| 00:07:50 | * | pregressive joined #nim |
| 00:09:10 | * | dashed joined #nim |
| 00:09:27 | keyle | it's pretty much the point |
| 01:14:26 | * | brson quit (Ping timeout: 272 seconds) |
| 01:14:53 | * | Demon_Fox quit (Quit: Leaving) |
| 01:17:36 | * | Demon_Fox joined #nim |
| 01:31:25 | * | endragor joined #nim |
| 01:48:29 | * | bendiken quit () |
| 01:55:46 | * | chemist69_ joined #nim |
| 01:58:54 | * | chemist69 quit (Ping timeout: 255 seconds) |
| 02:00:13 | * | ^aurora^ quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 02:25:13 | * | eliudnir joined #nim |
| 02:25:35 | * | eliudnir left #nim ("Leaving") |
| 02:53:44 | * | pregressive quit (Remote host closed the connection) |
| 03:02:31 | * | jck quit (Remote host closed the connection) |
| 03:02:51 | * | jck joined #nim |
| 03:14:35 | * | endragor quit (Remote host closed the connection) |
| 03:22:07 | * | dashed quit (Quit: Connection closed for inactivity) |
| 03:30:44 | * | gsingh93 quit (Ping timeout: 244 seconds) |
| 03:31:54 | * | jck quit (Read error: Connection reset by peer) |
| 03:32:35 | * | gsingh93 joined #nim |
| 03:33:01 | * | jck joined #nim |
| 03:37:25 | * | endragor joined #nim |
| 03:39:02 | * | endragor quit (Remote host closed the connection) |
| 03:40:01 | * | endragor joined #nim |
| 03:42:18 | * | endragor quit (Remote host closed the connection) |
| 03:50:38 | * | CryptoToad joined #nim |
| 04:01:48 | * | endragor joined #nim |
| 04:17:13 | * | endragor quit (Ping timeout: 250 seconds) |
| 04:17:43 | * | wh1t3r0s3 quit (Ping timeout: 244 seconds) |
| 04:19:53 | * | darkf joined #nim |
| 04:24:57 | * | NimBot joined #nim |
| 04:27:14 | * | endragor quit (Remote host closed the connection) |
| 04:27:42 | * | dom96 joined #nim |
| 04:42:30 | * | vega_nsk joined #nim |
| 04:44:13 | * | vega_nsk quit (Client Quit) |
| 04:44:31 | * | vega_nsk joined #nim |
| 05:13:21 | * | kilon quit () |
| 05:28:29 | * | Demon_Fox_ joined #nim |
| 05:30:30 | * | Demon_Fox quit (Disconnected by services) |
| 05:30:37 | * | Demon_Fox_ is now known as Demon_Fox |
| 05:37:01 | * | kilon joined #nim |
| 05:50:49 | * | jszymanski joined #nim |
| 05:56:29 | * | xcombelle joined #nim |
| 06:00:17 | * | ^aurora^ joined #nim |
| 06:22:48 | * | kilon quit () |
| 06:24:15 | * | chemist69_ is now known as chemist69 |
| 06:24:32 | * | kilon joined #nim |
| 06:26:21 | * | BitPuffin|osx quit (Ping timeout: 250 seconds) |
| 06:31:58 | * | Jesin quit (Quit: Leaving) |
| 06:34:10 | * | Jesin joined #nim |
| 06:51:30 | gokr1 | Morning! |
| 06:54:12 | * | kilon quit () |
| 07:03:16 | * | Ven joined #nim |
| 07:04:16 | * | Toad_ joined #nim |
| 07:07:44 | * | CryptoToad quit (Ping timeout: 244 seconds) |
| 07:10:15 | chemist69 | Morning gokr1 |
| 07:16:40 | * | NimBot joined #nim |
| 07:19:58 | * | BitR quit (Ping timeout: 250 seconds) |
| 07:39:12 | * | filcuc joined #nim |
| 07:41:06 | * | yglukhov joined #nim |
| 07:45:05 | * | yglukhov quit (Ping timeout: 240 seconds) |
| 07:53:10 | * | Toad_ quit (Quit: Leaving) |
| 07:59:32 | * | ^aurora^ quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 08:01:50 | * | endragor joined #nim |
| 08:03:48 | * | yglukhov joined #nim |
| 08:03:51 | * | yglukhov quit (Remote host closed the connection) |
| 08:04:24 | * | yglukhov joined #nim |
| 08:10:49 | * | Trustable joined #nim |
| 08:27:43 | OnO | morning |
| 08:27:51 | ekarlso | morn |
| 08:35:59 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 08:37:39 | * | Demon_Fox quit (Quit: Leaving) |
| 08:44:46 | * | jszymanski quit (Quit: computer sleeps...) |
| 08:51:09 | * | Ven joined #nim |
| 08:54:30 | * | X67r joined #nim |
| 08:55:38 | * | silven quit (Ping timeout: 265 seconds) |
| 09:11:36 | yglukhov | quick question. how do i convert ident to string? |
| 09:12:39 | yglukhov | template foo(op: expr): stmt = echo op; foo(`+`) |
| 09:18:28 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 09:26:52 | * | hbguy joined #nim |
| 09:26:58 | hbguy | hello all |
| 09:27:30 | hbguy | anyone alive? |
| 09:28:43 | * | X67r quit (Ping timeout: 268 seconds) |
| 09:29:07 | * | X67r joined #nim |
| 09:32:29 | * | hbguy quit (Quit: Page closed) |
| 09:33:39 | federico3 | *cough* |
| 09:41:40 | * | vega_nsk quit (Quit: Ухожу я от вас (xchat 2.4.5 или старше)) |
| 09:42:02 | * | vega_nsk joined #nim |
| 10:01:05 | OnO | yeah, alive I kicking |
| 10:13:24 | * | Ven joined #nim |
| 10:19:26 | keyle | ish |
| 10:19:34 | * | aziz joined #nim |
| 10:36:19 | NimBot | nim-lang/Nim devel 7dd263f Reimer Behrends [+0 ±1 -0]: Fix readLine handling of long lines. |
| 10:36:19 | NimBot | nim-lang/Nim devel 5a9d2d8 Andreas Rumpf [+0 ±1 -0]: Merge pull request #3266 from rbehrends/fix-readline... 2 more lines |
| 10:37:58 | NimBot | nim-lang/Nim devel a3cc519 JamesP [+0 ±1 -0]: bug fix: Switch state to dbgSkipCurrent for Eval, Local & Global commands.... 3 more lines |
| 10:37:58 | NimBot | nim-lang/Nim devel d614611 Andreas Rumpf [+0 ±1 -0]: Merge pull request #3280 from jlp765/endb... 2 more lines |
| 10:47:49 | * | redwyn joined #nim |
| 10:52:52 | * | jszymanski joined #nim |
| 10:57:49 | * | raza joined #nim |
| 11:00:00 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 11:02:57 | * | NimBot joined #nim |
| 11:03:22 | * | NimBot joined #nim |
| 11:15:12 | * | redwyn quit (Ping timeout: 246 seconds) |
| 11:25:11 | * | UberLambda joined #nim |
| 11:26:45 | * | drewsrem joined #nim |
| 11:28:30 | * | Kingsquee quit (Quit: http://i.imgur.com/EsXzoum.png) |
| 11:43:01 | * | Ven joined #nim |
| 11:53:26 | * | BitPuffin joined #nim |
| 11:55:58 | * | NimBot joined #nim |
| 12:04:37 | drewsrem | yglukhov, turns out the function-pointer casting I did yesterday works if you use nimcall, I suspect it's because nimcall actually generates a function-pointer whereas closures in nim generate structs |
| 12:08:12 | * | wuehlmaus joined #nim |
| 12:09:05 | * | filcuc quit (Ping timeout: 240 seconds) |
| 12:09:47 | NimBot | nim-lang/Nim devel 1b47941 JamesP [+0 ±1 -0]: fixes #3220 lib/impure/db_mysql.nim getValue()... 2 more lines |
| 12:09:47 | NimBot | nim-lang/Nim devel c3d399e Andreas Rumpf [+0 ±1 -0]: Merge pull request #3264 from jlp765/db_mysql... 2 more lines |
| 12:21:53 | * | ^aurora^ joined #nim |
| 12:22:13 | NimBot | nim-lang/Nim devel 02214d8 Yuriy Glukhov [+0 ±1 -0]: Fixed JS copying. |
| 12:22:13 | NimBot | nim-lang/Nim devel 580855d Andreas Rumpf [+0 ±1 -0]: Merge pull request #3273 from yglukhov/js-copy-fix... 2 more lines |
| 12:28:51 | * | redwyn joined #nim |
| 12:29:14 | yglukhov | drewsrem does it mean that casting closure to nimcall proc will loose the ctx? |
| 12:29:47 | drewsrem | yglukhov, oh, idk, I just forwent closures completely |
| 12:30:23 | yglukhov | oh ok |
| 12:30:47 | drewsrem | and I'll see how far I can go without them |
| 12:32:38 | * | NimBot joined #nim |
| 12:36:26 | Araq | yglukhov: https://github.com/nim-lang/Nim/pull/3274 is really a big problem |
| 12:36:49 | yglukhov | mixing js code to system? |
| 12:37:00 | yglukhov | or why? |
| 12:37:18 | Araq | the 'when defined(js)' is so fragile, it means when targetting JS, 'del' won't work at compile-time anymore |
| 12:37:44 | Araq | you need to patch "delete" and "insert" in the jsgen |
| 12:38:02 | yglukhov | argh, thats nasty |
| 12:38:22 | Araq | there is some mechanism for that via "stdlib.system.delete" |
| 12:38:42 | Araq | you can look at vmgen for how this can work |
| 12:39:10 | Araq | alternatively we need a new 'when' construct that keeps everything around until the backend gets to pick its branch |
| 12:39:41 | Araq | it's a tough problem, it's a 'when' where every branch is sem'checked |
| 12:39:53 | Araq | well in other words ... an 'if' ;-) |
| 12:40:11 | Araq | so ... codegens need to be able to deal with 'if defined(...)' |
| 12:40:12 | yglukhov | in D it is "if" |
| 12:40:23 | Araq | and now you know why |
| 12:40:54 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 12:44:25 | yglukhov | well, doing stupid "if(ctfe)" semchecks and codegen should not be really hard =). For now we could just forbid using ctfe in any expr other than if(ctfe) |
| 12:44:35 | yglukhov | or even when ctfe |
| 12:47:04 | Araq | well 'defined' already is a magic |
| 12:47:12 | Araq | if defined(js) should do it |
| 12:50:38 | * | redwyn left #nim (#nim) |
| 12:53:59 | * | saml_ joined #nim |
| 12:54:21 | * | NimBot joined #nim |
| 13:01:12 | * | nchambers is now known as dtscode |
| 13:06:38 | * | FedeOmoto joined #nim |
| 13:10:14 | yglukhov | ok, but defined(js) is already a valid construct now, and thus it can be used in complex expressions. E.g. var a = defined(js) and (b == c) or smth. and my proposal is to use some magic symbol that can be used ONLY in an "if magicSymbol" construct. that should simplify semantics and codegen, right? |
| 13:10:48 | * | vega_nsk quit (Quit: Ухожу я от вас (xchat 2.4.5 или старше)) |
| 13:11:30 | yglukhov | otherwise there would have to be some complex tracking of what this defined(js) may affect |
| 13:12:01 | yglukhov | and i'm not really good at fancy-compiler-recursive-algorithms |
| 13:14:03 | * | endragor quit (Ping timeout: 252 seconds) |
| 13:14:26 | Araq | so instead of a simple check that mDefined is not mixed with mOr you want this restriction in the grammar? |
| 13:14:45 | Araq | that's far more work |
| 13:15:01 | Araq | plus it doesn't compose: |
| 13:15:25 | Araq | template t(x): auto = defined(x) |
| 13:15:40 | Araq | if t(js): ... # works in my proposal |
| 13:15:50 | Araq | # breaks in yours :P |
| 13:17:40 | drewsrem | Araq, do you happen to have example-code for the virtual table dynamic dispatch thing you talked about yesterday? |
| 13:18:18 | * | BitR joined #nim |
| 13:19:14 | yglukhov | no i don't want to change grammar in any way. i want to introduce a magic symbol. let's call it "ctfe". then semantics when analysing "if" node checks if "if"'s only expression is "ctfe", and marks "ctfe" node as valid. when reaching unmarked "ctfe" node semantics will produce an error: "ctfe" may only be used in "if ctfe". |
| 13:19:54 | yglukhov | vm codegen transletes "if ctfe" nodes to thenBranch of the node. |
| 13:20:15 | yglukhov | cgen and js gen translate "if ctfe" nodes to elseBranch of the node |
| 13:21:20 | Araq | we already have this magic and it's called mDefined, handled by semDefined |
| 13:21:54 | yglukhov | but mDefined can be currently used everywhere, not only in ifs. |
| 13:23:20 | Araq | drewsrem: sorry, no, it's an exercise for the reader |
| 13:23:49 | yglukhov | drewsrem, Araq, isn't it what streams.nim are doing? |
| 13:23:54 | * | irrequietus joined #nim |
| 13:24:13 | * | keyle quit (Quit: http://twitter.com/keyle/) |
| 13:24:14 | Araq | yglukhov: you're right but that doesn't make it hard to implement |
| 13:25:11 | Araq | yglukhov: no, streams doesn't require multiple dispatch |
| 13:25:41 | yglukhov | ah, you're talking about that matrix dispatch thing. sorry |
| 13:26:46 | yglukhov | Araq, what if you require different VM behavior depending on C or JS target? |
| 13:27:26 | yglukhov | my approach. |
| 13:27:30 | yglukhov | if ctfe: |
| 13:27:35 | yglukhov | when defined(js): |
| 13:27:39 | Araq | what if I just require different backend behaviour? |
| 13:27:42 | yglukhov | dosmthforjs |
| 13:27:47 | yglukhov | else: |
| 13:27:52 | yglukhov | dosmthfoc |
| 13:27:55 | Araq | if defined(cpp): |
| 13:28:03 | Araq | elif defined(js): |
| 13:28:08 | Araq | elif defined(nimscript): |
| 13:28:24 | Araq | else: for_da_c_backend() |
| 13:28:53 | yglukhov | nimscript == vm? |
| 13:29:00 | Araq | yes |
| 13:30:07 | yglukhov | nimscript can be called when compiling to both js and c. how do you distinguish the target inside nimscript code? |
| 13:31:02 | Araq | js is still defined, you need to order and nest stuff properly |
| 13:31:30 | * | endragor joined #nim |
| 13:32:42 | * | endragor quit (Remote host closed the connection) |
| 13:34:52 | yglukhov | how will codegens skip the non-related parts? we'll have to track all of the expresisons if they are influenced by defined(nimscript)? |
| 13:35:42 | yglukhov | or analyse only "if defined(nimscript)" and disallow "defined(nimscript)" anywhere else? |
| 13:36:02 | * | endragor joined #nim |
| 13:36:04 | * | saml_ quit (Ping timeout: 272 seconds) |
| 13:38:01 | yglukhov | and what do we do with "const foo = defined(nimscript)" will it actually change it's value between ct/rt? |
| 13:40:54 | yglukhov | do you agree that we have to make this defined(nimscript) magic available only when it is a single expression of if node? |
| 13:41:33 | yglukhov | and maybe relax this requirement in the future. or maybe not. |
| 13:42:33 | Araq | ok, ok, you convinced me |
| 13:43:23 | yglukhov | errr... about using new magic or about not tracking defined(nimscript)? =) |
| 13:43:44 | Araq | about yet another magic with subtle semantic differences |
| 13:43:56 | Araq | people already have problems with declared vs defined |
| 13:44:22 | yglukhov | i think they should be unified. |
| 13:44:34 | Araq | no way |
| 13:44:34 | yglukhov | i mean declared == defined. |
| 13:44:38 | yglukhov | why? |
| 13:44:42 | Araq | no, we used to have that |
| 13:45:03 | Araq | introduced subtle bugs |
| 13:45:22 | yglukhov | ok, i'll join the camp who has problems understanding the principal difference =) |
| 13:45:53 | Araq | it's really not hard: defined(foo) checks for -d:foo |
| 13:46:07 | Araq | declared(foo) checks for a *Nim* symbol foo |
| 13:46:21 | Araq | when declared(strutils.replace): |
| 13:46:58 | yglukhov | just allow -d:foo="mystringvalue", with -d:foo resulting in const foo="" and thats it. no? |
| 13:48:09 | yglukhov | that way you'll have both defined and declared in a backward compatible manner, and just deprecate one of those |
| 13:49:44 | * | endragor quit (Remote host closed the connection) |
| 13:50:03 | Araq | and where is this const foo? |
| 13:50:05 | Araq | system.nim? |
| 13:51:38 | yglukhov | is it so important? well, maybe the best way to place it in a pseudo-module. e.g. nimCmdLine.foo or smth |
| 13:52:02 | yglukhov | * not the best, but the first solution that comes to mind |
| 13:52:28 | yglukhov | maybe even system.nim, why not? |
| 13:53:40 | * | ^aurora^ quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 13:54:17 | yglukhov | also, afair there's some magic to get nim's cmdline. why not remowing this defined in favour of a library function. |
| 13:54:46 | Araq | what? |
| 13:55:16 | Araq | do you have any idea what it would take to remove or even deprecate 'define' at this point? |
| 13:55:28 | Araq | *defined |
| 13:56:03 | * | NimBot joined #nim |
| 13:56:27 | yglukhov | well, i'm talking conceptually. "remove" may mean "forget about" in my case =) |
| 13:57:03 | yglukhov | also defined could become a lib function |
| 13:57:18 | Araq | anyway the new magic shall be called nimcodegen as codegen is a too useful name and we still don't have 'import system except codegen' |
| 13:57:50 | Araq | if nimcodegen(js) # a bit verbose though |
| 13:58:10 | yglukhov | why should it need an arg? |
| 13:58:21 | yglukhov | why not just nimcodegen? |
| 13:58:33 | yglukhov | or nimvm i like better |
| 13:58:46 | Araq | or perhaps nimTarget |
| 13:59:14 | yglukhov | nimTarget is even further from my understanding. |
| 13:59:23 | Araq | if nimvm ? but then how would i check for the C++ target? |
| 13:59:43 | yglukhov | if nimvm: |
| 13:59:51 | yglukhov | when defined(cpp): |
| 14:00:01 | yglukhov | elif defined(js): |
| 14:00:05 | yglukhov | etc |
| 14:00:39 | yglukhov | when defined(cpp) - that works already, doesn't it? |
| 14:00:49 | yglukhov | i've never compiled to cpp yet %) |
| 14:02:26 | Araq | you know what? you implement it, you get to pick the name |
| 14:03:27 | yglukhov | name is the easiest part. =). semantics is what i want to have your (and other nimers) approval for. |
| 14:04:20 | yglukhov | does "if nimvm" look good enough? i mean i know it looks clumsy, but i doubt it would be used often. |
| 14:04:40 | yglukhov | nim has gone a long way without it |
| 14:05:06 | Araq | well nimvm is fine but I wonder if 'when nimvm' wouldn't be easier in the backends |
| 14:05:22 | Araq | since they currently don't handle nkWhen at all |
| 14:06:18 | Araq | bbs |
| 14:06:19 | yglukhov | that depends, at what stage of semantics nkWhen gets stripped away. |
| 14:06:48 | yglukhov | and how much more semantics do we have to take this nkWhen through |
| 14:09:48 | yglukhov | but "when nimvm" looks better imho. |
| 14:10:43 | * | bendiken joined #nim |
| 14:26:14 | * | Ven joined #nim |
| 14:30:56 | * | raza quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) |
| 14:38:06 | * | raza joined #nim |
| 14:45:44 | * | anthgur joined #nim |
| 14:52:09 | * | anthgur quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 14:54:24 | * | Deltron joined #nim |
| 14:54:33 | * | Deltron left #nim (#nim) |
| 15:00:52 | * | aziz quit (Remote host closed the connection) |
| 15:25:13 | * | xcombelle quit (Remote host closed the connection) |
| 15:34:02 | * | dtscode is now known as DeerGod |
| 15:42:17 | * | irrequietus quit () |
| 15:43:00 | * | irrequietus joined #nim |
| 15:44:31 | * | xet7_ quit (Ping timeout: 246 seconds) |
| 15:44:57 | * | jszymanski quit (Quit: computer sleeps...) |
| 15:45:13 | * | xet7_ joined #nim |
| 15:45:38 | * | irrequietus quit (Client Quit) |
| 15:49:55 | * | irrequietus joined #nim |
| 15:55:47 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 15:56:45 | * | anthgur joined #nim |
| 16:12:42 | OnO | Araq: I've seen you're working on escape analysis, that's for skipping GC and using stack instead where possible? or maybe to detect unused variables? |
| 16:14:05 | * | durm left #nim ("Konversation terminated!") |
| 16:17:43 | Araq | OnO: good question. |
| 16:18:04 | Araq | currently I'm working out how to use this to make the GC a pure refcounting GC |
| 16:18:43 | Araq | so that no stack tracing is required, this would solve the "can I use Nim for hard realtime?" problem |
| 16:19:20 | Araq | as well as the "Nim's GC is thread local" problem |
| 16:19:43 | Araq | since elimination of RC ops automatically affects atomic reference counting as well |
| 16:21:22 | Araq | so yeah, it's not only a classic optimization. It's my secret Rust killer ;-) |
| 16:21:50 | yglukhov | Araq, "when nimvm" works! =) https://github.com/nim-lang/Nim/pull/3286 |
| 16:22:09 | Araq | since it infers immutability as well as ownership |
| 16:23:15 | Araq | we'll see how it works out, currently simple examples work and it adds 0.2s to the bootstrapping times |
| 16:24:30 | * | UberLambda quit (Ping timeout: 240 seconds) |
| 16:24:53 | OnO | sounds great |
| 16:24:56 | * | yglukhov_ joined #nim |
| 16:26:44 | Araq | well I have thought about this algorithm for 3 years now ... so I really hope it is great |
| 16:27:04 | * | anthgur quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 16:27:34 | * | UberLambda joined #nim |
| 16:27:51 | Araq | yglukhov_: nice, but it's wrong unless you spec the 'nimvm' as "keyword-like" |
| 16:27:55 | OnO | how does it differ from Apple's ARC? |
| 16:28:18 | OnO | would it be still able to tear down ref cycles? |
| 16:28:45 | * | yglukhov quit (Ping timeout: 255 seconds) |
| 16:29:29 | * | yglukhov_ quit (Ping timeout: 265 seconds) |
| 16:30:23 | * | anthgur joined #nim |
| 16:32:47 | * | anthgur quit (Client Quit) |
| 16:33:53 | * | endragor joined #nim |
| 16:35:04 | * | endragor quit (Remote host closed the connection) |
| 16:36:42 | Araq | ref cycles are not a problem for a purely RC'ing GC |
| 16:37:06 | Araq | it's called "trial deletion", has been known for a decade now |
| 16:37:34 | Araq | and even PHP's implementation uses it |
| 16:38:14 | Araq | not sure why everybody still acts like it's impossible to solve |
| 16:40:09 | Araq | it's a bit like "omg, malloc is slow because it uses freelists" ... well, O(1) alloc and dealloc (with freelists) is also a solved problem |
| 16:42:42 | * | jaco60 joined #nim |
| 16:42:53 | * | kilon joined #nim |
| 16:44:59 | NimBot | nim-lang/Nim devel aa8da7f Araq [+0 ±1 -0]: optimized writetracking |
| 16:44:59 | NimBot | nim-lang/Nim devel e90c155 Araq [+0 ±1 -0]: fixes 'line too long' warnings |
| 16:44:59 | NimBot | nim-lang/Nim devel a373bdd Araq [+0 ±3 -0]: improvements for NimScript support |
| 16:51:02 | OnO | btw. any hints to make asynchttpserver multithreaded? would new GC make it easier |
| 16:51:53 | OnO | I think it would be easiest to make it multithreaded at socket level, not request revel |
| 16:52:23 | OnO | I am vitually interested on having that working, since I am now trying to turn Nim into some nice server solution |
| 16:52:44 | Araq | well I started hacking on that problem but I didn't really know what I was doing |
| 16:53:16 | Araq | I added a shared hash table and made the code .gcsafe |
| 16:53:40 | Araq | the idea is that multiple worker threads read from the same shared hash table which socket are ready to process |
| 16:54:26 | * | vendethiel joined #nim |
| 16:54:51 | Araq | or rather, each thread gets a list of sockets to asyncronously work on |
| 16:55:53 | OnO | you mean that on accept socket is bound to thread? |
| 16:56:01 | Araq | yes |
| 16:56:48 | OnO | okay, then what do we need actually to pass? just a socket handle (int) or something more? |
| 17:00:06 | Araq | yeah an int but the async socket implementation uses 'ref' like there is no tomorrow :P |
| 17:03:44 | * | brson joined #nim |
| 17:09:55 | * | yglukhov joined #nim |
| 17:10:17 | * | anthgur joined #nim |
| 17:15:11 | Araq | yglukhov: got my message? |
| 17:18:13 | * | anthgur quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 17:21:31 | yglukhov | yes |
| 17:21:43 | yglukhov | now how do i do that... |
| 17:22:10 | yglukhov | this has to be a new token? |
| 17:42:03 | * | ^aurora^ joined #nim |
| 17:44:04 | yglukhov | Araq, do you mean that parser has to forbid nimvm? |
| 17:48:11 | * | icebattle quit (Quit: Page closed) |
| 17:53:09 | * | unclechu joined #nim |
| 17:57:33 | * | jszymanski joined #nim |
| 18:03:46 | * | UberLambda quit (Ping timeout: 240 seconds) |
| 18:06:24 | * | BitPuffin quit (Ping timeout: 264 seconds) |
| 18:08:20 | yglukhov | Araq, updated pr with lexer/parser changes. |
| 18:22:59 | * | xcombelle joined #nim |
| 18:23:03 | * | darkf quit (Quit: Leaving) |
| 18:23:12 | * | filcuc joined #nim |
| 18:23:23 | * | brson quit (Quit: leaving) |
| 18:25:12 | * | brson joined #nim |
| 18:30:47 | Araq | omg, no |
| 18:31:06 | Araq | you need to introduce a new magic and check for that after symbol lookup, so that: |
| 18:31:11 | Araq | const nimvm = 4 |
| 18:31:23 | Araq | when nimvm == 8 works |
| 18:36:40 | yglukhov | allow "nimvm" shadowing? |
| 18:36:53 | Xe | is nimvm nimscript? |
| 18:38:28 | yglukhov | Xe. kinda. It's a magic flag that let's you know whether you code is running in compile-time or runtime. |
| 18:39:07 | yglukhov | with "nimscript" notion this distiction gets blurred imho. |
| 18:39:22 | yglukhov | so i'd rather prefer to call it nimvm =) |
| 18:39:27 | yglukhov | D calls it CTFE |
| 18:40:33 | * | boopsiesisaway is now known as boopsies |
| 18:49:29 | federico3 | OnO: why adding multithreading to async? |
| 18:53:27 | OnO | because I want multithreaded web service with nim, currently asynchttpserver is superfast, but it is single threaded |
| 18:59:16 | Araq | OnO: speaking of which, I studied some more papers. Essentially the approach of our async implementation is really as good as it can get except of course our memory management is far from optimal |
| 19:00:49 | Araq | but our state machine generation is optimal |
| 19:03:28 | * | NimBot joined #nim |
| 19:03:56 | * | NimBot joined #nim |
| 19:21:15 | * | golak joined #nim |
| 19:33:24 | * | kilon quit () |
| 19:33:46 | * | raza quit (K-Lined) |
| 19:38:45 | * | UberLambda joined #nim |
| 19:44:42 | * | BitPuffin|osx joined #nim |
| 19:46:00 | * | UberLambda quit (Quit: GTG) |
| 19:49:41 | * | ^aurora^ quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 19:51:54 | * | Matthias247 joined #nim |
| 19:53:24 | * | askatasuna joined #nim |
| 19:55:15 | * | xcombelle quit (Remote host closed the connection) |
| 20:06:07 | * | Jehan_ joined #nim |
| 20:15:10 | * | golak quit (Ping timeout: 240 seconds) |
| 20:23:18 | * | filcuc quit (Ping timeout: 272 seconds) |
| 20:25:38 | * | dv- joined #nim |
| 20:25:52 | dv- | where did the repl (nim i) go? |
| 20:29:13 | * | Matthias247 quit (Read error: Connection reset by peer) |
| 20:29:15 | strcmp1 | dv-, i think its still there but as something nonobvious because it has many quirks or something |
| 20:37:38 | NimBot | nim-lang/Nim devel ef5f0b1 Adam Strzelecki [+0 ±1 -0]: times: Export & document countLeapYears/Years/Days... 6 more lines |
| 20:37:38 | NimBot | nim-lang/Nim devel 4fefdb4 Dominik Picheta [+0 ±1 -0]: Merge pull request #3281 from nanoant/patch/lib-pure-times-hotfix... 2 more lines |
| 20:39:08 | dom96 | dv-: its been deprecated |
| 20:46:11 | * | Demon_Fox joined #nim |
| 20:58:26 | federico3 | OnO: for a server environment I'd rather have multiple prpcesses |
| 21:01:07 | * | jszymanski quit (Quit: computer sleeps...) |
| 21:07:20 | Araq | dv-: 'nim secret' |
| 21:27:03 | dv- | oo, now i'm part of the nim cabal |
| 21:27:40 | strcmp1 | lol |
| 21:52:18 | Araq | dom96: why does nimble say "WARNING: No nimblemeta.json file found in blah/compiler-0.11.3" ? |
| 21:53:03 | dom96 | Araq: the compiler package was installed with a super old version of nimble? |
| 21:53:10 | dom96 | you deleted the nimblemeta.json file |
| 21:53:13 | dom96 | or a bug in nimble |
| 21:53:32 | dom96 | try nimble install compiler@#head |
| 21:54:08 | Araq | well I'm testing nimble with lexim.nims |
| 21:54:21 | Araq | and it complains about my hacked together compiler package |
| 21:54:32 | Araq | and I wonder if that's the correct behaviour |
| 21:55:04 | Araq | because lexim doesn't depend on the compiler package |
| 21:55:51 | * | askatasuna quit (Ping timeout: 244 seconds) |
| 21:57:01 | * | boopsies is now known as boopsiesisaway |
| 22:04:07 | dom96 | it will warn about everything |
| 22:04:13 | dom96 | even if the package isn't used |
| 22:09:15 | * | askatasuna joined #nim |
| 22:10:05 | * | yglukhov quit (Remote host closed the connection) |
| 22:15:34 | * | Kingsquee joined #nim |
| 22:21:56 | Araq | good night |
| 22:21:57 | * | brson quit (Quit: leaving) |
| 22:26:15 | * | askatasuna quit (Ping timeout: 260 seconds) |
| 22:32:17 | * | brson joined #nim |
| 22:35:12 | * | jaco60 quit (Quit: Leaving) |
| 22:39:07 | * | askatasuna joined #nim |
| 22:45:54 | * | NimBot joined #nim |
| 22:46:15 | * | Kingsquee quit (Read error: Connection reset by peer) |
| 22:46:30 | * | Kingsquee joined #nim |
| 22:46:51 | * | keyle joined #nim |
| 22:59:07 | * | ^aurora^ joined #nim |
| 23:01:26 | * | unclechu quit (Remote host closed the connection) |
| 23:04:51 | * | Jehan_ quit (Quit: Leaving) |
| 23:10:41 | * | yglukhov joined #nim |
| 23:15:02 | * | yglukhov quit (Ping timeout: 250 seconds) |
| 23:18:29 | * | stomplopster joined #nim |
| 23:28:50 | * | yglukhov joined #nim |
| 23:29:07 | * | yglukhov quit (Remote host closed the connection) |
| 23:31:46 | * | Trustable quit (Remote host closed the connection) |
| 23:33:41 | * | infinity0 quit (Remote host closed the connection) |
| 23:34:12 | * | infinity0 joined #nim |
| 23:37:14 | * | X67r quit (Quit: leaving) |
| 23:49:10 | * | vendethiel quit (Ping timeout: 240 seconds) |
| 23:53:56 | * | Dark_Fox quit (Quit: Leaving) |