<< 06-01-2020 >>

---

00:19:07	*	krux02_ joined #nim
00:21:56	*	krux02 quit (Ping timeout: 248 seconds)
00:27:19	*	filcuc quit (Quit: Konversation terminated!)
00:27:31	*	filcuc joined #nim
00:36:48	*	muffindrake quit (Quit: muffindrake)
00:38:56	*	muffindrake joined #nim
00:55:43	*	filcuc quit (Quit: Konversation terminated!)
00:59:02	*	moduledge joined #nim
01:01:54	moduledge	Hi there folks. I got a question, would this be a bug when using sequtils.all() on an empty sequence? I was trying to solve an Exercism exercise with something like this and it didn't make sense to me that it would return `true`: https://play.nim-lang.org/#ix=26xf
01:03:13	moduledge	I'm open to be corrected as well
01:09:03	disruptek	because all members of the sequence pass the predicate.
01:14:14	*	krux02_ quit (Remote host closed the connection)
01:17:29	moduledge	Oh I see. It's true by default unless it finds a false in the sequence.
01:19:19	moduledge	and any() would do the opposite.
01:35:45	*	endragor quit (Remote host closed the connection)
01:36:20	*	endragor joined #nim
01:41:37	*	dwdv quit (Ping timeout: 265 seconds)
01:44:35	*	abm joined #nim
01:44:45	*	lritter quit (Ping timeout: 268 seconds)
01:45:00	*	lritter joined #nim
01:47:15	*	lritter quit (Client Quit)
01:56:57	*	endragor quit (Remote host closed the connection)
02:05:33	disruptek	yay, i got a [FATAL] unpaired dealloc
02:05:39	disruptek	now we're on the road to a solution.
02:07:32	*	chemist69 quit (Ping timeout: 252 seconds)
02:08:42	*	chemist69 joined #nim
02:15:05	*	abm quit (Read error: Connection reset by peer)
02:33:26	FromGitter	<perfecto25> Hello, noob question
02:33:38	FromGitter	<perfecto25> how do I append to a JSON object?
02:33:48	FromGitter	<perfecto25> or remove a key:val from a json obj
02:34:50	disruptek	https://nim-lang.org/docs/json.html -- add, []=, delete
02:45:20	*	endragor joined #nim
02:56:27	*	Hideki_ joined #nim
03:01:18	*	Hideki_ quit (Ping timeout: 260 seconds)
03:08:28	FromGitter	<gogolxdong> `/usr/bin/ld: /mnt/c/Nim/nimcache/r_linux_amd64/stdlib_intsets.nim.c.o: relocation R_X86_64_32 against undefined symbol `NTI__PQrmFm4BXFkFSVOUl1ojSg_' can not be used when making a PIE object; recompile with -fPIC ⏎ /usr/bin/ld: final link failed: Nonrepresentable section on output ⏎ collect2: error: ld returned 1 exit status`
03:08:36	FromGitter	<gogolxdong> latest devel
03:18:37	*	Jjp137 quit (Read error: Connection reset by peer)
03:19:11	*	Jjp137 joined #nim
03:25:09	*	arecaceae quit (Remote host closed the connection)
03:25:53	*	arecaceae joined #nim
03:49:55	*	endragor quit (Ping timeout: 258 seconds)
03:59:57	*	muffindrake quit (Ping timeout: 260 seconds)
04:01:43	*	muffindrake joined #nim
04:10:37	FromGitter	<perfecto25> thanks
04:10:58	*	uu91 joined #nim
04:11:55	FromGitter	<perfecto25> is there a way to pass a Table to a proc, ie, ⏎ ⏎ proc add_to_table (mytable: Table): Table = ⏎ ⏎ ``` return table.add("newkey", "newval")``` [https://gitter.im/nim-lang/Nim?at=5e12b38bbe6bab58cd7e814e]
04:28:25	*	endragor joined #nim
04:36:07	*	chemist69 quit (Ping timeout: 260 seconds)
04:37:54	*	chemist69 joined #nim
04:49:31	*	nsf joined #nim
04:49:39	disruptek	yes, but add() doesn't return a value. use a TableRef.
04:49:44	disruptek	~tables
04:49:45	disbot	https://nim-lang.org/docs/tables.html
05:04:49	muffindrake	Is there a flag you can pass to the nim compiler to have the companion C compiler generate CPU-specific code a la -march=native?
05:05:19	disruptek	yeah, you can pass whatever you want through to the native compiler.
05:07:18	muffindrake	You didn't answer my question
05:07:28	disruptek	sure i did.
05:08:11	muffindrake	My question wasn't 'can you pass arbitrary flags to the C compiler'
05:08:48	disruptek	you asked if there was a flag. in the same sentence, you gave example of such a flag.
05:11:06	disruptek	try `nim --fullhelp` and note the --passC (and --passL) switches.
05:12:27	muffindrake	I was asking if I could give a specific switch to the nim compiler that doesn't have me write build scripts for every possible C compiler that could potentially be used, because all of them have differently named flags for this feature
05:13:34	disruptek	in that case, it's unlikely. nim supports multiple backends, so the compiler options are geared towards broader goals such as "size", "speed", etc.
05:14:16	disruptek	even for, say, a c backend, nim supports multiple compilers (gcc, clang, vcc, tcc, mingw, etc.)
05:15:20	disruptek	you can specify flags for particular compilers in a single nim.cfg or config.nims, though.
05:16:39	disruptek	~compiler
05:16:40	disbot	compiler: 11https://nim-lang.org/docs/nimc.html -- disruptek
05:55:59	*	ltriant quit (Quit: leaving)
06:11:35	*	dddddd quit (Ping timeout: 268 seconds)
06:14:02	*	beshr quit (Read error: Connection reset by peer)
06:58:27	*	Hideki_ joined #nim
07:02:48	*	Hideki_ quit (Ping timeout: 260 seconds)
07:05:41	*	ikan-keli_2 joined #nim
07:05:56	*	ikan-keli_ quit (Ping timeout: 265 seconds)
07:06:59	*	narimiran joined #nim
07:16:42	*	matic joined #nim
07:19:15	*	marmotini_ joined #nim
07:31:50	*	marmotini_ quit (Remote host closed the connection)
07:32:17	*	marmotini_ joined #nim
07:32:40	*	marmotini_ quit (Remote host closed the connection)
07:32:47	*	marmotini_ joined #nim
07:37:06	*	endragor_ joined #nim
07:40:46	*	PMunch joined #nim
07:41:00	*	endragor quit (Ping timeout: 268 seconds)
07:46:17	*	ptdel quit (Ping timeout: 260 seconds)
07:46:41	*	solitudesf joined #nim
08:00:00	*	gmpreussner quit (Quit: kthxbye)
08:05:03	*	gmpreussner joined #nim
08:07:59	FromDiscord	<Skaruts> from testing it seems like the operator `==` works by default for `objects` but not for `ref objects`
08:08:03	FromDiscord	<Skaruts> why is that?
08:12:32	Zevv	you are comparing the refs
08:12:57	FromDiscord	<Skaruts> ah, I thought the comparison would go through the refs
08:13:00	Zevv	which are basically pointers. Two different pointers to two objects are never the same
08:13:12	Zevv	dereference your objects: let same == a[] == b[]
08:13:16	FromDiscord	<Skaruts> indeed
08:13:31	Zevv	PMunch: ping
08:13:45	FromDiscord	<Skaruts> I implemented my own `==` to make it easier
08:13:59	*	Trustable joined #nim
08:14:42	Zevv	Sure
08:14:50	Zevv	this is nim, after all
08:15:03	PMunch	Zevv, pong
08:15:05	Zevv	oi!
08:15:28	PMunch	What's up?
08:15:36	Zevv	You've been doing the atmel stuff lately, right?
08:15:41	PMunch	Yup
08:15:42	Zevv	Did you use heap memory there?
08:16:07	PMunch	I think I did some manual allocation and freeing
08:16:14	PMunch	But I didn't set up the garbage collector
08:16:23	Zevv	hold on, phone, sorry
08:20:23	Zevv	back
08:20:44	Zevv	yeah, I was thinking to get nim-in-linux-kernel space up and running with --gc:arc
08:20:59	Zevv	but with useMalloc and os:standalon allocs always happen in 32Mb chunks
08:21:05	Zevv	so I was wondering what you did to get small allocs to work
08:24:53	Zevv	is your stuff public somewhere?
08:25:41	PMunch	Unfortunately not..
08:25:46	PMunch	Yet at least
08:26:03	PMunch	But I had a look and it doesn't actually seem like I'm doing an alloc in that code
08:26:07	Zevv	right :)
08:26:16	PMunch	Must've been thinking of another project..
08:26:49	*	marmotini_ quit (Remote host closed the connection)
08:27:01	PMunch	But have a look at https://github.com/gokr/ardunimo, it even works with the GC
08:27:14	Araq	-d:useMalloc and --os:standalone in combination is untested
08:27:16	*	marmotini_ joined #nim
08:27:52	Zevv	well, I can test that. But at this time it uses an additional layer of allocators which does large chunks
08:27:58	Zevv	I do see why, the comments are clear enough about that
08:28:08	PMunch	What does -d:useMalloc actually do?
08:28:21	Zevv	It makes 'theHeap' and does allocs from that
08:28:48	Zevv	last time my workaround was adding my own os: to Nim, but I'd like to avoid that
08:29:00	Zevv	so we can run on generic standalone targets and use default malloc()/free()
08:29:16	Zevv	PMunch: ardunimo did the same as I did: For the moment Ardunimo needs a patched Nim, copy the alloc.nim file to replace the one in your Nim installation:
08:29:37	PMunch	Yeah I was thinking that that might be the case
08:29:38	Araq	I think the problem is --os:standalone
08:29:50	Araq	ARC with -d:useMalloc is fine and ARC was designed for it
08:30:03	Zevv	yes but still we don't have an os
08:30:21	Zevv	no main, and a limb libc stumbling on one leg
08:31:58	*	marmotini_ quit (Ping timeout: 265 seconds)
08:34:15	*	marmotini_ joined #nim
08:38:28	*	cyraxjoe quit (Ping timeout: 260 seconds)
08:41:35	*	cyraxjoe joined #nim
09:06:56	*	Vladar joined #nim
09:50:12	*	Trustable quit (Remote host closed the connection)
09:53:06	*	leorize joined #nim
09:57:46	PMunch	Been programming in C# for ~2 hours today, already missing Nim..
09:58:20	FromDiscord	<Skaruts> is there any possible way to make a nim file implicitly imported by every other file in the same project?
09:59:10	PMunch	Not without doing something funky in your Nimble file
09:59:51	PMunch	Set up a build task that adds the import statement to every file before compiling
10:00:14	leorize	--import
10:00:18	leorize	but please don't do that
10:00:27	PMunch	Oh there is a flag for it?
10:02:57	FromDiscord	<Skaruts> I have a bit of a beef with nim's for loops (I never hit the dots right), so I made an iterator that I could potentially use to make the ranges easier for me to type, but it kinda sucks to have to import a file with this everywhere
10:03:52	FromDiscord	<Skaruts> I have a bit of a beef with nim's for loops (I never hit the dots or the '<' right), so I made an iterator that I could potentially use to make the ranges easier for me to type, but it kinda sucks to have to import a file with this everywhere
10:05:36	PMunch	Why not just use countup from the standard library?
10:05:37	leorize	maybe you should try to get used to it? :p
10:06:02	PMunch	https://nim-lang.org/docs/system.html#countup.i%2CT%2CT%2CPositive
10:06:09	leorize	there's also countup/countdown fwiw
10:06:13	FromDiscord	<mratsim> if you have things that you often import just create a "prelude" file where you import/reexport every common things
10:06:35	FromDiscord	<mratsim> countup is annoying because you always have to do -1
10:06:38	leorize	you can just `include` it if it only contains import statements
10:07:16	FromDiscord	<Skaruts> countup isn't so practical to type for me either, tbh.
10:07:24	FromDiscord	<Skaruts> I'm not sure it would conflict with system range, but I made it so I just do `for i in range(100)` (and overloads for `min` and `step`), which is Godot's loops, and I type that in literally 2 secs
10:08:13	leorize	import everywhere doesn't seem too bad tbh
10:08:31	*	marmotini_ quit (Remote host closed the connection)
10:08:37	leorize	it lets people that read your code knows where something might be from
10:08:53	FromDiscord	<Skaruts> but it requires worrying about it, and I know I'll forget 99999999 times 😄
10:09:01	*	marmotini_ joined #nim
10:09:05	*	dwdv joined #nim
10:10:13	leorize	if you don't import it, it won't work so don't worry :P
10:13:38	*	marmotini_ quit (Ping timeout: 258 seconds)
10:20:21	FromDiscord	<kodkuce> hi if someone bored and have time, can you check/review my code, i am thinking of rewriting it from start with using some kind of pattern or something cuz duno i think its starting to look like mess, this is poker websocket game http://ix.io/26z4
10:21:29	FromDiscord	<kodkuce> basicly my main issue comes i think from bridging websockets with game logic, i think i should use some patter or duno, open for suggestions
10:23:07	FromDiscord	<kodkuce> meybe some kind of command pattern or something
10:23:08	*	krux02 joined #nim
10:34:15	Araq	Skaruts: write a 'prelude' include file that has the imports you seek. import+export is not as nice as that confuses where the symbol actually comes from IMO
10:34:50	FromDiscord	<Clyybber> Araq: Morning, to answer your question: I didn't find any additional flaws, but I think what cooldome said is right, this should exclude global vars.
10:35:19	FromDiscord	<Clyybber> Good news: I think I know what the bug in my one pass find all approach is 😄
10:38:19	Araq	for global vars we need to distinguish between
10:38:22	Araq	var x = value
10:38:24	Araq	and
10:38:27	Araq	var x: T
10:38:29	Araq	x = value
10:38:44	Araq	it's bad to pessimise globals too much, people use them quite a bit
10:38:57	Araq	even worse
10:39:03	*	Vladar quit (Remote host closed the connection)
10:39:21	Araq	move-only types lack the '=' and yet var nonMovable = foo() must work
10:40:31	FromDiscord	<Clyybber> why must it work?
10:40:38	FromDiscord	<Clyybber> I think we should leave globals alone
10:41:23	Araq	var x = value # this actually doesn't move nor assign, it constructs 'x' with 'value'
10:41:46	FromDiscord	<Clyybber> does it not invoke sink?
10:43:00	FromDiscord	<mratsim> constructors!
10:44:59	*	Vladar joined #nim
10:46:58	*	marmotini_ joined #nim
10:52:00	FromDiscord	<Skaruts> @Arak how does that work? I make the prelude file and include it in the main one?
10:52:15	*	marmotini_ quit (Remote host closed the connection)
10:52:22	*	marmotini_ joined #nim
10:58:16	FromDiscord	<Skaruts> nvm, saw it in karax's codebase, it has to be included everywhere
10:59:59	*	Hideki_ joined #nim
11:00:26	*	dwdv quit (Quit: quit)
11:01:31	FromDiscord	<Skaruts> does help though
11:02:26	FromDiscord	<Skaruts> is there any added cost to importing a file in parts where it's not needed?
11:03:27	*	Vladar quit (Ping timeout: 260 seconds)
11:04:30	*	Hideki_ quit (Ping timeout: 268 seconds)
11:07:50	*	moduledge quit (Quit: Leaving)
11:08:59	*	nsf quit (Quit: WeeChat 2.7)
11:22:19	*	pbb quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
11:22:33	*	pbb joined #nim
11:23:52	FromGitter	<mratsim> Slight compilation time
11:25:58	*	Gnjurac joined #nim
11:28:13	*	Gnjurac quit (Remote host closed the connection)
11:46:26	*	Hideki_ joined #nim
11:57:59	FromDiscord	<Skaruts> ah nice
12:02:39	Araq	"This is a rocket"
12:03:10	Araq	I suppose it does mean some secret benchmark got faster
12:03:16	Araq	https://github.com/nim-lang/Nim/pull/13021
12:03:17	disbot	➥ Continue #13002
12:03:41	FromDiscord	<Clyybber> heh
12:03:56	FromDiscord	<Clyybber> I think my new analysis will improve it a bit
12:04:27	Araq	as I said, now wasMoved+destroy elision should have prio number one
12:04:28	FromDiscord	<Clyybber> since it will consider lastRead - firstWrite pairs too
12:04:42	Araq	ah yeah
12:09:28	FromDiscord	<Skaruts> are there any future plans to allow parent folder imports like this `import ../[things, stuff, beer]`?
12:09:50	FromDiscord	<Clyybber> hmm, does it not work currently?
12:10:12	FromDiscord	<Skaruts> hmm, I upgraded nim a few days ago and didn't check
12:10:57	FromDiscord	<Clyybber> ah then it probably doesn't work (yet), I faintly remember an issue about it, or at least a comment
12:11:51	Araq	I think at least
12:11:57	Araq	import ".." / [a, b]
12:11:59	Araq	does work
12:12:58	FromDiscord	<Clyybber> deech: Can you report those issues you encountered with the parser? I think it would be a good idea to have an issue for them to track
12:13:04	FromDiscord	<Clyybber> or at least discuss
12:14:21	FromGitter	<alehander92> `.. / services / [event_log_service, debugger_service, editor_service, callgraph_service, history_service, flow_service]` work
12:14:36	FromGitter	<alehander92> hm but `../` as well
12:14:39	FromDiscord	<Skaruts> oh yea that works
12:14:47	FromGitter	<alehander92> i thought it might be the space
12:14:54	FromDiscord	<Skaruts> why are the quotes needed though?
12:15:21	FromDiscord	<Clyybber> maybe its the space between `/` and `[`
12:15:49	Araq	what you need to know (and apparently nobody does?) is that parsing 'import'
12:15:51	FromGitter	<alehander92> ah i just realized mine is not ../ [ directly sorry
12:15:56	Araq	has no special rules whatsoever
12:16:26	Araq	it's just import <expr> (',' <expr>)*
12:16:29	FromGitter	<alehander92> yeah so `/[` is probably an error everywhere in import
12:16:33	FromDiscord	<Skaruts> this doesn't work though: `import ".."/[sfmlw, tilemap, asset_manager as am]`
12:16:35	FromDiscord	<Skaruts> `Error: undeclared identifier: 'am'`
12:17:01	Araq	but that's the parsing step, it's why sometimes you need ""
12:17:13	FromGitter	<alehander92> put spaces around `/` imho
12:17:16	Araq	the other side of the coin is then how we interpret the expression
12:17:23	FromGitter	<alehander92> hm nvm
12:17:44	Araq	and that intepretation has grown and supported more and more features
12:18:49	PMunch	Ugh, why is everything so convoluted in C# https://stackoverflow.com/questions/1500194/c-looping-through-lines-of-multiline-string
12:19:00	PMunch	What's wrong with for line in mystring.lines..
12:19:26	FromDiscord	<Skaruts> wanna know why I ran away from C++ like the plague? 😄
12:19:48	PMunch	Haha, don't worry I've been down that road as well
12:20:19	FromDiscord	<Skaruts> don't even mention stl error messages...
12:21:14	FromGitter	<alehander92> c# is usually not verbose
12:21:27	Araq	Zevv, should we fix --os:standalone or come up with a new switch?
12:22:49	federico3	a default panicoverride.nim would be nice
12:24:50	FromGitter	<alehander92> what is the problem with standalone
12:24:54	PMunch	alehander42, where do you get that from? I find C# very verbose
12:28:33	FromGitter	<alehander92> sorry
12:28:41	FromGitter	<alehander92> i remember c# being less verbose than java
12:28:51	FromGitter	<alehander92> but not very experienced with both :O
12:29:17	lqdev[m]	both are overly verbose imo
12:29:43	FromDiscord	<Skaruts> C# setters/getters?
12:29:59	FromGitter	<alehander92> yeah c# had some shortcuts didnt it?
12:30:06	FromGitter	<alehander92> and they added a simple form of pattern matching?
12:30:13	lqdev[m]	well, they're better than the non-existent Java getters and setters
12:30:15	FromGitter	<alehander92> but java also changed a lot of things .. so dont know
12:30:36	FromDiscord	<Skaruts> off the top of my head I think I remember setters/getters being a bit cryptic or verbose
12:30:56	FromDiscord	<Skaruts> but I hardly ever look at C# so I could easily be wrong
12:30:59	lqdev[m]	I ran away from Java to start using Groovy
12:31:06	lqdev[m]	it helped a bit, but not much
12:31:46	lqdev[m]	at least I didn't have to do ctx.setFillColor() everywhere, I could just use ctx.fillColor =
12:32:11	lqdev[m]	Groovy is quite a big improvement over Java, tbh
12:32:16	FromDiscord	<Skaruts> I ran away from java because of `org.this.that.theother.yetanother.otheragain.stuff.things.more`
12:32:32	lqdev[m]	ah shit, this too
12:32:48	lqdev[m]	wasn't that much of a problem with a proper IDE, but still
12:32:59	Araq	C# is excellent in the sense that it is the one production language I know about where featurism didn't cause quality problems
12:33:03	FromDiscord	<Clyybber> intellij saves your ass
12:33:19	FromDiscord	<Clyybber> yeah, C# is solid
12:33:27	Araq	C# has many many features and they all work well together
12:33:33	FromDiscord	<Clyybber> I also like that you can pretty easily reverse engineer C#
12:33:36	Araq	quite remarkable
12:33:49	FromDiscord	<Clyybber> take any unity game
12:33:51	FromDiscord	<Skaruts> just two days ago I was looking in some codebase and went to the depths of the earth to find a file that might be interesting, only to find 20 lines of code, 3 of which were imports from elsewhere... -.-
12:33:54	FromDiscord	<Clyybber> inspect its whole code
12:34:06	FromDiscord	<Skaruts> but it was kotlin not java, but same thing
12:34:54	Araq	contrast that with Scala where stuff also works well together afaict but compile-times suffered
12:35:55	PMunch	Everything works together, but at the penalty that everything seems to be wrapped in 8 layers of abstraction..
12:36:18	Araq	that's Java's influence on C#, sadly
12:36:31	PMunch	But yeah, Nim could definitely work a bit more to make features work together more seamless (which is mostly a stdlib problem I think)
12:37:13	Araq	it's mostly a dogfood issue, whenever we eat our own dogfood stuff works well
12:37:31	Araq	and when we don't (not nil, concepts) it's really not fleshed out enough
12:39:12	PMunch	Yeah, if concepts was a lot more fleshed out you could write the stdlib to work on concepts instead of specific types
12:39:37	Araq	a bit but it's complex
12:39:38	lqdev[m]	^
12:39:58	PMunch	So you could create your own version of something and it would just work because it shares the same "interface" as the type it was indented to work with
12:40:15	Araq	for example, proc countWords(input: string): CountTable[string]
12:40:27	Araq	ok, concepts instead:
12:40:28	lqdev[m]	I wanted to use concepts in rapid, but was forced to use the usual tuple approach for implementing interfaces, because concepts were buggy
12:40:44	PMunch	proc countWords[T: concept](input: T): CountTable[T]
12:40:46	FromDiscord	<mratsim> concepts kind of Work in Weave
12:40:55	Araq	proc countWords(input: StringLike): CountTable[StringLike]
12:40:59	FromDiscord	<mratsim> I use them in a couple of places
12:41:09	Araq	but this is still not enough...
12:41:11	Araq	so ...
12:41:16	*	FromGitter quit (Read error: Connection reset by peer)
12:41:17	Araq	proc countWords(input: StringLike): CountTableLike[StringLike]
12:41:34	PMunch	Well no, it could return a specific CountTable
12:41:34	*	FromGitter joined #nim
12:41:42	Araq	except we don't have return type inference for 'CountTableLike'
12:42:05	Araq	and also is the input string the same concept as in CountTable[string]?
12:42:39	FromDiscord	<mratsim> while I would love return type overloading that seems like opening pandora's box in the current state of generics/typedesc/semchecks/sigmatches issue
12:42:43	PMunch	Yeah if you do `proc countWords[T: StringLike](input: T): CountTable[T]`
12:42:48	PMunch	That way it would be the same type
12:42:48	lqdev[m]	kind of sucks that there isn't a runtime alternative to concepts, it could really help with some code using inheritance and methods
12:43:06	FromDiscord	<Clyybber> @mratsim Yeah, return type overloading would be great, fixing generic bugs even more so
12:43:12	FromDiscord	<mratsim> @lqdev, there is a commented out description of runtime concepts :p
12:43:12	FromDiscord	<Skaruts> are there printable versions of the manual and the tutorial?
12:43:14	Araq	maybe my input is 'cstring' though and the return type is CountTable[string]
12:43:45	Araq	even with a perfect concept implementation it's not simple
12:44:07	FromDiscord	<mratsim> @lqdev: https://raw.githubusercontent.com/nim-lang/Nim/a7aeabb9d2c70a5d9bd89abf3eb08372d2c6d9d0/doc/manual_experimental.rst scan for VTable
12:44:31	PMunch	I just want to be able to pass a cstring or a StringSlice to a procedure that does some simple things on a string.
12:44:49	PMunch	E.g. to `[]`(index: int)
12:45:28	Araq	we're getting openArray borrowing for that but I'm also considering to support O(1) slicing for Nim's builtin strings
12:45:31	PMunch	Would make using distinct types much more important though..
12:45:35	FromDiscord	<Clyybber> for now I think its a better idea to fix generic bugs first and then concept
12:46:03	Araq	'concept' is my weapon to make Nim's generics more robust, see my RFC about it
12:46:09	FromDiscord	<mratsim> How are methods nowadays though?
12:46:27	Araq	I think they are fine
12:46:42	Araq	we could use a better implementation now that the 'multi' aspect of them is dead
12:46:58	FromDiscord	<mratsim> I'm mostly using either compile-time stuff or enums but I suspects methods are needed in many high-level cases
12:47:20	Araq	for high level cases I use closures
12:47:27	Araq	they work really well
12:47:44	Araq	and even ARC supports them, yay
12:48:20	FromDiscord	<mratsim> They work so weel that when you want a higher-order function, you need to tag {.nimcall.} everywhere because otherwise everything is a closure ...
12:48:22	FromDiscord	<Clyybber> we should lure people into sigmatch
12:48:23	FromDiscord	<mratsim> well*
12:48:30	FromDiscord	<mratsim> it's a trap
12:48:35	FromDiscord	<Clyybber> "big pile of gold spotted at the bottom of sigmatch"
12:48:44	FromDiscord	<Clyybber> dig deep and dig safe!
12:49:04	Araq	sigmatch not as bad as semexprs.nim
12:49:15	FromDiscord	<mratsim> maybe you should use synthesis and rewrite sigmatch as a state machine 😉
12:49:16	FromDiscord	<Clyybber> oh
12:49:31	Araq	semexprs.nim is full of WTFs
12:49:31	*	Hideki_ quit (Ping timeout: 268 seconds)
12:49:36	FromDiscord	<mratsim> now if it could spit out a diagram that would be nice
12:49:43	FromDiscord	<Clyybber> @mratsim nah, I don't think so, sigmatch having state is one of the big problems
12:49:58	FromDiscord	<Clyybber> I think
12:49:58	Araq	I can guide you through the code and tell you
12:50:08	FromDiscord	<mratsim> state machine is not about removing state, it's about making it manageable.
12:50:20	Araq	"not in the spec, violating the spec"
12:50:22	FromDiscord	<mratsim> flags, booleans and what not
12:50:22	FromDiscord	<Clyybber> exactly, but I think sigmatch should have no state
12:50:29	FromDiscord	<Clyybber> Araq: lol
12:51:08	Araq	whereas sigmatch is mostly inline with the spec
12:51:23	FromDiscord	<Clyybber> hmm
12:51:27	Araq	sigmatch is complex, semexprs is wrong.
12:52:21	Araq	or "not even wrong".
12:53:52	FromDiscord	<mratsim> btw, low-priority as i have a workaround but why does the Nim compiler returns a "r" char instead of a nnkSym here? https://github.com/nim-lang/Nim/issues/13049#issuecomment-571089335 That's like WTF.
12:53:54	disbot	➥ typeof(default(T)[]) cannot be manipulated in macros ; snippet at 12https://play.nim-lang.org/#ix=26zO
12:55:24	FromDiscord	<mratsim> That's the strangest issue I've ever seen I think
12:55:28	FromDiscord	<Clyybber> because [r]
12:55:36	FromGitter	<alehander92> Araq i think `not nil` is mostly just lacking the final implementation
12:55:54	FromGitter	<alehander92> i dont think there is so much confusion about how it should work
12:57:04	Araq	alehander92: agreed, not any more
12:57:31	*	a_chou joined #nim
13:00:01	*	a_chou quit (Client Quit)
13:00:06	FromGitter	<sheerluck> No videos found at https://www.twitch.tv/araq4k/videos and that makes me sad
13:00:48	FromDiscord	<Stuffe> I just formatted and reinstalled OSX, on a fresh install Nim is missing "libmysqlclient.20.dylib" for the mysql module
13:01:07	FromDiscord	<Stuffe> Anyone have an idea why?
13:02:54	FromGitter	<sheerluck> @Stuffe I think it's `mysql-connector-c ` not Nim
13:02:57	*	nsf joined #nim
13:03:30	FromDiscord	<Stuffe> Something I need to install?
13:04:46	*	Kaivo joined #nim
13:04:59	FromDiscord	<Stuffe> Crap, I get these errors:
13:04:59	FromDiscord	<Stuffe> dlopen(libmysqlclient.dylib, 2): image not found
13:05:00	FromDiscord	<Stuffe> dlopen(libmysqlclient.20.dylib, 2): image not found
13:05:00	FromDiscord	<Stuffe> dlopen(libmysqlclient.19.dylib, 2): image not found
13:05:00	FromDiscord	<Stuffe> dlopen(libmysqlclient.18.dylib, 2): image not found
13:05:00	FromDiscord	<Stuffe> dlopen(libmysqlclient.17.dylib, 2): image not found
13:05:02	FromDiscord	<Stuffe> dlopen(libmysqlclient.16.dylib, 2): image not found
13:05:04	FromDiscord	<Stuffe> dlopen(libmysqlclient.15.dylib, 2): image not found
13:05:06	FromDiscord	<Stuffe> dlopen(libmariadbclient.dylib, 2): image not found
13:05:07	FromDiscord	<Stuffe> dlopen(libmariadbclient.20.dylib, 2): image not found
13:05:09	FromDiscord	<Stuffe> dlopen(libmariadbclient.19.dylib, 2): image not found
13:05:11	FromDiscord	<Stuffe> dlopen(libmariadbclient.18.dylib, 2): image not found
13:05:13	FromDiscord	<Stuffe> dlopen(libmariadbclient.17.dylib, 2): image not found
13:05:14	FromDiscord	<Stuffe> dlopen(libmariadbclient.16.dylib, 2): image not found
13:05:16	FromDiscord	<Stuffe> dlopen(libmariadbclient.15.dylib, 2): image not found
13:05:19	PMunch	Please don't paste stuff into Discord..
13:05:26	FromDiscord	<Stuffe> But I found libmysqlclient.21.dylib in a dir, I assume with my new mysql install
13:05:30	FromDiscord	<Stuffe> Ok sorry
13:05:48	FromDiscord	<Stuffe> Anyway, I guess it may be looking for older versions only
13:07:55	*	Romanson joined #nim
13:15:31	*	marmotini_ quit (Remote host closed the connection)
13:15:58	*	marmotini_ joined #nim
13:20:42	*	marmotini_ quit (Ping timeout: 258 seconds)
13:20:53	FromDiscord	<Stuffe> I tried installing mysql-connector-c with brew and I tried creating a symlink to libmysqlclient20 from libmysqlclient21, neither worked
13:22:02	FromDiscord	<Stuffe> I have no idea what it is about and google has nothing about it either :/
13:24:47	*	dddddd joined #nim
13:28:35	FromGitter	<alehander92> Araq btw
13:29:02	FromGitter	<alehander92> about standalone, i wonder if now it would be easier to use more of the stdlib in this mode (with arc)
13:30:46	FromGitter	<alehander92> i remember people complaining about not knowing which modules are avaiable with js backend, so some kind of docs feature where it can infer "show me a list of stuff available for X flag" might be useful
13:33:28	*	marmotini_ joined #nim
13:33:53	FromGitter	<alehander92> the discord bot can autodetect many lines added very quickly by timestamp maybe
13:34:07	FromGitter	<alehander92> maybe not so easy
13:35:44	Araq	alehander92: I figured it out
13:35:53	Araq	the design was unclear but now I know
13:49:38	Zevv	he once was blind but now can see
14:02:19	blackbeard420	when using asyncnet, is it possible to spawn off a long running task on the threadpool via spawn and send the response *from* that thread. The code compiled and ran however wouldnt send due to what im assuming is the threadlocal storage having different global dispatcers
14:05:53	FromDiscord	<mratsim> can't you pass the handle or whatever asyncnet provide to the spawned thread?
14:06:07	*	lritter joined #nim
14:06:20	blackbeard420	yeah passing the AsyncSocket is what i did
14:06:58	blackbeard420	however asyncCheck'ing it did nothing and trying a waitFor errored with an unregistered handle
14:09:37	disruptek	because the dispatcher is in the other thread.
14:10:58	blackbeard420	yeah thats what i discovered, is there any way to share the dispatcher between threads. Or how would i go about sending from another thread? Im used to c++ asio where you can post send/reads from other threads
14:16:08	FromGitter	<alehander92> Araq so what did you figure out
14:18:09	FromDiscord	<treeform> with --gc:arc the heap objects are shared between threads right? So should you be able to append to a regular string from all threads? (which requires reallocation some times)
14:21:17	FromDiscord	<mratsim> yes, but it's very discouraged
14:21:35	FromDiscord	<mratsim> I would need a lock that would be hammered by all threads
14:22:53	FromDiscord	<treeform> isn't the point of shared heap is that we can pass stuff? (provided we lock stuff)
14:30:38	disruptek	i thought sink annotation was optional. 🙁
14:35:48	FromDiscord	<mratsim> yes the points of shared heap is to pass stuff, but the point of threading is to handle more stuff in a similar amount of time.
14:35:48	FromDiscord	<mratsim> Locking goes against threading. parallel program are much much faster when they don't need to synchronize
14:36:15	FromDiscord	<mratsim> and if you need to synchronize, only require synchronization between 2 threads in a P2P fashion instead of N threads
14:36:35	FromDiscord	<mratsim> that's what channel/flowvar allow you to do
14:36:50	FromDiscord	<treeform> well let me time it then?
14:36:51	Araq	nevertheless it is supported
14:37:13	FromDiscord	<treeform> Is there a way to disable "is not GC-safe as it accesses"
14:37:14	Araq	treeform: with --threads:on the heap is shared by default and once you lock your stuff
14:37:21	Araq	it's safe
14:37:35	Araq	oh yeah, use {.gcsafe.}: block here
14:37:42	Araq	to make the compiler shut up
14:38:20	FromDiscord	<treeform> that did make it shut up, thanks!
14:38:49	FromDiscord	<treeform> segfaults with "execution of an external program failed" instead
14:39:02	Araq	can you share the gist?
14:39:14	FromDiscord	<mratsim> you often need gdb/lldb to backtrace and have the actuall Nim stacktrace
14:39:36	FromDiscord	<treeform> https://play.nim-lang.org/#ix=26Al
14:39:41	FromDiscord	<mratsim> I think the stacktrace doesn't have enough time to be printed to stdout otherwise or something like that
14:40:56	FromDiscord	<mratsim> works on my machine
14:41:25	FromDiscord	<mratsim> ah, with danger it crashes
14:41:42	FromDiscord	<treeform> with `--gc:arc --threads:on` ?
14:41:51	FromDiscord	<mratsim> 1: 15 ......................................................................
14:41:51	FromDiscord	<mratsim> 2: 16 .......................................................................
14:41:52	FromDiscord	<mratsim> 4: 16 ........................................................................
14:41:52	FromDiscord	<mratsim> 3: 16 .........................................................................
14:41:52	FromDiscord	<mratsim> [Thread 0x7ffff78ab700 (LWP 512962) exited]
14:41:53	FromDiscord	<mratsim>
14:41:55	FromDiscord	<mratsim> Thread 4 "abb" received signal SIGSEGV, Segmentation fault.
14:41:57	FromDiscord	<mratsim> [Switching to Thread 0x7ffff762a700 (LWP 512963)]
14:41:59	FromDiscord	<mratsim> 0x0000555555560295 in resizeString ()
14:42:00	FromDiscord	<mratsim> (gdb) bt
14:42:02	FromDiscord	<mratsim> #0 0x0000555555560295 in resizeString ()
14:42:03	FromDiscord	<mratsim> #1 0x0000555555566a81 in threadFunc__9cIvQ16Dlrt0c4oldW9b0Mcg ()
14:42:05	disbot	https://github.com/nim-lang/Nim/issues/1 -- 5Trojan horse reported by avira free-av during the setup (installer)
14:42:05	FromDiscord	<mratsim> #2 0x00005555555650d0 in threadProcWrapDispatch__xMRsJuvEkkgop9aeu6Zsxaw_2 ()
14:42:07	disbot	https://github.com/nim-lang/Nim/issues/2 -- 5Compiler segfaults ; snippet at 12https://play.nim-lang.org/#ix=22vr
14:42:07	FromDiscord	<mratsim> #3 0x000055555556519c in threadProcWrapper__oTnP9cUoE9cVTUL7iHAoIIAA ()
14:42:08	disbot	https://github.com/nim-lang/Nim/issues/3 -- 5methods cause compiler to segfault
14:42:09	FromDiscord	<mratsim> #4 0x00007ffff7e054cf in start_thread () from /usr/lib/libpthread.so.0
14:42:11	disbot	https://github.com/nim-lang/Nim/issues/4 -- 5high, len function with sequences
14:42:11	FromDiscord	<mratsim> #5 0x00007ffff7d342d3 in clone () from /usr/lib/libc.so.6
14:42:12	disbot	https://github.com/nim-lang/Nim/issues/5 -- 5compile pragma does not support cpp files
14:42:13	FromDiscord	<mratsim> let me try with arc
14:42:13	Yardanico	what is happening
14:42:41	FromDiscord	<treeform> neat
14:42:54	FromDiscord	<mratsim> no crash with arc
14:42:56	FromDiscord	<treeform> I though GDB output was issues
14:43:04	FromDiscord	<treeform> It though GDB output was issues
14:43:06	Araq	crashes for me too, interesting
14:43:10	FromDiscord	<Skaruts> is it at all possible to implement syntax like `i++`?
14:43:31	FromDiscord	<mratsim> Nim doesn't support postfix (except the export marker)
14:43:31	Araq	it also crashes with -d:useMalloc
14:43:54	FromDiscord	<Skaruts> would it be possible with a prefix?
14:43:55	FromDiscord	<mratsim> it doesn't for me :/
14:44:02	Araq	however
14:44:02	FromDiscord	<treeform> @manterolat what OS?
14:44:04	FromDiscord	<mratsim> yes @Skaruts
14:44:10	FromDiscord	<treeform> I am on windows
14:44:14	Araq	for me it also crashes with counter.add commented out
14:44:20	Araq	so it's something else
14:44:22	Yardanico	@Skaruts do you have a specific usecase for it? You can write a simple proc with a different name which will do the same
14:44:33	Araq	maybe thread creation wasn't ported properly
14:44:52	FromDiscord	<mratsim> are you on windows as well @Araq?
14:44:53	FromDiscord	<treeform> works for me without counter add
14:45:02	Araq	mratsim: yes
14:45:10	FromDiscord	<mratsim> a wild test case appears
14:45:36	FromDiscord	<Skaruts> I'm just having fun with templates, but didn't seem to work, I'm guessing this would be a thing for macros
14:45:41	FromDiscord	<mratsim> no
14:45:55	FromDiscord	<mratsim> proc `++`(i: var int) = inc i
14:46:20	Araq	treeform: works for me with --exceptions:setjmp
14:46:39	Araq	my exit handling for the new exceptions is not threadsafe
14:46:47	Araq	or something like that :-)
14:46:56	FromDiscord	<treeform> @Skaruts https://play.nim-lang.org/#ix=26Am
14:46:57	FromDiscord	<mratsim> ah, I didn't update devel the past week AFAIK
14:47:28	Yardanico	@treeform that's not i++ though :P
14:47:36	Yardanico	ah nvm
14:48:12	FromDiscord	<Skaruts> hah! the reason why it wasn't working for me is that I was doing `echo ++i`
14:48:47	FromDiscord	<mratsim> you can do proc `++`(i: var int): int {.discardable.} = inc i; i
14:48:55	FromDiscord	<treeform> @Skaruts https://play.nim-lang.org/#ix=26Ao
14:48:57	FromDiscord	<mratsim> if you really want the result to also be returned
14:48:59	*	marmotini_ quit (Remote host closed the connection)
14:49:01	Araq	"This approach worked fine until our application got threaded, at which point atexit() handlers stopped to work predictably.
14:49:01	Araq	We learned by trial an error that threads may already be dead in atexit() handler, and their stacks deallocated."
14:49:03	Araq	bummer
14:49:32	FromDiscord	<mratsim> Sonds like a great quote on a tomb
14:49:41	FromDiscord	<Skaruts> hmm works
14:49:59	FromGitter	<alehander92> mratsim please please dont paste
14:50:01	FromGitter	<alehander92> in discord
14:50:08	Araq	oh well I have a better way to do the same, I think
14:50:32	FromDiscord	<mratsim> @alehander42 noted
14:50:32	FromGitter	<alehander92> disruptek can disbot ignore `#<one digit>` maybe
14:50:33	FromDiscord	<treeform> Araq, how would you do it?
14:50:48	FromDiscord	<Skaruts> not even one liners should be pasted?
14:50:52	FromGitter	<alehander92> it seems that often they are not nim issues but just normal parts of markdown
14:50:55	FromDiscord	<mratsim> one liners are OK
14:51:03	FromDiscord	<mratsim> but it's the newline characters that is broken I think
14:51:25	FromGitter	<alehander92> Skaruts it seemed as many lines on my screen, + disbot got confused and it became pretty :P
14:51:31	FromGitter	<alehander92> but one-two liners are fine
14:51:47	FromDiscord	<Skaruts> yea I noticed it on irc the code gets all weird or cut out
14:51:48	FromDiscord	<mratsim> If I write something that spans
14:51:48	FromDiscord	<mratsim> multiple line it will come up ugly
14:51:48	FromDiscord	<mratsim> whether it's code or not
14:51:48	FromDiscord	<mratsim> See https://irclogs.nim-lang.org/
14:51:59	FromGitter	<alehander92> sorry, one day i guess there would be better bridges
14:52:07	Araq	treeform: via compiler magic
14:52:19	Araq	in fact, I already have the code
14:52:43	FromGitter	<alehander92> mratsim yeah a bit annoying no big deal
14:53:03	FromGitter	<alehander92> @michael72 are you on gitter
14:53:29	disruptek	yeah, i keep forgetting to fix disbot for short numbers.
14:53:47	disruptek	on the plus side...
14:53:53	FromDiscord	<treeform> Araq, what would happen if two threads decide to resize the same string at roughly the same time?
14:53:59	FromGitter	<alehander92> gdb stacktraces are full of those disruptek :P
14:54:02	disruptek	we have biographies now.
14:54:04	disruptek	~zevv
14:54:05	disbot	zevv: 11#nim's resident german. -- disruptek
14:54:14	FromGitter	<alehander92> ~disruptek
14:54:15	disbot	disruptek: 11a sexy fella with magic hands.
14:54:23	FromGitter	<alehander92> ~Araq
14:54:24	disbot	no footnotes for `araq`. 🙁
14:54:31	FromGitter	<alehander92> neat!
14:54:35	disruptek	~araq is 👑
14:54:36	disbot	araq: 11👑
14:54:49	FromDiscord	<treeform> "no footnotes" - only "fingernotes"
14:54:50	FromDiscord	<mratsim> @treeform, that's why you need a lock before add
14:55:00	FromGitter	<alehander92> especially the zevv one .. great banter :D
14:55:06	PMunch	~PMunch
14:55:07	disbot	no footnotes for `pmunch`. 🙁
14:55:13	PMunch	Aw :(
14:55:15	FromGitter	<alehander92> ~mratsim
14:55:16	disbot	no footnotes for `mratsim`. 🙁
14:55:20	FromGitter	<alehander92> ~disbot !
14:55:21	disbot	no footnotes for `disbot`. 🙁
14:55:24	disruptek	you have to add them, knuckleheads.
14:55:28	disruptek	~help
14:55:34	Araq	treeform: also try --tslEmulation:off
14:55:35	FromDiscord	<mratsim> ~welp
14:55:36	disbot	no footnotes for `welp`. 🙁
14:55:41	PMunch	~help
14:56:17	PMunch	~PMunch is testing out the new biography feature
14:56:17	disbot	no footnotes for `pmunch`. 🙁
14:56:27	FromGitter	<alehander92> disruptek it needs to be a stack that you push bio-lines to
14:56:32	disruptek	lowercase.
14:56:40	PMunch	~pmunch is testing out the new biography feature
14:56:41	disbot	pmunch: 11testing out the new biography feature
14:56:44	PMunch	Ey!
14:56:48	FromGitter	<alehander92> ~disruptek is dont replace the previous ?
14:56:48	disbot	disruptek: 11a sexy fella with magic hands.
14:56:48	PMunch	~PMunch
14:56:48	disbot	disruptek: 11dont replace the previous ?
14:56:49	disbot	pmunch: 11testing out the new biography feature
14:57:01	FromGitter	<alehander92> huh it seems it already is interesting
14:57:10	disruptek	you will replace your last footnote on the term, but not the footnote others have left.
14:57:13	PMunch	~disruptek
14:57:14	disbot	disruptek: 11a sexy fella with magic hands.
14:57:14	disbot	disruptek: 11dont replace the previous ?
14:57:20	PMunch	Oh, it adds them?
14:57:22	FromGitter	<alehander92> great bio
14:57:27	FromGitter	<alehander92> makes you think
14:57:29	PMunch	Oooh, right
14:57:34	FromGitter	<alehander92> asking questions about life
14:57:52	disruptek	~compiler
14:57:53	disbot	compiler: 11https://nim-lang.org/docs/nimc.html -- the Nim Compiler User Guide -- disruptek
14:57:53	PMunch	It should att the -- <name> part automatically
14:58:00	FromDiscord	<mratsim> I think you just unleashed black hole that will suck IRC soul
14:58:08	disruptek	it will omit others' names so that they don't get pinged.
14:58:10	FromGitter	<alehander92> no
14:58:22	disruptek	unless you speak to it privately, of course.
14:58:33	PMunch	Aah, that makes a lot of sense :P
14:58:43	FromDiscord	<mratsim> ~API is https://nim-lang.org/docs/theindex.html
14:58:44	disbot	no footnotes for `api`. 🙁
14:58:45	Araq	treeform: it's --tlsEmulation:on in combination with --exceptions:goto
14:58:50	disruptek	lowercase.
14:58:51	FromDiscord	<mratsim> ~api
14:58:52	disbot	no footnotes for `api`. 🙁
14:58:55	Araq	but I don't have a fix yet, sorry
14:58:59	FromDiscord	<mratsim> ~api is https://nim-lang.org/docs/theindex.html
14:59:00	disbot	api: 11https://nim-lang.org/docs/theindex.html
14:59:04	Araq	more complex that I thought it would be
14:59:05	disruptek	i guess that's the first bug.
14:59:11	FromGitter	<alehander92> ~mratsim is a french guy, aspiring go player and true man of baguette
14:59:11	disbot	mratsim: 11a french guy, aspiring go player and true man of baguette
14:59:18	FromGitter	<alehander92> ok yeah private makes more sense
14:59:35	FromDiscord	<mratsim> lol
14:59:50	*	marmotini_ joined #nim
14:59:58	PMunch	Hmm, I got "-- #nim" when I polled it for my bio in a private chat
15:00:13	disruptek	neat, the second bug.
15:00:31	FromGitter	<alehander92> ok 9disruptek
15:00:46	PMunch	Can I delete a bio?
15:00:58	disruptek	you can replace them but you can't delete them.
15:01:02	PMunch	Or will I be testig the new biography feature until I can think of something better to write? :P
15:01:03	PMunch	Ah..
15:01:18	disruptek	the problem with allowing deletion is that it makes people put garbage into the database.
15:01:18	FromGitter	<alehander92> ~help
15:01:22	FromDiscord	<treeform> Araq, it seems to work, but then it crashes some place in the middle for unknown reason.
15:01:24	disruptek	at least, that's my theory.
15:01:37	FromGitter	<alehander92> ~~
15:01:38	disbot	no footnotes for `~`. 🙁
15:01:42	Araq	treeform: for me it then doesn't crash
15:03:30	PMunch	disruptek, what do you mean?
15:03:37	PMunch	Now I can't remove the garbage I put in there..
15:03:49	disruptek	yeah, that's the way bugs are.
15:03:59	disruptek	they are like the opposite of features.
15:04:02	disruptek	they arrive broken.
15:04:23	FromDiscord	<treeform> Araq, ha, for me it crashes at the very end when all threads finish... and it should return from the joinThreads.
15:04:42	Araq	yes, when it does crash
15:04:50	Araq	it does crash at the end when we try to cleanup threads
15:05:53	*	PMunch quit (Quit: Leaving)
15:07:30	FromDiscord	<treeform> Araq, arc+thread bug?
15:07:48	Araq	for me it's a --tlsEmulation:on bug
15:08:07	Araq	and it's easy enough to see in gdb.exe
15:08:19	Araq	we access thread local storage after it was freed
15:08:29	Araq	in the implementation of --exceptions:goto
15:08:54	Araq	working on a fix but --tlsEmulation:on is a hack anyway, disable it
15:09:09	Araq	please disable it, native TLS is much better
15:10:35	*	ng0_ joined #nim
15:10:48	FromDiscord	<treeform> Araq, I do "--gc:arc --threads:on --exceptions:goto --tlsEmulation:on" still crashes: minimal example: https://play.nim-lang.org/#ix=26AB
15:11:06	Araq	no, use --tlsEmulation:off
15:11:25	Araq	alternatively use --exceptions:setjmp
15:11:32	FromDiscord	<treeform> Araq, yes off works, sorry miss understood you
15:11:47	Araq	but --exceptions:setjmp sucks and is obsolete
15:11:57	disruptek	lol
15:12:04	FromDiscord	<treeform> --exceptions:setjmp works
15:12:22	Araq	I know but I don't want to support it for much longer
15:13:10	disruptek	nim 2020: obsoleting features before they are even replaced.
15:13:16	FromDiscord	<treeform> i'll use the "--exceptions:goto --tlsEmulation:off" for now, I guess.
15:13:28	disruptek	treeform: resistance is futile.
15:13:37	*	ng0 quit (Ping timeout: 265 seconds)
15:13:38	FromGitter	<alehander92> https://news.ycombinator.com/item?id=21968686
15:13:46	Araq	treeform: yeah, that's what we strive for
15:13:53	FromGitter	<alehander92> dont know if my comments are correct feel free to correct me here
15:15:36	Araq	it's correct but you could elaborate a bit
15:16:11	Araq	given Nim's distinction between ref/ptr we have lots of flexibility how to map Nim onto an MM scheme
15:16:30	Araq	and we exploit that
15:17:13	*	marmotini_ quit (Remote host closed the connection)
15:17:36	FromGitter	<alehander92> well i am not entirely sure how what does this mean :D
15:20:50	FromDiscord	<treeform> "what would happen if two threads decide to resize the same string at roughly the same time?" - you get a crash... now we know.
15:23:23	*	ng0_ is now known as ng0
15:28:20	*	solitudesf quit (Remote host closed the connection)
15:28:27	*	solitudesf- joined #nim
15:31:28	FromGitter	<matrixbot> `grantmwilliams` Anyone know of a comparision between Nim and some of the other newer languages like crystal, V, zig, pony etc. ?
15:31:53	Yardanico	what do you mean by "comparasion"?
15:32:28	FromDiscord	<mratsim> there is an in-depth D comparison but AFAIK it's already obsolete 😛
15:33:21	FromDiscord	<mratsim> ultimately besides some deal breaker (do you need windows support?), the best would be to try each language to see what fit for you and your applications
15:34:20	FromDiscord	<mratsim> I honestly don't think there is one thing in those languages that cannot be done in Nim. So it all comes down to the library ecosystem.
15:35:09	FromDiscord	<mratsim> On another hand, there are several things that I do in Nim that I wouldn't be able to do in those languages. So it all depends on your use cases.
15:35:21	FromGitter	<matrixbot> `grantmwilliams` I dont really care about benchmarks or anything like that. I guess I'm mostly interested in more of the design. What do the language authors intend the language to be used for, what dependencies do the languages have, what platforms can they target etc.
15:36:29	Yardanico	well, Nim can target the most platforms out of all languages you asked, because it compiles to C and has support for quite a lot of OSes
15:36:50	Yardanico	Zig and Crystal use LLVM IR so the number of targets is more limited, although I really like cross-compilation in Zig
15:37:02	FromGitter	<alehander92> 1) browser/electron (with js backend and afaik maybe with nlmv's wasm support )
15:37:11	FromDiscord	<mratsim> AFAIK except from Pony which is heavily biaised towards distributed, usually financials, applications all aim to be a general-purpose languages, usually a better C
15:37:42	FromGitter	<alehander92> i'd say crystal sounds like a "fast typed ruby" , alternative to go to me
15:37:55	FromDiscord	<Rika> it even has the fitting name!
15:37:58	FromGitter	<alehander92> but yeah people use even it for system stuff indeed(as well as go)
15:38:10	FromDiscord	<treeform> Big LOL: https://play.nim-lang.org/#ix=26AM
15:38:19	FromDiscord	<mratsim> it's not even always typed, and when you use untyped Crystal apparently it slow downs compilation a lot
15:39:19	FromGitter	<alehander92> you mean inference? it is still statically typed
15:39:23	FromDiscord	<Rika> is it type inference or dynamic typing?
15:39:24	FromDiscord	<Rika> yhea
15:39:26	FromDiscord	<Rika> yeah*
15:39:35	FromGitter	<alehander92> but yeah, i've heard the inference might be slow sometimes indeed
15:39:36	FromDiscord	<mratsim> no crystal has a mixmatch of both AFAIK
15:39:39	FromGitter	<alehander92> for their design
15:39:43	FromGitter	<alehander92> mratsim i dont think so
15:40:30	FromGitter	<alehander92> they just skip many of the annotations but no real dynamic typing (except if they have something like our `.` JsObject kind of thing)
15:41:42	FromDiscord	<mratsim> look at the very last example here: https://crystal-lang.org/2015/03/04/internals.html
15:41:51	FromDiscord	<mratsim> apparently it's dynamic dispatch
15:42:17	FromDiscord	<Rika> so still static typing
15:42:34	FromDiscord	<mratsim> no, the def(x, y) example is dynamic
15:42:55	FromDiscord	<mratsim> or i didn't understand anything
15:43:24	*	solitudesf- quit (Quit: Leaving)
15:43:45	Yardanico	maybe it's just like "auto" in Nim?
15:43:45	*	solitudesf joined #nim
15:44:06	Yardanico	creates a new generic function for each combination of argument types
15:44:06	FromGitter	<alehander92> i think it just looks like generics or auto indeed
15:44:20	FromGitter	<alehander92> but i agree that "And this also works if the type of all arguments is not known at compile time." is confusing
15:44:28	FromGitter	<alehander92> but i guess this is about subclasses?
15:54:49	*	marmotini_ joined #nim
15:58:20	*	Trustable joined #nim
16:01:28	*	Vladar joined #nim
16:04:13	FromGitter	<FabienPoirierPrincipia> Hi all, I am new to Nim, found it great, I'd like to statProcess (osProc) with a specific PRIORITY (lower one), but it seems hard coded tp normal priority. Is it possible to run process in lower prority?
16:08:48	Yardanico	I think you have to use OS-specific APIs for now
16:09:37	*	beshr joined #nim
16:23:32	Araq	FabienPoirierPrincipia: PRs are accepted for the stdlib
16:24:24	Araq	and usually better than copying the existing lib in order to mutate it. though this is useful too
16:24:55	*	NimBot joined #nim
16:26:38	*	filcuc joined #nim
16:26:56	*	Romanson quit (Quit: Connection closed for inactivity)
16:35:24	disruptek	i'm getting an index error during compilation of a nimph test in 1.0.4. the error is absent in 1.0.5. 😢
16:36:29	Araq	time for 1.0.6 then
16:37:44	*	marmotini_ quit (Remote host closed the connection)
16:39:03	*	disruptek sighs.
16:40:39	narimiran	disruptek: now's the time then for "can you backport this too?" :)
16:41:02	disruptek	thanks, i'm good. 😉
16:46:10	*	Hideki_ joined #nim
16:50:46	*	Hideki_ quit (Ping timeout: 258 seconds)
16:54:21	disruptek	maybe the way to do `maybe` is with operators that visually indicate the addition of the `denil` "condom". a.<|x2.x3|> would allow the semantics to change without respect to the deref operator.
16:54:40	disruptek	looks nice with ligatures, also.
16:55:34	Araq	interesting idea
16:56:01	Araq	but I would use ?(a.b.c.d)
16:56:18	disruptek	that's even cheaper, yeah.
16:56:20	Araq	a ? prefix operator has little problems
17:00:50	Araq	in fact, I'd pick `?!`
17:01:04	Araq	it sticks out and is unlikely to be used for something else
17:01:22	Araq	if ?!x.foo.bar.baz == "a"
17:01:28	disruptek	nim-result is using `?` but i think i'm gonna fork it in any event.
17:01:52	disruptek	https://github.com/arnetheduck/nim-result/issues/10
17:01:52	disbot	➥ `$` is side-effect free yet calls `$` on all values
17:04:32	disruptek	() means you cannot pass a Maybe around. maybe that's good.
17:04:59	Araq	we have Option[T] for that
17:05:10	disruptek	i guess ?!foo is a Maybe.
17:06:06	*	sagax quit (Ping timeout: 258 seconds)
17:14:00	*	marmotini_ joined #nim
17:26:37	disruptek	maybe ?!foo is defined in optons, and its value is Option[typeof(foo)] and there's no Maybe. then it works correctly with the rest of stdlib and nil/notnil.
17:31:58	*	endragor_ quit (Remote host closed the connection)
17:33:52	disruptek	nim 1.0.5 doesn't ignore `sink` in default gc? and i get different behavior with and without `sink`.
17:35:43	Araq	sounds like a bug to me
17:35:47	Araq	'sink' is an optimization
17:36:00	Araq	(except for move-only types but we don't have them in the stdlib...)
17:36:30	disruptek	i know, but i'm having a devil of a time debugging between all these different behaviors.
17:37:16	Araq	that's because you're helping us to get --gc:arc stable
17:37:33	disruptek	there's a sucker born every minute.
17:37:57	Araq	I don't know what that means
17:38:33	disruptek	it's a quote of p.t. barnam, the famous showman.
17:38:58	disruptek	it means there's always another fool to dupe.
17:39:49	disruptek	adding sink got me around the bug i pm'd you about ~12hrs ago.
17:40:27	FromDiscord	<mratsim> when sink fixes bug it means it's ready for prime :p
17:43:03	disruptek	if i get a nil Cell it means i'm trying to dealloc something that we have no rc on, right?
17:43:29	FromDiscord	<mratsim> ping @Clyybber
17:44:58	Zevv	--gc:arc is totally done and finished. I switched it on, ran my tests, and all was well
17:45:01	Zevv	amen
17:45:31	disruptek	no more bugfixes, then.
17:46:13	*	filcuc quit (Ping timeout: 260 seconds)
17:47:47	*	nixfreak joined #nim
17:49:46	FromDiscord	<Clyybber> @mratsim pong
17:50:00	FromDiscord	<mratsim> see disruptek question about nil Cell
17:50:39	FromDiscord	<Clyybber> Cell is what?
17:50:49	FromDiscord	<Clyybber> An ref object I assume?
17:51:24	*	matic quit (Quit: Leaving)
17:55:33	disruptek	i had a nil deref issue where the pointer wasn't nil but head(p) was.
17:57:12	disruptek	an exhaustive case should error on else:
18:01:27	*	marmotini_ quit (Remote host closed the connection)
18:06:15	nixfreak	is it possible to create a buffer overlow using nim ?
18:07:16	Araq	yes. it's more difficult than in C though.
18:07:35	Araq	and if you use 'unsafe' memory it's even possible in Java
18:08:08	nixfreak	I'm just looking to use nim for exploit development , so trying to learn
18:09:38	FromDiscord	<mratsim> just use alloca
18:10:37	FromDiscord	<treeform> I think it's possible to create a buffer overflow in any general purpose language if you are creative enough.
18:10:47	Araq	yup
18:13:49	Zevv	if you can't do that, its not general purpose
18:14:07	Zevv	it should be hard and look ugly as hell, though
18:15:55	disruptek	this outputs "dumb `` 6"; result is string: echo "dumb `", result, "` ", result.len
18:18:21	disruptek	if i subsequently append to result, the new result is "6" & appended_string
18:32:32	disruptek	this is crazy. definitely broken, but it seems to have something to do with cstring.
18:33:19	narimiran	here are some numbers for arc-lovers: https://forum.nim-lang.org/t/5734#35803 (and a puzzle at the end for arc-debuggers :))
18:35:56	Araq	btw it's not "my" benchmark
18:36:09	narimiran	whose is it? :)
18:36:30	*	endragor joined #nim
18:36:34	Araq	it was a semi-official benchmark for latencies extracted from someone's production code
18:37:16	narimiran	ok, changed to "Araq has shared...." :)
18:38:00	FromDiscord	<treeform> When will --gc:arc will be default?
18:39:33	leorize	persumably when async supports it :P
18:40:25	FromDiscord	<sealmove> What exactly is async? Is it a stdlib module ?
18:41:38	*	endragor quit (Ping timeout: 260 seconds)
18:42:05	FromDiscord	<treeform> @sealmove yes its a set of modules asyncdispatch, asyncnet ... async*
18:42:25	FromDiscord	<treeform> @sealmove yes its a set of stdlib modules asyncdispatch, asyncnet ... async*
18:44:49	Araq	treeform: we said 'async' needs to work with it but *changing the default* are big words
18:45:07	Araq	and arguably the Nim compiler itself must work with it then
18:47:14	Araq	I got async to work on Unix, need to make the test part of the suite so we notice regressions
18:47:42	Araq	likewise we got GordonBGood's hell of a test to work
18:53:43	leorize	@treeform, please avoid editing on discord, it results in two messages on irc
18:55:56	Araq	narimiran, is it too late to fix/change your test results?
18:56:04	Araq	there is another bug in the code :-(
18:56:27	narimiran	update the gist and i'll re-run it and post the updated version
18:56:52	Araq	I don't have an update, the problem is this line
18:56:56	Araq	GC_setMaxPause(30)
18:57:15	Araq	the GC cannot handle this deadline so it never runs
18:57:19	Araq	I think
18:57:30	Araq	30µs is too low for it
18:57:47	Araq	maybe give it 500µs
18:58:30	Araq	but yeah, it does highlight a benefit of arc, there are no numbers to fill out
18:58:43	narimiran	yep, that gives much more reasonable results for old gc
18:58:47	narimiran	no long tail anymore
18:59:38	Araq	ok, so 500? or 400? or ...?
18:59:54	Araq	you need to tweak the value to the lowest sustainable latency
19:00:05	narimiran	60 is still bad, 70 is ok
19:00:13	Araq	in order to get meaningful results
19:00:46	narimiran	ha, 61 is bad, 62 is ok :D
19:01:00	Araq	it's machine specific :P
19:01:09	Araq	as you can see for me 30 was OK
19:01:19	Araq	otherwise why would the code contain it?
19:01:30	narimiran	excuses, excuses.... :D
19:01:43	Araq	sorry, well not too many have read it yet :P
19:01:56	Araq	so what number should my gist use? 100?
19:02:30	narimiran	yeah, i can run it with 100 to be on the safe side
19:07:49	narimiran	heh, now my conclusions are all wrong
19:08:43	Araq	sorry, I only spoted the problem
19:08:53	Araq	once you posted it
19:09:05	Araq	zct capacity: 1,008,895 # <-- 1 million entries?!
19:09:46	Araq	oh well, benchmarking is hard, everybody knows it
19:15:10	narimiran	btw, is the forum slow or is that on my end?
19:16:54	narimiran	there, updated my post, not as spectacular anymore
19:18:16	disruptek	but the misses delta grew in our favor.
19:21:07	*	luis_ joined #nim
19:21:52	disruptek	manual is a document worth reading at least once: https://nim-lang.org/docs/manual.html
19:21:57	disruptek	~manual is a document worth reading at least once: https://nim-lang.org/docs/manual.html
19:21:58	disbot	manual: 11a document worth reading at least once: https://nim-lang.org/docs/manual.html
19:23:42	disruptek	~arc is a new memory manager for Nim; see https://forum.nim-lang.org/t/5734
19:23:43	disbot	arc: 11a new memory manager for Nim; see https://forum.nim-lang.org/t/5734
19:25:19	*	luis_ quit (Client Quit)
19:25:53	*	luis_ joined #nim
19:28:37	leorize	disruptek: what is this new feature? :P
19:29:18	disruptek	just a little key/val database people can add to.
19:29:45	disruptek	try ~help
19:29:50	leorize	~help
19:30:07	leorize	nice
19:30:50	disruptek	we'll see...
19:32:18	FromGitter	<timotheecour> nim CI broke => https://github.com/nim-lang/Nim/pull/13052
19:32:19	disbot	➥ testament/important_packages dont run hts
19:33:00	disruptek	~leorize is responsible for an awesome NeoVim plugin at https://github.com/alaviss/nim.nvim
19:33:01	disbot	leorize: 11responsible for an awesome NeoVim plugin at https://github.com/alaviss/nim.nvim
19:33:42	leorize	make it `the awesome NeoVim plugin` :P
19:33:54	disruptek	~leorize is responsible for the awesome NeoVim plugin at https://github.com/alaviss/nim.nvim
19:33:55	disbot	leorize: 11responsible for the awesome NeoVim plugin at https://github.com/alaviss/nim.nvim
19:34:03	leorize	is it possible to make disbot print it out to the chat?
19:34:17	disruptek	print what out?
19:34:47	leorize	the added snippets?
19:34:54	*	filcuc joined #nim
19:35:08	disruptek	code snippets? from where?
19:35:39	disruptek	it migrates snippets from issues to the playground.
19:36:16	leorize	I mean the added texts to disbot? those ~term?
19:36:31	disruptek	you query for terms with just ~someterm:
19:36:33	disruptek	~leorize
19:36:34	disbot	leorize: 11responsible for the awesome NeoVim plugin at https://github.com/alaviss/nim.nvim -- disruptek
19:36:48	leorize	ooooh
19:36:58	leorize	I thought everything was going to be like ~help :P
19:37:36	disruptek	it was, originally, but usually people who know the terms don't need to know their definitions, so...
19:38:04	solitudesf	~manual
19:38:05	disbot	manual: 11a document worth reading at least once: https://nim-lang.org/docs/manual.html
19:40:42	FromDiscord	<has1> Can you create a dll file with nim? Instead of a binary?
19:40:57	disruptek	~concepts is user-defined type classes; documented in the experimental section of the manual: https://nim-lang.org/docs/manual_experimental.html#concepts
19:40:58	disbot	concepts: 11user-defined type classes; documented in the experimental section of the manual: https://nim-lang.org/docs/manual_experimental.html#concepts
19:41:08	disruptek	has1: yep.
19:41:22	FromDiscord	<has1> Something that you can call methods on easily?
19:41:54	disruptek	some good demonstrations of what's possible here: https://scripter.co/
19:41:58	FromDiscord	<has1> In .net you can load the assembly and they just call the methods of the dll file
19:42:37	disruptek	~kaushalmodi is a badass coder with a wealth of useful Nim articles on his blog, at https://scripter.co/
19:42:37	disbot	kaushalmodi: 11a badass coder with a wealth of useful Nim articles on his blog, at https://scripter.co/
19:43:13	FromDiscord	<has1> What am I looking at here
19:43:32	disruptek	i have no idea; hold the camera closer to the screen.
19:43:49	disruptek	it looks like porn to me.
19:43:55	disruptek	is that a horse?
19:44:01	disruptek	oh, it's a llama?
19:44:46	FromDiscord	<inv> Hello, duplicate from telegrams:
19:44:54	FromDiscord	<inv> I have a question about vs-code nim plugin. Isn't it right that it has tab for 4wide? Or something wrong with my vs.code
19:45:50	FromDiscord	<has1> Is that guy mental
19:46:07	FromDiscord	<has1> Why is he talking about porn in a programming discord
19:47:19	Zevv	has1: he is kind of bad publicity for the nim community every now and then
19:47:20	Araq	yeah disruptek, enough jokes for now
19:47:32	Zevv	hush, boy
19:48:10	Araq	has1: there is support for that via the dynlib.nim stdlib module
19:48:24	Araq	but it's not as easy as in .NET land I suppose
19:50:30	FromDiscord	<has1> This seems interesting, but what do I need to do to compile nim code so that I can call it this way?
19:52:18	Araq	https://nim-lang.org/docs/nimc.html#dll-generation
19:52:31	Araq	and read about the .dynlib and .exportc pragmas
19:52:56	Araq	and yeah, once again the results of our polls show documentation is lacking
19:53:04	Araq	so once again we'll improve it
19:53:54	disruptek	~dynlib is shared library creation; documented at https://nim-lang.org/docs/nimc.html#dll-generation
19:53:55	disbot	dynlib: 11shared library creation; documented at https://nim-lang.org/docs/nimc.html#dll-generation
19:54:16	FromDiscord	<has1> Thanks I'll check this out
20:01:07	*	hpyc9 is now known as Irving
20:02:04	*	Irving is now known as hpyc9
20:04:25	*	pbb quit (Ping timeout: 252 seconds)
20:07:27	leorize	Araq: btw on *nix Nim isn't hiding unexported symbols when building dlls
20:07:43	*	pbb joined #nim
20:07:56	leorize	should we just enforce -fvisibility=hidden then un-hide symbols like for windows?
20:10:58	*	pbb quit (Excess Flood)
20:11:06	*	pbb joined #nim
20:13:21	*	pbb quit (Excess Flood)
20:13:31	*	pbb joined #nim
20:15:38	*	nsf quit (Quit: WeeChat 2.7)
20:24:28	FromGitter	<deech> Does "there must be only one instance of the GC per process/address space" mean that I can't have a Nim executable that depends on more than one Nim library DLL?
20:28:49	shashlick	disruptek: that link is wrong for dynlib
20:29:06	shashlick	https://nim-lang.org/docs/dynlib.html
20:31:39	leorize	@deech no
20:31:46	leorize	it just means that don't mix pointers
20:32:08	leorize	a ref from a DLL is not compatible with one from the other afaict
20:32:46	leorize	things get messy once you start using threads though
20:32:58	shashlick	best to use boehm for this situation
20:33:09	leorize	I'd just recommend --gc:arc :P
20:33:14	shashlick	eventually yes
20:33:41	shashlick	feud has a main nim executable and a bunch of nim dlls all coexisting with memory moving around
20:36:00	*	pbb_ joined #nim
20:36:38	*	pbb quit (Ping timeout: 246 seconds)
20:37:19	*	filcuc quit (Ping timeout: 258 seconds)
20:44:54	*	luis_ quit (Quit: luis_)
20:46:13	*	marmotini_ joined #nim
20:47:12	nixfreak	does nim have a library like binascii with python?
20:47:50	*	Hideki_ joined #nim
20:49:35	FromDiscord	<mratsim> that seems like a pragmatic versioning scheme: https://calver.org/
20:50:39	FromDiscord	<mratsim> @nixfreak, you have some stuff in strutils and parseutils, otherwise in stew we have some parsers for hex
20:50:57	nixfreak	stew?
20:50:59	*	marmotini_ quit (Ping timeout: 265 seconds)
20:51:43	nixfreak	forgive my ignorance what is stew?
20:52:21	nixfreak	also does nim have a doc like javadoc ?
20:52:22	*	Hideki_ quit (Ping timeout: 268 seconds)
20:52:32	nixfreak	or just look online ?
20:52:49	*	pbb_ quit (Quit: No Ping reply in 180 seconds.)
20:54:06	*	pbb joined #nim
20:55:58	disruptek	!repo nim-stew
20:55:59	disbot	https://github.com/status-im/nim-stew -- 9nim-stew: 11stew is collection of utilities, std library extensions and budding libraries that are frequently used at Status, but are too small to deserve their own git repository. 15 11⭐ 2🍴
20:56:22	disruptek	there's no javadoc/pydoc yet, but you're welcome to contribute.
20:56:42	nixfreak	ok thanks
20:56:47	disruptek	!repo disruptek/ndoc
20:56:48	disbot	https://github.com/disruptek/ndoc -- 9ndoc: 11like pydoc but for nim 15 0⭐ 0🍴
20:57:49	*	tane joined #nim
20:58:22	*	sealmove joined #nim
20:59:03	*	pbb quit (Ping timeout: 252 seconds)
20:59:16	*	pbb joined #nim
21:00:42	nixfreak	ok ndoc is a WIP , but can't really use yet ?
21:00:53	disruptek	it's more of an RFC.
21:01:02	nixfreak	gotcha
21:04:01	sealmove	what's the standard equivelant to tables in C++?
21:04:28	sealmove	currently using `const char* myTable[]` duh...
21:04:46	sealmove	in combination with enums
21:05:55	sealmove	essentially I want a Table[myEnum, string]
21:07:02	sealmove	using const char* myTable[] I must be careful to not messup ordering...
21:07:05	disruptek	~dynlib is shared library creation; documented at https://nim-lang.org/docs/dynlib.html
21:07:05	disbot	dynlib: 11shared library creation; documented at https://nim-lang.org/docs/dynlib.html
21:07:05	sealmove	which is annoyin
21:07:19	leorize	sealmove: pretty sure they have unordered_map iirc
21:07:40	sealmove	so I should use STL?
21:07:59	disruptek	~nimrtl.dll DLL generation is documented briefly at https://nim-lang.org/docs/nimc.html#dll-generation
21:08:00	disbot	no footnotes for `nimrtl.dll`. 🙁
21:08:05	disruptek	~nimrtl.dll is DLL generation is documented briefly at https://nim-lang.org/docs/nimc.html#dll-generation
21:08:05	disbot	nimrtl.dll: 11DLL generation is documented briefly at https://nim-lang.org/docs/nimc.html#dll-generation
21:08:12	sealmove	leorize: thx
21:08:20	disruptek	sealmove: yes, use STL.
21:08:30	*	Ven`` joined #nim
21:08:55	FromDiscord	<Clyybber> ~disbot is a cool bot
21:08:55	disbot	disbot: 11a cool bot
21:09:07	disruptek	shashlick: footnotes are cumulative; you can add your own answers to terms.
21:09:19	disruptek	to correct knuckleheads such as myself.
21:10:30	disruptek	disbot only outputs the first three answers in channel, but privately can yield many more.
21:10:41	disruptek	clyybber: want to look at this arc bug?
21:10:50	sealmove	another thing.. is there a better way than stdarg.h to traverse command line arguments?
21:11:09	leorize	use nim :P
21:11:42	sealmove	i wish...
21:12:40	skrylar[m]	muffindrake: as requested https://gist.github.com/Skrylar/cb626152b7af8c413816eaa1f4f3377c
21:13:59	muffindrake	skrylar[m]: The issue was had was related to calling conventions, which was solved by a {.nimcall.} annotation
21:14:20	muffindrake	The issue I had*
21:14:34	muffindrake	Thanks, anyway.
21:14:48	*	Vladar quit (Quit: Leaving)
21:17:26	FromDiscord	<yewpad> @Araq: Do you have a clue what happened there? Does this happen because the CSS is now being written to a separate CSS file? https://github.com/nim-lang/Nim/issues/12998
21:17:28	disbot	➥ regression(1.04) `nim doc main.nim` generates broken html (no css) ; snippet at 12https://play.nim-lang.org/#ix=26CF
21:18:10	FromGitter	<timotheecour> @yewpad yes, that’s exactly what’s going on, shd be an easy fix, PR welcome
21:20:30	sealmove	lqdev: Just fyi I decided to abandom the idea of constructing a macro call in a macro. Yes, it would help me avoid having to implement a typechecker, but it messes up the overall code architecture and makes the code less readable...
21:20:57	sealmove	and less functional too
21:21:31	sealmove	was fun learning this technique though
21:21:39	*	pbb quit (Quit: No Ping reply in 180 seconds.)
21:22:56	*	pbb joined #nim
21:25:41	FromDiscord	<yewpad> It indeed is broken @timotheecour. Gonna look into it
21:47:14	*	narimiran quit (Ping timeout: 240 seconds)
21:49:16	*	krux02 quit (Remote host closed the connection)
21:55:29	*	vsantana joined #nim
22:00:01	*	sagax joined #nim
22:02:23	FromDiscord	<mratsim> wohoho I added dataflow graph parallelism to Weave
22:04:12	FromDiscord	<mratsim> this means that you can do continuations and trigger tasks only when their dependencies are resolved
22:10:55	skrylar[m]	soooo. futures. :3
22:11:47	*	Trustable quit (Remote host closed the connection)
22:14:23	FromGitter	<timotheecour> @leorize ⏎ ⏎ > leorize Araq: btw on *nix Nim isn't hiding unexported symbols when building dlls ⏎ ⏎ => see https://github.com/nim-lang/Nim/issues/13054 [https://gitter.im/nim-lang/Nim?at=5e13b13ee409da486cd75286]
22:14:25	disbot	➥ we need a cross-platform way to set dll symbol visibility ; snippet at 12https://play.nim-lang.org/#ix=26D0
22:14:45	FromDiscord	<mratsim> No futures are different @skrylar
22:15:01	FromDiscord	<mratsim> Futures are about control flow dependencies
22:15:11	FromDiscord	<mratsim> what I did was about data dependencies
22:16:00	FromDiscord	<mratsim> i.e. when you spawn a task with a future result, it's eagerly scheduled
22:16:01	disruptek	zedeus: regarding #12982, does your `$` actually work? looks like it's not defined for the new MultipartEntry type. also, wanna make that a ref and inherit from RootObj?
22:16:03	disbot	https://github.com/nim-lang/Nim/pull/12982 -- 3Implement file streaming for httpclient's MultipartData
22:16:39	FromDiscord	<mratsim> while what I did was allow tasks to be delayed until what they depends on is ready
22:19:11	zedeus	yeah, I modified the $ for MultipartData which is just a list of MultipartEntry, although the output probably isn't as useful as before
22:19:32	zedeus	why ref? and why inherit?
22:20:01	skrylar[m]	@mratsim well sounds neat anywya
22:20:24	skrylar[m]	was there still a need to fork rfc3339 to support iso dates
22:21:04	*	endragor joined #nim
22:21:56	disruptek	for extending the type.
22:25:24	disruptek	that cl constant could go into httpcore.
22:25:25	*	endragor quit (Ping timeout: 258 seconds)
22:28:53	zedeus	hmm sure, I can do that. you want it public too?
22:28:56	disruptek	prolly should have an addLine proc that includes the cl. anyway, this pr lgtm. nice upgrade; i wanted it numerous times.
22:29:21	disruptek	i bet it's a huge win in practical code.
22:30:19	disruptek	i wonder what else this knocks off my list of httpclient errors.
22:30:34	disruptek	!issue author:disruptek httpheaders
22:30:35	disbot	https://github.com/nim-lang/Nim/issues/12211 -- 3proposed changes to HttpHeaders as in HttpClient
22:31:31	FromGitter	<kaushalmodi> why is disbot outputting "3" before every message (I see this on gitter)
22:32:16	disruptek	it's an irc color code.
22:32:31	disruptek	actually, the color prefix. subsequent character is the color.
22:33:10	FromGitter	<kaushalmodi> those "3" and other numbers show in the IRC log too https://irclogs.nim-lang.org/
22:33:20	Yardanico	guess we need to fix NimBot too :P
22:33:39	*	tane quit (Quit: Leaving)
22:35:59	*	ptdel joined #nim
22:36:30	zedeus	disruptek: do you want the MultipartEntry type public? i'm hesitant to do that since it might cause users to try tinkering with its fields which they shouldn't
22:36:30	FromDiscord	<Clyybber> Araq: I'm not sure how to handle instructions in the cfg that don't have .sym set
22:37:08	disruptek	zedeus: it can start private and could be exposed later if necessary.
22:37:16	*	ltriant joined #nim
22:37:42	disruptek	nim's http is a mess that i don't want to fight about; anything is an improvement if it actually gets committed.
22:38:35	zedeus	sure, I'll let it stay non-ref for now then since it's passed around in a seq belonging to a ref object. I might be wrong but having the entries be refs as well shouldn't make any difference
22:38:55	disruptek	right, i think it's fine.
22:40:08	FromDiscord	<Clyybber> Araq: ping if you're here
22:40:46	FromDiscord	<Clyybber> `s/if/to see if
22:49:24	disruptek	why does arc want me to move() a str constant as in `newException(IOError, "str constant here")`?
22:58:05	*	endragor joined #nim
23:02:55	*	endragor quit (Ping timeout: 268 seconds)
23:04:10	FromDiscord	<Clyybber> can't reproduce
23:06:49	disruptek	gittyup is a source of bewilderment.
23:07:06	shashlick	Imagine how horses feel
23:07:27	FromDiscord	<Clyybber> or do you mean someConstStr and not "str literal" ?
23:07:27	*	nixfreak quit (Ping timeout: 260 seconds)
23:07:29	disruptek	double free or corruption (fasttop)
23:08:02	disruptek	i tried const foo = "bar" ... raise newException(... foo) ... same result.
23:08:35	disruptek	it seems that my problems are all in iterators.
23:11:15	FromDiscord	<Clyybber> hmm
23:13:37	FromGitter	<Albus70007> i have an hexadecimal string, but i need it to have the "\x" thing before every byte, how can i do it??
23:15:52	*	solitudesf quit (Ping timeout: 268 seconds)
23:16:08	leorize	@Albus70007 wdym?
23:16:15	*	marmotini_ joined #nim
23:18:31	disruptek	i will write a little thingy to translate irc color to pango and/or html. then someone and patch nimbot with it eventually.
23:18:39	disruptek	s/someone and/someone can/
23:19:19	*	pbb quit (Quit: No Ping reply in 180 seconds.)
23:20:36	*	pbb joined #nim
23:20:50	*	marmotini_ quit (Ping timeout: 240 seconds)
23:22:08	FromGitter	<Albus70007> @leorize i have a socket client that communicates with a python server, the server sends me data, and i need to send a buffer back, with a two bytes long header, hardcoding the header works (if the data stays the same obv), but if i try pass the header, the server wont interpret it as it should
23:23:27	nisstyre	Albus: not sure what the issue is. IF the size changes, you need to somehow encode that in the beginning of the data being sent
23:23:46	FromGitter	<matrixbot> `arkaroo` and the header needs to be bytes, not a string?
23:23:48	nisstyre	e.g. as 2 bytes representing a little endian integer or something
23:23:56	nisstyre	or however large you need it to be
23:24:22	leorize	arkaroo: join #freenode_#nim:matrix.org instead of relaying to gitter :P
23:25:17	FromGitter	<Albus70007> yes, i know all of that, i even have the proper header, but i need it to have "\x" at the beggining of each byte because if it doesnt, the server messes up decoding it
23:25:18	nisstyre	or are the 2 bytes the length already? in which case you calculate how large it needs to be
23:25:22	*	arkaroo[m] joined #nim
23:26:00	FromGitter	<Albus70007> the 2 bytes are the lenght yes
23:26:11	FromGitter	<Albus70007> want the code?
23:26:21	nisstyre	Sure
23:26:42	*	Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
23:27:20	*	Ven`` joined #nim
23:27:29	*	Ven`` quit (Client Quit)
23:27:31	FromGitter	<Albus70007> var inputs = bot.inputs ⏎ inputs.get_inputs(game_info) ⏎ let strinputs: string = $tojson(inputs) ⏎ let header: string = (len(strinputs) + 2).int16.toHex() ⏎ echo(header) ... [https://gitter.im/nim-lang/Nim?at=5e13c2630ba7b60f2035a359]
23:27:36	nisstyre	Albus: you might find this useful, https://github.com/weskerfoot/ADBTool/blob/b74aa9a91a5f7062b786b429009a02d1f184fcd5/src/adbtoolpkg/adb.nim#L60-L63
23:27:47	nisstyre	I wrote it a little while ago to do something similar
23:28:11	nisstyre	instead of 4 bytes you could do the same for 2
23:28:35	FromGitter	<Albus70007> how
23:28:43	nisstyre	the way it would be used would be like `text.len.unrollBytes`
23:28:45	*	uu91 quit (Remote host closed the connection)
23:28:50	nisstyre	which would give you a byte string
23:28:51	*	arecaceae quit (Quit: WeeChat 2.0.1)
23:28:56	nisstyre	that you could then send over the wire
23:29:42	*	arkaroo[m] left #nim (#nim)
23:30:04	FromGitter	<Albus70007> the main problem is that i need it to be in utf8 therefore have this "\x" between bytes
23:30:13	FromGitter	<matrixbot> `arkaroo` @leorize looks like the freenode room munges code formatting, while the gitter one respects it :P
23:30:26	nisstyre	FromGitter: yeah I see
23:30:27	FromGitter	nisstyre, I'm a bot, *bleep, bloop*. I relay messages between here and https://gitter.im/nim-lang/Nim
23:30:30	nisstyre	oops
23:30:46	FromGitter	<Albus70007> so this header; 0074 would look like this \x00\x74
23:32:19	FromGitter	<Albus70007> and i cant just do something like buffer.add("\x") because nim considers it as a typo
23:33:16	FromGitter	<matrixbot> `arkaroo` maybe I'm missing something here, but it seems like this needs to be solved on the python side? if python is respecting the \x syntax, doesn't that tell you tha python is interpreting the data as a string, not a byte sequence?
23:33:29	nisstyre	yeah that's the part I find a bit confusing
23:34:03	FromGitter	<matrixbot> `arkaroo` I'm not sure whether you're dispatching it as a byte sequence or a string from the nim end, but I'm pretty sure you're reading it as a string on the python end
23:34:44	FromGitter	<Albus70007> python decodes it with header = struct.unpack("H", recieved[:2])[0]
23:35:47	nisstyre	so it's interpreting it as an unsigned short
23:35:51	FromGitter	<matrixbot> `arkaroo` yeah, so you;re trying to unpack a string. if it's arriving as a byte sequence, you shouldn't need to unpack it
23:35:55	nisstyre	why would it be utf-8 then?
23:36:32	FromGitter	<Albus70007> thats how the python client seems to do it
23:36:54	FromGitter	<Albus70007> length = len(jString.encode('utf-8'))header = struct.pack("H",length)
23:37:16	nisstyre	so the data is utf-8
23:37:32	FromGitter	<Albus70007> at least in python, yes
23:37:49	nisstyre	and the length is a 2 byte sequence representing an unsigned short (in what byte order?)
23:38:33	FromGitter	<Albus70007> the opposite to the one nim uses, but i already flipped their position
23:44:01	nisstyre	Albus: Is this not what you want then? https://play.nim-lang.org/#ix=26DD
23:44:40	nisstyre	that gives you the length as 2 bytes (unrolled) and converted to a string which you can prepend to the data
23:44:57	nisstyre	which is exactly what that python is doing
23:45:10	nisstyre	except maybe not in the correct byte order
23:46:25	FromGitter	<Albus70007> i will try it, i have a very pythonic way of solving things and i am not used to dealing with bytes at all yet, starting in such a high level language is being painful now
23:47:25	nisstyre	no problem, and FWIW, there are probably better ways of doing this, I think there are a few libraries for dealing with parsing/serializing binary protocols
23:47:31	nisstyre	but it's probably overkill for this
23:48:19	FromGitter	<matrixbot> `arkaroo` cap'n proto has nim support I think?
23:48:31	nisstyre	capnp is great
23:48:43	FromGitter	<matrixbot> `arkaroo` >last updated 2 years ago ⏎ situation unclear :P
23:48:53	FromGitter	<matrixbot> `arkaroo` * > last updated 2 years ago ⏎ situation unclear :P
23:49:07	nisstyre	yeah I actually tried building it and it needed a compiler flag to build
23:49:13	nisstyre	I think I left a comment on the issue
23:49:35	nisstyre	https://github.com/zielmicha/capnp.nim/issues/8#issuecomment-562317259
23:49:36	disbot	➥ When using latest nim "[...] util.nim(81, 12) Error: 'nil' is now invalid for 'string';"
23:52:42	FromGitter	<Albus70007> i think it worked, thank you <3
23:52:48	nisstyre	awesome
23:53:26	*	ng0 quit (Quit: leaving)
23:55:33	*	marmotini_ joined #nim
23:55:54	*	Hideki_ joined #nim
23:56:30	FromDiscord	<Fern & Simula (They/Them)> why does so much of the stdlib define data types like this? https://github.com/nim-lang/Nim/blob/devel/lib/pure/ropes.nim#L33
23:56:46	FromDiscord	<Fern & Simula (They/Them)> why not just have a `Rope = ref object` ?
23:58:38	FromDiscord	<mratsim> Legacy
23:58:47	FromGitter	<Varriount> Because there are times where using an object is faster, (less heap allocation)
23:58:53	FromDiscord	<mratsim> before you coudln't do ref object followed by the field
23:59:46	FromGitter	<Varriount> Also, I believe that originally ref types could only be specified in procedure parameters and type bodies