<<01-09-2012>>

---

02:19:13	*	q66 quit (Quit: Quit)
06:58:36	*	EfTwelve quit (Remote host closed the connection)
09:32:03	*	Boscop quit (Disconnected by services)
09:32:05	*	Boscop joined #nimrod
10:03:13	shevy	oh a question
10:03:25	shevy	say, I would write something in nimrod-gtk
10:03:43	shevy	in 2 years, would that code still be expected to work? or may there be large changes to nimrod-gtk in between
10:27:37	*	q66 joined #nimrod
11:19:59	dom96	shevy: Hard to say. The wrapper shouldn't change too much, but someone might make a
11:20:08	dom96	nicer binding to it.
11:20:23	dom96	The only changes to the actual wrapper will be bug fixes though.
11:20:35	dom96	The interface shouldn't change, unless gtk actually changes.
11:20:50	dom96	But that's gtk3 so... yeah. It is expected to work :P
16:36:12	*	shevy quit (Ping timeout: 268 seconds)
16:48:51	*	shevy joined #nimrod
17:07:45	*	XAMPP_ joined #nimrod
17:17:39	*	XAMPP quit (Ping timeout: 246 seconds)
17:37:16	*	Trix[a]r_za is now known as Trixar_za
17:38:28	dom96	Trixar_za: Have you tried compiling Aporia recently?
17:38:38	dom96	It should compile for you now.
17:38:50	Trixar_za	Nope. I should try updating it and Nimrod
17:39:00	Trixar_za	I may just do that now
17:39:00	Trixar_za	:P
17:39:26	dom96	Let me know how it goes :D
17:39:50	Trixar_za	... still freaks me out that I know the web addresses by heart
17:40:50	dom96	hehe
17:46:23	Trixar_za	Right, here we go with bootstrapping Nimrod
17:46:24	Trixar_za	:P
17:57:53	Trixar_za	Nearly done
17:59:34	dom96	You know you could have just downloaded a build from nimbuild :P
17:59:58	Trixar_za	Yes, but this way I look sexy
18:00:22	Trixar_za	And when I work as a programmer, it will look like I'm actually doing some work
18:00:27	dom96	haha
18:05:50	Trixar_za	This last stage always takes a while
18:05:51	Trixar_za	:P
18:06:43	dom96	Your computer is quite slow :P
18:06:50	Trixar_za	Yeah :(
18:07:18	Trixar_za	It's a little faster now. At 9 minutes for me now instead of the usual 15
18:08:06	dom96	Interesting.
18:09:08	Trixar_za	btw
18:09:29	*	Trixar_za is listening to: Haters by HoN Raps
18:09:35	Trixar_za	I actually tracked down a copy of the song
18:09:36	Trixar_za	:P
18:16:44	Araq	dom96: try to scroll by clicking on the arrow in aporia
18:17:03	Araq	do that once a day
18:18:09	dom96	huh?
18:19:04	Araq	it often opens a new tab :P
18:19:14	Araq	which doesn't help when scrolling
18:19:20	Araq	it's really annoying
18:19:21	fowl	you mean doubble clicking on the scroll bar
18:19:31	Trixar_za	Aporia seems to be compiling
18:19:33	dom96	I see.
18:19:49	dom96	Araq: Come on, it's a known issue.
18:19:59	Trixar_za	lol
18:20:05	Trixar_za	Yeah, I discovered that on myself
18:20:07	Araq	fix it
18:20:12	Araq	fix it
18:20:13	Araq	fix it
18:20:14	Trixar_za	one*
18:20:45	fowl	what i need is some kind of layout/splitting options
18:21:33	dom96	Araq: You could always you know, use a scroll wheel like a normal person :P
18:21:41	fowl	i work best like this: http://minus.com/lFOq6tKBRTlco
18:21:47	Trixar_za	Yay, it compiled
18:21:53	Trixar_za	Did one of you fix it?
18:21:53	Trixar_za	:P
18:22:16	dom96	Trixar_za: I fixed it :P
18:22:36	dom96	fowl: Vertical split is in the todo list, not sure if I will get it in for 0.1.3 though
18:23:24	Trixar_za	I like piekno
18:23:38	dom96	Trixar_za: :D
18:31:55	Trixar_za	I, like all my bachelor forefathers, will now have breakfast foods for dinner
18:35:44	fowl	breakfast for dinner is the best
18:36:36	Trixar_za	I'm newly single, so celebrating like a guy
18:36:39	Trixar_za	:P
18:36:49	Trixar_za	That and I'm too broke to go on a drinking binge
18:51:00	shevy	lol
18:51:07	shevy	Trixar is one funny guy
18:55:31	fowl	Trixar_za: i used to have tons of fun for $5 i'd go to the gas station and buy two bottles of cough drops, they sold this brand that only had DXM in it so it didnt make you sick when you downed 40 pills
18:58:35	*	zahary joined #nimrod
19:02:43	Trixar_za	Probably will only work once on me. My body sucks because it builds tolerance to most things pretty quickly
19:47:21	*	EfTwelve joined #nimrod
20:06:36	Araq	hi EfTwelve, welcome back
20:09:36	EfTwelve	hi
20:29:22	Araq	ping zahary
20:30:06	Araq	semtypes.nim:830
20:30:16	Araq	result = copyType(result, getCurrOwner(), false)
20:30:18	Araq	for i in countup(1, n.len - 1):
20:30:19	Araq	result.rawAddSon(semTypeNode(c, n.sons[i], nil))
20:31:04	Araq	this is only supposed to be triggered if result.kind == tyTypeDesc, right?
20:36:23	reactormonk	Araq: would you accept some random overloads?
20:36:45	Araq	reactormonk: like?
20:37:31	reactormonk	Araq: random(int) 0..<int random(x,y) x..<y and same for floats
20:38:20	fowl	proc random(x: TSlice[int]): int = result = random(x.b - x.a) + x.a
20:38:39	Araq	random(int) already exists, right?
20:38:46	reactormonk	oh, they're called slices, not ranges...
20:38:49	reactormonk	Araq: yep
20:39:33	fowl	isnt there a standard randomfloat func we can import from c
20:39:40	reactormonk	fowl: sure there is
20:40:01	Araq	dunno I think random(int) should become a template
20:40:15	reactormonk	huh?
20:40:34	Araq	so you'd have: case random() mod 3
20:40:47	Araq	of 0, 1, 2: echo "yummy"
20:40:56	Araq	--> no 'else' required
20:41:10	Araq	compiler is smart enough to perform the interval analysis :P
20:41:30	Araq	it can't do this interval tracking through procs though
20:41:35	Araq	so we'd have:
20:41:50	Araq	template random(n: int): int = random() mod n
20:42:52	Araq	same for random(x.b - x.a) + x.a
20:44:56	Araq	thinking about it -- it's wrong for negative numbers :-|
20:45:11	Araq	-5 mod 2 == -1 according to the REPL
20:45:57	Araq	oh well, I'll fix it somehow
20:52:24	fowl	there's always %%
20:53:44	Araq	true
23:22:56	zahary	Araq, about your question: no, it would be true for tyExpr too
23:23:46	Araq	well sure, but was that your intention?
23:23:59	Araq	the code does not contain any check
23:24:37	Araq	and I changed it for anything != tyTypeDesc :P
23:24:44	Araq	to support:
23:24:55	Araq	expr{lit|call, noalias}
23:28:29	Araq	also you setting typ.n for tyExpr in sigmatch is evil
23:30:09	Araq	but it's a minor grief as I introduced typ.constraint for the new features
23:32:23	Araq	template optRe{re(X, Y)}(X: string{lit}, Y: varargs[expr]) =
23:32:24	Araq	let cached {.global, gensym.} = re(X, Y)
23:32:26	Araq	cached
23:32:39	Araq	--> problem: it triggers an endless recursion
23:33:04	Araq	as the resulting code is searched for TRMs again
23:33:20	Araq	and 're(X, Y)' matches just like it did the first time
23:41:45	Araq	zahary: well?
23:47:48	zahary	my reasoning was that this T{foo, bar} syntax is always likely to work by introducing children to the type whatever the implemented feature is so I didn't put any restrictions there. feel free to change it
23:48:01	zahary	anything != tyTypeDesc doesn't sound right tho
23:48:37	Araq	but it is: string{lit}
23:49:05	Araq	is natural to do
23:49:27	Araq	expr{lit} is often not good enough
23:49:34	Araq	and you can also have:
23:49:43	Araq	string{skVar}
23:50:15	Araq	which matches a node of kind nkSym with a symbol of kind skVar of type 'string'
23:51:15	Araq	in fact, I've implemented a stack based virtual machine for efficient pattern matching of these ...
23:51:17	zahary	do we disagree somewhere? feel free to add a switch statement based on the type of the constraint
23:51:34	Araq	well I did
23:51:46	Araq	and you said != tyTypeDesc doesn't sound right to you ;-)
23:51:53	Araq	which is my 'switch'
23:52:53	zahary	I meant result.kind != tyTypeDesc doesn't sound right (because tyExpr{string} should continue to work) - is that what you meant?
23:53:19	Araq	yes
23:53:35	Araq	I broke tyExpr{string}
23:54:02	Araq	why was that useful btw?
23:54:14	zahary	this is compile-time string value
23:54:43	Araq	hrm
23:55:35	zahary	we probably have some overlap now between string{lit}, expr{nkStrLit} and expr{string{
23:55:43	zahary	expr{string}
23:56:12	Araq	not really as the patterns only work in TR patterns
23:56:16	zahary	expr{string} is any kind of compile-time evaluatable expression that returns a string
23:56:38	zahary	expr{nkStrLit} is supposed to be stronger, requiring a real literal node
23:56:47	Araq	I see now why you used static{string} for that
23:56:58	Araq	well
23:57:13	Araq	I'll change it so that:
23:57:36	Araq	the [] will introduce a type class
23:57:47	Araq	and {} a pattern
23:58:02	Araq	typedesc[int] # no problem, right?
23:58:13	Araq	expr[string]
23:59:06	zahary	well, no problem, but they sound more pattern-like to me than generic-like (i.e. seq[int])
23:59:23	Araq	well T[] produces a type
23:59:33	Araq	and T{} does not
23:59:55	zahary	I mean expr[string] is more of a pattern to me than a type