<< 14-07-2018 >>

---

00:00:10	CodeVance	toif your lazy use converters if you aren't lazy then wrap the proc
00:00:27	CodeVance	exelotl
00:01:40	stefanos82	Araq, out of curiosity: did you ever use OCaml? If yes, what role did it play to your coding as a programmer, let alone as a researcher?
00:04:27	*	dddddd quit (Remote host closed the connection)
00:05:22	Araq	never really used it, its syntax was dynamite for me. However, I did use Standard ML
00:05:55	stefanos82	I have found many similarities to Nim's syntax in so many ways.
00:06:01	stefanos82	not all of them, but quite a few
00:06:14	stefanos82	so far Nim has the best syntax
00:06:26	stefanos82	I have even dare to say that Nim's syntax is way better than Python's
00:06:52	Araq	yay. :-)
00:07:17	Araq	the similarities to ML are due to ML's Pascal heritage and Pascal influenced Nim too.
00:07:39	Araq	other stuff like 'f a' for 'f(a)' were later additions
00:07:45	stefanos82	I know that you used Pascal to build the first versions of it, isn't that correct?
00:07:51	FromGitter	<kayabaNerve> Araq: What would you think about implicit conversion support as long as it doesn't make it into the stdlib/compiler?
00:08:14	Araq	kayabyNerve we have lenientops in the stdlib.
00:08:18	FromGitter	<kayabaNerve> Meant to ask that since I'm working with a bunch of different representations of the exact same data.
00:08:19	stefanos82	no implicit conversions -_- explicit is always better than implicit
00:09:13	Araq	stefanos82: I agree but I think it's fair if it's an explicit module import that gives you implicit conversions
00:09:34	FromGitter	<kayabaNerve> I mean like this: ⏎ proc myProc(x: typeX) ⏎ ⏎ implicit proc (x: typeX, y: typeY) ⏎ ... [https://gitter.im/nim-lang/Nim?at=5b493f3e641ecc065fa1571c]
00:09:45	FromGitter	<kayabaNerve> Completely ignore the syntax on the second proc please
00:10:16	stefanos82	Araq: fair enough. Are they labeled or marked as "partially unsafe" though? Because an implicit conversion could open so many issues for you behind the scenes, without you realizing it
00:10:44	FromGitter	<kayabaNerve> That said, I have to go change something to SomeInteger. Couldn't find it earlier but now I have it. Thanks.
00:10:53	stefanos82	this brought back some early C days with implicit value rounding and type casts -_-
00:11:39	FromGitter	<kayabaNerve> BTW, the reason I'm thinking of this is just because I have three number bases and it's a pain to constantly call conversions... I don't actually want it in the standard or not. Just wanted to ask about it
00:12:26	Araq	kayabaNerve: are you looking for Nim's "converter" keyword?
00:13:12	Araq	stefanos82: sometimes one needs to hack though, esp when interfacing with C that had no type discipline
00:13:19	FromGitter	<kayabaNerve> Possibly
00:13:26	Araq	and then you reach for lenientops or similar
00:13:41	stefanos82	Araq: I know; that's why I mentioned above my C memories :/
00:14:18	*	krux02 quit (Remote host closed the connection)
00:16:29	stefanos82	Araq: btw, what is the actual objective or should I say, your ultimate mission with Nim's future? Faster compilation times or stabilizing the language by fixing its semantic design until 1.0's release?
00:17:04	stefanos82	of course feel free to replace 1.0 with NG (next generation) if you like
00:17:48	FromGitter	<kayabaNerve> Araq: TIL! That's exactly what I was talking about. Thank you!
00:18:18	Araq	stefanos82: please read the github milestones for the plan :-)
00:19:14	stefanos82	Araq: don't you have any TLDR; or something similar? Some terminologies are too advanced for my silly brain :/
00:19:44	CodeVance	u can call it controlled implicit
00:20:22	CodeVance	There's also araq's musings @stefanso
00:20:26	CodeVance	stefanos82:
00:21:17	stefanos82	actually, thanks to Araq it's not implicit nor explicit; it should be considered as complicit!
00:21:29	stefanos82	because basically they are collaborating with you to do your job ;)
00:21:39	CodeVance	lol
00:21:46	CodeVance	I didn
00:21:57	CodeVance	didn't know of lenientops until now
00:22:05	stefanos82	me neither
00:22:12	stefanos82	but as you can see, I'm learning every day
00:22:26	CodeVance	nim isn't batteries included, It's super mech robot included
00:23:54	Araq	CodeVance: the musings are old though, I need to write new ones
00:24:20	CodeVance	I mentioned it in relation to your 'vision'
00:25:42	stefanos82	Araq: I have just read about v2 milestone, about mapping the macro system to a compiler plug-in. Basically, will it work as a background service that will be communicating back and forth with the core, much like we do with RPC protocol??
00:26:04	stefanos82	weird...I inserted one "?" and printed 2
00:26:11	FromGitter	<rayman22201> Pegs module is awesome btw. Exactly what I wanted. Much more clear than regex for huge log file parsing that needed.
00:26:39	Araq	stefanos82: that's one way of thinking about it
00:27:07	stefanos82	Araq: that could increase the building process by orders of magnitude in my opinion
00:27:31	FromGitter	<kayabaNerve> Thanks again for converters.
00:27:57	Araq	we'll see about that, stefanos82. it's all pretty cacheable.
00:28:47	stefanos82	I see a bright future for Nim. It goes to the right direction
00:29:11	stefanos82	the best thing you have decided to do with Nim is the whole metaprogramming concept that uses macros and templates
00:29:24	stefanos82	and generics, let's not forget those too!
00:30:26	Araq	generics should be mapped to macros :P
00:30:38	Araq	we'll get there...
00:32:40	stefanos82	without knowing much about macros yet (I have just started reading the chapter about them), I was going to say the same thing since macros allow you to do pretty much everything
00:38:39	FromDiscord	<exelotl> CodeVance: thanks, I'll give that a try (I'm nearly at the metaprogramming chapter in the nim book, wasn't expecting to have to use it yet haha)
00:39:46	CodeVance	exelotl that isn't meta programming yet
00:42:32	CodeVance	TBH I don't see how generics can be mapped to macros. Unless you make every macro able to generate custom generics. Some kind of compile time hook?
00:43:25	zacharycarter	I think it's possible
00:43:56	zacharycarter	I mean - if they're possible without macros they should definitely be possible with macros
00:44:06	zacharycarter	not like generics need runtime info to function
00:44:39	zacharycarter	macros are already evaluated at compile time so
00:45:15	FromDiscord	<exelotl> wait using a nim template isn't metaprogramming?
00:45:24	CodeVance	ya you can get the same functionality of generics procs by using `any` param types and when statements
00:45:27	stefanos82	the objective of Araq's is to move the macro system as a background service.
00:45:56	zacharycarter	templates are metaprogramming
00:46:18	zacharycarter	but you'll see the difference between them and macros when you get to macros
00:46:19	stefanos82	so logically the macro system will be already warmed up for us *before* we start compiling actual Nim code
00:46:26	zacharycarter	I find templates much more usable than macros honestly
00:46:31	stefanos82	therefore it *should be* considerably fast
00:46:35	zacharycarter	or at least - I find myself using them way more often
00:46:36	CodeVance	exelotl I said wrap. wrappers don't need templates. (though they might be cleaner)
00:46:49	Araq	no, not a "background" service. don't spread myths here.
00:46:58	Araq	but I need to sleep, we can talk later
00:47:02	Araq	good night
00:47:07	stefanos82	my apologies Araq, but when you said plug-in, that's how I interpreted it
00:47:13	CodeVance	night
00:47:36	CodeVance	I see templates as compiler copy paste
00:48:01	*	BitPuffin quit (Remote host closed the connection)
00:48:06	FromDiscord	<exelotl> oh right, yeah I could just wrap it in another proc of course, just wondered if there's a more idiomatic way to do it x)
00:48:40	FromDiscord	<exelotl> or maybe that *is* the idiomatic way
00:48:41	CodeVance	there is not idiomatic. just domain specific.
00:48:55	FromDiscord	<exelotl> ah fair enough
00:48:59	CodeVance	not for nim at least
00:50:25	CodeVance	So your target is to make the library easy to use. if there are a bunch of functions with 0 and 1 instead of true false then it would be easier to use macros
00:51:32	FromDiscord	<exelotl> I really like that philosophy but it makes learning difficult because you don't even know what possibilities are available to you until you have more experience with the language
00:51:49	CodeVance	exelotl BTW go back and read "statement list expression" in the manual. Unless you have already done so. That explains how all the nim statements and expressions are laid out
00:52:03	skrylar	if you can get away with a tempmlate you should do so, because macros require effort :p
00:52:10	skrylar	and also suffer breakage with compiler updates
00:52:16	CodeVance	no they don't. not really
00:52:27	CodeVance	Just don't make a macro in a current project
00:52:34	CodeVance	make a new file
00:52:39	CodeVance	then compiling is much faster
00:53:08	FromDiscord	<exelotl> well, I'm aware that the common advice is to use the least powerful tool for the job when it comes to metaprogramming in nim
00:53:08	skrylar	well in rebol you just invoke parse and splat out a DSL on the spot. in nim you have to sit there and dumpTree for a while and pick apart the ASTs you get (thus, effort)
00:53:37	CodeVance	skrylar: hint reprAstGen
00:53:49	skrylar	and krux was literally talking earlier about said breakages =p
00:53:50	CodeVance	astgenrepr
00:54:45	FromDiscord	<exelotl> but yeah I really want to make a game engine that leverages the power of macros but the macro system is really daunting to me, it's one of the main reasons I'm reading the book
00:54:48	*	ofelas joined #nim
00:57:05	skrylar	its handy but mysterious :\
00:57:13	CodeVance	hint astgenrepr. I don't think it gets easier. then hint repr result at the end of the macro. if it doesn't print its not valid
00:57:28	skrylar	it's too late to get it but it might have been nicer to just get the token stream instead inside a macro, because you can get weird shit
00:57:45	CodeVance	But like i said. depends on how many of the 0/1 true/false things there are
00:57:49	skrylar	i had a prototype of a FSM generator and the compiler got picky about pseudo function parameter groups
00:58:25	CodeVance	skrylar: psuedo function?
00:59:04	skrylar	it would accept foo(a, b, c: T, d: Y) but not foo(a, b, c: T; d: Y)
00:59:21	FromGitter	<rayman22201> you guys want a lisp reader macro system in Nim lol
00:59:47	CodeVance	rayman ther's already a list reader in nim . Nim itself XP
00:59:51	CodeVance	lisp
01:00:14	CodeVance	(echo 1,2,3,4)
01:00:21	FromGitter	<rayman22201> https://gist.github.com/chaitanyagupta/9324402
01:00:23	FromGitter	<rayman22201> no that
01:01:11	FromGitter	<rayman22201> It's a type of macro that operates on the lexer instead of the parser
01:04:10	CodeVance	ya. You'd have to change the import the compiler for that
01:04:13	CodeVance	XD
01:05:18	FromGitter	<rayman22201> It would be a big change. I don't think Araq would go for it lol. Maybe in V2 when the compiler is all plugin based :-P
01:08:55	*	stefanos82 quit (Quit: Quitting for now...)
01:34:16	FromGitter	<kaushalmodi> exelotl: If you are somewhat familiar with Nim, you can skip to the last "9 Metaprogramming" section of the book. It starts out with very basics of templates and then macros. I happen to have just reached that section; currently on 9.4.1, 8% of the book to go (Kindle).
01:43:54	*	vivus quit (Quit: Leaving)
05:08:47	*	csgator joined #nim
05:10:48	*	lompik joined #nim
05:10:55	*	miran joined #nim
05:17:26	*	fastrom joined #nim
05:36:02	*	nsf joined #nim
05:51:14	*	donzx quit (Remote host closed the connection)
05:51:36	*	donzx joined #nim
06:01:20	*	csgator quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
06:25:14	*	donzx quit (Remote host closed the connection)
06:25:53	*	donzx joined #nim
06:25:53	*	donlzx joined #nim
06:26:26	*	donlzx quit (Client Quit)
06:26:30	*	donzx quit (Client Quit)
06:26:50	*	lompik quit (Quit: WeeChat 2.1)
06:26:54	*	donlzx joined #nim
06:36:32	*	Vladar joined #nim
07:11:10	FromGitter	<tim-st> Can someone confirm that devel nim cannot compile filenames with umlauts and just hangs forever?
07:11:26	FromGitter	<tim-st> @Araq I think this is the problem from yesterday
07:12:26	*	fastrom quit (Quit: Leaving.)
07:12:32	FromGitter	<tim-st> nim git hash: dd47013017d75a2b1a8c80e12f273e18e86f7d72
07:22:29	FromGitter	<bevo009> Just tried out the jester sample code, I get the same error on windows 10 and linux subsystem (Nim 0.18.0)...can anyone confirm? ⏎ .nimble\pkgs\jester-0.3.0\jester.nim(5, 38) Error: cannot open 'macrocache'
07:24:52	*	Trustable joined #nim
07:25:15	FromGitter	<tim-st> I think I read this requires devel
07:25:24	FromGitter	<bevo009> just read that
07:25:57	FromGitter	<bevo009> how do i downgrade to upgrade to v0.2.1
07:26:47	FromGitter	<bevo009> nimble install jester 0.2.1?
07:26:50	FromGitter	<Varriount> @rayman22201 I'll be merging in that PR soon. I just need to test it locally.
07:27:51	FromGitter	<tim-st> Hm, devel can compile files with umlauts, there must be another problem :\
07:28:01	FromGitter	<tim-st> really weird
07:32:30	FromGitter	<tim-st> Maybe the problem is: vscode+umlautsFilename; can someone try it?
07:34:08	FromGitter	<bevo009> well `nimble install jester 0.2.1` doesn't work, I still get 3.0...what's the correct command?
07:34:24	FromGitter	<tim-st> maybe ==0.2.1 ?
07:35:22	FromGitter	<tim-st> no
07:36:55	FromGitter	<tim-st> here is the example: nimble install nimgame@"> 0.5"
07:37:36	FromGitter	<tim-st> nimble install [email protected]
07:38:02	FromGitter	<bevo009> yep ta, `nimble install jester@"0.2.1"`
07:38:15	FromGitter	<bevo009> thanks Tim ;)
07:38:41	FromGitter	<bevo009> maybe I could have dropped the quotes as well
07:40:43	FromGitter	<tim-st> yes, seems only needed for `>` etc
07:43:43	FromGitter	<tim-st> but jester < 0.3.0 has security problems, only use this version for testing
07:53:15	FromGitter	<bevo009> now i get this error: `routes.nim(3, 1) template/generic instantiation from here ⏎ lib\core\macros.nim(384, 57) Error: undeclared identifier: 'Future'`
07:56:07	FromGitter	<kayabaNerve> How do I run the debug compiler (getting SIGSEGVs and no idea why)
07:56:37	FromGitter	<kayabaNerve> Well, to be honest, it's because I'm butchering Nim. That said, how do I run the debug compiler?
07:57:48	FromGitter	<kayabaNerve> By debug compiler, I mean, isn't there a way under koch to print a stack trace/line trace when a SIGSEGV occurs?
08:01:56	*	nsf quit (Quit: WeeChat 2.1)
08:02:01	FromGitter	<tim-st> compile in debug mode?
08:02:30	FromGitter	<tim-st> I think only in release mode information is missing
08:03:28	FromGitter	<bevo009> ay caramba, why doesnt this work?
08:03:40	FromGitter	<bevo009> maybe I should downgrade to 2.0 again
08:08:22	FromGitter	<bevo009> nope, that didn't work either
08:08:24	FromGitter	<bevo009> Shouldn't this sample code work on jester 0.2.1?
08:08:36	FromGitter	<bevo009> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b49af836534bc5d2e5a276c]
08:26:38	FromGitter	<bevo009> got it working, this is the correct example code for version 0.18.0 / 0.2.1 ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b49b3bd66c1e833a9e76134]
08:27:42	Araq	kayabaNerve: koch temp c foo.nim
08:28:07	Araq	produces a debug version of the compiler, passes 'c foo.nim' to it and this debug version produces a stacktrace
08:34:02	*	xet7 joined #nim
08:36:13	FromGitter	<tim-st> Araq: I couldnt find out the problem, but there is definetly one, that for some code interrups compilation forever at devel
08:40:00	FromGitter	<tim-st> It seems if umlaut or ß or similiar in filename and something like this code, it fails: https://gist.github.com/tim-st/a7e4300a68e8c2b61381ff79ff71e1ca
08:40:36	FromGitter	<tim-st> when the filename is made normal, the same code echos 0
08:41:10	FromGitter	<tim-st> if not it compiles forever
08:45:00	FromGitter	<tim-st> I tried other source code with special filename like `echo 1` and it worked, both things filename and source code seems related to this thing
08:45:19	FromGitter	<tim-st> *special filename and source code `echo 1`
08:45:38	*	dorelix joined #nim
08:48:13	Araq	tim-st: which OS?
08:48:23	FromGitter	<tim-st> win7x64
08:48:39	FromGitter	<tim-st> gcc (x86_64-posix-seh-rev0, Built by MinGW-W64 project) 8.1.0
08:49:30	FromGitter	<tim-st> Nim Compiler Version 0.18.1 [Windows: amd64] ⏎ Compiled at 2018-07-13 ⏎ Copyright (c) 2006-2018 by Andreas Rumpf ⏎ ⏎ git hash: dd47013017d75a2b1a8c80e12f273e18e86f7d72 ... [https://gitter.im/nim-lang/Nim?at=5b49b91995e03e3d7b49a3ac]
08:52:24	FromGitter	<tim-st> I have deleted nimcache and used [Running] nim c --verbosity:0 --hints:off --opt:none -r
08:52:46	*	csgator joined #nim
08:53:08	FromGitter	<tim-st> the same source code with only-ascii filename works expected :\
08:54:07	*	csgator quit (Client Quit)
08:55:06	FromGitter	<tim-st> I remember there was a commit that allowed non-ascii filenames
08:59:07	FromGitter	<tim-st> this is a minimum not working code:
08:59:10	FromGitter	<tim-st> import re
08:59:11	FromGitter	<tim-st> let htm = "" ⏎ let m = htm.findAll(re"<P (.*?)</P>") ⏎ echo m.len
08:59:23	FromGitter	<tim-st> together with filename nonascii
09:00:38	*	donlzx quit (Remote host closed the connection)
09:01:04	*	donlzx joined #nim
09:01:40	FromGitter	<tim-st> with `regex` instead of `re`it works
09:02:04	*	ofelas quit (Quit: shutdown -h now)
09:02:08	FromGitter	<Lite5h4dow> ok, so i decided to dive into nuklear nim. however im having issues. i think im just dimb as fuck but i dont understand how binds work
09:02:19	FromGitter	<Lite5h4dow> do they change the format of the c commands?
09:02:46	FromGitter	<Lite5h4dow> because the source for nuklear nim liteally just includes the c srce and dependancies
09:02:59	FromGitter	<Lite5h4dow> yet nk_begin dosent even work
09:03:22	FromGitter	<Lite5h4dow> do i just have shit for brains or am i missing something here?
09:03:53	*	ofelas joined #nim
09:05:03	FromGitter	<kayabaNerve> Thanks Araq
09:05:38	*	donlzx quit (Remote host closed the connection)
09:06:05	*	donlzx joined #nim
09:06:40	FromGitter	<tim-st> @kayabaNerve are you on windows?
09:10:21	*	NimBot joined #nim
09:17:08	*	donlzx quit (Remote host closed the connection)
09:17:34	FromGitter	<bevo009> does jester have hot reload?
09:17:34	*	donlzx joined #nim
09:20:19	*	dddddd joined #nim
09:21:01	*	BitPuffin joined #nim
09:22:57	FromGitter	<Varriount> @Lite5h4dow What stage are you running into problems at? Compilation, or execution?
09:23:00	*	nsf joined #nim
09:23:38	FromGitter	<Varriount> @bevo009 For code, probably not. For static files, probably.
09:23:59	FromGitter	<bevo009> thanks
09:32:36	FromGitter	<Lite5h4dow> @Varriount writing. i keep getting linter errors saying this and that dosent exist
09:32:43	FromGitter	<Lite5h4dow> as a result it dosent even let me compile
09:33:06	FromGitter	<Lite5h4dow> (it does but it fails anyway)
09:34:56	FromGitter	<Varriount> @Lite5h4dow You are using some sort of IDE or build system?
09:35:19	FromGitter	<Lite5h4dow> atom with the nim addon and linter addon
09:35:51	FromGitter	<Varriount> Ok. And you are attempting to use https://github.com/zacharycarter/nuklear-nim?
09:36:31	FromGitter	<Lite5h4dow> correct
09:37:56	FromGitter	<Lite5h4dow> am i doing something wrong?
09:39:08	FromGitter	<Varriount> How did you install the package - through Nimble?
09:39:11	FromGitter	<Lite5h4dow> yes
09:39:24	FromGitter	<Lite5h4dow> `nimble install https://github.com/zacharycarter/nuklear-nim`
09:41:35	FromGitter	<Varriount> Hm, I'm wondering if Nimble checked out the sub module.
09:43:40	FromGitter	<Varriount> @Lite5h4dow Can you look in the Nimble directory and check if the nuklear submodule was downloaded?
09:44:02	FromGitter	<Lite5h4dow> hold on
09:44:34	FromGitter	<Varriount> Also, can you post a gist or paste on the errors that are being thrown during compilation?
09:44:44	FromGitter	<bevo009> nimble path nuklear-nim
09:47:04	FromGitter	<Lite5h4dow> `PS C:\WINDOWS\system32> nimble path nuklear-nim ⏎ ⏎ ``` Error: At least one of the specified packages was not found```` [https://gitter.im/nim-lang/Nim?at=5b49c698ba5f154b3b9e3f01]
09:47:15	FromGitter	<Lite5h4dow> but its installed
09:47:20	FromGitter	<Lite5h4dow> so im super confused rn
09:47:22	*	fastrom joined #nim
09:47:40	FromGitter	<bevo009> exact same name?
09:47:48	FromGitter	<Varriount> Try just "nuklear"
09:48:28	FromGitter	<Lite5h4dow> just the bind c file
09:48:37	FromGitter	<Lite5h4dow> didnt include the actual nuklear library
09:48:54	FromGitter	<Lite5h4dow> i take it i just need to download the nuklear source into that folder
09:49:25	FromGitter	<Varriount> No, you should just be able to do "nimble install nuklear"
09:49:55	FromGitter	<Lite5h4dow> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b49c742641ecc065fa23dd4]
09:51:07	*	fastrom quit (Client Quit)
09:53:00	*	fastrom joined #nim
09:55:28	FromGitter	<bevo009> Error: Specified directory (C:\Users\bevo\AppData\Local\Temp\nimble_17220\githubcom_vurtunnuklear\) does not contain a .nimble file.
09:55:45	FromGitter	<bevo009> Is that the problem?
09:56:09	FromGitter	<Lite5h4dow> how did you create that error
09:56:25	FromGitter	<Lite5h4dow> i want to try and replicate that to see ifd that is indeed the issue
09:56:37	FromGitter	<bevo009> I don't think there's a nimble file @ https://github.com/vurtun/nuklear
09:56:57	FromGitter	<bevo009> I ran this: nimble install https://github.com/vurtun/nuklear/
09:56:57	FromGitter	<Lite5h4dow> there wouldnt be
09:57:01	FromGitter	<Lite5h4dow> its a c livrary
09:57:25	FromGitter	<Lite5h4dow> im using https://github.com/zacharycarter/nuklear-nim
09:57:47	FromGitter	<Lite5h4dow> its a bind for the c library
09:57:51	FromGitter	<bevo009> fair enough, I thought it might depend on nuklear
10:00:19	*	PMunch joined #nim
10:02:11	FromGitter	<tim-st> anyone get this file compiled with the same filename: https://gist.github.com/tim-st/162abe7f0b498477ccd5391380d1fb55
10:02:45	*	Vladar quit (Quit: Leaving)
10:03:50	FromGitter	<kayabaNerve> @tim-st Yes, why?
10:04:17	FromGitter	<tim-st> on which os, nim version, compiler version?
10:04:23	FromGitter	<tim-st> for me it doesnt work...
10:04:40	FromGitter	<kayabaNerve> What doesn't?
10:04:56	FromGitter	<tim-st> it doesnt compile with filename `ää.nim`
10:05:01	FromGitter	<tim-st> only with `aa.nim`
10:06:06	*	nc-x joined #nim
10:06:15	nc-x	@tim-st doesn't compile for me on windows as well
10:06:33	FromGitter	<tim-st> nice, thanks!
10:06:42	FromGitter	<tim-st> so it's not my problem
10:06:43	nc-x	nim issues the following c command - gcc.exe -c -w -mno-ms-bitfields -IC:\Nim\lib -o C:\Users\alpha\AppData\Local\Temp\nimcache\ää\ää.c.o C:\Users\alpha\AppData\Local\Temp\nimcache\ää\ää.c
10:06:59	nc-x	in this command the path is wrong
10:07:04	FromGitter	<tim-st> nc-x: but try the same filename with just `echo 1` it works
10:07:05	FromGitter	<bevo009> gets to here then stalls: CC: ää
10:07:14	FromGitter	<tim-st> so the path isnt wrong
10:07:25	FromGitter	<tim-st> also when replacing by `import regex` it works
10:07:38	FromGitter	<tim-st> the problems is different, than just filepath
10:07:44	nc-x	yeah
10:07:52	FromGitter	<tim-st> really weird problem
10:08:22	FromGitter	<Lite5h4dow> @Varriount i downloaded the source and added it to the folder and it fixed it
10:09:15	FromGitter	<Lite5h4dow> wait no
10:09:16	FromGitter	<Lite5h4dow> nevermind
10:09:19	FromGitter	<Lite5h4dow> still broken
10:09:46	FromGitter	<Lite5h4dow> i wonder if @zacharycarter could shed some light
10:09:56	FromGitter	<kayabaNerve> I had working code; I spent 4 hours upgrading it; it doesn't work anymore.
10:10:15	FromGitter	<kayabaNerve> Never ever upgrade your code or change everything. Right now, I'm switch back to Ubuntu 8.04
10:10:32	nc-x	Actually sorry. The path is not wrong. Windows cmd does not display unicode properly. After chcp 65001 I see correct path.
10:11:03	FromGitter	<tim-st> maybe @GULPF knows more about it ( https://github.com/nim-lang/Nim/issues/7897 )
10:11:38	FromGitter	<tim-st> of course I used devel version, where it should work
10:11:39	nc-x	Compiling with the gcc command gives a lot of errors
10:11:58	FromGitter	<tim-st> didnt see one error with verbosity:3
10:12:19	nc-x	`error: stray '\244' in program` and same for '\303`
10:12:42	nc-x	@tim-st Compile with --listCmd and then copy the gcc command and use it
10:12:47	FromGitter	<tim-st> but why does it work with other source code? is it `re` only problem?
10:13:45	FromGitter	<tim-st> --listCmd shows no error at all
10:13:55	FromGitter	<tim-st> just stays compiling
10:14:34	nc-x	copy the gcc command and use it
10:15:02	FromGitter	<tim-st> oh, ok, I guessed verbosity:3 gave me this too
10:19:00	FromGitter	<Varriount> @Lite5h4dow I'm going to see if I can install that library
10:21:18	nc-x	@tim-st I tried with msvc. The compiler does not hang but give similar error to gcc. I think (all?) C compilers do not support unicode in identifiers.
10:22:03	FromGitter	<tim-st> but why does `ä.nim` -> `echo 1` work?
10:22:27	FromGitter	<tim-st> https://github.com/nim-lang/Nim/issues/8318
10:23:11	nc-x	Because the problem is with name mangling. The function name findall from regex is getting the `ää` appended to it. And the C compilers are choking on that.
10:23:45	FromGitter	<tim-st> ah, ok, but it should work easily when escaping filenames by hex
10:23:50	FromGitter	<tim-st> filename.toHex
10:23:50	*	Vladar joined #nim
10:24:03	FromGitter	<tim-st> at least for special chars
10:27:56	*	xet7 quit (Quit: Leaving)
10:28:52	FromGitter	<Varriount> @Lite5h4dow I was able to install the package, do you have an example I could run?
10:30:38	nc-x	@Varriount the nuklear-nim repo contains an example - https://github.com/zacharycarter/nuklear-nim/blob/master/examples/nuklear_nim_examples/glfw3_opengl3.nim
10:31:20	nc-x	And it contains a nimble file as well.
10:31:35	FromGitter	<Varriount> nc-x: Yes, however that fails to compiler due to something with the opengl package.
10:32:36	FromGitter	<Varriount> nc-x: `nimble build` for that directory gives me `Error: Nothing to build. Did you specify a module to build using the `bin` key in your .nimble file?`.
10:33:04	FromGitter	<Varriount> And compiling manually gives me `opengl.nim(391, 28) Error: undeclared identifier: 'stmt'`
10:33:28	Yardanico	stmt/expr were removed from devel
10:33:44	FromGitter	<Varriount> :\
10:34:03	Yardanico	because they were deprecated a long time ago
10:36:28	nc-x	I get `opengl.nim(229, 23) Error: ordinal type expected`
10:37:34	FromGitter	<Varriount> nc-x: https://github.com/nim-lang/opengl/issues/56
10:41:31	nc-x	Anyways `nimble` does download nuklear src code as well. So it should technically work.
10:42:03	*	nc-x quit (Quit: Page closed)
10:50:29	FromGitter	<Varriount> Good night.
10:50:52	FromGitter	<kayabaNerve> @Varriount You going to bed or saying that to someone?
11:11:39	*	Perkol joined #nim
11:13:37	FromGitter	<Lite5h4dow> @Varriount there is an example in the Nuklear Library
11:13:47	FromGitter	<Lite5h4dow> im guessing he went to bed
11:19:17	*	dorelix quit (Ping timeout: 268 seconds)
11:35:58	*	dorelix joined #nim
11:37:23	FromGitter	<Vindaar> is it by design or a bug that I can't catch an `IOError` when writing a file in nimscript? Raising it manually works just fine
11:45:08	*	krux02 joined #nim
11:51:31	*	stefanos82 joined #nim
12:09:27	dom96	Hello everyone
12:10:29	stefanos82	greetings dom96
12:28:56	*	csgator joined #nim
12:30:16	*	csgator quit (Client Quit)
12:34:01	FromGitter	<kayabaNerve> So I have a Nimble library about some C files. All the files are in installFiles. I use importc for the functions. I don't do any header/passL/passC/compile pragmas. It compiles. It links. The .out crashes missing the symbol from my C file. The nimcache has file.c.o which is not any text file.
12:34:08	FromGitter	<kayabaNerve> Ideas on how to fix this?
12:36:18	FromGitter	<kayabaNerve> I used to have a compile/header pragma but I couldn't get the paths right in the Nimble package since it was handling it relative to the nimcache directory which isn't... usable AFAIk
12:41:22	FromGitter	<tim-st> I had this too, for me it turnt out I needed to compile an addtional c file
12:46:10	FromGitter	<tim-st> for me relative works with compile pragma perfect like `{.compile: "../a/b/file.c".}`
13:01:23	FromGitter	<kayabaNerve> Thanks @tim-st
13:11:21	*	Perkol quit (Remote host closed the connection)
13:14:17	*	find0x90 joined #nim
13:18:25	*	find0x90 quit (Ping timeout: 248 seconds)
13:22:39	stefanos82	is it possible to use Nim procs inside pragmas? For instance, like the aforementioned {.compile: ... .}, can we use something like {.compile: getCurrentDir() & "a/b/file.c".} ?
13:26:14	*	find0x90 joined #nim
13:33:27	FromGitter	<tim-st> stefanos82: yes, just safe it inside a const string before and make `{.compile: s & "a/b/file.c".}`
13:34:20	FromGitter	<tim-st> or if you need `var` then the var needs to be compileTime var
13:34:37	FromGitter	<tim-st> or static block
13:35:36	dom96	wow, just implemented HTTP pipelining in httpbeast
13:35:59	dom96	The more threads I give to `wrk` the more req/s
13:36:07	Yardanico	dom96, yay \o/
13:36:15	dom96	418k req/s with 32 threads and 64 connections
13:36:16	FromGitter	<tim-st> this means number one?
13:36:32	dom96	Doubt it :P
13:36:46	FromGitter	<tim-st> does mofuw has this too?
13:36:47	dom96	I don't have a lot of confidence in my skills, but we'll see
13:37:05	Yardanico	mofuw has multithreading too
13:37:14	FromGitter	<tim-st> ok
13:37:31	dom96	mofuw gets 100k req/s for the same settings on my machine
13:37:34	Yardanico	dom96, wow
13:37:44	FromGitter	<tim-st> interesting :O
13:37:48	stefanos82	dom96: beware with wrk's results though; you need to figure out its formula to produce your server's peak performance
13:38:22	FromGitter	<tim-st> yes, it doesnt behave like real requests
13:38:25	dom96	Yeah, I think there is a bell curve with the amount of threads I throw at it
13:39:41	dom96	Max I can get is 490k
13:39:58	dom96	If I ask for more threads I start getting: unable to create thread 64: Too many open files
13:40:16	Yardanico	change ulimit :P
13:40:25	dom96	oh, 128 connections + 72 threads works
13:40:29	dom96	and got 600k req/s
13:40:30	dom96	lol wtf
13:40:50	FromGitter	<tim-st> but it needs to be stable
13:41:03	dom96	It's pretty stable so far
13:41:16	FromGitter	<bevo009> #dom96 https://phoenixframework.org/blog/the-road-to-2-million-websocket-connections
13:41:18	FromGitter	<tim-st> I think it depends much on how much files a system can open, I read rocksdb gets very unstable when too many files open
13:41:30	FromGitter	<tim-st> I think, lmdb doesnt have this open files problem
13:41:45	FromGitter	<tim-st> the more open files the easier the os gets unstable quickly
13:41:55	dom96	I'm planning on benchmarking frameworks on my RPI
13:42:02	dom96	I don't think anyone has done a comparison like that
13:42:16	FromGitter	<tim-st> which rpi?
13:42:19	FromGitter	<tim-st> 3?
13:42:27	dom96	So yeah, I could use some tips on how to get rid of the limits
13:42:29	dom96	yeah
13:42:35	dom96	Not the latest model
13:42:46	dom96	I could buy the latest model but I already have like 3 RPis lol
13:43:05	Yardanico	dom96, https://superuser.com/questions/433746/is-there-a-fix-for-the-too-many-open-files-in-system-error-on-os-x-10-7-1 for macos
13:43:06	stefanos82	dom96: there you go mate https://github.com/wg/wrk/issues/227
13:43:08	krux02	i used the rpi 3 today as well
13:43:11	Yardanico	AFAIK it works on macos sierra too
13:43:22	FromGitter	<tim-st> I only have rpi2, I think with my application in release mode a rpi3 can work as a server
13:43:22	krux02	(to play Tales of Phantasio on the PlayStation emulator)
13:43:32	FromGitter	<tim-st> hahaha
13:43:40	dom96	Yardanico: I went down this road already and it's annoying to set this on macOS
13:43:55	dom96	so meh
13:43:59	Yardanico	well, it's pretty straightforward :)
13:44:03	Yardanico	just two lines in a file
13:44:19	dom96	stefanos82: thanks
13:44:26	dom96	Yardanico: It's not
13:44:36	dom96	The later versions of mac make it more complex
13:44:48	Yardanico	dom96, IDK, this suggestion from SO worked for me on macOS Sierra
13:44:55	Yardanico	*high sierra
13:45:19	FromGitter	<tim-st> btw is it possible to limit the ram usage when many threads are enabled?
13:45:20	*	find0x90 quit (Quit: find0x90)
13:45:27	FromGitter	<tim-st> like set maximum to 75% ram?
13:45:28	dom96	Maybe, but it's pointless to benchmark on my mac anyway
13:45:41	dom96	I'd rather do it on my RPi
13:45:48	dom96	I even bought a nice SD card
13:48:05	FromGitter	<tim-st> krux02: maybe you know it: can I force my process to only use a max amount of ram?
13:48:17	krux02	yes you can
13:48:22	FromGitter	<tim-st> in nim?
13:48:37	krux02	I don't know if "in nim"
13:48:57	FromGitter	<tim-st> and what happens if it reaches a blocking state, because it needs more data in ram?
13:49:03	FromGitter	<tim-st> or waiting
13:49:18	krux02	In nim I would say, just allocate the maximum amount of ram you want to use and then don't allocate anything else anymor
13:49:36	FromGitter	<tim-st> yes, but if I do this, this memory would be shared
13:49:44	FromGitter	<tim-st> and I wanted to have multiple threads in one process
13:50:13	krux02	tim-st: I am not 100% sure what happens when a process reaches it's maximum allowed ram, but the usual think that happens it, the program crashes.
13:50:46	FromGitter	<tim-st> ok, thanks, the idea is good, I then have to look if different threads can access the same allocated memory in nim
13:51:12	krux02	when memory is allocated with malloc, but the process isn't allowed to get more memory, then malloc returns a null pointer. But almost no program checks for this possibility that malloc could return 0 and then you get a segmentation fault.
13:51:41	Araq	malloc doesn't return null reliably on your favorite OS.
13:52:03	krux02	Araq: what do you mean?
13:52:07	FromGitter	<Vindaar> asking again, because more people around :): when writing a file with nimscript, is it normal that I cannot catch an `IOError`? My except is just ignored
13:52:07	Araq	because of memory overcommitment.
13:52:15	FromGitter	<tim-st> ok, just wanted to get sure, that when I have a server like dom96 said and it has so many requests that it should never reach more than 80% memory
13:52:42	Araq	and memory overcommitment was invented to hack around fork()s inherent design flaws.
13:53:44	krux02	I have no idea what overcommitment is nor do I know the inherent design flaw of fork()
13:54:41	stefanos82	Araq: did you see Alexandrescu's video that explains how flawed free() is? I honestly didn't know that and gave me goosebumps :S
13:54:56	krux02	I did see it
13:55:09	krux02	but I almost forgot everything
13:55:13	Araq	stefanos82: nope, but let me guess. it doesn't take a 'size' parameter?
13:55:19	stefanos82	lol bingo
13:55:24	stefanos82	C++ fixed it
13:55:29	stefanos82	at least that's what he said
13:55:38	krux02	not really
13:55:43	Araq	yeah, I noticed that long ago too. ;-)
13:55:48	krux02	delete [] doesn't get a size parameter
13:55:57	stefanos82	it returns it for you behind the scenes
13:56:43	stefanos82	blimey, it's so hot that I want to GET TO DA CHOPPAAAA and fly away from here!
13:57:11	krux02	stefanos82, what is returned behind the scenes?
13:57:20	stefanos82	the size of the allocated memory
13:57:40	krux02	well the size of the allocated memory is what you pass to malloc
13:57:51	krux02	it isn't returned
13:57:59	krux02	also not behind the scenes
13:58:19	krux02	it is just stored in a table to manage the chunks of allocated memory blocks
13:59:37	stefanos82	well, they fixed it; https://en.cppreference.com/w/cpp/memory/new/operator_delete
13:59:40	krux02	or wait, do you mean the rounded up size of the allocated memory to block size?
14:01:35	Araq	stefanos82: https://github.com/nim-lang/Nim/blob/devel/lib/core/allocators.nim#L13 ;-)
14:02:14	stefanos82	Araq: no need for extra convincing that Nim is a beauty man, I already know :P
14:06:07	Araq	oh Nim is far from perfection, but I take some pride in the fact that most of its mistakes are all new. ;-)
14:09:34	stefanos82	Nim could gain traction if it had OCaml's quick compilation. I have no idea how they have implemented the language, but it's scary. You just say, for instance ocamlc -o fib fib.ml and boom; you have right in front of you a binary file ready for use
14:10:18	stefanos82	overall Nim is a beauty
14:10:33	*	find0x90 joined #nim
14:10:40	stefanos82	it has a bright future
14:11:02	zacharycarter	Nim does have pretty quick compilation
14:11:12	zacharycarter	But it's going to be based on the compiler speed of whatever backend you're targetting
14:11:58	zacharycarter	Nim can also do a lot of things OCaml can't
14:12:47	zacharycarter	GC for instance
14:13:24	zacharycarter	Ocaml has incremental m&s
14:13:26	zacharycarter	it will stop the world
14:13:53	Araq	why? did they add multi-threading finally?
14:14:15	Araq	it used to be thread local like Nim's iirc.
14:15:09	zacharycarter	I may be incorrect - but I was reading - https://v1.realworldocaml.org/v1/en/html/understanding-the-garbage-collector.html
14:15:20	stefanos82	Araq: I know of this http://ocamllabs.io/doc/multicore.html
14:15:41	Yardanico	from ocaml.org: "The OCaml GC is synchronous. It doesn't run in a separate thread, and it can only get called during an allocation request.
14:15:42	Yardanico	"
14:16:11	Yardanico	ah, that's not the relevant info
14:16:16	FromGitter	<Bennyelg> Why I can't do something like that? ⏎ ⏎ ```proc shouldRun*(o: Job): bool = ⏎ now() > o.nextRun``` ⏎ ⏎ both now() and nextRun are DateTime type [https://gitter.im/nim-lang/Nim?at=5b4a05af26aa91065e93c412]
14:16:31	zacharycarter	I think if completion of the NLVM backend becomes possible / a reality, Nim will outshine most of its competitors
14:16:38	Yardanico	zacharycarter, why so?
14:16:57	Yardanico	as Araq said before, NLVM doesn't have any real advantages over C/C++ backends
14:17:06	zacharycarter	because that opens the door for real WASM suppoert
14:17:07	zacharycarter	support
14:17:12	zacharycarter	w/o emscripten / binaryen
14:17:45	FromGitter	<Yardanico> @Bennyelg you can?
14:17:53	FromDiscord	<awr> NLVM could be interesting in the future if it ever supported MCJIT
14:18:09	FromGitter	<Bennyelg> @Yardanico I can't I convert it now to Time and now its working
14:18:24	FromGitter	<Bennyelg> Datetime is not covered too , I 'll have to re check the manual
14:18:38	FromGitter	<Yardanico> ```code paste, see link``` ⏎ ⏎ works for me [https://gitter.im/nim-lang/Nim?at=5b4a063d73026160f5a08135]
14:19:05	zacharycarter	```
14:19:06	zacharycarter	WebAssembly initially focuses on C/C++, and a new, clean WebAssembly backend is being developed in upstream clang/LLVM, which can then be used by LLVM-based projects like Emscripten and PNaCl.```
14:19:17	zacharycarter	ugh why did it copy / paste like that
14:19:25	zacharycarter	```WebAssembly initially focuses on C/C++, and a new, clean WebAssembly backend is being developed in upstream clang/LLVM, which can then be used by LLVM-based projects like Emscripten and PNaCl.````
14:19:35	Yardanico	zacharycarter, are you sure it will have any advantages over C+emscripten?
14:19:42	zacharycarter	whatever - that's the idea
14:20:21	zacharycarter	advantages? doubtful - but languages like Rust have this and you'll always get criticisms about how Nim doesn't have direct wasm support but Rust does
14:20:23	zacharycarter	blah blah blah
14:20:39	Yardanico	this is not really "direct"
14:20:46	Yardanico	because you still generate LLVM IR ;)
14:20:48	zacharycarter	I mean, having done some Nim -> wasm I know it's rather trivial once you know what you're doing
14:21:01	zacharycarter	haha well
14:21:08	stefanos82	hmmm...should I drink a milkshake or drink a milkshake? -_- there is a third option of drinking a milkshake. Any suggestions?
14:21:11	zacharycarter	you don't have to use something like emscripten I guess is the point
14:21:24	zacharycarter	inject that shit into your bloodstream
14:21:30	zacharycarter	go hard
14:21:56	FromDiscord	<awr> are there any plans to enhance the way inference works for generic procs
14:22:00	stefanos82	it's a homemade man therefore I will push it bloody hard into my veins!
14:22:01	FromDiscord	<awr> https://glot.io/snippets/f2wd7w3lbf
14:22:50	Yardanico	you can do "var n = newSeq[int](5)" so you won't have to specify "int" tiwce
14:22:51	Yardanico	*twice
14:22:54	zacharycarter	Araq: I'm looking at karax and trying to figure out where to attach this callback - would it be a good idea to allow buildHTML to take in a function pointer and when the VNode gets created, when it's first attached as a dom node - invoke it then?
14:23:28	zacharycarter	for now I've done this
14:23:58	FromDiscord	<awr> i'm not talking about combined declaration and initialization
14:24:06	*	zacharycarter sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/kQsoxWpeKpaNlazTcSmHMyxo >
14:24:41	*	zacharycarter sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/veTNwPHKyTBcZwFgnvWErocU >
14:25:17	Araq	awr: there is an RFC for that, accepted and won't require a language change.
14:25:25	zacharycarter	not sure if that makes sense or not though - or if there's a better approach
14:25:31	FromDiscord	<awr> cool!
14:25:44	Araq	zacharycarter: setRenderer() should just take yet another callback
14:26:25	Araq	don't mess the DOM creation events etc. Karax exposes these already but they are a PITA. wait for my new components prototype.
14:26:35	zacharycarter	okay
14:28:22	FromGitter	<Bennyelg> @Yardanico your code not work for me
14:28:26	FromGitter	<Bennyelg> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b4a088973026160f5a085f8]
14:28:40	FromGitter	<Yardanico> oh maybe you're not on devel ;)
14:28:45	FromGitter	<Yardanico> there's a lot of fixes in devel as always
14:29:07	FromGitter	<Yardanico> there's no real reason to use stable Nim versions before 1.0 IMO
14:29:48	FromGitter	<Bennyelg> :/
14:29:54	FromGitter	<Bennyelg> going to switch devel
14:30:01	FromGitter	<Yardanico> just compile devel every week, it's easy with choosenim
14:30:34	FromGitter	<citycide> can you not do imports for `runnableExamples`? would like to show some path joining in there
14:32:10	FromGitter	<Bennyelg> @Yardanico so every week you bin/nim c koch ?
14:32:39	Yardanico	I run it almost every day (when there's any significant fixes to the codegen/language features), or maybe once in 3 days :)
14:32:44	FromGitter	<tim-st> does the following `bin/nim c koch` also work as `bin\nim c koch` on linux/mac?
14:32:54	FromGitter	<Bennyelg> mac
14:32:56	FromGitter	<tim-st> on windows only second works
14:33:14	FromGitter	<tim-st> from here: https://github.com/nim-lang/Nim
14:33:26	zacharycarter	unix filepaths vs windows filepaths
14:33:32	FromGitter	<Bennyelg> yea Iknow
14:33:33	zacharycarter	windows uses backslashes
14:33:33	zacharycarter	unix uses forward slashes
14:33:34	FromGitter	<Bennyelg> almost finished
14:33:49	Yardanico	@tim-st you should use stdlib to join paths :)
14:33:51	FromGitter	<tim-st> I just ask to have a valid tutorial for all os
14:34:07	FromGitter	<tim-st> this is a nightly version tut from nim page
14:34:12	zacharycarter	bin\nim c koch does not work on osx
14:34:17	FromGitter	<tim-st> which doesnt work out of the box for windows
14:34:21	FromDiscord	<awr> that's interesting that choosenim doesn't use git to clone the repo
14:34:24	zacharycarter	bin/nim c koch does
14:34:33	FromDiscord	<awr> https://github.com/dom96/choosenim/blob/master/src/choosenim/download.nim
14:35:04	FromGitter	<tim-st> zacharycarter: ok, thanks
14:35:32	Yardanico	@awr https://github.com/dom96/choosenim/issues/12
14:37:11	FromDiscord	<awr> hah
14:37:56	FromDiscord	<awr> i just have a cloned repo and i pull from devel and manually compile every once in a while
14:38:01	zacharycarter	np
14:38:35	Yardanico	@awr me too
14:39:14	*	find0x90 quit (Quit: find0x90)
14:39:36	FromDiscord	<awr> there's some bodges i have made in the nim stdlib that I should make issues for but i'm lazy
14:39:56	FromGitter	<Bennyelg> works like charm thanks @Yardanico
14:42:18	dom96	https://github.com/dom96/httpbeast/commit/ecc7edecfe285c1359ba732ca8c2b3bbf0575127 :D
14:44:00	Yardanico	dom96, you're using "likely"?!
14:44:52	dom96	Of course
14:44:59	dom96	Optimising to the max
14:48:53	PMunch	What is this likely stuff?
14:49:03	PMunch	A hint to the branch predictor?
14:49:06	zacharycarter	if a branch is more likely to be taken?
14:49:09	zacharycarter	in a conditional?
14:49:09	dom96	yep
14:49:15	PMunch	Cool
14:49:26	zacharycarter	I suggested doing that one time in Nim and someone told me it wasn't worth it
14:49:27	dom96	No idea how much of a difference it actually makes
14:49:39	zacharycarter	but I don't remember who suggested that / how to actually benchmark it
14:49:50	Araq	probably me... :-)
14:49:52	dom96	It'd be interesting to check how the asm differs
14:49:59	zacharycarter	haha maybe
14:50:01	Araq	I don't believe in 'likely' ;-)
14:50:05	*	donlzx quit (Ping timeout: 276 seconds)
14:50:28	PMunch	Hmm, interesting concept: http://www.toccata.io/2017/10/No-Booleans.html
14:50:34	PMunch	A language without booleans
14:50:42	FromGitter	<tim-st> no one knows what it does
14:51:10	PMunch	(The linked article is also interesting
14:51:11	PMunch	)
14:51:16	*	PMunch quit (Quit: leaving)
14:51:27	FromDiscord	<awr> likely feels like it should be a pragma if anything
14:51:31	FromGitter	<tim-st> it's likely just a call to c's likely and also in c noone knows
14:51:54	stefanos82	Araq, I wanted to ask you: when someone says about using REPL, do they mean they have to depend on a bytecode mechanism or not?
14:52:41	dom96	https://stackoverflow.com/questions/109710/how-do-the-likely-and-unlikely-macros-in-the-linux-kernel-work-and-what-is-t
14:53:58	FromGitter	<tim-st> ok, that's good, but still noone calculated the probability number at which it makes definetly sense to use it
14:54:17	FromGitter	<tim-st> independent on the cpu
14:54:47	zacharycarter	https://github.com/hnes/libaco
14:56:17	FromGitter	<Lite5h4dow> @zacharycarter did you see my previous posts
14:56:32	zacharycarter	I didn't
14:56:35	zacharycarter	I havne't been as active lately
14:56:37	stefanos82	dom96: I guess your article should be combined with this one too https://stackoverflow.com/questions/11227809/why-is-it-faster-to-process-a-sorted-array-than-an-unsorted-array
14:56:38	zacharycarter	what were they about?
14:56:51	FromGitter	<Lite5h4dow> i cant get you bindings for nuklear to work
14:56:58	zacharycarter	oh
14:57:05	zacharycarter	what issue are you having?
14:57:24	zacharycarter	I haven't compiled them in a while - I need to set up CI jobs for all my bindings I plan on actually using / think others might use
14:57:30	FromGitter	<Lite5h4dow> after i installed it the compiler cant see any of the functions of vulcan
14:57:46	FromGitter	<Lite5h4dow> but nuklear is a valid package
14:57:56	zacharycarter	well - I don't have a vulkan implementation for a nuklear backend
14:57:58	zacharycarter	that's something you'd have to write
14:58:11	FromGitter	<Lite5h4dow> so ¯\_(ツ)_/¯
14:58:14	zacharycarter	nuklear is graphics API agnostic - so if you want to use it with Vulkan - you need to provide all the drawing code
14:58:15	FromGitter	<Lite5h4dow> nuklear sorry
14:58:21	zacharycarter	oh
14:58:33	zacharycarter	let me try to compile / run them - give me a few
14:58:38	FromGitter	<Lite5h4dow> tyty
14:58:42	FromDiscord	<awr> if you're going to do anything with vulkan i noticed a few bits of the vulkan nim package are broken
14:58:55	FromGitter	<Lite5h4dow> i misstyped
14:58:57	FromGitter	<Lite5h4dow> i meant nuklear
14:59:05	zacharycarter	I suggest anyone wanting to do graphics work at this point use BGFX
14:59:35	zacharycarter	if opengl deprecation hasn't proved something like BGFX is worth it - I don't know what will
15:00:17	FromGitter	<Lite5h4dow> damn ok
15:00:20	FromGitter	<Lite5h4dow> ill give that a go
15:00:26	FromGitter	<Lite5h4dow> i hadnt heard of BGFX
15:00:37	FromDiscord	<awr> there's also the vulkan portability subset thing
15:00:43	zacharycarter	there are bindings out there for it already - they're probably out of date, but I'll be updating mine once I start on my next graphics engine project
15:00:50	FromDiscord	<awr> https://www.khronos.org/vulkan/portability-initiative
15:07:04	FromGitter	<krux02> zacharycarte: OpenGL isn't deprecated.
15:07:20	zacharycarter	it is on osx
15:07:30	FromGitter	<krux02> There is a difference between deprecation and abandonment. Deprecation can only be done on something that you own.
15:07:40	zacharycarter	well whatever
15:07:44	FromGitter	<krux02> Apple isn't entitled to deprecate OpenGL
15:07:55	zacharycarter	I agree with you - but that's how they phrased it
15:08:32	FromGitter	<krux02> That is true. But that doesn't make it right.
15:08:34	zacharycarter	I guess they're deprecating their support for opengl
15:08:36	zacharycarter	is the better way to word it
15:09:20	zacharycarter	Lite5h4dow: the nuklear nim bindings are compiling / running fine for me on OSX
15:09:44	FromDiscord	<awr> apple "abandoned" openGL long before they made that annoucement imo
15:09:45	FromDiscord	<awr> https://support.apple.com/en-us/ht202823
15:09:48	FromGitter	<krux02> well support can not be deprecated
15:09:49	zacharycarter	make sure you run nimble install in the examples/nuklear-nim-examples
15:09:53	zacharycarter	and run the examples from that directory
15:10:03	FromDiscord	<awr> they're all 4.1
15:10:32	zacharycarter	also do a nimble install in the root of the project too
15:10:40	FromGitter	<krux02> yes, it is locked to version 4.1, because on never versions you could get better performance, but then metal wouldn't be that much of an improvement anymore.
15:11:03	FromGitter	<krux02> OpenGL is locked to version 4.1 to make Metal "shine"
15:29:05	miran	can a dictionary/table contain functions as values?
15:29:38	Yardanico	afaik yes
15:31:36	zacharycarter	function pointers sure
15:31:45	zacharycarter	procs - I don't think so
15:31:48	zacharycarter	but I could be wrong
15:32:06	miran	i'm trying to do `Table[string, (int, int) -> bool]`
15:32:25	Araq	miran: that works but might cause trouble in a compiletime context
15:32:52	miran	so it looks something like `{"less": `<`, "le": `<=`}`
15:33:08	Araq	`<` are not real procs
15:33:34	stefanos82	Araq: what inspired you to design the {. .} for pragmas? I'm curious.
15:34:03	miran	Araq: ok, then i won't force that style. i already have the alternative solution....
15:34:51	Araq	it was clear to me the 3 different bracket kinds () [] {} wouldn't do and so I came up with (. .) [. .] {. .}
15:35:37	Araq	for a systematic way of how to extend the brackets
15:36:04	Araq	but only {. .} are used for anything
15:37:02	Araq	and devel added [: ] because [. .] is not sexy enough :P
15:38:57	Araq	and (. .) can't be used because somebody said these look like boobs
15:39:09	Yardanico	lol
15:39:31	Yardanico	( . ) khm
15:39:38	Yardanico	!vanish
15:40:20	miran	what is the probability that we'll see pattern matching in nim?
15:41:27	stefanos82	Araq: I don't know whether it was coincidence or not, but Ocaml's flags are the same lol
15:42:08	stefanos82	sorry, ocamlbuild's tags
15:44:39	stefanos82	Araq: now I realized what they meant with boobs; I didn't see it at first ^_^
15:52:22	dom96	https://github.com/TechEmpower/FrameworkBenchmarks/pull/3933
15:52:33	dom96	I expect Jester to be rather slow
15:52:40	dom96	I have some ideas on how to fix that though
15:53:23	FromGitter	<Bennyelg> whats wrong? ⏎ ⏎ ```let t1 = Job(nextRun: now(), lastRun: now()).every(10).seconds().friday()``` ⏎ ⏎ Traceback (most recent call last) ... [https://gitter.im/nim-lang/Nim?at=5b4a1c731c0f906b1452a026]
15:53:39	FromDiscord	<awr> nim needs < > everywhere just like C++ /s
15:57:35	*	Trustable quit (Remote host closed the connection)
15:59:12	FromGitter	<Bennyelg> https://github.com/2vg/mofuw wonderful
15:59:21	Yardanico	@Bennyelg look at httpbeast too!
16:00:15	dom96	:D
16:00:50	FromGitter	<Bennyelg> Heheh Digging now, Looks so promising Nim is improving like hell and all thanks to you guys.
16:02:57	FromGitter	<Bennyelg> @dom96 any bentchmark ready ? for the httpbeast ?
16:03:38	dom96	https://irclogs.nim-lang.org/14-07-2018.html#13:35:36
16:03:48	*	csgator joined #nim
16:04:19	Araq	miran: pattern matching is not scheduled for v1, but there are nimble packages that implement it
16:05:30	miran	Araq: oh, yeah, i forgot about patty! haven't tried it yet, i might give it a chance
16:05:52	FromGitter	<Bennyelg> @dom96 freaking crazy
16:06:08	FromGitter	<Bennyelg> lol
16:17:59	*	donlzx joined #nim
16:28:31	FromGitter	<Bennyelg> how can I call any proc from a function? ⏎ e.g ⏎ proc a(func: proc()) = func
16:28:52	FromGitter	<Bennyelg> I want to wrap a procedure which receive procedure and run it
16:29:12	FromGitter	<Bennyelg> I tried something like ⏎ ⏎ ```proc exec*(o: var Job, fun: untyped) = ⏎ fun``` [https://gitter.im/nim-lang/Nim?at=5b4a24d866c1e833a9e84b68]
16:29:24	FromGitter	<Bennyelg> but I dont have actually any idea
16:32:19	dom96	func()
16:32:52	Yardanico	don't use "func" btw, it's a keyword
16:33:06	FromGitter	<Bennyelg> but if my func is need parameters ?
16:33:17	FromGitter	<Bennyelg> ok
16:33:24	dom96	oh yeah, good point
16:33:40	FromGitter	<Bennyelg> and func is a keyword for?
16:33:44	FromGitter	<Bennyelg> to be replace proc someday ?
16:33:45	Yardanico	for "func" :)
16:33:52	Yardanico	@Bennyelg - just call it like func(arg1, arg2)
16:33:56	FromGitter	<Bennyelg> but proc is func ? :O
16:35:16	FromGitter	<Bennyelg> i want to run something like this: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b4a2643b5efcd5aeefccd9b]
16:36:54	*	donlzx quit (Quit: Leaving)
16:37:10	FromGitter	<Bennyelg> another way which will be even better is : ⏎ ⏎ ```exec(helloworld, args1, arg2, arg3)``` [https://gitter.im/nim-lang/Nim?at=5b4a26b5b5efcd5aeefcce78]
16:41:47	FromGitter	<Vindaar> You can do it with templates like this for example (if I understand you correctly :) ): ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b4a27ca897d514b3c4e731d]
16:42:35	FromGitter	<Bennyelg> looks good, what was the magic here? just pass fun as an untyped instead of func ?
16:42:38	FromGitter	<Bennyelg> proc *
16:43:19	miran	@Bennyelg you can do it without templates, just by using procs
16:43:24	FromGitter	<Vindaar> it's template magic ;) replaces it at compile time. If you wanted to do it with a proc, you'd need to give the correct signature of the procs you need to wrap (although generics might help here too?)
16:43:49	miran	@Bennyelg @kaushalmodi has written about it couple of days ago
16:44:27	FromGitter	<Bennyelg> works like charm
16:44:42	*	csgator quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
16:44:48	FromGitter	<Bennyelg> you were right procs also work I forgot to remove discard word in the begining of my request
16:45:18	miran	here it is: https://scripter.co/notes/nim/#lambda
16:46:50	FromGitter	<Bennyelg> every day I am amazed by how nim is so elastic.
16:47:05	FromGitter	<Vindaar> it never stops. I still get the same feeling :D
16:47:22	FromGitter	<Bennyelg> :D
16:47:38	FromGitter	<Vindaar> (ok, I only started using Nim ~Oct 2017)
16:48:05	FromGitter	<Bennyelg> around my time but I am doing some fun with it My guys at work don't want to switch they afraid.
16:48:17	FromGitter	<Bennyelg> we are using go and I want to leave. looking now for another job
16:48:18	FromGitter	<Bennyelg> lol
16:48:25	*	skrylar is attempting to make some of that switching pain lessened
16:48:35	FromGitter	<Bennyelg> but No jobs on nim currently :(
16:48:42	skrylar	we at least have (albeit an ugly) GUI for instance ;p
16:49:07	FromGitter	<Vindaar> I can imagine. I'm only at Uni, but my colleagues aren't too keen on using my code so far either (let alone using Nim themselves..). But I'll get them to eventually :P
16:49:28	skrylar	actually i know fltk is ugly and doesn't port to mac well, but its functionality is really solid :( never had to deal with bugs and screwy event dispatch likei gtk
16:50:00	*	skrylar has had gtk flat out refuse to move objects in a box layout, despite using the API calls to do so
16:50:51	FromGitter	<alehander42> is `func` working now :O
16:51:04	miran	@Vindar i was sure you had more nim experience!!
16:51:25	miran	@alehander42: yeah, i'm using `func` regularly now (in v0.18.1)
16:52:16	FromGitter	<Vindaar> @miran: well, since then I've been coding in Nim almost daily, since I switched (almost) completely, haha. But thanks :)
16:52:55	miran	i should step up my game
16:52:59	FromGitter	<Vindaar> :D
16:58:30	FromGitter	<kaushalmodi> miran: Cool, now I need to play with func vs proc :)
16:58:41	skrylar	question: do we have a preferred way of running background tasks in nim (a.k.a. setTimeout)
16:59:14	*	dddddd quit (Remote host closed the connection)
16:59:47	miran	@kaushalmodi: it's easy, just change every proc to func, and then try to see why it doesn't work at some places :)
17:00:22	miran	that helped me to refactor some of my code to make it more elegant
17:00:23	*	fvs joined #nim
17:01:27	*	endragor joined #nim
17:07:56	*	Ven`` joined #nim
17:08:08	FromGitter	<Bennyelg> @kaushalmodi what is the different
17:08:17	FromGitter	<Bennyelg> between func and proc :d
17:08:44	FromDiscord	<exelotl> I would also like to know
17:10:31	Yardanico	func is {.noSideEffect.}
17:12:08	FromDiscord	<exelotl> ah I see, so that could be useful for optimisation reasons or just for clarity/intent?
17:13:15	FromGitter	<kaushalmodi> @Bennyelg I don't know. That's why I'll need to play with those.
17:13:59	miran	just clarity
17:14:07	Yardanico	as I said, the only difference now is just that func is {.noSideEffect.}
17:14:11	Yardanico	like real "function" in maths
17:14:34	miran	but it is nice to be "just" clearer :)
17:17:53	FromGitter	<kaushalmodi> miran: Oh, that means no echoes in there.
17:18:14	skrylar	debugEcho exists
17:18:21	*	endragor quit (Remote host closed the connection)
17:18:26	skrylar	it does an echo while pretending not to have side effects, for that very reason
17:18:30	miran	and no globals (except for `const`)
17:18:40	FromGitter	<kaushalmodi> TIL. Thanks
17:19:31	skrylar	dom96, Araq is there already a timer system for async stuff
17:19:56	*	Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
17:21:45	Araq	skrylar: yeah, there is, but I never used it :-) see eg. asyncSleep()
17:23:11	skrylar	https://nim-lang.org/docs/asyncdispatch.html#sleepAsync, hm
17:23:26	skrylar	these might work, will have to see
17:23:55	skrylar	bacon.js has this thing where you can make event streams either from timeouts or that take the value of a property on a timeout; looking in to what nonsense is required to support that nimside
17:28:29	*	develooper joined #nim
17:28:48	develooper	hello nim
17:29:37	*	develooper quit (Client Quit)
17:30:12	Yardanico	lol
17:35:01	FromGitter	<Vindaar> procrastination is strong in me again. problem: texlive is missing packages. solution: write nim program to compile tex file over and over again, parse output for missing package error and call `tlmgr` to install appropriate package. It worked, so I guess... yay?
17:42:43	*	Perkol joined #nim
17:44:25	miran	@Vindaar don't let the latex-haters to hear that!!
17:44:45	miran	btw, what packages did miss?
17:46:30	FromDiscord	<exelotl> lol that's a brilliant solution
17:47:03	Araq	Vindaar: :-)
17:47:11	Araq	miran: too late, I heard it.
17:47:31	miran	:)
17:47:46	FromGitter	<Vindaar> @miran: haha :D it was missing random stuff like mchem, siunitx etc. Freshly installed void linux, installed texlive 2018 from the repos. That only installs ~280 packages
17:48:16	FromGitter	<Vindaar> I mean I could have simply checked the header of the tex file and just parsed those or something :D this was a somewhat *inefficient* solution :D
17:48:20	miran	cant you install also all those texlive-xyz packages?
17:49:12	FromGitter	<Vindaar> In void linux there's only a minimum texlive package. I could have installed texlive-full manually of course
17:49:38	miran	but you're right - it is weekend - we need to make use of that extra free time :D
17:50:55	FromGitter	<Vindaar> It's not like next week is gonna be crazy again, haha (gonna reinstall our detector at our experiment at CERN starting Monday + I need to give a talk on Thursday. ughh)
17:52:17	miran	i just checked - in manjaro linux, you have huge amount of texlive packages, i thought the situation is the same in other distributions
17:55:33	FromGitter	<Vindaar> I think you're right for many distributions (although e.g. on Ubuntu these packages tend to be so old, that I always installed texlive manually anyways)
17:55:53	FromGitter	<Vindaar> but tbh `tlmgr` isn't such a bad tool actually :)
17:58:37	miran	yeah, old packages (and then using PPAs) is one of the reasons why i decided to leave linux mint (it is based on ubuntu LTS), and try my luck with manjaro
18:00:35	*	Trustable joined #nim
18:01:11	FromGitter	<Vindaar> hehe, same here. my desktop is still running mint 17 (too scared to do a `dist-upgrade` :P it works, so I'm happy...)
18:10:58	stefanos82	@Vindaar: under Windows at work, I used to use TeXnicCenter which had the option to automatically fetch missing packages for me
18:13:10	*	find0x90 joined #nim
18:17:20	FromGitter	<Vindaar> stefanos82: that's handy!
18:20:42	skrylar	i hope heaps of closures don't cause issues xD
18:21:25	miran	stefanos82: MikTeX has the same ability
18:22:58	*	find0x90 quit (Quit: find0x90)
18:24:05	*	endragor joined #nim
18:24:12	*	endragor quit (Remote host closed the connection)
18:28:56	stefanos82	miran: it's an IDE with MikTeX behind the scenes
18:29:04	miran	oh!
18:29:34	FromGitter	<Vindaar> https://gist.github.com/Vindaar/9ed80ee38faff8b6f3d3ce174d729f70 :D
18:29:39	stefanos82	I don't remember whether TeXworks offers this option or not
18:40:32	*	find0x90 joined #nim
18:46:14	*	Perkol quit (Remote host closed the connection)
19:18:36	skrylar	nothing stupid happens if i delete a seq entry during an iterator right
19:21:17	FromGitter	<Vindaar> you mean an element of the seq you iterate over?
19:24:01	skrylar	yeah it doesn't work
19:27:41	skrylar	its easy enough to replace because only the 'current' element can get deleted
19:27:59	skrylar	implementing something like bacon.js (functional reactive programming / FRP) to nim because eeeh
19:30:37	skrylar	i remember a few years ago reading somebody's demo that was suggesting FRP is really useful for event chains
19:30:48	skrylar	Rust has adopted the primitives for iterators
19:33:55	Yardanico	I should use "cast[ptr int](addr(arg))[]" to get address of `arg` as int?
19:34:12	Yardanico	not to interface with C, just to get address
19:35:35	krux02	Yardanico, you want the addr as int?
19:35:38	Yardanico	yes
19:35:53	krux02	then you should do ``cast[uint](arg.unsafeAddr)``
19:35:55	skrylar	addr(someint) will already be `ptr someint`
19:36:06	Yardanico	krux02, ok, thanks
19:36:30	krux02	you know when unsafeAddr is ok to use?
19:36:48	skrylar	~~never~~ during moments of extreme magic :b
19:36:57	Yardanico	krux02, I won't modify any memory by that address
19:37:03	Yardanico	I just need to get addr :P
19:37:06	krux02	yes
19:37:32	krux02	skrylar, unsafeAddr isn't any different than addr, it just compiler more often
19:37:54	skrylar	addr does a sanity test to see if you can write to the addr, unsafeaddr doesn't
19:38:07	skrylar	although either keyword should probably be in your "block and force an audit" list in phabricator ;)
19:38:14	krux02	and when you have to use it it either means you did not get your types correct, or you just write wrapper code
19:38:49	krux02	well I like to use addr, but I am very happy that in Nim I just don't need to do it
19:39:21	krux02	passing parameters by reference implicityly really saves me using the `addr` operator.
19:39:37	krux02	it is a simple and yet so powerful feature of the Nim programming language.
19:39:48	skrylar	i'm not anti addr (or even goto) ;p
19:39:56	skrylar	it's just a warning sign is all
19:40:10	krux02	ok, then we don't need a discussion :P
19:40:18	Araq	"force an audit" nails it.
19:41:09	skrylar	Araq, phabricator actually has (if you make everyone use arc) the ability to auto-reject pushes based on conditions and alerts some maintainer to go/no-go it. there is a version that does it without arc but the commits still go through w/ a warning
19:41:12	skrylar	its neat
19:41:20	Araq	'cast', 'ptr' and 'addr' are Nim's unsafe features. and everything else that is currently unsafe must be made safe :-)
19:42:50	skrylar	the sheer amount of closures being made by this module is giving my anneurisms :X
19:43:15	Araq	ugh, ticks in my leg
19:56:07	FromDiscord	<exelotl> ow that's not good, be safe
20:03:40	Araq	lol thanks, got a pair of tweezers
20:15:53	skrylar	first successful event piped through a filter :^)
20:16:31	skrylar	araq.filter(proc(x:thing): bool = return x.kind != tkTick)
20:16:47	skrylar	well, `keep`, but meh
20:26:00	FromDiscord	<exelotl> item_t* items = get_items(); // ptr to first item
20:26:04	FromDiscord	<exelotl> item_t* foo = &items[5]; // ptr to 6th item
20:26:10	FromDiscord	<exelotl> what's the equivalent in nim?
20:26:21	FromDiscord	<exelotl> let items:ItemPtr = getItems() # ptr to first item
20:26:27	FromDiscord	<exelotl> let foo = ????
20:27:44	FromDiscord	<exelotl> (where ItemPtr and getItems are imported from C)
20:35:11	skrylar	https://github.com/Skrylar/cobweb/blob/master/cobweb.nim :D
20:39:06	FromGitter	<tim-st> skrylar: regarding your iterator question I found this in `re` docs: `Note that since this is an iterator you should not modify the string you are iterating over: bad things could happen.`
20:39:44	skrylar	tim-st: if the length of a seq changes length, there is an assertion error (debug mode?)
20:39:57	skrylar	changed it for a while loop with conditional increment
20:40:05	FromGitter	<tim-st> I dont know, sry
20:40:54	skrylar	wasn't a question. wrote a test :)
20:41:12	FromGitter	<tim-st> btw it seems `re` is much faster than `regex` unfortunately :\
20:44:40	FromGitter	<tim-st> is it normal that this regex matches from <P to </P> ? `(?s)<P .*?#0007'>(.*?)</P>`
20:44:59	FromGitter	<tim-st> I assumed only in brackets matches
20:46:51	krux02	tim-st: I am not sure about nim regular expressions, but if you want to match the actual character <, it might be that you have to escape it.
20:47:02	*	miran quit (Ping timeout: 244 seconds)
20:47:05	krux02	but it depends on the regular expression library/mode
20:47:27	FromGitter	<tim-st> no I want to match (.*?) but it matches already at <P
20:47:35	FromGitter	<tim-st> I tried re and regex...
20:48:09	krux02	originally regular expressions have some characters that you had to escape and some you did not. that was a bit confusing, so it got changed, kind of. Now there are more regular expression modes around.
20:48:19	*	miran joined #nim
20:48:26	krux02	. is any character
20:48:47	krux02	\w is [0-9a-zA-Z_]
20:48:47	FromGitter	<tim-st> yes, I use it like in python and it always worked until now
20:49:09	krux02	well the python regular expressions are quite ok
20:49:42	FromGitter	<Vindaar> @tim-st can you clarify an example string you want to match and your regex?
20:50:08	FromGitter	<tim-st> it's a rtf converted to htm by word or something
20:50:11	FromGitter	<Vindaar> in general though, I find www.regex101.com super helpful
20:50:26	FromGitter	<tim-st> yes, thanks, I try it there
20:50:33	krux02	I recommend the emacs commend re-builder that lets you interactively build a regular expression and see what it matching within an entire file.
20:50:33	FromGitter	<tim-st> also used this before
20:50:42	krux02	command not commend
20:50:45	FromGitter	<Vindaar> if he uses emacs I second that :)
20:51:00	FromGitter	<tim-st> I dont touch emacs and vim and things
20:51:11	FromGitter	<rayman22201> https://regexr.com/
20:51:21	FromGitter	<rayman22201> super nice interface for building regex
20:51:35	FromGitter	<rayman22201> If you like huge JS apps in your browser that is :-P
20:52:31	FromGitter	<tim-st> krux02: there was indeed an unescaped character, funny that it matched though quite fitting
20:52:32	krux02	I wonder whis this this is so huge, regex is built into javascript and then there is text.
20:52:40	FromGitter	<data-man> @tim-st: There is nothing unexpected, PCRE has a long history of optimizations (decades!)
20:53:16	FromGitter	<tim-st> yes, I know, I also didnt expect it, just hoped gcc makes some magic and it's like 2 times slower only
20:53:22	FromGitter	<tim-st> but it seems like ten times
20:53:28	FromGitter	<tim-st> (guessed)
20:53:41	krux02	gcc doesn't do magic
20:53:48	krux02	don't expect the compiler to do magic
20:53:51	krux02	compilers can't do that
20:53:58	*	find0x90 quit (Quit: find0x90)
20:54:10	krux02	and the programmer is resposible to get the program run fast, not the compiler programmer.
20:54:12	skrylar	except stalin, who brutally optimizes
20:54:30	FromGitter	<tim-st> ok
20:56:58	krux02	rayman22201: somehow RegExr seems to be weird. Neither \< nor < match the beginning of a word
20:57:12	krux02	both just match the < character
20:58:08	FromGitter	<rayman22201> I don't think that is supported by JS regex
20:59:34	FromGitter	<rayman22201> I've actually never seen the `<` character used as a regex anchor. Is that a PCRE extension or something?
21:00:08	FromGitter	<Vindaar> skrylar: github.com/skrylar/*cobweb* -> stalin -> https://github.com/barak/stalin -> link in description: http://*cobweb*.ecn.purdue.edu/~qobi/so… -> full circle. I see what you did there :P (or maybe just a fun coincidence :D)
21:02:45	krux02	rayman22201: I don't know where the < character got introduced, but it is one of the most important charactors that I use in regular expressions
21:02:56	krux02	searching for word boundaries is very handy
21:03:15	FromGitter	<rayman22201> I can't even find it on google. there is the `\b' for word boundry?
21:03:42	krux02	just do ``man grep``
21:03:46	krux02	on the terminal
21:03:48	FromGitter	<tim-st> < is for matching groups I think, and ^ and $ are anchors
21:03:55	krux02	and everything is explaind to you
21:04:16	krux02	tim-st: no () is for matching groups
21:04:18	skrylar	Vindaar: well, cobweb was originally a dataflow thing (think Excel equations, or lisp 'cells')
21:04:28	krux02	but depending on the engine it can be \( \) or just ( )
21:04:30	FromGitter	<tim-st> krux02: named groups
21:04:37	krux02	no not names group
21:04:46	FromGitter	<Vindaar> skrylar: ahh, I had no idea :)
21:04:55	skrylar	Vindaar: i rewrote it so its now baconjs in native nim, but the whole point is to replace spaghetti code or GUI events and whatnot with a cobweb :p
21:05:07	FromGitter	<rayman22201> @krux02 That is interesting. Super non-standard
21:05:09	FromGitter	<tim-st> https://stackoverflow.com/questions/10059673/named-regular-expression-group-pgroup-nameregexp-what-does-p-stand-for
21:05:52	krux02	Just to make sure we are talking about the same thing here, when I talk about regular expressions I mean the ones that you can use in vim, sed, grep, pearl and all other unix tools, I am never talking about javascript.
21:06:12	FromGitter	<tim-st> this is pcre
21:07:11	FromGitter	<rayman22201> It sure as hell isn't in the PCRE man page: https://pcre.org/pcre.txt
21:07:52	FromGitter	<rayman22201> I believe you. It's in the grep man page. It's just something I haven't seen before, and I have done a lot of PCRE stuff
21:07:53	FromGitter	<tim-st> it is:
21:07:54	FromGitter	<tim-st> There are several different ways of writing back references to named ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b4a662ad1428e5aefe8a38d]
21:08:45	krux02	rayman22201: I am not 100% sure what you mean by non-standard, it works across vim emacs sed and grep. I tested them all.
21:08:59	FromGitter	<rayman22201> I consider PCRE the standard
21:11:07	FromGitter	<rayman22201> I'm not dogmatic about it though. I'm just surprised. It's super weird to me that all of these standard unix utilities have it but PCRE does not.
21:12:23	FromGitter	<rayman22201> It's also surprising to me that I can't find any history of it on google. (But I could just be googling the wrong things)
21:12:55	krux02	I try to find the specification of the pcre
21:13:01	krux02	but that document is just the C api
21:14:35	*	krux02 quit (Read error: Connection reset by peer)
21:15:01	FromGitter	<rayman22201> I did find this in the pcre man page on Ubuntu: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b4a67d46534bc5d2e5ba12b]
21:15:12	*	krux02 joined #nim
21:17:55	*	miran_ joined #nim
21:18:24	FromGitter	<tim-st> the regex doesnt work, although I tested it on the regex tester site
21:18:27	*	miran quit (Ping timeout: 240 seconds)
21:18:37	FromGitter	<tim-st> it always matches the full block
21:18:53	FromGitter	<tim-st> I use re.findAll(...)
21:19:39	FromGitter	<data-man> @krux02: https://www.pcre.org/original/doc/html/pcresyntax.html and https://www.pcre.org/current/doc/html/pcre2syntax.html
21:20:11	*	miran_ is now known as miran
21:21:47	FromGitter	<rayman22201> @tim-st do you have an example string?
21:22:11	krux02	aparently I am in the gnu regular expression land
21:22:17	FromGitter	<tim-st> yes, wait a moment I try it in python first, I trust pythons re
21:22:58	FromGitter	<rayman22201> lmao @krux02 you have been drinking too much Richard Stallman toe tea :-P j/k
21:23:09	FromGitter	<tim-st> `re"(?s)<P .*?#0007'>\s*(.*?)\s*<\/SPAN.*?<\/P>"` is my regex
21:23:26	krux02	https://www.gnu.org/software/grep/manual/html_node/The-Backslash-Character-and-Special-Expressions.html#The-Backslash-Character-and-Special-Expressions
21:23:27	FromGitter	<tim-st> this should only match the thing in brackets in the middle?
21:23:59	FromGitter	<tim-st> but it matches from <P to </P>
21:24:34	*	miran quit (Ping timeout: 260 seconds)
21:24:50	*	nsf quit (Quit: WeeChat 2.1)
21:24:55	krux02	tim-st python is not gpl, so I doubt it has gnu regular expressions
21:25:27	FromGitter	<tim-st> I'm using pcre in nim
21:25:50	FromGitter	<tim-st> `This module is implemented by providing a wrapper around the PRCE (Perl-Compatible Regular Expressions) C library.`
21:26:07	Araq	re.nim ?
21:26:12	FromGitter	<tim-st> yes
21:26:16	FromGitter	<tim-st> import re
21:26:35	Araq	my stuff is unkillable :P
21:27:11	FromGitter	<tim-st> here is the target string: https://gist.github.com/tim-st/be9441b3c1065797a6ce16a7baca6035
21:27:44	Araq	should have used my HTML parser ;-)
21:28:07	FromGitter	<tim-st> this is "htm" some weird ms office format
21:29:33	*	Jesin quit (Quit: Leaving)
21:29:51	FromGitter	<Vindaar> @tim-st it's working fine on my end, as far as I can tell?
21:30:07	FromGitter	<Bennyelg> how to sort list of elements by specific key
21:30:24	FromGitter	<tim-st> @Vindaar what's your output?
21:31:12	FromGitter	<Vindaar> oh, you said you're using `findAll` no?
21:31:33	FromGitter	<tim-st> which one are you using?
21:31:41	krux02	I think regular the use expressions is a sign of hackyness. But the fact that the regular expression library is just stable and dosn't need constant revisions and improvements is really showing how well the software works.
21:32:15	FromGitter	<tim-st> I want the same like pythons re.findall()
21:32:58	krux02	continue where you left off?
21:33:28	FromGitter	<Vindaar> `data` is your thing up there, `reg` your regex from above: https://gist.github.com/Vindaar/b555e2a347d6c1fb8fe3f4199f80d3fa
21:33:44	Araq	krux02: which library do you mean?
21:33:53	FromGitter	<Bennyelg> how to do something like this ? ⏎ ⏎ ```sortedByIt(s.listOfJobs, s.listOfJobs.atTime)``` [https://gitter.im/nim-lang/Nim?at=5b4a6c41897d514b3c4efd21]
21:34:58	krux02	Araq: I am referring the the pcre library.
21:35:15	FromGitter	<tim-st> @Vindaar lol, that doesnt work for me...
21:35:21	krux02	tim-st: I looked at re, and there is a re.findAll
21:35:37	krux02	what is the problem with that?
21:35:58	FromGitter	<tim-st> for match in htm.findAll(re"(?s)<P .*?#0007'>\s*(.*?)\s*<\/SPAN.*?<\/P>"): ⏎ echo match
21:36:21	FromGitter	<tim-st> echos full block not like the result from Vindaar
21:36:50	FromGitter	<Vindaar> as you see, `findAll` also does that for me. But using the `=~` template (or `match` manually) works fine
21:37:05	FromGitter	<Vindaar> no idea what the difference is with `findAll`
21:37:11	krux02	ah, you are right. a find should return a Match, not a string
21:37:31	FromGitter	<tim-st> @Vindaar ah, now I see it you had two echos^^
21:37:38	FromGitter	<tim-st> so this is a bug?!
21:37:48	FromGitter	<tim-st> findall gives wrong result
21:38:15	krux02	That is a valid criticism on the regular expression library.
21:38:37	FromGitter	<tim-st> no, pythons does it too, it's ok for findall
21:38:40	Araq	findAll does what its docs say it does
21:38:51	krux02	Araq: that is not the problem
21:39:05	Araq	krux02: well PCRE version 10 has a new API and we should port re.nim to use that...
21:39:09	krux02	the problem is that a match as subpatters that one might want ot access
21:39:20	krux02	yes
21:39:52	FromGitter	<tim-st> oh, ok, so it doesnt care if I use brackets or not
21:39:55	Araq	it also uses a JIT
21:40:07	Araq	and the pure Nim implementation doesn't
21:40:15	FromGitter	<tim-st> then it's not possible to solve this problem with `re`
21:40:44	Araq	tim-st: try Nim's scanf :-)
21:41:20	FromGitter	<tim-st> ok, thanks, will have a look
21:41:33	krux02	tim-st. This is a real problem that is solvable. You could try to improve the situation by porting the API over.
21:42:09	FromGitter	<tim-st> that's too time consuming
21:42:44	krux02	well it is doable
21:42:53	FromGitter	<tim-st> I currently implement a perfect lmdb version in nim, that's more important since nim only supports the 20 years old sqlite
21:43:12	krux02	what is lmdb?
21:43:24	FromGitter	<tim-st> kv database
21:43:59	FromGitter	<tim-st> 700k random reads per second on my laptop in debug mode
21:44:03	Araq	20 years old sqlite is still unmatched.
21:44:32	FromGitter	<tim-st> no, the author of lmdb implemented sqlite using lmdb and it's 20 times quicker
21:44:35	*	chopzwei joined #nim
21:44:45	FromGitter	<tim-st> and additonally lmdb is much easier to use like hashmap
21:50:51	*	Trustable quit (Remote host closed the connection)
21:53:25	FromGitter	<data-man> @Araq: should PCRE2's wrapper and nre2 (for e.g.) had the same API as re/nre?
21:54:09	Araq	re.nim should use pcre2.nim but its API shouldn't change
21:54:24	FromGitter	<Vindaar> @Araq is there a specific reason why `strutils.removePrefix/Suffix`only offer in place versions? Less bloat? It's trivial to implement for me, sure
21:57:00	FromGitter	<rayman22201> @tim-st ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b4a71ac582aaa63076ab726]
21:57:07	FromGitter	<data-man> @Araq: do you have PCRE2 on Windows?
21:58:37	Araq	Vindaar: removeSuffix can be mapped to setLen() and so has an efficient inplace version. removePrefix follows removeSuffix for consistency.
21:58:51	FromGitter	<tim-st> @rayman22201 yes, it yields the very same text
21:59:07	FromGitter	<rayman22201> Is that what you want?
21:59:09	Araq	expression based variants are not available because we don't want to have two versions for every text handling proc
21:59:11	FromGitter	<tim-st> no
21:59:16	krux02	Vindaar: don't ask questions about strutils. Many people had many ideas how that library should look, and now it is a mess.
21:59:30	FromGitter	<rayman22201> what is your goal then @tim-st ? I think I missed that part
21:59:55	Araq	krux02: it's not that bad. we only have remove 50% of the procs in there
22:00:03	FromGitter	<tim-st> I wanted only the content inside `(.*?)`
22:00:17	*	dddddd joined #nim
22:00:24	Araq	however, I need to figure out why I *always* need to import strutils.
22:00:37	Araq	something in there must be super useful.
22:00:48	FromGitter	<Vindaar> @Araq: ah, that makes sense :) ⏎ @krux02: haha :D
22:00:57	FromGitter	<tim-st> about strutils: the best way is to overload all to be in place and return new string with the same proc name, and this is always secure for the compiler to find out, but some people dont want this
22:01:26	FromGitter	<tim-st> like `a.strip; let b = a.strip`
22:01:43	FromGitter	<tim-st> first in place, second return new
22:01:55	FromGitter	<rayman22201> @tim-st that gives me only the inside content?
22:02:11	krux02	Araq: I think the super usefull stuff of strutils could be moved to system. I mean when it is used anyway all the time it could be available all the time.
22:02:25	krux02	moving things from strutils to system also doesn't really break something.
22:02:26	FromGitter	<tim-st> @rayman22201 I wanted this but it didnt work for me :(
22:04:01	FromGitter	<rayman22201> what version of Nim is running on the Nim playground? does anybody know?
22:04:11	FromGitter	<tim-st> echo NimVersion?
22:04:27	FromGitter	<tim-st> dunno how to start nim bot
22:04:29	FromGitter	<Vindaar> pretty sure it's 0.18.0
22:04:46	FromGitter	<rayman22201> duh. thanks @tim-st. Yeah, it's 0.18.0
22:04:47	Araq	krux02: no, system is too bloated already
22:05:18	FromGitter	<rayman22201> I think it's a regression. On Nim playground the regex works. On my machine with a nightly devel, it gives the whole string.
22:05:23	Araq	also, we need to find the useful parts first.
22:05:31	FromGitter	<tim-st> ah, I have nightly too
22:05:34	FromGitter	<tim-st> maybe this is the thing
22:05:46	FromGitter	<tim-st> or because of windows?
22:06:06	FromGitter	<tim-st> I also get whole string
22:06:10	FromGitter	<rayman22201> Well, I am Ubuntu, and it gives me the whole string
22:06:23	Araq	probably just toLower/toUpper, replace, replaceWord and `%`
22:06:29	*	Vladar quit (Quit: Leaving)
22:06:36	FromGitter	<rayman22201> I think it's a regression on devel. I will have to verify
22:06:57	FromGitter	<tim-st> Araq: contains
22:07:08	FromGitter	<tim-st> and maybe a few others
22:07:28	Araq	the rather uninspired allCharsInSet :-)
22:08:06	FromGitter	<tim-st> also I thought about some procs for strutils and sequtils should be the same
22:08:26	FromGitter	<tim-st> some time ago I wanted to concatenate seqs but there wasnt away
22:08:47	FromGitter	<tim-st> I think nearly everything from strutils can be used in sequtils too
22:09:16	FromGitter	<Vindaar> @tim-st https://nim-lang.org/docs/sequtils.html#concat,varargs[seq[T]] ?
22:09:37	FromGitter	<tim-st> when was this added?
22:09:41	FromGitter	<tim-st> :D
22:09:54	FromGitter	<Vindaar> ages ago I think. Ever since I've been using Nim it's been available afair
22:10:04	FromGitter	<tim-st> well, why is it called `concat` and not `add`
22:10:23	FromGitter	<tim-st> it's a problem in nim that every module introduces own styles
22:10:32	FromGitter	<tim-st> like sets -> `incl`
22:10:39	FromGitter	<tim-st> every `add` has an own name
22:11:25	FromGitter	<rayman22201> wow, 0.18.0 is showing the entire string as well... weird
22:11:31	FromGitter	<rayman22201> on my machine
22:11:39	FromGitter	<rayman22201> Nim Playground does the right thing though...
22:11:43	FromGitter	<Vindaar> because adding is something different from concatenating. `add` would add something to an existing `seq` in place, while `concat` concatenates different seqs to a new one?
22:11:44	FromGitter	<tim-st> yes, thought so, it's a compiled dll that does the work
22:11:59	krux02	there is strutils and sequilts and algorithm
22:12:05	FromGitter	<rayman22201> so maybe different PCRE version
22:12:14	krux02	I relly don't get the distinction between algorithm and sequtils
22:12:50	FromGitter	<tim-st> @Vindaar but I wanted `add` and it doesnt exists
22:12:50	krux02	maybe those two can be joined
22:13:20	FromGitter	<tim-st> this is actually the proc that was missing, now we found it
22:14:39	FromGitter	<Vindaar> `add` like this? https://nim-lang.org/docs/system.html#add,seq[T],openArray[T]
22:17:05	FromGitter	<tim-st> now I remember^^ it was `find` xD
22:17:28	FromGitter	<tim-st> I needed to see if a subseq is inside the seq and if so find the index
22:18:57	FromGitter	<tim-st> so find and contains are missing in sequtils
22:20:21	FromGitter	<Vindaar> ```let s = @[@[1, 2], @[2, 3], @[3, 5]] ⏎ let inS = @[2, 3] ⏎ ⏎ echo s.find(inS) ⏎ # 1``` [https://gitter.im/nim-lang/Nim?at=5b4a772466c1e833a9e8f38c]
22:20:27	FromGitter	<Vindaar> works just fine
22:20:46	FromGitter	<tim-st> no, you have seq[seq]
22:20:55	FromGitter	<tim-st> I'm talking about same seq types
22:21:17	FromGitter	<tim-st> @[1,2,3] in @[1,2,3,4] ? -> true
22:21:21	FromGitter	<Vindaar> ahhh!
22:21:43	FromGitter	<tim-st> I really needed this proc to build my super fast unicode collation algorithm
22:21:49	FromGitter	<tim-st> I ended up writing it myself
22:22:22	dom96	Reminder: answer our community survey here: https://goo.gl/forms/t2gWWBEDC1walw6s2
22:22:39	FromGitter	<Vindaar> I mean that is quite specific, no? not *everything* has to be in the stdlib after all
22:23:08	FromGitter	<tim-st> yes, but also sequtils is specific as whole module
22:27:12	FromGitter	<Vindaar> @dom96 since you're around: `nimble path` seems to show the location of the package installation *including* what's specified as `srcDir` in the nimble file. Although that location doesn't actually exist in the corresponding `.nimble` folder. Is it supposed to be like that?
22:28:22	FromGitter	<Vindaar> already answered the survey :)
22:29:05	FromDiscord	<exelotl> uh, this is annoying https://cdn.discordapp.com/attachments/239878713016188939/467819715394535440/unknown.png
22:29:47	FromDiscord	<exelotl> now any time I try to print a `cint`, it uses the `$` I defined for the `Version` type
22:30:06	dom96	Vindaar: It shouldn't be like that.
22:30:28	FromDiscord	<exelotl> I was hoping I could just define $ for the subtype and not have it affect the base type
22:30:31	FromGitter	<tim-st> just use cint.int
22:30:47	FromGitter	<Vindaar> ok, thought so! I could try to fix that in a few days
22:30:51	FromGitter	<tim-st> or distinct
22:31:31	FromDiscord	<exelotl> yeah I don't want to mess up cint for other code, but distinct is so annoying because I have to explicitly pull in all the operations I could ever want to use
22:32:03	FromGitter	<tim-st> hm^^
22:32:18	skrylar	lmdb sounds interesting
22:33:12	FromGitter	<tim-st> it is, but it has downside: the database size is not growing automatically, you can resize on closed db or when no transaction is open, but I cannot find out for sure if no transaction is open
22:33:25	FromDiscord	<exelotl> oh wait you're right it's not too bad, in this case I can just do e.g. `proc error*(v:Version):bool = v.cint < 0`
22:33:40	FromGitter	<tim-st> I implement all procs to find out loadfactor and resize to do it manually
22:34:49	*	jamesroseman joined #nim
22:39:48	*	NimBot joined #nim
22:45:49	*	NimBot joined #nim
22:46:49	*	jamesroseman quit (Ping timeout: 260 seconds)
22:46:59	*	jamesroseman joined #nim
22:51:45	*	NimBot joined #nim
22:57:41	*	NimBot joined #nim
23:03:35	*	NimBot joined #nim
23:09:42	*	NimBot joined #nim
23:10:20	*	brainpro1 is now known as brainproxy
23:15:44	*	NimBot joined #nim
23:21:31	*	NimBot joined #nim