<< 13-07-2013 >>

---

00:01:02	Max00355	Wow, nice :P
00:01:03	*	Sergio965 joined #nimrod
00:01:08	Max00355	http://github.com/Max00355
00:01:12	Max00355	That's mine.
00:02:19	dom96	yeah, some of my projects are also in https://github.com/nimrod-code
00:02:43	Max00355	Oh, wow, awesome.
00:05:29	dom96	hrm, BlooCoin sounds familiar.
00:05:46	dom96	so many *coin nowadays though heh
00:33:39	Araq	good night
00:36:26	DAddYE	hey there, another noob question
00:36:36	DAddYE	I'm unable to find any doc about that
00:36:46	DAddYE	if I've: proc poll_init*(loop: ptr TLoop)
00:37:09	DAddYE	what the best way to provide a pointer of TLoop
00:37:14	DAddYE	(allocated)
00:37:17	DAddYE	?
00:37:47	dom96	addr loop
00:38:02	dom96	But unless you are interfacing with C you shouldn't be using 'ptr'
00:38:13	dom96	use var or 'ref'
00:39:05	DAddYE	dom96: I'm interfacing with c
00:39:24	DAddYE	dom96: so you mean some: var loop = TLoop.new
00:39:40	DAddYE	pool(addr(loop))
00:39:41	DAddYE	?
00:39:42	dom96	new() only works for 'ref' I think
00:39:48	DAddYE	dom96: exaclty
00:40:48	DAddYE	so how I can get a pointer of a object?
00:44:08	dom96	DOesn't the C lib have a way to init this TLoop structure?
00:44:09	dom96	oh yeah
00:44:09	dom96	poll_init does the initing.
00:44:09	dom96	Is that a function from the C lib?
00:44:09	dom96	if so then:
00:44:09	dom96	var loop: TLoop
00:44:09	dom96	poll_init(addr loop)
00:44:44	DAddYE	dom96: lemme try
00:45:05	dom96	by doing 'addr obj'
00:45:11	dom96	or addr(obj)
00:46:24	DAddYE	SIGSEGV: Illegal storage access. (Attempt to read from nil?)
00:46:34	DAddYE	so should be allocated before
00:49:12	dom96	show me your code
00:51:51	DAddYE	https://github.com/DAddYE/gist/blob/master/uv.nim
00:53:39	dom96	where is the code where you are using the uv wrapper?
00:54:06	DAddYE	right now, I'm writing a test so
00:54:10	DAddYE	I'm initializing a timer
00:54:18	DAddYE	no more than one line
00:54:28	DAddYE	var timer = TTimer.new
00:54:37	DAddYE	let Loop = default_loop()
00:54:58	DAddYE	err = timer_init(loop, addr timer[])
00:55:04	DAddYE	^^^
00:55:20	dom96	huh?
00:55:26	dom96	You're not even using poll_init?
00:55:44	dom96	TTimer.new doesn't look right
00:56:23	DAddYE	dom96: was a copy/paste issue but the problem is the same
00:56:24	dom96	ok, so default_loop returns a ptr TLoop
00:56:37	DAddYE	dom96: exactly
00:56:42	dom96	I think you'd be safer using 'var' there
00:56:50	DAddYE	ok
00:57:22	dom96	and then remove the []
00:57:38	dom96	on which line does the segfault happen?
00:58:13	DAddYE	this was a try since addr(TTimer.new) => pointer to => ref TTimer
00:58:27	DAddYE	[] is a dereference so a pointer to pointer
00:58:30	DAddYE	(I guess)
00:58:45	dom96	lol, yeah. I don't think doing that is a good idea :P
00:58:55	dom96	var timer: TTimer
00:59:01	dom96	var loop = default_loop()
00:59:04	DAddYE	okj
00:59:10	dom96	err = timer_init(loop, addr timer)
00:59:12	dom96	Should work.
00:59:21	dom96	But i'm not familiar with libuv
01:00:36	DAddYE	yep, it does
01:00:38	DAddYE	thanks man
01:00:52	DAddYE	just because is day 2 for me in Nimrod
01:01:23	DAddYE	var timer: TTimer means => var timer = null of TTimer type ?
01:02:33	DAddYE	var timer: TTimer means => declares timer to have type TTimer
01:02:44	DAddYE	and timer is NULL
01:02:45	DAddYE	right?
01:03:11	dom96	kinda
01:03:42	dom96	var timer: TTimer # declares a variable of name `timer` and type `TTimer`
01:04:10	dom96	Each of the fields of the object are initialised to their defaults.
01:04:25	DAddYE	dom96: so after timer_init(loop, addr timer)
01:04:27	dom96	i.e. if TTimer object has an int field it is '0' (IIRC) not nil.
01:04:32	dom96	timer itself is not nil
01:04:47	DAddYE	what happen to the original timer var?
01:04:56	DAddYE	how I can see fields populated ?
01:05:09	dom96	I dunno what libuv does, but it possibly sets the fields to some values.
01:05:13	dom96	repr may work
01:06:02	*	Associat0r quit (Quit: Associat0r)
01:06:02	DAddYE	echo repr(timer)
01:06:04	DAddYE	SIGSEGV: Illegal storage access. (Attempt to read from nil?)
01:06:07	DAddYE	-.-
01:06:19	DAddYE	from the manual should fill fields
01:08:38	dom96	maybe it doesn't.
01:08:49	dom96	why do you want to see all the fields?
01:12:37	DAddYE	dom96: just for testing
01:13:09	dom96	I still feel that repr should never just crash.
01:13:31	dom96	Not sure what Araq's take was on that.
01:13:47	DAddYE	echo repr(timer)
01:13:51	DAddYE	before timer_init
01:14:02	DAddYE	report correctly all fields to their default value
01:14:06	DAddYE	after timer_init
01:14:09	DAddYE	just crash
01:22:11	*	q66 quit (Quit: Leaving)
01:29:12	*	DAddYE quit (Remote host closed the connection)
01:29:46	*	DAddYE joined #nimrod
01:33:05	dom96	anyway, I have to sleep.
01:33:07	dom96	Good night
01:34:19	*	DAddYE quit (Ping timeout: 256 seconds)
02:13:26	*	Sergio965 quit (Ping timeout: 240 seconds)
02:30:21	*	DAddYE joined #nimrod
02:34:48	*	DAddYE quit (Ping timeout: 245 seconds)
03:10:11	*	DAddYE joined #nimrod
03:33:19	*	Trix[a]r_za is now known as Trixar_za
03:56:16	DAddYE	Araq: big question, why nimrod is not case sensitive
03:56:46	DAddYE	?
04:36:32	*	OrionPK quit (Read error: Connection reset by peer)
04:40:08	Max00355	Because it isn't Java
04:43:45	Max00355	dom96: BLC sounds famiular you say? They were popular about two months ago. I created it myself frm scratch.
05:00:57	DAddYE	Max00355: what does it means?
05:01:05	DAddYE	Max00355: every language I know is sensible case
05:01:12	DAddYE	maybe Visual Basic isn't
05:58:37	*	comex quit (Quit: Coyote finally caught me)
05:58:45	*	comex joined #nimrod
07:07:44	*	Associat0r joined #nimrod
07:07:44	*	Associat0r quit (Changing host)
07:07:44	*	Associat0r joined #nimrod
07:26:02	*	zahary joined #nimrod
07:34:15	Araq	DAddYE: Lisp, Pascal, Ada, Eiffel, Basic are case insensitive
07:34:49	Araq	it's mostly some historic thing as keyboards only supported upper case
07:36:52	Araq	that said, nimrod is style insensitive because people spend too much time about underscore_crap vs camelCase vs PascalCase
07:37:01	Araq	*arguing about
07:39:08	DAddYE	Araq: I can agree, but let's say that modern language are case sensitive
07:39:18	DAddYE	then u can dictate your stile
07:39:22	DAddYE	then u can dictate your stYle
07:39:31	DAddYE	like we do with T and P
07:39:46	DAddYE	having a sensible case will help me a lot
07:39:55	DAddYE	so I can avoid TIo
07:40:08	Araq	yes but the idea is that your IDE/editor can render the identifiers as you prefer
07:40:26	Araq	though no editor does that yet ;-)
07:40:41	DAddYE	Araq: not at all we use and Ide like Eclipse/VS
07:40:50	DAddYE	I'm a poor man with vim
07:40:52	DAddYE	:D
07:41:03	Araq	consistency is better left to tools IME
07:41:12	DAddYE	I agree
07:41:14	DAddYE	but say that
07:41:20	DAddYE	TIo
07:41:38	DAddYE	if isn't written by you isn't under stable at first look
07:41:42	DAddYE	IO
07:41:49	Araq	*shrug* I can deal with your nick here, DAddYE
07:41:55	DAddYE	Araq: hahahah
07:41:57	*	Trixar_za is now known as Trix[a]r_za
07:42:04	DAddYE	I mean, I'm for freedom
07:42:17	DAddYE	but _ and camel case are for me the opposite
07:42:26	DAddYE	I've a style where
07:42:33	DAddYE	Types are CamelCase
07:42:37	DAddYE	constant CONSTANT
07:42:41	DAddYE	but
07:42:46	DAddYE	timer_start
07:42:49	Araq	I've programmed in enough languages to not give a shit about these things
07:42:53	DAddYE	shouldn't be the same of timerstart
07:43:13	*	Trix[a]r_za is now known as Trixar_za
07:43:50	Araq	using case to distinguish between different entities is just another form of hungarian notation
07:43:54	DAddYE	that's a thing, but consider that I need to spend time to figure out a different name
07:44:12	Araq	I agree it causes problems for wrappers
07:45:02	Araq	constants are CONSTANT because of the preprocessor
07:45:08	DAddYE	Time = object; TIME = enum; [1 .. 60];
07:45:14	Araq	we have no preprocessor so it doesn't make sense
07:45:40	DAddYE	proc time() = Time.now
07:46:02	DAddYE	is easier (remember that I'm super new)
07:46:05	DAddYE	to
07:46:17	DAddYE	TTime = object; ETIME = enum; [1 .. 60];
07:46:37	DAddYE	proc time() = TTime.now
07:46:41	Araq	there is no ETIME
07:47:08	DAddYE	I saw that in some wrappers
07:47:15	DAddYE	otherwise (like I've)
07:47:20	DAddYE	TTime = object; TIME = enum; [1 .. 60];
07:47:35	DAddYE	proc time2() = Time.now
07:47:49	Araq	and then we use TiMe for a time represented as a float?
07:48:00	Araq	and why is that a good idea?
07:48:16	DAddYE	Araq: I'm not a super expert to say that
07:48:17	DAddYE	but
07:48:33	DAddYE	ruby, python, go, clojure,
07:48:36	DAddYE	scala
07:48:40	DAddYE	java
07:48:45	DAddYE	and so on are in that way
07:48:52	DAddYE	so programmers like me
07:48:59	DAddYE	that have code for 10 years
07:49:17	DAddYE	have some difficulties to learn a different way
07:49:54	DAddYE	TIO, TTIME, TOS, looks more weird than Io Time TimeFloat
07:50:05	DAddYE	or OS
07:50:50	Araq	you can do 'var time: time' btw
07:51:05	Araq	the T vs P is about value vs reference types
07:51:35	Araq	which require different coding styles so I thought it is a good idea to keep that convention from delphi
07:51:37	DAddYE	but if I omit T/P
07:51:50	DAddYE	then when I need a procedure
07:51:57	DAddYE	I should invent a name
07:52:26	Araq	yeah I figured inventing names is a big problem for many people
07:53:28	Araq	btw Go -- does something special with case too
07:53:35	DAddYE	yes
07:53:41	*	Trixar_za is now known as Trix[a]r_za
07:53:50	Araq	python -- likes to be case insensitive
07:54:03	Araq	at least Guido wanted that once
07:54:24	DAddYE	go, has Camel for public
07:54:39	DAddYE	go is very strict about those things and people seems to like it
07:54:41	Araq	Java, Scala, Ruby - I don't think the creators of these languages really thought about it
07:54:51	DAddYE	hahahah, totally agree
07:55:40	DAddYE	can we argue that 90% of common used languages are case sensitive?
07:55:44	DAddYE	isn't a thing of style
07:55:53	DAddYE	I mean, you know that better than me
07:56:07	DAddYE	is more as a habit that we haven't
07:56:17	DAddYE	and sensible case reduce at least by 3
07:56:25	DAddYE	our naming rules
07:57:10	Araq	*shrug* 90% use {} too
07:57:33	DAddYE	not using it
07:57:41	DAddYE	is different to say
07:57:49	DAddYE	use ∫∫
08:00:07	Araq	http://www.reddit.com/r/programming/comments/1i3qi9/cereal_a_c11_library_for_serialization_binary_xml/cb0sijv
08:01:18	Araq	btw experiments showed snake_case to be more readable and yet I hate it
08:03:06	DAddYE	I'm for it in procedures
08:03:16	DAddYE	helps your readability
08:03:19	DAddYE	but like in javashit
08:03:26	DAddYE	I've no problems with CamelCase
08:03:41	Araq	so what do you do as a language designer? design for the Right Thing (TM) or design for yourself? ;-)
08:03:43	DAddYE	camelCase for functions
08:04:07	DAddYE	give up the choice
08:04:14	DAddYE	to do whatever u what
08:04:25	DAddYE	BUT
08:04:39	DAddYE	if you want to contribute in my repo, please read my code style guide
08:04:54	DAddYE	simple good and okay for most of us
08:05:05	DAddYE	http://uscilab.github.io/cereal/index.html
08:05:12	DAddYE	wrote in camel case he can
08:05:22	DAddYE	if he like it that's okay
08:05:28	DAddYE	but I can't
08:05:33	DAddYE	and that's my problem :D
08:05:52	Araq	but that problem is mitigated in Nimrod ;-)
08:05:53	DAddYE	only STUPID that should NO CODE
08:06:01	DAddYE	only STUPID that should write NO CODE
08:06:16	DAddYE	can do: timer_start then timerstart then TimerStart
08:06:34	DAddYE	fortunately never seen that code
08:06:35	Araq	you know ... it's all the same to me
08:06:50	Araq	it's something like timer-start
08:07:00	DAddYE	is lispy
08:07:05	DAddYE	I don't like it that much
08:07:15	DAddYE	but I wrote a lot of code in clojure
08:07:27	DAddYE	and it's okay... for lisp languages
08:07:54	DAddYE	maybe you can consider it stupid
08:07:57	DAddYE	but like lua
08:08:15	DAddYE	people complain about index that start from 1
08:08:26	DAddYE	there are good reason behind it
08:08:33	DAddYE	(index of array)
08:08:36	Araq	btw psychology tells us people remember the sound of words, not their spelling
08:09:01	DAddYE	infact here we came
08:09:07	DAddYE	TTime
08:09:20	DAddYE	titime?
08:09:28	DAddYE	ttime?
08:09:36	Araq	t_time ?
08:09:44	DAddYE	same
08:09:46	Araq	oh wait t_time is bad but time_t is fine
08:10:03	DAddYE	or _s if is a struct
08:10:05	Araq	because time_t comes from C so it's a classic
08:10:22	DAddYE	honestly I hate a lot prefixes
08:10:25	DAddYE	and suffixes
08:10:32	DAddYE	I prefer camelcase
08:10:42	DAddYE	UPPER => I know is a constant
08:10:59	DAddYE	Capitalize => That's is a type or a Object
08:11:03	Araq	what if it's read from a config file?
08:11:23	DAddYE	lower_maybe_underscore => method/procedure
08:11:34	Araq	is a "conceptually" constant CONST for you?
08:12:01	DAddYE	unless is a crappy json thing
08:12:03	DAddYE	yes
08:12:21	Araq	fair enough but then the compiler can't enforce your naming scheme
08:12:28	Araq	and that sucks imho
08:12:36	DAddYE	wait
08:13:10	DAddYE	are u sure that CIO is a constant?
08:13:20	Araq	no
08:13:20	DAddYE	so is the same
08:13:35	Araq	but then I don't program by guessing around in the first place
08:14:20	Araq	my mouse cursor should tell me what it is when moving over it
08:15:00	DAddYE	unless again is a config file
08:15:00	DAddYE	or like a lot of us
08:15:10	Araq	I know, I know, people don't use/have the proper tools
08:15:10	DAddYE	you code in vim/emacs
08:15:40	DAddYE	never tried aporia but
08:16:00	DAddYE	Eclipse sucks in all sense
08:16:00	Araq	so?
08:16:20	DAddYE	VS is good but I prefer lost an hand before using again windows
08:16:20	Araq	Visual Studio is bearable, Delphi was great
08:16:40	DAddYE	don't know that much on linux
08:16:55	DAddYE	but on mac stylish tools like sublime
08:16:55	DAddYE	python powered
08:17:10	DAddYE	or Texmate Ruby+c++
08:17:35	DAddYE	both good but again unless some weird plugin is a vanilla text editor
08:17:45	DAddYE	I mean
08:17:45	DAddYE	the point is
08:17:55	DAddYE	if you make nimrod case sensitive
08:17:55	DAddYE	you can dictate your coding style
08:18:20	DAddYE	you can still write PTime => The only case where I totally agree
08:18:30	DAddYE	if you like TTime you can write it
08:18:40	DAddYE	but if another one would like to use Time and time
08:18:50	DAddYE	should be free to do it
08:19:15	DAddYE	isn't a critique
08:19:15	DAddYE	but trust me
08:19:30	DAddYE	you are saying me
08:19:40	DAddYE	hey bro, you are right handed but now start to write with the left
08:20:10	Araq	type Time = TTime
08:20:10	Araq	and you're free again
08:20:30	DAddYE	no
08:20:40	DAddYE	proc time():
08:20:50	DAddYE	will be raise a redefined problem
08:21:15	Araq	we could change that, it's a bit buggy already
08:21:15	DAddYE	another important thing
08:21:25	DAddYE	is that in this way
08:21:25	DAddYE	IMHO
08:21:35	DAddYE	you are archiving the opposite problem
08:21:45	DAddYE	some code itself in nimrod
08:21:45	DAddYE	is written in CamelCase
08:22:00	DAddYE	some in snake_case
08:22:10	DAddYE	some without any of them
08:22:30	DAddYE	some like the lua wrapper
08:22:30	Araq	yeah and I can import it and write it so that it adheres to my style
08:22:40	DAddYE	even without T/P/E
08:23:00	Araq	lua_usesthisstyle
08:23:20	Araq	so ... how do you fix that in C?
08:23:20	Araq	you can't
08:23:30	DAddYE	infact
08:23:30	Araq	you're stuck with their bad taste
08:23:50	DAddYE	you're case sensitive
08:24:15	DAddYE	if you want to rename UV_Time function (all of them are lowercase btw)
08:24:30	Araq	also ... me/the compiler enforcing a naming scheme is a nice idea
08:24:30	Araq	however
08:24:40	Araq	I don't know of a working naming scheme
08:24:40	DAddYE	proc myCoolName: {.importc "lua_cool_name"}
08:25:20	DAddYE	I can't argue for procedures
08:25:20	DAddYE	whatever is okay to me
08:25:30	DAddYE	but types Imho
08:25:40	DAddYE	should start for CapitalLetter
08:25:50	DAddYE	const UPPER
08:25:50	DAddYE	then, IMHO up to the coder
08:26:15	DAddYE	99% know that code is like the os (unix)
08:26:15	DAddYE	case sensitive
08:26:25	DAddYE	1% don't know that isn't
08:27:35	DAddYE	and trust me we can see some differences between snake or camel but ONLY in procedures/methods/functions
08:27:35	DAddYE	never seen in my life
08:27:50	DAddYE	class foo < Bar; end
08:28:00	DAddYE	or type my_type
08:28:10	Araq	time_t
08:28:35	DAddYE	struct
08:28:35	DAddYE	or type
08:28:45	DAddYE	sorry type
08:28:55	DAddYE	but again
08:29:10	DAddYE	that's is up to the coder
08:29:10	DAddYE	python codebase
08:29:20	DAddYE	is in CamelCase
08:29:20	DAddYE	and is c
08:29:35	DAddYE	preventing case sensitive
08:29:55	DAddYE	you aren't avoiding those exception
08:30:11	DAddYE	but allowing TiMeIsNow
08:30:11	Araq	well
08:30:41	Araq	C doesn't prevent TiMeIsNow either it's just that you're stuck with it
08:31:01	Araq	should the guy writing the definition be drunk
08:31:16	DAddYE	exactly for the same reason
08:31:26	DAddYE	case insensitive should be unnecessary
08:31:36	DAddYE	other than a bit confusing
08:31:46	DAddYE	mostly removed 3 choices
08:31:46	DAddYE	for naming
08:31:56	DAddYE	that is 90% of the problem
08:32:11	Araq	I'm sorry but you don't make sense
08:32:36	Araq	ok, so it's myIdent in Python and in C it's my_ident
08:32:36	Araq	and I mix these languages
08:32:51	Araq	and it's no problem
08:32:51	Araq	but when I have the same thing all in Nimrod
08:33:01	Araq	it starts to be a problem
08:33:01	Araq	for you
08:33:11	Araq	and I have no idea why that is
08:33:21	DAddYE	the fact that I need to prefix everything
08:33:36	Araq	you don't
08:33:36	Araq	but it's the convention in the stdlib
08:33:51	Araq	oh wait, you don't like conventions that are not your own
08:33:51	DAddYE	otherwise I've redefinition problem
08:34:11	DAddYE	enum HOUR, MINUTE, SECOND
08:34:11	DAddYE	conflict with
08:34:11	Araq	you could give different things different names
08:34:21	DAddYE	proc hour
08:34:21	Araq	also there is .pure for enums
08:34:36	DAddYE	yea thanks to you I figure out the problem
08:34:51	DAddYE	but I spent 1 hour gsubbing last day
08:34:51	DAddYE	:D
08:35:41	DAddYE	in C I can write struct Foo {}; function foo(){};
08:35:41	DAddYE	in nimrod not
08:36:01	DAddYE	in python again class Foo:
08:36:01	DAddYE	def foo
08:36:11	DAddYE	nimrod not
08:36:11	DAddYE	that's the thing
08:36:31	Araq	some say functions should be verbs and types should be nouns ...
08:37:01	DAddYE	in C I can write struct Time {}; function time(){};
08:37:11	DAddYE	or I should
08:37:11	DAddYE	in C I can write struct TheTime {}; function time(){};
08:37:21	Araq	in nimrod I can type allinlowercasefordebugging
08:37:31	Araq	and I surely don't want to miss *that*
08:38:21	Araq	you know ... for the people who have no proper tools like a debugger ;-)
08:38:21	DAddYE	I think a macro/template debug allinlowercasefordebugging could work
08:38:36	Araq	I don't think so
08:38:56	DAddYE	or even better: autocomplete :D
08:39:16	Araq	I have no intellisense for the conditions in breakpoints
08:39:16	Araq	in visual studio
08:39:36	Araq	makes you realize how much case sensitivity does suck
08:40:01	Araq	it's only bearable due to autocomplete IMHO
08:40:21	Araq	of course the autocomplete itself is case insensitive
08:40:21	Araq	proving my point :P
08:40:41	DAddYE	Araq: the point is that u use windows
08:40:51	*	XAMPP_ joined #nimrod
08:41:11	DAddYE	and the system itself told you
08:41:11	DAddYE	in years that is insensitive
08:41:31	DAddYE	unix guys
08:41:31	DAddYE	is the opposite
08:42:01	*	Araq despises Unix
08:44:01	DAddYE	btw up to you, at end is your language :D
08:44:31	DAddYE	quick info, do u really like windows?
08:44:31	DAddYE	0.0
08:44:41	*	XAMPP quit (Ping timeout: 264 seconds)
08:45:36	Araq	I'm typing this on a Linux ...
08:46:21	*	Araq doesn't like any OS really
08:46:41	DAddYE	wow, for I moment I was scared :D
08:46:51	Associat0r	DAddYE: I use windows
08:48:01	DAddYE	Associat0r: we are talking about like/dislike os'es
08:48:16	DAddYE	Araq: back to a real code
08:49:01	Araq	DAddYE: there is 'repr' in the language and I have mRepr that is the magic value for it and opcRepr is its opcode in the VM I'm writing
08:49:16	DAddYE	https://github.com/DAddYE/gist/blob/master/uv.nim#L806-L807
08:49:16	Araq	with your way I'd have REPR for the magic and then need opcRepr anyway
08:49:41	Araq	or use .pure enums which I don't like
08:50:21	*	Araq doesn't mind prefixes, they always sneak into Araq's world
08:51:01	DAddYE	proc repr; const/enum M_REP, O_REP
08:51:21	DAddYE	but again turning on case sensitive will never change your stile
08:51:36	Araq	yeah prefixes ftw ;-)
08:51:36	DAddYE	turning of yes
08:51:46	DAddYE	turning off yes
08:51:56	*	DAddYE can't spell today
08:52:41	DAddYE	but I can create a type: Rep
08:53:11	Araq	I'm willing to implement something to lessen your pain but you'd better come with something that works
08:53:11	DAddYE	which 99.8% of coder should know that in c'ish language starting with Capital means class/struct/object/type
08:53:11	Araq	*come up
08:53:31	Araq	and why is that important?
08:53:41	Araq	we have the ':' to indicate "type following here"
08:53:56	DAddYE	for me is:
08:54:16	DAddYE	does that guy who wrote that wrapper used the prefix
08:54:36	Araq	again the convention is to skip the C prefixes
08:54:56	Araq	because you can do uv.setInterval()
08:54:56	DAddYE	but you can't create types without prefixes
08:55:06	Araq	yes you can
08:55:06	DAddYE	uv.ttimer
08:55:41	Araq	and you can also write tTime which may read better for you as it emphasizes the Time aspect better
08:56:01	DAddYE	I will kill my self before
08:56:01	Araq	but as I said
08:56:16	Araq	come up with a good proposal to "fix" this language wart
08:56:26	DAddYE	type/classes/etc in modern languages should be throated with Capital
08:56:26	DAddYE	Araq: okay
08:56:36	Araq	which it obviously is because I spent too much time arguing about it
08:56:36	DAddYE	will do this weekend
08:56:51	Araq	I have to go, see you later
08:56:51	DAddYE	Araq: https://github.com/DAddYE/gist/blob/master/uv.nim#L806-L807
08:57:21	Araq	what about it?
08:57:21	DAddYE	beside the fact I don't need do:
08:57:31	DAddYE	(was just a try)
08:57:41	DAddYE	doesn't work
08:58:01	Araq	write a bug report then
08:58:16	DAddYE	u're right
08:58:16	DAddYE	but not sure if is a bug
08:58:26	DAddYE	SIGSEGV: Illegal storage access. (Attempt to read from nil?)
08:58:56	DAddYE	since you talk about repr
08:58:56	DAddYE	on line https://github.com/DAddYE/gist/blob/master/uv.nim#L794
08:59:06	DAddYE	repr(time) raises the same problem
08:59:16	DAddYE	but
08:59:26	DAddYE	*timer
08:59:36	DAddYE	but I checked and is populated
08:59:36	DAddYE	so timer.timout => 1000
08:59:46	DAddYE	timer.data => pointer
08:59:46	DAddYE	and so on
08:59:56	DAddYE	I guess is GC thing
09:01:51	DAddYE	using discard works well
09:02:21	Araq	I don't think it's a Gc thing
09:02:31	Araq	but I'm away
09:02:41	DAddYE	oki
09:19:11	*	Associat0r quit (Quit: Associat0r)
09:53:11	*	DAddYE quit (Remote host closed the connection)
09:53:41	*	DAddYE joined #nimrod
09:57:51	*	DAddYE quit (Ping timeout: 245 seconds)
10:09:41	*	fowl quit (Quit: EliteBNC free bnc service - http://elitebnc.org)
10:29:16	*	fowl joined #nimrod
10:34:31	*	EXetoC joined #nimrod
10:36:31	*	q66 joined #nimrod
10:41:56	*	fowl quit (Quit: EliteBNC free bnc service - http://elitebnc.org)
10:55:42	*	DAddYE joined #nimrod
11:00:52	*	DAddYE quit (Ping timeout: 264 seconds)
11:12:37	*	Amrykid quit (Changing host)
11:12:37	*	Amrykid joined #nimrod
11:19:52	*	jbe_ joined #nimrod
11:21:42	dom96	hi jbe_
11:23:37	jbe_	hi dom96
11:23:57	jbe_	was just looking at your code
11:24:12	jbe_	what's the state of the sdl2 wrapper?
11:24:32	jbe_	or actually, sdl/sfml support in general
11:24:42	dom96	The state is pretty good I think.
11:25:57	jbe_	just started messing with Nimrod yesterday, love it so far
11:27:22	dom96	I haven't personally used fowl's sdl2 wrapper but i've used the sdl wrapper that is in the stdlib and it worked well. I think it wraps an older sdl version though.
11:27:42	jbe_	yeah
11:28:02	dom96	I've created a game using SFML for Ludum Dare and didn't have any problems with the wrapper
11:28:17	jbe_	cool, which one?
11:29:02	dom96	https://github.com/fowlmouth/nimrod-sfml
11:29:22	*	zahary quit (Quit: Leaving.)
11:29:52	dom96	jbe_: how did you find out about Nimrod?
11:30:02	*	nihathrael quit (Quit: Unknown Horizons - Open Source real-time strategy with the comfy 1602 feeling!)
11:30:22	*	nihathrael joined #nimrod
11:31:32	jbe_	hmm let me think... it was something very very random, i don't remember, i think i heard it mentioned somewhere and googled it :P
11:32:22	dom96	I see. I'm asking because i'm wondering what the best way to advertise nimrod is :P
11:36:22	jbe_	hmm.. i'm surprised the language isn't more famous, but i guess maybe that can be a good thing in the beginning..
11:36:52	jbe_	as soon as some awesome software gets written using it, perhaps...
11:37:22	dom96	I think most people don't use it because there is no big company behind it.
11:37:57	jbe_	good point :|
11:39:02	jbe_	could you help me understand a compile error?
11:39:22	jbe_	was just trying to use sdl_ttf from std
11:40:02	jbe_	scene/root.nim(45, 40) Error: type mismatch: got (PFont, cstring, TColor)
11:40:02	jbe_	but expected one of:
11:40:02	jbe_	sdl_ttf.RenderText_Blended(font: PFont, text: cstring, fg: TColor): PSurface
11:40:22	*	zahary joined #nimrod
11:41:42	dom96	yeah, that error is confusing.
11:41:52	dom96	(I've submitted an issue about on github IIRC)
11:42:02	dom96	in any case, I bet the problem is that you are passing a colors.TColor not a sdl.TColor
11:42:22	jbe_	oh man, of course.. thanks!
11:43:22	jbe_	i'm getting a little confused by module names some times
11:52:22	*	fowl joined #nimrod
11:52:52	*	zahary quit (Read error: Connection reset by peer)
11:53:02	*	zahary joined #nimrod
11:56:57	*	DAddYE joined #nimrod
11:57:22	*	q66 quit (Quit: Leaving)
11:58:22	*	q66 joined #nimrod
12:00:42	*	q66 quit (Client Quit)
12:01:22	*	q66 joined #nimrod
12:53:42	*	jbe_ quit (Read error: Operation timed out)
13:27:58	*	OrionPK joined #nimrod
13:31:53	*	Araq0 joined #nimrod
13:33:43	Araq0	Dom96 fix the logs
13:35:03	Araq0	People now depend on them
13:35:43	*	Araq0 only has a phone here...
13:35:43	dom96	I know
13:37:03	*	Associat0r joined #nimrod
13:37:03	*	Associat0r quit (Changing host)
13:37:03	*	Associat0r joined #nimrod
13:37:58	Araq0	Well? Whats the problem?
13:38:23	dom96	There is a bug.
13:38:33	dom96	I'm fixing it.
13:38:33	dom96	Chill man.
13:38:53	Araq0	Ok
13:38:53	Araq0	Bbl
13:39:03	*	Araq0 quit (Quit: Bye)
13:52:23	dom96	done
13:58:45	*	zahary quit (Quit: Leaving.)
14:01:14	*	gradha joined #nimrod
14:15:13	*	jbe_ joined #nimrod
14:26:14	*	zahary joined #nimrod
14:31:33	gradha	I'm trying to package a regex' parameters into a const tuple, to later create the object through re
14:31:56	gradha	so I have R_BLAH = (s: "somestring", flag: {reExtended})
14:32:22	gradha	then I want to write "let regex = re(R_BLAH)"
14:32:41	gradha	but it says the R_BLAH can't be called
14:33:18	gradha	is there any way to unpack it or do I suck it and pass R_BLAH[0] and R_BLAH[1]?
14:37:20	gradha	meh, [] doesn't do what I want it to do here, pity
15:03:11	gradha	oh, forgot, NimBot: high five!
15:04:32	dom96	I think you have to suck it up and do it the verbose way
15:04:52	gradha	maybe if I pester Araq enough...
15:05:56	gradha	in python you can do this prefixing the packed variable with *, maybe nimrod could figure this out at compile time and prevent the *, but then what if there's ambiguity
15:06:28	gradha	huh, maybe an unpack macro in stdlib?
15:07:05	gradha	the macro would replace PARAM with PARAM[0], ..., PARAM[n]
15:07:38	gradha	of course, how could I doubt the awesome nimrod, it's all there for me to play with
15:07:50	dom96	yeah, write a macro :P
15:08:43	gradha	hmm... the unpack would work for tuples and arrays with fixed lenght, but how to do it for seqs?
15:09:09	gradha	hmm... maybe if the macro could "know" it's being used inside a proc, and statically gather the input parameters, so it would us 0..n reading the proc's parameters...
15:11:21	gradha	is it possible for a macro to know it's parent in the AST where it is being invoked?
15:11:47	gradha	if that's possible you could write macros which modify outer scopes, scary
15:18:05	gradha	things that scare you more than death: you touch the power cord and the laptop suddenly says it's not plugged in
15:28:45	EXetoC	don't worry, we'll have immortality pills soon
15:29:21	gradha	should feed one to my laptop
15:31:08	EXetoC	get your priorities straight, man
15:31:42	gradha	right, download porn
15:31:55	EXetoC	give one to your freezer first
15:31:57	EXetoC	lol
15:45:15	Max00355	Can someone explain to me how to use os.walkDir()
15:45:30	Max00355	I can not seem to figure it out.
15:47:07	gradha	what kind of trouble you have?
15:47:24	gradha	here's an usage example of mine https://github.com/gradha/awesome_rmdir/blob/master/awesome_rmdir.nim#L42
15:48:32	Max00355	When you import os you don't call the function by doing os.walkDir?
15:49:01	gradha	that's only needed if you also imported something else or defined your own walkDir locally, so you need to desambiguate the call
15:49:23	gradha	if both modules a and b define the same proc, you would need to use a.proc b.proc if you import both
15:49:30	gradha	but if you only import a, it's fine without prefix
15:49:55	Max00355	Oh, cool
15:50:01	Max00355	Didn't know that.
15:50:10	gradha	the system module is imported by default, so you don't need to system.whatever for everything, unless you have a conflict
15:50:22	Max00355	Okay, I see.
15:50:42	Max00355	For Python is .split() or whatever for everything.
15:52:50	Max00355	So you can not echo within a function?
15:53:11	gradha	please ellaborate on that
15:53:38	Max00355	proc walk(path: string):
15:53:48	Max00355	for x in walkDir(path):
15:53:52	Max00355	echo(x)
15:54:02	Max00355	discard walk("/")
15:54:09	Max00355	You must return?
15:54:28	gradha	as you defined proc, it doesn't return anything, so no need for discard
15:54:40	gradha	should be "proc wal(path: string) =" though
15:54:40	Max00355	Ohhh, that's what discard is for
15:54:54	Max00355	Ah, another point, why =?
15:54:59	gradha	also, you can use the {.discardable.} pragma to tell the compiler not be stingy
15:55:12	gradha	the = is used for the proc body, the : is used for the return type
15:55:13	Max00355	Gotcha
15:55:25	Max00355	Sorry I am so noobish... I am just starting to really pick up Nimrod...
15:55:37	Max00355	I am not used to this static programming stuff and what not..
15:55:49	gradha	no problem, it's not hard really
15:56:17	Max00355	I know, but it is different.
15:56:25	gradha	I'm refactoring now some nimrod code and it's a *pleasure* to change a type and have the compiler tell you all the places its wrong, rather than fail at runtime some time later
15:56:46	Max00355	So the compiler is throwing an error that I have no return type declared/
15:57:08	gradha	likely due to the colon
15:57:09	Max00355	But I am not returning anything
15:57:20	Max00355	I took the colon away
15:57:28	gradha	did you replace it with the equal sign?
15:57:32	Max00355	I did
15:57:43	gradha	hmm.. something else is wrong then, can you post a snippet?
15:58:02	gradha	or is it just that proc walk as you wrote?
15:58:24	Max00355	Oh wait, I fixed it, my bad.
15:58:31	*	DAddYE quit (Ping timeout: 276 seconds)
15:58:32	Max00355	I was trying to return something.
15:58:46	Max00355	Okay, so when do you use a colon and when do you use an equal sign?
15:59:00	gradha	if your proc doesn't return, use an equal sign
15:59:14	Max00355	Ahhh, okay.
15:59:17	Max00355	Awesome.
15:59:18	gradha	if you return anything, you need "proc name(foo: type): return_type = body"
15:59:33	Max00355	Awesome
15:59:37	dom96	It's also a pleasure to see the types in the docs. It's annoying to have to scan the function's docstring to determine what should be passed to a function in Python.
16:00:04	gradha	dom96: when will Aporia display that info when completing something?
16:00:53	dom96	gradha: once I get some free time and get rid of my laziness :P
16:01:29	gradha	ah, cool, then don't buy deus ex fallen for ios, it's on sale now
16:01:45	dom96	I bought The Last of Us instead lol
16:02:59	gradha	cordyceps fungus, sounds like a programmer illness
16:03:10	Max00355	Does the developer of Nimrod ever come into this channel?
16:03:21	gradha	Max00355: nah, not gonna happen
16:03:42	Max00355	Haha
16:04:07	gradha	he never comes into this channel because he's always here (Araq)
16:04:41	dom96	Some say that the developer of Bitcoin has left the development of Bitcoin to pursue creating his own programming language :P
16:05:03	gradha	programming languages are clearly more profitable
16:05:35	Max00355	I really can't believe how much bitcoin exploded.
16:05:38	Max00355	It's incredible.
16:05:44	gradha	I wonder if his language will use the $ sign for something...
16:05:55	Max00355	PHP
16:05:57	Max00355	PERL
16:05:59	Max00355	GAY
16:06:08	Max00355	$_POST['name']
16:06:12	Max00355	I mean waht the fuck is that?!
16:06:19	Max00355	Global variables are nasty..
16:06:22	gradha	interesting, didn't know about the GAY programming language
16:06:30	Max00355	Really?
16:07:58	Max00355	Okay, so something I could never understand in any other language because I was spoiled by Python is random numbers.
16:08:50	gradha	what is to understand about random numbers? they are... random, so, like 4
16:09:02	Max00355	rand(100) returns 83 every time
16:09:13	gradha	mine is m/221/
16:09:20	gradha	sorry, https://xkcd.com/221/
16:09:47	Max00355	LOL
16:10:06	gradha	called rand(100) right now three times and got 7, 49, 73
16:10:13	gradha	clearly a poorer implementation of 4
16:10:24	dom96	gradha: Speaking of The Last of Us, it's pretty interesting that the zombie infection is caused by a fungus.
16:10:53	gradha	dom96: why is that interesting?
16:10:59	Max00355	Hey, that exists in the ant world!
16:11:22	gradha	in the human world it's maria which creates zombies, albeit temporarily
16:11:32	Max00355	Boom
16:12:38	Max00355	Malaria* btw
16:12:47	dom96	gradha: Dunno, never heard a zombie story which said that the infection is caused by a fungus.
16:13:17	gradha	never understood the people who try to rationalize zombies through pseudo science, it just doesn't make sense, better to think of zombies like something magic
16:13:33	Max00355	http://www.livescience.com/5631-zombie-ants-controlled-fungus.html
16:13:35	gradha	fungus, bacteria, virus, whatever
16:13:56	gradha	it's all just a gimmick to tell people "look, it's too complicated, treat it like magic", so better think it's magic anyway
16:14:12	dom96	Max00355: oh yeah. I remember reading that, forgot that it's caused by a fungus.
16:14:24	Max00355	Messed up shit isn't it?
16:15:54	dom96	gradha: Meh, magic. It's perfectly reasonable to consider this science fiction.
16:16:16	dom96	And if it can happen to ants, perhaps it can to humans as well?
16:16:48	Max00355	Let's hope it cant
16:16:56	Max00355	can't8
16:17:58	dom96	yeah, I prefer an apocalypse in the style of the Terminator.
16:18:09	dom96	Then maybe I can use my computer skills to fight the AI :P
16:18:36	gradha	or independence day, but you don't have a mac, so you would die
16:18:39	Max00355	Nimrod vs Skynet :P
16:18:55	dom96	lol independence day
16:19:03	gradha	or a dog, in independence day dogs survive impressive explosions
16:19:04	Max00355	gradha: Only the hipsters would be left.. my god..
16:19:11	*	Associat0r quit (Quit: Associat0r)
16:19:38	dom96	"let me just code this quick virus in Nimrod and upload it to the alien mothership!"
16:20:02	gradha	"darn, I didn't git pull before coming here, I'm doomed with this bug!"
16:20:48	dom96	"screw it, this segfaulting app will do well to disable the alien mothership"
16:21:48	gradha	oh, speaking about bugs, why don't you askubuntu how to compile nimrod?
16:22:19	gradha	you could ask like "hey, this compiles in other linux, why is ubuntu such crap?"
16:22:32	gradha	then you answer yourself how to solve it, like a sir
16:23:09	gradha	didn't know http://askubuntu.com was just stackoverflow disguised, heard about it the other day in their podcast
16:23:20	dom96	and then I would get banned
16:24:01	gradha	try to be polite then, "can somebody help me fix this shit, please?"
16:25:56	Max00355	What are you guys running
16:25:59	Max00355	?
16:26:19	gradha	I run with macosx, born hipster
16:26:28	Max00355	Oh my god...
16:26:35	Max00355	Disgusting
16:26:39	Max00355	dom96: You?
16:26:39	gradha	when macs become mainstream I'll jump to hurd
16:26:41	dom96	Arch Linux.
16:26:56	Max00355	gradha: They have been maib stream.
16:27:01	Max00355	dom96: Real men use Arch
16:27:10	dom96	:D
16:27:15	Max00355	So I am not a real man because I use Mint :P
16:27:23	dom96	lol
16:27:24	gradha	ios devices are mainstream, macs not so much, they are just a tiny drop of total computer machines
16:27:33	dom96	You'd be happy to hear that Araq uses Mint too :P
16:27:43	Max00355	Hahaha, well it is awesome.
16:27:59	dom96	I consider switching to a different distro sometimes though
16:28:10	Max00355	I tried Arch and hated it.
16:28:39	Max00355	It took me 4 hours to get everything the way I wanted, then I switched back to Mint in that same day :P
16:28:56	Max00355	I like stuff out of the box.
16:30:51	*	Associat0r joined #nimrod
16:30:51	*	Associat0r quit (Changing host)
16:30:51	*	Associat0r joined #nimrod
16:31:51	gradha	didn't nimrod have any proc to expand ~ in paths under unix?
16:32:22	dom96	getHomeDir()? :P
16:32:36	gradha	that's what I would call to replace the ~ in the string
16:33:05	dom96	dunno, look around in os.nim, if it doesn't exist there it likely doesn't exist at all
16:34:14	Max00355	I don't understand how Nimrod hasn't gotten more popularity.
16:34:36	gradha	people love crap
16:34:44	Max00355	Simple syntax that compiles to C...
16:34:54	Max00355	I wish I had heard about it when I started programming..
16:34:57	dom96	gradha: well put :P
16:35:09	Max00355	I would have made some really nice shit.
16:35:15	Max00355	I have hit a wall with Python.
16:35:20	Max00355	I can't do anything anymore.
16:35:36	gradha	did they ban you from programming?
16:35:38	dom96	Yep, you should rewrite all your projects in Nimrod.
16:35:52	Max00355	Heh, that's a lot of programs :P
16:36:07	Max00355	gradha: Yes, the Python community has shunned me.
16:37:00	gradha	Max00355: you are welcome here, we shun the python community instead
16:37:21	gradha	I'm also rewriting my python programs in nimrod
16:37:24	Max00355	Haha, I plan on staying for a bit.
16:37:33	Max00355	gradha: Do you have a github?
16:37:42	dom96	Max00355: You should tell all your friends about Nimrod!
16:37:47	Max00355	I have
16:37:51	Max00355	they laugh at me
16:37:56	dom96	what, why?
16:38:48	gradha	not hard to figure, https://github.com/gradha
16:40:22	gradha	nakefiles are nice, but the constant recompiling is really slow, to run a task it compiles twice
16:40:50	dom96	why twice?
16:40:51	gradha	well, I'm sure fowl will like me spamming issues
16:40:54	gradha	no idea
16:41:04	dom96	are you compiling the nakefile
16:41:07	gradha	I guess first nake compiles the nakefile, but then I see another compilation going on
16:41:17	dom96	and then running the nakefile which compiles something else?
16:41:32	gradha	yes, the first compilation is not needed
16:41:37	dom96	just compile the nakefile once?
16:41:49	gradha	and then use ./nakefile instead of nake? surely you jest
16:42:29	dom96	yep :P
16:42:54	gradha	I guess I'll replace nake in my path with a bash function which compares the date of nakefile.nim and nakefile and runs it or compiles it
16:45:43	gradha	ouch, copyFile doesn't preserve the execution bit for installation of binaries
16:52:14	EXetoC	most programmers obviously don't prioritize language features
16:55:10	*	DAddYE joined #nimrod
16:57:23	gradha	I've heard programmers prioritize java over alternatives due to having eclipse, which to me sounds so wrong on many levels
16:57:57	gradha	"here, this language is crap, but if you use these crutches, which are crap too, it suddenly becomes wonderful!"
16:59:38	EXetoC	most people probably just use whatever will get them employed some day
17:00:01	gradha	OTOH bundling language and IDE is a nice strategy, you get one holy war for the price of two
17:00:15	gradha	clever dom96, now I understand the meaning of Aporia...
17:00:16	*	DAddYE quit (Ping timeout: 276 seconds)
17:01:26	EXetoC	put it in a box together with a couple of OOP pamphlets
17:04:43	dom96	gradha: oh?
17:05:25	Max00355	dom96: They laugh because Pythonistas don't like static typing.
17:05:53	EXetoC	I got tired of dynamic typing after only a couple of days
17:06:10	Max00355	It is so much easier
17:06:23	gradha	dom96: a new planguage is pain, so better make it lang+ide so it is painful only once?
17:06:48	dom96	oh, I thought you were referring to the name
17:07:43	gradha	Max00355: it's only easier because it doesn't force you to think ahead of the data, yet that's what makes programs more efficient, or prevents out of range bugs and such
17:08:12	gradha	in fact, it's delying the inevitable
17:08:23	gradha	consider a python program which serializes to a database
17:08:39	gradha	you use a field to store integer values, and python will happily accept anything integer like, so a big number
17:08:56	gradha	but then you have to store it in sqlite, and wait, your table only accepts integers in a specific range...
17:10:23	gradha	then, dynamic typing is fine when you are writing programs, but can be a pain when you write libraries for other people
17:10:46	gradha	suddenly you have people not reading docs and calling your lib with parameters which are just not right, but due to coercion they kind of work and fail in obscure ways
17:11:02	gradha	so now you have to add to your functions manual parameter checks for the type, assertions and such
17:11:14	gradha	basically, manually implementing a type system
17:11:32	gradha	but of course, people who don't mind writing crappy code for others thing its ok to not care
17:13:19	gradha	I've yet to talk to a ruby programmer who checks input parameters for correctness or to avoid weird behaviours, they are like "what, why would I check the parameters?", lol
17:13:35	dom96	because Nimrod is so flexible with some hacky macro work you could probably actually implement dynamic typing :P
17:13:56	dom96	but it's probably not the best idea to be writing code like that.
17:14:15	gradha	the difference is allowing vs everything is dynamic
17:14:40	gradha	in objc you can never be sure of the type of input parameters, so to write really solid code you have to ask your parameters if they actually are the type they say to be
17:15:01	gradha	because everything is an object, and objects are wonderful things
17:43:56	Max00355	To each his own.
17:44:02	Max00355	I don't mind manual type checking.
17:44:12	Max00355	try:
17:44:25	Max00355	inp = input("> ")
17:44:30	Max00355	except TypeError:
17:44:35	Max00355	continue
17:44:37	Max00355	Not hard
17:45:46	gradha	well, string input sanitization doesn't have much to do with it
17:46:02	gradha	consider "def do_something(text):" in python
17:46:11	gradha	you can pass anything as text
17:46:12	Max00355	input() only take integers
17:47:32	Max00355	gradha: You can
17:47:33	gradha	Max00355: python docs say you are wrong, it returns whatever the string evals to
17:47:56	Max00355	Python3 docs say I am wrong, yes.
17:48:03	Max00355	Python 2.7 docs say otherwise
17:48:10	gradha	really? I'm on 2.7 here
17:48:17	gradha	input([prompt]) -> value
17:48:21	gradha	Equivalent to eval(raw_input(prompt)).
17:48:26	gradha	so no type information for value
17:48:34	gradha	raw_input returns a string
17:48:40	Max00355	input() returns an integer
17:48:53	Max00355	eval(raw_input(prompt)) is just silly
17:49:00	gradha	In [8]: repr(input(""))
17:49:00	gradha	"something"
17:49:00	gradha	Out[8]: "'something'"
17:49:06	gradha	that returned a string, AFAICS
17:49:27	Max00355	I gtg, i'll be on later, Bye!
17:49:32	gradha	see ya
17:56:01	*	DAddYE joined #nimrod
18:02:34	*	DAddYE quit (Ping timeout: 257 seconds)
18:08:54	*	jbe_ quit (Quit: Leaving)
18:24:11	*	Associat0r quit (Quit: Associat0r)
18:24:31	Araq	Max00355: I use the debian edition of mint though ;-)
18:24:54	Araq	and enjoy my X11 crashes ...
18:25:17	*	Araq just restarted X11 to get back a working mouse cursor
18:31:15	*	Associat0r joined #nimrod
18:31:15	*	Associat0r quit (Changing host)
18:31:15	*	Associat0r joined #nimrod
18:39:09	gradha	hah, found a but in parseHex
18:39:30	gradha	or rather, a weird behaviour, the proc depends on the input value of the temporal variable
18:39:44	gradha	so if I initialize the integer to something else, the resulting parsed value has nothing to do with input
18:41:02	gradha	Araq: how do you want me to "fix" that, document the behaviour or make parseHex initialize number before writing to it?
18:41:40	Araq	what's the difference between "fix" and "initialize number before writing to it"?
18:42:23	Araq	I think it shouldn't touch it in case of error; this way a default value is kept when a parsing error occurs
18:42:59	gradha	there's also a side benefit, you can concatenate several calls to parseHex and it works, though you quickly overflow
18:43:10	gradha	I'll document it
18:43:59	Araq	parseutils.parsehex, right?
18:44:05	gradha	yep
18:45:02	Araq	how does concatenate produce any meaningful result?
18:45:14	Araq	you need to shift it at least
18:45:24	gradha	parseHex("c3", temporal); parseHex("68", temporal) == parseHex("c368", temporal)
18:45:43	Araq	oh right
18:46:33	Araq	hmm
18:47:00	gradha	awesome, even parsing hex is fun in nimrod
18:48:10	Araq	well if it makes you happy, document it
18:48:24	Araq	I can't see a value in this "feature" for now though
18:48:54	Araq	on the other hand ... garbage in, garbage out is not too bad for a low level stdlib routine
18:51:51	gradha	oh, another unexpected behaviour in re.findAll
18:52:02	gradha	since it's an iterator, you shouldn't modify the string you are iterating over
18:52:17	gradha	for that it's better to toSeq the result
18:52:25	gradha	result was funny
18:54:07	Araq	hmm the compiler could easily warn about this
18:54:38	*	OrionPKM joined #nimrod
18:55:18	gradha	how can the compile detect such behaviour http://pastebin.com/1mhkQE6w ? usage of match variable?
18:55:44	gradha	aw, should use parseHex's start parameter instead of range
18:55:44	*	DAddYE joined #nimrod
19:02:48	Araq	gradha: simple: 'dest' is passed to an iterator so it's "const" for the for loop body
19:03:15	Araq	it's a very easy analysis
19:03:50	gradha	ok, doit
19:05:02	Araq	make a feature request :P
19:05:23	gradha	I thought people got irritated with me spamming issues
19:05:33	*	OrionPKM quit (Remote host closed the connection)
19:05:41	gradha	so I keep them now in my todo
19:11:20	Araq	what about closing issues and making a new issue "improve idetools"?
19:11:42	gradha	yeah, there's also that, unfulfilled expectations
19:12:31	DAddYE	Araq: do you think is the correct way to "instantiate" a pointer ?
19:12:33	DAddYE	Araq: proc timer_init*(a2: ptr TLoop; handle: ref TTimer)
19:12:35	DAddYE	sorry
19:12:41	DAddYE	Araq: https://github.com/DAddYE/gist/blob/master/uv.nim#L790-L794
19:13:01	Araq	ref TTimer is most certainly wrong
19:13:43	DAddYE	Araq: so, what do you suggest?
19:13:59	DAddYE	(unable to find any doc and I read 80% of all wrappers)
19:14:00	DAddYE	:D
19:14:07	Araq	dunno perhaps 'var TTimer'
19:14:19	DAddYE	Araq: here https://github.com/DAddYE/gist/blob/master/uv.nim#L677 ?
19:14:53	DAddYE	and then var t: TTimer = TTimer.new ?
19:15:02	DAddYE	timer_init(loop, t)
19:15:07	Araq	'new' is for 'ref' only
19:15:14	Araq	no 'ref', no 'new'
19:15:34	Araq	so it's only:
19:15:44	Araq	var t: ttimer; timer_init(loop, t)
19:15:58	Araq	though I would name it initTimer I think
19:16:22	Araq	but then you lose even more guessability in your wrapper
19:17:16	Araq	whether you can transform 'ptr T' to 'var T' mostly depends onto whether C allows for NULL
19:19:51	DAddYE	Araq: ok, thanks lemme try
19:22:41	DAddYE	Araq: in your way: type mismatch: got (TTimer) but expected 'ref TTimer'
19:22:55	DAddYE	proc timer_init*(a2: ptr TLoop; handle: var TTimer): cint {.libuv.}
19:24:25	Araq	don't declare your timer 'ref ttimer' then
19:24:46	DAddYE	Araq: I didn't var timer: TTimer
19:25:51	Araq	don't use TTimer.new then
19:26:51	DAddYE	Araq: I didn't!
19:27:18	Araq	well your gist doesn't contain any 'ref'
19:27:40	Araq	and if your example code doesn't either I can't see why the compiler would say "expected 'ref TTimer'"
19:30:55	DAddYE	Araq: https://gist.github.com/DAddYE/9f2ec7a876164d8fe93f
19:31:05	DAddYE	fuck
19:31:06	DAddYE	found
19:31:07	DAddYE	:D
19:32:44	Araq	btw you shouldn't use "block" in a template unless you know what you're doing and know about a codegen bug
19:33:01	DAddYE	Araq: infact SIGSEGV: Illegal storage access. (Attempt to read from nil?)
19:33:11	DAddYE	I don't what to share with the rest of my code
19:33:13	DAddYE	my vars
19:33:34	Araq	but since you don't use a dirty template it isn't
19:33:59	Araq	no need for 'block' here anymore
19:34:13	Araq	maybe we need to update the docs somewhere ...
19:35:04	DAddYE	Araq: haaaaaaaaaaaaa
19:35:45	DAddYE	Araq: http://build.nimrod-code.org/docs/manual.html#scoping-in-templates_toc
19:36:02	DAddYE	http://build.nimrod-code.org/docs/manual.html#scoping-in-templates
19:36:23	DAddYE	block only when is immediate
19:36:26	DAddYE	or dirty?
19:36:42	DAddYE	also there is no enough doc about {.dirty.} and {.inject.}
19:36:54	DAddYE	so I've few ideas what they really are
19:37:18	Araq	well if you use .dirty everything is injected
19:37:50	Araq	if you don't use dirty you need to explicitly inject symbols (unless in certain contexts)
19:38:27	Araq	immediate refers to overloading resolution, you can't overload an immediate template
19:38:33	DAddYE	btw removed block:
19:38:41	Araq	this however enables other possibilities
19:38:55	Araq	like passing a non existing symbol to the template
19:39:19	DAddYE	cool
19:39:33	Araq	most people use .immediate, dirty and call it a day though
19:39:47	Araq	this way you get quite C-like macros
19:40:02	DAddYE	Araq: removing block still have same sigsev illegal storage access
19:40:15	DAddYE	seems timer var get gc'd
19:40:44	Araq	so it fails at runtime?
19:41:31	Araq	be careful with the "proc" type please, you want "proc ()" instead
19:41:50	Araq	and yeah we should fix that, zahary
19:42:03	DAddYE	var t = set_timer(1000, 1000) do:
19:42:31	DAddYE	SIGSEGV: Illegal storage access. (Attempt to read from nil?)
19:42:35	DAddYE	at runtime
19:42:50	Araq	hmm
19:42:51	DAddYE	if I'll change it with discard set_timer ... works
19:43:45	Araq	well ... you got rid of the 'block', right?
19:44:00	DAddYE	yep removed
19:44:36	DAddYE	Araq: https://gist.github.com/DAddYE/9f2ec7a876164d8fe93f
19:49:23	DAddYE	tried also returning var TTimer but no luck
19:50:04	DAddYE	aaaaaaaaaaarg
19:50:07	DAddYE	I think I found
19:50:17	DAddYE	looks the compiler think timer is still NULL
19:50:20	DAddYE	or some weird
19:50:26	DAddYE	echo repr(timer)
19:50:38	DAddYE	=> crash
19:51:00	DAddYE	but echo timer.my_fields works
19:51:05	DAddYE	with updated results
19:51:42	DAddYE	that's why I did in past example TTimer.new
19:51:45	DAddYE	to avoid that problem
19:52:00	Araq	dunno it's weird
19:52:28	Araq	make a bug report
19:52:44	Araq	workaround should be to change your set_timer template
19:53:15	Araq	so that it takes the ttimer to declare
19:54:27	Araq	however ...
19:54:39	Araq	you do copy the ttimer maybe that doesn't work
19:55:40	DAddYE	Araq: here my example
19:55:41	DAddYE	https://gist.github.com/DAddYE/9f2ec7a876164d8fe93f
19:55:45	DAddYE	with the output
20:00:24	Araq	repr(timer) crashing doesn't mean much I think
20:00:38	Araq	it follows some pointer it shouldn't follow or something like that
20:01:01	DAddYE	ok
20:01:21	DAddYE	last to things and I'll stop to boring you
20:01:22	DAddYE	:D
20:01:41	DAddYE	can I have templates / procs with same name but different signature ?
20:01:49	Araq	yes
20:02:06	DAddYE	like template set_timer(timeout: uint64, repeat: uint64, timer: expr, actions: proc ()) =
20:02:12	DAddYE	and template set_timer(timeout: uint64, repeat: uint64, actions: proc ()) =
20:02:16	DAddYE	?
20:05:30	DAddYE	uv.nim(803, 12) Error: wrong number of arguments
20:09:14	Araq	they both can't be immediate for this to work
20:09:48	DAddYE	they aren't
20:12:18	Araq	works for me
20:13:58	DAddYE	Araq: https://gist.github.com/DAddYE/9f2ec7a876164d8fe93f
20:14:37	Araq	DAddYE: firstly one IS immediate
20:14:47	DAddYE	tried now
20:14:48	DAddYE	without
20:14:52	Araq	secondly you use set_timer version 2 before it's been declared
20:15:01	DAddYE	uv.nim(789, 32) Error: undeclared identifier: 't'
20:15:23	Araq	well yeah it can't work
20:15:35	DAddYE	0-0
20:15:59	DAddYE	ok I see why
20:16:04	DAddYE	I need to dup the code
20:16:10	DAddYE	inside my template
20:16:14	Araq	you know ... for overloading resolution to work the compiler needs *types*, an undeclared identifier has no type
20:16:44	Araq	so it can't work for your use case
20:17:00	DAddYE	sorry but
20:17:03	DAddYE	so why here works
20:17:05	DAddYE	set_timer(1000, 1000, t) do:
20:17:07	DAddYE	?
20:17:29	Araq	you can get lucky thanks to a somewhat hard to fix compiler bug ...
20:17:53	DAddYE	haaaaaaaa
20:18:00	DAddYE	okay here I'm
20:20:37	DAddYE	your compiler doesn't like me
20:20:53	DAddYE	if I put {.immediate.} on the second template
20:20:54	DAddYE	echo("Interval every 1s")()' cannot be called
20:20:57	DAddYE	-.-
20:24:08	Araq	you need to do 'actions' for immediate and actions() for non-immediate I think
20:24:27	DAddYE	Araq: yep figured now
20:24:58	DAddYE	doest it looks okay to you now
20:25:00	DAddYE	https://gist.github.com/DAddYE/9f2ec7a876164d8fe93f ?
20:26:24	Araq	no; as I said you merely get lucky if it works
20:26:45	Araq	the compiler shouldn't allow immediate and non-immediate overloads
20:27:31	DAddYE	ok
20:30:18	DAddYE	Araq: why in some example you returns stmt ?
20:30:25	DAddYE	like here http://build.nimrod-code.org/docs/manual.html#scoping-in-templates
20:30:34	DAddYE	there is a particular reason behind ?
20:31:58	DAddYE	maybe you can chain
20:39:53	Araq	it's mostly for historical reasons
20:40:18	Araq	': stmt' and ': void' and nothing all mean the same for template declarations
20:41:28	DAddYE	Araq: haaaaa ok
20:41:28	DAddYE	Araq: https://gist.github.com/DAddYE/9f2ec7a876164d8fe93f
20:41:28	DAddYE	this works
20:41:38	DAddYE	tell me that looks so good
20:41:38	DAddYE	to you
20:41:53	DAddYE	that you want to start coding in UV
20:41:53	DAddYE	:D
20:42:28	Araq	I would simply rename the one set_timer to decl_timer which makes more sense anyway
20:42:49	DAddYE	Araq: yep, totally agree
20:42:59	DAddYE	Araq: thanks for your time
20:44:39	Araq	you're welcome
20:44:54	Araq	now push your wrapper to babel
20:45:19	*	DAddYE quit (Remote host closed the connection)
20:49:59	gradha	dom96: nimbuild is having problems with the ftp upload, has something changed on the server?
20:50:29	gradha	I can telnet into port 21, but trying to use any ftp client just times out
20:50:49	gradha	I wonder if this is one of those ftp port thingies which have firewall problems
20:51:19	gradha	OTOH I don't have any problems connecting to an italian ftp I manage
20:52:49	Araq	port 21 for FTP is the standard though, right?
20:52:59	dom96	that's odd
20:52:59	gradha	yes, but ftp always opens a secondary port for data
20:53:34	dom96	the secondary port is only opened when data transfer begins
20:53:54	dom96	But you can't even connect with an ftp client, right?
20:54:04	gradha	I can't ls into it
20:54:14	gradha	nimbuild fails with Error: unhandled exception: Socket is not connected [EOS]
20:54:14	gradha	or Error: unhandled exception: Expected reply '226' got: 421 Data timeout. Reconnect. Sorry. [EInvalidReply]
20:54:29	*	DAddYE joined #nimrod
20:54:39	DAddYE	Araq: since u have a wrapper in the code base
20:54:49	DAddYE	isn't better update it
20:54:49	DAddYE	?
20:54:59	Araq	do I?
20:55:09	dom96	well I just noticed that nimbuild crashed
20:55:09	gradha	I'd understand if I could not connect to any ftp, but being able to connect to some makes it strange
20:55:59	Araq	DAddYE: but your wrapper might be incompatible, better make the stdlib's deprecated and refer to your babel package instead
20:56:14	DAddYE	Araq: okay make sense
20:56:14	Araq	we have this problem for other things too
20:56:39	Araq	IUP, GTK come to mind
20:56:39	gradha	dom96: looks like I can log in and ls with the ftp command, but the fancy lftp times out
20:56:59	dom96	lftp?
20:57:09	gradha	lftp - Sophisticated file transfer program
20:57:29	dom96	oh
20:57:39	gradha	so much for sophistication that it breaks connections
20:58:14	gradha	but yeah, something is preventing nimbuild from uploading the data, which is weird
20:59:19	dom96	ugh
20:59:19	gradha	using ftp I was able to upload a zero bytes log.txt file, though I can't remove it
20:59:39	gradha	during the ftp upload it said "entering passive mode"
20:59:39	dom96	Nimbuild doesn't deal well with builders reconnecting during a build
21:00:19	DAddYE	Araq: in my gist can't I change templates to be proc instead?
21:01:19	Araq	you can for the non-immediate one but it might cost you a closure
21:03:09	DAddYE	ok
21:03:34	dom96	In fact, nimbuild should never crash because the builder's send some incorrect data.
21:03:44	*	dom96 adds a try except and calls it a day
21:03:59	gradha	since you are at it add another one around the ftp data upload part
21:04:09	Araq	dom96: what about the raises: [] annotation?
21:04:39	dom96	gradha: ok
21:04:39	*	Araq slaps dom96 for adding try except and calling it a day
21:04:54	dom96	Araq: It's better than creating a restart script
21:05:04	dom96	which ignores ALL crashes
21:05:04	Araq	I disagree
21:05:19	Araq	a restart script can log the error reliably at least
21:05:49	Araq	and the memory protection of processes is nothing to give up easily
21:06:14	dom96	And Nimbuild can't?
21:06:39	Araq	well
21:06:59	Araq	*I* would write a nimrod restart program instead of a crappy shell script
21:07:14	dom96	Actually my plan isn't ideal. The builder should really terminate when this occurs.
21:07:29	dom96	Otherwise it may continuously be sending the same bad data.
21:07:59	dom96	Araq: ok, write me one then :P
21:09:19	Araq	I'm too busy with other things
21:09:29	dom96	yeah, me too.
21:14:19	dom96	oh cool, you get a watcher count on github now.
21:14:39	dom96	"Account for Github research", interesting.
21:15:29	Araq	hu? "watcher count"?
21:15:39	Araq	shouldn't that be a "stargazers count"?
21:16:09	dom96	There is a difference
21:16:39	Araq	yeah; there is no such thing as a "stargazer"
21:17:34	EXetoC	beep beep
21:17:54	Araq	oh wait ... there is ... lol
21:18:09	Araq	it's a real english word ... interesting
21:28:19	dom96	gradha: I'm guessing you want the builder to continue even when ftp upload fails
21:29:19	dom96	not sure I can do that.
21:29:49	dom96	I would have to tell the website that ftp upload succeeded even though it hasn't
21:30:34	gradha	wouldn't it be better if it tried to retry? or notify you of the error somehow
21:30:55	*	DAddYE quit (Remote host closed the connection)
21:31:33	dom96	yeah, well currently it crashes right?
21:31:43	gradha	yes, uncaught exception
21:31:59	gradha	I don't mind really, I use the infinite shell script approach
21:32:20	gradha	but maybe if the upload was kept it wouldn't try to recompile it, or advance it further
21:32:30	gradha	it seems that compilation and upload are separate things
21:33:02	dom96	yeah, but then you don't get the testresults
21:34:08	gradha	don't understand, imagine instead of ftp you used dropbox, putting files in a shared folder. nimbuild compiles, then moves the files somewhere, and keeps going
21:34:23	gradha	with the ftp it would keep going, and the upload would be left to a secondary nimbuild helper
21:35:00	dom96	yes, but that requires many changes to the code.
21:35:51	gradha	so I have now 33MB in the builds directory, useless?
21:36:08	dom96	yeah
21:36:24	dom96	there is no point in keeping the old builds
21:37:09	gradha	I guess there's no point in me starting nimbuild again since ftp upload doesn't work now
21:37:16	gradha	curse you FTTH!
21:40:19	dom96	hrm, I wonder what happens if I put a proc definition in a try except?
21:41:19	Araq	nothing? a proc definition is not executable code
21:41:58	dom96	well just in case I left it out of my try block
21:42:33	NimBot	nimrod-code/nimbuild master 3378960 Dominik Picheta [+0 ±1 -0]: Builder will no longer crash if FTP upload failed.
21:48:07	dom96	gradha: well I can successfully connect to the ftp server
21:48:57	gradha	I can too, with passive ftp, but nimbuild does something else which doesn't work for me now
21:50:13	dom96	perhaps nimrod's ftpclient module does not support passive ftp
21:50:21	*	dom96 can't remember
21:51:07	gradha	the thing is, I can still use lftp with an italian ftp, so it's not the passive stuff, but maybe something related
21:51:31	gradha	if suddenly "active" ftp was problematic I would be unable to connect to that other server too
21:53:46	gradha	ok, so lftp works when I specify a username in the login, but not if I first open the server then try to authenticate
21:54:48	gradha	so now the question remains: why is nimbuild refusing to upload if all other ftp tools work?
21:55:35	gradha	ok, recompiled the builder and starting another try
21:55:38	*	EXetoC quit (Read error: Connection reset by peer)
21:59:14	Araq	gradha: [23:58] NimBot macosx-x86_64: Build OK.
21:59:25	Araq	you should join #nimbuild
22:00:18	*	EXetoC joined #nimrod
22:20:36	*	gradha quit (Quit: bbl, need to watch https://www.youtube.com/watch?v=1ZZC82dgJr8 again)
22:30:59	fowl	Gradha
22:31:35	fowl	Put that nake stuff into nake.nim not a new bash script
22:31:58	fowl	Wth son.. lol
22:32:07	fowl	I don't know if ur even here
23:34:11	*	Associat0r quit (Quit: Associat0r)
23:48:16	*	Trix[a]r_za is now known as Trixar_za