| 00:01:56 | krux02 | yes efficient compiler caches would be nice, for example when I compile my project, it would be nice if the compiler could save the state when all the imports are processed, and just come back to exactly this state when recompilation is invoked |
| 00:02:04 | krux02 | I know it's a bit more complicated than that |
| 00:02:15 | krux02 | and I am quite happy with the compilation speed |
| 00:02:25 | * | derlafff quit (Remote host closed the connection) |
| 00:03:06 | * | derlafff joined #nim |
| 00:03:27 | Araq | that's what --symbolFiles do, but I need to rewrite this part, more or less |
| 00:03:42 | krux02 | I just imagine this to be implemented with fork, and the fork doing nothing and waiting for a command line invocation of nimc |
| 00:04:20 | Araq | hmm? |
| 00:04:57 | krux02 | don't worry too much, that's pretty dirty implementation and would only work on linux as far as I know |
| 00:05:17 | krux02 | but with fork you can get on linux an almost identical copy of a process |
| 00:05:33 | Araq | I know what fork does. |
| 00:06:31 | krux02 | It really is not that important, because that idea just came into my head |
| 00:06:44 | krux02 | and it is not very well thought through |
| 00:07:32 | krux02 | and it does not cover proper handling of files that actually change when the standard library has changes |
| 00:15:41 | * | Trustable quit (Remote host closed the connection) |
| 00:17:05 | * | devted quit (Quit: Sleeping.) |
| 00:18:54 | * | Gilga quit (Ping timeout: 272 seconds) |
| 00:27:08 | * | def-pri-pub quit (Quit: leaving) |
| 00:27:38 | * | def-pri-pub joined #nim |
| 00:34:15 | def-pri-pub | Qustion about the nimgrep tool: is it only a simple text search-and-replace tool? Or can it refactor specific things in a Nim source file/ project directory? |
| 00:35:37 | dom96 | only search and replac |
| 00:35:38 | dom96 | e |
| 00:35:52 | def-pri-pub | ah. Are there any more complex refactoring tools out there? |
| 00:44:49 | krux02 | def-pri-pub: there is nothing complex thar transforms source code representations |
| 00:56:44 | * | jinshil joined #nim |
| 01:00:38 | * | couven92 quit (Quit: Client disconnecting) |
| 01:01:18 | krux02 | Araq: the fact that I cannot do this on ast nodes is bugging me a lot lately: let x = arg[0][0][2][2]; for n in nodeIter(): x.add(n) |
| 01:04:11 | krux02 | I am very close to write my own reference to a NimNode that is a seq[int] internally just to work around this bug |
| 01:05:01 | * | yglukhov joined #nim |
| 01:09:38 | * | yglukhov quit (Ping timeout: 248 seconds) |
| 01:14:25 | * | bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…) |
| 01:16:31 | krux02 | Araq: and a fast `tail` implementation on `openarray` would also be great |
| 01:18:14 | * | bjz joined #nim |
| 01:18:26 | Araq | krux02: you could also fix the VM instea |
| 01:18:46 | dom96 | http://xkcd.com/1782/ |
| 01:46:08 | * | nsf quit (Quit: WeeChat 1.6) |
| 01:52:44 | * | yglukhov joined #nim |
| 01:56:53 | * | yglukhov quit (Ping timeout: 240 seconds) |
| 02:04:34 | * | vlad1777d quit (Remote host closed the connection) |
| 02:29:37 | * | brson quit (Quit: leaving) |
| 02:35:48 | * | pregressive quit (Remote host closed the connection) |
| 02:39:49 | * | chemist69 quit (Disconnected by services) |
| 02:39:54 | * | chemist69_ joined #nim |
| 02:52:02 | * | krux02 quit (Quit: Leaving) |
| 02:58:19 | * | jinshil quit (Quit: client terminated!) |
| 03:08:20 | * | pie__ joined #nim |
| 03:48:54 | * | pie__ quit (Changing host) |
| 03:48:54 | * | pie__ joined #nim |
| 04:31:59 | * | pregressive joined #nim |
| 04:35:00 | * | arnetheduck joined #nim |
| 04:42:17 | * | Snircle quit (Quit: Textual IRC Client: www.textualapp.com) |
| 04:43:53 | * | chemist69_ quit (Ping timeout: 240 seconds) |
| 04:48:31 | * | chemist69 joined #nim |
| 04:55:49 | * | ioreth joined #nim |
| 04:56:27 | * | ioreth left #nim (#nim) |
| 05:18:42 | def-pri-pub | Another question: When I have a complex type/object, are they passed by reference or value in a proc's argument? |
| 05:20:18 | def-pri-pub | e.g. the `Matrix3d` object in `basic3d`, are all of it's fields copied when it's being called onto a `proc`? |
| 05:24:58 | * | arnetheduck quit (Ping timeout: 246 seconds) |
| 05:27:54 | * | bjz_ joined #nim |
| 05:28:58 | * | bjz quit (Ping timeout: 258 seconds) |
| 05:39:11 | * | space-wi_ quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 06:23:37 | * | space-wizard joined #nim |
| 06:37:38 | * | bjz_ quit (Ping timeout: 248 seconds) |
| 06:37:46 | * | bjz joined #nim |
| 06:48:59 | * | def-pri-pub quit (Quit: leaving) |
| 07:19:31 | * | GustavoLapasta joined #nim |
| 08:10:57 | * | space-wizard quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 08:26:35 | * | ARCADIVS joined #nim |
| 08:47:31 | * | chemist69 quit (Ping timeout: 240 seconds) |
| 08:48:22 | * | chemist69 joined #nim |
| 08:50:48 | * | jinshil joined #nim |
| 09:42:08 | * | byte512 quit (Ping timeout: 255 seconds) |
| 09:52:45 | * | Trustable joined #nim |
| 09:57:43 | * | Trustable_2 joined #nim |
| 09:57:49 | * | Trustable_2 quit (Remote host closed the connection) |
| 10:07:35 | * | Matthias247 joined #nim |
| 10:09:32 | * | nsf joined #nim |
| 10:32:36 | ftsf_ | hmm how can i get a random element of a set? |
| 10:50:19 | * | cheatfate quit (Read error: Connection reset by peer) |
| 10:50:43 | ftsf_ | hmm can't filter a queue =( |
| 10:54:46 | * | Gilga joined #nim |
| 10:57:54 | * | couven92 joined #nim |
| 11:06:06 | Araq | there is a PR for that iirc |
| 11:19:45 | ftsf_ | cool, cool, worked around it for now |
| 11:19:56 | ftsf_ | nearly finished another game \o/ |
| 11:21:33 | ftsf_ | http://static.impbox.net/boxen2/boxen.mp4 |
| 11:47:17 | * | cheatfate joined #nim |
| 11:50:14 | Gilga | Hello, I'm new to nim and still am learning. I just wanted to ask Macros. How powerful are they? As far as I'm understanding with Macros you can expand the language. Can you for instance add pattern matching to the language? What about list comprehension? |
| 11:50:32 | Gilga | I haven't gotten too far into Nim, so this question is out of pure curiousity |
| 11:50:39 | ftsf_ | Gilga, list comprehension has been done |
| 11:51:08 | ftsf_ | tbh, i i haven't used macros at all and i'm still very productive in nim. |
| 11:51:20 | ftsf_ | (haven't created my own macros that is) |
| 11:51:55 | * | gokr joined #nim |
| 11:51:57 | Gilga | Nice, then it's quite powerful. I'd just like to know on an inntuitive level how powerful macros can get. |
| 11:52:58 | def- | Gilga: https://github.com/nim-lang/Nim/blob/master/lib/pure/future.nim#L128 |
| 11:53:04 | Gilga | Don't get me wrong, I'm not saying that you have to use Macros. Nim seems like a very expressive language without them. I'd just like to know more about them as I've never really gotten involved into them. |
| 11:53:10 | def- | Gilga: Pattern matching: http://www.drdobbs.com/open-source/nimrod-a-new-systems-programming-languag/240165321 |
| 11:53:52 | Gilga | hanks for the links, def. |
| 11:54:03 | Gilga | and for the answers ftsf |
| 11:55:05 | ftsf_ | i should probably start trying to think with macros |
| 11:56:32 | def- | ftsf_: I'd say always try to solve problems without macros first. they are too powerful for many problems |
| 11:57:07 | ftsf_ | hmm good point |
| 12:04:36 | * | Snircle joined #nim |
| 12:13:13 | * | gangstacat quit (Ping timeout: 240 seconds) |
| 12:18:38 | * | gangstacat joined #nim |
| 12:33:15 | * | gangstacat_ joined #nim |
| 12:34:11 | * | gangstacat_ quit (Client Quit) |
| 12:34:33 | * | gangstacat quit (Ping timeout: 240 seconds) |
| 12:37:30 | * | gangstacat joined #nim |
| 12:38:28 | * | gangstacat quit (Client Quit) |
| 12:40:20 | * | gangstacat joined #nim |
| 12:45:00 | * | mat4 joined #nim |
| 12:45:25 | mat4 | hello |
| 12:46:29 | ftsf_ | hi mat4 |
| 12:50:07 | mat4 | does there exist an option to disable garbage collection with the current Nim version ? |
| 12:50:46 | dom96 | --gc:none |
| 12:50:47 | ftsf_ | yep, --gc:none |
| 12:50:58 | ftsf_ | =( too slow |
| 12:54:14 | mat4 | no I mean temporary for specific source sequences |
| 12:55:08 | dom96 | http://nim-lang.org/docs/system.html#GC_disable, |
| 12:55:27 | dom96 | ftsf_: That looks good. Inspired by one of your previous games? It looks familiar |
| 12:56:01 | ftsf_ | dom96, yeah, based on an old android game of mine made in lua, the original code is a horrid mess and i can't bare to look at it again |
| 12:59:57 | ftsf_ | trying out nim-chipmunk |
| 13:03:05 | * | couven92 quit (Quit: Client disconnecting) |
| 13:12:19 | * | byte512 joined #nim |
| 13:21:56 | mat4 | dom96: sorry, I would require to temporary disable GC completly, not only the mark-and-sweep collector |
| 13:24:12 | * | enthus1ast joined #nim |
| 13:29:04 | * | mat4 quit (Quit: Verlassend) |
| 13:30:44 | * | beaumonta is now known as abeaumont |
| 13:31:02 | * | enthus1ast quit (Ping timeout: 245 seconds) |
| 13:31:30 | * | gokr quit (Read error: Connection reset by peer) |
| 13:35:53 | * | confundus joined #nim |
| 13:42:19 | * | jinshil quit (Remote host closed the connection) |
| 13:50:47 | * | confundus quit (Quit: confundus) |
| 14:06:15 | * | fredrik92 joined #nim |
| 14:13:37 | * | bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…) |
| 14:20:22 | * | fredrik92 quit (Quit: fredrik92) |
| 14:27:40 | * | subsetpark joined #nim |
| 14:31:02 | * | byte512 quit (Ping timeout: 245 seconds) |
| 14:31:48 | * | Trustable quit (Ping timeout: 248 seconds) |
| 14:32:18 | * | Trustable joined #nim |
| 14:47:48 | * | nsf quit (Quit: WeeChat 1.6) |
| 14:49:15 | * | dddddd joined #nim |
| 14:51:44 | * | confundus joined #nim |
| 15:15:25 | * | vlad1777d joined #nim |
| 15:19:37 | * | Parashurama joined #nim |
| 15:21:13 | * | Snircle quit (Quit: Textual IRC Client: www.textualapp.com) |
| 15:24:33 | Parashurama | Is there any way a make a template return value discardable? |
| 15:25:00 | Parashurama | I use this for now : |
| 15:25:00 | Parashurama | proc makeDiscardable[T](a: T): T {.discardable, inline.} = a |
| 15:25:00 | Parashurama | template foo(): bool = |
| 15:25:18 | Parashurama | makeDiscardable(true) |
| 15:35:43 | * | gangstacat quit (Quit: Ĝis) |
| 15:35:54 | * | yglukhov joined #nim |
| 15:40:28 | * | gangstacat joined #nim |
| 15:43:59 | * | confundus quit (Quit: confundus) |
| 15:44:46 | * | NimBot joined #nim |
| 15:45:09 | * | NimBot joined #nim |
| 15:45:34 | * | NimBot joined #nim |
| 15:46:04 | * | NimBot joined #nim |
| 15:46:21 | * | NimBot joined #nim |
| 15:49:37 | * | confundus joined #nim |
| 15:53:03 | * | Parashurama quit (Quit: Leaving.) |
| 15:53:21 | * | Parashurama joined #nim |
| 16:15:42 | * | yglukhov quit (Remote host closed the connection) |
| 16:48:14 | FromGitter | <Varriount> Gilga: Nim's macros are essentially just procedures that construct an abstract syntax tree, so they're pretty powerful. |
| 16:49:32 | FromGitter | <Varriount> def-pri-pub: Internally, unless a type/argument has been annotated with the bycopy pragma, all arguments are passed via pointer |
| 16:50:30 | FromGitter | <Varriount> The semantics must still slow for the possibility of copying though, especially for object types |
| 16:51:20 | * | gokr joined #nim |
| 16:51:39 | FromGitter | <Varriount> That being said, you should also be aware of assignment semantics: http://forum.nim-lang.org/t/2665#16487 |
| 16:52:01 | * | chemist69 quit (Ping timeout: 258 seconds) |
| 16:52:23 | FromGitter | <Varriount> ^ Good explanation for anyone new to lower-level programming |
| 16:52:36 | * | chemist69 joined #nim |
| 16:54:44 | * | space-wizard joined #nim |
| 16:55:14 | * | gokr quit (Read error: No route to host) |
| 16:56:34 | FromGitter | <Varriount> mat4: Why do you need to disable the garbage collector? |
| 17:03:06 | * | pie__ quit (Ping timeout: 240 seconds) |
| 17:19:47 | * | nsf joined #nim |
| 17:28:08 | * | confundus quit (Quit: confundus) |
| 17:32:28 | * | dddddd quit (Remote host closed the connection) |
| 17:32:42 | * | def-pri-pub joined #nim |
| 17:33:37 | def-pri-pub | I asked this in the IRC last night but I never got an answer for it, so here it is again: |
| 17:33:40 | def-pri-pub | 05:18:42 def-pri-pub Another question: When I have a complex type/object, are they passed by reference or value in a proc's argument? |
| 17:33:43 | def-pri-pub | 05:20:18 def-pri-pub e.g. the `Matrix3d` object in `basic3d`, are all of it's fields copied when it's being called onto a `proc`? |
| 17:33:46 | def-pri-pub | whopps. |
| 17:33:52 | def-pri-pub | Another question: When I have a complex type/object, are they passed by reference or value in a proc's argument? |
| 17:33:59 | def-pri-pub | e.g. the `Matrix3d` object in `basic3d`, are all of it's fields copied when it's being called onto a `proc`? |
| 17:53:52 | * | Gilga quit (Ping timeout: 240 seconds) |
| 17:58:16 | FromGitter | <Varriount> def-pri-pub: As I said, they aren't copied |
| 17:58:36 | FromGitter | <Varriount> Not unless you mark the type with the bycopy pragma |
| 17:59:08 | FromGitter | <Varriount> If you look at the C source code, you'll notice that arguments are always pointers |
| 18:00:19 | FromGitter | <Varriount> The pointer either pointing to the stack (in the case of object and var parameters) or to the heap. |
| 18:00:49 | * | sarlalian quit (Ping timeout: 258 seconds) |
| 18:03:07 | * | sarlalian joined #nim |
| 18:31:13 | * | Jesin quit (Quit: Leaving) |
| 18:31:53 | def-pri-pub | so what about in the case of a paramter that is a `ref object`? When it enters the scope of the function, will the reference count increase by one, then when it exits, decrease by one? |
| 18:36:14 | * | arnetheduck joined #nim |
| 18:37:10 | * | Jesin joined #nim |
| 18:46:24 | FromGitter | <Varriount> Semantically, yes |
| 18:47:16 | * | arnetheduck quit (Ping timeout: 248 seconds) |
| 18:47:41 | FromGitter | <Varriount> Def-pri-pub: although whether the reference is actually incremented depends on the garbage collector |
| 18:48:09 | def-pri-pub | Okay, thanks. I'm a little confused on the internals of Nim. Does this only apply to the C like targets? Or JS as well? |
| 18:49:35 | FromGitter | <Varriount> It depends on the backend |
| 18:50:29 | def-pri-pub | gotcha |
| 18:50:34 | * | Gilga joined #nim |
| 18:50:46 | FromGitter | <Varriount> def-pri-pub: If you're worried about reference counting being costly with regards to performance, it's not. Especially not with the default garbage collector |
| 18:51:08 | FromGitter | <Varriount> I mean, refcounting is what smartpointers in C++ use |
| 18:51:37 | Araq | " Internally, unless a type/argument has been annotated with the bycopy pragma, all arguments are passed via pointer" no, the compiler picks the most efficient parameter passing mode |
| 18:51:49 | Araq | and it has been this way since day 1. |
| 18:51:59 | FromGitter | <Varriount> ^ I stand corrected |
| 18:53:07 | FromGitter | <Varriount> Araq: Would you be happier if I said that *most* arguments are passed by pointer, internally? |
| 18:53:08 | def-pri-pub | So compiler knows best? |
| 18:53:54 | FromGitter | <Varriount> def-pri-pub: Yes. And if it offers to do your taxes for you, let it. |
| 18:54:35 | Parashurama | Araq: hey, can you look at earlier today in IRC I asked a question about template with discardable return value |
| 18:55:32 | * | hendi joined #nim |
| 18:56:05 | Araq | Parashurama: yeah, that's not possible |
| 18:56:20 | Araq | what's the use case? |
| 18:56:52 | Parashurama | It is an optimization move proc -> template |
| 18:57:20 | Parashurama | also for debugging since instaiationInfo only works in templates |
| 18:58:10 | Araq | why is it discardable? |
| 18:58:19 | Parashurama | but I can manage with a returning a proc like makeDiscardable[T](a: T): T {.discardable, inline.} = a |
| 18:58:39 | FromGitter | <Varriount> Parashurama: Why not use the inline pragma? |
| 18:59:12 | Parashurama | the return value is just a bool, to check if operation was succesful. |
| 18:59:36 | Parashurama | Varriount: see instanciationInfo in system.nim to get line info. |
| 18:59:47 | * | chemist69 quit (Ping timeout: 245 seconds) |
| 19:00:00 | Parashurama | but the return value can be safely ignored. |
| 19:00:09 | * | chemist69 joined #nim |
| 19:02:07 | Parashurama | Araq: btw would "when not (defined(nimvm) or defined(js) or defined(nimdoc)):" be sufficient to guard importc proc with header pragma |
| 19:03:41 | * | Amun_Ra quit (Ping timeout: 258 seconds) |
| 19:05:27 | Araq | Parashurama: maybe, what about the LLVM backend though? |
| 19:10:38 | * | Amun_Ra joined #nim |
| 19:14:04 | * | Gilga quit (Ping timeout: 255 seconds) |
| 19:25:03 | * | yglukhov joined #nim |
| 19:25:10 | * | yglukhov quit (Remote host closed the connection) |
| 19:31:55 | Parashurama | Araq: yes, I'm not sure what is defined by nimllvm. I think emscriptem support header pragma |
| 19:35:15 | * | yglukhov joined #nim |
| 19:39:20 | Parashurama | The goal being at the moment add support for memchr in strutils.find for supported platforms & systems. |
| 19:49:16 | Araq | ah cool |
| 19:53:24 | * | ARCADIVS quit (Quit: ARCADIVS) |
| 19:54:10 | Parashurama | Araq: Where do you think I could add compiler builtins like __builtin_clz, popcount and things like that. |
| 19:55:17 | * | odc_ joined #nim |
| 19:56:20 | Araq | depends on the implementation |
| 19:56:43 | Araq | but we still lack a intrisics.nim module ;-) |
| 19:57:18 | def-pri-pub | Araq: I'm working on some bindings for the HTML5 Canvas tag and drawing on it, for the JS target. When I'm done with it, would you be interested in merging it into the Nim repo? |
| 19:57:21 | def-pri-pub | https://gitlab.com/define-private-public/HTML5-Canvas-Nim |
| 19:57:58 | Araq | no, sorry. that should be a nimble package |
| 19:58:09 | def-pri-pub | gothca (it already is) |
| 19:58:13 | Parashurama | Araq: intrinsics.nim seem like a good idea, and also eventually a simd module with SSE and NEON intrinsics |
| 19:58:20 | Araq | we could make it an official nimble package eventually. |
| 19:59:05 | Parashurama | Araq: you mean for the intrinsics module? |
| 19:59:09 | Araq | so ... do we want a risky merge of the concepts branch in 0.16? |
| 19:59:23 | Araq | Parashurama: no, I meant the canvas |
| 19:59:51 | Araq | release 0.16 only got the code generation improvements |
| 19:59:55 | Parashurama | Okay, for concept branch what the difference with curent impl |
| 19:59:57 | def-pri-pub | Sounds cool to me. Are there any plans to make the OS module work for the JS target? I know that browser != operating system, but there are something things that depend on it that I want for the js target. (e.g. HttpClient) |
| 20:00:33 | Araq | def-pri-pub: that's pretty hopeless. you could try to make ospaths work for JS |
| 20:00:58 | Araq | but stuff like HttpClient needs to be reimplemented for JS afaict |
| 20:01:19 | * | dddddd joined #nim |
| 20:01:57 | def-pri-pub | I'll look into it eventually. Yeah, I'm aware that HTTPClient does need to use the JS's XMLHttpClient. I was thinking about working on that and submitting a patch, but then I saw HttpClient is dependant upon the os module. |
| 20:02:28 | Araq | that doesn't mean much. os.nim creeps into everything |
| 20:02:57 | def-pri-pub | Also, have you gotten a change to look at this PR yet? I know it's a little trivial, but I think it's important to mention in the docs: https://github.com/nim-lang/Nim/pull/5181 |
| 20:03:01 | Araq | I doubt the http client uses e.g. setting of file permissions |
| 20:03:45 | Araq | def-pri-pub: meh, I don't like the wording |
| 20:04:12 | def-pri-pub | Is there some better wording I could use? |
| 20:05:13 | def-pri-pub | Eh, I'll think of something and update the PR. |
| 20:05:26 | * | def-pri-pub is now known as def-pri-pub|afk |
| 20:05:29 | def-pri-pub|afk | (lunch time) |
| 20:37:57 | * | yglukhov quit (Remote host closed the connection) |
| 20:38:14 | * | yglukhov joined #nim |
| 20:44:14 | * | def-pri-pub|afk is now known as def-pri-pub |
| 20:45:39 | Araq | def-pri-pub: too late, I merged it manually |
| 20:49:29 | def-pri-pub | damn. |
| 21:03:51 | * | Sembei joined #nim |
| 21:33:01 | def-pri-pub | For the JS backend, what Nim type turns into a JavaScript array? |
| 21:34:30 | def-pri-pub | Would It be seq since the JS array is mutable? |
| 21:35:50 | Araq | seqs and arrays are both mapped to JS arrays |
| 21:40:00 | def-pri-pub | okay. |
| 21:59:29 | * | GustavoLapasta quit (Quit: Leaving) |
| 22:05:42 | Parashurama | Araq: hey, is there a way to check if nim version has `compile_option` Option( |
| 22:06:13 | Parashurama | something like: when compiles(compileOtion("option")): |
| 22:06:34 | Parashurama | but the previous expression crashes |
| 22:06:50 | Araq | I thnk you can only check for this via a version check |
| 22:07:10 | Parashurama | Thats what I though, oh well, thaks anyway. |
| 22:07:46 | Parashurama | btw portable compiler intrinsics are harder that I thought |
| 22:07:50 | Parashurama | :) |
| 22:09:59 | Parashurama | I am making some progress, but sometime even compiler intrinsics have awful performance . |
| 22:10:48 | * | MyMind joined #nim |
| 22:12:43 | * | Sembei quit (Ping timeout: 255 seconds) |
| 22:42:49 | Araq | Parashurama: well if it were easy, somebody would have done it already :-) |
| 22:43:21 | Parashurama | At least we already have Hamming weight function in math. thats something |
| 22:44:56 | Parashurama | I think the module will be focusing on bit instrinsics, most others are kinda random. |
| 22:46:12 | * | nsf quit (Quit: WeeChat 1.6) |
| 22:50:03 | * | Parashurama left #nim (#nim) |
| 22:51:51 | * | gangstacat quit (Quit: Ĝis) |
| 22:55:27 | * | Salewski joined #nim |
| 22:56:24 | Salewski | Parashurama, yes intrinsics/builtin support would be nice. |
| 22:57:47 | Salewski | I recently tested popcnt mentioned in a forum post, works nice with gcc, but portability may be difficult. |
| 22:58:29 | Salewski | For popcount there is also a pull request. |
| 22:58:35 | * | devted joined #nim |
| 22:59:36 | Salewski | It would be nice if it would be possible to code stuff like xxhash or clhash in pure Nim. |
| 23:00:35 | Salewski | xxHash is already available for many other languages than C. |
| 23:01:42 | Salewski | We may need 128 bit data types too, I think I have seen that in the clHash C code. |
| 23:03:15 | Salewski | http://programmingzen.com/so-you-want-your-programming-language-to-be-popular/ |
| 23:03:21 | Salewski | Bye. |
| 23:03:26 | * | Salewski left #nim (#nim) |
| 23:09:48 | * | pregressive quit (Remote host closed the connection) |
| 23:12:07 | * | yglukhov quit (Remote host closed the connection) |
| 23:19:00 | * | Matthias247 quit (Read error: Connection reset by peer) |
| 23:35:28 | cheatfate | Araq, it looks like macos before Leopard don't have api to set thread affinity |
| 23:35:44 | Araq | I noticed. |
| 23:35:50 | Araq | I disabled it for osx |
| 23:38:11 | cheatfate | could we check for macos version? |
| 23:39:48 | Araq | yeah but that would be a runtime check |
| 23:40:01 | cheatfate | i think this can be used https://github.com/witheve/cluster/blob/464e3d3b64c629639e82117e34fbc2f6dc8bba34/csrc/unix/darwin/thread.c |
| 23:40:52 | Araq | bloat, bloat, bloat everywhere :P |
| 23:41:06 | Araq | who cares about thread affinity on shitty osx |
| 23:41:37 | cheatfate | dom96 just created issue in my `wanted` about thread affinity :) |
| 23:41:49 | Araq | do you use pinToCpu? |
| 23:42:18 | cheatfate | yep |
| 23:42:33 | Araq | so close the bug, stdlib was fixed |
| 23:42:43 | cheatfate | and everything is works? |
| 23:42:56 | Araq | it's a nop for osx now |
| 23:43:20 | Araq | so your code compiles and works but it might be slower |
| 23:44:26 | cheatfate | dom96, does it works for you? |
| 23:48:22 | * | vlad1777d quit (Quit: Leaving) |