<< 02-05-2020 >>

---

00:01:58	Yardanico	metaprogramming be like https://i.imgur.com/y7F7jeB.png
00:02:34	Yardanico	and yes I took this print macro from nimpylib XD
00:08:54	FromDiscord	<KingDarBoja> https://github.com/nim-lang/Nim/wiki/Nim-for-Java-programmers
00:09:38	disruptek	ot
00:10:01	disruptek	er, there's nothing inherently wrong with discarding a future.
00:11:33	FromGitter	<alehander92> but there is
00:11:53	FromGitter	<alehander92> it seems that an error can just stay hidden
00:11:59	FromGitter	<alehander92> which seems as a bad thing
00:12:27	FromGitter	<alehander92> maybe i am wrong, but if that's wrong, then we need to fix the "Futures should *never* be discarded." docs
00:13:04	disruptek	unless they MUST never be discarded, i'm okay with it.
00:13:47	disruptek	should we also deny except:?
00:13:47	FromGitter	<alehander92> well, never makes it sounds like it should've been
00:13:51	FromGitter	<alehander92> must never
00:15:25	FromGitter	<alehander92> the problem is that i can easily imagine cases where you discard an async call and this leads to annoying non-obvious bug
00:15:42	FromGitter	<alehander92> and i cant imagine situation where you're like "oh no i need to discard this future"
00:16:02	disruptek	i guess i have a vivid imagination.
00:16:19	FromGitter	<alehander92> ok, give such a case
00:16:24	FromGitter	<alehander92> i am just trying to understand it, sorry
00:16:57	disruptek	i like to think that i'm god when it comes to software. unless i'm going to break the laws of physics, the machine should kowtow to me -- not the other way around.
00:17:27	disruptek	if i chose to type `discard` in front of a future, damnit what do you think i want to do with it?
00:17:52	FromGitter	<alehander92> but many users would just discard asyncCall() as the first obvious way
00:17:55	FromGitter	<alehander92> to run in background
00:17:59	FromDiscord	<Rika> @exelotl ok so after reading a bit of the article you sent, i have a question; how would i ensure that what objects i allocate on the heap are adjacent to each other?
00:18:03	FromGitter	<alehander92> thats just bad language design
00:18:11	disruptek	i'm going to give them all your cell-phone number so you can tutor them.
00:18:17	disruptek	isn't that what you want?
00:18:18	FromGitter	<alehander92> to make the easiest thing a buggy thing
00:18:21	disruptek	a better user-experience?
00:18:44	FromGitter	<alehander92> i want less bugs
00:18:48	disruptek	discard should be introduced later in the manual. that's a patch i'd support.
00:19:05	FromGitter	<alehander92> language design is not about "order in manuals"
00:19:11	FromDiscord	<Rika> then we'd have a lot of random people asking how to have an empty block
00:19:27	disruptek	it's not introduced in that context.
00:19:30	FromGitter	<alehander92> its about less "surprises which woudlnt be surprises if you read manual #2.7"
00:19:34	disruptek	better would be to really explain what it is.
00:19:40	FromGitter	<alehander92> dont we have enough with tax laws
00:19:59	disruptek	i think your argument might make sense for some languages, but not nim.
00:20:11	FromDiscord	<exelotl> if you have a seq of ref objects, where each one is allocated individually with new(), then you have no guarantee where each object will end up in memory
00:20:14	FromGitter	<alehander92> i dont want bugs in my programs in any language
00:20:22	FromGitter	<alehander92> its as simple as that
00:20:33	disruptek	then don't program.
00:20:34	FromDiscord	<exelotl> but if you have a seq of non-ref objects, then those objects are guaranteed to be contiguous
00:20:44	FromDiscord	<Never Listen To Beef> Damn it disruptek, stealing my jokes
00:20:47	FromDiscord	<Rika> but non-ref are stackk
00:20:50	FromGitter	<alehander92> the other option is to make discard asyncCall() actually do what asyncCheck does
00:20:51	FromDiscord	<Rika> ?
00:20:56	FromDiscord	<exelotl> not if they're in a seq
00:20:58	FromDiscord	<Never Listen To Beef> Was going to say "Cant have any bugs in your programs if you dont make any"
00:21:00	FromGitter	<alehander92> but to just swallow errors
00:21:03	FromDiscord	<Rika> well i never mentioned a seq
00:21:04	disruptek	if i cannot discard a future, i cannot use async.
00:21:09	disruptek	that's just nuts.
00:21:12	FromGitter	<alehander92> you cant philosophy your way out of it
00:21:21	FromDiscord	<Rika> okay, how would i make ref objects be contiguous with each other
00:21:33	*	xet7 quit (Quit: Leaving)
00:21:36	disruptek	use memory regions or perform your own alloc.
00:21:59	FromDiscord	<Rika> memory regions?
00:22:03	FromGitter	<alehander92> i dont care about the discard, i care about the "raise error from it if it fails" part
00:22:06	disruptek	--gc:regions
00:22:09	FromDiscord	<Rika> if i perform my own alloc then am i really using nim at this point
00:22:25	disruptek	asyncCheck is terrible. i never use it and i cannot imagine why it exists.
00:22:41	FromDiscord	<exelotl> if it's a worthwhile optimisation then sure rika
00:22:58	FromGitter	<alehander92> ok, i am stupid
00:23:00	FromGitter	<alehander92> so, what should happen
00:23:05	FromGitter	<alehander92> when you do discard a()
00:23:10	disruptek	having related data nearby in memory is almost always an optimization.
00:23:10	FromDiscord	<Rika> okay, ill just keep that in mind; i fear that i am totally doing premature optim. right now
00:23:12	FromGitter	<alehander92> and a completes with fail
00:23:28	disruptek	make it work, make it correct, make it fast.
00:23:35	FromDiscord	<Rika> disruptek: the thing is whether it is worthwhile or not
00:23:52	disruptek	if it's not worthwhile, don't even bother to make it work.
00:23:53	FromGitter	<alehander92> disruptek stop with the slogans
00:24:00	disruptek	that's how i hack on the compiler.
00:24:02	FromGitter	<alehander92> this isnt a socialist demonstration
00:24:05	FromGitter	<alehander92> even if its 1 may
00:24:05	FromGitter	<alehander92> :D
00:24:12	FromDiscord	<KingDarBoja> Who said socialism?
00:24:18	FromGitter	<alehander92> this really doesnt answer a question
00:24:21	FromDiscord	<KingDarBoja> ☭
00:24:26	disruptek	you haven't asked one worthy of answer.
00:24:28	*	xet7 joined #nim
00:24:40	FromGitter	<alehander92> and you can always do
00:24:48	FromGitter	<alehander92> let ignorename = future
00:24:50	FromGitter	<alehander92> here
00:24:57	FromGitter	<alehander92> you have it, discard workaround
00:25:08	disruptek	uh, no.
00:25:15	FromGitter	<alehander92> so this is bad
00:25:22	FromGitter	<alehander92> but *ignoring* errors on purpose
00:25:23	FromGitter	<alehander92> is fine
00:25:37	disruptek	that's why we have discard.
00:25:45	disruptek	because programmers know better than us sometimes.
00:25:52	FromGitter	<alehander92> discard is for ignoring *values*
00:26:01	disruptek	dude.
00:26:05	disruptek	read the code.
00:26:18	FromGitter	<alehander92> dude, you can still ignore an error if you want
00:26:18	FromDiscord	<exelotl> throwback to the time I saw discard in a shader and assumed it was the same as discard in nim
00:26:28	FromDiscord	<Never Listen To Beef> lol
00:26:39	FromDiscord	<Never Listen To Beef> It sorta is, you say you dont want that pixel 😛
00:26:40	FromGitter	<alehander92> the problem is that i've seen that in node
00:27:04	FromGitter	<alehander92> writing several random background calls, never seeing any actual error, random stuff breaks
00:27:07	FromGitter	<alehander92> its just super annoying bugs
00:27:39	FromGitter	<alehander92> you cant discard an exception in sync code
00:27:46	FromGitter	<alehander92> if you could, then ok
00:30:02	disruptek	sure you can.
00:30:03	FromDiscord	<Rika> try: code except: discard
00:30:05	FromDiscord	<Rika> ????
00:30:14	disruptek	nailed it.
00:30:22	FromGitter	<alehander92> and .. you use .. except
00:30:23	FromDiscord	<Rika> with that code you've essentially discarded an exception
00:30:26	FromGitter	<alehander92> totally the same code
00:30:31	disruptek	uh, no.
00:30:32	FromGitter	<alehander92> that's exactly what i am saying
00:30:33	FromGitter	<alehander92> make it raise
00:30:41	FromGitter	<alehander92> then yeah, use try except
00:30:43	FromDiscord	<Never Listen To Beef> ```
00:30:43	FromDiscord	<Never Listen To Beef> a
00:30:44	FromDiscord	<Never Listen To Beef> defer: discard
00:30:44	FromDiscord	<Never Listen To Beef> ```
00:30:44	FromGitter	<alehander92> somehow
00:30:48	FromDiscord	<KingDarBoja> From past disruptek: _basically, it's a nice feature and discard discards it. keep it and use it instead._
00:30:50	disruptek	but it's fine. i know araq agrees with me. 😁
00:30:56	FromDiscord	<Rika> @Never Listen To Beef what's that supposed t odo
00:30:59	FromDiscord	<KingDarBoja> Still no idea what he meant hahaha
00:31:17	FromGitter	<alehander92> but araq agreeing with you doesnt make it correct
00:31:34	FromDiscord	<Never Listen To Beef> i mean that'd put all the stuff inside a try block then discard the finally
00:31:36	disruptek	i'm saying that i'm not terribly motivated to convince you.
00:31:52	FromDiscord	<Never Listen To Beef> I mean it's silly cause it's silly
00:31:54	FromDiscord	<Rika> @Never Listen To Beef it wouldnt catch the exception
00:32:02	FromDiscord	<Never Listen To Beef> Doesnt he want to discard the exception?
00:32:09	FromDiscord	<Rika> thats what he said
00:32:12	disruptek	i want to discard /futures/
00:32:24	FromDiscord	<Rika> oh, you mean disruptek
00:32:44	FromDiscord	<Rika> well i dont understand why you want to but eh i might just not know
00:32:49	FromDiscord	<Never Listen To Beef> I mean i dont either
00:32:53	FromDiscord	<Never Listen To Beef> Exceptions exist to help you
00:32:55	FromGitter	<alehander92> which you can
00:33:08	FromDiscord	<Never Listen To Beef> Although the xlib library stuff i dont catch any of the errors directly
00:35:23	FromGitter	<alehander92> ok disruptek
00:35:30	FromGitter	<alehander92> you need to then solve for me
00:35:40	disruptek	unlikely.
00:35:43	FromGitter	<alehander92> how can i detect if "background" async calls
00:35:48	FromGitter	<alehander92> in my codebase
00:35:54	FromGitter	<alehander92> written from other random people
00:36:02	FromGitter	<alehander92> dont error
00:36:06	FromGitter	<alehander92> detect on runtime*
00:37:07	disruptek	how do you detect any runtime error?
00:37:20	FromGitter	<alehander92> by seeing the exception
00:37:24	FromGitter	<alehander92> or crash
00:37:34	disruptek	well, do that.
00:37:36	FromGitter	<alehander92> or *explicitly* handling it in try except
00:37:40	FromGitter	<alehander92> but i *cant* do that
00:37:45	FromGitter	<alehander92> if somebody just goes discard a()
00:37:49	disruptek	a future is not an exception.
00:38:10	FromGitter	<alehander92> but an async call can be
00:38:26	disruptek	show me your problematic code.
00:38:55	FromGitter	<alehander92> every code that does discard aCall()
00:39:18	disruptek	oh, i see. we should break discard for just this one library in stdlib.
00:39:22	disruptek	that's your position?
00:39:44	disruptek	because someone /could/ write something that has an unexpected effect.
00:39:53	disruptek	because they didn't read the manual.
00:39:57	FromGitter	<alehander92> every time you write discard asyncCall() # you just say "go silent errors"
00:40:04	FromGitter	<alehander92> the language doesnt work that way
00:40:13	FromGitter	<alehander92> it doesnt happen when you call sync calls
00:40:18	FromGitter	<alehander92> when you await normal calls
00:40:18	disruptek	if you don't understand it after your work on await, i don't think i can explain it to you.
00:40:26	FromGitter	<alehander92> oh, i understand it
00:40:33	disruptek	but, look. we don't /have/ to agree.
00:40:37	FromGitter	<alehander92> but does one need to be an expert in async
00:40:47	FromGitter	<alehander92> or to have been bitten 5 times by it?
00:40:54	disruptek	they do if they cannot discard a future, i guess.
00:41:04	FromGitter	<alehander92> ok, lets differentiate
00:41:12	FromGitter	<alehander92> i am ok with discarding a future variable
00:41:20	FromGitter	<alehander92> i dont want to discard an async call
00:41:27	disruptek	dude, i am trying to invent a new try/except. i really don't want to argue with you.
00:41:28	FromGitter	<alehander92> if it was possible to make the difference somehow
00:41:32	FromGitter	<alehander92> i would be ok with it
00:41:44	FromDiscord	<Rika> disruptek: do you think its a stupid idea to introduce a `new` proc that can be passed a count to initialize x ref objects of that type contiguously?
00:41:48	FromGitter	<alehander92> well if you dont want to, dont block useful additions
00:42:09	disruptek	there's no point to argue because it won't be changed.
00:42:14	leorize	Rika: newSeqWith?
00:42:16	disruptek	i'm not blocking anything.
00:42:35	FromGitter	<alehander92> ok, sorry for the wrong way to word it
00:42:50	FromDiscord	<Rika> leorize: ?
00:42:53	FromDiscord	<Rika> hm?
00:43:06	FromGitter	<alehander92> probably we still kinda misunderstand each other
00:43:07	disruptek	discarding an async call... that's another story.
00:43:16	disruptek	discarding a future... this is something else.
00:43:31	leorize	Rika: `newSeqWith` seems to be what you're looking for :P
00:43:32	FromGitter	<alehander92> i agree, sorry, i should've separated them early
00:43:51	FromGitter	<alehander92> but is there a way to detect that on compile time?
00:44:03	disruptek	you could have a macro that detects it.
00:44:29	FromGitter	<alehander92> maybe term rewriting macros can
00:44:32	FromDiscord	<Rika> leorize: i dont see how
00:44:36	FromGitter	<alehander92> but i am very unfamiliar with them
00:44:50	FromDiscord	<Rika> does it guarantee that the objects are near each other in memory?
00:45:04	disruptek	yes.
00:45:26	disruptek	the refs, anyway.
00:45:29	disruptek	not their values.
00:45:37	FromDiscord	<Rika> i mean their values
00:45:47	disruptek	use value types if you want values?
00:45:57	disruptek	value types are better.
00:46:30	FromDiscord	<Rika> that'd be a bit difficult
00:46:37	FromDiscord	<Rika> ill think about it more
00:47:22	leorize	if you want object next to each other, use a seq of values?
00:47:31	leorize	don't copy C's broken semantics
00:48:48	*	PMunch quit (Quit: leaving)
00:49:24	FromDiscord	<Never Listen To Beef> You say value types are better but my OOP brain says, suuuuure 😄
00:49:43	FromDiscord	<KingDarBoja> You mean object variants? 😄
00:50:16	FromDiscord	<KingDarBoja> Btw, what's the recommended way of documenting a type, right after the type Rika = enum line?
00:50:31	leorize	when I first pick up programming I avoided OOP at all costs :P
00:50:35	FromDiscord	<KingDarBoja> Or between type and MyHumbleType (separated lines)
00:50:49	leorize	one indent, then a doc comment
00:51:18	FromDiscord	<KingDarBoja> Okie Dokie 🙂
00:51:34	FromDiscord	<KingDarBoja> `type MyHumble = enum`
00:51:44	FromDiscord	<KingDarBoja> ` ## My docstirng`
00:51:57	FromDiscord	<exelotl> Pick two: by-reference semantics, contiguity, resizable container
00:51:59	FromDiscord	<KingDarBoja> ` north, west, etc..`
00:52:07	FromDiscord	<KingDarBoja> Like that?
00:52:26	leorize	yea
00:52:30	FromDiscord	<KingDarBoja> 😗
00:53:01	FromDiscord	<exelotl> by which I mean, you can have a seq of values and take pointers to those values, but all the pointers will become invalid if the sequence is ever reallocated
00:53:53	FromDiscord	<Rika> i dont think i need resizing
00:54:35	FromDiscord	<Rika> man this is a headache
00:55:41	leorize	what are you trying to do?
00:55:47	leorize	maybe I can figure out a way
00:56:03	leorize	are you by any chance attempting object pooling?
00:58:25	FromDiscord	<Never Listen To Beef> I am sorta interested on how you guys would design a system where Screens have positions, sizes, Workspaces and a selected workspace, and those workspaces have windows, and a selected window
00:58:45	FromDiscord	<Never Listen To Beef> *I feel like you guys wouldnt use ref objects* 😄
00:58:47	FromGitter	<alehander92> disruptek sorry again
00:59:03	FromGitter	<alehander92> and thanks for the direction, i think i found a way
00:59:05	disruptek	we're cool, dude.
00:59:10	FromGitter	<alehander92> to adapt this guy's macro
01:00:09	disruptek	you could use the effect system.
01:00:35	FromGitter	<alehander92> no https://github.com/nim-lang/Nim/issues/11912
01:00:36	disbot	➥ Add undiscardable pragma and forbid Futures from being discarded
01:00:38	FromGitter	<alehander92> it seems it works
01:02:35	FromDiscord	<Rika> i think i know what to do now
01:02:56	FromGitter	<alehander92> sorry https://github.com/nim-lang/Nim/issues/11912#issuecomment-622645281
01:02:56	disbot	➥ Add undiscardable pragma and forbid Futures from being discarded
01:02:57	FromDiscord	<Rika> `object`s in a seq act like a `ref object` no?
01:03:04	FromGitter	<alehander92> yeah it seems this sould be an ok version
01:09:01	FromGitter	<alehander92> do we have a way to say "dont rewrite that more"
01:09:40	leorize	Rika: no, they don't have ref semantics
01:09:58	leorize	you can "refer" to them via indices though
01:12:42	*	muffindrake quit (Quit: muffindrake)
01:14:15	disruptek	alehander92: the best solution is to not add the code in the first place.
01:18:41	*	muffindrake joined #nim
01:31:12	FromDiscord	<Rika> leorize: i misphrased
01:31:21	FromDiscord	<Rika> ill just do whatever and see if i get issues
01:38:36	FromGitter	<alehander92> disruptek its like public safety
01:38:58	FromGitter	<alehander92> sometimes its ok to forbid people entering the hazard zone
01:41:30	*	waleee-cl quit (Quit: Connection closed for inactivity)
01:42:18	FromGitter	<alehander92> otherwise i am ok with giving `ignoreErrors a()` instead of discard
01:42:22	FromGitter	<alehander92> so at least one can grep
01:44:21	FromGitter	<alehander92> (which one can still do now just by using try/empty except around await and run that in the background)
01:45:26	FromGitter	<alehander92> and this way we dont break one such "law" of physics: exceptions travel up to an except clause
01:45:41	FromGitter	<alehander92> not to an except clause or one case of discard
01:46:11	disruptek	discarding a future isn't an exception.
01:46:22	FromGitter	<alehander92> yes, i am not talking about this
01:46:29	FromGitter	<alehander92> i am talking about discard asyncCall()
01:46:50	disruptek	what's the difference?
01:47:00	FromGitter	<alehander92> it is that now suddenly my exception
01:47:05	FromGitter	<alehander92> never gets to an except clause
01:47:11	FromGitter	<alehander92> because it is .. discarded ??
01:47:16	FromGitter	<alehander92> it doesnt make sense
01:47:23	FromGitter	<alehander92> it breaks the laws of physics
01:47:48	disruptek	what exception?
01:47:58	disruptek	i'm talking about futures.
01:48:00	FromGitter	<alehander92> the one somewhere inside asyncCall children
01:48:04	disruptek	y'know, the value you want to discard.
01:48:06	FromGitter	<alehander92> i agreed with you for futures
01:48:13	FromGitter	<alehander92> i changed the title of my issue
01:48:18	disruptek	what do your async procs return?
01:48:30	FromGitter	<alehander92> futures
01:48:34	disruptek	okay?
01:48:48	FromGitter	<alehander92> so then
01:48:57	FromGitter	<alehander92> discard asyncCall() should register an error callback.
01:49:21	disruptek	what's the error?
01:49:53	FromGitter	<alehander92> possibly an exception raised inside
01:50:00	disruptek	so what?
01:50:19	FromGitter	<alehander92> so your whole argument is that discard should not be broken for one case
01:50:27	*	chemist69 quit (Ping timeout: 240 seconds)
01:50:29	FromGitter	<alehander92> and my whole argument is that i also want raise to not be broken for one case
01:50:30	disruptek	it shouldn't be broken at all.
01:50:34	FromGitter	<alehander92> i want my raise
01:50:38	FromGitter	<alehander92> to raise
01:50:42	leorize[m]	what is this discard argument?
01:50:43	disruptek	raise isn't broken.
01:50:49	FromGitter	<alehander92> and the exception to travel until an except
01:50:52	FromGitter	<alehander92> not until a discard
01:50:56	disruptek	it does.
01:50:58	leorize[m]	can someone fill me in on this?
01:51:00	FromGitter	<alehander92> it does not
01:51:08	disruptek	not really, it's silly.
01:51:10	FromGitter	<alehander92> i do asyncCall() = .. raise
01:51:11	disruptek	#11912
01:51:12	disbot	https://github.com/nim-lang/Nim/issues/11912 -- 3Add undiscardable pragma and forbid Futures from being discarded (EDIT: only async calls maybe)
01:51:17	FromGitter	<alehander92> discard asyncCall() # bam no error
01:51:23	FromGitter	<alehander92> no except:
01:51:26	FromGitter	<alehander92> still no error
01:51:30	FromGitter	<alehander92> broken law
01:51:45	disruptek	you just don't understand async.
01:52:00	FromGitter	<alehander92> so, explain it to me
01:52:00	disruptek	it's okay; i'm pretty sure dom doesn't understand it, either.
01:52:09	FromGitter	<alehander92> this is bullshit
01:52:19	FromGitter	<alehander92> with all my love for ya
01:52:31	disruptek	it's a closure. if you don't iterate over it, it doesn't run. if it does except, it stores the exception in the future.
01:52:43	disruptek	you can query it if you want, or don't.
01:52:49	*	chemist69 joined #nim
01:52:53	disruptek	you can raise it in sync code, or not.
01:53:08	disruptek	ima go play video games. maybe leorize can convince you.
01:53:15	FromGitter	<alehander92> i know well the code which is generated
01:53:16	disruptek	i need to learn PoE for work.
01:53:33	disruptek	gotta get those exalts.
01:53:39	FromGitter	<alehander92> but this doesnt address my concerns at all
01:53:43	FromGitter	<alehander92> you still have silent bugs
01:53:46	disruptek	i'm okay with that.
01:53:49	FromGitter	<alehander92> and raise still works unexpectedly
01:54:15	FromGitter	<alehander92> and it shoudln't have anything to do with "its a closure" imho
01:54:16	disruptek	there are lots of fair criticisms to be made of async.
01:54:26	FromGitter	<alehander92> closures dont know of "Futures"
01:54:32	FromGitter	<alehander92> this is not a property of closures
01:56:20	FromGitter	<alehander92> the whole "stores exception" thing is just an asyncmacro invention
01:57:14	FromGitter	<alehander92> and my problem is
01:57:36	FromGitter	<alehander92> i can see how let a = asyncCall() # ignore errors but maybe later do something about them somehow
01:57:39	FromGitter	<alehander92> can be useful
01:58:34	FromGitter	<alehander92> but discard asyncCall() just doesnt serve any useful purpose in my mind, it's just "lets not write let ignored = so people can suffer by silent bugs"
01:58:54	FromGitter	<alehander92> but maybe i am too emotional about it, sorry again, have a great night :D
02:01:50	*	Senketsu joined #nim
02:07:06	*	Senketsu quit (Quit: WeeChat 2.8)
02:08:20	*	Senketsu joined #nim
02:08:43	*	tlanger joined #nim
02:09:13	*	Senketsu quit (Client Quit)
02:11:15	*	ryan_ joined #nim
02:11:38	FromDiscord	<KingDarBoja> Rika, Yard?
02:13:05	*	Senketsu joined #nim
02:14:27	*	tlanger quit (Ping timeout: 258 seconds)
02:15:14	*	mono joined #nim
02:17:42	*	monok quit (Ping timeout: 260 seconds)
02:26:22	*	muffindrake quit (Ping timeout: 260 seconds)
02:28:20	*	muffindrake joined #nim
02:28:37	*	Senketsu quit (Quit: WeeChat 2.8)
02:31:49	FromDiscord	<Rika> hi
02:32:39	FromDiscord	<Rika> @KingDarBoja wat
02:45:35	FromGitter	<bung87> how to do seq equals?
02:48:01	FromDiscord	<Rika> ==?
02:50:33	FromGitter	<bung87> hmm, my bad , there's one item not equals.
03:01:59	*	zacharycarter quit (Ping timeout: 258 seconds)
03:12:36	*	ptdel quit (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.)
03:13:54	*	zacharycarter joined #nim
03:48:06	FromDiscord	<KingDarBoja> Rika, I am back, sorry 😄
03:51:36	*	GabbyWest joined #nim
03:52:36	FromDiscord	<KingDarBoja> Was going to ask if it is possible to create methods at runtime
03:53:58	*	GabbyWest quit (Client Quit)
03:54:32	FromDiscord	<bedwardly-down> @Never Listen To Beef , so i spent part of the day learning how parsing works and some of the weird basic stuff that’s actually pretty intuitive
03:55:03	FromDiscord	<Rika> @KingDarBoja but?
03:55:52	FromDiscord	<KingDarBoja> Just that, 🙂
03:56:23	FromDiscord	<KingDarBoja> I swear I need another monitor
03:56:36	FromDiscord	<KingDarBoja> Can't look at multiple windows, only two max and it is still small
03:56:48	FromDiscord	<Rika> you cannot
03:57:32	FromDiscord	<KingDarBoja> Ah, okay
03:57:48	FromDiscord	<KingDarBoja> Right now comparison 3 different implementations of same module
03:58:11	FromDiscord	<KingDarBoja> The original in JS, the Python one and GO one, this last one seems intuitive as it is strong typed
03:58:48	FromDiscord	<Rika> go will be easiest to port i assume
04:00:14	FromDiscord	<KingDarBoja> But it is the longest one lol
04:01:16	skrylar[m]	Go is cute. I didn't mind writing some of it
04:01:40	FromDiscord	<KingDarBoja> https://github.com/graphql-go/graphql/blob/master/language/visitor/visitor.go
04:02:00	FromDiscord	<KingDarBoja> Source implementation in JS -> https://github.com/graphql/graphql-js/blob/master/src/language/visitor.js
04:02:12	FromDiscord	<KingDarBoja> I feel like it is a mess hahaha
04:02:15	skrylar[m]	when i tried to look up how graphql worked what i found was that apparently its just the whole query string jammed in a GET request o.O
04:02:41	FromDiscord	<KingDarBoja> Oh yeah
04:02:47	FromDiscord	<KingDarBoja> It is
04:03:05	FromDiscord	<KingDarBoja> The parser and lexer works, just need to get the rest of stuff done
04:03:09	skrylar[m]	should probably experiment with it some time; never actually got around to using it for anything. :X
04:03:16	FromDiscord	<KingDarBoja> And ofc use object variants
04:03:25	skrylar[m]	was thinking of toying around with beeflang but the ide is required for some hcr stuff, and the ide doesn't work on linux, so rip.
04:03:39	FromDiscord	<KingDarBoja> It is awesome as it will give u better error responses and type safety
04:04:30	FromDiscord	<KingDarBoja> Did you saw the visitor on JS?
04:04:57	FromDiscord	<KingDarBoja> They using Flow with JS, I really hope the Typescript migration happens on the short term
04:04:59	FromDiscord	<KingDarBoja> x.x
04:05:11	skrylar[m]	whats wrong with flow
04:06:00	skrylar[m]	i gave typescript a light poke and was not impressed waiting 5 seconds for hello world to 'build'
04:06:01	FromDiscord	<KingDarBoja> They had to rely on workarounds for some type definitions
04:06:01	*	supakeen quit (Quit: WeeChat 1.9.1)
04:06:32	FromDiscord	<KingDarBoja> Remember that TS does Type checking and transpile into JS
04:06:43	*	supakeen joined #nim
04:07:01	FromDiscord	<KingDarBoja> So everything will take a bit to transpile into JS then execute
04:07:04	skrylar[m]	yeah so does nim and v
04:07:31	FromDiscord	<KingDarBoja> You're right
04:07:53	skrylar[m]	i still need to check what the opengl module is doing because when i imported it my import stack exploded and my builds went from 1s to 5-7s ... which is weird because it SHOULD just be a pile of innocent getproc's
04:08:33	FromDiscord	<KingDarBoja> Anyway, I will try to port that module, seems messy lol
04:09:20	skrylar[m]	it looks pretty gnarly which is weird for a visitor
04:09:42	FromDiscord	<KingDarBoja> gnarly?
04:11:00	FromDiscord	<KingDarBoja> Also, what you guys recommend for exporting proc / vars on a module
04:11:12	FromDiscord	<KingDarBoja> use `*` or declare at top level using export ?
04:11:18	skrylar[m]	it looks like a hot mess to me; i'm used to seeing visitors just like. switch on a type and then call accept_foo to recurse
04:11:52	skrylar[m]	pretty sure `export` has to do with making it C accessible and `*` is about making it public in the module to other nim code
04:11:58	leorize[m]	@KingDarBoja: export at top level is not designed for exporting proc inside your module
04:12:03	FromDiscord	<KingDarBoja> Ah ok
04:12:08	FromDiscord	<KingDarBoja> Thanks
04:12:24	FromDiscord	<KingDarBoja> skylar: at some part (the visit function) seems to do that but very messy
04:13:10	FromDiscord	<KingDarBoja> Also, do you know Go right? I think you mentioned that few comments ago
04:13:33	skrylar[m]	i dunno enough about what they are doing. i've just seen very clean visitors in Pharo/Squeak and i write them the same way they did elsewhere. it might be more clogged looking because they have to do some extra reflecting or something :shrug:
04:14:21	skrylar[m]	closures in nim are very nice and its pretty easy to do sax style visitors with a struct of them and then just call those.
04:14:40	*	FromDiscord <KingDarBoja> Too bad I am still noobish
04:15:00	FromDiscord	<bedwardly-down> How noobish are you? Day 2 here
04:15:11	FromDiscord	<KingDarBoja> 1 month I think
04:15:18	skrylar[m]	https://github.com/Skrylar/skcbor/blob/master/skcbor.nim#L66 =p
04:15:38	FromDiscord	<KingDarBoja> 2 months to be honest
04:16:15	FromDiscord	<KingDarBoja> Ahhhh
04:16:22	FromDiscord	<bedwardly-down> Nothing wrong with that if you're willing to learn and improve
04:16:23	FromDiscord	<bedwardly-down> 😄
04:16:37	FromDiscord	<KingDarBoja> Skrylar: pretty clever approach
04:17:00	FromDiscord	<KingDarBoja> I am worried about the visit function as it returns `Any` on JS-base implementation
04:17:12	FromDiscord	<KingDarBoja> Python does the same and Golang defines interface{}
04:17:21	FromDiscord	<KingDarBoja> Which I have no idea what that means on GO lol
04:17:32	skrylar[m]	`interface{}` is basically `void*`
04:17:58	FromDiscord	<Rika> so nim's `pointer` xd
04:18:06	FromDiscord	<bedwardly-down> Is that the same as Void in most languages where it returns nothing and cannot be returned?
04:18:21	FromDiscord	<Rika> no
04:18:24	skrylar[m]	no the star makes it a pointer
04:18:30	FromDiscord	<bedwardly-down> Most languages like Java or Haxe
04:18:37	FromDiscord	<Rika> it is the same as "literally any data type"
04:18:53	FromDiscord	<KingDarBoja> Wait calm down
04:18:58	FromDiscord	<Rika> and "good luck figuring out what this type is"
04:19:01	FromDiscord	<bedwardly-down> I missed the pointer comment. I was refering to Void
04:19:08	skrylar[m]	go interfaces have automatic membership and `interface{}` is an interface with no required method, thus every struct is a valid member
04:19:34	skrylar[m]	although go still type checks casting from `interface{}` so there is some safety and runtime overhead
04:19:34	FromDiscord	<KingDarBoja> JS -> Any , Python -> Any (it is just a type hint), Go -> interface{}, Nim -> ?
04:19:45	FromDiscord	<Rika> `pointer`
04:19:48	FromDiscord	<KingDarBoja> Ok
04:19:50	skrylar[m]	not quite
04:19:58	FromDiscord	<bedwardly-down> How do you guys find time to understand as many languages as you do and then apply them in this context?
04:19:59	FromDiscord	<Rika> its the closest thing
04:20:04	skrylar[m]	i think we do have an "any" module but its a little screwy
04:20:17	skrylar[m]	nim's equivalent would be ref rootobj
04:20:24	FromDiscord	<Rika> just dont use it, convert it into a proper static typing system
04:20:30	FromDiscord	<KingDarBoja> I do know Python and TS
04:20:32	skrylar[m]	because if you put something in a pointer it doesn't inform the gc what you did
04:20:35	FromDiscord	<KingDarBoja> Rika: yup
04:20:55	FromDiscord	<KingDarBoja> So anything like skrylar shared
04:21:06	FromDiscord	<KingDarBoja> right? I mean -> https://github.com/Skrylar/skcbor/blob/master/skcbor.nim#L66
04:21:12	skrylar[m]	this problem is why that skcbor module has a box type
04:21:32	*	zacharycarter quit (Ping timeout: 272 seconds)
04:21:47	skrylar[m]	but nims equivalent to `interface{}` is `ref RootObj`
04:22:55	FromDiscord	<KingDarBoja> So the idea should be
04:23:20	FromDiscord	<KingDarBoja> Use the kind (object variants) and call a proc per kind
04:24:19	leorize[m]	skrylar[m]: Nim do have typesafe conversion between ref object types
04:27:52	skrylar[m]	unrelatedly should probably bother someone that `after_code_reload:` causes segfaults on windows in devel
04:28:20	leorize	can you file an issue?
04:29:16	FromDiscord	<KingDarBoja> I like the Golang impl organization of the AST lol
04:29:42	skrylar[m]	go and v are cute.
04:29:51	FromDiscord	<KingDarBoja> https://github.com/graphql-go/graphql/blob/master/language/ast/node.go
04:30:07	FromDiscord	<KingDarBoja> I assume this is equivalent to the object variants
04:30:26	skrylar[m]	are you just parsing graphql?
04:30:51	FromDiscord	<KingDarBoja> By parsing what u mean?
04:30:54	FromDiscord	<KingDarBoja> Porting?
04:31:03	leorize	parsing
04:31:16	leorize	like read the language and understand it kind of thing :P
04:31:28	*	silvernode joined #nim
04:31:38	FromDiscord	<KingDarBoja> Yes
04:31:47	FromDiscord	<KingDarBoja> It does
04:31:52	silvernode	good morning
04:32:03	FromDiscord	<KingDarBoja> Or not sure if you mean reading the spec and doing that
04:32:10	skrylar[m]	if you're just trying to get a nim program to read graphql queries you're probably better off looking at references for parsing in static languages (like npeg or mratsim's thing) than trying to copy how dynamic languages do it
04:32:40	FromDiscord	<KingDarBoja> That's why I started to look at Go
04:33:21	FromDiscord	<KingDarBoja> But the source implementation was done in JS so that's why it gets harder to understand things and how stuff got implemented
04:33:38	skrylar[m]	yea parsing used to be my nemesis
04:34:01	leorize	if you write a parser, lexbase is an excellent base
04:35:14	FromDiscord	<KingDarBoja> It's not only a parser 😛
04:35:50	silvernode	Can anyone explain what scenario I might want to use a reference object? Still trying to make my grid map for my game and someone here recommened making a position type that takes a reference to the grid type but I am not really sure how to approach this or why I would want to.
04:36:00	FromDiscord	<KingDarBoja> But been able to overcome some stuff and learn Nim on the way
04:36:08	FromDiscord	<KingDarBoja> Thanks to you guys ofc
04:36:26	leorize	silvernode: you use a ref object when you need the reference semantics
04:36:42	skrylar[m]	well when it comes to ASTs they are purely there for the convenience of a program trying to understand stuff; there is often no point in trying to be exactly close to another AST
04:36:57	leorize	silvernode: the "I want this to reference the same thing when I 'copy' it" and not "I want to optimize"
04:37:52	skrylar[m]	@silvernode references are pointers but garbage collected
04:39:06	silvernode	hmm, this all sort of makes sense to me but I'll be honest, I am still trying to comprehend the concept of pointers/refs.
04:40:07	leorize	there's the low level and the high level aspect of them
04:40:57	leorize	let me see if I can still find this documentation Zevv wrote
04:41:18	silvernode	leorize: thanks
04:41:27	FromDiscord	<KingDarBoja> skrylar: https://graphql.org/code/ look at all the current graphql implementation languages
04:43:49	leorize	silvernode: http://zevv.nl/nim-memory/
04:44:43	silvernode	leorize: hopefully this will help me understand, thank you :)
04:46:06	leorize	I don't think I remember how I understood pointers, but I think what I did was that I played a lot with them
04:46:22	FromDiscord	<KingDarBoja> That's a neat doc
04:46:26	FromDiscord	<KingDarBoja> Bookmarked
04:46:33	leorize	~memory
04:46:39	leorize	disbot?
04:46:42	FromDiscord	<KingDarBoja> Rip
04:46:50	leorize	disruptek: your bot is dead again
04:48:24	*	zacharycarter joined #nim
04:49:44	leorize	playing with pointers seem to be the way. I've just read some of the materials I used back when I step in programming and none of them fully explain the concept
04:50:24	skrylar[m]	beef has some interesting memory operators
04:50:55	skrylar[m]	mostly around letting you muck with custom mallocs for certain purposes
04:51:17	skrylar[m]	but they are explicitly gc-less so they can get away with more silly things easier
04:53:12	*	zacharycarter quit (Ping timeout: 272 seconds)
04:53:22	leorize	Rika: we do actually have a way to allocate a block of GC-ed memory
04:53:26	leorize	not that I recommend it
04:53:37	FromDiscord	<Rika> which is
04:54:41	leorize	unsafeNew
04:56:45	leorize	but note that this is for allocating a block
04:56:54	leorize	not for "a bunch of refs side by side"
04:57:27	*	silvernode quit (Ping timeout: 260 seconds)
05:02:05	*	chemist69 quit (Ping timeout: 240 seconds)
05:02:56	leorize	what would be a sane default for a process spawning interface?
05:03:01	leorize	what do people use the most?
05:03:18	*	chemist69 joined #nim
05:04:44	FromGitter	<bung87> `execCmdEx` seems good
05:06:04	FromGitter	<bung87> it return exitcode and stdout , quite common use case.
05:10:02	FromDiscord	<Never Listen To Beef> If i see another person talk about the beef programming language i think i have to change my name
05:10:10	FromDiscord	<KingDarBoja> Beeeef
05:10:24	FromDiscord	<KingDarBoja> I can't belive it is actually a language
05:10:28	FromDiscord	<KingDarBoja> With a beef logo lol
05:11:59	FromDiscord	<Rika> ref seqs are identical to seqs in memory?
05:12:00	FromDiscord	<Rika> cool
05:13:44	FromDiscord	<Rika> ah nvm, zev's article is outdated?
05:14:35	leorize	not that outdated
05:15:15	leorize	actually it's super up to date regarding old runtime
05:15:59	FromDiscord	<Rika> hmm, i think its just how nim's repr prints seqs now
05:16:13	FromDiscord	<Rika> i dont think repr for seqs show that its from a ptr anymore?
05:16:38	leorize	!eval echo repr @[ 0x30, 0x40, 0x50 ]
05:16:41	NimBot	0x563e87311860@[48, 64, 80]↵
05:18:05	*	NimBot joined #nim
05:25:46	disruptek	~memory
05:25:47	disbot	no footnotes for `memory`. 🙁
05:25:57	disruptek	when was that ever a thing?
05:27:29	leorize	I was checking to see if it's free
05:27:30	*	zacharycarter joined #nim
05:28:09	FromDiscord	<Rika> so why doesnt nim show that seqs are pointers tho
05:28:13	leorize	~memory: The Nim memory model: https://zevv.nl/nim-memory/
05:28:13	FromDiscord	<Rika> in repr
05:28:14	disbot	no footnotes for `memory:`. 🙁
05:28:21	leorize	~memory The Nim memory model: https://zevv.nl/nim-memory/
05:28:22	disbot	no footnotes for `memory`. 🙁
05:28:22	FromDiscord	<Rika> its memory is
05:28:27	leorize	~memory is The Nim memory model: https://zevv.nl/nim-memory/
05:28:28	disbot	memory: 11The Nim memory model: https://zevv.nl/nim-memory/
05:28:33	leorize	thanks :P
05:32:01	*	zacharycarter quit (Ping timeout: 265 seconds)
05:51:20	FromGitter	<dawkot> Is VSCode Nim extension popular?
05:51:23	FromGitter	<dawkot> Seems pretty inactive on github
05:51:39	Yardanico	well it's ok
05:51:49	Yardanico	I think it's still the most popular choice for nim
05:52:31	FromDiscord	<Never Listen To Beef> Use gary's branch for nicer syntax highlighting
05:52:31	FromDiscord	<Never Listen To Beef> https://marketplace.visualstudio.com/items?itemName=garym.nim-alt
05:53:34	Yardanico	well syntax highlighting isn't the main issue with it
05:54:28	FromDiscord	<Never Listen To Beef> Sure, well i didnt say it was, just saying it has better syntax highlighting 😄
06:13:27	FromDiscord	<KingDarBoja> Installing rightnow
06:15:16	FromDiscord	<KingDarBoja> Done, seems ok
06:15:18	FromDiscord	<KingDarBoja> good night
06:17:44	*	solitudesf joined #nim
07:00:00	*	gmpreussner quit (Quit: kthxbye)
07:02:06	FromGitter	<bung87> how to do async test?
07:04:45	*	gmpreussner joined #nim
07:15:35	*	Vladar joined #nim
07:57:52	*	voltist joined #nim
08:07:04	FromDiscord	<Benumbed> Can anyone spot what I'm doing wrong here? https://play.nim-lang.org/#ix=2kqM as far as I undersood, default params are fine in templates...
08:15:32	FromDiscord	<Rika> @Benumbed `(ref exceptn)(class: class, meth: meth, code: code, msg: message)` becomes `(ref exceptn)(0: 0, 0: 0, 0: 0, msg: message)`
08:15:38	FromDiscord	<Rika> rename the variables to something ele
08:15:40	FromDiscord	<Rika> else*
08:17:43	FromDiscord	<Benumbed> oh my god I feel quite stupid, thanks @Rika
08:17:47	*	dddddd joined #nim
08:17:47	FromDiscord	<Benumbed> I should go to bed I think lol
09:26:00	*	Trustable joined #nim
09:34:11	*	fredrikhr joined #nim
09:39:12	*	dddddd quit (Ping timeout: 256 seconds)
09:46:20	*	krux02 joined #nim
09:48:42	voltist	It seemed only fitting that my website, made with Nim, should use the Dracula theme: https://liamsc.com/
09:51:57	*	krux02 quit (Remote host closed the connection)
09:54:01	*	fredrikhr quit (Read error: Connection reset by peer)
09:54:26	*	fredrikhr joined #nim
10:04:30	*	fredrikhr quit (Read error: Connection reset by peer)
10:04:55	*	fredrikhr joined #nim
10:30:21	*	filcuc joined #nim
10:31:01	*	zacharycarter joined #nim
10:31:29	*	fredrikhr quit (Read error: Connection reset by peer)
10:31:54	*	fredrikhr joined #nim
10:35:43	*	zacharycarter quit (Ping timeout: 256 seconds)
10:36:40	Prestige	voltist: bad gateway
10:37:20	voltist	Prestige: Seems fine to me
10:37:28	Prestige	huh weird
10:37:30	*	fredrikhr quit (Read error: Connection reset by peer)
10:37:49	Prestige	https://0x0.st/i2qs.png
10:37:58	*	fredrikhr joined #nim
10:40:18	solitudesf	502 for me too
10:42:30	*	fredrikhr quit (Read error: Connection reset by peer)
10:42:55	*	fredrikhr joined #nim
10:45:17	*	zacharycarter joined #nim
10:46:16	*	filcuc_ joined #nim
10:46:28	*	filcuc quit (Ping timeout: 256 seconds)
10:48:38	*	waleee-cl joined #nim
10:48:58	Araq	Yardanico: your bug is killing me... :P
10:49:52	Araq	ah no
10:49:58	*	fredrikhr quit (Read error: Connection reset by peer)
10:50:12	Araq	found it
10:50:14	Araq	lol
10:50:25	*	fredrikhr joined #nim
10:51:28	*	fredrikhr quit (Read error: Connection reset by peer)
10:51:56	*	fredrikhr joined #nim
10:54:28	*	fredrikhr quit (Read error: Connection reset by peer)
10:54:55	*	fredrikhr joined #nim
10:55:34	*	fredrikhr quit (Read error: Connection reset by peer)
10:56:00	*	fredrikhr joined #nim
11:00:52	*	filcuc_ quit (Ping timeout: 260 seconds)
11:14:14	*	lritter joined #nim
11:48:41	*	zacharycarter quit (Ping timeout: 258 seconds)
11:48:43	*	nsf joined #nim
12:02:13	FromDiscord	<Zed> is it faster to return a new object or edit the object passed into a proc as a variable?
12:02:31	FromDiscord	<Zed> is it bad practice to edit the parameter rather then returning a new object?
12:06:02	*	supakeen quit (Quit: WeeChat 1.9.1)
12:06:41	*	supakeen joined #nim
12:07:34	Araq	a good rule is that inplace mutation is fine if there is only a single owner
12:08:15	Araq	and in general inplace mutation is faster unless the optimizer can remove the spurious copies but it's hard to tell without seeing any code
12:08:57	FromGitter	<bung87> edit param is programming like c way, that depends, on the other hand programming like functional, just get input and put output, logical simple
12:09:44	Araq	in general optimizers in 2020 are still disappointing unless you're into heavy number crunching code
12:09:55	Araq	IME anyway, ymmv
12:10:39	Araq	on the other hand just today I fixed Nim's move optimizer to be less aggressive
12:11:54	dom96	good afternoon
12:12:00	Prestige	o/
12:12:16	FromGitter	<bung87> optimizer? is it about compiler?
12:13:42	Araq	bung87: yes. Nim does optimize too, not only the C step, but you can't disable it
12:14:49	dom96	Why are we still arguing about disallowing `discard`?
12:14:55	FromDiscord	<Zed> thanks Araq!
12:15:28	Araq	dom96, we aren't, the discussion is over :P
12:15:29	FromGitter	<bung87> oh,I have no idea about compiler optimizer, I just know react and parcel's optimization
12:17:07	dom96	Araq, so your problem is with how it's implemented
12:17:18	dom96	I don't really care how, do you at least agree that discarding futures should be disallowed?
12:17:35	dom96	The bottom line is, either we disallow it or we make `asyncCheck` implicit
12:17:44	Araq	I don't agree that it should be disallowed
12:17:50	Araq	not anymore.
12:17:58	dom96	I'm pretty sure the latter was tried and it caused bugs
12:18:21	Araq	it would only introduce a subtle case where 'let _ = f()' is not the same as 'discard f()' anymore
12:18:22	dom96	This is an obvious source of bugs
12:18:42	dom96	no, and that should be disallowed as well.
12:19:00	Araq	I don't agree, today I can write 'discard posix_api_call()' and it's a code smell too
12:19:02	dom96	but really if someone tries `discard future` and they see "Do not use this in async procs, it's a stupid thing to do"
12:19:09	dom96	then if they work around it then it's their own stupid fault
12:19:15	FromGitter	<bung87> discard async proc will make it like js behavior?
12:19:19	dom96	how is someone meant to know that discarding futures is bad without that message?
12:19:25	Araq	it's all their own stupid fault to begin with
12:20:11	Araq	it's like "'cast' needs more hand holding from the compiler", no it doesn't need that, it's a dangerous sharp weapon.
12:20:39	dom96	Okay, it's a fair argument to correlate it to discarding posix API return values (which can indicate errors)
12:23:20	Araq	having said that, can we improve the tooling in other ways? how about we add future leak detection at runtime. or "warn me for blocking calls" at runtime
12:25:30	*	zacharycarter joined #nim
12:25:48	dom96	I think in the end we may need to be pragmatic here. The compiler's current error messages are not good enough.
12:26:59	FromDiscord	<dom96> https://play.nim-lang.org/#ix=2kse
12:27:16	FromDiscord	<dom96> Error: expression 'foo()' is of type 'Future[system.void]' and has to be discarded
12:27:23	FromDiscord	<dom96> The compiler leads the user in the wrong path
12:27:37	Araq	aha
12:27:40	FromDiscord	<dom96> Implying the use of `discard`
12:27:48	Araq	'and has to be *used* '
12:28:11	dom96	ideally we should mention `asyncCheck`
12:28:29	Araq	that's acceptable, library specific error messages / filters should be a thing
12:29:01	zedeus	+1 for a pragma to specify custom "dont discard" message as the reason can vary
12:29:27	Araq	but in the general case replacing 'discarded' with 'used' would be a nice improvement
12:29:39	dom96	sure, seems like a quick and easy fix
12:30:35	zedeus	'discarded' serves as a hint to new users that may not be very familiar with the discard keyword
12:31:56	Araq	zedeus, yeah that was the idea but for async is counter productive advice
12:32:41	zedeus	yeah I agree, it's a pretty ironic error message :)
12:34:08	dom96	I updated #11912 to reflect what we discussed
12:34:24	dom96	disbot down?
12:34:49	zedeus	can anything be done for cases where async is used, but the return value is optional and rarely used? example is Redis where for most commands 0 or 1 is returned, and you very rarely need it. {.discardable.} doesn't work since that would apply to the future. I'd like to avoid `discard await ..` everywhere, gets hairy quickly
12:35:17	zedeus	I guess it's less of a problem now with await being a template
12:35:33	dom96	how does `await` being a template affect this?
12:36:04	zedeus	you could make a `dawait` template that just does `discard await`
12:36:05	dom96	yeah, that's a problem, because you want the result of future.read to be discardable not the result of the async call
12:36:14	zedeus	yep
12:36:15	dom96	which we have no way to model, and likely won't :/
12:36:25	dom96	zedeus, please don't do that :(
12:36:41	Araq	zedeus, do that, it's cool :-)
12:38:06	zedeus	I might.. it almost kept me from using redis because 10 lines of `discard await` isn't nice to read
12:39:37	*	filcuc joined #nim
12:39:47	zedeus	oh you're gonna love this one. for my chat bot I made a `safeAsync` proc that's basically `asyncCheck` but it ignores exceptions :)
12:41:46	dom96	yeah.. have fun with that
12:42:45	dom96	hopefully by "ignores" you mean it at least logs them
12:43:59	zedeus	sure. the problem with asyncCheck exceptions is it crashes the program with nothing to catch them
12:45:01	*	zacharyc1rter joined #nim
12:45:26	zacharyc1rter	is it possible to do `amd64.macosx.gcc.linkerexe` in a nims file? or on a nim.cfg file?
12:45:42	zacharyc1rter	can I just use - `switch("amd64.macosx.gcc.linkerexe", "clang++")` ?
12:46:43	dom96	zedeus, yep, the intention is actually to allow people to create their own `asyncCheck` variants
12:47:18	dom96	we should probably add some common ones, I for example have one in my game that uses the logging module to log the exceptions
12:47:54	*	filcuc quit (Ping timeout: 240 seconds)
12:48:00	zedeus	that would be great, asyncCheck is useful as a way to say "go do that elsewhere", but also often useless because of the exceptions
12:48:58	*	fredrikhr quit (Read error: Connection reset by peer)
12:49:26	*	fredrikhr joined #nim
12:52:22	dom96	btw you could have defined a dawait before await became a template, you just need to use `yield` in the template
12:52:56	Araq	zacharyc1rter, I think so, yes, try it
12:53:24	Araq	'dawait', come on, it's awesome. it almost sounds like a real word
12:53:41	zacharyc1rter	Araq: I think I'm doing something wrong anyway as I'm trying to get clang++ to link my program but whether I use nim.cfg or config.nims it doesn't work
12:54:13	Araq	use 'nim cpp --cc:clang'
12:54:27	zacharyc1rter	well I'm compiling with clang but I want to link with clang++
12:54:37	zacharyc1rter	since I have an objectfile that was comiled with C++
12:54:45	zacharyc1rter	compiled*
12:55:26	zacharyc1rter	I'm not sure it even matters because things seem to work even when I just link with clang
12:55:34	zacharyc1rter	but I thought it would...
12:56:21	Araq	had the same problem with DrNim and my solution was 'nim cpp'
12:57:06	zacharyc1rter	unfortunatley that breaks nimterop and what I"m using that for :/
12:57:36	zacharyc1rter	I guess I could compile that code separately and link it
12:57:36	Araq	you can also link with C++'s stdlib, somehow
12:57:54	zacharyc1rter	oh yeah I'm doing that already `--passL:-lc++`
12:58:09	Araq	well then you should be fine
12:58:13	zacharyc1rter	ah okay - thanks
12:58:53	*	xcm quit (Ping timeout: 246 seconds)
12:59:26	FromDiscord	<bedwardly-down> Tbf, dawait sounds very catchy and rolls well off the tongue. It also describes how async can feel, since you have “the wait” until it hits
12:59:43	*	tane joined #nim
12:59:51	FromDiscord	<bedwardly-down> I like puns in code i read
12:59:59	*	idf joined #nim
13:00:03	*	xcm_ joined #nim
13:00:05	Araq	1 showstopper bug left :-)
13:00:10	zacharyc1rter	\o/
13:00:33	disruptek	the bot isn't stable on orc/arc. 🙁
13:00:40	FromDiscord	<bedwardly-down> Nice. Morning all. O7
13:00:55	Araq	disruptek, fixed an important bug
13:00:56	zacharyc1rter	morning
13:01:17	disruptek	cool, i will try new builds.
13:01:26	Araq	wait until I merged my PR
13:01:37	FromDiscord	<bedwardly-down> Got a link? Disruptek was a pleasant host yesterday during his stream
13:01:40	dom96	Araq, you asked me a while ago about which Nimble to bundle with the new release
13:01:41	disruptek	i saw you fixed the cyclic error.
13:01:49	dom96	never replied to my questions, which version are you releasing?
13:02:05	Araq	disruptek, it was not just a cycle error, it was a 'sink' parameter bug
13:02:17	Araq	dom96, read the top of koch.nim
13:02:26	disruptek	oh wow.
13:02:28	Araq	but we already updated it
13:02:36	disruptek	that's kinda critical.
13:02:45	Araq	disruptek, I know :-)
13:03:00	dom96	you're bundling a non-released Nimble commit?
13:03:16	Araq	dom96, in devel
13:03:42	dom96	okay, what about the release, that was what your question was about AFAIK
13:03:43	Araq	not in an official release
13:03:55	Araq	well make a Nimble release please for us
13:04:24	Araq	gah, it's raining again. :-(
13:04:46	Araq	anyhow the real question I wanted to ask if we should ship the new cycle collector in 1.2.2
13:05:16	*	ftsf joined #nim
13:05:18	dom96	sure, why not?
13:05:36	Araq	dom96, because some might say it's a new feature
13:06:06	dom96	so release 1.4.0?
13:06:49	dom96	or 1.3.0 or however the versioning works these days
13:07:12	disruptek	if you don't cut a new minor then other software cannot easily /require/ your new feature.
13:07:18	disruptek	that's the point.
13:09:19	Araq	dom96, it would be 1.4.0 indeed and the versioning works as it always worked
13:10:35	disruptek	"my async code leaks", "what ver?", "1.3", "use >=1.4.0 and --gc:orc", "ah, cool, thanks"
13:13:12	FromDiscord	<dom96> https://github.com/nim-lang/nimble/releases/tag/v0.11.2
13:13:29	Araq	disruptek, *shrug* what about it, we never said --gc:arc is production ready for async. we're working hard to make it production ready
13:14:00	disruptek	costs us nothing to use fresh integers; there are plenty left.
13:14:10	dom96	Araq, be sure to include that Nimble 0.11.2 in the new release.
13:17:34	Araq	what's new in Nimble 0.11.2?
13:17:51	Araq	never mind, I was blind
13:19:15	FromDiscord	<mratsim> I'm really interested in your plans to make closures gc-safe so that we can pass them through channels.
13:20:32	*	filcuc joined #nim
13:25:13	*	nsf quit (Quit: WeeChat 2.8)
13:25:36	FromDiscord	<bedwardly-down> I love the changelog. It's slim enough that it's easy to parse and understand but detailed enough that I can decide whether or not to upgrade since it's such a minor release. \o/
13:26:19	Araq	mratsim: with orc we have a new tool 'proc isIsolated[T: ref|closure](x: T): bool' available. isolated subgraphs can be transferred to a different thread safely
13:26:41	*	Vladar quit (Quit: Leaving)
13:27:02	Araq	that's the plan.
13:28:55	dom96	bedwardly-down: thanks, glad someone appreciates my changelogs :)
13:29:17	zedeus	it would be real nice if you could do a new jester release
13:29:41	disruptek	zacharyc1rter: anyone can make new rooms on the mumble server so they have low-latency voice chat on their stream.
13:29:53	disruptek	i
13:30:02	disruptek	i'm happy to see it get used by the community.
13:31:53	zacharyc1rter	that's awesome thanks mannnnn
13:32:01	zacharyc1rter	I will do that today when I start streaming
13:32:10	disruptek	cool.
13:32:57	zedeus	what are you going to do on stream?
13:33:25	zacharyc1rter	continue working on game engine dev from yesterday - still trying to figure out why the f this code works in one project and not in this new one
13:33:37	zacharyc1rter	I don't want to bore users with this debugging
13:33:37	zedeus	ah cool
13:33:54	disruptek	it's not that easy, is it?
13:34:06	zacharyc1rter	streaming?
13:34:11	disruptek	yeah.
13:34:16	FromDiscord	<bedwardly-down> dom96, are you the core maintainer of nimble? I appreciate the TAB cycling approach for going through its interactive options. Not many devs have that down.
13:34:17	zacharyc1rter	yeah it's a task for sure
13:35:02	dom96	bedwardly-down: I'm the one that originally created it, nowadays I have very little time to maintain it, but I do what I can. That particular feature was contributes by our community :)
13:35:15	dom96	*contributed
13:35:50	FromDiscord	<bedwardly-down> Zachary, if you're streaming, I'd be interested in seeing some debugging stuff. Debugging is very much where I struggle in all languages I dabble in
13:37:03	zacharyc1rter	disruptek: I think I need to be an admin to create channels
13:37:06	zacharyc1rter	could you create one for me?
13:37:13	zacharyc1rter	bedwardly-down: wish granted - I'm about to start my stream
13:37:31	FromGitter	<dawkot> Is there a way to make compiler check if a method has been overloaded on all subclasses?
13:37:38	disruptek	yeah, i was hoping you'd figure it out for me. 🤣
13:37:43	zacharyc1rter	lol
13:38:18	zacharyc1rter	https://www.twitch.tv/zachary_carter
13:38:37	FromDiscord	<bedwardly-down> I'm there. Zach
13:39:11	FromDiscord	<bedwardly-down> @Zachary Carter https://gettc.xyz/ that's pleasantly easy to setup on Linux
13:39:46	FromDiscord	<bedwardly-down> It's a straight Appimage. Just make sure you close out the stream preview or else you'll get audio feedback
13:39:52	*	filcuc quit (Ping timeout: 256 seconds)
13:44:28	*	filcuc joined #nim
13:45:58	*	fredrikhr quit (Read error: Connection reset by peer)
13:46:26	*	fredrikhr joined #nim
13:49:45	*	Vladar joined #nim
13:51:08	shashlick	@zacharycarter why is nimterop breaking
13:51:16	zacharyc1rter	it only is when I'm using C++
13:51:19	zacharyc1rter	or at least it was in previous versions
13:51:46	shashlick	Any details you can share? Am curious
13:52:00	zacharyc1rter	I can put up an example later today
13:52:06	zacharyc1rter	with sokol
13:52:56	shashlick	Ok thanks
13:53:00	zacharyc1rter	np!
13:57:40	*	dddddd joined #nim
14:01:32	Araq	any love for type (Node, NodeKind) = ref object ?
14:01:42	Araq	see https://github.com/nim-lang/RFCs/issues/219 for the context
14:01:45	disbot	➥ Object variant kind field type inference ; snippet at 12https://play.nim-lang.org/#ix=2ksV
14:01:54	disruptek	thanks, i hate it.
14:04:16	dom96	no, if you want to improve it then go all the way and add some syntax inspired by sum types in other langs
14:04:47	dom96	maybe we can come up with something that feels like idiomatic Nim
14:05:15	disruptek	one field two branches is the more important feature.
14:05:30	companion_cube	sum types ♥
14:06:14	Araq	dom96, we can also simply add patty to the stdlib
14:06:26	*	filcuc_ joined #nim
14:06:31	dom96	disruptek, +1
14:06:35	Araq	but it's worth exploring how to change today's syntax
14:06:45	*	filcuc quit (Ping timeout: 272 seconds)
14:07:18	Araq	type
14:07:25	Araq	T = case kind
14:07:35	Araq	of intVal: x: int
14:07:43	Araq	meh
14:08:12	Araq	the ref vs non-ref distinction alone indicates the old syntax is good
14:10:12	dom96	heh, nice, we're actually mentioned in https://en.wikipedia.org/wiki/Tagged_union#2000s
14:11:18	dom96	Maybe we should copy Haxe's syntax: https://play.nim-lang.org/#ix=2ksZ
14:11:35	companion_cube	it's a bit cheating because nim doesn't check for invalid field access :>
14:11:54	dom96	in fact, I really like that syntax
14:13:19	Araq	companion_cube, it does check, use warn[ProveField]:on
14:13:43	FromDiscord	<bedwardly-down> Amen to that, dom96. There are things that drive me nuts about Haxe and its development at times, but the syntax is spot on, at least to me
14:14:03	FromDiscord	<bedwardly-down> That's part of why I love using it as my main
14:14:08	companion_cube	ok, but why is it opt in?!
14:14:31	*	Zevv joined #nim
14:14:39	Araq	companion_cube, because a day only has 24 hours
14:14:56	FromDiscord	<bedwardly-down> dom, `/usercode/in.nim(3, 12) Error: identifier expected, but got '('`
14:15:08	dom96	The English language should have a word for "because a day only has 24 hours"
14:15:11	FromDiscord	<bedwardly-down> that's what running it as C gets me in the link
14:15:28	dom96	Many things are not so because they simply haven't been implemented yet
14:15:42	disruptek	~features
14:15:43	disbot	features: 11begin in the unfortunate state of /unimplemented/ -- disruptek
14:15:44	dom96	bedwardly-down: it's a syntax proposal, of course it doesn't work
14:15:47	companion_cube	but in this case it's implemented, right?
14:16:03	dom96	companion_cube, presumably it needs refining and testing
14:16:16	dom96	Araq, what do you think about that syntax?
14:16:38	Araq	it's implemented but probably needs more refinement if we enable it for everybody
14:16:40	zedeus	dom96: that syntax is pretty much what patty does, it's great. I think it doesn't support multiple branch values (of a, b: ...) and it makes it more strict since you can't have other shared fields in the object
14:17:21	Araq	dom96, as I said, how to make the 'ref object' vs object distinction with your syntax?
14:17:31	dom96	hm
14:17:34	dom96	`ref enum` :P
14:17:51	Araq	Swift uses 'indirect enum' fwiw
14:18:05	Araq	'ref enum' could work though
14:18:06	dom96	what does the `indirect` imply?
14:18:10	disruptek	i don't want to define both at once.
14:18:49	Araq	dom96, that it's a 'ref' ;-)
14:19:07	disruptek	i personally won't use the inference, either. i like that it's explicit, unambiguous.
14:19:21	FromDiscord	<bedwardly-down> Gotya, dom. 😓
14:20:09	dom96	Araq, yeah, meh, I prefer `ref`
14:20:23	dom96	but I suppose the point is that Swift also has this additional keyword
14:20:36	dom96	using `ref` for us would be nicely consistent
14:21:22	Araq	indeed, looks lovely
14:21:29	dom96	on the other hand, it would put the Nim philosophy of "Small core and macros" in the coffin :P
14:21:37	dom96	We have a macro for this, why don't we use it?
14:21:55	dom96	Language features are more official and therefore are easier to adopt
14:22:05	*	silvernode joined #nim
14:22:28	Araq	true but maybe it's time to grow up and accept that Nim is more about what is practical rather than theoretically perfect
14:22:42	dom96	yep
14:22:52	disruptek	what a lame argument.
14:22:59	Araq	and we don't do type system extensions via macros
14:23:17	Araq	we never did that, it's too hard to let the macro writers do it
14:23:31	Araq	type systems are beasts
14:23:42	disruptek	i'm doing it in my persistence project.
14:23:56	disruptek	and
14:23:58	disruptek	!repo nesm
14:23:59	disbot	https://github.com/xomachine/NESM -- 9NESM: 11NESM stands for Nim's Easy Serialization Macro. The macro that allowing generation of serialization functions by one line of code! (It is a mirror of https://gitlab.com/xomachine/NESM) 15 40⭐ 2🍴
14:24:23	disruptek	technically, that's not a type extension.
14:24:32	Araq	exactly
14:25:01	Araq	but you're right, the big problem is how to do serialization of sum types in a composable manner
14:25:03	disruptek	write it as a macro.
14:25:36	disruptek	prove that it cannot be done correctly. 😉
14:25:37	dom96	that's not a problem
14:25:52	Araq	dom96, it is
14:26:04	dom96	what does "in a composable manner" mean?
14:26:17	Araq	I know code that uses .push fieldChecks: off in order to accomplish it
14:26:40	Araq	dom96, it means you can construct via 'obj.kind = X; obj.field = value'
14:26:58	Araq	instead of 'obj = Constr(field: value)'
14:27:01	dom96	hrm
14:27:23	zedeus	json's `to` can do it
14:27:47	*	filcuc_ quit (Ping timeout: 246 seconds)
14:27:53	Araq	zedeus, it can do it because Nim's case objects allow for it
14:28:15	Araq	we need to be careful that the new, superior way can also do it...
14:28:30	dom96	json's `to` would be able to do it if you allowed expressions in these Constructor nodes
14:28:33	*	solitudesf quit (Ping timeout: 260 seconds)
14:28:39	zedeus	hm, msgpack4nim can't do it so I'm a little confused
14:28:42	dom96	but json's `to` doesn't really scale well
14:29:28	dom96	but yeah, this does pose a problem
14:29:36	Araq	dom96, I cannot easily allow it if we're after more static checking
14:29:45	dom96	My thrift deserializer uses fieldPairs so it also suffers from this
14:30:04	disruptek	i'm all for improvements, but this isn't a priority right now and the community could make more progress with an RFC until the priority increases.
14:30:18	Araq	disruptek, indeed. bbl
14:30:26	dom96	true
14:31:20	FromDiscord	<bedwardly-down> Nimble init does something goofy when a project directory has a '-' in it. It creates a subdirectory with the root directory's name but replaces '-' with '_'
14:31:46	disruptek	imports cannot contain '-'
14:31:47	FromDiscord	<bedwardly-down> Example: `nimble-echo` becomes `nimble-echo/nimble_echo`
14:31:52	FromDiscord	<bedwardly-down> Gotya
14:32:21	disruptek	s/imports/identifiers/
14:32:36	FromDiscord	<bedwardly-down> Substitute?
14:33:12	FromDiscord	<bedwardly-down> And in the global config I'm guessing?
14:34:55	FromDiscord	<bedwardly-down> Solved the issue by just passing it as an argument when initting
14:35:26	FromDiscord	<Zed> has anybody had any trouble using nim with emacs?
14:35:27	FromDiscord	<Zed> I cant seem to get indentation working properly
14:36:18	zedeus	hey other Zed, yeah I'm using nim with emacs, you're probably missing the nimsuggest settings
14:36:47	dom96	bedwardly-down: you shouldn't force the naming, you will need to use underscores
14:36:53	silvernode	So I got a Raspberry Pi 4 and tried a Nim nightly version. I think 1.3.0, but it complained about my nimble file that works fine on earlier versions. Not a big deal, just wanted to share my experience. I can load it up again if someone wants the error.
14:36:56	zedeus	@Zed: http://ix.io/2kta
14:36:58	FromDiscord	<Zed> hey other zed :), i downloaded nim-mode from melpa, is it not included?
14:36:59	FromGitter	<Vindaar> @Zed: I'm using NIm with emacs without nimsuggest. Aside from a couple of cases where indentation is messed up it works well
14:37:07	dom96	silvernode, yes please
14:37:22	FromGitter	<Vindaar> @zedeus: does nimsuggest work well with emacs nowadays?
14:37:34	zedeus	yes, sometimes, sort of
14:37:38	FromDiscord	<Zed> ok you know whats creepy
14:37:42	silvernode	dom96: Oh ok, I should have it going in a few minutes here then. Will report back.
14:38:00	FromDiscord	<Zed> other zeds home directory is the same as mine
14:38:11	dom96	silvernode, please report such things on the issue tracker, I might not always be here to ask for details :)
14:38:12	FromDiscord	<Zed> are you my fbi agent?
14:38:17	zedeus	seeing a Zed here made me feel the same way, man
14:39:01	silvernode	dom96: I suspected the problem was just a known change between versions or something
14:39:18	FromDiscord	<Zed> lol, it's good to meet you, i have never met anybody else called zed before
14:40:08	zedeus	yeah, we're a rare breed
14:40:10	FromDiscord	<bedwardly-down> Maybe you two are twins from different parents? 😄
14:40:23	FromDiscord	<bedwardly-down> Or even clones? 😮
14:40:55	FromDiscord	<Zed> lol maybe
14:41:03	zedeus	what are you working on?
14:41:38	FromDiscord	<Zed> i am trying to build a ray tracer with nim and and set up bloody spacemacs to work with nim properly.
14:41:38	FromDiscord	<Zed> you?
14:42:07	dom96	when I see the name Zed I think of Zed Shaw. You're not that Zed right? :)
14:42:34	zedeus	oh cool. my main project is nitter.net but I'm working on a syncplay/cytube replacement (90% done), that supports mpv and a web client
14:42:46	FromDiscord	<Zed> no
14:42:52	FromDiscord	<Zed> i cna change my name if you want?
14:43:43	FromDiscord	<Zed> what is nitter.net?
14:43:54	zedeus	https://github.com/zedeus/nitter
14:44:12	zedeus	afaik the most popular Nim project besides the compiler
14:44:26	FromDiscord	<Zed> damn that's impressive
14:44:47	FromDiscord	<Zed> looks really clean as well!!
14:45:01	zedeus	thanks :) i'll be back later, food is calling
14:46:26	FromDiscord	<Zed> haha, sleep is calling me
14:46:52	zacharyc1rter	shashlick: nimterop isn't working for me with sokol anymore, the latest version
14:47:05	*	solitudesf joined #nim
14:47:21	zacharyc1rter	I saw all of the overrides that get generated and passed to toast, but I"m running into an error that one of the overridden types is undeclared
14:47:28	zacharyc1rter	will post an example later
14:49:08	FromDiscord	<Zed> hey @vindarr how have you got emacs setup?
14:55:08	FromGitter	<Vindaar> @Zed: pretty much close to default (regarding nim specific settings that is). Added a couple of minor modes, but other than that using just normal `nim-mode` and having set the `nimsuggest-path` to `""` via `M-x customize-variable`
14:55:40	FromGitter	<Vindaar> mostly rely on helm functionality to find stuff tbh
14:56:57	FromGitter	<Vindaar> my emacs settings are here: https://github.com/Vindaar/emacs.d/ ⏎ But probably rather check out @kaushalmodi's settings: https://github.com/kaushalmodi/.emacs.d
14:58:16	dom96	zedeus, got any nitter usage numbers you're willing to share? :)
14:59:52	FromDiscord	<Zed> awesome thanks, ill test them out and see what works
15:03:18	*	chapl joined #nim
15:04:00	*	chapl quit (Remote host closed the connection)
15:04:22	*	chapl joined #nim
15:16:11	*	filcuc_ joined #nim
15:18:30	*	fredrikhr quit (Read error: Connection reset by peer)
15:18:55	*	fredrikhr joined #nim
15:20:24	shashlick	@zacharycarter sounds good, will fix once we have some detail
15:21:47	FromDiscord	<bedwardly-down> I'm live at https://www.twitch.tv/bedwardly_down . I'm not really a streamer, but I'm trying to get back into it.
15:25:58	shashlick	@zacharycarter latest version of nimterop or sokol
15:26:04	zacharyc1rter	shashlick both
15:26:17	shashlick	Head?
15:26:23	shashlick	Of nimterop
15:26:24	zacharyc1rter	whatever nimble installs
15:26:37	shashlick	Ok that's old
15:26:58	shashlick	Then it sounds like a sokol change that's breaking
15:27:06	shashlick	But 0.5.0 is coming very soon
15:41:11	zedeus	dom96: sure, according to nginx, since the 12th of april there has been 400k unique visitors, 36m requests
15:42:00	*	FromGitter quit (Remote host closed the connection)
15:42:18	dom96	zedeus, any DAU/MAU numbers?
15:42:20	*	FromGitter joined #nim
15:43:14	*	setenforce joined #nim
15:54:49	*	ChantalZale joined #nim
15:55:12	*	ChantalZale quit (Client Quit)
15:56:30	*	setenforce quit (Quit: WeeChat 2.8)
15:57:50	*	Lord_Nightmare quit (Quit: ZNC - http://znc.in)
15:58:59	FromDiscord	<Broadwell> I'm trying to write unit tests for my macros, and it seems that I cannot (1) run unittest at compile time or (2) manipulate NimNodes at runtime. Are there any tools to facilitate this?
15:59:13	*	sagax quit (Excess Flood)
16:02:27	zedeus	dom96: about 15k daily
16:04:21	FromDiscord	<Recruit_main707> https://play.nim-lang.org/#ix=2ktM is that go code the same as the nim one? (both in there)
16:05:58	zedeus	maybe. depends on what "Unique visitors per day" means
16:06:24	*	Lord_Nightmare joined #nim
16:06:25	dom96	zedeus, I guess you don't have Google analytics on it?
16:06:31	zedeus	indeed
16:10:58	*	fredrikhr quit (Read error: Connection reset by peer)
16:11:23	*	fredrikhr joined #nim
16:17:03	*	filcuc_ quit (Ping timeout: 260 seconds)
16:27:52	zedeus	and ofc that doesn't include usage stats from the other 10 instances, I know some of them also get a nice amount of traffic
16:28:28	*	ptdel joined #nim
17:01:28	*	fredrikhr quit (Read error: Connection reset by peer)
17:01:55	*	fredrikhr joined #nim
17:14:47	*	ftsf quit (Ping timeout: 260 seconds)
17:19:36	*	filcuc_ joined #nim
17:25:59	*	NimBot joined #nim
17:34:20	disruptek	broadwell: run your ct tests in a static: in your unittest file or w/e.
17:36:24	FromDiscord	<bedwardly-down> How types and variables of those types work with nim is definitely not something I've seen before. I'm going to have to stew on them a bit
17:36:47	FromDiscord	<bedwardly-down> Not terrible but definitely different to me
17:44:28	zacharyc1rter	hmm - so I'm importing a module in my host app and my plugin and now I get errors about a class being implemented in both places
17:44:55	disruptek	nodecl?
17:45:08	zacharyc1rter	well the module being imported is using nimterop
17:46:03	zacharyc1rter	actually I'm importing a module that's importing a module that uses nimterop
17:46:12	zacharyc1rter	in the plugin
17:46:30	zacharyc1rter	I tried using `import except` but that doesn't solve the problem either
17:46:58	zacharyc1rter	but yeah maybe I can just redefine the type with nodecl
17:50:10	*	fredrikhr quit (Quit: Client Disconnecting)
17:52:21	shashlick	any details?
17:52:54	zacharyc1rter	I have a single nimterop module that creates bindings for `sokol_app.h` and `sokol_gfx.h`
17:53:22	zacharyc1rter	I used one module because I need to compile the implementations for these header files - which are in the header files
17:54:16	shashlick	are both pulling in the same header
17:54:23	zacharyc1rter	so in the nimterop module i use `cCompile` and I have a `.m` file which includes those two header files and defines some preprocessor directives like `#define SOKOL_IMPL` so that the implementations are included when I compile the header files
17:54:54	zacharyc1rter	I included both APIs in a single nimterop module, because I only want to define `SOKOL_IMPL` once and breaking them up into two nim modules seems to break everything
17:55:18	zacharyc1rter	here let me commit and I can point you to the relevant files
17:56:07	zacharyc1rter	https://github.com/zacharycarter/frag/blob/master/lib/sokol.nim
17:56:09	shashlick	but what's the actual error
17:56:31	zacharyc1rter	the way I have it now I don't get an error from nimterop
17:56:58	zacharyc1rter	the problem is when I import a module in my plugin code that imports this module which uses nimterop
17:57:06	zacharyc1rter	I get an error about the same class being implemented twice
17:57:28	zacharyc1rter	I'd break the nimterop module into two modules and only import the one I need - except them the nimterop module fails to compile
17:57:34	shashlick	output
17:57:41	zacharyc1rter	one sec I need to redo splitting it
17:57:44	zacharyc1rter	will take a moment
17:57:48	shashlick	no problem
18:03:45	zacharyc1rter	I think it was user error - I wasn't defining one of the preprocessor directives in both source files and I needed to - now it's compiling
18:05:50	shashlick	okay cool
18:05:56	zacharyc1rter	thanks for offering to help
18:06:02	shashlick	you're still using v0.4.4 right? not nimterop@#master
18:06:09	zacharyc1rter	0.4.3 I think
18:06:20	zacharyc1rter	I'll try with master later today
18:06:27	shashlick	okay big release coming soon
18:06:31	zacharyc1rter	and report any issues I encounter
18:06:34	shashlick	need to publish the changelog
18:06:34	zacharyc1rter	okay I'll just wait for that :)
18:08:55	*	krux02 joined #nim
18:10:45	*	jwm224 quit (Quit: WeeChat 2.8-dev)
18:13:06	*	filcuc_ quit (Ping timeout: 272 seconds)
18:14:21	federico3	https://github.com/nim-lang/Nim/wiki/BuildServices updated
18:29:38	FromDiscord	<Broadwell> disruptek: https://play.nim-lang.org/#ix=2kuP
18:33:51	FromGitter	<Vindaar> shashlick: just tried out nimterop for the first time. Have to say I'm *very* impressed!
18:35:59	*	lain quit (Quit: brb)
18:40:26	disruptek	broadwell: you cannot use unittest for it; just write your tests differently. you can also check a result value at runtime after storing it in a const, etc.
18:41:13	zacharyc1rter	first example works with hot reloading now - https://github.com/zacharycarter/frag/blob/master/examples/00-minimal/minimal.nim - on osx anyway
18:49:29	*	chapl quit (Ping timeout: 272 seconds)
18:52:47	*	lain joined #nim
18:53:17	FromDiscord	<KingDarBoja> Hi guys
18:53:44	*	disruptek quit (Read error: Connection reset by peer)
18:54:05	*	disbot quit (Read error: Connection reset by peer)
18:54:59	*	disbot joined #nim
18:55:59	*	disruptek joined #nim
19:04:52	*	chapl joined #nim
19:05:35	*	chapl quit (Remote host closed the connection)
19:05:57	*	chapl joined #nim
19:09:54	*	chapl quit (Remote host closed the connection)
19:12:39	*	chapl joined #nim
19:25:37	FromDiscord	<Weaponized Lego> Hey guys. I'm s little curious about nim, i wanted to pick up a novelty language as i haven't really had the need to pick up a second one as JavaScript and node.js pretty much solves all my problems right now.
19:25:37	FromDiscord	<Weaponized Lego>
19:25:37	FromDiscord	<Weaponized Lego> What are some ideas / problems where nim can be applied? I prefer to work on issues presented and work from there rather than video tutorials.
19:26:10	FromDiscord	<clyybber> just work an what you want. I for example use it for gamedev
19:26:36	FromDiscord	<clyybber> its a general purpose language though, so you should be able to use it for pretty much everything
19:26:44	FromDiscord	<clyybber> from webdev to embedded
19:27:24	FromDiscord	<Weaponized Lego> I'll start with data manipulation then, I have a few CSV files I need to process
19:29:54	FromDiscord	<KingDarBoja> Welcome Zoiberg!
19:30:01	*	FromDiscord <KingDarBoja> I love your profile pic
19:30:08	*	def- quit (Quit: -)
19:30:16	FromDiscord	<KingDarBoja> And yeah, pretty much you can use it for whatever you want
19:30:19	*	def- joined #nim
19:30:42	FromDiscord	<KingDarBoja> Someone said it was the lego language as you can build powerful stuff using macros and all that
19:33:15	*	chaplchapl joined #nim
19:33:25	*	chaplchapl quit (Client Quit)
19:37:08	FromDiscord	<KingDarBoja> Arrghh I lost the example of nested object kinds, anyone can share it?
19:42:12	FromDiscord	<KingDarBoja> Nvm I think I got it
19:42:23	planetis[m]	Imo, case object syntax is fine as is. Biggest nuisance is not being able to define same object field for different branches. Can this ever be fixed?
19:43:10	leorize	currently it's fixable with a macro, as long as you different branches share the same type for that field
19:44:13	leorize	there were proposed syntax but I think no one was bothered enough to actually work on it
19:44:14	*	nsf joined #nim
19:45:17	*	idf quit (Ping timeout: 260 seconds)
19:45:39	planetis[m]	Does patty implements it?
19:45:51	planetis[m]	Or any other macro
19:46:17	zedeus	patty does not, but you can use the intended names with `match`
19:47:30	zacharyc1rter	stream is back online - https://www.twitch.tv/zachary_carter
19:47:47	FromDiscord	<Varriount> @Zachary Carter Are you on Mumble?
19:47:56	FromDiscord	<Varriount> Or zacharyc1rter
19:49:14	FromDiscord	<KingDarBoja> https://github.com/nim-lang/RFCs/issues/19
19:49:44	FromDiscord	<KingDarBoja> leorize: too bad you have to use the same type on the same field for different branches :/
19:54:50	planetis[m]	Oh my god, idont care about the syntax i want that 🥺
19:56:20	*	silvernode quit (Ping timeout: 272 seconds)
19:57:58	*	inv2004 joined #nim
19:58:29	inv2004	Hello, is it possible to point nimble requires to local directory ?
19:59:04	FromDiscord	<Never Listen To Beef> Why would you want to?
19:59:09	zedeus	run `nimble develop` in the directory then add the package name to your .nimble
19:59:44	*	Trustable quit (Remote host closed the connection)
20:01:28	*	Trustable joined #nim
20:03:25	FromDiscord	<mratsim> inv2004: export NIMBLE_DIR=path/to/local/nimbledir
20:03:59	FromDiscord	<mratsim> ah sorry misunderstood, do as zedeus said
20:06:04	inv2004	`nimble develop` in package directory I am going to require ?
20:06:17	zedeus	yes
20:07:12	leorize[m]	KingDarBoja: it's a reasonable limitation
20:07:18	leorize[m]	same field of the same type with different type per branch is a nightmare
20:08:35	FromDiscord	<KingDarBoja> Still I like the suggestion of 4raq of shared fields, which seems to fit my need
20:08:39	leorize	no strongly typed language have return type overloading for a good reason
20:09:23	FromDiscord	<KingDarBoja> Getting used to the strong typing, I like it
20:09:38	FromDiscord	<Never Listen To Beef> IIRC dont you come from python?
20:09:47	FromDiscord	<Never Listen To Beef> That's strongly typed 😛
20:09:59	FromDiscord	<KingDarBoja> I come from Python lol
20:10:09	FromDiscord	<KingDarBoja> I do TypeScript but ain't the same at all
20:11:20	FromDiscord	<Never Listen To Beef> I come from C# so i have the type system down, the big issue i have is C# is a heavy OOP language and nim is certainly not
20:11:29	FromDiscord	<KingDarBoja> right
20:11:54	FromDiscord	<Never Listen To Beef> *that and compile time vs runtime*
20:12:26	FromDiscord	<KingDarBoja> TS is more like C#
20:12:34	FromDiscord	<KingDarBoja> But ofc has the JS things
20:12:46	FromDiscord	<Never Listen To Beef> Yea i know what TS is, not that i've ever touched it
20:13:00	FromDiscord	<KingDarBoja> That's why they said "Composition oevr Inheritance"
20:13:09	FromDiscord	<KingDarBoja> Just saying 🙂
20:13:32	leorize	I come from pascal, which has a lot of concepts in common with Nim
20:13:50	*	vqrs quit (Ping timeout: 246 seconds)
20:13:52	leorize	I'd say that makes Nim really easy to pick up for me
20:14:09	FromDiscord	<Never Listen To Beef> I tend to write nim code like i would in C# then realize that C#'s OOP has more features so then have to reevaluate
20:14:29	FromDiscord	<KingDarBoja> Yeah, if you go OOP in Nim, you need to be careful with child types
20:14:41	Yardanico	well nim OOP has it's advantages too
20:14:44	Yardanico	like multimethods
20:14:55	FromDiscord	<Never Listen To Beef> I mean my biggest gripe is runtime interfaces, since im used to having runtime mutable things
20:14:58	leorize	multimethods is deprecated
20:14:59	FromDiscord	<Never Listen To Beef> Yea but it's foreign to me
20:15:10	*	Hexeratops joined #nim
20:15:39	*	vqrs joined #nim
20:16:04	*	Senketsu joined #nim
20:16:22	inv2004	next question: the requires from nimble tries to compile the develop package, but cannot find .h header
20:16:26	FromDiscord	<KingDarBoja> I hate casting everytime I need to get a property of a child type as the code on cimpile time will thell me it doesn't exist
20:16:30	inv2004	Locally everything is ok in the package
20:16:40	FromDiscord	<KingDarBoja> But I know it does but it is using the base type
20:16:54	leorize	casting? you means converting, right? :P
20:16:59	*	Senketsu quit (Client Quit)
20:17:38	FromDiscord	<KingDarBoja> Not really `cast` but doing `childType(myVarW)`
20:17:45	leorize	yea
20:17:48	FromDiscord	<KingDarBoja> 😛
20:17:50	Yardanico	well that's very different from "casting"
20:17:59	FromDiscord	<Never Listen To Beef> Coming from C# you're wrong 😛
20:18:20	*	Senketsu joined #nim
20:18:38	leorize	the C family of language bundles all kind of conversion into one thing
20:19:13	leorize	iirc only C++ have different operators for different kinds of conversions
20:20:39	FromDiscord	<KingDarBoja> As `myVarW` is truly `childType` butbeing recognized as the base type due to proc call
20:20:44	FromDiscord	<Never Listen To Beef> Yea idk why you get a compiletime error, i know my ECS converted components to their true type seemlessly
20:21:11	Yardanico	Araq: thanks, with #14159 fixed nimes works with sink inference on, although nimes for some reason generally doesn't perform that well with arc (it's like 2-3x slower than with default gc)
20:23:26	Yardanico	although 70% of the time is still spent in __strstr_sse2 from libc
20:24:33	FromDiscord	<KingDarBoja> @Never Listen To Beef too lazy to explain now
20:25:01	leorize	note that Nim's proc use static binding, so "seamless" conversion isn't a thing here
20:25:35	FromDiscord	<KingDarBoja> Pretty much what leorize said
20:25:49	FromDiscord	<KingDarBoja> Doing the object variant conversion now
20:26:05	*	Senketsu quit (Quit: WeeChat 2.8)
20:29:34	FromDiscord	<mratsim> you are supposed to use methods on inherited stuff
20:29:52	Yardanico	oh seems like strstr is called by isObj
20:30:37	Yardanico	hmm maybe it due to nimes using inheritance for different mappers, lemme try without them
20:32:25	leorize	shashlick: #14180 is merged :)
20:33:35	Yardanico	lol wtf
20:33:39	Yardanico	I was right
20:33:42	*	sagax joined #nim
20:33:53	Yardanico	after removing inheritance for mappers it now performs like 3x better (same as with default gc)
20:34:43	FromDiscord	<7815> if i'd want to statically link with sdl2, what do i do?
20:35:05	leorize	what sdl2 wrapper are you using?
20:35:17	FromDiscord	<7815> the "standard one"
20:35:26	leorize	nim-lang/sdl2?
20:36:14	FromDiscord	<7815> the one that comes with `nimble install sdl2`
20:36:39	leorize	then --dynlibOverride:SDL --passL:"<flags for linking with sdl2 here>"
20:36:49	FromDiscord	<7815> very thanks
20:38:46	*	lritter quit (Quit: Leaving)
20:39:44	Yardanico	Araq: sorry for pinging again, so is it expected that --gc:arc performs much slower with ref objects with inheritance than default GC? in nimes it's a hot path since for PPU and for some CPU addresses it has to call a [] method for Mapper (so for one of its inherited objects)
20:40:04	Yardanico	a lot of time is spent in strstr (which is called from isObj) with that
20:40:33	Yardanico	without inheritance it performs just as well as in the default gc (I don't have specific benchmarks, but 5fps vs 30 fps can be noticed by a naked eye :D)
20:42:06	Yardanico	although I guess it won't be hard to fully remove inheritance for mappers from nimes
20:42:45	FromDiscord	<Never Listen To Beef> Well leorize my components were just data holder so i was accessing the properties on child not using procs on them
20:43:15	FromDiscord	<KingDarBoja> Hey guys, how I can call a type which is going to be declared later on the object variant?
20:43:28	Yardanico	wdym "call a type"?
20:43:36	Yardanico	well you can have recursive dependencies in types if you declare them in the same type block
20:43:55	FromDiscord	<KingDarBoja> I am using the same type block
20:43:57	FromDiscord	<KingDarBoja> But hold on
20:43:58	leorize	Yardanico: I think it's kinda unexpected. reading refs_v2 it seems that it's expected that the performance will be the same
20:44:22	Yardanico	leorize: well yeah I was initially suprised too, literally 69% of the time was spent in __strstr_sse2 from libc
20:45:10	FromDiscord	<KingDarBoja> https://imgur.com/VprFPUu
20:45:11	Yardanico	in C there are lines like "if (!((m) && (isObj((*m).Sup.m_type, "|nimes.mapper3.Mapper3:ObjectType|nimes.types.Mapper:ObjectType""
20:45:23	Yardanico	and isObj does T1_ = strstr((*obj).name, subclass);
20:45:37	leorize	turns out pointer tracing is still faster than substring search :P
20:45:47	Yardanico	@KingDarBoja then you'll have to declare these types separately
20:45:56	Yardanico	you couldn't do what you showed in a screenshot
20:46:04	Yardanico	can't*
20:46:15	FromDiscord	<KingDarBoja> Ugh
20:46:31	leorize	KingDarBoja: you can't enforce which variant goes where
20:46:44	leorize	at least you can't do that in the type system
20:47:16	FromDiscord	<KingDarBoja> So I must declare them separately before using
20:47:23	FromDiscord	<KingDarBoja> On a separate object variant type
20:47:57	leorize	your type system is creepily complex
20:48:10	Yardanico	leorize: it's a port from JS afaik
20:48:15	Yardanico	of graphql
20:48:20	FromDiscord	<KingDarBoja> ye
20:48:29	FromDiscord	<KingDarBoja> It works with OOP as I mentioned
20:49:04	*	chapl quit (Quit: Leaving)
20:49:21	FromDiscord	<KingDarBoja> Just seeing if I can replace the OOP approach with the object variants
20:49:27	leorize	well your system is just too complex
20:49:31	leorize	simplify it to the barebones
20:50:39	leorize	conversion from inheritance to variants requires rethinking some of the approach
20:51:09	leorize	don't be overspecific, be "just specific enough"
20:51:39	FromDiscord	<Rika> its really hard for me to convert inheritance to variants
20:51:54	Yardanico	just think in the nim way :P
20:52:01	FromDiscord	<KingDarBoja> How is the Nim way?
20:52:04	leorize	yea, you don't "convert"
20:52:10	FromDiscord	<Rika> it'd be easier with the "same field same type" thing for 2 different variants
20:52:35	FromDiscord	<KingDarBoja> I am once again requesting your help 😄
20:52:38	leorize	I can help with that
20:52:41	FromDiscord	<Rika> like, much easier for me
20:52:44	FromDiscord	<KingDarBoja> Hold on
20:52:46	leorize	"same field same type" thing
20:53:12	FromDiscord	<KingDarBoja> https://github.com/graphql/graphql-js/blob/master/src/language/ast.js#L135 scroll down and see what I can't see
20:53:41	FromDiscord	<Rika> leorize: see only "Individual" and "Overall"
20:53:41	FromDiscord	<Rika> https://cdn.discordapp.com/attachments/371759389889003532/706247061271740487/unknown.png
20:53:44	FromDiscord	<KingDarBoja> So far I made it with OOP but the drawback is that my proc calls are using the base type for some of them
20:54:05	Yardanico	@KingDarBoja I'm not saying it's easy, but do you really think that if Nim's own AST can be constructed with object variants, graphql's AST can't? :)
20:54:08	FromDiscord	<Rika> `column` and `column count` are intentionally (meant to be) different
20:55:30	FromDiscord	<KingDarBoja> Wdym Yard?
20:55:45	*	FromDiscord <KingDarBoja> reading it slowly
20:56:11	FromDiscord	<KingDarBoja> I mean, it can be, just probably lacking the thinking
20:56:27	leorize	Rika: yea I have an idea
20:56:56	FromDiscord	<Rika> other than renaming the two fields? ;;
20:57:02	*	Vladar quit (Quit: Leaving)
20:57:19	FromDiscord	<Rika> maybe ill just take the "rename the two fields" or smth
20:57:38	leorize	Rika: use genSym to generate a field that can't be accessed normally, then generate an accessor template that yield it for the fields you'd like
20:57:57	FromDiscord	<Rika> oh yeah, macros exist
20:58:01	FromDiscord	<Rika> what a big brain
20:58:10	FromDiscord	<Rika> forgive me for i am a brainlet
20:58:31	leorize	I've seen no one actually attempted that path though
20:58:33	FromDiscord	<Rika> ill prolly even just do it manually...
20:58:49	Araq	accessor templates are totally a thing
20:59:27	*	chapl joined #nim
21:01:26	FromDiscord	<7815> can you even link sdl2 statically? i'm getting tons of `__declspec(dllimport)` errors
21:01:50	leorize	yes you can
21:02:07	leorize	how did you link and what's the full output?
21:02:16	Araq	you can also just ship software that consists of more than a single file though
21:02:44	Araq	static linking is the new cargo cult
21:03:05	FromDiscord	<7815> `--dynlibOverride:SDL2 --passL:-lSDL2`
21:03:05	FromDiscord	<7815> libSDL2 is in my directory
21:03:11	FromDiscord	<7815> `--dynlibOverride:SDL2 --passL:-lSDL2`
21:03:11	FromDiscord	<7815> libSDL2 is in my cwd
21:03:18	Araq	number of software I use daily consisting of a single file: 0
21:03:20	FromDiscord	<7815> `--dynlibOverride:SDL2 --passL:-lSDL2`
21:03:20	FromDiscord	<7815> libSDL2.a is in my cwd
21:03:58	FromDiscord	<7815> araq, i don't think you forcing your ideals on others is a helpful trait for anybody
21:04:09	Araq	I'm saving your time
21:04:19	Araq	and that's helpful for anybody
21:04:54	FromDiscord	<7815> dependencies are certainly fun to deal with
21:05:11	Araq	copy sdl.dll next to your .exe, done.
21:05:13	*	Senketsu joined #nim
21:06:44	FromDiscord	<7815> https://hastebin.com/eyinoxurex.shell
21:06:58	inv2004	zedeus, Do you know what to do with C-header from develop package ?
21:07:51	FromDiscord	<Rika> how do i make a fail on compile time?
21:07:57	Yardanico	{.error.}
21:08:09	FromDiscord	<Rika> okay thanks
21:08:11	Yardanico	!eval {.error: "no".}
21:08:12	NimBot	Compile failed: /usercode/in.nim(1, 8) Error: no
21:08:23	Yardanico	@7815 well if you cross-compile for windows you should have a static version of SDL2 for windows
21:08:34	Yardanico	also you need to pass --os:windows
21:09:40	FromDiscord	<7815> it links correctly with my zig project
21:09:46	Yardanico	well nim is not zig
21:09:52	Yardanico	you should also tell nim that you're compiling to windows
21:10:08	Yardanico	and in the paste you only tell zig to compile to windows, and nim still thinks you're compiling for linux
21:11:12	FromDiscord	<7815> `--os:windows` doesn't stop it from compiling `src\timer\windows\SDL_*.c`
21:12:29	FromDiscord	<7815> `--os:windows` doesn't give `src\timer\windows\SDL_*.c` its necessary link symbols...
21:12:52	Yardanico	Araq: would you like a relatively short code example for inheritance where --gc:arc is ~5x slower?
21:13:11	Araq	Yardanico, ok
21:13:13	inv2004	how to include imported c-header into package?
21:14:26	Yardanico	Araq: https://gist.github.com/Yardanico/61384b0fe3a7913eb231cc0c9d3b996a
21:14:35	Yardanico	the slowest part there is ppu.mapper.step()
21:14:48	Yardanico	(for arc)
21:15:38	Yardanico	although it's not so high priority since I think it can be reworked to avoid methods
21:15:52	Araq	any idea why it's 5x slower?
21:16:01	Yardanico	Araq: because it calls isObj for each step call
21:16:06	Yardanico	and isObj calls strstr
21:16:49	Yardanico	inv2004: it should work by default, is your header in the src folder too?
21:16:54	Yardanico	Araq: https://i.imgur.com/MdMqIki.png there
21:17:03	shashlick	@Vindaar: glad to hear
21:17:08	shashlick	what were you wrapping
21:17:13	shashlick	and which version of nimterop
21:17:13	Yardanico	and full c src https://gist.github.com/Yardanico/6173219bfd29544087d055650e13388d
21:17:36	inv2004	@Yardanico, yep, here it is https://github.com/inv2004/kdb_nim
21:17:51	inv2004	I just removed bin in nimble if it matters
21:18:00	Araq	Yardanico, 'won't fix'? :P
21:18:16	Yardanico	Araq: well I'm really not sure if that's intended behaviour or not :D
21:18:21	Araq	but hmm my suspicion was that isObj is fast enough in practice
21:18:33	Araq	strstr vs pointer chasing
21:18:41	Yardanico	well nimes is using methods in an a bit surprising way
21:18:59	Yardanico	for each PPU step or some CPU step it calls a method of a mapper (NES roms can have different mappers)
21:19:00	Araq	though the old implementation also was tuned to remember the hot path
21:19:06	Yardanico	oh
21:19:40	Araq	maybe it's time to embrace VTables
21:20:04	FromDiscord	<KingDarBoja> leorize, u around?
21:22:00	Yardanico	inv2004: read https://forum.nim-lang.org/t/2668
21:22:05	FromDiscord	<Rika> man this new template thing idea works SWELL
21:22:09	Yardanico	https://forum.nim-lang.org/t/2668#22111
21:22:58	inv2004	@Yardanico, thx
21:23:01	Yardanico	also you don't have to do importc, header for every proc
21:23:11	Yardanico	you can just do {.push importc, header: "yourheader.h".}
21:23:24	Yardanico	importc by default uses the same nim proc name for the relevant C function
21:24:48	FromDiscord	<mratsim> Before embracing VTables I want private imports.
21:24:52	Yardanico	:P
21:25:03	FromDiscord	<mratsim> Having proc public just for testing is so so annoying when refactoring
21:25:20	Araq	so write an include file for testing
21:26:10	Araq	"I need X to be public for testing" is so far among my top ten of the least convincing issues ever brought up
21:29:19	FromDiscord	<mratsim> As I said, include hides generic bugs
21:30:39	FromDiscord	<mratsim> This bug is caught with import but not with include: https://github.com/nim-lang/Nim/issues/6387
21:30:58	FromGitter	<Vindaar> @shashlick: I wanted to call a function from the GNU scientific library and thought nimterop might make that easier. It certainly did.
21:31:02	FromGitter	<Vindaar> https://github.com/Vindaar/gsl-nim
21:31:38	FromGitter	<Vindaar> I'll turn that into a full wrapper when I have a better idea on how to properly use nimterop.
21:32:34	FromDiscord	<Rika> ???
21:32:34	FromDiscord	<Rika> https://cdn.discordapp.com/attachments/371759389889003532/706256845517619212/unknown.png
21:32:39	FromDiscord	<Rika> oops
21:32:46	FromDiscord	<Rika> it says "expected float"
21:33:07	Yardanico	show errors from the nim compiler itself :)
21:33:13	Yardanico	when compiling
21:33:13	FromDiscord	<mratsim> template + return is a recipe for bugs
21:33:22	Yardanico	ah you're in a template
21:33:26	FromDiscord	<Rika>
21:33:26	FromDiscord	<Rika> https://cdn.discordapp.com/attachments/371759389889003532/706257064120418334/unknown.png
21:33:27	FromDiscord	<mratsim> you should use a proc, otherwise you will return from the caller
21:33:27	*	tane quit (Quit: Leaving)
21:33:35	Yardanico	you can't call return in templates
21:33:38	FromDiscord	<Rika> i know
21:33:42	Yardanico	it'll be inserted as-is to the original proc
21:33:57	FromDiscord	<Rika> but im unsure why it wouldnt give me the Duration w/o return
21:34:00	FromDiscord	<Rika> which is odd
21:34:01	Yardanico	you should remove "return" statements from the template
21:34:04	Yardanico	then it might work
21:34:06	FromDiscord	<Rika> it just "expects a discard
21:34:42	FromDiscord	<mratsim> remove the "return" and replace discard by `@[]`
21:34:43	FromDiscord	<Rika>
21:34:43	FromDiscord	<Rika> https://cdn.discordapp.com/attachments/371759389889003532/706257388843565107/unknown.png
21:34:51	FromDiscord	<Rika> hmm
21:35:13	FromDiscord	<Rika> seems good
21:35:19	FromDiscord	<Rika> ill runtime error the discard instea
21:35:26	Yardanico	you can compile-time error it
21:35:29	FromDiscord	<mratsim> also template chaining via method call syntax is buggy because one of the template is not properly triggered
21:36:02	FromDiscord	<Rika> compile time erroring it doesnt work i dont think
21:36:08	FromDiscord	<mratsim> i.e. holdEndTimes.anyIt might give you undeclared identifier
21:36:11	FromDiscord	<Rika> i tried it and it doesnt work
21:36:19	Yardanico	what does it say?
21:36:34	FromDiscord	<Rika> it just compile time errors no matter correct usage or not
21:36:37	FromDiscord	<mratsim> are you writing a Stepmania clone?
21:36:40	FromDiscord	<Rika> no
21:37:18	Yardanico	@mratsim https://github.com/de-odex/circa
21:38:08	FromDiscord	<Rika> i really havent committed in a while
21:38:22	FromDiscord	<Rika> because i've been traumatized by git too much
21:38:42	Yardanico	well I really see no issue with git even if you do spontaneous commits and stuff like that
21:38:49	Yardanico	just commit at the end of the day lol
21:38:59	Yardanico	it's still better than not commiting
21:39:01	FromDiscord	<Rika> i dont know, im not a fan of not properly using it
21:39:04	Yardanico	commits help you discover regressions and stuff
21:39:26	FromDiscord	<Rika> if there were a good sourcetree equivalent on linux that isnt gitkraken
21:39:37	FromDiscord	<Rika> i'd deffo fix it up right now
21:39:49	Yardanico	i don't think it's really hard to use git from terminal for simple commits
21:39:52	FromDiscord	<KingDarBoja> Having big brain time trying to figure out object variant simplification lol
21:39:53	Yardanico	VSCode has GitLens too btw
21:40:05	FromDiscord	<Rika> i can use TUI, not git TUI
21:40:13	FromDiscord	<Rika> git subcommand names make 0 sense to me
21:40:38	Yardanico	use vscode with gitlens plugin then
21:40:42	FromDiscord	<Never Listen To Beef> I tend to use git cli and then use vscode's git tools to ensure im not pushing useless files cause it's a nice easy method imo
21:40:52	FromDiscord	<Rika> will do, just installed
21:41:03	FromDiscord	<Rika> does it have gitflow stuff?
21:41:11	FromDiscord	<Rika> not that i cant do it myself
21:41:15	FromDiscord	<KingDarBoja> Gitlens is awesome imo
21:42:19	Araq	mratsim: I don't see the connection as the symbol binding rules within generics do not depend on private vs public
21:45:52	FromDiscord	<exelotl> "git subcommand names make 0 sense to me" saaaame
21:46:24	FromDiscord	<exelotl> been using git for like 7 years now and I still have to google how to do the most basic things
21:46:34	FromDiscord	<Rika> struggle of a non-english native i guess?
21:46:55	FromDiscord	<exelotl> nah I'm a native speaker
21:47:03	FromDiscord	<Rika> :monkaHmm:
21:47:31	FromDiscord	<exelotl> it's actually ridiculous how bad the interface is
21:49:49	*	zacharyc1rter quit (Ping timeout: 246 seconds)
21:50:20	*	zacharycarter quit (Ping timeout: 272 seconds)
21:55:38	FromDiscord	<KingDarBoja> lol
21:59:47	Araq	exelotl: but it's pretty simple, 'git checkout' does everything ;-)
22:00:32	Araq	for a system created by a guy who cannot stand GCs or overloading git's commands are surprisingly overloaded and git also has a GC...
22:00:38	FromDiscord	<Rika> i cant seem to stage a folder for some reason...
22:02:39	*	Senketsu quit (Quit: WeeChat 2.8)
22:05:58	FromDiscord	<Never Listen To Beef> I mean linus started git, he went pretty hands off rather quickly from my understanding
22:07:46	*	nsf quit (Quit: WeeChat 2.8)
22:08:30	FromDiscord	<Never Listen To Beef> No clue when the GC was implemented so my point is irrelevant yes
22:12:45	leorize	to be fair, linus is not that against a gc
22:12:52	leorize	the kernel have a gc itself
22:13:24	leorize	KingDarBoja: you called?
22:14:01	Araq	leorize, which kind of GC does the kernel have?
22:14:24	Araq	so ... number of showstopper bugs: 0 :-)
22:14:41	leorize	simple ref counting the last time I checked
22:14:55	Araq	leorize, doesn't count
22:15:16	Yardanico	Araq: new 1.2 release soon? :P
22:15:33	leorize	1.4 when? :P
22:15:52	Araq	Yardanico, who knows. maybe we should make 'koch boot --gc:orc' work and release 1.4
22:15:58	supakeen	gc:orc
22:16:00	supakeen	mrm
22:16:04	supakeen	i like it
22:16:10	FromDiscord	<KingDarBoja> Yeah, was going to ask something but I think I got it
22:16:15	FromDiscord	<KingDarBoja> The idea of simplify stuff
22:16:18	shashlick	@Vindaar: awesome - let me know when you need any help with nimterop
22:16:35	shashlick	And feedback obviously to make it easier to get up to speed
22:17:02	FromDiscord	<KingDarBoja> Instead of specifying the node kind on some child, I just put the base type
22:17:10	FromDiscord	<Rika> haa, pushed to temp-all branch now
22:17:16	FromDiscord	<KingDarBoja> `Node` and let the proc handle the leaf kind 🤔
22:17:45	Yardanico	in the meantime I guess I'll fork nimes and replace inheritance with something else + use computedGoto for CPU opcodes to possibly speed it up even more (right now it uses an array of procedures, I guess that's already quite fast)
22:18:09	Araq	embrace computedGoto
22:18:23	Yardanico	yeah it's a very cool optimization
22:18:26	FromGitter	<Vindaar> @shashlick: Thanks, will do!
22:18:49	Yardanico	but why nim vm has it disabled? :P https://github.com/nim-lang/Nim/blob/devel/compiler/vm.nim#L543
22:19:13	leorize	Yardanico: have you reported the async bug with gc:orc for the irc module?
22:19:22	Yardanico	leorize: it turns out it's a bug with sink inference
22:19:28	Yardanico	not gc:arc or gc:orc themselves
22:19:32	Yardanico	it works with --sinkInference:off
22:19:43	leorize	so... did you report it? :P
22:20:14	Yardanico	no, actually lemme try to run it again, who knows, maybe it was solved by #14183
22:21:54	Yardanico	btw what happened with disbot? :P
22:22:01	Yardanico	it doesn't reference PRs now
22:22:09	leorize	it bugs out once in awhile
22:22:13	leorize	ping disruptek I guess
22:22:35	Araq	disbot was built with --gc:arc, that's why :-)
22:23:06	Yardanico	well it's async, doesn't it leak memory this way
22:23:31	Yardanico	also yeah, irc module doesn't work in async mode with gc:orc, works with sink inference disabled
22:23:35	Yardanico	although the error is different than before
22:23:46	Araq	well we'll soon build it with orc
22:24:20	leorize	guess you should report that bug now so we can tackle it
22:24:27	Araq	Yardanico, produce a reduced test case please, it really speeds up our development
22:24:35	Yardanico	well I'll try :P
22:24:41	Yardanico	irc module itself is ~700 loc
22:24:57	Araq	lol but async is a big lib
22:25:02	Yardanico	ah xd
22:25:06	Araq	anyhow, good night
22:25:09	Yardanico	good night
22:25:19	leorize	g'night
22:25:39	*	Hexeratops quit (Quit: Take care :))
22:26:05	leorize	why does fork() on linux have to be so ridiculously inefficient...
22:26:51	*	inv2004 quit (Read error: Connection reset by peer)
22:32:09	FromDiscord	<Rika> huh, what does "expression has no address" mean now...
22:33:04	leorize	that means you're trying to take an address of a const
22:35:29	FromDiscord	<Rika> but im not?
22:35:31	FromDiscord	<Rika> hm
22:36:15	leorize	can I see the code?
22:36:40	FromDiscord	<Rika> https://play.nim-lang.org/#ix=2kwB
22:36:57	FromDiscord	<Rika> error at `case`
22:37:25	FromDiscord	<Generic> that is an unfortunate short coming
22:37:25	leorize	can I have the typedef too?
22:37:59	FromDiscord	<Generic> https://github.com/nim-lang/Nim/issues/13033
22:38:01	leorize	though this is probably just due to the `var` thingy
22:38:33	leorize	try result =
22:38:36	leorize	should work
22:39:11	FromDiscord	<Rika> https://play.nim-lang.org/#ix=2kwD
22:39:31	FromDiscord	<Rika> bah, the result fixed it
22:40:09	FromDiscord	<Rika> i see
22:40:39	leorize	can you merge maniaIndividual and maniaOverall?
22:41:01	leorize	the diff seems to be the column thingy only :P
22:41:48	leorize	though I'm for certain don't understand how this code work or what it's used for
22:42:07	leorize	why does this look like an osu! parser :P
22:42:30	FromDiscord	<Rika> it is.
22:44:58	FromDiscord	<KingDarBoja> 😄
22:45:08	FromDiscord	<KingDarBoja> From desperation comes innovation
22:45:26	*	solitudesf quit (Ping timeout: 272 seconds)
22:46:54	leorize	well I still don't understand osu! beatmaps so I can't say much about this code :P
22:46:55	FromDiscord	<Rika> leorize: maniaIndividual and maniaOverall are totally separate
22:47:17	FromDiscord	<Rika> this isnt parsing the beatmap its the difficulty calculation code
22:47:29	FromDiscord	<Rika> im still having difficulty converting the heavy OOP of C#
22:47:55	leorize	I don't know how the game work is the point actually :P
22:48:21	FromDiscord	<Rika> i see
22:48:52	FromDiscord	<Rika> i still dont know if this code will work
22:49:50	Yardanico	ok I started minimizing it, now I managed to use expandMacros to get output of the async macro stuff
22:49:57	Yardanico	and yes I managed to compile it after removing tons of `gensym things
22:53:46	Yardanico	right now it's 600 lines but at least it only needs asyncdispatch for the dispatcher and asyncfutures for futures, now I need to start removing functionality :P
22:55:46	Yardanico	compiled C is 6.4k lines though
22:56:27	FromDiscord	<KingDarBoja> I am mixing OOP with object variants
22:56:30	FromDiscord	<KingDarBoja> and so far, so good
22:56:40	FromDiscord	<KingDarBoja> Using it to group leaf types
22:56:52	FromDiscord	<KingDarBoja> And OOP to keep parent types inherit from a single base
22:57:34	FromDiscord	<KingDarBoja> Also, I noticed I can't declared shared fields using multiple kinds on same case then another different for a specific field
22:57:37	FromDiscord	<KingDarBoja> 🤔
22:58:01	Yardanico	https://github.com/nim-lang/RFCs/issues/19
22:58:23	FromDiscord	<KingDarBoja> Yeah I saw that
22:59:21	*	chapl left #nim (#nim)
22:59:30	FromDiscord	<KingDarBoja> https://github.com/nim-lang/RFCs/issues/19#issuecomment-173529871 this comment is what I am looking for
22:59:34	FromDiscord	<KingDarBoja> But still RFC 😢
23:01:50	FromDiscord	<KingDarBoja> https://imgur.com/4met1K4 pretty much had to use a different name for that
23:02:33	FromDiscord	<KingDarBoja> But after running the test, I will get an error as I am trying to access `value` on `IntValueNode`
23:02:50	FromDiscord	<KingDarBoja> But the funny thing is that the terminal doesn't point me to the error :S
23:09:41	FromDiscord	<bedwardly-down> Not even with —verbose flag on?
23:09:55	FromDiscord	<KingDarBoja> Had no idea about that flag
23:09:57	FromDiscord	<KingDarBoja> Let me try
23:10:04	FromDiscord	<KingDarBoja> `nimble tests --verbose` ?
23:10:17	FromDiscord	<bedwardly-down> Should be
23:10:23	FromDiscord	<KingDarBoja> Nope
23:10:54	FromDiscord	<bedwardly-down> I’m not able to test but nimble —verbose c *somenimfile* works for me
23:11:12	FromDiscord	<bedwardly-down> That stinks.
23:11:40	FromDiscord	<KingDarBoja> I have defined it like `exec "nim c -r tests/language/test_parser"` on my packageName.nimble file
23:14:13	FromDiscord	<KingDarBoja> @bedwardly-down your approach works
23:14:15	FromDiscord	<KingDarBoja> Thank you 🙂
23:14:37	FromDiscord	<KingDarBoja> `Error: undeclared field: 'strValue' for type ast.NameNode` this is what I was looking for
23:15:17	*	pbb quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
23:15:26	FromDiscord	<bedwardly-down> Depending on what system you’re on, you could possibly create an alias for your shell where nim and nimble always run with that flag on when you call them
23:15:56	FromDiscord	<KingDarBoja> I use Windows
23:16:03	*	pbb joined #nim
23:16:09	FromDiscord	<KingDarBoja> All my tests are defined at the .nimble file with the above script
23:16:47	FromDiscord	<bedwardly-down> Using Powershell, you could possibly do the same. I’m just not sure on specifics there
23:19:01	FromDiscord	<KingDarBoja> using `nimble test --verbose` works lol
23:19:08	FromDiscord	<KingDarBoja> I had a typo
23:19:51	FromDiscord	<bedwardly-down> Did you put 1 - instead of 2?
23:22:49	*	zacharyc1rter joined #nim
23:22:49	*	zacharycarter joined #nim
23:23:49	FromDiscord	<KingDarBoja> It is 1 by default
23:32:33	*	Trustable quit (Remote host closed the connection)
23:42:25	*	stefantalpalaru quit (Changing host)
23:42:25	*	stefantalpalaru joined #nim
23:46:50	*	nim669 joined #nim
23:47:21	*	nim669 quit (Client Quit)
23:49:25	FromDiscord	<Rika> can i make an object variant based on a field of a field? i know the answer's pretty much no, but i'd like to know of alternatives
23:49:27	FromDiscord	<bedwardly-down> For -v in most apps yeah
23:49:39	Yardanico	?
23:49:54	FromDiscord	<bedwardly-down> @KingDarBoja
23:50:09	Yardanico	"object variant based on a field of a field" the only way to know that is to embed the other object variant in your own object variant if I understood you correctly
23:50:14	Yardanico	show code :P
23:50:17	FromDiscord	<KingDarBoja> Wait
23:50:36	FromDiscord	<KingDarBoja> Object variant based on a filed of a field ? Nested?
23:51:00	FromDiscord	<Rika> one moment
23:51:30	FromDiscord	<bedwardly-down> How deep can you even go with that anyway?
23:53:49	FromDiscord	<KingDarBoja> Deeper as long as you don't need previous object variant kinds
23:54:06	FromDiscord	<Rika> something like this? https://play.nim-lang.org/#ix=2kx0
23:55:23	Yardanico	yeah as I said the only way for that is to embed modeHitObject definition inside of your DifficultyHitObject
23:56:06	FromDiscord	<Rika> but theyre separate objects
23:56:21	Yardanico	then no
23:56:41	FromDiscord	<Rika> okay
23:57:56	*	krux02_ joined #nim