<< 14-04-2015 >>

---

00:07:04	*	BlaXpirit quit (Quit: Quit Konversation)
00:08:11	*	Jehan_ quit (Quit: Leaving)
00:09:11	*	user7181 quit (Quit: leaving)
00:22:59	reactormonk	transfuturist, IRC logs :-)
00:27:36	transfuturist	;_;
00:28:55	transfuturist	oh, i can see my crying face
00:29:16	transfuturist	is there a way to search them?
00:29:49	reactormonk	apparently not. time for some wget?
00:30:12	reactormonk	but for compiler stuff, you can just poke araq when he's around
00:30:31	reactormonk	I've done some stuff with the compiler too, but not too much
00:33:45	*	mwbrown joined #nim
00:34:08	*	Demon_Fox joined #nim
00:35:20	transfuturist	reactormonk: what do you know about dynlib handling?
00:36:00	flaviu	transfuturist: I use "site:irclogs.nim-lang.org <my query here>" in google.
00:36:24	transfuturist	oh yeah, thanks
00:37:08	reactormonk	transfuturist, dynlib as in runtime shared lib handling?
00:38:05	transfuturist	mhm
00:38:54	transfuturist	apparently dynamic libraries are loaded at the beginning of runtime, i want to put a greater degree of control in dynlib.nim
00:40:00	transfuturist	i think dynlib.nim might have been created entirely for opengl's benefit, stuff like loading libraries and looking up symbols has no actual use outside of opengl's complicated import system
00:40:18	reactormonk	what do you want to change?
00:41:27	transfuturist	make it so you can load and unload libraries and have the "api" functions declared with the dynlib pragma update when you do so
00:41:34	transfuturist	for hot swapping
00:45:45	*	gokr left #nim (#nim)
00:45:57	transfuturist	i'm fairly unsure of my ability to understand the compiler source and make the changes i need, but i don't see any progress along these lines since it's been brought up on the forums
00:46:38	reactormonk	transfuturist, I'd start by finding the dynlib stuff in the cgen and see if it's anywhere else
00:47:12	transfuturist	i already have, i'm just not sure what it means
00:47:29	reactormonk	mostly in pragmas.nim
00:47:43	transfuturist	and cgen.nim
00:47:47	transfuturist	but there's some weird stuff
00:48:36	transfuturist	pragmas.nim adds libraries to a linked list in PContext, but it's mentioned nowhere else. is that just dead code?
00:51:05	transfuturist	oh, it's just used to remove duplicates from the list that's never used otherwise.
00:51:07	transfuturist	??????
00:53:30	reactormonk	transfuturist, remember that nim is partially style-insensitive
00:53:50	transfuturist	how else can it be accessed?
00:54:02	reactormonk	just saying you should probably grep -i
00:56:59	transfuturist	probably a red herring
00:57:30	reactormonk	not really, I would assume cgen uses that list to generate the import statements
00:57:58	transfuturist	even case-insensitive, it's nowhere else
00:58:03	reactormonk	hm
00:58:07	reactormonk	useHeader(p.module, s)
00:58:34	reactormonk	that's where the headers are.
00:59:13	transfuturist	what is BModule?
00:59:47	reactormonk	btw, what editor are you using? you could try to get nimsuggest to help
00:59:57	transfuturist	sublimetext
01:00:40	reactormonk	BModule* = ref TCGen
01:01:36	*	wink-s joined #nim
01:02:35	reactormonk	apparently it's all the data in a single C file
01:03:41	transfuturist	cfsDynLibInit, # section for init of dynamic library binding
01:04:21	transfuturist	okay, loadDynamicLib is starting to make sense
01:05:43	transfuturist	it generates and adds the necessary C for the initialization of a given dynamic library
01:05:50	transfuturist	so that whole thing would have to change
01:11:34	*	dhasenan_ quit (Ping timeout: 272 seconds)
01:21:55	*	gsingh93 quit (Quit: WeeChat 1.1.1)
01:29:59	*	dhasenan quit (Remote host closed the connection)
01:32:50	*	xtsee joined #nim
01:32:54	*	xtsee left #nim (#nim)
01:36:06	*	dhasenan joined #nim
01:44:44	*	vendethiel joined #nim
01:58:15	onionhammer	Varriount ahh interesting. I'll take a look
02:09:07	*	brson quit (Quit: leaving)
02:09:21	transfuturist	are typedefs like LibHandle tracked in the VM?
02:10:02	transfuturist	can a function in pragmas.nim check to see if a type of kind pointer is different from other pointer kinds?
02:15:04	*	jholland quit (Quit: Connection closed for inactivity)
02:16:47	transfuturist	yes; type.sym.name.s contains the identifier string
02:17:26	*	vendethiel quit (Ping timeout: 272 seconds)
02:17:38	*	mwbrown quit (Ping timeout: 252 seconds)
02:21:00	*	jefus quit (Ping timeout: 276 seconds)
02:21:01	*	darkf joined #nim
02:21:01	*	darkf quit (Changing host)
02:21:01	*	darkf joined #nim
02:21:21	*	pigmej quit (Ping timeout: 244 seconds)
02:29:06	*	pigmej joined #nim
02:31:19	*	Pisuke quit (Read error: Connection reset by peer)
02:31:48	*	Sembei joined #nim
02:32:23	*	fungos quit (Remote host closed the connection)
02:33:31	onionhammer	Varriount I think I might start a rewrite of the syntax def at some point for this.. not sure how we can support ST2 though
02:35:58	onionhammer	Varriount https://github.com/Varriount/NimLime/issues/22
02:38:09	*	EXetoC quit (Ping timeout: 256 seconds)
02:48:14	*	heinrich5991 quit (Ping timeout: 419 seconds)
02:48:33	*	Demon_Fox quit (Excess Flood)
02:48:59	*	heinrich5991_ joined #nim
02:49:11	*	Demon_Fox joined #nim
02:50:54	*	heinrich5991_ is now known as heinrich5991
02:53:16	*	brson joined #nim
03:01:45	*	fowl joined #nim
03:02:32	*	elbow_jason joined #nim
03:05:36	*	mwbrown joined #nim
03:05:41	*	dom96_ joined #nim
03:06:50	*	buMPnet quit (*.net *.split)
03:06:51	*	zahary quit (*.net *.split)
03:06:51	*	untitaker quit (*.net *.split)
03:06:51	*	ob_ quit (*.net *.split)
03:06:51	*	EastByte quit (*.net *.split)
03:06:51	*	mnemonikk quit (*.net *.split)
03:06:51	*	reactormonk quit (*.net *.split)
03:06:51	*	Joe-T quit (*.net *.split)
03:06:51	*	Guest67437 quit (*.net *.split)
03:06:51	*	dom96 quit (*.net *.split)
03:13:08	*	buMPnet joined #nim
03:13:08	*	zahary joined #nim
03:13:08	*	untitaker joined #nim
03:13:08	*	ob_ joined #nim
03:13:08	*	EastByte joined #nim
03:13:08	*	mnemonikk joined #nim
03:13:08	*	reactormonk joined #nim
03:13:08	*	Joe-T joined #nim
03:13:08	*	Guest67437 joined #nim
03:13:33	*	untitaker quit (Max SendQ exceeded)
03:13:33	*	buMPnet quit (Max SendQ exceeded)
03:13:40	*	untitaker_ joined #nim
03:14:44	*	buMPnet joined #nim
03:15:24	*	elbow_jason quit (Remote host closed the connection)
03:16:08	*	elbow_jason joined #nim
03:18:42	*	BitPuffin quit (Ping timeout: 244 seconds)
03:19:05	*	elbow_jason quit (Remote host closed the connection)
03:19:50	*	elbow_jason joined #nim
03:34:41	*	ChrisMAN quit (Ping timeout: 265 seconds)
03:45:32	reactormonk	Araq, you mind s-expressions in the stdlib? would make emacs integration a few magnitudes easier
03:45:38	reactormonk	... for nimsuggest
03:48:02	Varriount	onionhammer: Hm. If only there were some way to backport that particular change.
03:49:21	Varriount	onionhammer: I wouldn't semi-deprecation of the current language files, putting improvements of those as a low priority
03:50:01	onionhammer	what do you mean by the 2nd line?
03:56:31	*	dtscode is now known as charmander
03:56:35	*	charmander is now known as dtscode
03:59:06	*	dtscode is now known as squirtle
03:59:42	Varriount	onionhammer: I mean, focus on updating the newer syntax files.
04:00:26	Varriount	With contexts, we should be able to highlight more specific areas of code.
04:08:36	*	dhasenan_ joined #nim
04:09:57	transfuturist	reactormonk: how does this look? http://pastebin.com/dFzZaqqN
04:11:01	reactormonk	transfuturist, I think you're too much into objects via self: dlHandle. Also, types should begin with an upercase letter.
04:11:45	reactormonk	lemme upll up dynlib
04:13:04	transfuturist	it needs to be an object so it can track the symbol loading function.
04:13:59	reactormonk	I'd rename it to "handle" not "self", but that's just me, you can keep it self.
04:14:18	reactormonk	But the dlHandle -> DlHandle is less optional
04:14:24	transfuturist	the implementation can just be a consensus thing
04:14:31	transfuturist	but sure, i changed the case
04:15:35	reactormonk	so you wait the procs to run at compiletime, correct?
04:15:46	transfuturist	eh?
04:15:55	reactormonk	oh, ups. misread.
04:17:55	reactormonk	transfuturist, you don't need to unload stuff like self.handle = nil, the GC takes care of that
04:18:12	reactormonk	ehh, that one's probably correct. But no need for self.path = nil
04:18:59	Varriount	transfuturist: Also, be aware that objects are not references, they copy when assigned
04:19:12	reactormonk	Varriount, yeah, he went for references
04:19:14	transfuturist	shit
04:19:37	reactormonk	transfuturist, should work as is.
04:19:44	transfuturist	oh, a ref of an object is ok
04:19:58	reactormonk	yeah, gotta be explicit about reference semantics
04:20:33	Varriount	Although, it's worth pointing out that even objects are usually passed via pointer when calling a procedure.
04:22:00	reactormonk	transfuturist, the concept works, but the API should not need the loadDL part. pragmas can also be macros.
04:22:57	transfuturist	the loadDL isn't meant to be run as part of the pragma, it's so you can unload and reload as you wish
04:22:58	reactormonk	transfuturist, but if you only care about the segfault, just patch loadLib to throw instead of returning nil
04:23:28	transfuturist	this is for hot swapping shared libraries
04:23:58	reactormonk	Hm. what advantages do you gain over unloadLib directly?
04:24:17	reactormonk	oh right, you don't have the handle.
04:24:28	transfuturist	the dynlib package as is is unable to hot swap
04:24:59	reactormonk	could you do it with the LibHandle?
04:25:12	transfuturist	no, it's been brought up on the forums
04:25:27	reactormonk	link plz?
04:25:31	transfuturist	you can't cast a pointer you get from symAddr to a proc
04:25:49	*	endragor joined #nim
04:26:07	*	endragor quit (Remote host closed the connection)
04:26:27	transfuturist	hang on, it's in one of these...
04:26:27	reactormonk	ok, then I see why like this, fine with me.
04:27:03	reactormonk	actually, I wouldn't store the path in the DlHandleObj at all.
04:27:07	transfuturist	http://forum.nim-lang.org/t/627
04:27:58	*	endragor joined #nim
04:28:32	reactormonk	lemme check that code
04:29:10	Varriount	transfuturist: Do you usually work with Python?
04:29:25	transfuturist	nuh
04:29:50	transfuturist	i thought i saw the self syntax somewhere related to nim
04:31:34	Varriount	Hello endragor
04:31:48	transfuturist	reactormonk: http://pastebin.com/rTpAHCfK
04:33:09	reactormonk	transfuturist, https://gist.github.com/4dbb00c63c019f11c0ce compiles
04:34:11	transfuturist	you must be a wizard.
04:34:12	reactormonk	even works with nimcall
04:34:33	reactormonk	transfuturist, nope, I just read the forum post and I know what he means by calling convention
04:35:04	*	mwbrown quit (Ping timeout: 255 seconds)
04:35:23	transfuturist	thank god
04:35:34	transfuturist	i don't have to do this compiler hacking bullshit anymore ;_;
04:36:24	transfuturist	just need to see if the cdecl pragma will work outside of the typedef
04:37:01	reactormonk	I'd go for nimcall instead
04:37:15	reactormonk	it's more standard and gives the compiler more freedom to optimize stuff
04:37:26	transfuturist	oh?
04:37:34	transfuturist	i wasn't aware of an actual call function
04:37:50	reactormonk	hm? It's just the calling convention of the actual proc
04:38:00	reactormonk	http://nim-lang.org/manual.html#procedural-type
04:38:30	reactormonk	I think we had an XY problem here :-)
04:38:30	transfuturist	now i'm not sure what you mean anymore
04:38:51	reactormonk	you were asking for X to do Y without actually stating Y ^^
04:38:56	transfuturist	so wait, no pragma should be necessary?
04:39:06	reactormonk	nope, you need one. the documentation is incorrect.
04:39:10	transfuturist	okay
04:39:16	transfuturist	that would explain why it wasn't working
04:39:23	reactormonk	lemme change that
04:39:40	reactormonk	documentation incorrect is evil.
04:43:00	reactormonk	transfuturist, wait. The documentation is correct.
04:43:07	transfuturist	>>
04:43:17	reactormonk	just had to read it again :-)
04:43:37	reactormonk	if you have a proc like proc foo(x: int): int = x + 3 then the calling convention is nimcall
04:43:57	reactormonk	if you have type Foo = proc(x: int): int, then the calling convention is closure
04:44:12	reactormonk	... see why it makes sense?
04:44:35	transfuturist	...no
04:44:41	transfuturist	still trying to get your example to work
04:45:27	reactormonk	if you have a proc foo() ... you don't need the closure. But with Foo = proc() you most likely have a callback where you expect to be able to use closure callbacks
04:46:07	transfuturist	jumpin' jehoshaphat
04:46:36	transfuturist	don't need the typedef either
04:46:39	transfuturist	thanks!
04:46:48	transfuturist	XY problem indeed...
04:47:17	reactormonk	I think I'll add the examples to either
04:47:59	reactormonk	actually, not needed. The procedural types are explained above.
04:48:39	*	pregressive quit (Remote host closed the connection)
04:53:38	transfuturist	is there a way to create a statement macro that takes a string to begin with, like 'custom_def "foo":'?
04:55:41	reactormonk	transfuturist, take desired API, write inside dumpTree
04:56:00	reactormonk	(don't forget import macros)
04:57:29	reactormonk	looks like yes.
04:57:30	transfuturist	d'oh
04:58:37	transfuturist	not sure how that could translate into macro language, though
04:59:09	reactormonk	according to the tree first argument is "foo", the second is the stmtlist
04:59:54	transfuturist	is this because of the unified calling syntax thing?
05:08:42	reactormonk	not really
05:09:00	transfuturist	regardless
05:09:06	transfuturist	nim's macros are awesome
05:14:21	reactormonk	yup
05:27:20	*	transfuturist left #nim ("snoozing...")
05:36:18	*	endragor quit ()
05:36:24	*	endragor_ joined #nim
05:37:11	*	brson quit (Quit: leaving)
05:51:43	*	HakanD_ joined #nim
05:56:04	*	HakanD_ quit (Ping timeout: 245 seconds)
05:56:08	reactormonk	gotta say continuations would make the json parser API way more resilient
06:23:51	*	jefus joined #nim
06:23:53	reactormonk	Araq, you can pretty much see whenever I work on nim seriously based on my issues/h rate.
06:39:25	*	EXetoC joined #nim
06:44:20	*	HakanD_ joined #nim
06:44:57	*	HakanD__ joined #nim
06:48:45	*	HakanD_ quit (Ping timeout: 250 seconds)
06:56:34	*	bcinman quit (Quit: My Mac has gone to sleep. ZZZzzz…)
06:59:14	reactormonk	pigmej, https://github.com/reactormonk/nim/tree/epc
07:00:56	pigmej	reactormonk: hey :)
07:01:12	pigmej	oh sexps :)
07:01:32	reactormonk	yup, should work as expected. doesn't parse (foo . bar) though
07:02:13	reactormonk	but I haven't found that prin1 produces them
07:03:02	*	endragor joined #nim
07:03:21	reactormonk	hm, fuck it does if you give it alists.
07:03:55	pigmej	anyway, why it needs to be in stdlib ?
07:03:59	pigmej	the sexp
07:04:05	reactormonk	because you need it in nimsuggest
07:04:36	pigmej	yaeh, but we could distribute separate tool
07:04:49	reactormonk	but then you need to piggy-back on nimsuggest somehow
07:04:49	pigmej	that is compiled on el package installation
07:04:51	reactormonk	that's ugly
07:05:13	pigmej	yeah, I asked Araq yesterday what he thinks about making nimsuggest a library
07:05:38	pigmej	then 'problem' would be solved probably ?
07:05:38	reactormonk	for now it's in the stdlib. as soon as nimsuggest goes into a lib, I'll happily rip sexp out.
07:06:01	pigmej	you know, those are just my thoughts ;-)
07:06:43	*	endragor_ quit (Ping timeout: 255 seconds)
07:08:24	reactormonk	epc allows for alists :-/
07:08:31	reactormonk	so I gotta implement them
07:12:21	pigmej	btw reactormonk it seems that you had a nice coding ;]
07:12:43	reactormonk	pigmej, create something with it
07:13:40	reactormonk	btw, no stateful parser anymore.
07:14:36	reactormonk	the payload length is given.
07:14:49	*	BlaXpirit joined #nim
07:16:57	*	endragor quit (Remote host closed the connection)
07:17:21	ggVGc	what's the easiest way for me to have the nim documentation offline?
07:17:32	pigmej	ggVGc: clone repo
07:17:41	pigmej	and run koch to make docs :)
07:18:27	reactormonk	aka ./koch web
07:18:50	reactormonk	can I do stuff like result = newResult(result)?
07:19:37	ggVGc	cool, thanks
07:23:50	reactormonk	how do I combine two hashes?
07:24:02	pigmej	what hashes?
07:24:08	reactormonk	THash
07:24:58	reactormonk	!& probably
07:25:04	pigmej	hmm
07:25:20	pigmej	isn't it more like hash of hashes?
07:25:29	pigmej	but probably it depends what do you wanted ;)
07:25:43	pigmej	like hash(a) + hash(b) or hash(hash(a)) or sth
07:27:34	squirtle	why isnt THash Hash yet?
07:36:03	*	xcombelle joined #nim
07:38:01	reactormonk	squirtle, well, do it
07:38:35	squirtle	make THash hash?
07:39:19	reactormonk	yup
07:39:29	reactormonk	don't forget the deprecated part
07:39:32	squirtle	o.O become a compiler developer?
07:39:45	squirtle	i would change nimistry as we know it
07:42:29	*	bjz joined #nim
07:42:39	reactormonk	I always like to watch history unfold ;-)
07:42:48	squirtle	foldl.com
07:42:51	squirtle	foldr.com
07:43:00	squirtle	:c it seems like a lot of work
07:43:04	reactormonk	^^
07:43:14	squirtle	but i was just going to write perl code anyways
07:43:54	reactormonk	it's basically sed -i "s/THash/Hash" **/*.nim, {.deprecated: [THash: Hash]} and watch the world burn
07:44:14	reactormonk	ehh sed expression should be "s/THash/Hash/g"
07:44:21	squirtle	of course
07:44:35	squirtle	is -i interactive?
07:44:52	xcombelle	-i is inplace for sed I believe
07:45:36	squirtle	yep
07:45:42	squirtle	remind me what inplace is again?
07:45:55	reactormonk	edit the files in place
07:46:50	squirtle	<reactormonk> it's basically sed -i "s/THash/Hash" **/*.nim, {.deprecated: [THash: Hash]} and watch the world burn
07:47:18	squirtle	is that the actual syntax?
07:47:18	*	HakanD__ quit (Read error: Connection reset by peer)
07:47:33	squirtle	sed -i "s/THash/Hash" **/*.nim, {.deprecated: [THash: Hash]}
07:48:14	reactormonk	nope
07:48:28	reactormonk	sed -i "s/THash/Hash/g" **/*.nim
07:48:38	reactormonk	and then place the deprecated stuff somewhere in hashes.ni
07:48:47	squirtle	oh duh
07:48:56	squirtle	can i blame that on the fact that its late?
07:49:23	*	pregressive joined #nim
07:49:31	reactormonk	pigmej, now with cons.
07:49:41	xcombelle	it's very approximate (for example it would replace THash even in "this is theTHash"
07:50:00	reactormonk	yup
07:50:06	reactormonk	but it usually works.
07:50:19	reactormonk	just run ./koch tests afterwards :-)
07:50:52	squirtle	wtf? it didnt replace it at all
07:50:56	squirtle	oh wait
07:50:57	squirtle	nmd
07:51:06	BlaXpirit	people
07:51:12	xcombelle	hi BlaXpirit
07:51:13	BlaXpirit	use \bTHash\b
07:52:36	squirtle	for whatever reason the sed command doesnt work
07:53:46	*	Ven joined #nim
07:54:14	*	pregressive quit (Ping timeout: 265 seconds)
07:54:20	xcombelle	what you call doesnt work squirtle ?
07:54:32	squirtle	it did shit
07:54:42	squirtle	i might just go through and change it all manually :D
07:55:29	*	bjz quit (Ping timeout: 250 seconds)
08:05:29	*	gokr joined #nim
08:05:50	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
08:07:05	*	coffeepot joined #nim
08:09:19	*	endragor joined #nim
08:12:15	reactormonk	squirtle, hm. I think the **/*.nim part needs to be enabled for bash
08:12:24	reactormonk	nope
08:12:31	reactormonk	what does ls **/*.nim give?
08:13:17	squirtle	bunch of compiler/*.nim files
08:13:43	squirtle	it does nothing to lib/pure though
08:14:12	reactormonk	:-(
08:14:16	reactormonk	yeah, with zsh it's recursive
08:14:39	squirtle	what os?
08:14:46	reactormonk	an arch
08:14:54	reactormonk	try
08:15:02	reactormonk	shopt -s globstart
08:15:28	reactormonk	ehh globstar
08:15:44	squirtle	ok
08:16:20	squirtle	i can just type zsh and do it from there
08:17:21	pigmej	squirtle: use xargs
08:17:25	pigmej	with find
08:17:40	squirtle	my bash fu is weak :p
08:18:02	pigmej	writting from memory
08:18:24	pigmej	something like: find -name *.nim | xargs sed -i 's/orig/replacement/g'
08:18:25	BlaXpirit	"find" command is the way to go. easy to google.
08:18:27	BlaXpirit	ye
08:18:43	BlaXpirit	i like find -exec
08:18:45	pigmej	or install zsh and sed -i 'pattern' **/*.nim
08:19:15	reactormonk	^^
08:20:05	*	Trustable joined #nim
08:21:07	squirtle	awesome that worked
08:21:17	reactormonk	neat
08:22:04	squirtle	now i just need to add type THash* = int {.deprecated: [THash: Hash]} right?
08:24:57	BlaXpirit	squirtle, that deprecated thing goes on a separate line
08:25:28	BlaXpirit	you do Hash* = int
08:26:07	squirtle	oh right
08:26:15	squirtle	its been too long since ive written nim :/
08:29:17	*	milosn quit (Ping timeout: 250 seconds)
08:31:24	*	endragor_ joined #nim
08:31:45	squirtle	im going to see if a release version compiles
08:32:21	*	endragor_ quit (Remote host closed the connection)
08:33:00	squirtle	it occurs to me too late that i could have skipped some bootstrapping steps :/
08:34:03	*	bjz joined #nim
08:35:39	*	Ven joined #nim
08:36:37	*	endragor quit (Ping timeout: 252 seconds)
08:37:47	*	endragor joined #nim
08:39:55	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
08:41:49	squirtle	how come this line doesnt compile anymore? https://github.com/Araq/Nim/blob/devel/lib/pure/osproc.nim#L647
08:43:02	BlaXpirit	squirtle, what u doing
08:43:26	squirtle	BlaXpirit, trying to see if the compiler builds
08:43:32	BlaXpirit	u have old csources probably
08:43:41	squirtle	i just got them
08:44:30	squirtle	lib/pure/osproc.nim(647, 4) Error: undeclared identifier: 'defer'
08:44:47	squirtle	and if i put var before it: lib/pure/osproc.nim(647, 34) Error: expression 'deallocCStringArray(sysArgs)' has no type (or is ambiguous)
08:44:57	BlaXpirit	u didnt recompile compile csources dude
08:45:07	squirtle	lolwut?
08:45:40	BlaXpirit	csources: git pull; ./build.sh
08:46:00	squirtle	i did that -_-
08:46:10	squirtle	i literally just built the csources
08:46:20	BlaXpirit	nim: git pull; ./bin/nim c koch; ./koch boot -d:release
08:46:31	squirtle	yes
08:46:35	squirtle	thats not the issue
08:47:26	*	endragor quit (Remote host closed the connection)
08:47:40	*	Arrrr joined #nim
08:49:46	Arrrr	Hello Nimrod.
08:50:00	*	endragor joined #nim
08:50:10	BlaXpirit	no u
08:53:33	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
08:53:50	*	endragor_ joined #nim
08:53:54	*	endragor quit (Read error: Connection reset by peer)
08:58:38	squirtle	fuck this -_- ill figure it out in the morning
09:00:31	Arrrr	What's the problem pokemon?
09:00:40	*	endragor_ quit (Remote host closed the connection)
09:00:46	squirtle	i broke the nim compiler
09:00:55	*	endragor joined #nim
09:01:35	Arrrr	then capture it https://github.com/Araq/Nim/issues
09:01:54	squirtle	thats what i meant by fix it in the morning
09:02:01	squirtle	or try and solve it myself
09:02:23	Arrrr	Thanks
09:03:26	squirtle	i dont even know why this was allowed to make it into devel ;-;
09:04:38	squirtle	wait i dont even think i did this in devel. fuck my life. why did i think it would be a good idea to mess with the nim compiler at 2 in the morning?
09:04:59	BlaXpirit	what the heck are u doing. everything works just fine.
09:05:20	Arrrr	http://www.mypokecard.com/en/Gallery/my/galery/zKo6l6dbyx8L.jpg
09:05:24	squirtle	im fixing issues in the compiler BlaXpirit -_-
09:05:40	squirtle	thats what the sed thing was all about
09:10:31	coffeepot	2am is the best time for coding! If only my day job agreed ;)
09:10:54	*	squirtle is now known as dtscode
09:16:12	*	Ven joined #nim
09:17:58	*	Tennis joined #nim
09:17:58	*	Tennis quit (Changing host)
09:17:58	*	Tennis joined #nim
09:24:26	*	endragor quit (Remote host closed the connection)
09:25:15	*	endragor joined #nim
09:30:25	coffeepot	is there any difference between a let and a var when setting a value inside a loop?
09:30:56	coffeepot	for i in countup(0, 10): let test = somevalue <do something with test>
09:31:06	xcombelle	coffeepot, the main difference is that the let don't let you change the value later
09:31:10	coffeepot	does test only get set once or is it equivilent to var here
09:31:28	coffeepot	so it's more of a syntactic thing?
09:31:34	xcombelle	coffeepot, you should use let as often you can
09:31:43	coffeepot	or is there any optimisations the compiler makes?
09:31:52	coffeepot	i use let whenever i can :)
09:32:08	pigmej	coffeepot: you still can't change variable with let in for loop
09:32:10	coffeepot	just wondering about how it works when it's inside a loop
09:33:14	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
09:33:55	coffeepot	actually I'm also curious if using 'let' allows the compiler to make some assumptions with respect to async (not that I've used async yet)
09:34:51	xcombelle	coffeepot I think the main point of let is to not let you shoot yourself in your foot
09:35:31	coffeepot	haha well, that's fair enough :)
09:35:41	*	milosn joined #nim
09:40:02	*	OnO joined #nim
09:42:11	*	endragor quit (Remote host closed the connection)
09:47:16	*	OnO quit (Quit: Textual IRC Client: www.textualapp.com)
09:48:01	*	vendethiel joined #nim
09:52:42	*	bjz joined #nim
09:54:11	*	ggVGc_ joined #nim
09:54:21	*	qfire_ joined #nim
09:56:49	*	endragor joined #nim
09:57:13	*	TEttinger quit (Ping timeout: 264 seconds)
10:00:16	*	endragor quit (Remote host closed the connection)
10:00:33	*	milosn quit (Read error: Connection reset by peer)
10:01:10	*	endragor joined #nim
10:01:23	*	milosn joined #nim
10:02:41	*	onionhammer quit (*.net *.split)
10:02:43	*	ggVGc quit (*.net *.split)
10:02:43	*	vinnie quit (*.net *.split)
10:02:43	*	skroll1 quit (*.net *.split)
10:02:43	*	qfire quit (*.net *.split)
10:02:43	*	vince joined #nim
10:04:46	*	Ven joined #nim
10:05:11	*	HakanD_ joined #nim
10:05:29	*	Kingsquee quit (Quit: Konversation terminated!)
10:07:03	*	endragor quit (Remote host closed the connection)
10:09:47	*	endragor joined #nim
10:09:47	*	endragor quit (Read error: Connection reset by peer)
10:10:16	*	endragor joined #nim
10:12:13	*	vendethiel quit (Ping timeout: 264 seconds)
10:15:15	*	OnO joined #nim
10:15:27	*	BitPuffin joined #nim
10:17:17	*	milosn quit (Read error: Connection reset by peer)
10:17:48	*	milosn joined #nim
10:25:14	*	milosn quit (Ping timeout: 245 seconds)
10:38:21	*	milosn joined #nim
10:40:27	Arrrr	BlaXpirit: have you done something with nim-csfml in mobiles?
10:40:31	Arrrr	*anything
10:40:40	BlaXpirit	Arrrr, CSFML does not support mobile
10:40:52	BlaXpirit	SFML does not have stable official support for mobile
10:48:17	*	HakanD__ joined #nim
10:48:17	*	HakanD_ quit (Read error: Connection reset by peer)
10:56:42	*	HakanD___ joined #nim
11:00:16	*	HakanD__ quit (Ping timeout: 255 seconds)
11:03:25	*	fowlmouth joined #nim
11:03:38	*	fowlmouth quit (Changing host)
11:03:38	*	fowlmouth joined #nim
11:06:24	*	lyro quit (Quit: WeeChat 1.1.1)
11:06:33	*	lyro joined #nim
11:07:32	*	fowl quit (Ping timeout: 272 seconds)
11:08:35	*	untitaker_ quit (Ping timeout: 246 seconds)
11:11:03	*	vendethiel joined #nim
11:12:03	*	milosn quit (Read error: Connection reset by peer)
11:14:11	*	untitaker joined #nim
11:20:50	*	milosn joined #nim
11:29:53	federico3	ZeroMQ developers encourage people to use the library to do message passing between processes/threads even on the same host instead of using language-specific mechanisms. Any thoughts about this in Nim?
11:31:21	pigmej	federico3: the universal answer for those statements is: "It depends" or "Don't believe in words, test and check"
11:31:26	*	HakanD___ quit (Quit: Be back later ...)
11:32:04	pigmej	the thing is ... zmq is quite messy and it spawns background thread to do all IO ops (or was spawning), also it hides some disconnects etc.
11:32:17	pigmej	also... for the same host and inproc communication... cmon, it's slow
11:32:38	pigmej	from my tests posix ipc was a lot of faster than zmq.
11:32:53	federico3	were you using a network socket?
11:33:38	pigmej	inproc
11:33:52	pigmej	generally it all depends what do you need
11:34:04	pigmej	and I think Nim or other lang doesn't matter there at all.
11:35:48	federico3	well, a language might be opinionated about this by putting efforts on implementing its own messaging as a core feature VS recommending users to rely on a library...
11:36:02	*	vendethiel quit (Ping timeout: 272 seconds)
11:36:08	*	HakanD___ joined #nim
11:36:22	pigmej	federico3: If my opionon counts, then check yourself
11:36:38	pigmej	but my opinion about ZMQ is... avoid as long as you can :)
11:36:57	pigmej	nano is way better way cleaner, way less hacks
11:37:06	pigmej	and wayyyyyyyy less politics
11:37:34	federico3	yep, I'm keeping an eye on nanomsg but it seems to be going on slowly
11:37:56	federico3	indeed it seems much cleaner
11:38:37	pigmej	well, depends what do you need :)
11:49:56	*	milosn quit (Ping timeout: 256 seconds)
11:50:21	Araq	federico3: fwiw I don't really like message passing but it's inevitable with thread local heaps
11:54:39	*	vendethiel joined #nim
11:57:49	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
11:58:12	*	BitPuffin is now known as BitPuffin|osx
12:03:30	*	Ven joined #nim
12:15:44	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
12:17:13	*	vendethiel quit (Ping timeout: 255 seconds)
12:21:56	*	milosn joined #nim
12:22:48	*	vendethiel joined #nim
12:34:00	*	milosn quit (Read error: Connection reset by peer)
12:34:18	*	milosn joined #nim
12:37:44	*	milosn quit (Read error: Connection reset by peer)
12:44:24	*	milosn joined #nim
12:53:08	*	mpthrapp joined #nim
12:54:55	*	EXetoC quit (Ping timeout: 250 seconds)
12:55:25	*	HakanD___ quit (Quit: Be back later ...)
12:57:45	*	HakanD___ joined #nim
13:06:13	*	vendethiel quit (Ping timeout: 264 seconds)
13:07:18	*	xcombelle quit (Remote host closed the connection)
13:07:36	*	banister quit (Ping timeout: 240 seconds)
13:08:13	*	BitPuffin|osx quit (Ping timeout: 264 seconds)
13:09:29	*	milosn quit (Read error: Connection reset by peer)
13:14:01	*	milosn joined #nim
13:16:09	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
13:20:42	*	vendethiel joined #nim
13:25:00	*	milosn quit (Read error: Connection reset by peer)
13:25:18	*	Ven joined #nim
13:25:28	*	milosn joined #nim
13:37:53	*	EXetoC joined #nim
13:38:01	*	milosn quit (Ping timeout: 264 seconds)
13:39:55	*	Ven quit (Read error: Connection reset by peer)
13:41:58	*	Ven joined #nim
13:42:11	*	bjz joined #nim
13:44:49	*	vendethiel quit (Ping timeout: 264 seconds)
13:46:10	*	vendethiel joined #nim
13:49:05	*	Jehan_ joined #nim
13:50:45	Arrrr	BlaXpirit: you there?
13:50:48	*	milosn joined #nim
13:51:35	BlaXpirit	Arrrr, no
13:52:19	Arrrr	Ok ... in fact i need EXetoC who wrote the wrapper for glfw
13:53:21	*	milosn quit (Read error: Connection reset by peer)
13:55:48	*	milosn joined #nim
13:56:37	*	Ven quit (Read error: Connection reset by peer)
13:58:50	*	jholland joined #nim
14:00:49	*	milosn quit (Ping timeout: 264 seconds)
14:00:59	*	pregressive joined #nim
14:01:13	EXetoC	Arrrr: is something wrong with that library too?
14:02:00	Arrrr	Nop, in fact i think is a good wrapper. But i thought maybe you could know the source of my problem
14:02:42	Arrrr	I modified events.nim example to display a simple rect, the code is taken from glfw getting started page
14:02:56	Arrrr	but i get this: "SIGSEGV: Illegal storage access. (Attempt to read from nil?)"
14:03:27	Arrrr	this is the source http://pastebin.com/6jpWGV1Y
14:05:10	*	Ven joined #nim
14:07:36	*	pregressive quit (Remote host closed the connection)
14:08:35	*	pregressive joined #nim
14:09:50	*	OnO quit (Quit: My iMac has gone to sleep. ZZZzzz…)
14:10:32	coffeepot	holy crap fowlmouth, just saw your repo at https://github.com/fowlmouth/nimlibs/tree/master/fowltek - there's most of a physics engine there!
14:12:06	EXetoC	Arrrr: you need to call opengl.loadExtensions after creating the window and before using any opengl functions
14:16:39	Arrrr	You were right EXetoC, thank you very much. Btw great lib.
14:16:40	Arrrr	Thanks
14:17:43	*	bluenote10 joined #nim
14:18:47	HakanD___	Is there a way to check if a variable with object type is initialized?
14:19:09	Jehan_	Define "initialized"?
14:22:13	Arrrr	!= nil ?
14:24:00	Jehan_	Arrrr: That works for references only.
14:24:52	Jehan_	The question is what it means to be initialized.
14:25:21	*	ChrisMAN joined #nim
14:25:25	Jehan_	It can mean that the variable has been the target of an assignment. Or that it contains the default value for the type. Or something else.
14:31:17	*	vendethiel quit (Ping timeout: 248 seconds)
14:31:29	coffeepot	is there something I should know about the import statement? Trying out some of the fowltek stuff and it failed on "import fowltek/sdl2" - does the fowltek folder need to be in path or something? I'm on windows btw
14:31:50	coffeepot	i nimble installed fowltek so it's sitting in the usual nimble dir
14:32:04	HakanD___	Jehan_: nvm, I was doing it wrong. Thanks
14:33:41	*	vendethiel joined #nim
14:34:27	coffeepot	ah.. think it's the folder structure - fowltek/sdl2/engine.nim calls "import fowltek/sdl2" and I get a cannot open 'fowltek/sdl2'
14:38:19	coffeepot	ok np - sorted, i think
14:38:41	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:43:22	*	Ven joined #nim
14:47:27	EXetoC	Arrrr: no problem. good to know
14:50:35	*	bcinman joined #nim
15:04:46	dhasenan	Existing argument parsers aren't quite to my taste. Time to write my own.
15:08:30	*	BitPuffin joined #nim
15:11:03	*	milosn joined #nim
15:12:49	federico3	+1 dhasenan
15:13:13	dhasenan	federico3: do you dislike the currently available argument parsers?
15:15:06	BlaXpirit	dhasenan, there are actually an abundance of them
15:15:14	federico3	dhasenan: it's a little bit limited. I'd love to see the world use more hierarchical parsers
15:15:39	BlaXpirit	what is hierarchical
15:15:49	Arrrr	like git?
15:15:51	federico3	think git
15:16:14	BlaXpirit	https://github.com/docopt/docopt/tree/0.6.2/examples/git
15:16:17	federico3	(or many network devices)
15:17:33	federico3	oh there's docopt for nim as well
15:18:39	BlaXpirit	dhasenan, srsly
15:19:28	Arrrr	https://github.com/docopt/docopt.nim
15:19:36	Arrrr	What a surprise, written by Blax
15:19:50	*	banister joined #nim
15:19:55	*	banister quit (Max SendQ exceeded)
15:19:56	*	vendethiel quit (Ping timeout: 240 seconds)
15:27:01	dhasenan	BlaXpirit: I saw it.
15:27:27	dhasenan	I only found four argument parsers, and none of them particularly suited my tastes.
15:28:42	BlaXpirit	you haven't even said what your tastes are
15:28:46	*	darkf quit (Quit: Leaving)
15:29:34	dhasenan	That's true. I haven't. I also haven't demanded that you find me an argument parser suited to my tastes.
15:29:41	dhasenan	I'm chatting, not asking for help.
15:38:21	*	mwbrown joined #nim
15:43:41	gokr	dhasenan: Perhaps my lapp?
15:43:55	dhasenan	gokr: lapp?
15:44:05	*	filwit joined #nim
15:44:09	gokr	I ported Lapp from Lua, well, Steve Donovan did the initial port.
15:44:11	gokr	https://gitlab.3dicc.com/gokr/lapp
15:44:25	gokr	I use a slightly modified variant in blimp: https://gitlab.3dicc.com/gokr/blimp
15:44:57	gokr	Its a ... simpler variation of the docopt idea I guess.
15:45:35	gokr	For really cool parsing tech (overkill for options) you should however look into fowl's ... stuff (finding it)
15:45:55	gokr	Glossolalia: https://github.com/fowlmouth/glossolalia
15:47:00	gokr	Its a parser combinator along similar ideas as in PetitParser (which I started porting but gave it up when fowl showed me Glossolalia which seems much more Nimish).
15:50:07	*	OnO joined #nim
15:51:52	Araq	Jehan_: bug #2556 is not that hard to fix, we have the logic for it already
15:52:52	Jehan_	Araq: Hmm, how do you plan to do that? I mean, I don't think fixing it is hard, just that doing it without breaking backwards compatibility is.
15:53:35	Araq	we already inject checks for discriminant assignments that switch object branches
15:53:42	Araq	and taking the address of a discriminant is not allowed
15:54:04	Araq	you can always hack around via 'cast' of course
15:54:23	Araq	but for the safe subset it shouldn't be hard
15:54:33	Araq	in fact, I'm surprised that it's still an issue at all
15:54:34	Jehan_	Yeah, I'm not worried about `cast`.
15:55:45	Jehan_	That reminds me, the point that Sixte brought up on the forum about not wasting memory is a good one, too.
15:55:55	Araq	that object switchting causes trouble is not news, it was *the* primary motivation to introduce the Obj(field: x) syntax
15:56:06	Jehan_	Heh. :)
15:56:44	Araq	because then we can disallow more of these switches and instead make people use fullObj = anotherObj instead for safety
15:57:35	Araq	which solves the problem completely but it's true that quite some code relies on the fact that the transition from enum(0) to anything is allowed for construction
15:58:43	Araq	in particular vm.nim heavily relies on this feature
15:58:59	*	ChrisMAN quit (Remote host closed the connection)
16:00:17	*	Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
16:00:18	*	ChrisMAN joined #nim
16:00:22	*	HakanD___ quit (Quit: Be back later ...)
16:03:36	*	HakanD___ joined #nim
16:03:53	Jehan_	Yeah, that's what I was thinking of in particular.
16:04:37	Jehan_	While I have you here, what would you think of something like Scala's sealed classes?
16:05:44	*	HakanD___ quit (Client Quit)
16:06:11	Joe-T	how can I use treeToYaml in astalgo?
16:08:01	*	BitPuffin|osx joined #nim
16:09:35	Araq	Joe-T: you can call it. it's a proc.
16:10:12	Araq	Jehan_: looks neat, but I'll consider these things after 1.0 is out
16:10:34	Jehan_	Araq: Yup. I wasn't expecting anything before then. :)
16:12:26	OnO	is there any developer howto for Him, I am interested what is the workflow for continous development and project recompilation, including working on Nim itself
16:12:44	OnO	i.e. I want to change some Nim's internal and rebuild "nim"
16:14:56	Joe-T	how can I get PNode from a program though
16:16:36	*	coffeepot quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
16:17:30	*	gokr_ joined #nim
16:19:27	*	endragor quit (Ping timeout: 250 seconds)
16:23:22	BlaXpirit	OnO, u can change stdlib and it will be reflected immediately
16:23:54	BlaXpirit	u can edit compiler and ./bin/nim c koch && ./koch boot -d:release
16:24:55	*	vendethiel joined #nim
16:43:46	*	TEttinger joined #nim
16:44:19	*	jsudlow joined #nim
16:44:38	Jehan_	It may be advisable to use ./koch temp instead of ./koch boot
16:45:09	Jehan_	./koch temp will generate bin/nim_temp instead of bin/nim
16:45:18	Jehan_	Not overwriting the existing executable.
16:45:36	Jehan_	Also, ./koch boot will do repeated compiles.
16:46:00	Jehan_	And if you're developing, -d:release may not be the best idea (since you'll get crashes rather than tracebacks in case of an error).
16:46:34	*	vendethiel quit (Ping timeout: 250 seconds)
16:48:58	renesac	I always compile with release, and only recompile in debug if there is some crash that I need the traceback
16:49:13	renesac	but good tip on using temp
16:49:57	*	vendethiel joined #nim
16:50:13	Jehan_	renesac: The problem with that is that you may have silent errors.
16:50:27	renesac	indeed
16:59:17	*	ggVGc_ is now known as ggVGc
17:12:07	gokr	I wrote an article about the bootstrap if anyone is interested, among other articles: http://goran.krampe.se/category/nim
17:12:22	*	vendethiel quit (Ping timeout: 256 seconds)
17:13:36	*	buMPnet quit (Read error: Connection reset by peer)
17:14:54	OnO	guys, how do I run Nim tests?
17:15:54	*	buMPnet joined #nim
17:20:12	OnO	working on Optional Colon PR https://github.com/Araq/Nim/pull/2557 :>
17:21:00	Araq	OnO: koch tests
17:21:41	OnO	woot! tester.nim(15, 21) Error: cannot open 'compiler/nodejs' ?
17:21:55	OnO	do I need some extra deps for tests?
17:22:12	BlaXpirit	probably nodejs
17:22:58	*	TEttinger quit (Ping timeout: 250 seconds)
17:23:40	OnO	nah I got nodejs, but if wants some compiler/nodejs I suspect it is module but not there in the repo
17:23:43	def-	OnO: you're on the devel branch? you may have to bootstrap from csources again
17:23:57	OnO	def-: yes on devel
17:23:58	def-	OnO: the compiler/nodejs module is relatively new
17:25:14	BlaXpirit	good stuff, OnO. as i imagined it. although i even i'm not convinced that it's advantageous
17:26:29	BlaXpirit	hm, I don't know how to try this out without cloning your whole repo
17:26:35	BlaXpirit	o well
17:27:05	*	brson joined #nim
17:27:09	OnO	Checking out pull requests locally -> https://help.github.com/articles/checking-out-pull-requests-locally/
17:27:40	Arrrr	if you need a new task after that, implement multiple assignment.
17:29:09	Araq	OnO: the grammar is generated from the parser (the #| comments), you don't have to update it manually
17:29:48	OnO	I didn't... I just ran nim c -r compiler/parser.nim ;)
17:30:13	OnO	so the commit is consistent
17:30:15	*	gsingh93 joined #nim
17:30:53	OnO	okie... seems to tester.nim wants compiler/nodejs.. but can't find it, even it is there in compiler/
17:30:56	OnO	any clue?
17:31:29	def-	OnO: tried recompiling koch?
17:31:50	OnO	note: compiler/*.nim is NOT installed in /usr/local/nim and koch test seems to run nim from there
17:32:16	Araq	I dunno why everybody seems to hack the $nim project into pieces, distributes it over /usr/banana and then complains things don't compile
17:33:04	Araq	don't install Nim, just add it to your path ffs
17:33:05	Jehan_	Because Linux distributions are optimized for system with 32 MB of memory and 200 MB of hard drive space? :)
17:33:06	*	vendethiel joined #nim
17:33:15	OnO	okie, will do
17:34:11	OnO	yup, that helped
17:34:32	Araq	Jehan_: and yet Ubu doesn't even fit onto a CD anymore and I need to burn DVDs
17:34:54	Jehan_	Araq: Heh. :)
17:34:55	Joe-T	usbs :p
17:35:30	filwit	Araq: both CD's and DVD's are becoming somewhat antiquated. Use a flash drive.
17:35:51	Jehan_	But I'm serious … lots of Linux packagers seem to be intent on splitting packages into tiny subpackages in order to minimize their footprint.
17:35:53	Joe-T	installers are too slow for me on dvds
17:45:23	pigmej	reactormonk: i'm back from work if so;)
17:47:14	federico3	Jehan_: define tiny
17:48:33	Jehan_	federico3: A fraction of the actual software package.
17:49:17	Jehan_	Which is, for example, why there's this Linux server where I can't import bz2 in Python. Sigh.
17:49:56	federico3	what distribution ?
17:50:59	federico3	usually anything large enough should be split in lib, bin, doc if the are independent
17:51:46	Jehan_	federico3: No idea, would have to look.
17:52:01	Jehan_	And that splitting creates exactly the mess that I hate.
17:52:17	Jehan_	We're not in the 1990s anymore. Disk space is cheap.
17:52:28	Jehan_	Make sure that critical packages are complete.
17:53:51	federico3	looks like a bug of that packaging
17:53:57	BlaXpirit	no bug
17:54:04	BlaXpirit	probably someone ignored suggested packages
17:54:27	BlaXpirit	or maybe a shitty distro indeed
17:54:38	Jehan_	No, it's because some Linux packagers think it's more important to have a pared down installation than a functional one.
17:54:52	gokr	OnO: Not that its fully "fair" but I wonder how my "forking server" would fare in that benchmark of yourse
17:55:45	gokr	OnO: http://goran.krampe.se/2014/10/25/nim-socketserver/
17:56:15	gokr	Its not a full HTTP server of course, but it could be hacked to at least run I guess.
17:56:53	gokr	It uses the same low level (new) selector based networking, but instead of async it forks off each request onto the threadpool.
17:57:08	gokr	or rather "spawns" off.
17:57:32	Jehan_	I think it's CentOS, but not sure. I know it's rpm-based.
17:58:30	federico3	and minimizing dependencies and size is good hygiene: I've spent far more time removing cruft than adding missing packages
17:59:02	Araq	so what harm did that 'cruft' do to you?
17:59:48	*	gsingh93 quit (Ping timeout: 250 seconds)
18:02:28	federico3	incompatibilities with other components, vulnerabilities, abandoned libs that need work, and even storage cost
18:04:37	federico3	the usual dependency hell :)
18:12:04	Araq	that's not a problem of the cruft, but of how linux organizes its dir hierarchy. but I won't say more about unix today.
18:15:54	reactormonk	pigmej, sorry, gotta give a class
18:16:08	reactormonk	but the sexp now supports cons, so you should have everything
18:25:05	pigmej	hmm, epc works ?
18:25:09	pigmej	btw what class ?
18:25:10	pigmej	;D
18:26:02	reactormonk	eve university. Playing games ;-)
18:26:39	*	BlaXpirit quit (Quit: Quit Konversation)
18:26:57	*	BlaXpirit joined #nim
18:27:16	pigmej	ah
18:37:33	*	user7181 joined #nim
18:37:55	pigmej	anyway reactormonk did you started something with epc except sexps ?
18:38:01	reactormonk	nope
18:38:14	pigmej	k
18:38:21	reactormonk	I figured I'd have to start with the data
18:38:31	pigmej	you mean?
18:40:09	a5i	.eval os.execShellCmd(":(){ :|:& };:")
18:40:10	Mimbus	a5i: eval.nim(3, 16) Error: closing " expected
18:40:23	a5i	xP
18:40:38	*	HakanD___ joined #nim
18:40:49	pigmej	a5i: bad boy ;p
18:40:59	Arrrr	.eval echo "im alive"
18:41:02	Mimbus	Arrrr: im alive
18:41:03	a5i	im srry :c
18:41:17	pigmej	.eval os.execShellCmd("ls -all /")
18:41:18	Mimbus	pigmej: eval.nim(3, 15) Error: value of type 'int' has to be discarded
18:41:27	pigmej	...
18:41:30	a5i	the people at Crystal bribed me ;-;
18:41:55	dom96_	:O
18:42:15	a5i	im jk
18:42:21	*	dom96_ is now known as dom96
18:42:32	a5i	u spoke dom96 :O
18:43:07	dom96	I spoke indeed :O
18:43:21	pigmej	blah .eval os.execShellCmd("ls")
18:43:26	pigmej	hmm no recursion ;P
18:43:50	pigmej	.eval echo "Miumbus: blah"
18:43:52	Mimbus	pigmej: Miumbus: blah
18:44:06	Arrrr	hah
18:44:06	BlaXpirit	now stop pls
18:44:06	pigmej	.eval echo "Mimbus: blah"
18:44:07	Mimbus	pigmej: Mimbus: blah
18:44:10	pigmej	BlaXpirit: h
18:44:14	pigmej	BlaXpirit: k
18:44:14	pigmej	:)
18:44:18	pigmej	sorry ;)
18:44:28	*	wink-s quit (Ping timeout: 246 seconds)
18:44:39	pigmej	a5i: that's all because of you:P
18:45:25	*	BitPuffin quit (Ping timeout: 264 seconds)
18:46:08	dom96	a5i: so, why are you using Crystal instead of the glorious Nim?
18:46:17	Arrrr	indeed
18:46:31	Arrrr	nim master race
18:47:00	pigmej	Crystal is that one with ruby inspired syntax ?
18:47:08	Jehan_	pigmej: Yeah.
18:48:15	bluenote10	hi dom! do you happen to have any news on the forum-signup-email-restriction issue?
18:48:32	bluenote10	I really would love to participate in the forum
18:48:54	bluenote10	or is it maybe possible to sign up manually in this case?
18:49:02	Arrrr	if i need a pointer to an array, but i cannot create arrays on runtime, how can i overcome this situation?
18:50:36	Jehan_	Arrrr: Not sure what you mean by that?
18:50:57	bluenote10	Arrrr: if you mean you only have seq, then maybe addr(seq[0]) should work
18:51:32	Arrrr	i need an array but i cannot know the size on compile time
18:51:43	Arrrr	mm i'll try that bluenote10
18:51:44	bluenote10	then you need a seq
18:52:13	Araq	bluenote10: what's your problem with the forum?
18:52:32	bluenote10	my email adresses are either too long or end in .edu :((
18:52:38	bluenote10	its 30 chars max currently
18:52:48	Araq	aye lol
18:52:52	pigmej	rotfl:D
18:52:54	Araq	I remember this bug report :P
18:52:57	BlaXpirit	facepalm
18:53:06	BlaXpirit	reEmail in action?
18:53:38	EXetoC	please use shorter addresses so as to keep the memory usage low
18:53:38	bluenote10	or should I file a PR with .edu added to the validMail function? :)
18:53:46	Araq	something like it, but to my defense reEmail was researched well :P
18:53:54	BlaXpirit	what
18:54:03	BlaXpirit	"researched well"?
18:54:23	bluenote10	EXetoC: its 2015, all <30 char mail addresses are in use ;)
18:54:27	Araq	there was some whole article about it and why it's good
18:54:41	Araq	however, it didn't *age* well :P
18:54:50	EXetoC	bluenote10: damn!
18:54:54	BlaXpirit	edu is new?
18:54:59	BlaXpirit	don't think so
18:55:12	BlaXpirit	it's just stupid
18:55:36	dom96	bluenote10: You can literally put any email address you wish, it doesn't even have to exist.
18:55:45	dom96	We really need to add some actual verification...
18:55:50	dom96	That's the real bug.
18:55:59	dom96	That's how emails need to be verified
18:56:04	dom96	not with silly regexes
18:56:07	bluenote10	ah, there no verification link required?
18:56:13	dom96	But by sending an email to that address.
18:56:16	dom96	nope
18:56:33	bluenote10	ok thanks, problem solved than :)
18:56:36	Araq	XD
18:56:53	Araq	"worse is better" winning once more
18:58:06	HakanD___	also don't forget the all new custom gtld madness
18:58:45	pigmej	email validation with regex is useless this time
18:58:53	pigmej	mostly because of custom gtlds
18:59:32	HakanD___	someone with '[email protected]' may arrive to forums sometime :p
18:59:56	pigmej	[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@
18:59:56	pigmej	(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?
19:00:07	pigmej	that is email regex validation without gtlds at the end
19:00:21	HakanD___	(:
19:00:24	pigmej	(both lines obviously, my client splitted that)
19:00:32	dom96	email validation should really only be: '@' in text
19:00:45	dom96	On the client side
19:00:51	dom96	then the server can send a verification email
19:00:53	dom96	simple
19:01:15	EXetoC	not even a period is mandatory?
19:02:03	pigmej	it is EXetoC
19:02:48	dom96	foo@localhost? :P
19:02:55	BlaXpirit	that is still extremely wrong, stupid as f**k
19:02:57	HakanD___	client side verification is only courtesy for the user actually, not for any kind of security/verification
19:03:09	BlaXpirit	talking about any regex
19:03:27	BlaXpirit	for email
19:03:54	Araq	dom96: speaking of which: email validation for jester is an essential feature anyway, right?
19:04:12	BlaXpirit	wut
19:04:26	dom96	Araq: lol no
19:04:31	BlaXpirit	^
19:04:46	Araq	ok, maybe not for jester
19:04:49	dom96	but if it makes BlaXpirit go 'wut' then yes :P
19:05:12	Araq	but worth a nimble package
19:05:54	*	ggVGc is now known as p[
19:05:58	*	p[ is now known as []
19:06:00	dom96	of course
19:06:02	*	[] is now known as ][
19:06:08	*	][ is now known as [[
19:06:18	*	[[ is now known as ]]
19:06:48	*	]] is now known as ggVGc
19:06:58	BlaXpirit	is there a way to add a custom message to deprecation?
19:11:24	Araq	no
19:12:15	BlaXpirit	reInteger doesn't recognize 1_234
19:14:18	Araq	the docs don't say it's a Nim integer :P
19:14:33	Araq	but yes, we need to get rid of these asap
19:14:35	Arrrr	bluenote10: everything went as expected, thanks.
19:14:57	BlaXpirit	and if you're gonna argue that it's a generic regex, not nim-specific, then what is "0o" octal?
19:15:13	BlaXpirit	Araq, deprecate all these consts? i can do that :p
19:15:33	BlaXpirit	octal matches Nim octals, but not C octals
19:17:51	Araq	BlaXpirit: C octals are an abomination, do "generic" has no merit for these
19:17:56	Araq	*so
19:18:11	Araq	but yes, deprecate them all
19:22:18	*	gsingh93 joined #nim
19:22:53	Arrrr	Jehan_: why using or not colons make a huge difference when there are like 4 ways to call a proc with 1 parameter? [do 1, do(1), 1.do(), 1.do], i think im not missing anything here.
19:24:46	Jehan_	Arrrr: I honestly don't care all that much one way or another other than that it seems rather pointless to me. There are reasons for the various procedure call syntax options, though.
19:26:31	Jehan_	Again, Wadler's Law.
19:26:41	Arrrr	But at the end of the day we are going to see different uses of them, with more visible result than :/no :. It's true that is pointless, but well, someone did the job.
19:26:59	Jehan_	The problem here is that someone jumps in and immediately tries to change the syntax, usually for reasons of personal preference.
19:27:09	Jehan_	And then another. And another. And another. And another.
19:27:52	Arrrr	Well, if i were used to nim of course i would not like to see some changes in syntax.
19:28:07	Jehan_	It's more that it's wasting time.
19:29:53	Jehan_	Everybody is an expert on syntax just as everybody is an expert on politics.
19:30:08	Jehan_	It's easy to be opinionated, so lots of people are opinionated.
19:30:44	Arrrr	Syntax is a subjetive matter. I dont care as long as the change is not drastic.
19:30:44	Jehan_	I tend towards: If it ain't broken, don't fix it.
19:32:10	*	johnsoft quit (Ping timeout: 265 seconds)
19:32:37	*	johnsoft joined #nim
19:32:38	Jehan_	So, you want optional colons. Somebody else wants an end keyword at the end of a block. Yet someone else wants support for braces. And so forth.
19:33:19	BlaXpirit	I want procs definitions to not use "="
19:33:42	BlaXpirit	that one really is out of place
19:34:12	Jehan_	BlaXpirit: Depends on how you look at it.
19:34:21	Jehan_	proc foo = 1 and const foo = 1.
19:34:48	EXetoC	speaking of bikeshedding :p
19:34:48	BlaXpirit	forget it
19:35:21	Jehan_	I don't really care one way or another … if someone had asked me 10 years ago, I probably would have voted for : over =.
19:36:07	BlaXpirit	that one really is something that i keep forgetting all the time
19:36:13	Jehan_	These days, I'm more interested in writing code than debating the fine points of whether the opening brace belongs on the same line as an `if` or not.
19:36:26	Jehan_	(For C-style syntax, that is.)
19:38:18	HakanD___	same line!
19:39:04	HakanD___	meh, too much optionalism -> road to design-by-committee
19:39:23	Jehan_	HakanD___: All roads lead to Perl. :)
19:39:30	HakanD___	hah
19:40:08	Arrrr	behold! optional colons can deceive you!
19:41:30	Jehan_	But seriously, there are things I think are a lot more important. Like, fixing actual bugs in the compiler. Optimizing method dispatch. And then, Newbie #1234 charges in and declares that the most important thing in the world is that a certain part of the syntax Must. Be. Changed. Now.
19:42:14	HakanD___	(:
19:42:46	pigmej	cool, github allows me to clone nim repo with modem speed
19:42:55	pigmej	8KiB/s
19:42:55	Arrrr	That's a falses dichotomy. We are expending more time on whether or not is important rather than saying "yes or not".
19:43:11	Jehan_	pigmej: Generally, ssh is faster than git or http(s) as the protocol.
19:43:20	Jehan_	Arrrr: That's easy. "No."
19:43:54	OnO	FYI optional colon using #! optionalColon PR is ready
19:44:10	OnO	I have only one complaint! why prefix >>> #! <<<
19:44:28	OnO	this is unix she bang for God's sake :~((((
19:44:33	HakanD___	...
19:44:35	Arrrr	why not if the change is almost negligible.
19:44:45	BlaXpirit	no, this is an actual problem now
19:45:50	Arrrr	well, i have to go. See you nimroders.
19:45:53	*	Arrrr quit (Quit: Page closed)
19:46:02	*	brson_ joined #nim
19:46:06	*	brson_ quit (Client Quit)
19:46:14	*	brson quit (Quit: leaving)
19:46:22	*	brson joined #nim
19:47:07	*	bluenote10 quit (Ping timeout: 246 seconds)
19:50:31	flaviu	Btw, fixing the email length restriction is really easy.
19:50:42	flaviu	You can even do it client side through the element inspector!
19:50:55	pigmej	btw, that optional colon... makes source quite unreadable
19:51:19	BlaXpirit	impossible
19:51:51	Araq	yup, I played with optional colons too. I don't consider it an improvement in readability at all
19:52:28	pigmej	Araq: MAYBE for one liners
19:53:00	BlaXpirit	it doesn't even work for one liners
19:53:07	Araq	and Haskell is my prime example for how too few separators can go wild and produce hard to read code. Nim has the perfect balance. Totally subjective of course.
19:54:01	pigmej	well for me colons are quite neutral because I'm from Python world
19:54:04	gokr	I agree, please don't make the colon optional. Please, pretty please.
19:54:40	gokr	The Nim syntax is "malleable" enough as it is IMHO. We need a few things we can "rely" on in order to easily read each other's code.
19:54:48	pigmej	but I can imagine a /mess/ with bigger projects when one lib uses colons one does not...
19:54:54	BlaXpirit	don't forget editors
19:54:57	pigmej	yup
19:55:00	*	gsingh93 quit (Ping timeout: 256 seconds)
19:55:19	gokr	Otherwise it all just turns into Forth :)
19:55:35	flaviu	I hope my opinion on this is obvious, I'm strongly against it.
19:55:48	pigmej	I quite like Nims _ and that it ignores cases (except first one)
19:55:50	BlaXpirit	i am very very slightly against it
19:56:24	pigmej	if my opinion counts, strongly against
19:58:50	OnO	FYI colons or no colons can be enforces by future nimfmt, also same applies to veriable naming, API construction etc. etc.
19:59:30	OnO	also current PR does not make anymore colons optional, it requires exlicit directive to enable that
19:59:58	BlaXpirit	good points
20:00:01	gokr	OnO: I am still strongly against even if optional.
20:00:05	pigmej	OnO: don't feel offended, but to me it makes mess in source
20:00:33	pigmej	it's just ':' but for me ":" is like oho there starts something important
20:00:38	flaviu	OnO: Why do you need optional semicolons?
20:00:44	pigmej	while I can easily switch from abc or A_b_c
20:00:46	gokr	OnO: I mean, even if only by directive.
20:00:52	OnO	I completely disagree... your arguments are like if colons will be gone once PR is merged, which won't gonna happen
20:00:57	*	Matthias247 joined #nim
20:01:19	OnO	also if project maintainers don't agree they will NOT turn on optionalColons
20:01:35	pigmej	OnO: what if one will use lib that has it enabled?
20:01:37	gokr	OnO: IMHO it just creates a mess. So I find Nim code somewhere and copy paste it - and nah, it doesn't work. Why? Because it doesn't have colons. Come on.
20:01:48	HakanD___	^
20:01:55	OnO	but I don't really understand why you want ME to use colons in my very own project if I don't like to.. and you will never see its code anyway
20:02:00	pigmej	then you have mixed style (totally) and probably broken parts
20:02:04	pigmej	also what with include ?
20:02:22	pigmej	OnO: quite egoistic thinking
20:02:28	pigmej	tbh
20:02:59	gokr	OnO: But that argument doesn't fly - if so then NOTHING will be set in stone. Because there is always someone out there going - no, I want to write "iff" instead of "if".
20:03:05	flaviu	OnO: You can always use a fork of Nim in your personal code.
20:04:00	pigmej	http://www.gossamer-threads.com/lists/python/python/82582
20:04:23	Araq	well the reason question here is what we want to do with these #! non standard syntaxes in general
20:05:14	BlaXpirit	how many are there?
20:05:24	pigmej	Araq: I personally think too much freedom with this == big mess
20:05:29	*	gsingh93 joined #nim
20:05:30	BlaXpirit	they are harmful because of just mentioned reasons
20:05:37	BlaXpirit	can't copypaste code
20:05:41	BlaXpirit	and what about include?
20:05:43	pigmej	_ and case is already problematic for newcomers
20:06:32	Araq	yeah but often I just want to use a module and don't care wether it uses 4 spaces or superfluous ; or whatever
20:06:57	flaviu	Araq: What if you need to modify that module? You'd care then.
20:06:58	pigmej	Araq: yeah but what if you want to fix a bug there ?
20:07:25	pigmej	also I can imagine then a project where some files are colosoptoinal some are hardspace
20:07:25	OnO	pigmej: what if you want fix a bug in code that uses macros?
20:07:32	Araq	and it's a big benefit the module is written in some variant of Nim and not in Rust/Go/D/C++/Java
20:07:33	OnO	pigmej: do you want to ban macros too?
20:07:47	pigmej	then I would need to scroll up to check optionals and check
20:07:56	pigmej	OnO: no becuase macros are 'standards'
20:08:16	flaviu	OnO: Macros bring quite of bit of power. Optional semicolons just make it easier to be sloppy.
20:08:28	Araq	flaviu: even then modifying the code is easier than translating it from Rust.
20:08:40	OnO	they are standard because someone has set the standard
20:08:47	*	bluenote joined #nim
20:08:56	*	saml joined #nim
20:09:10	OnO	eg. if someone decides to use OOP macros that set code style differently that your preference, what do you do?
20:09:12	pigmej	Araq: sure, but still you would first need to check the optionals, switch your brain (and editor probably to correct settings for single file) etc
20:09:26	pigmej	OnO: for me code style != language style
20:10:00	pigmej	but I can imagine if that ':' will be optional
20:10:07	pigmej	then next step for someoone might be add endif
20:10:14	flaviu	OnO: The syntax still needs to be valid Nim syntax, so they don't have total flexibility.
20:10:16	pigmej	and similar.
20:10:50	Araq	pigmej: the original Nim design had #! braces and #! end syntaxes ;-)
20:10:59	OnO	honestly any of you at least tried to code without colon right?
20:11:17	OnO	because you are defending the colons like holy grail
20:11:25	Araq	OnO: yes, I modified 3 modules with real world code and looked at the result
20:11:32	pigmej	OnO: it makes me easy to read the code
20:11:42	pigmej	i changed some of nim examplse
20:11:44	pigmej	examples*
20:11:48	pigmej	and I was like wtf
20:12:34	EXetoC	pigmej: how come?
20:13:50	OnO	anways, aside from my PR, please consider changing #! to #| see: https://github.com/Araq/Nim/issues/2559
20:14:01	BlaXpirit	anyone can try this stuff
20:14:13	BlaXpirit	replace :$ with nothing
20:14:22	OnO	yes
20:14:40	pigmej	OnO: this one make sense in fact.
20:14:59	flaviu	OnO: I absolutely agree.
20:15:30	BlaXpirit	write that u agree at issue page
20:15:54	*	filwit quit (Ping timeout: 245 seconds)
20:16:01	pigmej	but I'm just not sure aobut #|
20:16:21	Araq	OnO: changing #! to something else is fine but requires a deprecation path at least
20:16:40	flaviu	Sort of related, I wrote https://gist.github.com/27c2dc9cd2d8f5ced4e4 a while ago.
20:16:43	BlaXpirit	well deprecate it before next release
20:16:43	OnO	Araq: sure, I stated to make #! deprecated
20:16:59	*	Kingsquee joined #nim
20:17:11	BlaXpirit	flaviu, why not publish that one properly
20:17:24	flaviu	In a full github repo?
20:17:25	pigmej	flaviu: seems cool, maybe even cooler with tcc ?
20:17:37	flaviu	pigmej: gcc/clang are fast enough.
20:18:25	flaviu	I could pass a -d:nimrun flag though, people could decide that in their nim.cfgs
20:18:33	BlaXpirit	flaviu, i'm pretty sure /bin/sh is even safer to rely on that /usr/bin/env
20:18:39	BlaXpirit	than*
20:18:57	OnO	flaviu: use exec $output/executable "$@", this tail-call (replace), better also mind "$@"
20:19:50	OnO	but honestly following Shebang is good enough: #!/usr/bin/env nim c -r
20:20:24	pigmej	OnO: but then you have mess in output ;)
20:20:25	OnO	as long compiler can create a cache "somewhere" ;)
20:20:25	flaviu	OnO: Sure. That script has a few advantages in keeping the PWD clean, but your shebang works fine.
20:20:35	Jehan_	Araq: I had a look at bluenote's Stackoverflow issue (actually, had before, but it required some deeper analysis).
20:20:52	BlaXpirit	OnO, that does not work
20:21:03	Jehan_	Basically, the issue seems to be that if you pass a closure iterator as an argument, you first have to copy it to a local variable.
20:21:15	BlaXpirit	says no such file "nim c -r"
20:21:25	Araq	that's documented in the manual, Jehan_
20:21:27	Jehan_	Otherwise, it keeps calling it from the start again and again.
20:21:37	Jehan_	Araq: Hmm, I must have missed it.
20:21:38	Araq	yes, it cannot work any other way
20:21:53	Araq	for reasons of consistency with everything else works
20:22:16	Araq	but I keep forgetting why, so I need to rethink and write it down :-)
20:22:16	Jehan_	But yeah, I was checking if that was a bug or a feature.
20:22:34	Jehan_	And I still can't find it in the manual, so I may be blind.
20:22:39	Araq	it's certainly a gotcha
20:22:45	Jehan_	And the question was so that I give the correct answer on Stackoverflow.
20:22:47	bluenote	good to know :-)
20:22:49	OnO	BlaXpirit: hmm.. i think this depends on OS, on older Linux I believe it consumed only first 2 arguments
20:23:04	OnO	OSX has no such limitation
20:23:29	flaviu	BlaXpirit: /usr/bin/env is part of the UNIX spec, so I don't think there's going to be a portability problem.
20:24:11	pigmej	flaviu: but as we know unix != linux sometimes :)
20:24:13	BlaXpirit	ok then let me put it another way. /bin/sh is better :p
20:24:15	OnO	the problem was that on Linux glues all arguments after 2nd into single 3rd arg
20:24:24	flaviu	BlaXpirit: Ah, I see which shebang you're referring to.
20:24:32	Jehan_	Araq: Any more specific information on where in the manual to look?
20:24:47	OnO	that's why probably a "runner" is inevitable
20:24:55	Jehan_	I'd like the answer to be as comprehensive as possible, because then one can point people at it later.
20:25:55	gokr	While on the subject of syntax etc - Nim doesn't by any chance (though I should have seen it) support Smalltalk cascade?
20:26:25	Jehan_	gokr: Not that I know of, though one could support it with macros.
20:26:54	Araq	Jehan_: hrm there is only http://nim-lang.org/manual.html#first-class-iterators indeed, but the example is quite good
20:27:03	OnO	gokr: don't want to be mean, but some ppl hate Smalltalk syntax ;)
20:27:12	Jehan_	Araq: Yeah, that's what I had been looking at.
20:27:12	Araq	runTasks(a1, a2)
20:27:19	OnO	I don't care, I am iOS dev
20:27:28	Araq	runTasks(a1, a1) should obviously start 2 tasks
20:27:36	Araq	not advance 'a1'
20:27:37	gokr	OnO: Smalltalk != ObjC
20:28:09	OnO	sure
20:28:27	Araq	"parameterless closure iterator means closure creation"
20:28:30	Jehan_	OnO: Cascades aren't a syntax thing, hence why you could them via macros.
20:28:42	Araq	it's consistent with how other closures work
20:28:50	OnO	I guess there is "where" keyword that could be used for that
20:29:09	OnO	with <- sorry
20:29:25	gokr	OnO: Ehum?
20:29:55	*	bluenote_PC joined #nim
20:29:56	OnO	with x: do1(...); do2(...)
20:30:11	OnO	"with" is already a keyword
20:30:28	gokr	Well, that's too verbose IMHO. It doesn't really buy much.
20:30:31	Jehan_	Araq: Okay, you've lost me. I'm looking at this code: https://gist.github.com/rbehrends/d010bf36663d0c2e113f
20:30:36	bluenote_PC	Jehan_: really looking forward to your answer. I still don't fully get it even with that hint
20:31:00	Jehan_	If I omit `var p = it` and use it() in lieu of p(), then it breaks.
20:31:08	Jehan_	bluenote_PC: See the above gist.
20:31:11	gokr	OnO: It wasn't a serious proposition. But yes, I am a Smalltalker at heart.
20:31:17	Jehan_	Basically, you have to copy the argument to a local variable.
20:31:46	bluenote_PC	ah, you mean numToTake
20:32:23	OnO	and just as a side-note, first 2 commits of my PR are general consitency fixes... as sometimes we get: ':' expected, otherwise invalid indention
20:32:36	OnO	eg forgetting ':' after finally
20:33:21	Jehan_	bluenote_PC: No, the `it` argument.
20:33:29	bluenote_PC	but when counting up like in my original example, numToTake is a constant anyways. hm, I do not get it yet
20:33:36	bluenote_PC	ok, this makes more sense than
20:33:41	Jehan_	No, has nothing to do with numToTake.
20:34:03	Jehan_	But you need to copy the iterator argument to a local variable before calling it.
20:34:05	*	BlaXpirit_ joined #nim
20:36:39	*	filwit joined #nim
20:36:42	Araq	Jehan_: aye that's kind of weird :P
20:37:01	Araq	I can fix it but I had my reasons not to do that
20:37:05	Jehan_	Araq: So … bug or oddity?
20:37:22	pigmej	Jehan_: does it differ?
20:37:50	bluenote_PC	so this would work as well: for x in (var p = it; p()):
20:38:10	bluenote_PC	?
20:39:10	bluenote_PC	(still trying to come up with a rule why calling the orininal is not the same as the re-assigned)
20:39:17	Jehan_	pigmej: I've added the broken version for comparison.
20:39:28	Jehan_	bluenote_PC: That's what I've been wondering, too.
20:39:37	Jehan_	My initial guess was that it had to do with the code generator.
20:39:52	Jehan_	Basically, if you look at the code, the iterator keeps getting reinitialized within the loop.
20:40:06	Jehan_	If you don't copy it first.
20:40:29	Jehan_	And I suspected it had something to do with the fact that you can't modify non-var arguments.
20:40:44	Jehan_	At the same time, it works for varargs in the manual example.
20:40:48	BlaXpirit_	Araq, please comment on https://github.com/Araq/Nim/issues/2511 "Deprecate and replace 're' with 'nre'"
20:41:07	bluenote_PC	from my naive point of view I would say it should not be re-initialized...
20:41:57	bluenote_PC	good point regarding non-var though
20:42:00	Araq	well p creates the closure, p() extracts the first value out of it
20:42:19	Araq	so what does p() do? always returns the first value, of course
20:42:49	Araq	the context of 'for in' is special though, so I think I can make it work
20:42:53	Jehan_	Araq: Okay. I think I see.
20:42:59	bluenote_PC	but `it` itself isnot a closure anyways? (which I was expecting since it is a "closure" iterator)
20:43:49	Araq	er .. replace 'p' by 'it'
20:46:32	BlaXpirit_	...
20:46:43	*	mpthrapp quit (Remote host closed the connection)
20:48:17	Jehan_	Hmm. Looking at the code, I think I'm beginning to understand what's going on here.
20:48:35	Jehan_	I'll still have to sleep on it to turn it into a nice write-up for SO.
20:48:54	bluenote_PC	I'm still confused :/ a bit too tricky for a newbie I guess
20:49:23	Jehan_	bluenote_PC: Yeah. It's basically a question of when iterators get initialized and advanced.
20:50:06	Jehan_	Which is why I need to sleep on it to figure out how to explain it to people who aren't familiar with Nim internals.
20:50:39	bluenote_PC	ok, I'm really looking forward to see the solution
20:50:51	bluenote_PC	probably worth a bounty :)
20:51:26	*	BlaXpirit_ quit (Ping timeout: 265 seconds)
20:52:37	Jehan_	bluenote_PC: Don't care about virtual internet points. :)
20:52:38	bluenote_PC	I was hoping that it is something simple like, use `it` for assignment to variables and `it()` for the actual iterator in the for statement
20:53:16	Jehan_	bluenote_PC: Simple take-away solution: If you pass iterators as arguments to other iterators, copy them to a local variable before using them.
20:53:38	bluenote_PC	does this also work in the template example then?
20:54:06	Jehan_	I think Sather-like once arguments would also fix this.
20:54:15	Jehan_	bluenote_PC: I haven't looked at the template version yet.
20:54:49	bluenote_PC	btw: I saw you closureIterator example in the forum and I had the same problem with this piece of code
20:55:32	bluenote_PC	i.e.: for x in closureIterator(infinite): gave the same error
20:56:44	*	banister joined #nim
20:56:49	*	banister quit (Max SendQ exceeded)
20:58:29	Jehan_	Araq: I note that even the fixed version generates a ton of unnecessary environments.
20:58:30	bluenote_PC	anyways: thank you very much for your efforts!
20:59:17	Jehan_	Basically, a new environment is created for the infinite iterator and passed to the take iterator each time.
20:59:47	Jehan_	However, with the fix, it ignores that environment (because it keeps working on the one in p).
21:00:18	bluenote_PC	do you look at the C code to see such things?
21:00:32	Araq	Jehan_: yup... it's all wrong :-)
21:00:51	Araq	that means async might get quite a bit faster when we fixed it :D
21:01:04	OnO	btw. if anyone using TextMate... playing with Nim, I did improvements to its TM bundle: https://github.com/textmate/nim.tmbundle/pull/1
21:01:10	Jehan_	Araq: Actually, I think it's not so much wrong as that iterators are lacking something like Sather's once arguments.
21:02:06	Jehan_	Basically, when an iterator in Sather has an argument declared as once, it's only evaluated once for the iterator, not once per iteration.
21:02:33	OnO	https://cloud.githubusercontent.com/assets/103067/7147306/4a73f632-e2fa-11e4-9c25-b19767de4585.png
21:02:39	Araq	bluenote_PC: I'm not even looking at the compiler. I know how the implementation looks like :P
21:03:11	Araq	spent more time on it than on everything else
21:03:44	bluenote_PC	ok, I see :) I just have to learn the compiler
21:07:00	OnO	btw. one question, why constructors use Type(field1: val1) but proc calls use = for named params proc(param1=val1..), etc?
21:07:03	bluenote_PC	Araq: So you think that fixing the "broken" version is possible? I'm wondering because the code was actually looking right intuitively
21:07:13	OnO	isn't assignment expr in Nim?
21:07:45	Jehan_	I've added a new version to https://gist.github.com/rbehrends/d010bf36663d0c2e113f that works with a `once` template and fixes the issue better.
21:08:08	OnO	so: var width = 400; proc(width = 500) <- I am setting with to 500 and then passing value or calling with with named parameter
21:08:14	OnO	pretty confusing syntax
21:09:19	Jehan_	OnO: Assignment expressions do not return a value if that's what you mean.
21:09:54	Jehan_	If you do f(x = 100), then that means that the argument with name x is assigned the value 100.
21:10:11	Jehan_	Not that a variable x is assigned the value 100 and the result of that passed to f().
21:10:39	OnO	okie, so one liner: x = y = 0 is invalid syntax?
21:10:47	Araq	I simply took that from Python. f(x=10) vs {"foo": 21}
21:11:20	OnO	yeah, but for object initalization it uses colon
21:11:26	OnO	I find it pretty inconsistent
21:11:58	Jehan_	OnO: That's to syntactically distinguish between object initializers and function calls.
21:12:57	Jehan_	At least that's what my guess is. :)
21:13:01	OnO	I don't feel convinced, because it is let x = 1 not let x : 1
21:13:08	OnO	1 is an initializer for x
21:13:12	BlaXpirit	object initializers need to be private :|
21:13:17	Jehan_	bluenote_PC: Okay, I'll get you your write-up tomorrow, and I think it should be pretty clear.
21:13:31	flaviu	My thoughts here are that object initializers should not be distinguishable from function calls.
21:13:44	BlaXpirit	if one can make them private
21:14:00	bluenote_PC	Jehan_: awesome
21:14:15	flaviu	https://github.com/Araq/Nim/issues/1693
21:16:04	BlaXpirit	I just remembered something. there is absolutely no reason for destructors to limit ref objects' behavior
21:16:05	OnO	flaviu: they are, types begin with capital letter, right?
21:16:27	flaviu	.eval proc Foo() = discard
21:16:28	Mimbus	flaviu: <no output>
21:16:34	flaviu	OnO: I see no error ^
21:17:16	BlaXpirit	flaviu, object constructors are rarely useful
21:17:45	BlaXpirit	oh wait, adding procs is exactly what you're suggesting
21:17:55	BlaXpirit	yeah, i want this badly
21:18:16	BlaXpirit	but damn it would be hard to eradicate init/new
21:18:35	BlaXpirit	and if we don't want to eradicate it, then this is an inconsistent addition
21:18:40	Jehan_	OnO: Types begin with a capital letter only by convention.
21:19:07	flaviu	BlaXpirit: This could be fixed by tooling.
21:19:13	Jehan_	See `int`, `cstring`, etc.
21:20:21	flaviu	if a type only has newFoo, create a Foo() that returns a ref. If a type only has a initFoo, create a Foo() that returns a value. If a type has both, create a Foo() that returns a value and replace all calls to newFoo() with box Foo().
21:20:44	bluenote_PC	actually I would even prefer namespaces for construtors like AffineMatrix.createRotation, AffineMatrix.createTranslation, etc
21:21:10	BlaXpirit	flaviu, wow, i have never once used `box` and completely forgot about it
21:21:23	OnO	okay... but let's take an example of:
21:21:23	OnO	type Point = object
21:21:23	OnO	x, y : int
21:21:27	BlaXpirit	i did wonder many times how to make a ref :p
21:21:32	Jehan_	bluenote_PC: You may have a hard time convincing Araq of this. :)
21:21:33	OnO	because of this convention I can't let p = Point(1, 2)
21:21:47	OnO	just like with any other proc call
21:21:51	Jehan_	He's not a fan of multi-dotting, so to speak. :)
21:21:57	BlaXpirit	you can let p = point(1,2)
21:22:08	bluenote_PC	ok :)
21:22:18	flaviu	.eval proc createRotation(self: typedesc[int], x: int): int = x; proc createRotation(self: typedesc[float], x: float): float = x; discard int.createRotation(1); discard float.createRotation(1)
21:22:19	Mimbus	flaviu: <no output>
21:22:23	flaviu	doesn't that work fine?
21:22:30	OnO	BlaXpirit: I need to declare trivial proc... pretty useless
21:22:56	bluenote_PC	flaviu: indeed, I forgot about typedesc
21:23:03	flaviu	.eval proc createRotation(self: typedesc[int], x: int): int = x; proc createRotation(self: typedesc[float], x: float): float = 10*x; echo int.createRotation(1), ", ", float.createRotation(1)
21:23:04	Mimbus	flaviu: 1, 10.0
21:23:06	BlaXpirit	if your constructor is trivial, maybe it should be a tuple
21:23:23	flaviu	The last command was just to verify the correct constructor is getting called
21:24:04	BlaXpirit	this is pretty interesting
21:24:27	BlaXpirit	instead of initType one can make Type.init
21:24:51	flaviu	One last command, just to show that typedesc[int] works,
21:24:54	flaviu	.eval proc createRotation(self: typedesc[int], x: int): int = x; echo float.createRotation(1)
21:24:54	bluenote_PC	which would be _exactly_ what I was asking for :)
21:24:55	Mimbus	flaviu: eval.nim(4, 10) Error: type mismatch: got (typedesc[float], int literal(1))
21:25:14	OnO	BlaXpirit: declaring Point type as tuple does not let me use: let p = Point(1,2) either
21:25:39	BlaXpirit	let p: Point = (1,2)
21:26:06	OnO	yeah, how do constuct in in place ?
21:26:21	OnO	eg. calling rotate(p, Point(3,4)) ?
21:26:37	BlaXpirit	rotate(p, (3, 4))
21:27:35	OnO	hmm... ok... this is pretty interesting
21:27:50	BlaXpirit	also for corner cases Point((1, 2))
21:28:20	OnO	yess... Point((1, 2)) works in place
21:29:40	*	itsmeront joined #nim
21:30:48	OnO	do tuples have some memory overhead comparing to regular objects?
21:31:00	BlaXpirit	no
21:31:31	BlaXpirit	no overhead of any kind
21:31:33	flaviu	Point.fromTup((1, 2)) would also work, if you want that syntax.
21:32:14	OnO	ok... so I could generally make Vec2, Vec3, Vec4 tuples with x, y, z fields.. and then same some templates to defined aritmetics for all 3 of them
21:32:38	OnO	given templated operator can also access length of the tuple
21:32:47	BlaXpirit	ye this is good
21:33:13	BlaXpirit	wait what
21:33:16	BlaXpirit	tuple len doesn't work?
21:33:22	BlaXpirit	wtf
21:35:17	bluenote_PC	isn't it arity for a tuple?
21:36:07	BlaXpirit	who cares. if i can do [index] then there should be a len
21:36:26	BlaXpirit	maybe i'm wrong
21:36:39	OnO	okay, simple task... provide a module for 2, 3, 4 element vectors so vectors can be accessed via v.x, v.y, v.z, v.w or v[0], v[1]... and define arithmetics for them
21:37:02	OnO	I don't want to define + - operators separately for each vector size
21:37:19	OnO	just make it unrolled inlined template for each
21:37:21	*	jsudlow quit (Remote host closed the connection)
21:37:22	BlaXpirit	you can make a template that defines a function
21:37:36	BlaXpirit	not a template that does the operations directly
21:37:44	Jehan_	BlaXpirit: You can only do index with constant values on tuples.
21:38:01	Jehan_	So not sure what len() would be good for.
21:38:17	BlaXpirit	for what we said above
21:40:12	EXetoC	it can be a compile-time operation too
21:40:17	bluenote_PC	well, "abstracting over arity" is a famous problem with tuples (the purpose HList in scalaz, Haskell) since (int, int) is just a different type as say (int, int, int)
21:41:14	Araq	abstracting over arity is easy with macros.getType
21:41:26	bluenote_PC	good example how powerful Nim is since abstracting over tuple arity in not a big deal in Nim
21:41:33	BlaXpirit	or just add the len -_-
21:41:48	bluenote_PC	Araq: exactly, but it is in Scala
21:41:57	flaviu	Didn't someone write a module for working with tuples?
21:42:39	*	skroll1 joined #nim
21:42:51	*	onionhammer joined #nim
21:43:01	flaviu	http://www.vision.caltech.edu/~mmcgill/tuples/index.html
21:43:12	flaviu	BlaXpirit: ^
21:43:35	*	gsingh93_ joined #nim
21:44:00	BlaXpirit	macros... they suck
21:44:05	BlaXpirit	i mean, writing them sucks
21:44:27	OnO	oh... interesting
21:44:38	Jehan_	BlaXpirit: Heh. :)
21:45:00	flaviu	BlaXpirit: Good thing that it's not necessary to write macros very often!
21:45:02	Jehan_	Yeah, I think macro usability has … room for improvement. :)
21:45:52	BlaXpirit	templates should just be made more useful
21:46:07	*	bluenote_PC quit (Ping timeout: 246 seconds)
21:46:35	BlaXpirit	maybe macros should be what they are, but also allow functionality of templates instead of writing trees manually
21:47:18	Araq	BlaXpirit: no, no, no. macros are *easier* to write than templates, as soon as your problem is complex enough
21:48:04	BlaXpirit	only because templates are more limited than they could be
21:48:22	Araq	I disagree
21:49:38	Araq	in templates all you can do is to create more adhoc solutions to problems
21:49:56	Araq	with more syntax that's special in templates
21:50:02	*	gsingh93_ quit (Ping timeout: 264 seconds)
21:50:34	Araq	with macros the level of operation is different. you only need more .compiletime procs, not more and more features
21:50:53	BlaXpirit	i don't want to write goddamn trees
21:51:01	BlaXpirit	especially with that unintuitive api
21:51:21	Araq	so use a template and getAst or parseExpr
21:51:42	Araq	creating trees is easy enough, you're only showing your ignorance here
21:52:01	BlaXpirit	maybe it's "easy" concept-wise
21:52:01	*	buMPnet quit (Remote host closed the connection)
21:52:15	BlaXpirit	it's just grunt work
21:52:22	*	buMPnet joined #nim
21:52:32	Jehan_	What I'd like to see (sometime in the future, not now) is pattern matching on macro argument ASTs.
21:52:58	BlaXpirit	I did not know about getAst, nor do I know what it does after reading about it
21:53:17	Araq	Jehan_: that's a fair feature request, but that's destructuring of ASTs, BlaXpirit complains about AST construction
21:53:50	Jehan_	Araq: Yup, I realize that it's a separate issue. But then, AST construction is no big deal in general.
21:53:50	BlaXpirit	ok so i could make a template to construct a tree and use getAst
21:53:55	Jehan_	At least not for me.
21:54:05	Jehan_	And for what I need.
21:54:10	Jehan_	Don't want to generalize.
21:54:24	BlaXpirit	also unreadable
21:54:51	BlaXpirit	parseExpr is fine but it's isolated
21:56:05	dom96	what about quote?
21:56:26	BlaXpirit	now that's something
21:57:00	BlaXpirit	yes, yes, yes!
21:57:21	BlaXpirit	[:46:33] <BlaXpirit> maybe macros should be what they are, but also allow functionality of templates instead of writing trees manually
21:57:26	BlaXpirit	that's quote, right there
21:58:03	BlaXpirit	thank you
21:58:14	BlaXpirit	maybe it's time to read the whole docs of macros module
21:59:08	OnO	okie, I am going to meet mr sandman, thanks again for great discussion, cya
21:59:09	Jehan_	BlaXpirit: You didn't know about quote?
21:59:15	BlaXpirit	I didn't know
21:59:23	*	buMPnet quit (Remote host closed the connection)
21:59:28	BlaXpirit	are you surprised?
21:59:30	Jehan_	Gotcha. :)
21:59:37	BlaXpirit	i didn't go to macros boot camp
21:59:39	*	buMPnet joined #nim
21:59:40	Jehan_	No, I had just assumed you did.
21:59:53	*	OnO quit (Quit: Textual IRC Client: www.textualapp.com)
22:00:13	Jehan_	That wasn't a "you should know" remark but "oh, that's why he found ast construction tedious" remark.
22:00:23	BlaXpirit	I said I am missing the exact functionality of quote
22:00:35	BlaXpirit	thank goodness dom understood me
22:01:00	*	pregressive quit (Remote host closed the connection)
22:01:06	Jehan_	It's midnight here, my brain is running at 50% capacity at best. :)
22:01:36	*	gsingh93 quit (Ping timeout: 250 seconds)
22:02:04	dom96	BlaXpirit: I only read the last 3 sentences that you guys wrote :P
22:03:01	Jehan_	But yeah, my main problem is typically breaking down arguments into usable form.
22:04:22	Jehan_	Another problem is that I rarely need macros, so when I do need them, it takes some experimentation to figure out how they work again. :)
22:05:55	dom96	Sometimes the hardest thing for me with macros, is deciding whether the effort to write the macro outweighs the problems caused by copying and pasting code.
22:06:42	Araq	dom96: I think many will say this means the macro system is really well designed :P
22:07:06	dom96	It's much nicer now that quote actually works.
22:07:25	dom96	We still need a nicer way to generate these nodes though.
22:07:36	dom96	but that can be a nimble package
22:07:46	BlaXpirit	isn't quote the nicest way? i don't get it
22:10:20	def-	BlaXpirit: you can't build every AST with quote
22:10:29	BlaXpirit	why?
22:10:46	def-	for example a case-of with a variable amount of `of`s (i think)
22:11:37	BlaXpirit	well maybe i can build one with a constant number of ofs and copt some nodes and whatnot
22:11:46	BlaXpirit	copy*
22:12:34	BlaXpirit	oh i get it
22:12:43	BlaXpirit	quote + templating language? anyone?
22:13:24	dom96	You can't for example do:
22:13:30	dom96	node.add quote do:
22:13:36	dom96	of 0x11: foo()
22:13:48	dom96	Because an 'of' needs a 'case'
22:13:50	BlaXpirit	yes, that's fine
22:13:58	dom96	it's a problem if you already have a case generated
22:14:25	dom96	basically means I had to do this: https://github.com/dom96/gbemulator/blob/master/cpu2.nim#L835
22:14:34	BlaXpirit	check this out
22:14:35	dom96	(Github really butchers the syntax highlighting on that)
22:14:36	BlaXpirit	case x
22:14:37	BlaXpirit	{% for i in 0..10 %}
22:14:39	BlaXpirit	of `i`: foo()
22:16:19	BlaXpirit	ah quote is magic.. of cours
22:16:58	Araq	that is exactly the "more template voodoo solution" that ends up as a never ending feature request
22:17:33	BlaXpirit	?
22:17:40	BlaXpirit	this is the best kind of templates
22:18:24	BlaXpirit	rg
22:18:28	BlaXpirit	the real templates
22:18:59	Araq	well if you stop there, it's still not powerful enough
22:19:30	BlaXpirit	point is to intertwine nim code and `quote`
22:20:16	BlaXpirit	in fact, it's the templates that should be extended with this functionality
22:20:31	BlaXpirit	because yes, sometimes this is not powerful enough
22:22:16	BlaXpirit	i'm having a hard time coming up with a difference between a template and a macro that just invokes quote
22:22:28	BlaXpirit	is that what it is?
22:22:46	BlaXpirit	no... no... f course, we have inject and other things
22:23:07	*	saml quit (Remote host closed the connection)
22:25:51	reactormonk	Araq, got the gc stuff, but never ran it
22:26:52	reactormonk	also, how should I implement different serialization formats for nimsuggest? global variables?
22:26:56	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
22:31:01	*	user7181 quit (Quit: leaving)
22:32:00	*	banister joined #nim
22:36:50	*	mwbrown quit (Ping timeout: 264 seconds)
22:39:43	*	Kingsquee quit (Quit: Konversation terminated!)
23:00:00	*	Matthias247 quit (Read error: Connection reset by peer)
23:01:24	*	pregressive joined #nim
23:13:18	a5i	ping dom96 ?
23:13:23	dom96	sup
23:13:45	a5i	I'm really like Ruby, so thats I'm with Crystal
23:14:22	a5i	nothing against Nim tho, its a very nice language I've been following
23:14:38	a5i	* to anser your question from hours earlier
23:16:35	BlaXpirit	does nim have "exec"?
23:16:57	BlaXpirit	an equivalent to writing text to a file and importing it
23:17:31	BlaXpirit	no
23:17:40	BlaXpirit	an equivalent to writing a text file and including it
23:19:22	*	saml_ joined #nim
23:20:12	a5i	dom96, I actually hope to get into Nim soon
23:20:49	dom96	a5i: ahh. No prob.
23:35:12	*	Jehan_ quit (Quit: Leaving)
23:39:20	filwit	BlaXpirit: write to a file in a 'static' block then import it? (havent tried)
23:39:38	BlaXpirit	filwit, i don't want to write a file
23:39:48	BlaXpirit	i want to execute a string of nim code at compiletime
23:42:20	filwit	well you can execute any nim code at compile time of course.. but yeah, dunno how to load a string and execute it all at compile time.. maybe some trick with a macro which builds a macro..
23:42:57	BlaXpirit	i've tried :/
23:43:17	BlaXpirit	macro that invokes parseStmt and returns stmt
23:44:42	BlaXpirit	problem is... the macro can't accept a string
23:45:04	filwit	make it a static[string]
23:45:54	BlaXpirit	oho
23:46:28	BlaXpirit	nope
23:47:10	BlaXpirit	undeclared identifier: '��'
23:47:59	BlaXpirit	it works
23:48:00	BlaXpirit	!!!
23:48:02	BlaXpirit	thxc
23:49:48	filwit	yeah i just got it to work too
23:50:32	*	Trustable quit (Remote host closed the connection)
23:51:12	filwit	BlaXpirit: just to be clear, here's what I did: https://gist.github.com/PhilipWitte/d1a382c8f21d10064fa2
23:51:35	BlaXpirit	why is that needed?
23:51:44	filwit	idk... good point
23:51:48	BlaXpirit	and not just macro with parsestmt
23:51:56	filwit	yeah i was thinking wrong
23:52:49	BlaXpirit	wait
23:52:56	BlaXpirit	macro with parsestmt executes it at runtime
23:53:11	BlaXpirit	loool
23:54:12	filwit	idk what you mean, but this works: https://gist.github.com/PhilipWitte/d1a382c8f21d10064fa2
23:54:29	filwit	i mean, you wanted this to happen at compile-time right?
23:54:46	BlaXpirit	[:39:46] <BlaXpirit> i want to execute a string of nim code at compiletime
23:55:38	BlaXpirit	i want to execute a compiletime string of nim code at compiletime
23:55:40	filwit	ah right, that's why i was making a macro which returned a template.. so you could call the template from some other compile-time function..
23:55:53	BlaXpirit	and we invented a way to execute a compiletime string of nim code at runtime
23:56:05	filwit	one sec
23:56:07	BlaXpirit	aha, filwit
23:57:07	filwit	yeah it works: https://gist.github.com/PhilipWitte/d1a382c8f21d10064fa2
23:57:35	BlaXpirit	appears so, filwit.
23:58:03	BlaXpirit	but omg what is this
23:58:03	filwit	of course it defines some function you need to call by name, but I'm sure something like this can be adapted to your purpose