<< 24-03-2017 >>

---

00:01:09	*	elrood quit (Quit: Leaving)
00:01:12	FromGitter	<Varriount> Well, vtables are references to a dedicated object, which may or may not be a global
00:02:41	FromGitter	<Varriount> Anyway, if multimethods are taken away, there's no adequate substitute.
00:02:42	FromGitter	<brechtm> Why doesn't this this output "inner A" and "inner B"? https://glot.io/snippets/eo9snwfmo1
00:04:28	*	user0____ joined #nim
00:06:59	FromGitter	<Varriount> @brechtm Seems like a bug.
00:07:36	FromGitter	<brechtm> returning a string instead will always return nil
00:08:13	Demos	I think the idea would be to switch to single dispatch methods
00:09:23	FromGitter	<Varriount> @brechtm By the way, what are you using multimethods for?
00:09:55	FromGitter	<brechtm> @Varriount to see how I could translate that to using generics instead :-)
00:10:50	FromGitter	<Varriount> I don't follow.
00:12:55	FromGitter	<brechtm> @Varriount you didn't ask that in relation to the code snippet?
00:14:31	FromGitter	<brechtm> bug logged: https://github.com/nim-lang/Nim/issues/5599
00:18:20	FromGitter	<Varriount> Hm, this is interesting. Generics don't seem to take into account procedure overloads defined after the generic.
00:19:20	FromGitter	<Varriount> @brechtm https://glot.io/new/nim
00:19:32	FromGitter	<Varriount> https://glot.io/snippets/eo9t5heq1k
00:20:07	FromGitter	<Varriount> I find it odd that the symbol has to be forced open with mixin, but w/e
00:20:23	FromGitter	<brechtm> @Varriount yup, that's what I came up with also
00:20:33	FromGitter	<brechtm> @Varriount but what's the mixin line for?
00:21:04	FromGitter	<Varriount> https://nim-lang.org/docs/manual.html#generics-symbol-lookup-in-generics
00:21:14	FromGitter	<brechtm> It seems I didn't need that line
00:21:28	FromGitter	<brechtm> But this was in my application, not in this sample code
00:21:50	FromGitter	<Varriount> It works without mixin if you rearrange the procedure implementation orders
00:23:48	FromGitter	<brechtm> Ah, the difference was that I didn't define inner(obj: Base)
00:24:48	FromGitter	<Varriount> Generics tend to work intuitively, except when they don't
00:25:17	*	Matthias247 quit (Read error: Connection reset by peer)
00:25:18	FromGitter	<brechtm> :-)
00:26:45	FromGitter	<Varriount> Actually, I wonder if @Araq would be willing to explain why symbols need to be closed at all.
00:27:45	Araq	proc complexAlgorith[T](x) = partOne(x); partTwo(x); partThree(x)
00:28:02	Araq	in what world do you want partOne() etc to be open?
00:28:41	FromGitter	<Varriount> One in which you want the user to supply an overload?
00:28:53	Araq	it's not part of the contract of 'T' to have a partOne proc, it's an implementation detail
00:29:07	Araq	hence you want it "private" aka "closed".
00:29:29	FromGitter	<Varriount> @Araq Yes, but in the posted snippet, the generic wasn't considering the implementations defined after it
00:29:53	Araq	so?
00:30:59	FromGitter	<Varriount> So, I'd consider the generic symbol lookup to consider the entire module, since a module usually contains an entire implementation.
00:32:04	Araq	why? Nim doesn't consider the entire module at all. it requires forward decls.
00:32:48	Araq	you can argue that forward decl suck but you cannot argue that it's inconsistent behaviour.
00:33:16	FromGitter	<Varriount> No, just unintuitive for people coming from other languages.
00:34:20	Araq	*shrug* what isn't.
00:34:36	Araq	everything is always confusing for programmers.
00:35:47	FromGitter	<Varriount> I wonder what Swift does
00:36:18	Araq	it considers the class context, what else.
00:36:53	Araq	but Nim has no class context :P
00:37:04	FromGitter	<Varriount> It does have module context though.
00:42:06	*	themagician quit ()
00:45:13	FromGitter	<Varriount> Araq: So if you remove multi-methods, what's going to replace them?
00:46:11	Araq	I've already explained it.
00:46:55	FromGitter	<Varriount> Oh, single-dispatch
00:47:44	FromGitter	<Varriount> @Araq: Isn't that what is currently done? I though multimethods dispatched on the type of the object.
01:00:49	FromGitter	<Varriount> @brechtm How are you finding Nim so far?
01:10:11	*	def-pri-pub joined #nim
01:10:11	*	def-pri-pub quit (Changing host)
01:10:11	*	def-pri-pub joined #nim
01:15:32	*	Demos quit (Quit: Konversation terminated!)
01:59:59	*	sleepy998 joined #nim
02:03:42	*	user0___ joined #nim
02:03:42	*	user0____ quit (Read error: Connection reset by peer)
02:04:36	sleepy998	is there any proc that can process other completed async 'threads' if there are any, else do nothing? currently i'm using sleepAsync(1), but i want zero added latency.. and sleepAsync(0) seems to not let other 'threads' run
02:06:22	*	rauss quit (Quit: WeeChat 1.7)
02:06:51	*	user0___ quit (Client Quit)
02:09:18	*	xet7 joined #nim
02:11:00	sleepy998	i also tried using hasPendingOperations() to conditionally sleepAsync(1) but that apparently returns true even if nothing else has completed (unless i overlooked something)
02:14:11	sleepy998	if sleepAsync(0) relinquished, that would suffice. i'm not sure why it doesn't.
02:29:01	sleepy998	oh nevermind, poll(0) works :o
02:30:57	FromGitter	<Varriount> sleepy998: You might want to wait until dom96 is awake, he's the expert on async stuff
02:38:28	FromGitter	<Varriount> sleepy998: Shouldn't the framework just complete other asynchronous tasks once you yield?
02:39:01	*	chemist69 quit (Ping timeout: 260 seconds)
02:39:47	sleepy998	Varriount: yes; the problem was how to yield an infinitesimally small amount of time
02:39:48	FromGitter	<barcharcraz> Idea: c++ffi via clang pch files
02:40:48	FromGitter	<Varriount> @barcharcraz Is that format documented anywhere?
02:41:31	*	user0___ joined #nim
02:45:58	FromGitter	<barcharcraz> I think so
02:46:37	FromGitter	<barcharcraz> https://clang.llvm.org/docs/PCHInternals.html
02:46:54	FromGitter	<barcharcraz> I think the clangserialization library can even real em
02:47:01	FromGitter	<barcharcraz> *read
02:49:58	FromGitter	<barcharcraz> Ine thing idk about is how to instantiate a cop Templar with a nim type, that seems complicated
02:50:15	FromGitter	<barcharcraz> Jesus iPhone autocomplete is bad
02:50:46	FromGitter	<barcharcraz> Also see dlang's calypso project
02:51:03	*	user0___ quit (Quit: user0___)
02:52:38	*	chemist69 joined #nim
02:58:56	def-pri-pub	Has anyone yet made any JS bindings for audio playback or keyboard events yet?
03:01:02	*	libman quit (Quit: Connection closed for inactivity)
03:02:48	def-pri-pub	Nevermind, might be able to do everything with the `dom` package, but it's not going to be "elegant."
03:09:10	*	bjz joined #nim
03:27:05	FromGitter	<barcharcraz> I mean tbh creating a video or audio element is how you do it
03:51:45	*	Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
03:58:39	demi-	is it possible to build a nimble package as an executable binary as well as dynamic library?
03:59:35	FromGitter	<vegansk> @demi-, use ``when isMainModule: callMainFunctionHere()``
03:59:50	demi-	yeah i'm doing that and get problems :\
04:00:11	FromGitter	<vegansk> Then show the code and the problem
04:00:58	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
04:11:04	*	Serenitor joined #nim
04:11:43	demi-	i think i have worked around it for now
04:12:06	demi-	now to work out how to get the existing set of environment variables
04:13:25	FromGitter	<vegansk> https://github.com/nim-lang/Nim/blob/master/lib/pure/os.nim#L814
04:14:14	*	def-pri-pub quit (Quit: leaving)
04:48:03	*	Nobabs27 quit (Quit: Leaving)
05:00:10	*	vivus quit (Quit: Leaving)
05:19:46	*	bjz joined #nim
05:39:48	*	Tiberium joined #nim
05:41:49	*	Calinou quit (Read error: Connection reset by peer)
05:52:25	*	Serenitor quit (Quit: Leaving)
05:53:34	*	yglukhov joined #nim
05:55:57	*	BitPuffin|osx quit (Ping timeout: 260 seconds)
05:56:08	*	yglukhov quit (Remote host closed the connection)
05:56:23	*	yglukhov joined #nim
06:00:13	*	vlad1777d joined #nim
06:00:31	*	enthus1a1t joined #nim
06:02:47	*	enthus1ast quit (Ping timeout: 264 seconds)
06:23:19	Tiberium	ehm, noob question - if I have an "url" variable in global section, can I modify it from while loop section?
06:25:25	FromGitter	<vegansk> @Tiberium, yes unless you iterating over it
06:25:53	FromGitter	<TiberiumPY> @vegansk no, I'm not iterating over it
06:26:04	FromGitter	<TiberiumPY> so how to do it?
06:27:02	FromGitter	<vegansk> ```var url = "x" ⏎ for x in 1..5: ⏎ url &= url & $x``` [https://gitter.im/nim-lang/Nim?at=58d4bc34a84f611959da1012]
06:27:17	FromGitter	<TiberiumPY> ah, I'm stupid
06:27:23	FromGitter	<TiberiumPY> I was doing it with "var" like
06:27:47	FromGitter	<TiberiumPY> ```var url = "" ⏎ while true: ⏎ var url = "something else"``` [https://gitter.im/nim-lang/Nim?at=58d4bc618fa16339540f0b17]
06:27:48	FromGitter	<TiberiumPY> thanks
06:35:05	*	girvo quit (Quit: Lost terminal)
06:40:35	*	nsf joined #nim
06:51:12	*	zachcarter joined #nim
06:53:20	*	SusWombat_ quit (Remote host closed the connection)
06:55:02	*	SusWombat joined #nim
06:59:01	SusWombat	Morning
06:59:19	SusWombat	Anyone a idea why lineDir doesnt work for me on linux?
07:05:16	*	sleepy998 quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
07:12:39	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
07:14:21	*	Tiberium quit (Remote host closed the connection)
07:14:25	zachcarter	nope :/
07:15:29	zachcarter	are you talking about with the compiler?
07:15:33	zachcarter	the switch?
07:15:38	zachcarter	SusWombat^
07:28:36	*	bjz joined #nim
07:35:53	*	Vladar joined #nim
07:42:10	*	tankfeeder joined #nim
08:07:59	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
08:09:14	*	bjz joined #nim
08:09:20	SusWombat	zachcarter, yeah
08:09:29	zachcarter	sec
08:09:39	SusWombat	zachcarter, i think we even discussed this together like 1-2 weeks ago :D
08:09:50	zachcarter	ah possibly
08:09:55	zachcarter	I remember a similar discusison
08:10:01	zachcarter	still no luch huh?
08:10:08	SusWombat	no :/
08:10:27	zachcarter	hrmmm
08:10:35	zachcarter	you’re using gdb right?
08:10:38	SusWombat	i compile with lineDir on but i still get some "random" code when i press l in gdb/lldb
08:10:39	zachcarter	have you tried lldb?
08:10:44	zachcarter	ah okay
08:10:48	zachcarter	have you tried linedir
08:10:51	zachcarter	not camel case?
08:11:23	zachcarter	nim c --linedir:on --debuginfo c hello
08:12:58	SusWombat	yeah same
08:13:10	SusWombat	zachcarter, the double c is a tyo right?
08:13:16	SusWombat	typo*
08:13:16	cheatfate	nim c --debugger:native --linedir:on hello
08:13:40	zachcarter	I copied and pasted that from a blog
08:13:42	zachcarter	but yeah probably
08:13:52	SusWombat	cheatfate, same :/
08:14:26	cheatfate	what is `l` command doing?
08:14:51	SusWombat	should list my code no?
08:15:05	SusWombat	cheatfate,
08:15:46	SusWombat	http://i.imgur.com/PCZG0Yw.png
08:16:09	SusWombat	the file only contains " echo "Hello World" "
08:17:56	cheatfate	SusWombat, you need to execute `b NimMainModule`
08:17:58	cheatfate	r
08:18:02	cheatfate	then `r`
08:18:08	cheatfate	then one `n`
08:18:14	cheatfate	and you are at your code
08:19:27	*	couven92 joined #nim
08:20:45	SusWombat	cheatfate, http://i.imgur.com/UUiNZdT.png
08:20:50	SusWombat	thats what happens for me
08:21:09	cheatfate	yeah i know
08:21:12	cheatfate	make one step
08:21:19	cheatfate	with `n`
08:21:22	*	Andris_zbx joined #nim
08:21:43	cheatfate	this is workaround :)
08:22:01	cheatfate	of course you can make `b helloworld.nim:<linenumber>`
08:22:23	cheatfate	so you can avoid walking through stub code (which is very very unclear)
08:22:56	SusWombat	cheatfate, ok after next step its works O: Do you happen to know why that happens?
08:23:41	SusWombat	it*
08:23:52	cheatfate	SusWombat, nope, and i have asked Araq, he dont know too :)
08:24:15	SusWombat	cheatfate, do other debuggers work without that workaround?
08:24:46	cheatfate	SusWombat, nope this workaround for gdb/lldb and windows debuggers too
08:25:15	*	couven92 quit (Quit: Client disconnecting)
08:25:32	SusWombat	cheatfate, as it seems it works for zachcarter out of the box i guess because he uses mac?
08:25:44	*	couven92 joined #nim
08:26:17	cheatfate	SusWombat, i dont think it workd better for zachcarter
08:26:36	cheatfate	but i think he make `b somecode.nim:<someline>` before executing
08:26:44	cheatfate	so he never seen stub code
08:26:51	SusWombat	zachcarter, could you comment on that?
08:27:06	zachcarter	hrm? I don’t have to do much to get debug line info in lldb
08:27:23	zachcarter	I pretty much do this -
08:27:53	zachcarter	nim —linedir:on —debuginfo c hello.nim
08:28:20	zachcarter	on osx with devel version of nim
08:28:44	SusWombat	zachcarter, and if you open the binary with lldb and press l whats the output?
08:29:09	zachcarter	sec
08:30:00	zachcarter	https://gist.github.com/zacharycarter/d36edf74f5c30690da2b0bc02cdd4a6f
08:30:42	SusWombat	ah well you have the same problem :/
08:30:55	SusWombat	btw thanks cheatfate zachcarter
08:31:16	zachcarter	sure
08:31:28	zachcarter	sorry I’m totally distracted haha, trying to get shaders to copmile for dx11 on my windows vm
08:31:39	SusWombat	no problem
08:38:09	*	vivus joined #nim
08:39:26	vivus	in this lib: https://nim-lang.org/docs/db_mysql.html . am I limited to searching for strings by what mysql offers?
08:45:23	zachcarter	vivus - I don’t understand your question
08:45:40	zachcarter	I don’t think that module adds any functionality to mysql
08:45:44	zachcarter	if that’s what you’re asking
08:46:47	vivus	zachcarter: say I want to search for pattern-matched strings in a mysql table. does the db_mysql library add any functionalities for doing this or is the library just a wrapper for what mysql provides?
08:47:01	zachcarter	I’m guessing the latter
08:47:11	zachcarter	but maybe not
08:47:16	zachcarter	it does say higher level wrapper
08:47:28	zachcarter	I’m not a user so a user would be better to comment, but thank you for clarifying your question
08:48:36	vivus	zachcarter: will it be possible to fetch each row from the table, treat it as a string/list/nims-equivalent-of-a-python-dictionary, do the things I want to do and then re-add each row back to its original place (after being modified) ?
08:49:37	zachcarter	vivus: I have no idea, I’ve never done any DB work with Nim. Looking at the module, it looks like doing something similiar to that should be possible
08:49:48	zachcarter	maybe try the .rows iterator
08:49:50	zachcarter	see what you get
08:50:01	vivus	yeah I'm going to try fetching each row to see what happens
08:50:11	vivus	gotta crawl before I can walk in nim :)
09:00:45	*	Tiberium joined #nim
09:05:38	zachcarter	true
09:07:11	SusWombat	what are the default values for object members? and can i set them when declaring a type?
09:09:02	*	Ven joined #nim
09:09:25	*	Ven is now known as Guest14586
09:12:59	cheatfate	SusWombat, var a = ObjectType(field1: value1, field2: value2, ...)
09:13:12	Tiberium	I'm wondering - is there any website like codecombat or codingame with Nim support?
09:14:21	cheatfate	SusWombat, var a = ObjectType() initializes memory of object to `0x00` bytes
09:14:44	SusWombat	cheatfate, ok thanks
09:15:16	SusWombat	Tiberium, if you find one pls be so kind and tell me about it :)
09:15:53	Tiberium	SusWombat, ok, will spend some time googling :)
09:18:20	Tiberium	ehh, codingame contains many languages, but no "Nim" in list, but there's even a Swift, C, C++, and so on
09:23:10	vivus	Tiberium: for an online nim compiler? I believe there is 1
09:23:25	Tiberium	vivus, no, I know there's a Nim compiler
09:23:35	Tiberium	vivus, I want to find a coding game like "codingame"
09:23:42	Tiberium	with Nim support
09:25:33	FromGitter	<brechtm> @Varriount How I am finding Nim so far? Ease of coding seems to be very close to Python. But I'm still struggling to let go of OOP and get used to types (and refs).
09:35:30	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
09:37:09	*	bjz joined #nim
09:41:22	FromGitter	<dom96> Someone from codingame emailed me about adding Nim support
09:41:28	FromGitter	<dom96> so it might happen
09:41:44	zachcarter	anyone with a windows machine want to help me out and test my game framework on their machine?
09:41:58	zachcarter	I’m getting some weird results in my DX11 shaders and I’m wondering if it’s because I’m using a VM
09:42:02	zachcarter	to test with
09:42:15	*	xet7 quit (Quit: Leaving)
09:45:44	zachcarter	I’m going to chalk it up to it being DX11 and me running a vm :P
09:46:03	*	nhywyll joined #nim
09:48:54	*	nhywyll quit (Client Quit)
10:01:26	FromGitter	<TiberiumPY> @dom96 how long time ago they emailed y
10:15:45	*	Guest14586 quit (Ping timeout: 268 seconds)
10:16:13	FromGitter	<Varriount> zachcarter: I might be able to.
10:29:24	*	couven92 quit (Quit: Leaving)
10:29:41	*	couven92 joined #nim
10:32:18	*	arnetheduck joined #nim
10:45:13	FromGitter	<brechtm> The following crashes on Nim devel on macOS: https://glot.io/snippets/eoaac145ag
10:45:43	FromGitter	<brechtm> But not if line 9 is uncommented
10:45:57	FromGitter	<brechtm> Can someone try this on Linux or Windows?
10:46:15	Araq	I don't think we claim you can echo typeinfo
10:47:14	FromGitter	<brechtm> Error: execution of an external program failed: 'clang -c -w -g -O3 -I/usr/local/Cellar/nim/HEAD-0d8a503/nim/lib...
10:48:29	FromGitter	<brechtm> @Araq The C compilation step errors, I mean. Is that normal?
10:49:09	zachcarter	Varriount: if you’d still be willing please lmk, sorry went to eat breakfast
10:49:50	FromGitter	<TiberiumPY> It doesn't crash for me
10:49:56	FromGitter	<TiberiumPY> (linux gcc)
10:49:59	FromGitter	<TiberiumPY> will try with clang now
10:50:17	FromGitter	<TiberiumPY> it's fine
10:50:24	FromGitter	<Varriount> zachcarter: I can test in the morning, I'm going back to sleep now.
10:50:31	zachcarter	Alright thanks :D
10:50:35	zachcarter	have a good night!
10:50:41	FromGitter	<brechtm> @TiberiumPY strange
10:51:26	Araq	brechtm: (value: ...rawTypePtr: ...) # produces this output for me
10:51:39	Araq	which is ok, the fields are hidden
10:52:17	FromGitter	<brechtm> @Araq I get this: http://pastebin.com/tPS2qJpe
10:52:42	FromGitter	<TiberiumPY> my clang version is 3.9.1
10:53:38	*	krux02 joined #nim
10:54:15	FromGitter	<brechtm> Hrm, it does seem to work now that I updated my Nim to current devel
10:54:35	FromGitter	<TiberiumPY> that's it :)
10:54:36	FromGitter	<brechtm> Nope, it's something to do with VSCode... nm
10:54:43	FromGitter	<TiberiumPY> ?
10:54:46	FromGitter	<TiberiumPY> I use VSCode too
10:55:33	FromGitter	<brechtm> I have no idea what's going on
10:59:01	zachcarter	what’s the error brechtm?
10:59:50	FromGitter	<brechtm> Ah, it's because in VSCode I'm passing more options to nim
11:00:42	FromGitter	<brechtm> @TiberiumPY try passing -d:release
11:01:38	FromGitter	<brechtm> Yup, that makes it fail on glot.io as well
11:02:24	*	Snircle joined #nim
11:02:30	FromGitter	<TiberiumPY> can you provide exact command?
11:02:50	FromGitter	<TiberiumPY> so I need to uncomment commented line
11:02:54	FromGitter	<TiberiumPY> and run with clang and release
11:02:58	FromGitter	<brechtm> nim c -d:release file.nim
11:02:58	FromGitter	<TiberiumPY> it works for me...
11:03:04	FromGitter	<brechtm> no, don't uncomment
11:03:07	zachcarter	https://glot.io/snippets/eoaac145ag
11:03:10	zachcarter	works with -d:release
11:03:11	FromGitter	<brechtm> uncommenting "fixes" it
11:03:18	FromGitter	<TiberiumPY> ah
11:03:19	FromGitter	<TiberiumPY> it is
11:03:23	FromGitter	<TiberiumPY> fails with comment
11:03:25	zachcarter	whoops
11:03:30	zachcarter	hrm how do I update your snippet
11:03:35	FromGitter	<TiberiumPY> both on gcc and clang
11:03:55	zachcarter	https://glot.io/snippets/eoaawusoex
11:03:56	zachcarter	there
11:04:00	zachcarter	that works with -d:release
11:04:01	FromGitter	<brechtm> @zachcarter You need to fork it first, then change it, then save it again... annoying, huh?
11:04:13	zachcarter	you guys can’t echo that
11:04:19	zachcarter	use repr
11:04:30	FromGitter	<TiberiumPY> but it's compile-time rror
11:04:32	FromGitter	<TiberiumPY> in C
11:04:44	FromGitter	<TiberiumPY> which is always bad
11:04:48	zachcarter	Araq: I don't think we claim you can echo typeinfo
11:05:16	zachcarter	I think that’s what he’s referring to
11:05:20	zachcarter	could be wrong
11:05:51	FromGitter	<brechtm> @zachcarter But it doesn't make sense that it works when uncommenting line 9
11:06:11	zachcarter	kind is a string?
11:06:15	FromGitter	<brechtm> And as @TiberiumPY says, I don't suppose this should generale C-compile error
11:06:23	zachcarter	probably not
11:06:38	Araq	it shouldn't. but it also doesn't for me.
11:06:55	zachcarter	I don’t know why the release define causes that
11:06:57	FromGitter	<brechtm> Araq: neither with -d:release?
11:07:34	Araq	ah, it does
11:07:58	FromGitter	<brechtm> @Araq let me know if you want me to create an issue for this
11:08:14	Araq	yeah sure go ahead.
11:08:29	Araq	not seeing how it's invalid C code
11:12:59	*	user0___ joined #nim
11:15:08	*	bjz quit (Ping timeout: 246 seconds)
11:16:04	*	bjz_ joined #nim
11:22:29	*	user0___ quit (Quit: user0___)
11:22:44	*	user0___ joined #nim
11:26:48	*	bjz joined #nim
11:27:02	FromGitter	<brechtm> How can I go from an Any (akObject) to the typedesc?
11:27:08	*	user0___ quit (Client Quit)
11:28:13	*	bjz_ quit (Ping timeout: 260 seconds)
11:29:13	Tiberium	zachcarter, I'll reboot to Windows now (I have windows 10), and I'll test your app. Will you precompile it?
11:29:18	*	Tiberium quit (Remote host closed the connection)
11:29:28	zachcarter	sure I already have it precompiled
11:29:34	zachcarter	let me zip it up fo ryou thank you Tiberium
11:30:20	FromGitter	<TiberiumPY> OK :)
11:32:28	zachcarter	going to figure out how to get this off my vm then I’ll have to you :)
11:32:47	FromGitter	<TiberiumPY> you can use services like transfer.sh
11:32:54	FromGitter	<TiberiumPY> to upload and share files
11:32:59	FromGitter	<TiberiumPY> it doesn't require any registration
11:33:05	FromGitter	<TiberiumPY> https://transfer.sh
11:35:00	zachcarter	thanks
11:38:28	zachcarter	https://transfer.sh/xtAsS/zach%27s%20dx11%20app.zip
11:39:25	FromGitter	<TiberiumPY> wait, I need to download SDL2, I can do this manually :)
11:39:37	zachcarter	ah crap forgot about that
11:39:43	zachcarter	okay thanks :)
11:40:38	FromGitter	<TiberiumPY> ah
11:40:43	FromGitter	<TiberiumPY> you forgot to send me assets :)
11:40:47	zachcarter	whoops
11:40:51	FromGitter	<TiberiumPY> C:\Users\Tiberius\Desktop\../assets\textures/test01.png does not exist.
11:40:59	zachcarter	let me do that real quick
11:41:20	FromGitter	<TiberiumPY> no, I do not hurry
11:42:14	*	Ven joined #nim
11:42:37	*	Ven is now known as Guest63163
11:42:41	zachcarter	https://transfer.sh/jyeGi/assets.zip
11:42:46	zachcarter	:P
11:43:00	zachcarter	quick is relative anyway
11:44:05	FromGitter	<TiberiumPY> so I started it
11:44:27	FromGitter	<TiberiumPY> http://image.prntscr.com/image/fed3a3a6bcc64df49690fcd8a6654790.png
11:44:36	FromGitter	<TiberiumPY> it is that I get
11:44:38	zachcarter	okay thanks that’s what I”m seeing too on my screen
11:44:42	zachcarter	do you have DX11>
11:44:43	zachcarter	?
11:44:45	FromGitter	<TiberiumPY> yeah
11:44:52	FromGitter	<TiberiumPY> GTX 750
11:45:03	zachcarter	gotcha, okay thank you for trying that I appreciate it very much!
11:45:28	FromGitter	<TiberiumPY> I always ready to help with game development projects :)
11:45:37	zachcarter	well I’m welcoming contributors to mine :D
11:45:46	zachcarter	https://github.com/fragworks/frag
11:45:54	FromGitter	<TiberiumPY> no, I'm actually a newbie in Nim
11:46:13	zachcarter	ah well you don’t have to be a pro to help
11:46:20	zachcarter	so much work to be done
11:46:42	FromGitter	<TiberiumPY> but I'm really a newbie, I've came from Python
11:46:50	zachcarter	:D well glad to have you
11:47:07	zachcarter	Nim is very cool
11:47:17	FromGitter	<brechtm> Hrm, getting nowhere trying to get the base type for an object type using typeinfo.base
11:47:26	FromGitter	<brechtm> Might as well get on with the real work...
11:50:28	*	couven92 quit (Quit: Client disconnecting)
11:58:41	*	elrood joined #nim
12:17:11	zachcarter	http://imgur.com/a/JG9y7
12:17:13	zachcarter	got it working
12:18:18	FromGitter	<TiberiumPY> wow, nice
12:19:51	zachcarter	thanks
12:24:23	*	kunev quit (Quit: е те!)
12:28:24	*	_yeeve quit (Quit: Leaving)
12:28:40	*	_yeeve joined #nim
12:28:40	*	_yeeve quit (Client Quit)
12:29:02	*	yeeve joined #nim
12:31:15	*	kunev joined #nim
12:33:08	*	Tiberium joined #nim
12:39:40	krux02	brechtm, are you still trying to implement super with a macro?
12:39:56	FromGitter	<brechtm> @krux02 I gave up
12:40:18	FromGitter	<brechtm> @krux02 well, I was now trying to use typeinfo.base, so no macro
12:50:41	*	zachcarter quit (Quit: zachcarter)
12:52:05	krux02	http://ix.io/pdP
12:52:11	krux02	forget about typeinfo
12:52:30	krux02	brechtm: I implemented super for you
12:52:54	krux02	you still need to do proccall, but the super does the casting for you
12:53:01	FromGitter	<timeyyy> is cmake what neovim uses for building all the things in .deps, ?
12:53:09	krux02	and it is written in a way that it assumes a ref type, and casts to a ref type
12:53:45	krux02	timeyyy well cmake is the default build system for c++ projects
12:53:54	krux02	well not the default, but it is getting there
12:54:11	FromGitter	<timeyyy> but vim is c :)
12:54:12	krux02	I don't know about .deps,
12:54:19	krux02	it also works for c
12:54:42	krux02	It is a horrible language that makes builds complicated
12:54:44	krux02	I hate it
12:55:09	krux02	but it get's the job done for most big projects and it allows cross platform builds
12:55:33	FromGitter	<timeyyy> i want to automatically download system dependencies for my python project
12:56:00	krux02	I have no idea about python, but cmake normally does not download dependencies
12:56:13	krux02	I think you can write a cmake script that does it, but normally it does not do that
12:56:32	krux02	and what do you mean with downloading system dependencies?
12:56:45	krux02	aren't system dependencies those things you can't download
12:56:59	FromGitter	<timeyyy> well when i do `make` with neovim
12:56:59	krux02	otherwise they would not be called "system" dependencies
12:57:11	krux02	I only know vim
12:57:32	krux02	and in vim make would call make by default, but you could replace it with any other command
12:57:39	FromGitter	<timeyyy> yeah i mean, when i ship my final product to the user it will depend on xclip for example
12:57:45	krux02	cmake generates (normally) makefiles
12:57:56	FromGitter	<timeyyy> or lets say luajit
12:58:07	krux02	so when you have a cmake project and you execute cmake already, make should be enough
12:58:26	krux02	those makefiles are written in a way that they call cmake when needed
12:58:41	krux02	doesn't always work though and you still have to call cmake from time to time
12:59:17	krux02	well you have some options
12:59:38	krux02	when you have a git repository, you can put your dependencies in a git subtree
12:59:59	krux02	then you ship with your dependencies, but then they are also not system dependencies anymore
13:00:11	krux02	cmake normally helps you detect system dependencies
13:00:18	FromGitter	<timeyyy> ok hold on
13:00:35	FromGitter	<timeyyy> so i want to bundle the deps in a reproducible way
13:00:48	krux02	yes then I recommend you do git subtree
13:01:01	FromGitter	<timeyyy> 1) download, 2. compile if required
13:01:38	krux02	with git subtree, your build does not download the dependencies, the dependencies are bundled in the download of the project
13:02:15	FromGitter	<timeyyy> as source?
13:02:18	krux02	yes
13:02:41	krux02	and then cmake can be configured to trigger the build of all dependencies
13:03:03	krux02	you can also bundle binaries, but that is much less reproducable
13:03:45	vivus	I defined a seq like this: dbstore: seq[string] ... and then ran this code: for x in fetchdb.fastRows(sql("select content from posts")): dbstore.add(x) , but it gives me this error: SIGSEGV: Illegal storage access. (Attempt to read from nil?) ... what am I doing wrong ?
13:04:12	FromGitter	<timeyyy> ok thanks for the help
13:04:21	krux02	dbstore is nil
13:04:32	krux02	timeyy, no problem
13:05:08	vivus	krux02: what do I do instead?
13:05:19	krux02	timeyyy, but those dependencies you mentioned seem to be very stable dependencies
13:05:35	krux02	timeyyy, so you could rely on them
13:05:51	krux02	cmake will trigger an error when they are not found, and users can install them
13:06:20	FromGitter	<timeyyy> well i am making an appimage
13:06:27	krux02	I would rather suggest bundling dependencies that are less stable and might introduce incompatible api changes
13:06:32	FromGitter	<timeyyy> so the idea is a bundle everything and it will just work
13:06:44	krux02	yes then git subtree is your friend
13:07:15	krux02	vivus, well I never work with sql, but you have a segfault
13:07:26	krux02	and what I said was just a guess
13:07:32	FromGitter	<timeyyy> now that you mention it i think neovim has all it's deps in a git repo
13:07:35	krux02	one of your variables you use is not initialized
13:07:58	krux02	and most of the time this happens is, when you try to append to an nil seq or a nil string
13:08:19	krux02	so before you call add on a seq, you should call newSeq(0) on it
13:08:36	krux02	or initialize it with = @[]
13:08:59	vivus	krux02: basically, I want to store the strings from the DB into a sequence/array/dictionary to do some 'work' on them within the program
13:09:22	vivus	krux02: should I use a sequence or some other advanced-type ?
13:09:37	krux02	seq is a very fine type
13:09:42	krux02	you can't overuse it
13:10:14	vivus	which type should I then use for a dynamic array-like type ?
13:10:16	krux02	array is too static for this context (fixed length)
13:10:29	krux02	vivus, you should use seq
13:10:33	krux02	that is your type
13:10:47	krux02	but you have to initialize it
13:10:52	vivus	alright I see the docs shows initialization, so I need to do that
13:11:32	krux02	I think it's sad that this needs to be initialized, but that is how it is
13:12:20	vivus	krux02: why is the initialization necessary? so that it can be placed on the heap for GC?
13:12:36	vivus	and do I initialize it in a proc or at the global level?
13:15:13	krux02	vivus, it's best when you initialize where you also declare it
13:15:26	demi-	dom96: when you are around i think i found a significant bug in nimble and not sure how to proceed with fixing it.
13:17:16	krux02	reactormonk, do you still maintain nim-mode?
13:17:26	reactormonk	krux02, nope.
13:17:34	krux02	hmm
13:17:37	FromGitter	<timeyyy> hey krux02, i thought i was in the neovim channel lol!
13:17:48	FromGitter	<timeyyy> now it makes sense ^^
13:18:02	krux02	np, was a bit off topic though
13:18:42	krux02	reactormonk, that was a quick reply, but can you give me some advice, on how I could take over the maintenance of the package?
13:18:57	krux02	because I have the feeling nim-mode is orphaned
13:19:44	krux02	since I rely on it it would be very nice if I would at least be able to fix problems with it
13:19:57	*	SusWombat quit (Remote host closed the connection)
13:21:35	*	SusWombat joined #nim
13:21:36	*	SusWombat quit (Remote host closed the connection)
13:21:57	reactormonk	krux02, not sure I can still vet PRs on that, if Araq agrees, I can just give you collab.
13:22:02	krux02	timeyyy, when you look at the YouCompleteMe plugin, then you can see how you could deal with system dependencies, that plugin bundles boost
13:22:18	*	user0___ joined #nim
13:22:23	reactormonk	and yuutayamada for that matter.
13:23:00	krux02	reactormonk, it's not about getting access rights, for now I am ok with creating pull requests, as long as they are actually merged, but I need some information on how to actually develop an emacs package
13:23:10	krux02	what are the right tools to do it, how to test changes etc
13:23:29	krux02	and where to put code that I want to be executed
13:23:41	vivus	thanks krux02 that worked
13:24:46	*	SusWombat joined #nim
13:25:41	krux02	@brechtm: did you test http://ix.io/pdP ?
13:28:20	reactormonk	krux02, check the makefile
13:40:52	demi-	hmmm, i'm trying to run a process using osproc's startProcess; that is all working fine but the process i'm starting is interactive, how do i make it "take over" while it is running?
13:41:05	*	zachcarter joined #nim
13:42:35	flyx	demi-: well, you'd need to dump the process' output to stdout and feed stdin to the process' input
13:43:26	demi-	and that isn't what "poParentStreams" is for?
13:43:39	SusWombat	Can someone tell me what im doing wrong here? https://hastebin.com/kolubazoki.nim
13:43:53	SusWombat	i get: SIGSEGV: Illegal storage access. (Attempt to read from nil?)
13:44:26	flyx	demi-: hmm, it should
13:45:10	demi-	yeah i am not seeing that
13:45:15	Araq	demi-: poParentStreams is only implemented for Windows :-)
13:45:23	demi-	
13:45:38	vivus	which lib should I use for reading/writing strings to files ?
13:45:50	demi-	Araq then it seems that the majority of osproc is only implemented on windows
13:46:38	flyx	vivus: depends on your use-case. there are simple procs in system, and there is the streams module
13:46:47	Araq	you can use os.execShellCmd that does poParentStreams
13:47:05	vivus	SusWombat: I had a similar problem above. you need to initilize something in your code
13:47:10	Araq	also majority of osproc my ass, I only touch this code when I'm drunk for a reason
13:47:14	SusWombat	Is there somewhere a list on stuff thats only implemented for windows?
13:47:29	Araq	it's a never ending pita in posix land
13:47:35	demi-	haha
13:47:40	demi-	ain't that the truth
13:47:57	vivus	so when calling osproc and it bricks your OS, blame Araq for writing code while drunk
13:47:59	*	user0___ quit (Quit: user0___)
13:48:22	flyx	SusWombat: it is really a bad idea to give a field the same name as its type. use Books for the type.
13:48:33	flyx	SusWombat: also Shelf
13:48:49	SusWombat	flyx, thats just stupid example code but thanks :)
13:48:53	Araq	SusWombat: only osproc poParentStreams comes to mind.
13:49:04	SusWombat	Araq, ok thank you
13:49:12	demi-	Araq: that would work but i have no way to modify the env of the command being invoked then :\
13:49:15	flyx	SusWombat: then do var bookShelf = Shelf(books: Books())
13:49:37	SusWombat	flyx, o.O ok thanks why isnt nim doing that?
13:49:37	Araq	demi-: I think if you say "please, please" cheatfate might give it a shot
13:49:47	flyx	SusWombat: why should it?
13:50:13	demi-	heh, ok i will give that a shot
13:50:14	SusWombat	flyx, cause i declared it as a member or not?
13:50:29	flyx	SusWombat: yes, and default value for ref types is nil
13:50:30	krux02	reactormonk, I now have cask and I can run the cask commands. And I now did some local changes, how do I test them?
13:51:02	krux02	just add the folder of my fork to load-path and add require in my init.el?
13:51:03	flyx	SusWombat: you need to tell Nim explicitly to allocate an object of Books, it won't do that on its own
13:51:14	SusWombat	ok thanks
13:51:40	reactormonk	krux02, in emacs or via cask?
13:51:47	reactormonk	via emacs, just eval-buffer
13:51:50	reactormonk	via cask, make test
13:51:55	flyx	SusWombat: imagine you have a tree structure `type Tree = ref object ; left, right: Tree`. do you see why it's a bad idea now?
13:52:35	*	user0___ joined #nim
13:53:20	krux02	reactormonk, so the test is only non-interactive?
13:53:27	reactormonk	krux02, yeah.
13:54:46	krux02	and with eval-buffer you mean the buffers of the test cases?
13:55:06	SusWombat	flyx, yeah
13:55:21	krux02	reactormonk, I was thinking more of interactive testing
13:56:12	reactormonk	krux02, define interactive testing
13:56:16	krux02	meaning to open a file in nim-mode, and do thinks I would normally do when I edit that file
13:56:43	reactormonk	ah nope, cask is explicitly about isolation.
13:57:03	krux02	well then how do I reaload a package from emacs?
13:57:08	krux02	just restart emacs?
13:57:25	krux02	I mean isolation tests are good
13:57:38	krux02	but more important is how it works in the editor
13:58:35	reactormonk	for reloading it in your editor, eval-buffer
14:00:23	*	Guest63163 quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:00:40	*	Ven joined #nim
14:01:03	*	Ven is now known as Guest67847
14:01:17	*	Guest67847 quit (Client Quit)
14:02:58	*	zaquest quit (Read error: Connection reset by peer)
14:03:11	*	zaquest joined #nim
14:04:48	SusWombat	reactormonk, do you use nim in emacs?
14:04:54	SusWombat	the nimmode i mean
14:04:57	reactormonk	SusWombat, I don't use nim at all anymore :-(
14:05:02	SusWombat	ah ok
14:05:21	reactormonk	mainly Scala, with Haskell on the side.
14:15:11	krux02	reactormonk, well I went in the other direction I used to write a lot of scala and now it's nim
14:15:21	krux02	main reason JVM
14:16:04	krux02	haskell only in university courses
14:16:14	vivus	I've opened a file like this: var fs = newFileStream(filename, fmWrite) . which option do I use to write each entry of my sequence to the file ?
14:17:05	krux02	vivus, the simplest thing you can do, is to write everything you do into a string, and then at the end you just write that string to a file in one operation
14:17:30	krux02	vivus, this file streaming stuff is a very ancient technology when memory was short and files used to not fit in memory
14:18:23	vivus	krux02: that is what I want to do. So I have this operation: for x in fetchdb.fastRows(sql("select content from posts")): ... and now I want to add each entry to a file. what would I do here?
14:18:39	demi-	for str in str_seq: file.write(str)
14:18:53	krux02	demi-, well not entirely
14:19:12	krux02	vivus, you need a separator, otherwise you wont't be able to parse the content back
14:19:27	reactormonk	krux02, actually got a haskell application here in the company, but it's only a smaller project :-)
14:19:28	krux02	the simplets separator is just a newline
14:19:50	vivus	krux02: should I use csv operations then ?
14:20:00	krux02	reactormonk, what is haskell actually be used for except for writing compilers?
14:20:14	reactormonk	krux02, servant makes for nice microservices
14:20:15	krux02	vivus, that would be an option
14:21:45	*	byte512 quit (Quit: Bye.)
14:21:59	krux02	reactormonk, I always liked this pure fuctional style of haskell, but neither the world nor the computer is pure functional. And I prefer procedures over IO monads
14:22:53	reactormonk	krux02, they're not too different. However, you can do funny things with eff monads.
14:23:03	reactormonk	... also, applicative do basically gives you some erlang power.
14:23:06	krux02	what is eff monads
14:23:58	reactormonk	Scala implementation: https://atnos-org.github.io/eff/
14:24:06	*	bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
14:24:13	reactormonk	Monad Transformers are inflexible. That's why eff exists.
14:25:05	Tiberium	how can I define tables which looks like this: {"peer_id": 123, "message": "some text"} ?
14:25:22	*	byte512 joined #nim
14:25:44	krux02	Tiberium, that is not a table
14:25:46	vivus	is this the only place to view CSV docs: https://github.com/achesak/nim-csv/blob/master/documentation.html ?
14:25:57	Tiberium	krux02, {"peer_id": 123, "message": "some text"}.toTable()
14:26:09	Tiberium	I don't know how to use objects like this without toTable
14:26:20	Tiberium	I need something like python's dict
14:26:24	krux02	Tiberium, well it is badly typed or better as a struct/object/tuble
14:26:28	krux02	tuple
14:26:43	krux02	well there is a package called tables
14:26:48	krux02	but they are typed
14:26:53	Tiberium	yeah I see
14:26:58	krux02	so you should have a consistent type on the right side
14:27:30	krux02	I think you can use the any type, but that is very dirty
14:27:31	Tiberium	ok, thankfully there's web API, so it accepts JSON, so I can safely convert all parameters to string
14:27:50	krux02	there is also a json package
14:28:03	Tiberium	I know
14:28:11	Tiberium	But it requires to unpack everything too
14:28:13	krux02	so maybe want to work with json objects instead of tables
14:28:17	Tiberium	like object.str
14:28:28	vivus	hows this lib: https://github.com/unicredit/csvtools . much cleaner usage
14:28:45	vivus	awesome to see a big bank using nim
14:29:03	krux02	well I would always be careful with 19 commits packages.
14:29:38	vivus	https://github.com/achesak/nim-csv 25 commits
14:29:44	krux02	either it's is very much proof of concept, or it is written by the most mazing programmer who can program in a way that covers all corner cases naturally
14:29:53	krux02	it's more likely the first one
14:30:05	krux02	yea you git the idea
14:30:13	krux02	test it get the idea of it
14:30:27	krux02	but eventually it means you have to write your own version of it
14:30:49	*	rauss joined #nim
14:30:53	krux02	I don't want to scare you off. I just want you to be careful
14:31:34	krux02	Tiberium, web API are written with javascript in mind, not with a typed programming language
14:31:41	Tiberium	krux02, I know sadly
14:31:53	krux02	but when you have a specification of the web api
14:31:55	krux02	is some form
14:32:09	Tiberium	I have them
14:32:12	krux02	you should be able to transform that specification into nim code that you can just call as functions
14:32:40	krux02	In is very powerful, it could generate all the glue code for you
14:32:45	Tiberium	they have JSON Schema describing their API
14:32:52	Tiberium	so like all the types
14:33:02	krux02	that's pretty great
14:33:14	krux02	that means you could transform this schema into a wrapper API
14:33:58	krux02	I don't know that API, but web api are better when they are async
14:34:08	krux02	so maybe you have several void procs
14:34:21	Tiberium	their json schema is located on GitHub: https://github.com/VKCOM/vk-api-schema
14:39:16	krux02	Tiberium, I have very little knowledge about web apis, but I know that it is best for performance, when the api allows to bundle requests
14:39:27	Tiberium	krux02, you mean send them as one?
14:39:33	krux02	yes
14:39:47	Tiberium	there's that possibility too, you can send up to 25 requests using special web api method
14:40:10	Tiberium	and api itself is limited for 3 requests per second, so the best you can get is 75 req/s
14:40:15	Tiberium	but it's not really needed
14:40:42	krux02	oh yea what speeds
14:41:00	Tiberium	it's largest social network in Russia and in CIS
14:41:01	krux02	but anyway, you should some request struct
14:41:11	krux02	and then an apicall that accepts an openarray of that
14:41:17	Tiberium	krux02, I will try to parse this json schema
14:41:22	krux02	and then the API itself can be sycronous
14:42:20	krux02	VK isn't that this russian Facebook with the German name?
14:42:32	Tiberium	it's russian facebook, but with russian name :)
14:42:56	krux02	well Kontakte is a German word :P
14:43:02	Tiberium	nooo
14:43:10	Tiberium	"VKontakte" means "In contact"
14:43:20	Tiberium	if you translate from russian to english
14:43:38	Tiberium	it's just english letters instead of russian one's
14:43:43	Tiberium	originally it's "ВКонтакте"
14:44:34	vivus	that would be an awesome domain to own
14:44:36	*	Andris_zbx quit (Ping timeout: 240 seconds)
14:44:36	FromGitter	<brechtm> @krux02 Thanks. Good to know it's possible!
14:45:07	FromGitter	<brechtm> @krux02 But I'm sure you don't mind I'll try to use generics instead ;-)
14:45:18	Tiberium	Where do I start learning, how to generate Nim code? Or maybe I shouldn't do it now, because I'm a newbie?
14:45:30	Tiberium	I've done some auto-binding generation in Python using pycparser and cffi
14:45:39	*	andris_ joined #nim
14:45:42	krux02	brechtm: I don't use inheritance at all, it's all generics in my code
14:46:09	krux02	Tiberium, I started with code generation, too
14:46:18	krux02	the nice thing is, that you can do it all in a macro
14:47:05	krux02	macro parseApi(): untyped = let api = staticRead("schema.json") ...
14:47:13	krux02	don't forget to import macros
14:47:19	krux02	and then you have quote do
14:47:30	krux02	let x = quote do: ...
14:48:38	vivus	Araq: is there a better place for documentation for this: https://github.com/achesak/nim-csv/blob/master/documentation.html ?
14:48:41	krux02	this means you don't need an extra build step thet generates your generates sources, it's all integrated in the default nim build
14:49:34	Tiberium	krux02, yeah, I've read some articles about Nim's compile-time possibilities
14:49:57	krux02	those compile time possibilites are the reason I am here
14:50:10	vivus	Araq: ignore that question, I just realized there is more than 1 CSV lib
14:50:11	krux02	without them I would never have touched this language
14:52:39	krux02	reactormonk, I have a bug with the newest nimsuggest version here: https://github.com/nim-lang/nim-mode/issues/154
14:56:42	*	yglukhov quit (Remote host closed the connection)
14:58:03	Tiberium	krux02, and how can I generate something based on this json schema using macro like you'd suggested? I can't use json module at compile time
14:58:13	Tiberium	or wait
14:58:29	vivus	krux02: in a sequence of strings, is there a way to preserve the \n characters when writing to a file ?
14:58:53	Tiberium	nah, I'm not so familiar with Nim to do such complicated things right now :)
15:01:32	*	BitPuffin|osx joined #nim
15:08:39	Tiberium	but maaaybe
15:13:17	krux02	vivus, I think when you use \n in string literals, nim replaces them by the system dependent version of that
15:13:22	krux02	I am not so sure about that
15:13:30	*	Ven joined #nim
15:13:53	*	Ven is now known as Guest28837
15:13:56	vivus	I'm hoping to preserve them, if possible
15:14:17	Tiberium	so I somehow need to parse json at compile-time, hmm
15:14:31	krux02	well best is, if you don't use newline at all
15:14:36	krux02	json can be one line
15:14:44	krux02	(eg)
15:14:49	vivus	krux02: the data itself is stored with newlines
15:14:56	krux02	or you have newline as your separator of strings
15:15:04	krux02	that also does work
15:15:21	krux02	but newline in the strings from arbitrary sources I am not so sure about that
15:15:35	stisa	Tiberium: I think it should work in a `static:` block
15:15:48	Tiberium	stisa, ah, let me try
15:16:23	krux02	Tiberium, stisa well in a static block you can generate NimNodes, but you can't do anything with them there
15:16:31	Tiberium	so something like static: let api = parseJson(staticRead("objects.json"))
15:16:32	Tiberium	?
15:17:05	krux02	Tiberium, that should work
15:17:23	krux02	but you still have to call a macro somewhere that generates the actual procedures
15:18:16	Tiberium	krux02, compiler says "lib/pure/lexbase.nim(155, 13) Error: VM is not allowed to 'cast' "
15:18:43	krux02	well then you are out of luck :(
15:18:53	Tiberium	krux02, "lib/pure/lexbase.nim(155, 13) Error: VM is not allowed to 'cast'<newline> L.buf = cast[cstring](alloc(bufLen * chrSize))"
15:19:41	krux02	that sucks
15:19:57	*	tankfeeder left #nim ("Leaving")
15:20:40	krux02	what you still can do is write a file that parses the json at runtime and prints an actual source file
15:21:14	Tiberium	hmm, there's https://github.com/cjxgm/jsonob
15:21:23	krux02	and then you can do staticExec("nim c -r generateApi.nim", cache="true")
15:21:36	krux02	that is much more classical code generation
15:22:00	krux02	then you don't have nimnodes, but a textfile you generate
15:22:28	Tiberium	yeah, I know
15:24:19	krux02	(waring very experimental) maybe you can find an easy transformation rule that allows te transform the json with regular expression replace into something the nim compiler does understand
15:25:21	krux02	then you would not need a json parser anymore
15:25:42	Araq	some procs of marshal.nim are available at compiletime
15:25:49	krux02	you could include the transformed json in your nim file and have a representation that is parsed already from the nim compiler
15:25:51	Araq	and marshal uses JSON, so there.
15:26:37	Araq	that said, I would use staticExec
15:27:17	Araq	its caching feature is just unbeatable for quick compiletimes. otherwise you do this complex compile time transformation over and over again
15:28:41	Araq	the VM is for macros, macros are for AST to AST transformations, not for Json to AST transformations
15:30:09	Araq	and your Json to AST tool doesn't have to use string concats to generate Nim code, import compiler/ast, compiler/renderer and let the compiler API do its job
15:32:02	Tiberium	Araq, can I find some docs on those?
15:33:38	Araq	no but the compiler's AST is the same as in macros.nim
15:33:48	Araq	except that nnkX becomes nkX
15:35:01	Araq	for reasons that excape me there is no macro to transform Nim code into compiler/ast construction
15:46:04	Tiberium	I have to wrap my head around macros and all the AST stuff until I can do something like auto-generation...
15:48:54	*	yglukhov joined #nim
15:53:02	*	yglukhov quit (Ping timeout: 246 seconds)
15:56:24	Araq	well you can also just interpret the json at runtime
15:57:35	Tiberium	Araq, thanks for all answers! there's a a newbie question again: how can I do "512 & somenumber" like in c? bitwise and
15:58:14	*	nsf quit (Quit: WeeChat 1.7)
15:58:46	Tiberium	ah
15:58:53	Tiberium	I thought I need to use them in if'
15:58:55	Tiberium	but it doesn't work
15:59:03	Tiberium	only as additional variable
15:59:09	*	Guest28837 quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
16:00:10	Tiberium	but no
16:00:58	Tiberium	so how can I use something like in c: " if (x & 512) {stuff} else {another stuff}"
16:01:44	Tiberium	ah, nevermind
16:03:05	def-	!= 0?
16:03:15	def-	there is no implicit int to bool conversion in nim
16:04:17	*	vlad1777d quit (Ping timeout: 258 seconds)
16:15:21	krux02	Tiberium, when you want to check for bits in integers, maybe the set is the more ideomatic nim way to do the same thing
16:15:39	krux02	at set is basically a bitmasked unsigned integer
16:15:52	Tiberium	krux02, can you tell me more?
16:15:52	*	brson joined #nim
16:15:54	krux02	it's not a hash set
16:16:17	krux02	well very often integers are used to enable flags
16:16:21	Tiberium	yeah
16:16:23	Tiberium	in my case too
16:16:27	krux02	for a nice c api
16:16:47	krux02	glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT)
16:17:19	krux02	so when you put all your flags into an enum
16:17:32	krux02	you can use set[MyEnum] instead
16:17:33	*	vlad1777d joined #nim
16:17:44	*	libman joined #nim
16:17:58	Tiberium	krux02, so I will need to check for every enum vs my integer value?
16:18:45	krux02	Tiberium, well the idea is that you then convert everything to the set, where you mean the set, and you don't work with integers anymore
16:20:41	krux02	and your example " if (x & 512) {... " would become "if enumValue in x: ..."
16:20:58	Tiberium	krux02, for example I have a enum like type Flags = enum unread=1, outbox=2, replied=4, important=8, chat=16, friends=32, spam=64, deleted=128, fixed=256, media=512 . And I've got a int 1+4, so it will set unread and replied flags
16:21:52	Tiberium	krux02, ah
16:21:55	Tiberium	krux02, thanks :)
16:22:06	krux02	well then you should remove the assignments in your enum
16:22:47	krux02	and then you can check what echo(int({unread,replied})) prints
16:23:03	krux02	(should be 5)
16:23:55	Sentreen	Is it possible to return an anonymous proc from within the body of a template? I'm trying to do something like this:
16:23:57	Sentreen	template lazy* (s: untyped): untyped =
16:23:59	Sentreen	proc [T](): T = s
16:24:43	Sentreen	So the template should just wrap the expression s in a function, so I can delay execution.
16:25:36	krux02	sorry in that case you would need a cast[int](...)
16:25:43	krux02	but I just tested it, it works
16:25:55	krux02	http://ix.io/pe8
16:26:26	krux02	Sentreen, yes that is possible
16:27:11	krux02	Tiberium, this C patters has become a language feature in Nim and is called set ;)
16:27:22	Tiberium	krux02, it says Type mismatch: got (int, Flags)
16:27:25	krux02	s/patters/pattern/
16:27:50	krux02	Tiberium, well without code C can't help
16:28:01	Sentreen	krux02: That code gives me an identifier expected error though, on the proc that I am returning. I should have specified that
16:28:44	krux02	Sentreen, well can you create a paste on something that I can execute?
16:29:29	krux02	I mean I can't help, not C
16:30:12	Tiberium	krux02, like this https://glot.io/snippets/eoajwn1c1x
16:32:46	Sentreen	krux02: http://pastebin.com/Y8FXH2fE
16:34:11	krux02	if unread in cast[set[Flags]](3):
16:34:20	Tiberium	oh ok
16:34:37	krux02	Tiberium, but try to do it without cast
16:34:57	krux02	I mean actual code
16:35:08	krux02	for doodling around and testing limits this is fine
16:36:13	vivus	is there an example of using this proc from the 're' lib: proc find(s: string; pattern: Regex; matches: var openArray[string]; start = 0) ?
16:36:51	Tiberium	krux02, but it doesn't work without cast, hm
16:37:16	vivus	or better yet, if somebody could explain to me what I should/can insert into these proc-examples
16:38:33	vivus	find("somestring", "some", matches, 0 ) ??
16:38:53	Araq	Sentreen: just give the proc a name and have that name the return expression of the template
16:39:19	Araq	the proc should be annotated as {.gensym.}
16:40:40	vivus	or do I need to instantiate an openArray before naming it in find() ?
16:42:05	vivus	here is my basic code: https://www.zerobin.net/?706d126f3817d415#wAcDGxijsnhl8n5eBuO5Oc+zeEkItjSZJfGDQwQEjrk= ?
16:44:10	*	Trustable joined #nim
16:48:37	Sentreen	@Araq, like this?
16:48:39	Sentreen	template lazy* (s: untyped): typed =
16:48:41	Sentreen	proc wrapper [T](): T {.gensym.} = s
16:50:32	Sentreen	Actually returning wrapper in that context doesn't seem to work. Using it like this results in a "has no type (or is ambiguous)" when I actually make a lazy value
16:50:53	krux02	well I found out the lazy templates doesn't work, because nim does not parse a proc() ... as a lambda expression
16:51:06	krux02	http://ix.io/peb
16:51:10	krux02	this one does work
16:51:45	krux02	I could also create the lambda expression manually
16:54:25	vivus	wow, its like I'm fighting with the documentation at times
16:54:56	Sentreen	krux02: that seems to work, thanks a lot!
16:55:07	Sentreen	I should dive into the macro documentation to get a better idea as to why now :)
16:55:21	*	andris_ quit (Remote host closed the connection)
16:57:56	krux02	Sentreen, well you need to call echo ast.treeRepr a lot
16:58:21	*	Tiberium quit (Remote host closed the connection)
16:58:22	krux02	and sometimes very similar looking ast things look very different
16:58:34	krux02	I have to admit, macro development is not just write it down
16:58:41	krux02	it is a lot of try end error
16:58:50	krux02	but it is better that in almost any other language
16:59:09	krux02	because not nice it much better than not at all
17:01:10	Sentreen	Okay, that makes it more clear :). The only thing I don't get is why you need to create the `let foo =` part and extract it later?
17:02:20	*	Nobabs27 joined #nim
17:02:29	Sentreen	I agree that the macro system (from what little I've seen) seems to be pretty powerful. Though I kind of get the obsession lispers have with s-expressions after making a (relatively simple) macro
17:04:35	vivus	krux02: how do I access the openarray here: proc match(s: string; pattern: Regex; matches: var openArray[string]; start = 0): bool {..} ?
17:05:10	krux02	Sentreen, the alternative is to create the syntax tree with NimNodes manually
17:05:26	krux02	with nnkLambda.newTree(...)
17:06:17	krux02	vivus, with matches[i] ?!
17:06:19	dom96	What we need is a library/module that builds on top of the macros module.
17:06:39	krux02	dom96, for what?
17:06:49	Sentreen	Oh right, because otherwise the parser gets confused with the proc, I get it now
17:06:51	Sentreen	thanks again!
17:06:52	dom96	To make building ASTs easier
17:07:34	vivus	krux02: so in my match statement I write: var check = match(string1, regex1, matches) ?
17:07:37	krux02	dom96, I think it would be nice to have a getAst that allows to pass in a context
17:08:14	krux02	for example it is not possible to create expr colon expr without the context of a contructor call
17:08:30	krux02	sorry I mean quote
17:11:47	krux02	so that I can say myObjConstr.add quote nnkObjConstr do: `myIdent` : `mySymbol`
17:12:46	krux02	dom96, I don't think that another abstraction/different-view on how to generate syntax trees is going to help anybody
17:21:56	demi-	dom96: if you have a minute, i ran into what seems to be a nasty bug with nimble last night
17:23:49	dom96	demi-: sure, i'm here
17:26:10	dom96	krux02: Maybe you're right. From the top of my head `quote` definitely needs to be improved (maybe it has been already but the last time I used it it had issues).
17:26:50	dom96	It would also be nice to have some sort of aliases for different node kind's children so that AST code isn't just a bunch of indexes.
17:27:24	dom96	Writing node[0][4][i][2] isn't very clear
17:28:23	demi-	so i'm working with this package: https://github.com/samdmarshall/rune and this is what i'm seeing: https://gist.github.com/samdmarshall/778c2fdf7458769d8ac85608ca632481 (notice the diff between line 11 and 47)
17:29:09	*	zachcarter quit (Ping timeout: 240 seconds)
17:31:38	dom96	So it's srcDir = "src" vs. srcDir = "src/"
17:31:48	dom96	i.e. the `/` makes it fail?
17:31:57	*	Tiberium joined #nim
17:32:08	demi-	yeah
17:32:24	demi-	if you do a clear install of a package it fails in the same manner
17:32:57	demi-	the paths that get resolved in that stack-trace are identical except for the trailing slash on the directory name as well
17:33:44	dom96	Report it please. It sucks but it's not too serious. Do you agree?
17:34:04	dom96	(Feel free to link to the IRC logs in your issue)
17:34:18	krux02	dom96, well for me it turned out that I neve use those names, because they are not visible from treeRepr
17:34:33	demi-	ok, i spent an inordinent amount of time trying to figure out what i was doing wrong last night
17:34:54	krux02	dom96, anything that cannot be inferred from treeRepr I tend to just not use at all
17:35:02	krux02	for example all thos constructor procs
17:35:14	dom96	demi-: If Nimble fails with anything other than a NimbleError then that's a bug.
17:35:47	demi-	ok i'll open a new issue
17:35:55	dom96	krux02: Not sure what you mean. What "names" are you referring to?
17:36:03	krux02	I used to write those that were missing for my use case, and then eventually I tended to always use nnkStmtList.newTree(...) because I could do it consistently whit any node kind
17:36:59	Araq	that might be a valuable advise. rather than adding more arbitrary AST constructors we could improve treeRepr's output to be more helpful
17:38:48	dom96	oh
17:38:57	*	dom96 just got what you meant
17:39:49	dom96	If `quote` gets significantly better we won't need constructor procs.
17:40:44	krux02	yes
17:41:10	Araq	what's wrong with 'quote'?
17:41:31	dom96	Maybe nothing today. I haven't used it in too long.
17:41:33	krux02	Araq, it always returns a stmtList
17:41:47	dom96	I can't remember the limitations I faced.
17:42:05	krux02	whenever I want to use it for something else I have to index into the result
17:42:07	Araq	you know you can write your own 'quote', right?
17:42:15	krux02	let tmp = quote do: ...
17:42:25	krux02	tmp[1][2][3] ...
17:42:31	Araq	krux02: gotcha.
17:42:59	dom96	Araq: Let's not encourage hundreds of 'quote' implementations.
17:43:02	Araq	but again, write your own quote and once it's pleasant, contribute it back to macros.nim
17:43:15	Araq	dom96: not my point.
17:43:29	Araq	my point is: it doesn't require compiler support.
17:43:46	dom96	Okay, so ask people to improve it.
17:43:48	Araq	the current quote is only a builtin to make it faster to process
17:44:23	Araq	ofc we never benchmarked it. zahary thought I would be concerned about compile times and so he added it as builtin
17:44:52	krux02	I wrote it earlier, but I thought of something like this: quote nnkObjConstr do: a : `someValue`
17:45:22	krux02	and I think that one does compiler assistence
17:45:36	krux02	because the argument would not be parsable without giving context
17:46:38	Araq	macros.quote doesn't parse anything either
17:46:39	dom96	Why does it need the 'do' keyword btw?
17:47:01	Araq	dom96: because of a grammar deficiency. already on our roadmap, remember?
17:47:43	dom96	I must be missing something, it doesn't take two stmtLists, does it?
17:49:36	*	ludocode quit (Ping timeout: 240 seconds)
17:52:10	krux02	well currently I am very busy with fixing the emacs nim-mode
17:53:12	dom96	I would say that while there is room for improvement it's already good enough.
17:54:59	*	zachcarter joined #nim
17:55:41	dom96	So fixing it is low priority.
17:59:34	*	zachcarter quit (Client Quit)
18:08:20	*	elrood quit (Quit: Leaving)
18:08:46	krux02	reactormonk, what is nim-compile-command-checker-functions for?
18:08:57	krux02	it's undocumented and unused
18:09:05	krux02	and makes my life hard
18:15:00	jivank_	created a simple file server with jester https://github.com/jivank/nim-simplefilemanager/blob/master/simplefilemanager.nim
18:15:08	vivus	Araq: in a proc like this: proc match(s: string; pattern: Regex; matches: var openArray[string]; start = 0): bool {..} ... can you explain to me what this part means: matches: var openArray[string] ... or rather, what should I be adding into the proc for that option ?
18:15:13	jivank_	maybe it will become a manager in the future
18:19:07	Sentreen	I'm trying to figure out why I get a `lib/system/profiler.nim(92, 23) Error: undeclared identifier: 'framePtr'` whenever I try to compile with profiler:on so that I can submit a proper issue report. However, I can't seem to reproduce this issue with a minimal example. Does anybody have any idea what I should look out for if I want to reproduce this (in terms of imports etc)
18:20:47	jivank_	dom96 is there some sort of lazy readFile?
18:20:58	jivank_	not sure if that would help though
18:22:02	dom96	jivank_: that is what async file essentially is
18:23:46	reactormonk	krux02, I have no idea. :-(
18:23:55	reactormonk	delete them and see what happens ^^
18:24:06	reactormonk	(make sure to restart emacs / unset the function)
18:24:27	vivus	where is the 're' library for nim hosted? I find nothing on github
18:25:01	*	nsf joined #nim
18:25:26	dom96	vivus: in the stdlib
18:27:36	*	brson quit (Ping timeout: 240 seconds)
18:28:42	Sentreen	Okay, so for some reason --profiler:on does not work when there is a .nims file in the same directory as the .nim file you are compiling. There is already an open issue though
18:36:15	vivus	dom96: maybe you might be able to answer here: in a proc like this: proc match(s: string; pattern: Regex; matches: var openArray[string]; start = 0): bool {..} ... can you explain to me what this part means: matches: var openArray[string] ... or rather, what should I be adding into the proc for that option ?
18:37:27	dom96	var matches: array[MaxSubpatterns, string]
18:37:34	dom96	IIRC
18:38:03	vivus	dom96: so I need to explicitly create an openArray ?
18:38:11	dom96	We have the whole regex spec in the documentation for 're' but no actual examples of the module...
18:38:27	dom96	vivus: yeah.
18:38:36	krux02	reactormonk, that is what I am doing
18:38:45	vivus	once I can figure it out, I will gladly add examples to all the libs I am using
18:39:01	dom96	vivus: that would be brilliant :)
18:39:02	krux02	it seems to be complicating things unnecessarily
18:39:39	*	dmi0 joined #nim
18:40:18	krux02	how does the project structure of nim projects really work
18:40:39	krux02	I mean when I have a project, how do I detect it, and how do I compile it?
18:40:43	dom96	vivus: If you do please make the examples short. Similar in style to the ones in: https://nim-lang.org/docs/httpclient.html
18:41:14	vivus	dom96: within a proc, what is the difference of saying proc foo(s : string, box: var openArray[string]) ? why does box get "var" ?
18:41:28	dom96	There is far too many "Large example"-style examples. For example here: https://nim-lang.org/docs/db_mysql.html#examples (there are also some nice short examples here too though).
18:41:37	krux02	nim-mode has currently logic to detect the project/project.nim but then still uses the current buffer-file-name as argument for nim build
18:42:01	Tiberium	waaait, simple social network bot which connects to long polling endpoint and pulls events, and sends message back to user (if user've sent a message) - 80kb stripped and upx'd
18:42:02	*	Matthias247 joined #nim
18:42:03	Tiberium	just WOW
18:42:28	dom96	vivus: The proc 'foo' is able to modify the `box` parameter
18:42:53	dom96	Tiberium: Nice. But why are you paying so much attention to the size?
18:43:00	vivus	dom96: in principle, I prefer examples over the specs provided below said-examples. to 'see' actual code helps a lot more than reading specs, IMHO
18:43:13	dom96	vivus: I agree.
18:43:15	Tiberium	dom96, because you can't simply distribute a program in python for example
18:43:19	Tiberium	it will be huge
18:44:08	vivus	Tiberium: isn't python universal on most OSes? you could distribute just the script
18:44:17	dom96	vivus: Long examples are also not good because users are almost always looking for how to do a specific task.
18:44:20	Tiberium	vivus, except for Windows :)
18:44:23	dom96	Scanning through a long piece of code is not easy
18:44:30	Tiberium	vivus, and there's also need to install packages
18:44:36	Tiberium	and users can have different python versions etc
18:45:26	vivus	dom96: what I think might work for the docs, (which is why I asked for where `re` is stored), is that the /tests/ for nim-code could be used as short examples. unfortunately, tre.nim had minimal tests for the procs
18:46:20	dom96	vivus: Some of them could be, yes.
18:46:59	dom96	Tiberium: Ahh. :)
18:47:15	vivus	I learnt that nifty trick when I was using a small lib written by 1 dude. not much example code, but quite a few tests showing his procs/functions at work
18:47:41	Tiberium	dom96, I've started with social network bot in Nim because I've done it in Python already
18:47:50	dom96	I never think about the size of dependencies, more about their tendency to fail in incredibly annoying ways.
18:48:15	Tiberium	https://github.com/TiberiumPY/LOLBot (sorry for russian language and off-topic)
18:52:13	dom96	Tiberium: Cool. Are you rewriting this in Nim? :)
18:52:22	Tiberium	dom96, I want to :)
18:52:31	Tiberium	so I will learn some Nim in process
18:52:52	Tiberium	Basically I already have working kernel of the bot
18:52:58	Tiberium	But I'm not happy with the code
18:53:46	vivus	dom96: I think I mentioned it earlier, but a lot of Python coders see (or would like to see) Nim as the language they can move to, instead of Go
18:53:54	dom96	Tiberium: You'll be using async right? Let me know if you need any help
18:54:39	Tiberium	dom96, firstly I'll try to make simple working synchronous implementation.
18:54:56	Tiberium	vivus, personally I don't like error handling in Go
18:55:05	Tiberium	it seems very...
18:55:12	dom96	vivus: I know I want that. But do a lot of Python coders really also would like to move to Nim?
18:55:54	dom96	Tiberium: Sounds like a good idea. You should be able to use the multisync macro to add an asynchronous implementation easily.
18:56:20	Tiberium	dom96, I'll use only httpclient with json I think, so yes
18:56:43	vivus	dom96: its not something that should be optimized for, but if the syntax/features make it appealing for pythonistas to use nim, that wouldn't hurt the progression (instead of python programmers optimizing around cython, pypy, etc.)
18:58:40	Tiberium	by the way, is 2 spaces or 4 spaces preferred in Nim?
18:58:50	vivus	2?
18:58:52	Vladar	2
18:59:10	Tiberium	ah, ok
19:01:07	dom96	Regarding Go error handling. I'm the same, I like exceptions.
19:01:18	*	hendi joined #nim
19:01:50	euantor	I don't mind Go's error handling, though it is repetitive
19:02:30	euantor	I like that you have to explicitly ignore an error, rather than simply missing it by not catching an exception that might get thrown
19:02:44	dom96	I prefer Nim's opt-in checked exceptions.
19:02:52	federico3	any lib to generate a gifs?
19:02:53	euantor	Though I also like the approach of using checked exceptions like java and Nim
19:04:35	*	athaudia quit (Ping timeout: 255 seconds)
19:04:53	euantor	Unchecked exceptions are the thing I dislike most about C#
19:05:43	*	athie joined #nim
19:06:22	dom96	federico3: not that I know of, what are you working on?
19:06:29	*	dmi0 quit (Ping timeout: 260 seconds)
19:06:53	*	brson joined #nim
19:08:33	Araq	euantor: huh? it's easy to call function() in go and ignore the return value(s). right?
19:08:46	Araq	I never heard of it having Nim's explicit "discard"
19:08:49	euantor	Yes, but if you want a return value, you have to ignore the error
19:08:55	federico3	dom96: a couple of useful and some toy projects
19:09:06	euantor	So if a function returns (x, error), you have to discard `error` (using `_`)
19:09:22	euantor	Which is the only good thing about their error model
19:09:35	federico3	generating a little gif animation
19:10:49	Araq	so it's easier to ignore errors than exceptions, got it :P
19:11:24	Tiberium	is it fine to have two http clients in two modules of one app, or I can reuse them?
19:11:49	euantor	It's easy to ignore unchecked exceptions at the low level where they should be handled and to let them propagate to a too high level to handle them properly
19:12:25	euantor	EG: ignore an exception in a library used by some other library and don't catch it until the main app, by which point you can't take proper corrective action in some cases
19:12:30	dom96	Tiberium: It's fine, but you can reuse them.
19:13:29	*	michael_campbell quit (Quit: ZNC 1.6.1 - http://znc.in)
19:16:19	*	tankfeeder joined #nim
19:16:25	*	tankfeeder left #nim (#nim)
19:16:26	Araq	euantor: never seen this happening.
19:16:46	euantor	Lucky you ;)
19:16:59	Araq	it's mostly a semiconscious process of getting shit done more quickly :P
19:17:21	euantor	I've seen it all too many times in poorly programmed C# libraries
19:17:45	Araq	and Java's model is just "ok, I'm logging this exception here"
19:17:54	euantor	Usually by people who came from writing old VB6 code, then went to vb.net, then to C#
19:17:58	Araq	so everything in Java depends on a logging framework
19:18:09	euantor	Which our company has had a couple of
19:18:30	Araq	I think logging is a disease.
19:19:09	euantor	I think it has its place
19:19:29	euantor	But again, I think it depends on what kind of software you work on most of the time
19:19:54	euantor	Write what works best for your scenario, and hope that it works
19:20:25	Araq	it's over used and doesn't come with any kind of policy. so, ok your log has 4000 errors. when will somebody look at it? what does it mean? is it 4000x the same error? is it a bug or a real error?
19:21:09	Araq	if it's important data, why is it in a textfile? shouldn't it be structured data that was put into a database instead?
19:21:26	euantor	What if you're logging that your DB's broken?
19:21:36	euantor	Good luck putting that in the DB :)
19:21:46	Araq	then you fix your DB.
19:21:47	euantor	Again, do what works best for your scenario
19:22:02	Araq	what if the logging IO operation fails? how do you log that?
19:22:04	euantor	And if it's on a remote machine you have no physical access to in morocco?
19:22:09	Araq	same problem, logging doesn't solve it.
19:22:20	Araq	you use a DB that's local.
19:22:26	euantor	No, but at that point your pretty much screwed
19:22:44	euantor	The DB is local t the machine that's deployed, it's an embedded system that cannot touch the internet
19:23:05	euantor	Anyway, I would say we're going off topic a bit and I don't want to argue
19:23:32	vivus	so I tried this: var check = match(string2, regex2, matches: var openArray[string]) and got this error: regex.nim(9, 38) Error: invalid expression: 'matches: var openArray[string]'
19:24:25	*	sz0 quit (Quit: Connection closed for inactivity)
19:26:36	Araq	vivus: var matches: array[10, string]; let check = match(s, r, matches)
19:27:48	vivus	Araq: so you have to define an array of a fixed size ?
19:28:05	Araq	you can also use a seq, but why bother
19:28:24	Araq	you should know your pattern and the number of captures in it
19:28:36	vivus	Araq: when I run the regex to match, I'm not sure how many I will get (consider the data hostile)
19:28:46	Araq	that's not how it works.
19:30:14	Araq	r(\w+)=(\w+).* matches a=b;c=d but matches only captures a and b
19:30:34	Araq	you get the c=d part the next itereation
19:30:59	Araq	with the starts int parameter or whatever it's called.
19:31:05	vivus	Araq: so as a rule of thumb, should the array size be the size of the database rows?
19:31:31	Araq	as a rule of thumb you shouldn't use regexes.
19:35:52	*	ludocode_ joined #nim
19:46:34	krux02	args, I made tea for myself, but someone before me abused the container for coffee, now the tea smells horribly like coffee
19:46:38	Tiberium	this is that I've written for my bot: https://gist.github.com/TiberiumN/96b252808798af7656d05f2412c0c489
19:46:49	Tiberium	I'm still not happy with that JSON stuff :(
19:47:17	Tiberium	(and, be aware, I started learning nim only ~3 days ago)
19:47:38	krux02	Tiberium, well yea but you have experience programming
19:48:12	Tiberium	krux02, but not in compiled languages :)
19:48:19	krux02	for me learning nim was mostly learning syntax, and a few new concepts here and there
19:48:48	Tiberium	Actually Nim feels like dynamic language sometimes - you don't need to put types everywhere
19:48:52	Tiberium	and compiler is very smart
19:48:52	krux02	Tiberium, well compilation is only a mental thing
19:49:09	krux02	the boundary between compiled languages and interpreted ones in blurry
19:49:15	krux02	in the middle in julia
19:50:05	Tiberium	I have a question - how I need to format long definitions of object instances (like MessageUpdate in mainLoop() in main.nim)
19:50:18	Tiberium	The way I've formatted it, or there's another styling preferred for Nim?
19:50:18	krux02	and python also has the preference to poop pyc files everywhere but still likes to call itself interpreted :P
19:51:18	dom96	MessageUpdate(
19:51:20	dom96	foo: 42
19:51:21	dom96	)
19:51:40	Tiberium	dom96, ah, ok
19:51:40	dom96	Also, you should put all your types under one type section
19:51:43	vivus	well my code compiles and runs, but my regex is not loved :P
19:51:50	Tiberium	dom96, thanks
19:52:24	dom96	Tiberium: Why aren't you happy with the JSON?
19:52:28	krux02	Tiberium, the types in one type section has the advangage that within one type section the types can reference each other
19:52:39	Tiberium	dom96, I thought it's a bad code
19:52:46	krux02	with different types sections you can only reference previous types
19:52:59	Tiberium	but yeah, I know that processing JSON in compiled languages is hard
19:53:17	dom96	Does this work? event.extract(_, msgId, flags, peerId, ts, subject, text, attaches)
19:53:27	krux02	this is horrible this tea just smells like coffee, think I have to throw it out, I can't enjoy it :(
19:53:47	dom96	Didn't realise `_` could be used like that. Did I accidentally implement this or did somebody else explicitly implement support for this feature?
19:53:54	Tiberium	dom96, it works for me via unpack package?
19:54:13	Tiberium	I've also used this underscore in Python
19:54:17	demi-	i've used that in nim a bunch
19:54:29	demi-	as a "discard this element because i don't want it"
19:54:37	Tiberium	in python it may be _, _, useful_data, _ = some_tuple
19:54:49	dom96	unpack package?
19:54:51	Tiberium	yeah
19:54:54	dom96	got a URL?
19:54:56	Tiberium	but wait
19:55:26	dom96	Yeah, I am aware that it works for tuple unpacking
19:55:49	dom96	But this is as a proc param
19:55:52	*	Trustable quit (Remote host closed the connection)
19:56:00	demi-	O_o
19:56:06	dom96	or maybe, extract is a template?
19:56:18	Tiberium	dom96, oh wait
19:56:21	vivus	Araq: did I get the formatting wrong: var reg2 = re"<a href=\\.*? style=\\.*?\\>.*?<\/a>" ?
19:56:25	Tiberium	It's a macro that I've copied
19:56:33	Tiberium	Second answer here - http://stackoverflow.com/questions/31948131/unpack-multiple-variables-from-sequence
19:57:06	dom96	vivus: Tony the Pony is coming.
19:57:27	dom96	Why are you parsing HTML using regex?
19:57:35	dom96	Use htmlparser
19:57:40	Tiberium	And in what order I should define variables? constants first, then let's, and after that typical vars?
19:58:05	vivus	dom96: because this is what I am looking for in the DB :) specific patterns of text
19:58:32	Tiberium	by the way: it would be a nice small addition to have automatic url encoding in httpclient
19:58:55	Tiberium	and possibility to pass query parameters :)
19:58:58	dom96	vivus: hrm, alright. Does that snippet not compile?
19:59:24	*	rauss quit (Ping timeout: 260 seconds)
19:59:28	vivus	dom96: it compiles and executes. I can show you the full code if you are interested?
19:59:35	dom96	Tiberium: That would be a nice addition to the uri module (if it's not already in there). PRs encouraged :)
19:59:49	dom96	vivus: I guess it doesn't give you the result you expect?
20:00:05	dom96	What pattern should that be? The `\\` look suspicious to me
20:00:22	vivus	dom96: it shows me the else part of my if/else (which is fuark :) )
20:00:48	dom96	Tiberium: The order only matters in the global scope in my opinion.
20:01:08	*	rauss joined #nim
20:01:14	vivus	dom96: <a href=\"http://somesite.net/\" style=\"text-decoration:none;color:#676c6c\">Link</a>
20:01:18	dom96	Tiberium: And for that case the order you mentioned sounds alright
20:02:02	dom96	vivus: I'm a bit lost now :)
20:02:16	vivus	dom96: do you use sublime text?
20:02:26	dom96	nope
20:02:34	dom96	VS Code mostly nowadays
20:02:41	vivus	do you use any text editor that does regex checking?
20:02:55	dom96	Never seen that in a text editor
20:03:26	vivus	I am 99% certain my pattern matches. verified it as well
20:06:08	*	zachcarter joined #nim
20:06:09	dom96	vivus: are you using the find or the match proc?
20:06:15	vivus	match
20:06:24	Tiberium	And what's preferred max line length in Nim?
20:07:43	demi-	i think that is personal preference
20:08:03	demi-	but there is very few things that yield long lines in nim
20:08:07	dom96	vivus: try 'find'
20:08:41	dom96	Max len is 80 for the stdlib/compiler
20:10:18	vivus	lots of code mods dom96 :P now the mind is tired :P
20:11:06	dom96	Araq: Any chance you could give https://github.com/nim-lang/Nim/issues/5589 a look?
20:20:08	Araq	I replied.
20:21:06	Araq	I don't look into bug reports that make me write my own programs with my own test data which may result eventually in the same reported wrong behaviour.
20:21:21	krux02	./koch tools -> Error: cannot open 'tools/nimsuggest/nimsuggest.nim'
20:21:30	Araq	krux02: nim c koch
20:22:03	krux02	ok thanks it works now
20:24:33	dom96	Araq: What do you mean? The bug report includes source code.
20:27:30	*	hcorion joined #nim
20:29:21	vivus	Araq dom96 here is the type of situation I am dealing with: https://www.zerobin.net/?593b3c78b10da57a#UOjWAxOudKUEZzyENzqXWdc/GpuJirr2sW+lRRov+e8= . fully working example
20:29:48	Araq	dom96: sorry, I really missed that then :-)
20:30:20	dom96	vivus: Did you try 'find'?
20:30:31	vivus	dom96: yep, same problem
20:30:46	dom96	did you try getting rid of the '\\'?
20:31:40	krux02	well nim-mode shows this: Error while checking syntax automatically: (error "Server may raise an error. Use \"M-x epc:pop-to-last-server-process-buffer RET\" to see full traceback:
20:32:04	krux02	and this process buffer shows this: Process epc:server:5 finished
20:32:12	krux02	the server just dies
20:33:34	vivus	dom96: yep, that too. my mind is doing a SEGFAULT now :P
20:36:24	vivus	dom96: the non-matching doesn't make sense to me because nim supports PCRE
20:37:36	*	rauss quit (Quit: WeeChat 1.7)
20:38:24	dom96	vivus: this seems to work: <a\ href=.*\ style=.*>.*<\/a>
20:38:30	dom96	The spaces need to be escaped, weird.
20:38:34	dom96	Maybe there is a re option for that
20:38:45	vivus	heh
20:38:52	dom96	(The best way to figure out what's going on is to reduce the regex pattern until it does work)
20:40:23	vivus	dom96: can you copy/paste your code in zerobin please?
20:41:14	dom96	I hope gist is fine: https://gist.github.com/dom96/091bc3e8526844e7f6bdcc4bb1aa3083
20:42:30	krux02	well the best way to test regular expressions, is with eithe vim the / (search) or in emacs with re-builder
20:42:58	krux02	not default vim I guess
20:43:08	krux02	maybe neovim highlights all matches
20:43:30	vivus	krux02: sublime supports it also
20:44:16	krux02	all you need is to have highlighting of all matches and a fill with all cases you want to cover
20:44:40	krux02	then you can realtime edit the regular expression and see how the highlight changes
20:44:49	dom96	That doesn't help in this case
20:45:00	dom96	Unless vim/whatever follows the same rules that 're' does
20:49:44	krux02	what does epc stand for?
20:50:26	vivus	no idea what is going on with re now.
20:51:54	vivus	trimmed it down to: var reg2 = re"<a.*>.*<\/a>"
20:54:05	krux02	I really feel a bit stinding in the rain, when I want to fix problem with nim mode
20:54:47	dom96	PMunch: I submitted your article (I think it's yours at least) to HN, hope you don't mind :)
20:58:45	Tiberium	Is it fine to have one types.nim file?
20:58:54	Tiberium	so all types are stored there
21:00:17	Tiberium	it doesn't work for me for some reason...
21:00:45	krux02	Tiberium, it should be
21:00:46	dom96	Sure it's fine. But do whatever works.
21:00:52	vivus	how are you importing?
21:00:58	Tiberium	vivus, at the top of the file
21:01:01	Tiberium	import types
21:01:06	Tiberium	should I include instead?
21:01:40	Tiberium	I have 3 files, main.nim, types.nim, api.nim main.nim imports types.nim and api.nim, api.nim imports types.nim
21:02:11	dom96	So a basic version of my little snake game is working: http://picheta.me/snake/
21:02:16	dom96	Let me know your thoughts :)
21:03:09	Tiberium	dom96, very nice game
21:03:17	Tiberium	I played up to 10 score
21:03:21	Tiberium	ah
21:03:23	Tiberium	it's endless
21:03:48	dom96	Yeah, maybe i'll add other levels in the future
21:04:30	dom96	Glad you think it's nice :D
21:04:37	Tiberium	so for some reason compiler can detect that I've imported type, but can't detect it's field
21:04:38	Tiberium	hm
21:04:41	krux02	it.s a bit easy
21:04:52	dom96	Tiberium: Did you export the field?
21:05:03	Tiberium	dom96, I didn't export, I just exported objects?
21:05:12	dom96	Tiberium: You need export each field as well
21:05:27	dom96	krux02: Snake typically is in the early game
21:05:41	dom96	krux02: But i'm open to ideas on how to make it more challenging earlier
21:05:50	vivus	nice game
21:05:52	Tiberium	dom96, so place a star for every field?
21:05:54	krux02	yes it's cool
21:05:58	krux02	it easy and fun
21:06:05	dom96	Tiberium: yeah
21:06:12	dom96	Tiberium: field*: string
21:06:16	Tiberium	I understand
21:06:25	vivus	the most clear example for snake would be for making the snake go faster after each bob it eats
21:06:34	Tiberium	dom96, do I need to do the same for enums?
21:06:40	Tiberium	thanks god no
21:06:42	dom96	vivus: Yep, that's what happens.
21:06:55	vivus	doesn't feel like it got faster
21:07:06	dom96	vivus: Maybe I made it a bit too subtle
21:07:15	dom96	But it definitely gets faster
21:07:24	krux02	snake always reminded me of the game I played when I was very young: http://s.uvlist.net/l/y2012/01/92323.jpg
21:07:38	dom96	It moves every (200ms - score)
21:08:11	krux02	it's basically snake with nondeadly walls keys doors traps and different things to collect
21:08:25	dom96	krux02: cool, never seen it before.
21:08:39	krux02	well it is also hard to google for it
21:08:51	krux02	it's called VEGI and exists only in German
21:09:04	krux02	but in that game language doesn't really matter
21:12:08	hcorion	@dom96, small bug, but I'm sure your aware, sometimes the fruit will spawn on the snake, so you can't see it
21:12:18	krux02	it's basically a lost game
21:12:22	hcorion	Score of 50 btw
21:12:36	dom96	Yeah, will fix that :)
21:13:33	krux02	reactormonk, when you are still online, you can merge my pull request in nim-mode
21:13:50	krux02	or check if you are ok with the changes
21:14:02	dom96	I can't remember my high score but I definitely got to at least 80.
21:14:34	dom96	Source code here in case you want to see how it's implemented: https://github.com/dom96/snake
21:15:48	*	_apg joined #nim
21:15:50	Xe	Why does the {.procvar.} pragma exist?
21:15:55	Xe	procvar pragma
21:15:56	Xe	The procvar pragma is used to mark a proc that it can be passed to a procedural variable.
21:18:15	dom96	Xe: https://irclogs.nim-lang.org/15-09-2016.html#18:02:52
21:18:50	dom96	And related issue https://github.com/nim-lang/Nim/issues/2172
21:19:09	dom96	There was a more recent discussion about this as well but I can't find it
21:20:04	*	hendi quit (Quit: Connection closed for inactivity)
21:22:33	Tiberium	how can I catch SIGINT and SIGTERM?
21:22:40	Tiberium	to end my app gracefully
21:22:52	dom96	A nice project might be using markov chains on the IRC logs to answer Nim questions
21:23:21	dom96	Tiberium: https://nim-lang.org/docs/system.html#setControlCHook,
21:23:43	krux02	dom96, that is how I feel when I enter the #emacs channel
21:24:02	krux02	they do it
21:24:16	dom96	they use markov chains?
21:24:44	krux02	I don't know if they use markov chains, but just do /join #emacs and ask a what is question
21:26:18	Tiberium	haha
21:27:13	dom96	I tried :P
21:27:48	*	kier quit (Remote host closed the connection)
21:28:03	hcorion	Very cool
21:28:29	dom96	hcorion: https://github.com/nim-lang/nimbot
21:28:32	dom96	Help welcome :)
21:29:02	*	kier joined #nim
21:29:07	vivus	dom96: if you can create a tar.gz dump of the logs until end-2016, someone might be able to work on that
21:30:02	vivus	the problem with answering a question on IRC is "where does the answer end?"
21:30:13	Tiberium	wait, does nim ssl module depends on system openssl library?
21:30:18	Tiberium	or it's embedded into binary?
21:30:33	dom96	vivus: They're available online in JSON-ish format: https://irclogs.nim-lang.org/04-02-2017.logs
21:30:50	dom96	Tiberium: System openSSL unfortunately
21:31:01	Tiberium	dom96, ah, dynamic loading
21:31:26	krux02	NimBot, how are you?
21:31:32	krux02	hmm
21:31:32	Tiberium	NimBot, hi
21:31:49	krux02	NimBot doesn't seem to respond very well
21:31:52	vivus	NimBot: sudo rm -rf /
21:32:07	dom96	!ping
21:32:07	NimBot	pong
21:32:10	dom96	!lag
21:32:10	NimBot	1ms between me and the server.
21:32:31	vivus	!nimble
21:32:43	krux02	!beer
21:32:46	dom96	That's about all the commands it knows really
21:32:57	dom96	Adding a dictionary would be easy (hint hint)
21:33:11	vivus	!ping
21:33:11	NimBot	pong
21:33:31	krux02	well I know that the emacs bot get's updates from the channel
21:33:33	hcorion	Sure is good at ping pong
21:33:43	krux02	when the bot doesn't know something they can tell him
21:34:13	dom96	Yeah. It's a simple "dictionary" feature
21:34:31	dom96	You could implement a !set command for NimBot
21:34:37	dom96	!set nimble Nimble is awesome
21:34:41	dom96	!nimble
21:34:44	dom96	>Nimble is awesome
21:34:55	hcorion	But only allow certain people to set it
21:35:26	dom96	of course
21:35:44	vivus	am I allowed to write the IRC bot in python for the nim channel?
21:35:52	Xe	no
21:35:59	Xe	that goes against the point of it
21:36:05	dom96	Why would you?
21:36:21	hcorion	There is already a nim bot written in Nim
21:36:23	vivus	shipping to market
21:36:32	vivus	MVP :P
21:36:35	dom96	Extending NimBot should be trivial
21:36:55	krux02	the bot should detect what is questions
21:36:57	hcorion	Yeah, nice clean code
21:37:08	hcorion	What is love?
21:37:18	vivus	there is an easy way to achieve what dom96 said without using intense mathematics/deep-learning
21:37:22	krux02	how do I compile nimsuggest in debug mode (with stack traces)?
21:37:48	dom96	nim c nimsuggest isn't possible?
21:37:49	vivus	the bot just needs to detect when a question is asked
21:38:09	krux02	I build nimsuggest with koch
21:38:40	dom96	And it needs to reply to hcorion's question with "Baby don't hurt me"
21:38:58	dom96	That feature is a must :)
21:39:33	dom96	But honestly, writing an IRC bot in Nim should be a breeze
21:39:34	hcorion	With emojis!  Baby don't hurt me 
21:39:40	Tiberium	Is there any example on plugin system for nim? I'd like to set-up something like handlers for an event, and then activate this event from main loop, so plugins will process it?
21:39:46	dom96	hcorion: OOOH YES :D
21:40:05	vivus	until your bot needs to actually detect when this "?" is used
21:40:18	dom96	Tiberium: You mean you want to load pre-compiled plugins at runtime?
21:40:29	krux02	well my font doesn't have those emojis
21:40:32	Tiberium	dom96, no, no, I will compile them all at the same time
21:40:40	vivus	nor mine krux02
21:40:45	Xe	Tiberium: why not compile them into the app then?
21:40:50	hcorion	You should get better fonts
21:40:56	Tiberium	Xe, I'm asking how to do this better
21:40:57	krux02	but I can do this: ✔✘†€‣♠♥♦⊂∫∀∈∃∀∈∃Σ√ΛℂΩ×ιαοωχεακψλχοηπβσj{`}/{([]^…_›[₁₂₃♀♂♀⟨⚥⟨₀⟩∅∝∞∥∡→∨∧⊥∡∥∨∧⊥∡∧⊥∡¹²³›³›¢›‹¢‚‘
21:41:16	Xe	Tiberium: why do you think that loading .so files is better than loading the final binary?
21:41:30	Tiberium	Xe, nooo
21:41:31	vivus	Xe: how is your regex?
21:41:43	krux02	aren't so files just binaries?
21:41:45	Xe	vivus: (not!good)
21:41:45	Tiberium	Xe, I mean, I will compile all plugins with the main app, so they'll be in one binary
21:41:53	vivus	was worth a try :P
21:41:54	Xe	oh i see waht you mean
21:42:16	dom96	Tiberium: Well... I'm not sure there is any examples of that.
21:42:36	dom96	If you show how you would implement it in Python I might be able to help
21:42:59	jivank_	by default --gc:markAndSweep is used?
21:43:15	Xe	Tiberium: https://github.com/Heartmender/sobot/blob/master/modules/hello.nim https://github.com/Heartmender/sobot/blob/master/src/command.nim friend of mine got this far, but ended up ditching this idea
21:43:58	dom96	jivank_: nope
21:44:10	Xe	Tiberium: you want something like this though right? https://gist.github.com/Xe/382817a00026f5eb68554c482c946e83
21:44:29	Tiberium	Xe, yeah
21:44:49	Tiberium	Xe, so my plugins will subscribe to commands or some other events
21:45:34	Xe	Tiberium: you need some kind of "default" global bot command demux or something
21:45:43	Xe	kinda like how Go's net/http.DefaultServeMux works
21:49:07	dom96	jivank_: I asked Araq earlier to look into your issue and I think he will once he gets a chance.
21:49:39	dom96	oh would you look at that. My HN submission is on the front page :D
21:50:26	hcorion	Which one?
21:51:03	dom96	Check the front page and you'll see
21:51:09	hcorion	http://serv.peterme.net/devlog/cross-platform-guis-and-nim-macros.html
21:59:04	vivus	I'm calling it now, someone in an HN comment will say "wow, what a cool language, never heard of nim before"
21:59:12	hcorion	Issue for automated responses for nimbot: https://github.com/nim-lang/nimbot/issues/13'
22:00:56	dom96	hcorion: Nice! Thanks :)
22:01:24	dom96	vivus: hah yes.
22:02:19	jivank_	dom96 thanks! , i tested your devel branch but the results seemed similar to the ones i just posted to the issue, is that the asyncfile gc bug you talked about yesterday?
22:02:34	dom96	Somebody submitted John Novak's blog post as well and it's also on the front page.
22:03:02	dom96	jivank_: Yeah. I'm pretty sure both are affected by the same bug.
22:03:14	dom96	Unused memory simply isn't freed.
22:03:25	krux02	hcorion, I just read the article, that is very interesing and I do like it
22:03:59	krux02	except that this site requires javascript so that I can read the blog
22:04:19	krux02	should not be a requirement
22:04:35	dom96	John Novak's blog post is nice too http://blog.johnnovak.net/2016/05/29/cross-platform-gui-trainwreck-2016-edition/
22:04:38	dom96	(Another Nim user)
22:04:45	dom96	Also, his blog looks awesome.
22:04:56	Tiberium	is there a list on what packages from stdlib are support on JS target (not emscripten)?
22:05:21	Tiberium	*are supported
22:05:30	dom96	Tiberium: Not as far as I know.
22:05:36	Tiberium	dom96, for example httpclient
22:05:59	dom96	I can probably name the modules which /do/ support JS
22:06:08	dom96	There isn't many
22:06:22	dom96	hrm, well no, maybe I can't
22:06:34	dom96	I was surprised a while back that deques worked fine
22:06:42	dom96	But httpclient doesn't support it
22:06:45	dom96	definitely
22:07:13	hcorion	Does it work with emscripten?
22:07:24	*	vlad1777d quit (Remote host closed the connection)
22:07:27	Tiberium	I think httpclient isn't that hard to port to JS
22:07:38	dom96	hcorion: no idea
22:07:59	dom96	Tiberium: I don't think it can be fully supported (in the web browser)
22:08:06	Tiberium	dom96, just simple get
22:09:12	dom96	That might be easy to get working.
22:09:29	*	brson quit (Ping timeout: 260 seconds)
22:09:33	dom96	I'll accept a PR implementing it
22:11:40	*	nsf quit (Quit: WeeChat 1.7)
22:13:22	*	brson joined #nim
22:32:33	*	bjz joined #nim
22:34:57	*	hcorion quit (Quit: Konversation terminated!)
22:37:13	*	Vladar quit (Quit: Leaving)
22:38:55	*	gokr joined #nim
22:39:30	*	Jesin quit (Quit: Leaving)
22:41:04	Tiberium	is memory that massif reports - real memory usage of my app?
22:43:37	*	vivus quit (Quit: Leaving)
22:44:15	Tiberium	it says my app (bot) used ONLY 20kb's of heap
22:44:21	Tiberium	(it's maximum value)
22:44:52	*	Jesin joined #nim
22:57:59	*	shashlick quit (Ping timeout: 256 seconds)
22:59:51	*	shashlick joined #nim
23:01:03	*	libman quit (Quit: Connection closed for inactivity)
23:09:12	*	adeohluwa joined #nim
23:16:43	*	vendethiel joined #nim
23:18:12	*	_apg left #nim ("rcirc on GNU Emacs 24.5.1")
23:25:03	krux02	dom96, maybe you can help me
23:25:17	krux02	what is the best pattern to detect a project root?
23:25:39	krux02	my projects are all customized with whatever works
23:25:51	krux02	but I don't know how to properly detect a project root
23:26:10	krux02	I need a project root specification
23:26:42	dom96	what do you mean by "project root"
23:26:44	dom96	?
23:27:27	demi-	this is about nimsuggest isn't it?
23:27:28	krux02	well I mean the important argument to pass to the nimsuggest process
23:27:33	krux02	demi-, yes
23:27:41	krux02	I try to fix nimsuggest for emacs
23:28:08	krux02	the current state of nim-mode is more like no maintenance anymore
23:28:30	demi-	it doesn't seem that difficult at all
23:28:38	krux02	well then you can answer me
23:28:53	krux02	at seems there are some weird patterns
23:28:58	krux02	project/project.nim
23:29:01	demi-	there is a directory that has a nim.cfg or .nimble file that is the top-most directory
23:29:10	krux02	project.nim + project.nim.cfg
23:29:16	krux02	project.nimble
23:29:43	krux02	well there can also be the project.nim.cfg
23:29:57	*	adeohluwa quit (Remote host closed the connection)
23:30:06	demi-	yeah there are 3 patterns, what is so complicated about this?
23:30:20	krux02	demi-, the question is, it that all?
23:30:27	krux02	what about nakefile
23:30:36	krux02	and project.nims
23:30:37	demi-	nakefile isn't a language supported thing
23:30:44	demi-	that is a third party project
23:32:36	krux02	demi-, what is with the nimscipt variant of nimble?
23:32:47	demi-	hm?
23:33:04	demi-	you mean nimble tasks?
23:33:30	Tiberium	can I output the name of the filename where exception happened? In except block
23:34:10	demi-	Tiberium: yeah, look at https://nim-lang.org/docs/system.html#instantiationInfo,
23:37:14	*	Tiberium quit (Remote host closed the connection)
23:41:51	FromGitter	<TiberiumPY> Thanks
23:44:35	krux02	didn't nimble once introduce the "new" format for package descriptions that was a .nims file?
23:46:21	krux02	well I got an answer, yes nimble has a nimscipt fromat but the file is still called .nimble
23:50:23	dom96	indeed
23:51:43	dom96	This project file business seems like it should just be handled by Nimsuggest
23:51:49	zachcarter	wewt frag has texture atlas / spritesheet support now
23:52:05	dom96	Assuming that the finding of this file always follows the same algorithm (and it seems it does)
23:53:57	krux02	args, it feels there are 20 different implementations of find project root in nim-mode
23:54:05	krux02	the code really is bad quality
23:54:33	krux02	dom96, well not really
23:54:56	krux02	dom96, when I need to find the root of a project manually otherwise nimsuggest doesn't work
23:55:32	krux02	dom96, nimsuggest want the project root as an argument otherwise it just doesn't do anything
23:55:35	*	rauss joined #nim
23:55:51	*	rauss quit (Client Quit)
23:56:07	dom96	but why couldn't nimsuggest find the project root on its own?
23:57:42	*	rauss joined #nim
23:58:36	*	yglukhov joined #nim