPuffi1 quit (Quit: WeeChat 0.4.3)
04:10:37	Varriount	renesac: The problem with 128 bit ints is that out of all the compilers I know, only GCC running on a 64 bit platform has them.
04:12:24	Varriount	Or rather, natively supports them.
04:15:57	renesac	yeah... I will see that and other issues
04:19:23	*	shodan45 quit (Quit: Konversation terminated!)
04:40:44	*	brson joined #nimrod
05:00:51	*	rob3 joined #nimrod
05:02:42	*	r0b2 quit (Ping timeout: 252 seconds)
05:02:45	Varriount	Damn. I got a 70% on my latest Calculus test.
05:03:25	Skrylar	is that a good or bad 'damn'
05:03:44	Skrylar	(sidenote: i kind of like how foobar2000 does its layout editing. more programs should work this way.)
05:05:56	Varriount	Skrylar: A bad one. I studied hard on that test.
05:06:44	Varriount	And what's what's worse is that now I'm going to be haunted by that grade for the rest of the day/night. :/
05:07:25	*	brson quit (Quit: leaving)
05:15:48	Skrylar	I studied math earlier and then Anki told me to stop
05:23:20	*	brson joined #nimrod
05:23:39	Varriount	Skrylar: Anki?
05:25:11	Skrylar	Varriount: the flash card program?
05:26:39	Skrylar	it does this thing where if you go over a limit of cards per day in a subject it tells you to stop, because there's research that shows studying too hard is unhelpful
05:26:51	Skrylar	you can bypass it but meh
05:30:21	*	BitPuffin joined #nimrod
05:44:59	fowl	Varriount, where are you located
05:49:04	Varriount	fowl: In the US
05:49:26	Varriount	fowl: Do you want me to be more specific?
05:54:31	Skrylar	ugh
05:54:51	Skrylar	I hate when people bug me about helping them set up cash shop crap for web games
05:55:18	Varriount	Cash shop?
05:55:33	Skrylar	"only 0.99$ for cosmetic things on your avatar!"
05:55:43	Varriount	Oh. Eeew
05:56:21	Skrylar	this one guy I know *knows* that I hate cash shops in games, and yet he keeps bothering me about running a server that subsists on selling 1$ cosmetic crap because "oh well people might buy stuff even if they don't stay very long"
05:56:37	Varriount	"We're programmers, not robots. Amazingly enough, we *do* have standards"
05:56:59	Skrylar	I'm one of those dinosoars that would rather publish a cheap monthly thing than 'free' to play crap
05:59:51	Skrylar	I disagree with Joel that SaaS is better; there's usually something you can do to improve a program even when you have everything and the kitchen sink; sometimes that improvement is just moving things to make them more usable or maybe waste less memory
06:00:10	Varriount	Joel?
06:00:14	Varriount	Skrylar: Well to be fair, most successful MMORPG's are free to play, and require a monthly subscription for premium play
06:00:17	Skrylar	the "joel on software" blog
06:00:34	Skrylar	ex-microsoft guy
06:00:44	fowl	Varriount, just wondering because its about midnight in the midwest
06:00:46	Skrylar	Varriount: depends on your definition of 'success'
06:01:08	Skrylar	I forget who did MapleStory, NXsomething
06:01:13	Varriount	fowl: Yep. I live in the midwest. I'm burning the midnight oil, getting a stupid essay done.
06:01:26	Varriount	Er, mideast actually
06:01:37	Varriount	VA, to be precise.
06:01:42	Skrylar	IMO 'free to play' is basically newb-lube for bad game designers; can't make a good game? lower the standards!
06:01:44	fowl	mideast? there is no mideast, only east, midwest and west
06:02:13	fowl	Skrylar, the best model is everybody can play free but paying gives you VIP status
06:02:30	Skrylar	you mean like FunCom sort of does it now?
06:03:16	Skrylar	I don't mind the FunCom method for anarchy online personally; to a lesser extent I don't mind the way they do DLC/Subs in DCUO, all though in all cases I abhor stupid things like charging real money for respecs
06:03:25	Varriount	Like RuneScape and Dofus? (The only two online rpg games I've played)
06:03:37	Skrylar	Runescapes wasn't bad
06:03:52	Skrylar	the worst is when you have 'f'2p, subscription, AND still cash shop the VIPs
06:05:31	fowl	vip status means when you want to join a full server some free-player gets booted
06:05:53	Skrylar	well DC Univers Online does this thing where they release new content as DLC you can buy, or if you have a sub you just get to go there
06:06:00	fowl	oh and they get 50% off the cash items
06:06:01	fowl	:p
06:06:18	Skrylar	all though they also do that stupid lockbox crap
06:06:27	fowl	whats that
06:06:39	Skrylar	you know the MANNco crates from valve?
06:06:54	Varriount	Skrylar, fowl: After Dofus, I realized that RPG games take up too much of my time, time better spent reading and programming.
06:07:00	Skrylar	its basically a ripoff where you get loot in the form of a cash shop box that says "sorry you don't actually get loot, pay real world money to unlock this and get something from the drop table"
06:07:16	Skrylar	more often than not its craploot too
06:07:45	Skrylar	Varriount: MMOs yeah. I like some offline RPGs still
06:08:11	Skrylar	MMOs are one of those weird genres where you can innovate nothing and still get game of the year every year o_O
06:08:38	Skrylar	"and this year, we made 50$ DLC which adds two dungeons and pandas!"
06:09:13	Skrylar	Sadly you can't make a sandbox MMO or else the Something Awful forums and /b/tards come in and ruin it
06:10:15	fowl	the pool has aids
06:12:37	Skrylar	lol habbo hotel
06:12:54	Skrylar	naaah that was funny at least :)
06:13:07	Skrylar	the first time anyway
06:14:46	Varriount	Skrylar: I like RPG's that have a good story. Then again, almost the only games I do like are ones that have good stories.
06:15:06	Varriount	Probably comes of reading too many books.
06:16:28	*	isenmann joined #nimrod
06:16:47	Skrylar	i always thought rpgs were sad to develop because you can't enjoy them as the developer
06:16:55	Skrylar	you already know everything about the story
06:18:41	Varriount	Hm. Well, you could probably enjoy them in the same way an author enjoys people reading a book he/she has written.
06:22:24	Skrylar	tonight will be the first time i actually use nimrod on windows
06:22:30	Skrylar	i had just been using putty to a linux box this whole time
06:22:44	Skrylar	also I'm curious
06:23:04	Skrylar	Why do the instructions say to checkout csources? Why is that not just a git-submodule
06:23:46	Skrylar	oh. the instructions now say to use --depth 1. maybe thats why :derp:
06:24:47	Varriount	Skrylar: Technically, you don't have to use csources, except if running on a previously unbooted platform
06:25:05	Skrylar	Varriount: technically, i need to prod araq with toothpics
06:25:20	Skrylar	"unknown operating system: mingw32_nt-6.1"
06:25:45	Varriount	Skrylar: That's odd, I use Mingw, and it works for me.
06:25:52	Varriount	(Yes, I'm on windows)
06:32:03	Skrylar	Oh good. Trying out this open source notekeeping thingy and it locked up.
06:32:23	Skrylar	The best thing you want out of programs you trust to store copious notes is that they lock up in the first ten minutes of use *nodnods*
06:32:43	Varriount	Skrylar: What is the software?
06:32:56	Skrylar	mindraider
06:33:17	Skrylar	i was giving it another try because semantic web software sounds like a cool idea
06:34:09	fowl	Skrylar, git submodules are difficult to maintain
06:34:10	Varriount	Looks like more trouble than it's worth for everyday stuff. Might be useful for projects though.
06:34:34	Skrylar	Varriount: more trouble than its worth period. lol
06:34:55	Skrylar	apparently i managed to create a corrupt node that locks up the program, and there's no "delete" option
06:35:02	*	Demos quit (Read error: Connection reset by peer)
06:36:10	Skrylar	Varriount: personalbrain had a neat idea where its basically a concept map stuffed in a database, so you can click'n'drag to make a new topic and it shows how they are related
06:36:22	*	almo140 left #nimrod (#nimrod)
06:36:38	Skrylar	which is kind of a neat idea, since you could just click through the web from one thing to the next... but 200$ for a subscription license is "no"
06:38:54	Varriount	Skrylar: I'll never understand why software licenses are so high (for individual users at least)
06:39:23	Varriount	I can see 200$ being a reasonable price for say, a business, possibly.
06:39:30	Skrylar	Varriount: not very many people tend to buy information management tools, and they tend to require a lot of special stuff
06:39:49	Skrylar	for instance the outliner i want to make requires a lot of custom code :/
06:39:57	Skrylar	unless i janked it on to libcef i guess
06:40:30	Varriount	Skrylar: Heh, I've tried creating a nimrod wrapper for libcef. Don't know if it works though
06:41:00	Varriount	One of these days I should have another go at it.
06:44:44	Skrylar	i've lost a bit of steam for GUI stuff as of late; i really just want to start doing work that shows a difference
06:44:54	Skrylar	writing infrastructure code for six months gets absurdly tedious
06:45:50	Varriount	Skrylar: GUI stuff doesn't show a difference?
06:46:00	Skrylar	Varriount: haven't gotten to do any of it in a while
06:46:16	Skrylar	the ncurses stuff I worked on blocked on the gap buffer code, which blocked on the skutf git
06:46:32	Varriount	Skrylar: SDL2?
06:46:48	Skrylar	you still need font layout code and unicode-safe text buffers
06:46:56	Skrylar	SDL just gives you a place to render on
06:48:22	Skrylar	IUP or Wx are decent targets for wrapping, though those don't really solve many of my troubles lol
06:48:39	Varriount	Skrylar: If it makes you feel any better, your skutf stuff will probably help with a nimrod implementaion of diff-match-patch
06:48:56	Skrylar	i need to poke araq about merging that
06:56:59	*	nequitans_ joined #nimrod
06:57:26	nequitans_	Hi all, I wrote a 'consider nimrod' blog post: http://geetduggal.wordpress.com/2014/03/03/consider-nimrod/ I'd appreciate if you notice any inaccuracies/corrections
06:58:36	*	Skrylar puts on glasses
07:00:50	Skrylar	I'm not sure Julia was really ever intended for low-level coding
07:02:15	Skrylar	i am horrified that you have used string concatenation in a call to shell calls (but you did say it was just a shell script replacement)
07:03:12	Skrylar	nequitans: might want to be more specific about how the iterators work
07:03:26	Skrylar	" Nimrod uses a uniform call syntax that implies these two calls are equivalent:"
07:03:44	Skrylar	which then shows the iterators, which have a different syntax (maybe i'm being nitpicky)
07:07:38	nequitans_	Skrylar, so maybe I shouldn't use iterators as an example of the uniform call syntax?
07:08:51	Varriount	nequitans_: Spelling error, "after a nearly three weeks"
07:09:36	nequitans_	Varriount, thankx
07:09:44	fowl	nequitans, in the TGroup example you do not need to importc each field (unless they have a different name than the nimrod name)
07:10:05	nequitans_	(I'm making corrections as you give them)
07:10:16	Varriount	nequitans_: No problem. I'm still looking.
07:10:55	fowl	nequitans, matrixToGroup(openarray) seems like a weird example for importc
07:11:05	nequitans_	fowl, thanks -- changed
07:11:55	Skrylar	nequitans_: i wouldn't use iterators as a demonstration of uniform call syntax because while the code is correct, iterators aren't really the same thing as calls
07:12:03	nequitans_	fowl, lol, yea, actually i'm finding a maximum density subgraph of a matrix using network flows iteratively,
07:12:06	Skrylar	the syntax to invoke them is the same, but they're kind of a fancy inline
07:12:51	fowl	nequitans, i mean because it has openarray
07:13:03	Skrylar	Varriount: got it working using the bat file; odd that nimrod doesn't work inside msys though
07:13:25	Varriount	Skrylar: You mean the msys shell?
07:13:45	Skrylar	Varriount: yep, i couldn't run the bootstrapped nimrod because it crashed with a 'cant create process' error
07:13:58	Skrylar	wait
07:14:14	Skrylar	how did the mingw installer install msys without actually setting the directories for anything oO
07:14:30	Varriount	Skrylar: Be sure to post bug reports.
07:14:44	Skrylar	Varriount: that was stupidity in the msys installer
07:16:24	Varriount	Skrylar: What I did to install msys with mingw64 was to first install mingw64, install msys in a separate directory, copy msys over to the mingw dir, and then adjust the PATH variable
07:16:35	nequitans_	fowl, what's the issue with the openarray?
07:16:42	Skrylar	i just installed the 32-bit version, Varriount
07:16:50	Skrylar	I don't really need 64-bit builds for the time being
07:17:18	Varriount	Skrylar: Well, if you need help setting up a dual build, I'm available
07:17:27	Skrylar	alright
07:17:29	fowl	nequitans, because of the way it expands to (T*, int)
07:17:54	Skrylar	pet peeve: koch mentions gnu readline
07:18:04	Skrylar	someone ought to add libeditline :)
07:18:09	Skrylar	the BSD-equivalent
07:18:13	Varriount	nequitans: Calls are actually pass by reference or pass by copy (it's determined by the nimrod compiler)
07:18:32	fowl	nequitans, that can be useful when a c library uses that pattern to pass arrays around, but its not explained in your article (nor should it be) i just think its confusng. i would use printf or something like that as an example
07:18:47	Varriount	You force one or the other by using a ref type, a pass by reference pragma, or a pass by copy pragma
07:19:06	Varriount	At least, that's what I've been told.
07:19:15	nequitans_	Varriount, I see -- so how does the compiler determine which one to use then?
07:19:25	Varriount	nequitans: No idea.
07:19:36	nequitans_	kk
07:19:44	Varriount	nequitans: Again, I might be wrong.
07:19:47	fowl	nequitans, the rule is something like if this type is bigger than 8 bytes pass by reference
07:20:33	nequitans_	but in my example the integer passed as a var is modified, does that mean it's more than 8 bytes?
07:21:36	Varriount	nequitans_: "offers the programmer type classes" needs to have a ' - ' after it, or possibly a comma
07:21:45	fowl	nequitans, by using 'var int' you enforce pass-by-reference
07:21:49	nequitans_	i think i might get rid of the link to C example
07:21:55	nequitans_	fowl, I see
07:22:59	fowl	since you write about 'traits' so much you should throw in some stuff about typeclasses
07:23:14	fowl	oh nm i see that part now
07:24:06	nequitans_	fowl, ah, yea, i'm curious as to your thoughts on that -- this relationship to traits has just been brewing in my mind the last few days and I'm not sure I'm quite solid on it, but i def have experienced them causing library design difficulties in a language like scala
07:25:27	Varriount	nequitans_: You mention Rob Patro showing you nimrod - is there a specific link to that?
07:26:24	nequitans_	Varriount, he and I just had in person discussions
07:26:40	Varriount	Varriount: Ah.
07:26:52	Varriount	*nequitans_
07:27:06	Varriount	Sorry, it's 2:30 in the morning.
07:27:42	nequitans_	lol
07:27:45	nequitans_	you're in my time zone :)
07:27:50	nequitans_	(i'm in pittsburgh)
07:28:58	Varriount	Virginia.
07:28:58	nequitans_	fowl, so for the openarray example (i removed it from the current draft of the post), is there a better more idiomatic way to do that link to C?
07:30:10	fowl	nequitans, actually that is a pretty awesome idea for interfacing with C, i just dont think its a great example
07:32:02	nequitans_	i see -- yea, i was ambivalent about putting it as an example, but it also took me some time to figure it out through trial and error so i thought i'd post it. but i agree, it's a little confusing and i don't like putting the array[0...1000000] b/c it's more of an idiosyncrasy that would cause confusion
07:32:16	fowl	in my wrappers i leave the function as (T*, int) but honestly that is cumbersome, this would make passing array/seq to these functions easier than calling them with (array[0].addr, array.len)
07:32:18	nequitans_	i ended up removing it
07:33:13	nequitans_	so what do you guys think about my brief comparison with other langs?
07:33:52	Varriount	nequitans_: Pretty good. Could use more examples (typeclasses) but I'm probably just nitpicking.
07:34:11	nequitans_	Varriount, noted.
07:34:30	Varriount	By examples, I mean, not of just nimrod, but of other langs
07:34:40	nequitans_	I see
07:35:41	fowl	jesus.. chromium is dominating my memory
07:38:08	Varriount	fowl: It's a web browser, what do you expect?
07:39:03	fowl	maybe it realizes "hey im using all this guys memory and swap maybe i should back off a bit"
07:39:15	Varriount	fowl: How many tabs do you have open?
07:39:22	fowl	hundreds :(
07:39:35	Varriount	fowl: Install the toomanytabs extension.
07:39:51	Varriount	Either that, or go to a support group for tab abusers
07:40:04	Varriount	"Tabaholics Anonymous"
07:41:05	fowl	ha
07:41:14	nequitans_	i bet fowl has a memory map in his brain for how to get to a specific tab instantly. yet the computer can't handle the tabbage!
07:41:25	fowl	i do
07:41:36	fowl	i keep the tabs grouped nicely and all
07:41:54	nequitans_	i've been using pinned tabs a lot recently
07:42:04	Varriount	Anyone here actually read how swap memory works? It's quite interesting.
07:42:41	Varriount	I've read how it's done in Windows, I don't know how linux does it though.
07:42:45	fowl	Varriount, 6000 years ago god gave us linux, i havent questioned it since
07:43:17	Varriount	fowl: That's the thing. I've found many books about the internals of windows. Not so many about the internals of linux.
07:44:00	fowl	because you can access the code
07:44:09	fowl	and read it yourself :p
07:44:27	Varriount	fowl: Shame on you. You know how much effort it takes to both read *and* understand code.
07:46:13	fowl	i opened chromium again, i have three tabs open. there are 40 chromium processes open now with memory usage ranging from 3.7% to 0.8%
07:46:44	Varriount	fowl: How many extensions do you have installed?
07:48:29	fowl	5
07:49:03	*	brson quit (Ping timeout: 244 seconds)
07:50:21	Varriount	fowl: Maybe they are zombie processes?
07:50:26	nequitans_	thanks for your input on the post! i made the mods and i'll try to add some more examples as time passes. i'm considering it a 'living' document :-P
07:50:46	fowl	Varriount, no i checked after i closed it, they were all closed
07:50:47	*	brson joined #nimrod
07:50:59	fowl	no prob nequitans
07:52:57	Varriount	nequitans_: Yeah, it's great.
07:53:35	Varriount	nequitans_: Don't forget to post it on reddit when you feel that it's ready
07:53:52	Varriount	Oh, an hacker news
07:55:05	Varriount	fowl: I only have a single chrome process open per tab, and per extension.
07:55:32	Varriount	Well, plus the 1-3 others that make up the core stuff.
07:55:40	fowl	you're using chrome or chromium
07:55:47	Varriount	Oh. Chrome.
07:56:16	fowl	oh i have 2 gigs of updates to install. nice
07:56:28	Varriount	fowl: Nom nom
07:57:41	Varriount	fowl: It could be worse, at least it ain't Visual Studio
07:57:58	fowl	tr00
08:01:57	nequitans_	Varriount, thanks -- will do
08:02:02	nequitans_	Speaking of chrome
08:02:12	nequitans_	I've been using this little tiny Chromebook 11 for a while now
08:02:24	nequitans_	as an 'additional computer'
08:04:20	nequitans_	aside from better performance (which you can't ask for too much for a $270 comp), I would really like chrome desktop apps to allow me to adjust the font size. e.g I'm using CIRC right now and i'm in this tiny font
08:06:38	*	zielmicha quit (Quit: Connection closed for inactivity)
08:14:50	Varriount	nequitans_: So is there anything you feel that Nimrod could improve upon, without drastically altering the language?
08:16:36	Varriount	nequitans_: Also, I think I read your "pypy vs shedskin vs ..." a while back, it was pretty good.
08:18:11	nequitans_	Varriount, lol, that was a big what i call 'fungus' -- i was trying to get some really simple code for a class project to run fast in python and then a sleepless night of tests :-P
08:18:34	nequitans_	re: Nimrod improvements
08:19:02	nequitans_	it's not a really big issue
08:19:04	nequitans_	but
08:19:42	nequitans_	if the pass by ref/pass by val issue could be made more clear. i.e. it helps me to know as programmer that when i'm passing in a matrix class it's just really a pointer to it rather than a copy
08:20:24	nequitans_	btw
08:20:44	Varriount	Yeah. That's a documentation issue that should really be made more clear. You don't want to pass a huge struct and end up doubling memory.
08:20:57	fowl	there is a pragma for that
08:21:18	fowl	Varriount, why do people think this happens?
08:21:37	Varriount	fowl: Because C/C++
08:21:53	nequitans_	i also noticed that i had to import modules at the root level that i never actually included, but included in some subordinate module and now it oddly requires to be included in the root module
08:21:56	fowl	nimrod has its own semantics
08:22:00	Varriount	fowl: I think he's talking about the behavior when no pragma/var/ref is forcibly used.
08:22:11	nequitans_	(maybe this is part of the design, but it took be by surprise)
08:22:26	fowl	nequitans, ah that is fixed now
08:22:52	nequitans_	fowl, cheers
08:22:58	fowl	nequitans, that is because when a generic is instantiated your module might not have access to the module that generic needs, this is what the export stmt is for
08:23:56	fowl	so if you have this problem still, complain to the librarymaster to fix it
08:25:34	fowl	Varriount, since parameters are const by default, everything can be pass-by-ref and still be secure that the parameter wont be modified
08:26:16	nequitans_	fowl, ic
08:26:17	*	brson quit (Quit: leaving)
08:26:34	Varriount	fowl: That fact should still be made clear somewhere
08:27:16	Varriount	Also, if that fact is so, why do we have modules with ref versions of every type?
08:27:39	Varriount	See: The original asyncio module
08:27:55	fowl	Varriount, do you know what ref is?
08:28:05	fowl	its a managed pointer (garbage collected)
08:28:08	Varriount	fowl: It's a traced reference.
08:28:14	fowl	"ref" is nothing to do with pass by reference
08:29:19	Varriount	fowl: I thought that a ref was analogous to a pointer in other languages.
08:30:47	fowl	i guess so
08:30:50	fowl	this is #nimrod though
08:30:58	fowl	this. is. #nimrod!
08:31:04	*	fowl kicks Varriount into the pit
08:31:18	fowl	good night
08:32:02	Varriount	Good night. When can I get out of the pit?
08:34:45	Varriount	nequitans_: "When C came about, preprocessors were a good idea. Then again, so was paper tape readers and pet rocks. "
08:37:03	nequitans_	Varriount, lol, i remember this from some internets post a while back, tho i forgot what it was about
08:37:17	nequitans_	something with Jav
08:37:18	nequitans_	*a
08:37:26	Varriount	nequitans: "Why I am not a Java Programmer"
08:37:50	Varriount	I've read the article before, but it always manages to put a smile on my face.
08:38:16	nequitans_	lol, yea, that's it
08:38:37	Varriount	I know its arguments aren't 100% substantial, but... it's such fun to read.
08:39:43	Varriount	I've used Java before and still do, but it's has always felt entirely too restrictive, like a nuerotic, OCD businessman
08:39:54	Varriount	*it has
08:42:02	nequitans_	Yea, I think that in that respect Scala esp. brought a fresher air to the JVM world. When I first started coding in Scala for a while, I wondered: will Java catch up to this or will a language like Scala just take over?
08:43:27	Varriount	nequitans_: I haven't tried Scala yet. I've had the manual/guide on my reading list for a while.
08:44:08	Varriount	I'd like to think that I'm more open to learning new programming languages than most programmers. If nothing else, I read the manuals/tutorials
08:44:27	nequitans_	It's a pretty solid language, but I've def run into efficiency issues. In some cases, it ran on par if not *faster* than my C code. But in other cases (esp with many threads on multicore machines, it was prohibitively slow)
08:45:08	Varriount	nequitans_: Such is the problem with JIT compiling, especially when the one for the JVM was designed for Java specifically.
08:46:30	Varriount	nequitans_: I've always wondered why you can't JIT compile (or rather, optimize) assembly code, eg, C/C++/Any other statically compiled language
08:47:30	Varriount	Is it because assembly is just too low-level to be optimized at run-time?
08:47:56	nequitans_	Varriount, yea, I think it's a good quality of a computer scientist to be able to pick up on a language's strengths and weaknesses and be able to use it effectively. I used to be on the DSL train: a good computer programmer/scientist adapts/develops DSLs for all sorts of useful tasks and can live in that world. Now I think that a single language could possibly allow the user to express the DSLs they want efficiently -- that's w
08:48:12	nequitans_	re: JIT assembly, that's an interesting idea
08:56:18	Varriount	nequitans_: Ah, here's a paper that explores the subject. I don't know if it's the same paper that I once saw, but it's related -> http://groups.csail.mit.edu/commit/papers/03/RIO-adaptive-CGO03.pdf
09:00:39	nequitans_	Varriount, interesting
09:12:49	*	nequitans_ quit (Ping timeout: 240 seconds)
09:20:00	profmakx	srsly
09:20:03	profmakx	do they write those in word?
09:20:21	Varriount	profmakx: What, the papers?
09:20:32	profmakx	yeh
09:20:45	Varriount	Who knows.
09:20:52	*	profmakx was forced to hand in a proposal in word. cost me more than a day getting it into decent shape
09:21:28	Varriount	profmakx, nequitans_: Any idea what "Notes potential avenues for additional research analysis" means in the context of an essay which analyzes another essay?
09:22:59	profmakx	Varriount it means "blah"
09:23:13	profmakx	vacuous phrase. people in research love those for some reason
09:24:26	Varriount	profmakx: Tell that to my composition professor.
09:24:54	profmakx	Varriount what is "composition"?
09:25:43	profmakx	Varriount it probably means that it mentions research directions somewhere. I am surprised that people still get away with these overly vague formulations
09:25:45	Varriount	profmakx: Language, english, writing.
09:26:18	Varriount	profmakx: The class is intended to teach effective skills for essay analysis and writing.
09:28:21	Varriount	profmakx: Here's the rubric for this particular essay, if you're interested -> https://docs.google.com/document/d/1SyvhE4kGKfh9Eo-F_anyPEwVuRVpbBN-NvMeyoQnQck/edit?usp=sharing
09:28:44	Varriount	I've highlighted the things in green that I'm sure my essay has done properly.
09:31:07	profmakx	I probably suck at writing :P
09:31:38	Varriount	profmakx: Of the few skills I have, writing and reading are my best.
09:31:49	Varriount	Also, Google says that you're an anonymous giraffe.
09:33:06	profmakx	I have no idea what that even is
09:33:12	profmakx	oh
09:33:27	profmakx	you mean on the doc?
09:34:32	*	CarpNet joined #nimrod
09:36:01	Varriount	profmakx: Yes
09:36:07	Varriount	Hi cark
09:36:12	Varriount	Er, CarpNet
09:37:15	CarpNet	hello
09:42:56	Skrylar	bonus
09:43:07	Skrylar	utf tests didn't mysteriously fail just because i built them on windows
09:49:03	Varriount	Skrylar: :O
09:49:38	Varriount	Hm..
09:49:51	*	Varriount is now known as MagiKarpNet
09:49:57	MagiKarpNet	:D
09:50:30	*	MagiKarpNet is now known as Varriount
10:01:33	*	io2 joined #nimrod
10:01:40	Varriount	oo, this is interesting -> https://github.com/xnko/libapi
10:07:58	Skrylar	heh. coroutines are pretty neat
10:08:22	profmakx	Skrylar do you think?
10:08:42	Skrylar	profmakx: i was acknowledging Varriount's link :)
10:09:07	Skrylar	I'm not sure why GCs are considered "must have next gen" but first class coroutines which are as old as assembly still get completely ignored
10:09:29	Skrylar	even though there have been benchmarks showing they're way better than how most people derp about IO
10:11:01	*	ponce__ left #nimrod (#nimrod)
10:12:05	Varriount	Skrylar: The only problem is that most implementations I've seen have to guess about how much memory to allocate for the stack
10:13:03	Varriount	I was wondering, when nimrod gets coroutines, if it could at least somewhat accurately guess the stack size (you know, being a compiler and all)
10:14:00	*	odc joined #nimrod
10:14:34	*	odc left #nimrod (#nimrod)
10:20:47	Skrylar	Varriount: probably, since it knows how much space the function needs for scratch
10:20:58	Skrylar	its kind of like how GCs work better when the compiler has some idea where to put the safe points, lol
10:21:27	Varriount	Safe points?
10:21:38	Skrylar	all though i wonder if thats the kind of thing macros should be able to do; seems like they should be able to read the AST for the function and check that kind of detail
10:21:45	Skrylar	though i don't think macros get access to resolved type details :/
10:22:11	*	faassen joined #nimrod
10:22:25	Skrylar	(before i went to rust and then to nimrod, i was working on a compiler that was basically just the Io scripting language with a code emitter out the back; so i'm a big fan of the microkernel+macros-for-everything model)
10:22:39	Varriount	Hi faassen
10:22:52	Varriount	Skrylar: Have you seen the compiler internals?
10:22:59	Skrylar	not for nimrod
10:23:10	Varriount	Skrylar: Most of the compiler is a macro, essentially
10:23:16	Skrylar	i live in enough of a small hell just fixing up this unicrap
10:23:39	Varriount	Skrylar: The methods you would use for a macro are pretty much the same methods the compiler would use.
10:23:55	Varriount	The compiler just has access to a lot more information
10:24:08	Skrylar	yeah that was pretty much the model i was going for
10:24:20	Skrylar	well, i had some smartass things to go along with it that i never got to
10:24:53	Varriount	Skrylar: Although, I doubt anyone will ever create a macro as obtuse as the lambda lifting compiler code.
10:25:09	Skrylar	i was going to have it so you could access a meta object in the object tree, which would let you put lisp-style declaims on things like methods so the slow-interpreter would ignore them but a JIT would be able to use them effectively
10:25:35	Skrylar	things like "don't use full message passing mechanics on this function, it doesn't need it"
10:26:04	Skrylar	(Io can do crazy things like look at who just called a method and then modify the caller object even though you didn't ask for that)
10:28:01	Varriount	Skrylar: Yeah, I've heard interesting stories about Io and the other smalltalk-like languages
10:28:41	Skrylar	i love that language
10:28:53	Skrylar	Io is basically Lisp with the parenthesis shoved over one token
10:29:42	Skrylar	they also have things like a binary message is just anything from this set of characters, so you can use +===---===+ as an operator if you feel pathological
10:29:45	Skrylar	IIRC
10:31:08	Varriount	Skrylar: What time zone are you in?
10:31:13	Skrylar	CST
10:32:15	Varriount	So.. I'm not the only one up in the wee hours
10:32:32	Skrylar	nerp
10:36:10	Skrylar	i must be a weirdo because i hate presets
10:36:21	Varriount	Presets to what?
10:36:26	Skrylar	presets in general
10:36:40	Skrylar	they felt like adding a bunch of preset views to MLO which I guess is good for newbies
10:36:50	Varriount	Skrylar: MLO?
10:36:53	Skrylar	but i almost never actually use the "tasks completed by project" preset
10:37:05	Skrylar	Varriount: http://mylifeorganized.net/
10:37:22	Skrylar	in music stuff people actually SELL presets too
10:38:07	Varriount	Skrylar: I think the aversion to presets is somewhat stereotypical of programmers - I mean, look at linux
10:38:25	Skrylar	I like the concept, I just don't tend to use them
10:38:51	Skrylar	older versions of mlo didn't have any and i think a lot of people were scared of the filter screen
10:39:02	Skrylar	you can ask it to do all sorts of arbitrary filing for you
10:39:16	Varriount	Skrylar: My way of organizing things is to use the sticky notes program built into windows, google keep, and google calender
10:40:05	Skrylar	i seem to end up with crippling amounts of TODOs when i start tracking them with anything
10:40:34	Varriount	Skrylar: Same
10:40:40	Skrylar	what i like with MLO and on the mac i used omnifocus for it
10:40:56	Skrylar	you can put everything in the outline view and have this scary huge list
10:41:05	Skrylar	then go to a filter view that only shows "things that are open RIGHT NOW"
10:41:32	Skrylar	so if i set something to depend on araq being awake, and i tell it when that usually is, it won't show those
10:41:43	Varriount	Hah.
10:42:07	Varriount	You have Araq's sleep schedule?
10:42:12	Skrylar	no
10:42:23	Skrylar	i thought about putting that in if i end up doing too much stdlib work though
10:44:38	*	ddl_smurf joined #nimrod
10:45:31	Varriount	Whenever I see ddl_smurf's username, I think "Dance-Dance Revolution Smurk" (yes, I know it's an L, not an R)
10:45:38	Varriount	*Smurf
10:48:04	ddl_smurf	(sorry)
10:48:18	Varriount	ddl_smurf: Sorry for what?
10:49:04	Skrylar	bleh. string slicing is a module thats going to have to be written isn't it
10:49:22	Varriount	Skrylar: You should already be able to slice strings
10:49:45	ddl_smurf	sorry you think of ddr
10:50:16	Skrylar	Varriount: well i wrote this painful code to do line wrapping but instead of returning copies of strings it just gave the start/end indices
10:51:37	Varriount	Skrylar: And..?
10:52:41	Skrylar	i didn't see a lot of stdlib code that dealt with indices
10:52:51	Skrylar	most of the string routines i saw returned copies of tiny strings
10:55:19	Varriount	Skrylar: True.. But modifying those procedures to return indices is probably not that hard.
11:01:08	Skrylar	i have noticed in my local nimroding that i have started getting too many 'skblah' folders
11:01:30	Skrylar	i should probably just rename it all under one library named 'skylights' instead of trying to keep this crap separate
11:06:17	Varriount	Skrylar: Does it all concern a single thing?
11:12:08	Skrylar	Varriount: no, its just various things i need
11:12:59	Skrylar	but a gap buffer deals with text, so i put it in skText; but text needs unicode, so it references skutf, but then if a gap buffer wants to use bitfields for efficiency i have to put it somewhere else
11:13:18	Skrylar	which means i get a dozen small modules and end up having to use a meta-folder with git submodules to do anything
11:13:31	Varriount	Skrylar: I could have sworn that nimrod had a bitfields module
11:13:39	Skrylar	that was an example
11:13:53	Skrylar	i was entertaining the idea of using an integer where the high bit was used as a dirty flag
11:14:48	Skrylar	since a gap position of -X makes no sense, it would store the value of a signed int and use the extra bit to see if the value changed; so it would only lazily move the gap
11:15:11	Skrylar	which would entail making a module to avoid spaghetti code, which has to go somewhere
11:27:57	Varriount	Araq: I modified the testament/babel-integration pull request so that the babel package tests aren't run when 'all' is specified.
11:53:40	*	BitPuffin quit (Ping timeout: 252 seconds)
12:15:48	Skrylar	Varriount: http://i.imgur.com/7B41peV.jpg sometimes i think i don't use my time very wisely :/
12:18:21	Varriount	Skrylar: *shrug* I like planning things. What software is that?
12:19:03	Skrylar	yEd
12:19:46	Skrylar	I got interested in it again because someone was asking about articy:draft alternatives on reddit and someone linked to this article where an indie made an RPG using yEd and some custom XML parsing
12:20:05	Skrylar	basically made his own tool that ate yEd output and turned it in to dialog script
12:20:40	Varriount	Interesting.
12:20:42	Skrylar	all though it also has a pretty nice automatic layout button for flowcharts which is pretty sweet
12:21:05	Skrylar	yay suspiciously useful freeware. lol
12:30:45	Skrylar	i wonder why gvim on windows installs with these assinine default settings
12:31:03	Skrylar	"your backspace key is going to have this really retarded restriction where it doesn't work half the time :D"
12:38:55	*	BitPuffin joined #nimrod
12:40:10	Varriount	Skrylar: I don't suppose there's a "beginners guide to vim"?
12:40:31	Varriount	Or do I just have to plunge in head first, and hope I don't drown?
12:42:53	EXetoC	I never read any guides, but I did start off with vimtutor the second time around, and that helped
12:46:28	EXetoC	http://www.viemu.com/a_vi_vim_graphical_cheat_sheet_tutorial.html
12:51:02	Skrylar	i don't remember how i learned vim, all though i barely know much of it
12:51:21	Skrylar	there's also vimcasts which is good
13:00:57	Skrylar	as boring as making flowcharts and mindmaps can get, it sure does make the actual coding process really simple
13:05:30	BitPuffin	ping Araq
13:05:47	Araq	png
13:06:06	BitPuffin	Araq: why are we adding an alias for python programmers
13:06:39	Araq	er ... what?
13:07:01	BitPuffin	Araq: Yeah of pow
13:07:05	BitPuffin	Araq: saw that on githerb
13:07:16	Skrylar	lol githerb
13:07:58	Araq	well I dunno about python but nimrod had ^ for pointer deref and got x[] instead so ^ becomes available
13:08:25	BitPuffin	Araq: well then why are we not going with ^
13:08:37	Araq	well but we are
13:08:41	BitPuffin	I am referring to https://github.com/Araq/Nimrod/pull/878
13:08:56	BitPuffin	Araq: okay well then why alias it. Considering it is what you were bashing linagl for :P
13:09:31	BitPuffin	granted it's not merged
13:09:38	Araq	yes I got your point
13:09:39	BitPuffin	but I don't see any discouraging
13:09:52	Araq	I didn't really look at this PR tbh ...
13:10:14	Araq	I just assumed it implements ^ and perhaps deprecates 'pow'
13:10:29	BitPuffin	well I didn't llook at the code lol
13:10:36	BitPuffin	but it doesn't sound like it by the title
13:10:49	BitPuffin	oh
13:10:50	BitPuffin	lol
13:10:52	BitPuffin	it adds **
13:10:57	BitPuffin	you knwo how you do x ** y
13:11:00	BitPuffin	in python for pow
13:11:05	Araq	also aliases are fine now when they are not introduced by BitPuffin ...
13:11:29	BitPuffin	Araq: lol, fuck off :D
13:11:31	Araq	because you added 'toString' and that is a different alias
13:11:45	Araq	toString is from Java hell
13:11:51	BitPuffin	well no you bashed a bunch of the other aliases too
13:11:53	Araq	;-)
13:12:16	BitPuffin	and I don't see why we'd need an alias for python nerbs, since we can have ^
13:12:18	BitPuffin	which makes sense
13:12:30	BitPuffin	think ruby might use ** too iirc
13:12:44	Araq	well more seriously, we should decide about an "alias policy"
13:13:15	*	Varriount throws an idea into the pot
13:13:16	BitPuffin	agreed
13:13:34	Araq	I don't mind to introduce -d:pythonlike and -d:rubylike to lessen the pain for scripters
13:13:39	Varriount	Why not have a single file per language which implements aliases for that language?
13:13:48	Araq	Varriount: or that, yes
13:14:00	BitPuffin	what Varriount said sounds a lot better
13:14:12	BitPuffin	because having a bunch of conditional compilation is gonna bloat up the code imo
13:15:41	Varriount	The downside to the alias file solution is that it can't override nimrod's own stuff.
13:16:02	BitPuffin	Varriount: hmm, you mean stuff that is in system right?
13:16:15	Varriount	BitPuffin: As well as the rest of the stdlib
13:16:34	BitPuffin	for seriousness I don't really think that a python/rubyfier should be in the main repo
13:16:40	BitPuffin	better as a babel pkg
13:16:46	Varriount	BitPuffin: Personally, alias files should be a seperate pack-
13:16:54	Varriount	Yeah, what you said.
13:16:58	BitPuffin	since I don't think we actually want to encourage people writing python in nimrod
13:17:05	BitPuffin	Varriount: hah! :D
13:19:42	Varriount	Such things, like programming styles, should realy be confined to the use of personal scripts and so-on
13:21:46	Varriount	Araq: Did you get my message about the testament/babel-integration update?
13:22:47	Araq	Varriount: sure, will merge it later
13:23:10	Varriount	Ok, just wanted to make sure.
13:23:53	Araq	"- On the VLIW architecture (below HD7700) there is no GOTO instruction. Only Loop, If/Else and exit."
13:24:17	Varriount	Wat
13:24:28	Varriount	Is that a Graphics card?
13:24:35	BitPuffin	sounds like AMD
13:24:44	BitPuffin	or wait
13:24:48	BitPuffin	maybe a new intel thing?
13:25:05	BitPuffin	or old I guess
13:25:08	BitPuffin	I dunno xD
13:28:39	Varriount	BitPuffin: http://devgurus.amd.com/thread/167192
13:28:54	Varriount	That's where Araq got it from.
13:29:16	Araq	hey ..
13:29:29	Araq	but yes
13:30:34	*	rndbit quit (Ping timeout: 245 seconds)
13:31:39	BitPuffin	Araq: you just got scroogled
13:32:28	Skrylar	Araq: i have to nap now, but later i'll need to poke you about symbol names in the skutf repo on github
13:32:48	BitPuffin	I'm gonna use nimrod to write a mumble bot eventually hehe
13:32:51	Skrylar	i'll want to make sure everything has acceptable names before shoving it in stdlib
13:32:59	*	rndbit joined #nimrod
13:33:01	Araq	Skrylar: excellent
13:33:06	filwit	fowl: where did you say your assimp example was again?
13:33:25	BitPuffin	filwit: probably in fowltek repo?
13:33:35	filwit	BitPuffin: been looking for it
13:33:38	BitPuffin	filwit: are you spanking the imp's ass?
13:33:46	Skrylar	Araq: https://github.com/Skrylar/skUtf if you want to browse through it in the interim; there's obviously a few parts that are going to get thrown out because they already exist, though the test cases are still handy
13:33:46	filwit	trying too.. ?
13:33:49	filwit	lol
13:34:34	BitPuffin	https://github.com/fowlmouth/nimlibs/blob/master/fowltek/assimp.nim
13:34:40	BitPuffin	da banding
13:34:55	filwit	yeah i know, he said there was an example of it somewhere
13:35:02	BitPuffin	https://github.com/fowlmouth/nimlibs/blob/master/fowltek/musings/nim3d/lib/models.nim
13:35:05	BitPuffin	imports assimp
13:35:27	filwit	ah, thought i looked through musings, thanks that's probably it
13:36:04	BitPuffin	and https://github.com/fowlmouth/nimlibs/blob/master/fowltek/musings/nim3d/examples/testmodel.nim
13:38:01	*	darkf quit (Ping timeout: 265 seconds)
13:39:21	BitPuffin	Araq: btw did this get fixed? https://github.com/Araq/Nimrod/issues/941
13:39:47	Araq	BitPuffin: it is still open for a reason
13:40:55	Varriount	Araq: You mentioned that there were important things that I could do, besides implementing stat()?
13:41:28	Araq	indeed. bug #941 is creepy
13:43:04	Araq	BitPuffin: just got an idea of how to hunt down this bug, but it may take a while
13:44:44	*	ddl_smurf quit (Quit: ddl_smurf)
13:48:54	Varriount	Araq, BitPuffin: Works for me using a nimrod built using latest devel
13:51:42	Araq	yes it's macosx specific and hence might be a corruption
13:52:04	BitPuffin	Varriount: yeah it's a mac thing
13:52:15	BitPuffin	Araq: oh, what is it?
13:52:40	Araq	some node gets recycled too early, but that's just an educated guess
13:52:58	Araq	the node then gets re-used and the compiler produces weird stuff
13:53:42	Araq	with decent watchpoint support not hard to track down
13:54:14	Araq	but nimrod supports watchpoints via software emulation ... and GDB's watchpoints are broken beyond repair
13:57:34	Varriount	Araq: How are GDB's watchpoints broken?
14:00:47	Araq	Varriount: sorry, will tell you later
14:03:29	*	reloc0 quit (Ping timeout: 245 seconds)
14:03:45	*	reloc0 joined #nimrod
14:04:51	*	ddl_smurf joined #nimrod
14:09:07	Araq	BitPuffin: system/gc.nim in 'growObj' there is a call to 'dealloc'. Disable that line, bootstrap again and try to compile your snippet please
14:16:48	*	nequitans_ joined #nimrod
14:22:46	Varriount	Hi again nequitans_
14:23:07	nequitans_	hello Varriount!
14:24:15	Araq	Varriount the tester needs to produce more JSON ... not only the differences between 2 commits
14:24:22	Araq	if you want, you can work on that
14:29:53	Varriount	Araq: Can you be a bit more specific?
14:30:13	Varriount	Do you mean, the differences between multiple commits?
14:30:28	Varriount	Or the date that the tests were made?
14:30:43	Araq	currently it produces the diff between the latest commit and the commit before that
14:32:41	*	awestroke joined #nimrod
14:33:22	Varriount	Hi awestroke
14:35:31	awestroke	Varriount: hello!
14:43:06	Araq	Varriount: or you could fix the remaining lambda lifting bugs ... ;-)
14:46:16	*	zielmicha joined #nimrod
14:52:18	*	nequitans_ quit (Ping timeout: 240 seconds)
15:27:45	BitPuffin	Araq: ok will do it later
15:42:01	*	Varriount|Mobile joined #nimrod
15:44:33	*	r0b4 joined #nimrod
15:45:59	*	rob3 quit (Ping timeout: 240 seconds)
16:16:11	dom96	helooo
16:33:07	dom96	nequitans: Thank you for that great articles :)
16:33:10	dom96	*article
16:35:13	skyfex	nequitans
16:37:36	skyfex	Huh, this IRC client has a weird bug where if you use paste in the search field it pastes into the message field and focuses it
16:37:50	skyfex	Should've used irssi i suppose ;)
16:41:42	dom96	I wish Nimrod would at least be mentioned here: http://blog.fourthbit.com/2014/03/01/the-best-programming-language-or-how-to-stop-worrying-and-love-the-code
16:43:58	filwit	hey dom96
16:44:04	dom96	hi filwit
16:45:07	BitPuffin	yoyo dom96
16:45:39	*	OrionPK joined #nimrod
16:51:03	dom96	BitPuffin: Still playing tha dotas?
16:51:13	dom96	Why do I even ask. Of course you are.
16:51:15	dom96	:P
16:54:01	*	BitPuffin quit (Ping timeout: 240 seconds)
17:08:12	*	faassen left #nimrod (#nimrod)
17:10:24	*	zahary1 quit (Read error: Operation timed out)
17:10:34	*	zahary1 joined #nimrod
17:23:19	*	Demos joined #nimrod
17:27:44	*	Matthias247 joined #nimrod
17:28:28	*	brson joined #nimrod
17:29:42	*	xtagon joined #nimrod
17:46:34	*	DAddYE joined #nimrod
17:47:53	*	BitPuffin joined #nimrod
17:50:36	*	zahary1 quit (Ping timeout: 264 seconds)
17:51:23	*	zahary1 joined #nimrod
18:17:23	*	carum joined #nimrod
18:26:21	*	q66 joined #nimrod
18:29:07	*	meguli joined #nimrod
18:42:22	dom96	hi meguli
18:43:02	fowl	filwit, did you find the example
18:44:58	*	carum quit (Remote host closed the connection)
18:46:08	meguli	hi everyone
18:46:13	meguli	just checking the language
18:48:41	EXetoC	hello
18:48:47	EXetoC	examine it. embrace it. write awesome code
18:54:59	*	vendethiel joined #nimrod
19:02:19	*	meguli quit (Quit: Page closed)
19:37:13	nequitans	dom96, hello and np. it's a pretty awesome language!
19:48:23	dom96	Araq: Read it so that we can reddit it. :)
19:48:29	*	filwit quit (Ping timeout: 269 seconds)
20:02:12	*	OrionPK quit (Ping timeout: 244 seconds)
20:03:30	*	brson quit (Quit: leaving)
20:07:04	*	brson joined #nimrod
20:09:43	Demos	How do people feel about code completion coming up while writing comments?
20:10:05	dom96	That would be annoying.
20:10:21	Matthias247	Can be helpful when you add examples
20:11:05	Matthias247	but ideally it would only come up in code sections of comment blocks :)
20:11:56	Demos	right now it comes up ANYWHERE if you hit '.', and anywhere if you hit ctrl-space. I think turning off '.' in comments is a good idea, I like having access to it
20:12:31	Demos	I am also having path related problems, I could not get idetools to work if the project file was different from the file I ran my query on
20:19:07	*	Duck_ joined #nimrod
20:24:16	*	Duck_ left #nimrod (#nimrod)
20:25:45	*	Duck_ joined #nimrod
20:26:04	*	Endy joined #nimrod
20:26:08	Araq	hi Duck_ welcome
20:26:16	Duck_	good evening
20:42:11	*	Varriount-Mobile joined #nimrod
20:42:54	Araq	hey nequitans I'm reading your article right now :-)
20:43:55	*	Varriount_ joined #nimrod
20:45:37	*	Varriount quit (Ping timeout: 240 seconds)
20:46:07	*	Varriount|Mobile quit (Ping timeout: 244 seconds)
20:47:40	*	Varriount-Mobile quit (Ping timeout: 244 seconds)
20:48:22	Demos	that matrix example not that great... ideally we would use fixed length arrays for a matrix.
20:49:23	Araq	Demos: that depends
20:49:41	Demos	well yeah, if you want a dyanmic matrix you need dynamic memroy
20:49:45	Demos	*memory
20:50:01	Araq	I keep arguing that a single matrix type simply doesn't cut it, generic or not
20:50:07	Demos	and it is quite hard to do fixed length matrices like we want without static[T]
20:50:11	Demos	really?
20:50:32	Araq	a 4x4 matrix is inherently different from a NxN matrix
20:50:44	Araq	for like every imaginable algorithm
20:50:55	Demos	right, you need to specialize
20:51:39	Demos	then there are various sorts of sparse matrices
20:51:53	*	carum joined #nimrod
20:52:18	Demos	the definitions of most algorithms are the same for 4x4 matrices as they are for NxN ones
20:52:28	Demos	assuming that the operations makes sense on any square matrix
20:52:43	*	carum quit (Remote host closed the connection)
20:58:46	Discoloda	nimrod has something else going for it compared to rust, its taking me over 3 hours to compile it
21:00:13	*	Duck_ quit (Ping timeout: 245 seconds)
21:00:21	Araq	"Nimrod also supports multiple dispatch which allows dynamic binding to variables." er ... no
21:00:40	Araq	it allows dynamic binding of methods
21:03:14	Matthias247	Discoloda: for compiling rust or nimrod?
21:03:30	EXetoC	Rust ofc :>
21:03:48	EXetoC	it's an implementation issue, but then again there are no alternatives
21:04:39	EXetoC	I tried to compile it with 1GB of RAM. It was slow :>
21:05:27	Matthias247	yeah, rust takes a while. Especially when you have to build llvm. But I think it's about 45min here
21:05:28	Discoloda	compiling librustc took 2 hours 47 seconds, for stage 0, now its doing it again for stage 1 :/
21:05:40	Matthias247	and Nimrod is < 5min
21:06:51	Matthias247	it will build llvm only once, so the further stages will be faster. But 3 hours still sounds slow
21:07:11	Matthias247	notebook?
21:07:36	Discoloda	im on a small VPN, 512MB of ram, 2GB of swap file
21:08:15	Matthias247	ah ok, that are quite limited resources :)
21:08:29	Discoloda	yeah
21:08:57	Matthias247	I'm glad I bought 16gb ram when it was cheap :)
21:09:07	dom96	Don't they have nightly builds of the compiler?
21:09:33	Matthias247	depends on the platform
21:10:02	Discoloda	i wish i bought more ram for my gaming computer when it was cheaper
21:10:57	dom96	I wish the RAM I bought would work.
21:11:29	Matthias247	at work we now have a workstation with 128gb - don't know what to do to utilize that :)
21:12:01	*	Endy quit (Ping timeout: 240 seconds)
21:12:13	Discoloda	make a ramdisk and play some games with zero load time
21:12:34	EXetoC	put all the files on it
21:12:46	Matthias247	guess the GPU is too shitty for gaming :)
21:13:03	*	carum joined #nimrod
21:15:39	Demos	I have seen matlab use amounts of ram that would make you run for cover
21:16:01	Matthias247	oh yeah
21:16:46	Discoloda	you dont need super pretty graphics to enjoy zero load time
21:16:59	Matthias247	I did signal processing simulations with Matlab in an internship were I exceeded 4Gb (was much back then), all constantly swapped and the simulation took a whole weekend...
21:17:43	Matthias247	probably the same thing would now run with julia in 2 hours
21:17:49	Demos	yeah the workstation I had was equipped with 64gb
21:17:53	Demos	and matlab was swapping
21:18:58	*	carum quit (Read error: Connection reset by peer)
21:19:18	*	carum joined #nimrod
21:20:04	EXetoC	:E
21:20:12	Matthias247	but currently I'm doing no more signal processing. Mostly compiling stuff. But even compiling a whole Linux system is pretty much CPU and I/O limited and not on RAM
21:20:34	EXetoC	Demos: did you analyze the universe?
21:21:03	Demos	data from RNA sequenceing
21:21:06	Demos	asstons of data
21:21:22	Demos	and matlab is pathetic
21:22:09	Matthias247	we did analyzing RF transmission channels. That's also tons of data
21:24:00	*	brson_ joined #nimrod
21:24:04	*	brson_ quit (Client Quit)
21:25:22	*	brson quit (Read error: Operation timed out)
21:39:55	Demos	hey fowl: How good is your assimp wrapper?
21:40:50	Demos	for some reason it does not specify {.dynlib.} and does not seem to import all the relevant functions
21:44:28	EXetoC	cdecl and dynlib are pushed/popped
21:48:30	Demos	ah, missed that
21:48:37	Demos	it looks pretty good
21:48:55	Demos	aside from not wrapping aiReleaseImport
21:52:06	*	carum quit (Remote host closed the connection)
21:52:42	*	carum joined #nimrod
21:57:48	*	carum quit (Ping timeout: 264 seconds)
22:12:52	Matthias247	I think I finally succeeded in parsing IPv6 addresses...
22:17:18	Araq	ping Skrylar
22:18:39	Araq	nequitans: you should mention that the multithreading you're using is low level and soon there will available be a much better one
22:19:30	Araq	ping Varriount_
22:34:22	*	Varriount_ is now known as Varriount
22:34:27	Varriount	Araq: Yes?
22:34:53	Araq	I want you to split the babel handling
22:35:13	Araq	the packages under nimrod-code should be tested with 'all', the others should not
22:36:24	Varriount	Araq: How would one specify to test "the others"?
22:38:24	Araq	"official" vs "addon" perhaps
22:39:10	Varriount	Araq: I mean, on the command line
22:39:30	Varriount	Would it be two seperate categories?
22:39:40	Araq	yes
22:53:44	Varriount	Araq: testament/babel-integration PR updated.
22:55:26	Araq	hmm that was quick
22:55:37	Araq	now I need to review it
22:55:53	Araq	did you test it?
22:56:55	Varriount	Araq: Yes.
22:58:22	Varriount	Araq: I already had some functionality to filter packages, I just extended it a bit.
23:03:50	Matthias247	Araq, dom96: Any objections against adding an IP Address structure to the standard library? https://gist.github.com/Matthias247/9336472
23:05:49	Araq	Matthias247: use sets, c in {'0'..'9'}
23:05:53	Varriount	Matthias247: IIRC, object's with case members are unions underneath, so you might want to make those array members refs
23:06:33	Matthias247	Varriount: that is the expected behavior :)
23:07:14	Matthias247	Araq: for style or does it really make a difference?
23:07:31	Araq	for style
23:08:11	Matthias247	I would expect the set to be slower, because it sounds like you have to compare against object instead of only comparing two numbers
23:10:12	*	io2 quit ()
23:10:22	Varriount	Sets are bit vectors. Also, micro-optimization is an unnecessary evil here.
23:11:46	dom96	Matthias247: I like it. Add it to the new net module. Will need something to convert into a form that can be used by connect/bind though.
23:13:02	Matthias247	dom96: yeah, the socket procs should all take an IPAddress instead of the current string as a parameter ;)
23:13:12	dom96	I agree with Araq though, you should also use case statements instead of those if statements for parsing
23:13:38	*	DAddYE quit (Read error: Connection reset by peer)
23:13:54	*	DAddYE joined #nimrod
23:14:22	dom96	Matthias247: What if I want to specify a domain then? :P
23:14:24	Araq	Matthias247: the compiler optimizes away the set construction etc. it ends up producing the same code in your case
23:14:37	Matthias247	dom96: then you use DNS :)
23:14:54	dom96	Matthias247: Connect/bind currently handles that automatically
23:14:57	Matthias247	Araq: ok, then I'll trust you and change it :)
23:15:03	Araq	in other cases the set is translated to a lookup table and is faster than a bunch of comparisons
23:15:59	Matthias247	dom96: you basically want an overload. One proc that directly takes an address and another one that takes a string and does DNS if necessary
23:17:27	dom96	Sure. That will work.
23:19:20	dom96	Is there a use case for parsing an IP and then modifying the resulting TIpAddress?
23:19:31	Skrylar	Sure
23:19:38	Skrylar	If you're pinging a range of IPs you might want to inc() to the next one
23:21:32	dom96	oh right
23:22:01	Matthias247	I'm also not sure if i should make the inner address fields private and provide gettter methods for it
23:22:32	Matthias247	would prevent setting an imcompatible family
23:22:54	Skrylar	True. This would also be done for you by using a variant object
23:23:05	Skrylar	At least locking address objects to the right family type
23:24:00	Araq	Varriount: just fyi your iterator is an inline iterator and you have 3 yields in it --> code bloat
23:24:01	Matthias247	Yes, that's what I have and what it is doing
23:24:18	Matthias247	but in the other way you can throw a user-defined conversion error
23:24:30	Araq	but I don't care enough, so ...
23:24:40	Matthias247	that doesn't work: return (c in {'0'..'9'}) or (c in {'a'..'f'}) or (c in {'A'..'F'})
23:24:42	NimBot	Araq/Nimrod devel 9f81e48 Clay Sweetser [+0 ±2 -0]: Integrated 'babel' with testament
23:24:42	NimBot	Araq/Nimrod devel 5876e62 Clay Sweetser [+0 ±1 -0]: Removed babel package tests from those run when 'all' is specified.
23:24:42	NimBot	Araq/Nimrod devel 7267635 Clay Sweetser [+0 ±1 -0]: Allowed specification of what babel packages to install via new categories - 'babel-core', 'babel-extra', and 'babel-all'
23:24:42	NimBot	Araq/Nimrod devel 1b80d31 Andreas Rumpf [+0 ±2 -0]: Merge pull request #970 from Varriount/testament/babel-integration... 2 more lines
23:24:46	Skrylar	Araq: i have an iterator with 6+ yields in a row in it
23:24:56	Skrylar	Admittedly it was one of those few times a goto ladder would have been more efficient :P
23:25:44	Araq	Matthias247: what do you mean? compiler bug?
23:26:10	dom96	Matthias247: You should use: case address_str[i]: of {'0' .. '9'}, {'a' .. 'f'} ...: # isXDigit
23:26:12	Araq	also it's c in {'0'..'9', 'a'..'f', 'A'..'F'}
23:26:14	Matthias247	Araq: it also accepts 'z' and other chars
23:26:52	Araq	report it
23:27:05	Araq	(if it's not your fault of course :P )
23:27:19	Matthias247	c in {'0'..'9', 'a'..'f', 'A'..'F'} -> same
23:27:27	Matthias247	so I'll report it
23:27:33	Matthias247	and use the other method
23:28:52	Matthias247	oh, mine also accepts it
23:28:58	Matthias247	seems to be my fault ;)
23:29:06	Araq	well the compiler uses things like these and it works so yeah
23:29:22	Araq	would be surprised if there is a bug left for that
23:29:46	Matthias247	think I have refactored the stupid IPv6 parsing method too often :)
23:31:10	Skrylar	i. hate. parsing.
23:31:24	Matthias247	ok, works now
23:31:48	dom96	I used to hate parsing, but then Araq enlightened me.
23:31:58	Skrylar	PEGs?
23:32:16	dom96	while loop + case statement ftw
23:32:18	Skrylar	For me it was parsec; I can actually get a parser combinator to obey.
23:32:34	dom96	parseutils is nice too
23:33:04	dom96	Similar to Haskell's parser combinator.
23:34:12	Varriount	My best experience with binary stream parsing (custom protocols) was using a python library called "construct"
23:34:30	Skrylar	time to bother zah... vim-nimrod on windows still tries to open /tmp/nimrod.log
23:34:44	Matthias247	if protocols are designed in a useful manner you don't need to parse them ;)
23:34:53	Varriount	Matthias247: Says you.
23:34:57	Skrylar	actually binary protocols aren't bad; i wrote a msgpack-v5 parser in rust a while back
23:35:26	Matthias247	yeah, they are in most cases much easier than text protocols
23:35:38	Skrylar	once you know how to do the bit fiddling
23:35:43	*	zahary quit (Quit: Leaving.)
23:35:49	Matthias247	yup
23:35:57	Skrylar	msgpack uses unsigneds :>
23:36:14	dom96	Skrylar: Now write one for Nimrod.
23:36:15	Skrylar	well, has both. it only uses signed ints for negatives; but its a line-protocol so it doesn't matter
23:36:22	Skrylar	dom96: el oh el
23:36:37	Matthias247	had an intern last year that did addr[0] + addr[1] + addr[2] + addr[3] to decode 32bit values ;)
23:36:55	Skrylar	dom96: i might at some point.
23:37:09	Skrylar	I found out that like nobody supports the latest msgpack... after writing code that supported the extension fields
23:37:39	Varriount	dom96: Why does NimBot no longer announce commits?
23:37:46	Matthias247	I'm doing something like messagepack (but much more advanced) at work. But unfortunatly can't make it OSS :(
23:37:56	dom96	Varriount: It did like 10 minutes ago.
23:38:09	Varriount	Oh, nevermind.
23:38:41	dom96	Varriount: If you just pushed and it doesn't announce it then keep in mind that GitHub may not always instantly send the data to NimBuild.
23:39:09	Skrylar	i donno.. msgpack is pretty good nowadays
23:39:21	Skrylar	only tools i might want for it would be a msgpack<->json transcoder
23:39:39	Skrylar	(inb4 pandoc clone for xml/jason/yaml markups)
23:39:53	dom96	I think i've only worked with binary protocols once
23:40:21	Matthias247	serialization is one thing. The other one is building a good RPC library around it. With good performance and kinds of error handling
23:40:24	dom96	I am a master at parsing the IRC protocol though
23:40:38	Skrylar	isn't the IRC protocol something designed to be handled with strtok ;p
23:41:06	dom96	Sure. It's pretty easy to parse.
23:41:33	*	Skrylar glares at vim unhappily... I use the help function for 'backspace' and it says nothing about how to fix it on windows. I use Google, THERE IS A HELP TOPIC TAGGED BACKSPACE. -_-
23:41:33	dom96	But as with any procotol there are some complications.
23:41:39	dom96	*protocol
23:41:45	Skrylar	dom96: like unicrap \o/
23:42:27	dom96	Can't say I ever had problems with Unicode.
23:42:48	Skrylar	I guess an IRCbot wouldn't
23:43:02	Skrylar	Its no fun if you're writing a text editor though
23:43:42	NimBot	Araq/Nimrod devel 7c7ed1c Varriount [+0 ±1 -0]: Fix categories.nim
23:43:48	*	carum joined #nimrod
23:44:03	dom96	Can't recall any major problems that I had with unicode when writing Aporia either.
23:44:14	EXetoC	Skrylar: backspace in all modes?
23:44:36	Varriount	dom96: GTK probably handled it all for you.
23:45:31	Varriount	Skrylar: Have you ever managed to get GTK built on Windows (without Cygwin)?
23:45:50	dom96	Varriount: Yeah, definitely.
23:46:02	Skrylar	Varriount: i do not touch gtk :S
23:46:08	Skrylar	gobject is sadness
23:46:22	Varriount	Skrylar: Good. Your sanity is intact then.
23:46:37	Varriount	GTK = Circular Dependancies
23:46:45	Skrylar	i concur with GTK handling his unicode problems
23:46:52	Skrylar	doesn't he use gtksourceedit?
23:47:00	Varriount	Yes
23:47:07	dom96	GtkSourceView. He does.
23:47:23	Skrylar	Y'know.. I.. Never really liked Scintilla-ish things
23:47:36	Skrylar	I don't know why because the text component works really well, it just feels like nobody uses it to its full potential
23:48:00	dom96	My sanity is unfortunately not intact.
23:48:13	dom96	Gtk may handle unicode well, can't say it handles much else well.
23:48:37	Skrylar	Y'know I was actually considering building my morph gui over cairo as of late
23:48:40	dom96	And GIMPNet/gtk+ rarely helps :(
23:48:52	Skrylar	I think you can static link to MPL1.1
23:49:26	dom96	I am still considering writing a GUI on top of OpenGL.
23:49:59	Skrylar	GL is kind of like working on a ship in a bottle.
23:49:59	dom96	Maybe we should work on a GUI together.
23:50:06	Discoloda	dom96: i support that
23:50:07	EXetoC	how about cairo?
23:50:27	Skrylar	EXetoC: cairo is nice but you still need to plug a font layout engine over it
23:50:46	Skrylar	Which you can admittedly do with Pango on linux, and probably hook Windows' native glyph support there
23:50:57	Skrylar	I need to recheck if MPL1.1 allows static linking
23:51:12	Skrylar	I'm somewhat allergic to unnecessary LGPL dependencies :S
23:51:14	EXetoC	right
23:51:16	dom96	Are there no other font layout engines?
23:51:25	Skrylar	I think there are a few lesser known ones
23:51:36	Skrylar	All though if the OS has one there's not a huge downside to using it
23:51:38	Matthias247	Ok, I've set up a PR for the net module
23:51:39	Skrylar	Lowers your footprint
23:51:40	Discoloda	freetype isnt hard to use
23:51:46	Skrylar	Discoloda: not a layout engine :)
23:51:56	Discoloda	ah
23:51:59	Araq	good night
23:52:08	Skrylar	Freetype is a rasterizer.
23:52:14	Skrylar	It doesn't position the diacritic marks for you
23:53:00	Skrylar	EXetoC: yeah. on windows the backspace mode for gvim doesn't delete past where you start editing
23:53:28	Skrylar	and using :help backspace doesn't take you to the documentation on how to fix, even though there is a help topic for *backspace* *bs* which does tell you :|
23:54:15	EXetoC	ok
23:54:26	dom96	Skrylar: I have a brilliant solution for you: delete vim, install sublime or something other than vim.
23:54:37	Skrylar	dom96: lololol
23:54:39	Matthias247	hehe
23:54:46	EXetoC	Skrylar: doesn't position them how?
23:54:59	Skrylar	EXetoC: if i press insert, then backspace, it beeps at me
23:55:11	EXetoC	(FT)
23:55:13	Skrylar	"NO YOU SAID INSERT YOU CANT DELETE BEFORE THE INSERT POINT" "this is not 1990 and you are not vi)
23:55:43	Skrylar	on linux they usually turn off the pedantic options but for some reason gVim on windows has them on by default :B
23:56:04	Skrylar	fixed
23:56:05	Discoloda	vim on windows has retarded defaults
23:56:14	dom96	Anyway. Good night.
23:56:19	Skrylar	dom96: so what you are saying is we need to jump on the SublimeText cloning bandwagon in nimrod
23:56:20	Skrylar	g'night
23:56:38	dom96	Skrylar: Yep.
23:57:05	dom96	Skrylar: Sublime uses some custom toolkit which uses OpenGL right?
23:57:11	Skrylar	his own stuff, yeah
23:57:25	Skrylar	Sublime 1 was neat, then he removed the toolbars :(
23:57:54	dom96	I really can't bring myself to pay for it, or to use the "free" version.
23:58:04	Skrylar	well, i won't buy the upgrade to 3
23:58:06	dom96	Especially after spending so much time on Aporia.
23:58:11	Skrylar	he hasn't done anything i care about to the new version
23:59:03	Skrylar	Could always do that weird business model Ardour does
23:59:05	Matthias247	the plugin api is a little better. When when everyone tries to stay compatible with v2 it won't help
23:59:12	dom96	I think a Delphi/WinForms-like toolkit with an IDE which includes a GUI designer would be awesome for Nimrod.
23:59:22	Skrylar	i was working on something like that
23:59:26	Skrylar	Well, not so much winforms as morphic
23:59:26	Matthias247	and it respects the DPI settings