<< 16-01-2015 >>

---

00:00:01	BlaXpirit_	not funny -_-
00:00:27	Triplefox	ah yeah, it does cover that case where many apis will return a -1 if missing
00:04:02	Triplefox	ah yeah, it does cover that case where many apis will return a -1 if missing
00:04:08	Triplefox	oops, arrowkeys
00:04:55	Triplefox	interesting how negative index "works" for slice and not for regular indexing
00:05:19	Triplefox	e.g. you can do @[1,2,3][-1.. -1] to get @[3]
00:06:07	EXetoC	Araq: you did provide some arguments in favor of inclusive intervals, but it was still debated at great length
00:06:19	Triplefox	i guess the error for a single index would stop some bugs
00:06:35	EXetoC	it can indeed be confusing to either have to add or subtract 1 in certain cases, but I must admit to sucking at such logic
00:06:42	*	Trustable quit (Remote host closed the connection)
00:07:00	BlaXpirit_	EXetoC, Python fixes brain very quickly
00:07:17	BlaXpirit_	but again, it is way too late to change this convention
00:07:25	BlaXpirit_	negative indices should go though
00:07:51	BlaXpirit_	BUT maybe there should be a special syntax inside square brackets
00:08:04	BlaXpirit_	that would be alias for this.len
00:08:24	Triplefox	i can tolerate the current as long as there is a desugared equivalent for the half-inclusive situation
00:08:33	BlaXpirit_	[2..^-2]
00:09:20	BlaXpirit_	no, that's not good
00:09:47	BlaXpirit_	.eval var s = "abcdef"; s[2 .. s.high-2]
00:09:50	Mimbus	BlaXpirit_: eval.nim(4, 1) Error: value of type 'string' has to be discarded
00:10:02	BlaXpirit_	you're supposed to say "cd"
00:10:15	Triplefox	then again, looking at actual code i've written i've already fallen into the size zero bug without having it surface
00:10:20	BlaXpirit_	s[2 .. ^-2] etc
00:10:38	BlaXpirit_	Triplefox, yup, and it's in standard library everywhere, i bet
00:11:02	BlaXpirit_	heck, this may even be possible with term rewriting macro
00:11:11	BlaXpirit_	i tried term rewriting template but it didnt seem to work
00:12:02	Triplefox	.eval var z : seq[int] = @[]; echo(z[0..high(z)])
00:12:06	Mimbus	Triplefox: @[]
00:12:16	Triplefox	okay, well that works i guess .-.
00:12:28	BlaXpirit_	it works because z[0..1000] would give @[] too
00:12:37	BlaXpirit_	as does z[0..0]
00:12:46	Triplefox	actually it out of bounds it when it's too small
00:12:55	Triplefox	.eval var z : seq[int] = @[]; echo(z[0..1000])
00:12:58	Mimbus	Triplefox: Error: unhandled exception: index out of bounds [IndexError]
00:13:45	BlaXpirit_	:o
00:13:53	Triplefox	so in the common case where you iterate over a sequence of unknown size and it happens to be size zero...no bug
00:14:28	EXetoC	that is neat
00:14:41	BlaXpirit_	neat that bug hides from you?
00:14:49	BlaXpirit_	ah high == -1
00:14:51	BlaXpirit_	ok
00:14:53	BlaXpirit_	...
00:15:16	EXetoC	that in isolation anyway
00:19:54	Triplefox	what i get from this is that in nim [0.. 0] is basically "magical" and has at least three contextual meanings
00:20:16	Triplefox	as an array declaration, it's the array of unknown length
00:20:39	Triplefox	on a slicable thing that is length zero, it's empty
00:20:45	flaviu	Triplefox: No, it's an array of size 1.
00:21:12	Triplefox	on a slicable thing that is length greater than zero, it's the whole thing
00:21:20	flaviu	{.unchecked.} is of unknown length.
00:21:52	Triplefox	okay, then it has four meanings and who knows why you'd ever have an array of size 1
00:21:55	BlaXpirit_	.eval echo(newSeq[int]()[0..0])
00:21:59	Mimbus	BlaXpirit_: Error: unhandled exception: index out of bounds [IndexError]
00:22:07	BlaXpirit_	Triplefox, it's not magical
00:22:20	BlaXpirit_	you wrote 0 .. -1 before thinking its 0 .. 0
00:22:38	Triplefox	oh, you're right
00:22:57	BlaXpirit_	inclusive indices :>
00:23:13	BlaXpirit_	i did it!
00:23:18	BlaXpirit_	term rewriting template
00:23:46	EXetoC	are they not supposed to keep the semantics?
00:24:03	BlaXpirit_	term rewriting template in case negative indices are removed
00:24:06	flaviu	Yeah, the solution is worse than the problem IMO
00:24:55	BlaXpirit_	imagine we removed negative indices. if you had s[1 .. -1], you make that s[1 .. ^ -1]
00:25:27	Triplefox	so then it's... array of size 1, array of unknown length(unchecked), whole if not empty, error if empty
00:30:11	Triplefox	good when you want an entire slice, then ugly in the case where you control the slice variables and have to insert emptyslice checks
00:30:53	Triplefox	"entire" = (everything in the original)
00:31:35	Triplefox	although it is still consistent...
00:31:59	Triplefox	.eval echo(@[1,2,3][1.. 0])
00:32:03	Mimbus	Triplefox: @[]
00:32:45	dtscode	hey Araq... how do you get the compiler to compile .tmpl files/
00:33:53	dom96	dtscode: Just include them.
00:34:09	dtscode	thats it? huh
00:34:15	dtscode	thanks dom96
00:34:41	BlaXpirit_	https://gist.github.com/BlaXpirit/f063fe6706201504580d
00:34:48	BlaXpirit_	"Replacement for negative indices"
00:35:28	dtscode	dom96, is there a page documenting .tmpl files?
00:35:39	dom96	http://nim-lang.org/filters.html
00:35:44	dtscode	ty
00:36:21	BlaXpirit_	so guys, what do you think?
00:36:52	flaviu	I don't like it, I think the better option is to use exclusive upper bounds.
00:37:31	def-	BlaXpirit_: looks like an abuse of Term rewriting. If you used this in a library, the user could disable TR.
00:37:36	BlaXpirit_	flaviu, and break all existing code?
00:37:51	flaviu	BlaXpirit_: ye.
00:37:57	BlaXpirit_	def-, whatever, then add it to compiler -___-
00:38:17	flaviu	I'm not actually proposing that the stdlib be changed
00:39:03	def-	BlaXpirit_: I'd just write a proc "sub" that behaves as you want
00:39:15	BlaXpirit_	explain
00:39:28	BlaXpirit_	ah x.sub
00:39:29	BlaXpirit_	whatever
00:39:44	gunn	what am I doing wrong here? or is it a bug? https://gist.github.com/gunn/27e8ae56e5ffebc05a26
00:39:48	BlaXpirit_	the point of that thing i sent is to easily fix code
00:40:02	BlaXpirit_	you can just add 1 symbol and everything is fixed
00:40:29	flaviu	did someone just get bitten by inclusive ranges?
00:40:37	BlaXpirit_	nah it's just empty array
00:41:08	flaviu	gunn: 1 sec, let me check the source
00:41:23	BlaXpirit_	flaviu, -_-
00:41:54	BlaXpirit_	i'm telling you, just 0-row result
00:42:33	dtscode	oh wow... so this is kind of like c's preprocessor system
00:44:25	flaviu	BlaXpirit_: ah, "row number 0 is out of range 0..-1" is coming directly from postgres
00:46:11	Triplefox	i've now lost interest in thinking about slices
00:47:27	BlaXpirit_	Araq, please do consider removing negative indices and introducing something like this syntax https://gist.github.com/BlaXpirit/f063fe6706201504580d to easily fix code that needs it
00:47:52	flaviu	please, not another symbol.
00:48:04	Varriount	^
00:48:18	BlaXpirit_	yes, yes, ^ :p
00:48:27	BlaXpirit_	can be anything else though
00:48:39	def-	gunn: the code looks ok to me. you don't need to pass user and pw?
00:49:03	BlaXpirit_	def-, looks like 0 rows result
00:49:15	BlaXpirit_	or maybe it's just because of previous error
00:49:32	gunn	def-, flaviu, BlaXpirit_: yes, the tweets table was empty, explaining most of the result
00:49:34	Varriount	Alternately, we could allow negative indices only for literals.
00:49:41	gunn	it works fine on a populated table
00:49:47	BlaXpirit_	Varriount, naaah, that won't do
00:50:00	Varriount	Too much magic, I guess.
00:50:02	BlaXpirit_	people sometimes need to just count from end -_-
00:50:22	BlaXpirit_	all i'm saying is, these negative indices should be explicit
00:50:30	def-	gunn: still a strange error for an empty table
00:50:49	BlaXpirit_	requiring a '^' is just one of the countless ways to achieve it
00:52:10	flaviu	ranges could also be less stupid.
00:52:27	BlaXpirit_	no, they couldn't without breaking all the code
00:52:28	flaviu	ie, say if it should be exclusive or inclusive and if bounds wrap
00:52:34	BlaXpirit_	for sure
00:52:43	BlaXpirit_	that's what's good
00:52:49	*	brson quit (Quit: Lost terminal)
00:52:50	BlaXpirit_	but it is impossible to introduce now
00:52:55	BlaXpirit_	(not literally)
00:52:57	*	brson joined #nim
00:53:08	flaviu	rewrite rewrite :P
00:54:08	gunn	def: yes, it looks like it's returning an empty row as per https://github.com/Araq/Nim/blob/master/lib/impure/db_postgres.nim#L162-L163 I wonder why it would SIGSEV on trying to iterate over it?
00:54:33	BlaXpirit_	you have 0 rows
00:54:42	BlaXpirit_	you take the first of the 0 rows
00:54:45	BlaXpirit_	do i need to continue
00:55:03	BlaXpirit_	i probably do
00:55:07	gunn	BlaXpirit_: row.len == 11?
00:55:13	BlaXpirit_	the pointer to 0 rows is not initialized
00:55:23	BlaXpirit_	gunn, row.len is taken from database columns
00:55:36	BlaXpirit_	it is stored separately for efficiency probably
00:56:13	BlaXpirit_	so yeah, when you take uninitializedpointer[0] - that's where the error happens
00:56:22	gunn	So is this wrong: "If the query doesn't return any rows, this proc will return a TRow with empty strings for each column." ?
00:56:36	BlaXpirit_	no, wait, i'm completely wrong here
00:56:43	Varriount	that was changed, I think
00:56:49	BlaXpirit_	gunn, that may indeed be a bug
00:57:24	Varriount	IIRC, something was changed so that something regarding databases returned 'nil' instead of a string. I can't remember specifics
00:57:33	BlaXpirit_	forget the part i said about pointers
00:57:57	flaviu	Varriount: When the DB contains null, it returns nil.
00:58:01	flaviu	instead of returning ""
00:58:08	Varriount	Ah, that was it, yeah.
00:58:42	*	BlaXpirit_ quit (Quit: Quit Konversation)
01:27:26	*	Demon_Fox joined #nim
01:33:58	*	JinShil joined #nim
01:51:49	dtscode	what does the standard filter do?
01:56:03	EXetoC	isn't it documented?
01:57:36	dtscode	i couldnt find it
01:57:40	dtscode	ill look again though
01:58:39	EXetoC	you mean stdtmpl here http://nim-lang.org/filters.html ?
01:58:51	EXetoC	oh, that
01:59:05	EXetoC	no idea. don't bother I guess
01:59:20	dtscode	ok. awesome
01:59:42	dtscode	and yeah i know what stdtmpl does. i was wondering what stdtmpl | standard does
01:59:52	dtscode	standardizes the html maybe. idk. oh well
02:01:22	EXetoC	the only reference I've found seems to indicate that it simply uses the standard syntax (rather than begin/end for example)
02:01:34	EXetoC	does |standard work?
02:01:58	dtscode	what do you mean? like appended to stdtmpl? or just |standard
02:02:23	EXetoC	stdtmpl|standard
02:02:38	EXetoC	strongspaces might work too
02:03:03	dtscode	itmight. let me finish and ill try to compile it
02:04:07	*	flaviu quit (Read error: Connection reset by peer)
02:04:07	*	Mimbus quit (Read error: Connection reset by peer)
02:06:21	*	flaviu joined #nim
02:10:12	*	Hakaslak joined #nim
02:15:03	*	Hakaslak quit (Ping timeout: 264 seconds)
02:20:14	*	phi_freenode is now known as phira
02:25:17	*	jefus__ joined #nim
02:28:53	*	jefus_ quit (Ping timeout: 246 seconds)
02:38:00	dtscode	can someone tell me why this repo https://github.com/DTSCode/dtscodeio is causing: https://bpaste.net/show/44b831bf7e8c
02:39:00	EXetoC	I dunno but it is known
02:40:03	Araq	dtscode: title or content that you pass to genHTMLPage is nil?
02:40:11	dom96	dtscode: You need to init result inside genHtmlPage
02:40:26	*	VinceAddons quit (Read error: Connection reset by peer)
02:40:28	dom96	result = "" on line 3
02:40:31	dom96	in the .tmpl
02:41:20	Araq	ah dom96 is right
02:41:30	Araq	and I should sleep
02:41:32	Araq	good night
02:41:40	dtscode	ah, forgot to do that. thanks :)
02:50:21	*	skyfex quit (Read error: Connection reset by peer)
02:50:22	*	JinShil quit (Read error: Connection reset by peer)
02:50:58	*	skyfex joined #nim
02:52:18	*	vendethiel quit (Ping timeout: 265 seconds)
02:53:35	*	JinShil joined #nim
02:54:54	*	vendethiel joined #nim
02:56:57	*	darkf joined #nim
03:01:47	*	keyle quit (Quit: leaving)
03:03:08	gmoore	dom96: if you're still here i had a debugging question. i've been trying nim out this week and really like it so far. one of the things ill need to evaluate shortly is how well i can debug a runtime problem. we're a microsoft shop so everything needs to play nice in MSVC more or less.
03:03:55	gmoore	what's the workflow for "oh no my program is doing weird things" for nim code? is there a way to get a sane mapping of the generated C code to your lines of nim and walk through line by line?
03:03:55	dom96	i'm still here
03:04:37	dom96	The compiler can generate line numbers for debuggers like gdb via the --lineDir:on option.
03:06:02	*	kapil__ joined #nim
03:07:00	dom96	good night
03:07:40	gmoore	ah ok so embedded #line is the primary way to debug the running C code then. i believe msvc does support that. i'll give it a shot tomorrow. thanks!
03:08:49	flaviu	gmoore: I do --debuginfo --linedir:on
03:08:55	flaviu	and then it runs fine in GDB, excluding some mangling.
03:10:03	gmoore	do you find that you're able to get enough info back to debug easily? i'm exploring nim as an alternative to writing C code when we don't need to, so everyone is very familiar with C, but the generated code is pretty unfriendly looking
03:12:58	flaviu	Easily enough for me. You have to use C syntax in print and some of the local variables are named LOC1, LOC2, ...
03:13:29	flaviu	Structure members are normally named
03:13:36	gmoore	so i take it there is a bit of a learning curve in deciphering the generated code
03:13:45	flaviu	yep, definitely.
03:14:09	flaviu	but GDB shows you the nim code and nim line numbers
03:14:24	gmoore	imo i'd rather spend a day writing some crap and 2 more days debugging it than spend an entire week just to do the code writing in C. just a matter of convincing my team, hah
03:14:36	gmoore	right ill have to see if i can get the same functionality out of msvc
03:15:42	*	vendethiel quit (Ping timeout: 245 seconds)
03:17:05	*	vendethiel joined #nim
03:19:00	*	gunn_ joined #nim
03:20:17	*	gunn quit (Ping timeout: 245 seconds)
03:33:40	*	NimrodBot joined #nim
03:34:58	*	NimrodBot quit (Remote host closed the connection)
03:35:01	*	Demon_Fox quit (Ping timeout: 272 seconds)
03:38:56	*	brson quit (Ping timeout: 244 seconds)
03:39:39	*	vendethiel quit (Ping timeout: 256 seconds)
03:40:17	*	johnsoft quit (Ping timeout: 245 seconds)
03:40:53	*	johnsoft joined #nim
03:41:59	*	Mimbus joined #nim
03:51:24	*	yonson quit (Read error: Connection reset by peer)
03:54:18	ldlework	flaviu: can you write up a short thing on how you do debugging?
04:01:09	flaviu	ldlework: Sure, I'll do it tomorrow.
04:21:11	dtscode	just dont write bugs
04:22:03	fowlmouth	If debugging is the process of removing bugs, then programming must be the process of putting them in.
04:22:55	dtscode	hmmm... i dont like this default font my site uses. i might fix that
04:33:06	*	JinShil quit (Quit: Konversation terminated!)
04:53:17	*	ARCADIVS joined #nim
05:00:26	dtscode	is aporia still being worked on?
05:04:39	EXetoC	dtscode: slowly
05:05:02	dtscode	damn :/
05:09:36	*	gunn_ quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
05:11:14	*	lws joined #nim
05:12:08	fowlmouth	4 npc sidewinders calling me a dreg
05:21:51	*	vendethiel joined #nim
05:23:13	dtscode	hrmmmm jester doesnt want to see my favicon...
05:24:18	ldlework	hi lws welcome :)
05:25:00	lws	hello
05:25:13	lws	ldlework: Don't get too excited, I'm not a convert yet :D
05:25:25	dtscode	thats what they all say
05:26:11	dtscode	then they see the UFC and they are like ZOMG GET ME NIM NOW
05:26:18	lws	UFC?
05:26:25	dtscode	uniform function call
05:26:32	lws	What is that?
05:26:40	dtscode	basically x(y) is the same as y.x()
05:27:09	dtscode	and, you can also do x y (ie if the function has 0 or 1 arguments, you dont need the ()'s)
05:27:14	lws	UFCS is what that's called in other systems BTW
05:27:33	fowlmouth	the S is for syntax..
05:28:13	dtscode	meh. ive seen it called ufs as well
05:28:23	ldlework	what would that stand for
05:28:34	fowlmouth	universal function call syntax
05:28:45	ldlework	that would be ufcs
05:28:50	dtscode	uniform function syntax
05:28:55	ldlework	aha
05:29:10	fowlmouth	unifunctax
05:29:43	fowlmouth	lws, what programming language do you most often use
05:29:48	dtscode	i was going to say imo its like procs/routines/functions/etc, and then you said that and now really want to call functions funkytions
05:29:54	lws	At work, or for pleasure?
05:30:04	dtscode	both
05:30:18	ldlework	lws is a D contributor
05:30:22	lws	I use to use C# for work, just started at a Go shop, and D is my hobby language
05:30:44	dtscode	>D
05:30:44	lws	contributor is being generous. Haha :)
05:30:56	fowlmouth	go shops exist? my soul just shrunk a little bit
05:31:02	ldlework	fowlmouth: Docker :/
05:31:04	lws	fowlmouth: Yes, unfortunately.
05:31:08	dtscode	>Go
05:32:03	lws	Go makes me miss C#..
05:32:20	lws	And that statement makes me really sad.
05:32:31	dtscode	i dont like go :}
05:32:35	dtscode	or D for that matter
05:32:43	ldlework	I found Nim because as a Python programmer I yearned for a nice language like C# in the FOSS space
05:32:44	dtscode	but thats just because im a n00b
05:32:45	ldlework	tada, Nim
05:33:03	lws	C# is nice because of the framework behind it
05:33:12	ldlework	That isn't the core reason no.
05:33:13	fowlmouth	lws, C# seems nice (i've only used it with visual studio though, intelliwhatever helps a lot)
05:33:14	lws	Although it's a pretty decent language in it's own right.
05:33:25	fowlmouth	i was able to jump right into it
05:33:32	ldlework	C# is nice because not only does each feature work well, all the features seem to work well together.
05:34:15	ldlework	intellisense does help a lot
05:34:34	ldlework	I tried to get into Mono .. but meh
05:34:35	lws	The reason all those features work well is that the standard library is tightly coupled to the language...
05:34:55	dtscode	can someone tell me why my favicon isnt working? https://github.com/DTSCode/dtscodeio
05:35:05	lws	The language syntax is aware of specific interfaces in .NET and if you support those, you get lots of nice syntactic sugar.
05:35:49	fowlmouth	dtscode, put http in the href or use a relative url (/images/favicon.ico)
05:36:05	fowlmouth	is my guess
05:36:10	dtscode	hmmm ill try that
05:37:09	lws	dtscode: FYI: http://www.drdobbs.com/cpp/uniform-function-call-syntax/232700394
05:37:21	lws	I'm pretty sure that article coined the term. Note the author.
05:37:50	dtscode	lws, im not saying im write. im just saying ive seen it called all 3 before.
05:37:53	dtscode	slang if you will
05:38:34	lws	dtscode: I'm not arguing with you, I was pointing out the inventor of the concept.
05:38:53	dtscode	ag
05:39:00	dtscode	ah*
05:39:05	dtscode	thanks for the article btw
05:39:26	dtscode	fowlmouth, didnt work :/
05:39:51	lws	dtscode: I don't blame you for not liking D. There's a lot of really cool stuff in the language and some of it was sort of "discovered by accident" such as UFCS. I like that Nim takes these ideas and designs with many of them in mind from the start.
05:40:42	lws	I would actually encourage you guys to hang out in the D usenet forum for reference. Because some of the features actually are causing a lot of heated debates due to bugs they cause in code.
05:40:56	lws	It would be good to be aware of those types of things when making those features first class citizens.
05:41:52	lws	e.g. Optional () around functions that take no parameters actually causes ambiguity when you want to refer to the function itself.
05:41:59	fowlmouth	dtscode, this says you should use rel="icon" http://www.w3.org/2005/10/howto-favicon
05:42:33	lws	Caused a lot of debate in the D forums. Optional () are going to be deprecated in D because of it, and they will only apply to @property methods and then they will not be allowed.
05:43:41	ldlework	.eval proc foo() = echo "hello"; foo
05:43:44	Mimbus	ldlework: eval.nim(4, 0) Error: value of type 'proc ()' has to be discarded
05:44:07	ldlework	wut
05:44:13	*	vendethiel quit (Ping timeout: 272 seconds)
05:44:25	ldlework	.eval proc foo() = echo "hello";foo
05:44:28	Mimbus	ldlework: eval.nim(4, 0) Error: value of type 'proc ()' has to be discarded
05:44:33	fowlmouth	nim requires parens if there are no arguments
05:44:43	ldlework	ah I see, it wants me to discard foo itself
05:45:09	dtscode	fowlmouth, oops my mistake. i dont know why i thought it was 0 or 1 arguments
05:45:09	lws	fowlmouth: ? Parents are option if there ARE arguments?
05:45:16	lws	parens*
05:45:21	dtscode	lws, only if there is one
05:45:27	dtscode	ie:
05:45:54	dtscode	.eval proc foo(x: int) echo x; foo 1; foo(2);
05:45:57	Mimbus	dtscode: eval.nim(3, 21) Error: invalid indentation
05:46:05	ldlework	no spaces
05:46:10	dtscode	.eval proc foo(x: int) echo x; foo 1;foo(2);
05:46:13	Mimbus	dtscode: eval.nim(3, 21) Error: invalid indentation
05:46:13	dtscode	oops
05:46:16	lws	.eval proc foo(x: string) = echo "hello"; foo("hello")
05:46:17	fowlmouth	lws, yes but only for the outtermost expression
05:46:20	Mimbus	lws: hello
05:46:27	fowlmouth	.eval echo 1,2,3
05:46:28	dtscode	.eval proc foo(x: int) = echo x; foo 1;foo(2);
05:46:30	lws	.eval proc foo(x: string) = echo "hello"; "hello".foo
05:46:30	Mimbus	fowlmouth: 123
05:46:33	Mimbus	dtscode: 1
05:46:36	Mimbus	lws: hello
05:46:44	lws	.eval proc foo(x: string) = echo "hello"; echo &"hello".foo
05:46:46	Mimbus	lws: eval.nim(4, 13) Error: expression 'foo("hello")' has no type (or is ambiguous)
05:46:56	dtscode	& is string catting in nim
05:46:56	lws	.eval proc foo(x: string) auto = return "hello"; echo &"hello".foo
05:46:59	Mimbus	lws: eval.nim(3, 24) Error: invalid indentation
05:47:53	lws	.eval proc foo(x: string):string = return "hello"; echo "hello".foo
05:47:56	Mimbus	lws: hello
05:47:56	lws	how do I take an address?
05:48:19	fowlmouth	lws, addr, but you couldn't do it that way
05:48:23	lws	.eval proc foo(x: string):string = return "hello"; echo addr "hello".foo
05:48:26	Mimbus	lws: eval.nim(4, 5) Error: expression has no address
05:48:36	lws	wah?
05:49:06	lws	Ah, I need to make foo a method on a struct to see the behavior
05:49:08	fowlmouth	.eval var x = "hello"; x.addr.echo
05:49:10	Mimbus	fowlmouth: eval.nim(4, 2) Error: type mismatch: got (ptr string)
05:49:17	fowlmouth	.eval var x = "hello"; x.addr.repr.echo
05:49:20	Mimbus	fowlmouth: ref 0x645b20 --> 0x7f6ae4543050"hello"
05:49:20	lws	.eval proc foo(x: string):string = return "hello"; echo addr "hello".foo.addr
05:49:23	Mimbus	lws: eval.nim(4, 22) Error: expression has no address
05:49:27	lws	.eval proc foo(x: string):string = return "hello"; echo "hello".foo.addr
05:49:30	Mimbus	lws: eval.nim(4, 17) Error: expression has no address
05:49:38	lws	Interesting
05:49:47	lws	How does nimbus work? Is it compiling to javascript or something?
05:49:51	lws	mimbus
05:50:23	ldlework	lws: why do you think it must be compiling to js?
05:50:41	lws	There's a direct nim interpreter?
05:51:14	ldlework	lws: Nim compiles very fast
05:51:32	fowlmouth	lws, you can only get address of an l-value, not an r-value, and inside the proc you can't get x's address
05:51:53	lws	ldlework: Seems like a bad idea to compile nim to C and C -> asm and then execute it and give me the result here?
05:52:13	lws	What's it doing, firing up a container real quick to sandbox it too?
05:52:21	ldlework	it is sandboxed, yes
05:53:33	ldlework	lws: try to write something
05:55:58	lws	.eval echo system.open("/dev/urandom").readLine
05:56:01	Mimbus	lws:
05:56:13	lws	:(
05:56:26	lws	.eval echo system.open("/etc/hosts").readLine
05:56:29	Mimbus	lws: #
05:56:36	lws	.eval echo system.open("/etc/shadow").readLine
05:56:39	Mimbus	lws: Error: unhandled exception: cannot open: /etc/shadow [IOError]
05:56:43	lws	:(
05:56:46	lws	lol
05:59:51	lws	.eval echo osproc.execCmd("mount")
05:59:54	Mimbus	lws: eval.nim(3, 5) Error: undeclared identifier: 'osproc'
05:59:56	lws	.eval osproc.execCmd("mount")
05:59:59	Mimbus	lws: eval.nim(3, 0) Error: undeclared identifier: 'osproc'
06:00:08	lws	.eval import osproc; osproc.execCmd("mount")
06:00:11	Mimbus	lws: eval.nim(4, 14) Error: value of type 'int' has to be discarded
06:00:18	lws	.eval import osproc; echo osproc.execCmd("mount")
06:00:22	Mimbus	lws: proc on /proc type proc (rw,nosuid,nodev,noexec,relatime)
06:00:22	*	Mimbus quit (Excess Flood)
06:00:29	*	Mimbus joined #nim
06:01:15	*	gunn joined #nim
06:02:03	lws	.eval import osproc; var x, y = osproc.execCmdEx("mount"); echo x
06:02:07	Mimbus	lws: (output: proc on /proc type proc (rw,nosuid,nodev,noexec,relatime)
06:02:07	*	Mimbus quit (Excess Flood)
06:02:13	*	Mimbus joined #nim
06:03:31	lws	.eval import osproc; import strutils; var x = osproc.execCmdEx("mount"); echo x.output.splitLines[2]
06:03:35	Mimbus	lws: dev on /dev type devtmpfs (rw,nosuid,relatime,size=4052072k,nr_inodes=1013018,mode=755)
06:03:39	lws	mwahaha
06:03:43	lws	.eval import osproc; import strutils; var x = osproc.execCmdEx("mount"); echo x.output.splitLines[3]
06:03:46	Mimbus	lws: run on /run type tmpfs (rw,nosuid,nodev,relatime,mode=755)
06:03:49	lws	.eval import osproc; import strutils; var x = osproc.execCmdEx("mount"); echo x.output.splitLines[4]
06:03:52	Mimbus	lws: /dev/mapper/linux on / type btrfs (rw,nodev,relatime,compress=lzo,space_cache)
06:04:36	lws	Seems like a neat language
06:05:03	lws	Very easy to get some stuff working
06:05:57	lws	ldlework: Is mimbus your bot?
06:06:04	ldlework	lws: no dts
06:06:31	ldlework	dtscode
06:07:01	lws	It seems like it'd be possible to overwrite mimbus with a .eval though
06:07:09	lws	Which would just be annoying I suppose
06:07:20	dtscode	hmmmm?
06:07:28	lws	I assume the bot is running in the same container?
06:07:31	dtscode	.eval and Mimbus: do the same thing
06:07:34	Mimbus	dtscode: eval.nim(3, 3) Error: expression expected, but found 'keyword and'
06:07:39	ldlework	hehe
06:07:44	dtscode	oh you tricked me
06:07:49	dtscode	im suing
06:07:56	ldlework	dtscode: he's saying that you could get Mimbus to overwrite itself
06:08:01	dtscode	how so?
06:08:17	lws	Is mimbus in the same container as what it's evaling?
06:08:22	ldlework	dtscode: is Mimbus running in a Docker container?
06:08:47	fowlmouth	.eval echo(hostOS, " ", hostCPU)
06:08:51	Mimbus	fowlmouth: linux amd64
06:08:58	dtscode	ldlework, not yet. its kind of on the backburner right now due to more important projects
06:09:07	ldlework	just wondering
06:09:09	dtscode	lws, no its not
06:09:39	lws	.eval echo system.open("/proc/cgroup").readLine
06:09:42	Mimbus	lws: Error: unhandled exception: cannot open: /proc/cgroup [IOError]
06:09:54	lws	.eval echo system.open("/proc/self/cgroup").readLine
06:09:57	Mimbus	lws: 8:memory:/
06:10:45	lws	.eval var f = system.open("/proc/self/cgroup"); f.readLine; echo f.readLine
06:10:48	Mimbus	lws: eval.nim(4, 1) Error: value of type 'TaintedString' has to be discarded
06:11:03	lws	.eval var f = system.open("/proc/self/cgroup"); var x = f.readLine; echo f.readLine
06:11:06	Mimbus	lws: 7:devices:/user.slice
06:11:31	lws	memory is not capped?
06:11:49	fowlmouth	can we move the Mimbus spamming to #nim-offtopic
06:11:55	dtscode	uhhh no. it looks like flaviu is running a very early release
06:11:55	lws	Sorry :)
06:45:04	lws	dtscode: Early release of?
06:48:02	vegai	3 overlapping ways to make function calls?
06:48:18	vegai	oops, sorry. Commented on something in the backlog
06:55:13	dtscode	lws, Nimbus
07:08:55	dtscode	fowlmouth, that didnt work eitehr
07:09:04	dtscode	Araq, dom96 do either of you have an idea?
07:11:53	*	nande quit (Remote host closed the connection)
07:36:57	*	vendethiel joined #nim
07:39:03	*	gour joined #nim
07:42:25	*	Demon_Fox joined #nim
07:52:00	*	SplinterOfChaos quit (Ping timeout: 264 seconds)
07:58:48	*	valberg quit (Changing host)
07:58:48	*	valberg joined #nim
07:59:51	*	vendethiel quit (Ping timeout: 244 seconds)
08:08:02	*	rkj-b joined #nim
08:08:30	*	BlaXpirit joined #nim
08:12:48	*	rkj-b quit (Quit: ChatZilla 0.9.91.1 [Firefox 34.0.5/20141126041045])
08:20:56	*	gsingh93 quit (Quit: Connection closed for inactivity)
08:34:46	*	Sphax joined #nim
08:41:44	ekarlso	c2nim --header foo.h -o foo isn't that correct ?
08:44:08	*	Sphax quit (Quit: ZZZZZzzzzz)
08:44:43	EXetoC	no, you're missing the inputfile parameter as you can see in the help output
08:45:29	dtscode	any day now linus will be finding out about c2nim and will be like "Fuck yeah. i can port the kernel to nim now."
08:48:00	ekarlso	EXetoC: so you pass it .c files ? ^
08:49:15	ekarlso	ahh
08:49:16	ekarlso	nice
08:49:49	EXetoC	headers too
08:50:59	*	Sphax joined #nim
09:07:46	*	vendethiel joined #nim
09:11:12	*	Trustable joined #nim
09:14:51	*	seemsindie joined #nim
09:30:25	*	vendethiel quit (Ping timeout: 256 seconds)
09:33:49	*	EXetoC quit (Ping timeout: 252 seconds)
09:45:38	*	vendethiel joined #nim
09:45:55	BlaXpirit	just realized it's impossible to extend `re` library to have the functionality I need
09:45:57	BlaXpirit	sigh
09:46:02	*	Demon_Fox quit (Quit: Leaving)
09:50:09	ekarlso	hmm, is there a way to have a object that inherits from Table ?
09:50:19	ekarlso	like Claims = Table[string, JsonNode] ?
09:50:52	ekarlso	I tried that but then when I do var claims = Claims(); claims["foo"] = %"foo" it goes invalid storage
09:51:01	BlaXpirit	ekarlso, heh, i wanted almost exactly the same thing yesterday
09:51:06	BlaXpirit	but i wanted to inherit from seq
09:51:21	BlaXpirit	ah ekarlso, wait
09:51:27	BlaXpirit	you actually can extend from it?
09:51:34	BlaXpirit	you probably need to call "new"
09:51:37	BlaXpirit	or newtable
09:51:39	BlaXpirit	or whatever it is
09:51:50	BlaXpirit	show code tho
09:51:54	BlaXpirit	always show code
09:53:07	BlaXpirit	ok whatever
09:54:06	ekarlso	BlaXpirit: was getting coffee :p
09:54:19	ekarlso	https://bpaste.net/show/03debcdf211a
09:54:43	BlaXpirit	ekarlso, basically my rule of thumb is
09:54:48	BlaXpirit	never create objects directly
09:54:55	BlaXpirit	always use a constructor function
09:55:53	BlaXpirit	TableRef... sigh
09:55:57	ekarlso	BlaXpirit: yeah, was wondering if I should just proc newclaims that returned a table..
09:56:24	BlaXpirit	but it doesn't look like Table is a ref object, so "new" is probably not needed for it itself
09:56:44	BlaXpirit	however, it contains a seq, which does need to be initialized
09:56:53	BlaXpirit	basically call v
09:56:55	BlaXpirit	inittable
09:57:26	ekarlso	?
09:57:29	BlaXpirit	ekarlso, oh to generalize a bit more
09:57:38	BlaXpirit	this type you created is not the problem here
09:57:47	BlaXpirit	you wouldn't be able to use a normal table like this too
09:57:58	BlaXpirit	because you should not create it directly!
09:58:01	BlaXpirit	initTable is the thing
09:58:41	ekarlso	I guess I can just skip declaring the claims type vs just creating a new table
09:59:57	ekarlso	yap, worked wonderfully
09:59:59	ekarlso	:D
10:00:16	ekarlso	it's allmost stupid how easy nim is sometimes..
10:00:40	BlaXpirit	easy?
10:00:50	BlaXpirit	i wouldn't say it's so easy
10:00:59	ekarlso	BlaXpirit: I mean, for me coming from a dynamic languages background vs learning C
10:16:21	*	VinceAddons joined #nim
10:58:24	*	matkuki joined #nim
10:59:20	*	ARCADIVS quit (Quit: ARCADIVS)
10:59:57	*	JinShil joined #nim
11:15:32	*	akir joined #nim
11:16:50	akir	Hi, a username "akiradeveloper" is kicked by ChanServ. Did I do anything wrong?
11:18:03	akir	plz let "akiradeveloper" join this chat. I am not sure why I am kicked out
11:18:39	*	gour_ joined #nim
11:18:48	*	gour quit (Read error: Connection reset by peer)
11:19:44	*	jm116_ quit (Quit: Leaving)
11:20:51	akir	I am going out for an hour but keep this connection. Please tell me if I did bad thing for kicking
11:22:25	*	gour_ is now known as gour
11:34:45	*	gokr quit (Ping timeout: 276 seconds)
11:41:01	BlaXpirit	ok i think i have found a bug here
11:41:13	BlaXpirit	minimal example time!
11:43:29	BlaXpirit	damn, cant reproduce
11:49:20	flaviu	akir: I checked the logs, and I see no reason why you would be banned. It's probably a mistake, glitch, or misunderstanding; hopefully dom96 can figure out how to fix it whenever he comes back.
11:51:21	ekarlso	so flaviu yes a jwt needs to be decoded on the server side..
11:51:26	ekarlso	hs256 that is.
11:51:52	ekarlso	flaviu: unless alg is none that is
11:51:59	flaviu	ekarlso: sorry, I can't stay and chat. I've got to go.
11:52:08	ekarlso	flaviu: ok :)
11:52:45	ekarlso	how do I get the compiler to know of a library path ?
12:05:28	dom96	akir: I don't see you being kicked anywhere.
12:12:33	*	Sphax quit (Quit: ZZZZZzzzzz)
12:14:09	akir	I am back now
12:14:09	*	keyle_ joined #nim
12:15:12	akir	dom96: really? I am using web IRC chatting. When I try to log into this room, ChanServ kicks me out
12:15:18	akir	I will try again
12:15:21	*	akir left #nim (#nim)
12:16:38	*	akir joined #nim
12:16:50	akir	[21:15] == akiradeveloper was kicked from #nimlang by ChanServ [Invite only channel]
12:16:56	akir	this is the log
12:17:25	dom96	Join #nim instead.
12:17:36	*	akir left #nim (#nim)
12:18:20	*	akiradeveloper joined #nim
12:18:26	akiradeveloper	Yeah
12:18:49	akiradeveloper	Thanks. Changed from nimlang to nim?
12:19:06	*	Sphax joined #nim
12:19:14	*	gokr joined #nim
12:19:18	dom96	yes
12:20:57	akiradeveloper	I see..
12:24:09	akiradeveloper	Please tell me why is this code compiled? T isn't obviously a subclass of float https://gist.github.com/akiradeveloper/2a5a6fed3246fc30a565
12:24:30	*	keyle_ is now known as keyle
12:28:35	BlaXpirit	does anyone know how to make function "any"?
12:28:42	BlaXpirit	and why it is not in standard lib
12:29:11	BlaXpirit	akiradeveloper, seems to me that ldlework brought this up yesterday
12:30:16	BlaXpirit	http://irclogs.nim-lang.org/15-01-2015.html 05:45:42
12:30:54	BlaXpirit	21:06:47 better
12:35:13	akiradeveloper	BlaXpirit: I use = instead of : but the same problem per se?
12:35:24	BlaXpirit	probably
12:35:35	BlaXpirit	i dont know this stuff
12:38:22	akiradeveloper	it's not a big thing for me and ok
12:45:51	*	BitPuffin joined #nim
12:46:13	BlaXpirit	what is up with json module
12:46:41	BlaXpirit	i hoped it would be so clever
12:46:51	BlaXpirit	to learn from it
12:50:19	BlaXpirit	how to get argv?
12:51:33	BlaXpirit	commandLineParams
12:52:00	*	gour quit (Remote host closed the connection)
12:59:32	*	gour joined #nim
12:59:45	Triplefox	Feeling pretty good about this spritesheet library... I think I've got the internal data sorted now so I'm starting work on some algorithms
12:59:49	*	gour quit (Changing host)
12:59:49	*	gour joined #nim
13:00:17	Triplefox	It's designed so that you will be able to write little scripts to do whatever is needed
13:00:39	*	EXetoC joined #nim
13:06:02	*	keyle quit (Quit: Morbo: Chitchat achieved!)
13:11:14	*	matkuki quit (Quit: ChatZilla 0.9.91.1 [Firefox 34.0.5/20141126041045])
13:14:22	*	BitPuffin quit (Ping timeout: 240 seconds)
13:19:38	gokr	Btw, for optparsing - I did tweak and get Lapp going (similar to docopt, but probably less advanced).
13:20:24	gokr	I am using it in my blimp utility. Here is Lapp: https://gitlab.3dicc.com/gokr/lapp
13:23:16	*	Demon_Fox joined #nim
13:24:02	*	BitPuffin joined #nim
13:30:13	gokr	Who controls nim-lang.com? Is it you Araq?
13:30:22	gokr	Eh, nim-lang.org I mean
13:31:45	*	jpoirier quit (Quit: Leaving)
13:33:38	BlaXpirit	type Argument = ref object of LeafPattern
13:33:47	BlaXpirit	why can't I convert Argument(leafpattern)
13:35:04	ekarlso	is it 2 spaces or 4 spaces that's wanted in .nim files ?
13:35:10	ekarlso	I see some with 2 and some with 4..
13:35:26	EXetoC	2 is more common
13:36:15	BlaXpirit	please explain to me: https://bpaste.net/show/cae091e63e53
13:38:35	BlaXpirit	what can I do to make this work?
13:40:34	BlaXpirit	pls
13:42:13	BlaXpirit	:(
13:42:19	BlaXpirit	this halts all my work
13:43:10	BlaXpirit	T_T
13:44:57	EXetoC	by keeping another reference? but I assume it shouldn't be needed. do you really want to downcast though?
13:45:20	BlaXpirit	you call it "down"?
13:45:24	BlaXpirit	yes, well, i do really want
13:46:52	EXetoC	works with a nil base object apparently :p
13:47:30	BlaXpirit	EXetoC, inheritance only works with non-final objects -__-
13:47:56	BlaXpirit	or i don't know what you mean
13:50:03	BlaXpirit	this works in c++
13:52:37	EXetoC	BlaXpirit: casting seems to work
13:52:50	BlaXpirit	but it's bad, mkay
13:52:59	BlaXpirit	what if there is non-empty memory
13:53:08	BlaXpirit	im pretty sure it writes trash to other members or something
13:53:14	BlaXpirit	although that shouldnt be a problem
13:53:24	BlaXpirit	no wait
13:53:28	BlaXpirit	writing to extra members
13:53:33	BlaXpirit	will write to uncharted territory
13:57:04	BlaXpirit	http://ideone.com/x17i0X vs https://bpaste.net/show/430a86cd7859
13:57:06	EXetoC	you might be right
14:00:17	EXetoC	but it's not a 'B' to begin with
14:01:15	EXetoC	compare your statement with this: "discard B(A(B()))", which works
14:02:06	BlaXpirit	EXetoC, but c++ does something with it
14:02:06	EXetoC	you don't want an object variant though? just asking because of the downcast. it's discouraged I think, but I don't know what your code is doing
14:02:25	*	BlaXpirit is trying to port a Python library to Nim without understanding it
14:02:25	ekarlso	hmm, can you go from varargs > seq ?
14:02:32	BlaXpirit	@
14:02:40	EXetoC	right, because it's a static cast, and not a dynamic cast
14:02:45	BlaXpirit	facepalm
14:02:55	EXetoC	well
14:03:01	BlaXpirit	so what if i want a static cast
14:03:17	BlaXpirit	c++ crashes, makes sense
14:03:36	akiradeveloper	I have two files a.nim b.nim. b is the main file and a has library function. What compile option do I need to import a from b?
14:03:41	BlaXpirit	crashes on dynamic_cast, that is
14:03:47	EXetoC	then you cast like I said, which I'm not encouraging now that I know a little more about this, and hope that it doesn't explode
14:03:56	BlaXpirit	akiradeveloper, why not import or include it??
14:04:13	EXetoC	yes, and B(A()) can be considered a failed dynamic cast
14:04:21	EXetoC	there's also the "x of y" test
14:04:25	*	gokr_ joined #nim
14:04:32	BlaXpirit	ok, EXetoC, then let me put this differently
14:04:33	*	askatasuna joined #nim
14:04:55	akiradeveloper	BlaXpirit: what do you mean? I want to import a.nim from b.nim
14:04:58	BlaXpirit	how to inherit constructor of ref object?
14:05:14	BlaXpirit	akiradeveloper, then write import a
14:05:55	akiradeveloper	We don't need to create .o file like in C?
14:05:58	EXetoC	you must explicitly upcast and then pass it to the base "constructor", which is one aspect that OOP macros simplify I think
14:06:07	BlaXpirit	EXetoC, https://bpaste.net/show/39b65bf105f6
14:06:30	EXetoC	that's what I recall for my brief use of inheritance over a year ago
14:06:48	BlaXpirit	this is one of the problematic places
14:06:57	akiradeveloper	yeah succeed, thanks
14:07:36	EXetoC	BlaXpirit: I think you must start by constructing the leaf class
14:07:55	EXetoC	since it does indeed encompass the base type as well
14:08:10	akiradeveloper	maybe, I am ready to implement nim-fuse. since I am newbie, it will be a long fight.
14:08:59	EXetoC	this can be tested by adding a field to each class and then stringifying a base class object (dereference first)
14:09:56	BlaXpirit	D:
14:10:00	BlaXpirit	I don't understand
14:10:49	*	akiradeveloper left #nim (#nim)
14:11:12	BlaXpirit	but don't try too hard to explain, i just might not need this all stuff
14:11:25	BlaXpirit	just gonna rewrite constructors, big deal
14:11:33	EXetoC	hence why "B(A(B()))" works, but "B(A())" doesn't, because the runtime type of the object passed to B in the latter case is an A and not a B
14:11:39	BlaXpirit	and hope it is not anywhere else in the code
14:11:52	BlaXpirit	EXetoC, well, yes, that totally makes sense
14:12:05	BlaXpirit	thank you
14:13:06	BlaXpirit	next up, attempt to read from nil
14:16:10	*	jefus__ is now known as jefus
14:23:23	ekarlso	https://bpaste.net/show/b627c94b1743 < does that API seem fair ?
14:25:57	*	jefus_ joined #nim
14:28:54	*	jefus quit (Ping timeout: 244 seconds)
14:32:52	ekarlso	https://bpaste.net/show/0fbb41758be5 < shouldn't #72 work there ? claims.nim(72, 24) Error: type expected
14:35:05	*	jefus_ is now known as jefus
14:35:23	EXetoC	EXP is not a type. perhaps you want T:static[ClaimKind]. it might or might not work, as static is experimental
14:35:37	EXetoC	but why is it not a normal parameter?
14:36:46	EXetoC	the variant type can be selected at run-time
14:37:02	ekarlso	good call
14:37:11	EXetoC	it can be done statically by making the type generic as well
14:37:23	*	Sphax quit (Quit: ZZZZZzzzzz)
14:38:20	ekarlso	that worked more nicely EXetoC :)
14:38:21	*	Sphax joined #nim
14:38:35	EXetoC	c(:)|<
14:38:51	*	sillesta joined #nim
14:40:48	*	gokr_ quit (Ping timeout: 245 seconds)
14:42:14	*	Demos joined #nim
14:46:51	*	Demos quit (Ping timeout: 264 seconds)
14:51:04	Araq	gokr: yes. why?
14:51:15	*	gokr_ joined #nim
14:57:19	Araq	BlaXpirit: I'm willing to do something about negative indexing (never liked it too much in the first place), but less noise from you can only help you as I'm about to /ignore you.
14:59:54	gokr	Hey guys
15:00:06	gokr	Araq: I am itching to set up a planet for Nim
15:01:57	Araq	gokr: nice :D
15:07:53	*	Guest51690 is now known as adam12
15:33:14	novist	a planet?
15:34:52	*	SplinterOfChaos joined #nim
15:44:53	*	faineance joined #nim
15:49:25	gokr	novist: Yeah. A blog aggregator. At say planet.nim-lang.org
15:49:50	gokr	I think I will use the Pluto tool - it seems the most maintained variant these days.
15:50:06	gokr	Will set one up this weekend I think.
15:50:21	novist	never heard of such things ^_^
15:50:42	*	nimnoob joined #nim
15:50:51	gokr	There are TONS of them: http://www.planetplanet.org
15:51:16	gokr	Here is the one for the Mono people: http://www.go-mono.com/monologue
15:52:16	gokr	And here is one using the software I will use: http://planetruby.herokuapp.com
15:53:34	novist	last one reminds me of hackernews
15:55:20	EXetoC	is there a planet-planet-planet too?
15:56:58	*	jpoirier joined #nim
15:59:18	*	faineance quit (Quit: leaving)
16:00:13	*	darkf quit (Quit: Leaving)
16:00:38	*	faineance joined #nim
16:01:35	*	kapil__ quit (Quit: Connection closed for inactivity)
16:04:45	*	nimnoob quit (Remote host closed the connection)
16:08:51	*	MajorWork joined #nim
16:20:39	*	nimnoob joined #nim
16:22:44	*	brson joined #nim
16:23:29	*	UberLambda joined #nim
16:33:17	*	nimnoob quit (Ping timeout: 252 seconds)
16:43:22	*	seemsindie quit (Ping timeout: 240 seconds)
16:54:36	*	BlaXpirit quit (Read error: Connection reset by peer)
16:54:59	*	BlaXpirit joined #nim
17:05:27	*	gokr_ quit (Ping timeout: 264 seconds)
17:29:35	*	JinShil quit (Quit: Konversation terminated!)
17:29:50	*	Sphax quit (Quit: CYA!!)
17:45:48	*	gmpreussner___ joined #nim
17:49:37	*	gmpreussner___ is now known as gmpreussner
17:53:17	*	faineance quit (Remote host closed the connection)
17:54:44	*	gsingh93 joined #nim
17:54:55	*	UberLambda quit (Quit: Leaving the Matrix)
17:55:54	*	MajorWork quit (Quit: Leaving)
18:23:46	def-	Is this the first time Nim(rod) appeared on this Programming Language Rankings?: http://redmonk.com/sogrady/2015/01/14/language-rankings-1-15/
18:25:38	*	yonson joined #nim
18:25:48	gmpreussner	def-: wohoo... gonna be number one soon
18:26:13	gmpreussner	Araq: maybe i'm doing it wrong, but the endians module seems rather inconvenient to use since it requires pointers
18:26:41	gmpreussner	i often have proc parameters (int16 etc.) that need to be converted
18:27:01	gmpreussner	but the addr operator doesn't accept those, so i have to copy to local vars first?
18:27:12	Araq	yes
18:27:45	Araq	well it just needs some more convenience procs
18:28:22	gmpreussner	if you tell me what they should be, i can add them and submit a PR
18:28:31	gmpreussner	but i'd like it to not require an extra copy
18:29:02	ekarlso	how would u go about doing something like >https://github.com/square/go-jose/blob/master/shared.go#L79 in nim +
18:29:15	gmpreussner	especially since the values will just be passed through a lot of times
18:30:00	Araq	gmpreussner: unsafeAddr is planned to allow taking the address of a parameter but for scalars at least the copy is optimized away anyway
18:30:08	def-	ekarlso: type SignatureAlgorithm = distinct string?
18:30:09	gmpreussner	ekarlso: type SignatureAlgorithm = string
18:30:14	ldlework	def-: I don't see Nim on that page
18:30:22	gmpreussner	or distinct, yes
18:30:31	def-	ldlework: it's on the picture, bottom left corner, as Nimrod
18:30:44	gmpreussner	Araq: ok, i will copy then
18:50:49	dom96	def-: nope
18:51:17	*	yonson quit (Quit: Leaving)
18:59:23	EXetoC	Araq: as in a parameter that is not annotated with 'var'?
19:00:31	Araq	EXetoC: yes exactly
19:01:29	EXetoC	great
19:09:59	EXetoC	anyone using nim for DSP?
19:11:30	gmpreussner	how do i properly do pointer arithmetic? i have to convert some C crap... i need to add sizeof(SomeType)' to some 'ptr uchar' and return the new 'ptr uchar'
19:12:31	BlaXpirit	gmpreussner, i did something here https://github.com/BlaXpirit/nim-csfml/blob/55b09f6c3251cd963790bcd71c217b03b6534b77/src/csfml_graphics.nim#L217
19:13:00	BlaXpirit	gmpreussner, basically u cast to int, then add, then cast back
19:13:10	gmpreussner	ah ok, cool thx
19:13:15	BlaXpirit	there are other, maybe better ways
19:13:23	BlaXpirit	if you plan to extensively use pointer arithmetic
19:13:30	*	Varriount|Busy joined #nim
19:13:31	BlaXpirit	there was a library https://github.com/fowlmouth/nimlibs/blob/master/fowltek/pointer_arithm.nim
19:13:35	gmpreussner	nah, it's just a few one offs
19:13:48	Varriount|Busy	Hi guys!
19:13:54	gmpreussner	o hai
19:17:30	*	rpag joined #nim
19:38:51	*	Guest70720 quit (Quit: WeeChat 1.0.1)
19:44:10	*	Demos joined #nim
19:46:36	*	sampwing joined #nim
19:48:12	*	joecodes joined #nim
19:56:25	gmpreussner	can someone tell me what this C syntax is? http://pastebin.com/mJJrZKiM
19:56:30	gmpreussner	is 'bar' a dynamic array?
19:56:40	gmpreussner	how would i translate this to nim?
19:59:01	*	Fusion_Gaming is now known as FusionGaming
20:00:36	Araq	gmpreussner: .unchecked pragma, look it up, it's documented
20:01:06	Araq	it's an array of unspecified length, valid only as last element in a struct
20:01:29	gmpreussner	interesting, i have never seen that before
20:01:49	gmpreussner	thanks Araq, the docs have what i need
20:05:46	*	nande joined #nim
20:09:07	*	NimBot joined #nim
20:11:59	dtscode	hey dom96... i did as you said, and moved the image to a public folder (so now its under $SiteRoot/public/images), but the favicon still isnt showing up
20:36:58	BlaXpirit	dtscode, i can look at it if it isn't something uncommon
20:45:33	*	sampwing quit (Ping timeout: 252 seconds)
20:49:17	*	Mat4 joined #nim
20:49:31	Mat4	hello
20:49:39	Varriount|Busy	Hi Mat4
20:50:01	Mat4	Hi Varriount, what's new ?
20:51:02	Varriount|Busy	Mat4: Well, I'm writing a string based linked-list of strings, using xor and tags to save memory
20:51:19	Varriount|Busy	It's interesting work.
20:51:37	dtscode	BlaXpirit, ok. give me a sec to update the repo
20:53:14	Mat4	Varriount|Busy: Reads nice
20:53:32	Mat4	I'm fighting with the bochs emulator and its debugger today
20:54:15	Mat4	(it's really a primitive monitor intended for OS development)
20:54:25	dtscode	https://github.com/DTSCode/dtscodeio BlaXpirit
20:55:35	BlaXpirit	ok, gonna pm you
20:56:00	ekarlso	do we have something like play.rust-lang.org ?
20:56:19	Varriount|Busy	ekarlso: Not really.
20:56:25	ekarlso	-,,-
20:56:30	ekarlso	maybe something to create :D
20:57:29	flaviu	dtscode: Do you have a new version of Nimbus?
20:58:05	dtscode	flaviu, no not yet :/ its kind of on the backburner right now.
20:59:30	flaviu	dtscode: Ok. I saw "dtscode | uhhh no. it looks like flaviu is running a very early release" in the logs, is that something I should fix or.. ?
21:00:04	*	yonson joined #nim
21:00:16	*	wan joined #nim
21:01:20	*	gour left #nim ("Leaving")
21:01:52	dtscode	it doesnt do gists
21:02:05	dtscode	which is why i thought that
21:02:11	ekarlso	Mimbus: is irc only atm no ?
21:02:27	dtscode	what else would it be?
21:02:39	flaviu	gokr: lapp looks cool, can you add it to nimble? It might also be a good idea to paste the example in the README so it's readily accessible.
21:02:48	flaviu	ekarlso: It's very tightly coupled to IRC.
21:02:59	ekarlso	shouldn't be that hard to make it be both..
21:03:08	ekarlso	dtscode: do you do somekind of isolation atm with it ?
21:03:29	flaviu	ekarlso: I would be the correct person to ask that, and I've set it up in it's own linux account, with ulimit.
21:03:31	dtscode	ekarlso, not yet. flaviu has managed it. when i get back to it nimbus will run with docker
21:03:53	ekarlso	hmm, does docker have cbindings ?
21:03:59	*	askatasuna quit (Ping timeout: 245 seconds)
21:04:00	dtscode	what other platform would nimbus run on?
21:04:05	dtscode	cbindings?
21:04:16	flaviu	I've set filesystem permissions so that it can't access anything sensitive, and ulimit manages memory usage and CPU time.
21:05:34	*	rpag quit (Quit: Leaving)
21:09:13	ekarlso	dtscode: yeah if docker has bindings for C
21:09:15	ekarlso	..
21:09:21	ekarlso	so you could call docker from nim
21:09:58	dtscode	no. thats not what docker really is
21:10:06	dtscode	its a program not a library
21:11:53	gokr	flaviu: Yeah, I will add lapp to nimble. I also use it in blimp, which I also should add to nimble I guess.
21:12:40	dtscode	what is lapp?
21:12:59	*	dtscode is guessing it has something to do with smalltalk
21:13:25	gokr	https://gitlab.3dicc.com/gokr/lapp
21:13:49	gokr	And no, its a Nim library. I just revived it, was written by Steve Donovan (Lua fame) - he ported it from Lua.
21:14:10	dtscode	ooo this looks nice
21:14:17	gokr	The idea is to let the help text for a command line utility double as the parser grammar.
21:14:31	gokr	I tweaked it a bit also.
21:14:37	gokr	I use it in blimp
21:14:43	*	sampwing joined #nim
21:14:51	gokr	https://gitlab.3dicc.com/gokr/blimp
21:14:51	dtscode	why is it called LAPP?
21:14:56	gokr	Dunno :)
21:15:16	gokr	I checked with Steve and he was fine with me taking over that code.
21:15:41	dom96	gokr: add it to nimble :)
21:15:58	gokr	It can be improved - but its already quite nice.
21:16:13	gokr	And blimp is a utility I wrote in Nim to handle big files in git.
21:16:32	gokr	Its very similar to git-fat.
21:17:13	*	irrequietus joined #nim
21:17:14	gokr	We use blimp for some of our own git repos and it works great.
21:17:44	gokr	dom96: Yeah, I will.
21:18:30	dom96	gokr: I can do it for you if you'd like :)
21:18:41	gokr	dom96: Yeah, sure
21:18:43	ekarlso	would be easier with pkgs :P
21:18:49	Varriount|Busy	Hm.. What is that hack that you can use to change the discriminant of an object variant?
21:18:50	ekarlso	nimble register foo
21:19:01	Araq	dom96: can you also add the compiler's source code as a nimble package for me?
21:19:38	Araq	Varriount|Busy: assign it the new value
21:19:45	dom96	Araq: perhaps
21:19:54	dom96	Araq: I'm assuming this is for c2nim?
21:19:54	Araq	but if debug mode complains then you really MUST not do it
21:20:01	Araq	yup
21:20:02	gokr	dom96: I actually made nimble files for both those utilities.
21:20:08	Varriount|Busy	Araq: Why not?
21:20:29	dom96	gokr: Want me to add blimp too?
21:20:38	gokr	Yeah, go ahead
21:20:49	Araq	Varriount|Busy: because you change how the GC sees the data structure, you will get nice memory corruptions
21:21:28	Varriount|Busy	Araq: Oh... yay
21:22:34	*	shodan45 joined #nim
21:23:04	Varriount|Busy	Araq: What if the object variant is, at bottom, just a struct with an int8 and an int64 member?
21:23:09	Varriount|Busy	Araq: https://gist.github.com/Varriount/ea6fec5a0558e81b433f
21:25:05	Araq	Varriount|Busy: you can modify the discriminant with pointer hacking. that said you data structure is completely pointless as it's always 16 bytes.
21:25:16	Araq	*your
21:25:30	Varriount|Busy	Araq: Yes, I know that. But not in when encoded.
21:25:57	Araq	then you can always decode it to int64
21:26:13	Araq	and save 8 bytes per entry
21:28:06	sillesta	StringStream from the 'streams' module is deprecated. is there any other way to get a empty string with defined size?
21:28:36	Araq	it's not deprecated
21:28:47	sillesta	maybe i misread..
21:29:06	dom96	Araq: I will make a package called "compiler", that means c2nim will need to import compiler/ast instead of just ast for the compiler modules.
21:29:07	dom96	Sound good?
21:29:26	Araq	yup
21:29:29	ldlework	cool
21:30:52	def-	sillesta: newString(size)?
21:30:59	Araq	sillesta: however there is also system.newStringOfCap() and system.newString()
21:31:16	sillesta	gotcha
21:31:39	Araq	dom96: beware that the compiler also depends on doc/*.txt though
21:32:03	dom96	Araq: why?
21:32:06	Araq	since we read the command line argument descriptions at compile time
21:32:45	dom96	That complicates things.
21:33:21	dom96	What directory prefix does it to try to read those files from?
21:34:00	Araq	doc/
21:34:41	Araq	dom96: well we can move the 2 files to compiler/doc
21:34:49	Araq	and adapt the docgen instead
21:35:11	dom96	Where will the compiler look for those files?
21:35:18	dom96	$moduleDir/../doc?
21:35:49	dom96	if that's the case then it will work
21:36:40	Araq	nah it's --path:"$projectPath/.."
21:37:13	Araq	and then I do slurp("doc/basicopt.txt") and slurp("doc/advopt.txt") iirc
21:37:24	ekarlso	why is there both map and mapIt ?
21:37:33	def-	ekarlso: mapIt is more convenient to use
21:37:44	def-	but map can be used with regular procs
21:37:45	dom96	Araq: That's unfortunate.
21:38:30	flaviu	ekarlso: map(proc (val: T): R = ...) is a huge pita
21:38:47	Araq	dom96: actually it's not. this means we can move these 2 files to compiler/doc and 'koch boot' should continue to work
21:38:48	ekarlso	.mao(|s| {})
21:38:50	ekarlso	would be awesome
21:39:42	dom96	Araq: alright
21:40:27	def-	ekarlso: you have seen the future module?
21:40:56	def-	.map((s) => ...) should work
21:41:22	onionhammer	(sometimes)
21:41:29	def-	yes
21:41:32	flaviu	def-: I honestly spend more time wrestling with bugs in that module than just writing out the proc...
21:42:08	Araq	I'm not aware of any unfixed (!) bugs for =>
21:42:16	ekarlso	one thing that would be even more cool .map((s) => (var i = foo; foo...))
21:42:41	onionhammer	.map((s:type) => ...)
21:43:00	onionhammer	vs .map((s) => ...)
21:43:03	*	jokra joined #nim
21:43:25	Araq	onionhammer: we have better type inference now. :type is not necessary
21:43:34	onionhammer	since today?
21:43:38	onionhammer	because its still happening for me
21:43:49	Araq	since months
21:44:06	onionhammer	welp, then i guess i should submit a but
21:44:08	onionhammer	bug*
21:44:17	dtscode	please dont submit a butt
21:44:29	onionhammer	test2.nim(14, 4) Error: type mismatch: got (seq[int], proc (GenericParam): auto)
21:44:32	onionhammer	but expected one of:
21:44:34	onionhammer	system.map(data: var openarray[T], op: proc (var T){.closure.})
21:44:36	onionhammer	system.map(data: openarray[T], op: proc (T): S{.closure.}): seq[S]
21:44:47	ekarlso	+1 to onionhammer
21:45:05	flaviu	https://gist.github.com/029cc41dc59c6f0758ba
21:45:39	onionhammer	yeah, I get that error with flaviu's code snippet
21:45:49	*	sillesta quit (Read error: Connection reset by peer)
21:49:48	*	sillesta joined #nim
21:50:20	*	nimnoob joined #nim
21:51:09	Araq	hrm strange. maybe I misremember
21:51:30	Araq	I saw code in the compiler to infer these things
21:51:48	onionhammer	it works sometimes
21:51:50	onionhammer	.filter(x => x mod 2 == 1)
21:51:54	onionhammer	but not map()
21:52:00	*	superfunc joined #nim
21:55:27	Araq	aha, it's because 'map' itself is generic
21:55:37	Araq	for a concrete mymap it works ... yay
21:56:23	*	rpag joined #nim
21:56:31	gmpreussner	i have a C function that takes a pointer to an array of Foo: void Foobar(Foo*** foo); what would be the best way to map this to nim? pointer to an unchecked array?
21:56:56	gmpreussner	or do unchecked arrays have some special format, i.e. a count in front?
21:57:21	Araq	they do not, but they are not pointers either
21:57:53	gmpreussner	is there a way to express this?
21:57:56	Araq	is that 3 or 2 stars?
21:58:00	gmpreussner	three
21:58:08	gmpreussner	it's an array out parameter
21:58:29	Araq	2 dimensional array then?
21:58:37	gmpreussner	nope
21:58:48	gmpreussner	i know, it's gross
21:59:15	gmpreussner	http://libusb.sourceforge.net/api-1.0/group__dev.html#gac0fe4b65914c5ed036e6cbec61cb0b97
22:00:14	Araq	ah so it's a pointer to an array of pointers to Foo
22:00:31	gmpreussner	yep
22:00:43	gmpreussner	array of Foo*
22:00:59	gmpreussner	i forgot the star above
22:01:40	gmpreussner	well, let me ask you this first: how would i model an array of Foo*?
22:01:51	gmpreussner	i think the next step would then just be a ptr to that
22:02:45	gmpreussner	it would be nice if it mapped to some array type in Nim. but if that's not possible, that's ok, too. then i just need to know how to access the values.
22:03:24	dom96	Araq: nimble install c2nim@#head, try it.
22:03:44	Araq	ptr UncheckedArray[ptr Foo] perhaps
22:04:21	BlaXpirit	dom96, c2nim installed successfully
22:04:24	BlaXpirit	and works
22:04:31	Araq	where type UncheckedArray {.unchecked.} [T] = array [10_000, T]
22:04:43	dom96	BlaXpirit: :)
22:04:47	BlaXpirit	thx
22:05:08	sillesta	okay, maybe i'm missing something, but isn't this supposed to print something out? https://gist.github.com/sillesta/af314983eab9865af89d
22:05:19	sillesta	just getting a newline
22:05:34	BlaXpirit	sillesta, sey hello to unicode :(
22:05:44	BlaXpirit	å is 2 bytes
22:05:49	BlaXpirit	reversing them gives gibberish
22:05:51	sillesta	removed them, same thing
22:06:31	def-	i have a pull request for reversing unicode strings
22:06:31	gmpreussner	Araq: ok, i'll try that
22:06:31	Araq	sillesta: use countdown(size-1, 0)
22:06:35	def-	it's even performant!
22:06:49	*	superfunc quit (Ping timeout: 246 seconds)
22:07:15	BlaXpirit	utf-8 strings?
22:07:25	sillesta	Araq: nada
22:07:29	BlaXpirit	bah, I just wish Rune was more ubiquitous
22:07:30	def-	BlaXpirit: yes
22:08:01	*	nimnoob quit (Remote host closed the connection)
22:08:06	ldlework	utf-8 != unicode
22:08:42	gokr	sillesta: In this article I also play with reversed, and also notice the unicode issue: http://goran.krampe.se/2014/10/29/nim-and-oo/
22:08:49	def-	https://github.com/Araq/Nim/pull/1842/files
22:09:02	Araq	gmpreussner: though it's possible that there is a codegen bug lurking here as 'ptr array' is mapped to a single star iirc
22:09:08	*	yonson quit (Ping timeout: 245 seconds)
22:09:33	Araq	gmpreussner: maybe you need to use 'ptr ptr ptr Foo' then ...
22:10:15	BlaXpirit	sillesta, there are mistakes in indices
22:10:25	Araq	or ptr ptr UncheckedArray[ptr Foo] ...
22:10:54	BlaXpirit	sillesta, this works :| https://bpaste.net/show/e273ff5ddffa
22:11:15	gmpreussner	Araq: ok, so 'ptr UncheckedArray[ptr Foo]' maps to 'Foo**' ?
22:11:16	ldlework	Araq: lol
22:11:26	sillesta	BlaXpirit: works for me too
22:11:41	sillesta	suspecting i may have copied the null terminator to the start maybe ;)?
22:11:43	*	Varriount|Busy quit (Ping timeout: 246 seconds)
22:11:53	BlaXpirit	sillesta, that is actually most likely the case
22:11:58	BlaXpirit	I haven't considered this
22:13:33	*	Demon_Fox quit (Quit: Leaving)
22:15:51	sillesta	gokr: interesting article, is there a way to treat utf8 characters as a single entity?
22:16:09	BlaXpirit	unicode module; Rune
22:16:33	gokr	Well, def- just pointed out his pull request that handles it.
22:16:57	gokr	sillesta: I do have more articles btw: http://goran.krampe.se/category/nim
22:16:57	Araq	Rune is just as wrong as char iirc
22:17:02	BlaXpirit	that's not wise to implement all character-based algorithms on bytes
22:17:18	sillesta	hehe, i'm not used to seeing "live" changes to languages
22:17:26	BlaXpirit	Araq, how is it wrong? what then?
22:19:13	ekarlso	sillesta: try Rust on for size :p
22:19:32	Araq	something like é can be 2 utf-32 characters
22:19:36	sillesta	i did, wasn't my style
22:19:42	ekarlso	sillesta: +1
22:19:44	Araq	and when you reverse it you get ´e
22:19:49	sillesta	couldn't get past lifetimes
22:20:18	def-	Araq: that should be fixed in my reverse
22:20:31	def-	ah, if you reverse runes
22:20:39	Araq	yep.
22:20:45	def-	yeah, has to be "blocks of runes"
22:20:54	def-	because each character can have a number of combining characters
22:21:11	def-	not like any software handles this correctly
22:21:16	BlaXpirit	as if that wasn't difficult enough, you attempt it on utf-8 :p
22:21:30	def-	BlaXpirit: first i did it with runes, but it was terribly slow
22:21:47	Araq	BlaXpirit: with utf-8 it's at least obviously wrong when you test it with ü
22:22:15	flaviu	It's easiest to never reverse any strings.
22:22:39	def-	flaviu: but you see that 2 people just tried it and ran into problems
22:22:43	*	nimnoob joined #nim
22:22:45	BlaXpirit	let's just console outselves with the fact that in real world you never need to reverse strings (and that's when I remembered right-to-left script)
22:23:17	ekarlso	what does the "distinct" mean ?
22:23:36	def-	ekarlso: http://nim-lang.org/manual.html#distinct-type
22:23:56	Araq	BlaXpirit: how about you blame me instead for not copying Java/C# blindly? or maybe another ramble of how the stdlib is useless
22:24:35	Araq	*rant
22:25:06	BlaXpirit	Araq, Java/C# is irrelevant here because they use unicode strings
22:25:26	renesac	def-: by the way, shouldn't you use 'distinct tuple' for your rational numbers library?
22:25:29	Araq	they use utf16...
22:25:36	gokr	In Pharo there is actually quite advanced support for multibyte strings - and it turns out, Unicode is not the whole answer.
22:25:40	renesac	now that tuples are more compabile with each other
22:26:36	def-	renesac: i don't know
22:26:38	BlaXpirit	Araq, well congrats, the situation is already better than C# and Java. nothing to rant about in such case
22:28:02	BlaXpirit	i don't pick these languages as examples anyway o.o
22:28:20	flaviu	Araq: There's no reason to flame, it won't take the discussion anywhere useful.
22:28:22	renesac	I think yes, as your type isn't any 2 integer tuple
22:29:32	renesac	and I think the shuffle and other random things would be better in a dedicated random library
22:29:50	renesac	of course, right now it don't exists
22:29:54	BlaXpirit	renesac, what shuffle are you talking about?
22:29:57	BlaXpirit	renesac, it exists
22:30:04	BlaXpirit	just that nobody knows about it and nobody cares
22:30:11	renesac	https://github.com/Araq/Nim/pull/1840/files
22:30:19	renesac	BlaXpirit: that one you did based on python?
22:30:26	BlaXpirit	yes
22:30:38	BlaXpirit	what is that shuffle doing there? :D
22:30:53	BlaXpirit	nim's random() is not even usable
22:31:12	renesac	ok
22:31:18	renesac	found my integer log function
22:31:22	def-	BlaXpirit: why not?
22:31:39	Araq	sillesta: btw you need to iterate up to 'len div 2'
22:31:45	BlaXpirit	https://github.com/Araq/Nim/issues/1506 "Problems with random number generation"
22:31:49	renesac	BlaXpirit: what behaviour you want when the operand is zero?
22:32:01	renesac	for the log base 2 function
22:32:30	fowlmouth	BlaXpirit, dont get discouraged if nobody uses your libraries, instead write things that you use
22:32:58	BlaXpirit	fowlmouth, yeah, just gonna finish up 3rd and final library nobody's gonna use :D
22:33:00	fowlmouth	if someone else finds it useful thats just a bonus
22:35:14	BlaXpirit	the gist of "Problems with random number generation" is random() will give a random number between 0 and 32767 even if you ask it for a random number between 0 and a billion
22:35:14	renesac	BlaXpirit: can't you make a random function that accepts a typedesc as parameter?
22:35:30	EXetoC	what libs again?
22:35:33	BlaXpirit	renesac, that's... not good
22:35:39	renesac	instead of many different functions
22:35:40	renesac	why?
22:35:46	EXetoC	just advertise, add readme's and so on
22:35:47	BlaXpirit	ifyou mean what i think you mean
22:36:10	EXetoC	3. ?????? 4. PROFIT
22:36:21	*	flaviu quit (Remote host closed the connection)
22:39:08	Araq	BlaXpirit: we mostly just wrapped ansi C here, so you're complaining about libc's implementation as far as I'm concerned
22:39:21	BlaXpirit	no, Araq
22:39:34	BlaXpirit	libc's implementation just gives you a random number between 0 and RAND_MAX
22:39:39	BlaXpirit	no strings attached, no hard feelings
22:40:07	BlaXpirit	Nim's functions based on it, however, are inherently bugged
22:42:56	*	flaviu joined #nim
22:43:02	def-	BlaXpirit: they can be fixed though?
22:43:12	def-	BlaXpirit: worst case, by getting multiple random numbers?
22:43:13	BlaXpirit	https://github.com/Araq/Nim/blob/3119fe087d28344a34f296f1548cd8d852aea2f8/lib/pure/math.nim#L221 for reference
22:43:18	BlaXpirit	workarounds are possible, yes
22:43:27	dom96	BlaXpirit: please fix it
22:43:28	ekarlso	would anyone wanna add base64url support for base64 ?
22:43:46	renesac	def-: you are better off implementing a better random number generator yourself
22:43:47	BlaXpirit	dom96, cryptographers will cringe
22:43:56	BlaXpirit	but sure, i can try
22:44:17	dom96	cryptographers should create specialised libraries
22:44:58	dom96	The current random's use case is basic things like card games.
22:45:04	dom96	It works ok for that I think.
22:45:10	renesac	dom96: right now, no
22:45:19	def-	shuffle is good for card games too :P
22:45:19	BlaXpirit	you can't have seeded games, for instance
22:45:21	renesac	or badly
22:45:53	dom96	Is there some other problem apart from it only generating random numbers between 0 and MAX_RAND?
22:45:55	*	Demos quit (Quit: Leaving)
22:46:07	BlaXpirit	yes
22:46:11	BlaXpirit	horrible inconsistency
22:46:34	dom96	inconsistency with what?
22:46:44	BlaXpirit	across systems and compilers
22:46:59	BlaXpirit	inconsistency with itself
22:47:00	dom96	ahh, MAX_RAND is always different.
22:47:11	dom96	Change it so that it doesn't use libc then.
22:47:16	BlaXpirit	not only it, the algorithm itself can be different and whatnot
22:47:23	BlaXpirit	dom96, well, that requires a dedicated library!
22:47:36	BlaXpirit	which i have made a long time ago
22:47:41	dom96	Add it to the stdlib then.
22:47:44	dom96	Call it random.
22:47:48	renesac	dom96: not so fast
22:47:51	BlaXpirit	i suggested that in september
22:48:01	BlaXpirit	but first people must review it and improve it
22:48:04	Araq	dom96: not so fast
22:48:20	dom96	That is what PRs are for.
22:48:26	BlaXpirit	it is not ready
22:48:32	BlaXpirit	it was the first thing i wrote in nim basically
22:48:36	EXetoC	not so fast
22:48:38	ekarlso	flaviu: u ot a clue on decoding hs256 ?
22:48:39	dom96	then what are we arguing about?
22:48:54	dom96	EXetoC: not so slow
22:49:12	BlaXpirit	i'm arguing about lack of action on the problems with math.random
22:49:26	BlaXpirit	i should indeed at least fix the glaring bug
22:49:30	renesac	I'm trying to write the integer log 2 function for BlaXpirit replace the float one in his library
22:49:51	Araq	tbh I don't really agree it's a bug. but that's just me. nobody should use this for anything serious anyway. and the fact that it says 'int' is rather meaningless
22:49:54	*	saml quit (Quit: Leaving)
22:49:57	flaviu	ekarlso: I'm not sure what you're asking.
22:50:06	BlaXpirit	Araq, please don't be like this
22:50:14	Araq	it's obviously underspecified, or is -4 a valid value for it?
22:50:40	BlaXpirit	returns a value in range "a .. b-1"
22:50:56	BlaXpirit	okay, maybe it does, but some values are impossible to obtain
22:51:09	gokr	For random I think this is brilliant: http://xkcd.com/221/
22:51:21	ekarlso	flaviu: so you're supposed to encrypt a string with hs256 but also supposed to decrypt / decode it ye ?
22:51:46	*	tinAndi joined #nim
22:51:56	flaviu	ekarlso: let me bring this into PM, it's not directly about nim
22:52:08	ekarlso	ok
22:52:09	BlaXpirit	but can we at least agree that shuffle() shouldn't be added to math, and that it shouldn't be based on random()?
22:52:46	renesac	https://github.com/Araq/Nim/issues/936 <-- Araq: have you read my last answer to this topic?
22:52:49	Araq	yes, we can agree on that
22:53:04	renesac	do you think what I propose to do with int literal is possible?
22:53:35	dom96	BlaXpirit: You're upset that Araq is not making random his priority when he has lots of other things to do?
22:54:08	gokr	gnite folks
22:54:11	EXetoC	"one year ought to be enough for the design and implementation of a programming language" - Bill Gates
22:54:30	dom96	We don't have a big company backing us. In the open source world if you want to get something fixed then in most cases you need to do it yourself.
22:54:51	ldlework	I think we should all take BlaXpirit not so seriously
22:54:57	ldlework	err that came out wrong
22:55:07	ldlework	I mean his immediacy and hand-waving
22:55:22	ldlework	he just cares. I think he knows Araq is busy.
22:55:41	BlaXpirit	something like that...
22:55:46	Araq	renesac: cannot see where this solution differs from what we do
22:56:04	EXetoC	just need to express it more calmly
22:56:12	BlaXpirit	i always start calm
22:56:15	Araq	5 already has the type 'int literal(5)' and is compatible with uint8, 300 is not
22:56:16	ldlework	lol
22:56:22	gokr	BlaXpirit: You wrote a different PRNG in Nim or did I misunderstand?
22:56:30	renesac	Araq: humm
22:56:41	renesac	I will try to run skyfex fork then
22:56:44	BlaXpirit	it's just that people always try to argue with my obviously correct statements :p
22:56:49	renesac	and add those proc overloads
22:57:10	renesac	but I will be away for a week first
22:57:14	BlaXpirit	gokr, i made a library that utilizes the same Mersenne Twister that Python uses
22:58:07	gokr	BlaXpirit: Neat. I also saw a huge number of other langs relying on the twister. Which apparently is good for simulation, albeit worthless for crypto I think I read.
22:58:07	ldlework	BlaXpirit: nice
22:58:10	BlaXpirit	and it should support different implementations of PRNG but i can't be sure the API is good enough for that
22:58:23	Araq	BlaXpirit: I'd fix it by s/int/range[0..10_000]
22:58:27	gokr	Julia uses it for one newish language.
22:58:50	flaviu	I'd like to post www.pcg-random.org
22:58:56	flaviu	That website is excellent.
22:58:56	BlaXpirit	Araq, if you want to make assumptions about RAND_MAX, yeah, that's not so bad
22:58:57	gokr	BlaXpirit: Guess this is it: http://forum.nim-lang.org/t/533
22:59:16	renesac	flaviu: I wanted to make a random library based on that xorshift*
22:59:18	BlaXpirit	yes. i mention it in https://github.com/Araq/Nim/issues/1506 too
22:59:25	dtscode	fowlmouth, wheres your ircd again?
22:59:31	renesac	but I had other priorities
22:59:35	flaviu	renesac: I'm excited about PCG these days!
22:59:35	BlaXpirit	renesac, there is no problem with having both
22:59:52	EXetoC	use several popular engines and pick a default
22:59:54	flaviu	Mersenne Twister actually fails a few statistical tests.
23:00:18	fowlmouth	dtscode, https://github.com/fowlmouth/nimlibs/tree/master/fowltek/musings
23:00:27	flaviu	LCG 96/32 is higher quality, smaller, and faster, but I can't find the multiplier anywhere.
23:00:39	renesac	BlaXpirit: yeah, the ideal is a library where you can use a uniform interface for different types of random number generators
23:00:42	renesac	that is what you tried to do
23:00:49	BlaXpirit	yes
23:01:11	Araq	"Which apparently is good for simulation, albeit worthless for crypto" gokr +10
23:01:13	renesac	I'm just not sure if it is the best interface (only looked quickly)
23:01:14	BlaXpirit	the library did kinda get abandoned (due to lack of attention and time)
23:01:14	dtscode	ty
23:01:34	renesac	BlaXpirit: some things you copied too much from python
23:01:51	renesac	by default, the random function spits a float
23:01:56	renesac	not an int that would be expected
23:02:03	ldlework	why would an int be expected?
23:02:04	BlaXpirit	Araq, my library exposes operating system's random byte source
23:02:18	EXetoC	have an engine interface
23:02:18	renesac	that kinda makes sense in python, because python don't have fixed lenght integers
23:02:20	BlaXpirit	which is "worthy" for crypto
23:02:24	EXetoC	but you probably know more than me
23:02:41	Araq	so then the stdlib is good for simulation but perhaps bad for crypto ... and this is exactly why a stldib's random is a toy IMHO
23:02:42	ldlework	renesac: Python is by far the only language that defaults to floats for random functions...
23:02:43	gokr	flaviu: Squeak has (since the dawn of time) an LCG variant (http://en.wikipedia.org/wiki/Lehmer_random_number_generator)
23:03:01	BlaXpirit	Araq, it can be a toy but without bugs at least
23:03:39	Araq	BlaXpirit: ok, but more elaborate stuff HAS to be a nimble package
23:03:49	BlaXpirit	probably
23:03:54	Araq	otherwise nobody trusts it anyway
23:04:01	Araq	well shouldn't trust it.
23:04:05	gokr	flaviu: Nice site you found.
23:04:32	flaviu	gokr: Thanks, that's interesting! re. the site, the paper is also excellent reading. I found it very accessible.
23:05:26	flaviu	The idea that it's possible to just change c to some arbitrary number to obtain a new sequence sounds useful.
23:07:02	def-	t
23:07:11	def-	(oops)
23:07:17	BlaXpirit	I sure would like to keep up the topic of RNG, but it's sleepy time
23:09:23	fowlmouth	def-, can you pull that big sdl2 PR
23:09:29	renesac	ldlework: do you mean "not the only"?
23:09:54	fowlmouth	merge*
23:11:46	def-	fowlmouth: huh?
23:11:48	*	brson quit (Quit: leaving)
23:12:38	fowlmouth	def-, https://github.com/nim-lang/sdl2/pull/17
23:13:15	def-	fowlmouth: into my own fork or what?
23:13:45	sillesta	are TType and PType legacy code?
23:14:00	sillesta	(the naming)
23:14:02	Araq	sillesta: yes.
23:14:47	fowlmouth	def-, i'll make you a contributor 1 sec
23:14:59	Araq	I don't think I'll ever update the compiler itself to get rid of T/P though
23:16:34	dtscode	ircd.nim(378, 8) Error: invalid indentation
23:16:42	sillesta	no magical regex to save the day ;)?
23:16:43	dtscode	do you know why it would be saying that?
23:16:50	dtscode	since it looks proper to me
23:17:36	dtscode	oh wait never mind
23:17:40	BlaXpirit	dtscode, it's often unmatched brackets on prev lines
23:18:00	dtscode	BlaXpirit, in this case it was an empty of
23:18:10	fowlmouth	dtscode, that section needs a discard
23:18:22	dtscode	yeah. just realized :/ silly me
23:20:18	*	BlaXpirit quit (Quit: Quit Konversation)
23:22:02	*	JinShil joined #nim
23:23:45	dtscode	ircd.nim(423, 34) Error: type mismatch: got (proc (AsyncSocket){.closure.}) but expected 'proc (AsyncSocket){.closure, gcsafe.}' # does this mean that i just need to find the function and add the gcsafe pragma to it?
23:24:24	Mat4	ciao
23:24:28	*	Mat4 left #nim (#nim)
23:24:33	*	sillesta quit (Ping timeout: 276 seconds)
23:24:48	*	nimnoob quit (Ping timeout: 265 seconds)
23:27:43	*	sillesta joined #nim
23:28:28	dtscode	fowlmouth,
23:29:36	*	zahary_ quit (Ping timeout: 246 seconds)
23:29:41	fowlmouth	yea that probably work
23:30:47	def-	fowlmouth: should be fine, i hope
23:30:58	dtscode	hmmm for whatever reason there isnt anything like that on that line
23:31:03	dtscode	no functions or anything
23:31:26	*	akiradeveloper joined #nim
23:32:26	def-	fowlmouth: my examples still don't compile though, can I pull in my pull request as well?
23:33:11	flaviu	`"abc".match(re"abc", endpos = 0)`, "Error: invalid expression: 'endpos = 0'"
23:33:11	fowlmouth	sure
23:33:28	flaviu	I'm I don't understand why that's an error.
23:33:42	flaviu	endpos is the correct name.
23:34:52	fowlmouth	def-, btw that importc wasnt missing, there was a {.push importc: "SDL_$1".} before it (those can go away now since the function names are lowercase)
23:40:52	rpag	fowlmouth, ive never seen you be so serious
23:41:08	*	tinAndi quit (Quit: ChatZilla 0.9.91.1 [Firefox 35.0/20150108202552])
23:42:14	Araq	flaviu: report it
23:44:15	def-	fowlmouth: hm, the big PR broke the gfx.nim
23:46:11	fowlmouth	def-, did he not update that file?
23:47:01	*	brson joined #nim
23:47:20	def-	fowlmouth: or it was broken even before, huh
23:47:46	fowlmouth	rpag, stop spying on me >_>
23:48:45	rpag	fowlmouth, i'm not, i check this channel from time to time and you happened to be speaking.
23:51:14	flaviu	looks like a unittest bug.
23:51:24	def-	fowlmouth: he renamed aapolygonColor to aaPolygonColor for example, now the importc fails. I'm fixing the importc instead of renaming back, ok?
23:52:31	dtscode	can someone tell me what is going on here?
23:52:50	dtscode	oops
23:52:51	dtscode	https://bpaste.net/show/543a78511efc
23:53:19	fowlmouth	def-, thats fine thank you, feel free to make any fixes you like
23:55:23	fowlmouth	dtscode, add gcsafe to that function? cant tellwhich it is