00:04:15 | * | lll joined #nim |
00:16:03 | * | vbtt_ quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) |
00:16:19 | * | kumool quit (Ping timeout: 255 seconds) |
00:20:34 | * | pregressive joined #nim |
00:24:27 | * | Sembei joined #nim |
00:24:54 | * | taesoo quit (Ping timeout: 256 seconds) |
00:25:12 | * | vaedd joined #nim |
00:25:49 | reactormonk | Araq, btw, kick the JS backend and go for emscripten? |
00:26:11 | * | boydgreenfield quit (Quit: boydgreenfield) |
00:28:33 | * | brson quit (Ping timeout: 258 seconds) |
00:29:18 | fowl | just make it emit typescript or something |
00:30:53 | reactormonk | hmm |
00:32:21 | * | pregressive quit (Remote host closed the connection) |
00:34:26 | * | MyMind joined #nim |
00:36:35 | * | Sembei quit (Ping timeout: 244 seconds) |
00:38:22 | * | wedowmaker joined #nim |
00:48:06 | * | havenwood joined #nim |
00:49:23 | * | TEttinger quit (Ping timeout: 246 seconds) |
00:50:18 | * | gsingh93 quit (Ping timeout: 265 seconds) |
01:04:15 | federico3 | is there a mailing list and/or is the forum delivered as a mailing list? |
01:08:24 | federico3 | is there a mailing list and/or is the forum delivered as a mailing list? |
01:14:04 | * | vendethiel joined #nim |
01:19:38 | * | adu joined #nim |
01:25:14 | * | wb quit (Ping timeout: 245 seconds) |
01:34:41 | flaviu | federico3: there is a mailing list, the forum is not delivered as a mailing list. |
01:35:10 | flaviu | There were plans to do make the forum work like a mailing list, but I guess no one got around to it. |
01:35:34 | flaviu | The volume on the mailing list is extremely low, I don't think I've gotten anything this month. |
01:37:31 | flaviu | http://www.freelists.org/list/nim-dev |
01:38:08 | flaviu | Varriount: Are you an admin on the forum? http://forum.nim-lang.org/t/1254 is spam. |
01:54:34 | * | adu quit (Ping timeout: 265 seconds) |
01:55:10 | * | ddl_smurf quit (Quit: ddl_smurf) |
02:22:31 | * | havenwood quit () |
02:22:54 | * | havenwood joined #nim |
02:23:15 | * | banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
02:29:11 | * | pipeep joined #nim |
02:33:48 | * | vendethiel quit (Ping timeout: 265 seconds) |
02:33:54 | * | tripleplus joined #nim |
02:34:17 | tripleplus | how do i echo() without a trailing newline? |
02:34:53 | tripleplus | i've tried stdout.write, but the behavior there was not exactly what i was looking for. |
02:35:19 | tripleplus | i'm on osx 10.10, fwiw |
02:35:28 | * | havenwood left #nim (#nim) |
02:35:47 | tripleplus | nim v0.11.2 |
02:39:44 | reactormonk | nope, echo includes the trailing newline. that's kinda the definition. |
02:50:39 | * | darkf joined #nim |
02:57:16 | tripleplus | reactormonk: fair enough. how do i output text without an implied newline? |
02:58:02 | tripleplus | btw, where is that definition (i assume you mean nim's since that's definitely not definitive behavior in all contexts for, say, bash)? |
02:58:11 | fowl | Write a function that calls stdout.write then stdout.flush |
02:58:17 | tripleplus | i looked wround for echo and write documentation, but wasn't able to find it. |
02:58:33 | tripleplus | i'm sure i'm being braindead. |
02:58:41 | tripleplus | fowl: thanks, will try that now. |
03:00:19 | fowl | tripleplus: http://nim-lang.org/docs/theindex.html covers the stdlib and manual |
03:00:44 | Varriount | flaviu: Unfortunately, no, I am not an admin on the forum. |
03:02:40 | * | banister joined #nim |
03:04:29 | Varriount | fowl: You need, like, a 'Macro Master' award. |
03:09:03 | * | saml_ quit (Ping timeout: 250 seconds) |
03:09:54 | tripleplus | fowl: i'm having a difficult time getting stdout.flush to work (Error: undeclared identifier: 'flush') |
03:10:12 | fowl | Varriount: macros are definitely my favorite feature |
03:10:29 | fowl | tripleplus: sorry its called flushFile maybe, let me look |
03:10:54 | fowl | Yea that's it |
03:11:06 | tripleplus | ah, k, ty sir |
03:11:20 | tripleplus | or ma'am, sorry, showing my bias |
03:14:37 | tripleplus | fowl: yep, sure enough, that did it |
03:31:57 | * | banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
03:32:29 | * | kumul joined #nim |
03:35:23 | * | kumool joined #nim |
03:39:10 | * | kumul quit (Ping timeout: 252 seconds) |
03:52:06 | * | tripleplus quit (Ping timeout: 265 seconds) |
03:52:36 | * | taesoo joined #nim |
03:54:11 | * | pregressive joined #nim |
04:19:14 | * | gsingh93 joined #nim |
04:38:29 | * | intra quit (Ping timeout: 258 seconds) |
04:48:29 | * | tripleplus joined #nim |
04:53:04 | * | tripleplus quit (Ping timeout: 255 seconds) |
04:56:23 | * | adu joined #nim |
05:00:24 | * | Zerthis joined #nim |
05:02:48 | * | Zerthis_ALT joined #nim |
05:06:03 | * | Zerthis quit (Ping timeout: 250 seconds) |
05:10:44 | * | taesoo quit (Ping timeout: 264 seconds) |
05:18:21 | * | BitPuffin|osx quit (Ping timeout: 258 seconds) |
05:34:32 | * | strcmp2 joined #nim |
05:36:54 | * | strcmp1 quit (Ping timeout: 245 seconds) |
05:46:00 | * | gsingh93 quit (Ping timeout: 272 seconds) |
05:47:13 | * | pregressive quit (Remote host closed the connection) |
05:49:20 | * | tripleplus joined #nim |
05:53:49 | * | tripleplus quit (Ping timeout: 255 seconds) |
05:55:08 | * | intra joined #nim |
06:25:44 | * | nande quit (Remote host closed the connection) |
06:50:11 | * | tripleplus joined #nim |
06:55:21 | * | tripleplus quit (Ping timeout: 276 seconds) |
07:00:28 | * | strcmp2 quit (Ping timeout: 252 seconds) |
07:16:00 | * | adu quit (Ping timeout: 265 seconds) |
07:35:58 | * | Zerthis_ALT left #nim ("Textual IRC Client: www.textualapp.com") |
07:40:21 | * | darkf_ joined #nim |
07:43:30 | * | darkf quit (Ping timeout: 256 seconds) |
07:48:48 | * | filcuc joined #nim |
07:50:57 | * | tripleplus joined #nim |
07:53:09 | * | synthmeat quit (Quit: WeeChat 1.3-dev) |
07:54:53 | * | synthmeat joined #nim |
07:55:21 | * | tripleplus quit (Ping timeout: 240 seconds) |
07:58:19 | * | kumool quit (Quit: Leaving) |
07:59:48 | * | ddl_smurf joined #nim |
08:05:31 | * | IronY is now known as sadasdas |
08:05:41 | * | sadasdas is now known as IronY |
08:15:45 | * | Senketsu quit (Quit: Leaving) |
08:26:52 | * | IronY is now known as test |
08:27:10 | * | test is now known as IronY |
08:32:54 | * | boopisaway is now known as boop |
08:45:54 | * | BlaXpirit joined #nim |
08:47:57 | * | pregressive joined #nim |
08:51:40 | * | tripleplus joined #nim |
08:51:57 | * | filcuc quit (Quit: Konversation terminated!) |
08:52:09 | * | pregressive quit (Ping timeout: 240 seconds) |
08:56:38 | * | tripleplus quit (Ping timeout: 272 seconds) |
09:01:16 | * | strcmp2 joined #nim |
09:21:20 | ggVGc | what nim type maps to const char*? |
09:21:52 | ggVGc | I need to implement a function taking const char* as an argument in Nim and use it from C |
09:22:52 | * | wb joined #nim |
09:23:58 | dom96 | ggVGc: cstring |
09:27:47 | ggVGc | dom96: thats char* though. how do i make it immutable |
09:29:44 | dom96 | it'll be immutable in Nim by default |
09:29:50 | * | xcombelle joined #nim |
09:33:37 | Araq | ggVGc: doesn't matter for ABI compatibility |
09:34:12 | ggVGc | Araq: yeah, but it makes other code error. Do I have to wrap the nim function? |
09:35:22 | * | vendethiel joined #nim |
09:37:41 | Araq | I have not enough context to answer that question |
09:37:42 | * | Matthias247 joined #nim |
09:39:27 | ggVGc | Araq: all I wonder is if there's any datatype in nim that maps to const char*. otherwise I'll just wrap the generated function from nim in one that takes const char* so it doesn't break with already existing code |
09:39:47 | * | boop is now known as boopisaway |
09:41:37 | Araq | no idea. "const char* " is 'cstring', "char*" is 'cstring' too. Nim has no notion of C's const |
09:52:23 | * | tripleplus joined #nim |
09:56:49 | * | tripleplus quit (Ping timeout: 255 seconds) |
10:04:48 | * | IronY is now known as irony |
10:04:54 | * | irony is now known as IronY |
10:06:34 | * | Ven joined #nim |
10:11:21 | * | jbomo quit (Ping timeout: 240 seconds) |
10:33:11 | * | pkoretic joined #nim |
10:39:35 | * | Gonzih joined #nim |
10:53:09 | * | tripleplus joined #nim |
10:57:20 | * | tripleplus quit (Ping timeout: 246 seconds) |
11:02:02 | * | juturna joined #nim |
11:02:42 | * | darkf_ is now known as darkf |
11:07:27 | * | filcuc joined #nim |
11:15:29 | * | Matthias247 quit (Quit: Matthias247) |
11:22:56 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
11:28:46 | * | Trustable joined #nim |
11:30:34 | * | banister joined #nim |
11:30:38 | * | banister quit (Max SendQ exceeded) |
11:41:53 | * | banister joined #nim |
11:47:15 | * | Kingsquee quit (Quit: Konversation terminated!) |
11:47:35 | * | dddddd joined #nim |
11:52:44 | * | Sembei joined #nim |
11:53:58 | * | tripleplus joined #nim |
11:56:53 | * | JinShil joined #nim |
11:58:09 | * | tripleplus quit (Ping timeout: 245 seconds) |
12:02:49 | * | Ven joined #nim |
12:09:12 | * | xcombelle quit (Ping timeout: 252 seconds) |
12:19:28 | * | johnsoft quit (Ping timeout: 252 seconds) |
12:20:31 | * | johnsoft joined #nim |
12:24:56 | * | gokr joined #nim |
12:38:35 | * | milosn quit (Quit: leaving) |
12:54:42 | * | tripleplus joined #nim |
12:59:20 | * | tripleplus quit (Ping timeout: 276 seconds) |
13:01:21 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
13:07:08 | * | Jesin quit (Read error: Connection reset by peer) |
13:17:21 | * | syncn quit (Ping timeout: 240 seconds) |
13:19:25 | * | syncn joined #nim |
13:21:08 | * | taesoo joined #nim |
13:24:11 | * | kumul joined #nim |
13:27:51 | * | Ven joined #nim |
13:30:08 | * | Jesin joined #nim |
13:35:04 | * | boopisaway is now known as boop |
13:43:26 | * | TEttinger joined #nim |
13:44:26 | * | BlaXpirit-UA joined #nim |
13:47:27 | * | BlaXpirit quit (Ping timeout: 256 seconds) |
13:48:26 | * | JinShil quit (Quit: Konversation terminated!) |
14:02:10 | * | OnO joined #nim |
14:02:27 | * | milosn joined #nim |
14:09:37 | * | flaviu quit (Remote host closed the connection) |
14:11:44 | * | flaviu joined #nim |
14:32:41 | * | OnO quit (Quit: My iMac has gone to sleep. ZZZzzz…) |
14:43:50 | * | kumul quit (Ping timeout: 272 seconds) |
14:53:21 | * | stefantalpalaru joined #nim |
14:55:59 | stefantalpalaru | Hi. What's the simplest way of getting an identifier from a string in regular code? i.e. go from "foo" to foo where foo is a variable. Do I have to write a macro or there's something similar already? |
14:56:23 | * | tripleplus joined #nim |
14:56:46 | * | kumul joined #nim |
14:57:42 | * | kumool joined #nim |
14:58:21 | dom96 | stefantalpalaru: Macro is the way to go I think |
14:59:48 | * | gsingh93 joined #nim |
15:00:39 | * | tripleplus quit (Ping timeout: 245 seconds) |
15:01:55 | stefantalpalaru | OK, thanks |
15:02:11 | * | kumul quit (Ping timeout: 276 seconds) |
15:03:20 | * | darkf quit (Quit: Leaving) |
15:05:52 | * | jefus quit (Ping timeout: 272 seconds) |
15:17:07 | * | johnsoft quit (Quit: Leaving) |
15:22:23 | * | superfunc joined #nim |
15:22:24 | * | Siecje joined #nim |
15:25:20 | * | gsingh93 quit (Ping timeout: 256 seconds) |
15:31:38 | * | stefantalpalaru quit (Quit: Leaving) |
15:33:10 | * | banister quit (Read error: Connection reset by peer) |
15:35:08 | * | banister joined #nim |
15:48:18 | * | gsingh93 joined #nim |
15:50:12 | * | johnsoft joined #nim |
15:52:59 | * | pregressive joined #nim |
15:55:22 | * | sepisoad joined #nim |
15:56:59 | * | tripleplus joined #nim |
15:57:32 | * | boop is now known as boopisaway |
16:01:45 | * | tripleplus quit (Ping timeout: 256 seconds) |
16:09:14 | * | BlaXpirit-UA quit (Quit: Quit Konversation) |
16:11:19 | * | BlaXpirit joined #nim |
16:12:47 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
16:13:39 | * | filcuc quit (Ping timeout: 256 seconds) |
16:14:06 | * | sepisoad_ joined #nim |
16:17:13 | * | sepisoad quit (Ping timeout: 264 seconds) |
16:22:15 | * | stefantalpalaru joined #nim |
16:23:05 | stefantalpalaru | How can I use typedesc as table values? |
16:24:29 | * | boopisaway is now known as boop |
16:26:39 | * | Matthias247 joined #nim |
16:39:02 | * | boop is now known as boopisaway |
16:40:16 | flaviu | stefantalpalaru: You can make them strings. |
16:40:38 | flaviu | I'm pretty sure that the actual typedesc value is only available at compile time. |
16:40:44 | * | boopisaway is now known as boop |
16:41:21 | stefantalpalaru | OK, and how do I go from string to typedesc afterwards? |
16:41:50 | flaviu | At runtime? You can't. |
16:42:14 | flaviu | You might be able to serialize the interesting parts of the typedesc at compile time and access those at run time. |
16:45:45 | * | kumool quit (Ping timeout: 240 seconds) |
16:45:50 | * | fowl quit (Remote host closed the connection) |
16:53:57 | * | TEttinger quit (Ping timeout: 264 seconds) |
16:57:03 | Varriount | stefantalpalaru: You might be more interested in the typeinfo module |
16:57:18 | * | Siecje1 joined #nim |
16:57:19 | * | notfowl joined #nim |
16:57:24 | * | Siecje quit (Ping timeout: 252 seconds) |
16:57:43 | * | tripleplus joined #nim |
16:57:47 | * | Siecje1 left #nim (#nim) |
16:58:06 | onionhammer | Varriount u figure out the doc comment stuff? |
17:02:21 | * | tripleplus quit (Ping timeout: 265 seconds) |
17:03:23 | flaviu | What's the status of the rodwrite/rodread? |
17:03:48 | * | tripleplus joined #nim |
17:08:53 | * | tripleplus quit (Quit: WeeChat 1.1.1) |
17:13:03 | * | Ven_ joined #nim |
17:13:30 | * | vendethiel quit (Ping timeout: 276 seconds) |
17:13:47 | flaviu | It hasn't been touched since 2013-05-04, so I guess I'll draw my own conclusions. |
17:14:43 | * | vendethiel joined #nim |
17:30:35 | * | jbomo joined #nim |
17:30:40 | * | jbomo quit (Client Quit) |
17:33:52 | flaviu | So rodfiles use their own custom serialization format? :O |
17:36:25 | flaviu | And there's a YAML backend? |
17:38:13 | flaviu | The AST is manually serialized to YAML :/ |
17:41:26 | sepisoad_ | how to immediately get notified when a thread calls send on a channel without having to check the channle using tryRecv function |
17:42:33 | sepisoad_ | like defining a callback function and get it called when the thread is done |
17:43:20 | sepisoad_ | do we have a native event loop in nim? |
17:43:43 | sepisoad_ | so the we can poll for events without wasting cpu time |
17:46:55 | Araq | sepisoad_: yes, we have async io |
17:47:11 | Araq | but theads don't work with it yet :P |
17:47:13 | sepisoad_ | but it's deprecated as stated in the documents |
17:47:22 | sepisoad_ | oh, noo |
17:47:25 | Araq | nah, it's not |
17:47:33 | * | xcombelle joined #nim |
17:48:00 | sepisoad_ | yes, it is |
17:48:00 | sepisoad_ | http://nim-lang.org/docs/asyncio.html |
17:48:01 | Araq | asyncio is deprecated, asyncdispatch is not |
17:51:58 | flaviu | Araq: How does rodwrite/rodread work with the rest of the compiler? I thought it just did rodfiles, but it looks like lots of other things depend on it. |
17:54:33 | lll | in C wrappers i can just push {.raises: [].} i guess? since c has no exceptions? |
17:55:16 | Araq | flaviu: not sure I understand the question but incremental builds require more than just serialization of the AST |
17:55:29 | Araq | lll: iirc you cannot push .raises due to a bug |
17:55:37 | lll | :( |
17:55:49 | Araq | but you can declare your own pragma that's a shortcut for that |
17:56:04 | lll | and push my own pragma? |
17:56:16 | Araq | he, nah |
17:56:18 | lll | :P |
17:56:24 | lll | would have been too easy lol |
17:58:01 | * | Jesin quit (Ping timeout: 264 seconds) |
17:59:17 | flaviu | Araq: https://github.com/Araq/Nim/wiki/Compiler-module-reference says 'Reader of rod files, the symbol file mechanism', but it seems to do much more than that. |
18:00:51 | Araq | it also does module dependency tracking to decide whether a rod file can be used |
18:01:01 | Araq | or whether the module needs to be recompiled |
18:01:20 | * | kumul joined #nim |
18:01:23 | flaviu | Ok, thanks! |
18:01:27 | * | Jesin joined #nim |
18:07:02 | lll | the manual says: Every call to a proc q which has an unknown body (due to a forward declaration or an importc pragma) is assumed to raise system.Exception unless q has an explicit raises list. |
18:07:15 | lll | but it doesnt seem to be true at all, unless i missunderstood |
18:10:47 | bassa | does nim have an equivalent to #define in c? |
18:10:59 | * | Sembei quit (Ping timeout: 265 seconds) |
18:11:40 | lll | i know it has a defined(), and you can define with a compiler param |
18:11:42 | * | Pisuke joined #nim |
18:12:09 | lll | the exact mechanism, i dont know |
18:12:10 | sepisoad_ | is there a code sample to show how to create an event loop with multiple threads running in a program? |
18:12:26 | bassa | thanks I'll check those out, its not necessary for my purposes but would be cleaner |
18:12:44 | lll | i know a lot of library use it |
18:12:49 | lll | like opengl, sdl etc |
18:12:53 | lll | (wrappers) |
18:13:14 | * | BitPuffin|osx joined #nim |
18:13:45 | * | MyMind quit (Ping timeout: 240 seconds) |
18:14:47 | bassa | good idea I checked the opengl wrapper and it uses the same thing I was doing before |
18:15:30 | bassa | for defining type alias like type GLint = int32 |
18:16:02 | * | vendethiel- joined #nim |
18:16:33 | * | vendethiel quit (Ping timeout: 276 seconds) |
18:16:39 | * | Pisuke quit (Max SendQ exceeded) |
18:16:54 | sepisoad_ | please check this code on github gist: https://gist.github.com/sepisoad/f61e4127f6ca5d0b14da |
18:17:02 | * | Ven__ joined #nim |
18:17:21 | sepisoad_ | I want to turn the busy loop into something more elegant |
18:17:52 | * | Pisuke joined #nim |
18:18:24 | sepisoad_ | istead of sleeping and checking the condition in a busy loop I want to wait for an event to be signaled |
18:18:32 | * | Ven_ quit (Ping timeout: 272 seconds) |
18:20:58 | bassa | about to do something very similar sepi, perhaps using await? http://nim-lang.org/docs/threadpool.html |
18:21:35 | bassa | or sync if you need results from all |
18:23:11 | * | Sembei joined #nim |
18:24:52 | * | Pisuke quit (Ping timeout: 255 seconds) |
18:28:11 | * | MyMind joined #nim |
18:29:39 | lll | I think i should PR the opengl wrapper someday, theres some type that i think are wrong |
18:30:28 | lll | some of the gl ptr types are int32 instead of int (or something machine dependant) as per opengl spec |
18:32:48 | * | gsingh93 quit (Ping timeout: 276 seconds) |
18:39:26 | * | coffeepot joined #nim |
19:08:03 | * | xcombelle quit (Quit: Leaving) |
19:15:58 | * | kumul quit (Ping timeout: 256 seconds) |
19:18:33 | * | kumul joined #nim |
19:18:55 | * | Ven__ quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
19:19:07 | lll | https://github.com/nim-lang/nimble/issues/128 |
19:28:27 | * | Ven joined #nim |
19:45:25 | * | gsingh93 joined #nim |
20:01:46 | * | coffeepot quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) |
20:02:16 | * | coffeepot joined #nim |
20:09:43 | * | johnsoft quit (Ping timeout: 255 seconds) |
20:17:53 | * | TEttinger joined #nim |
20:19:24 | * | gsingh93 quit (Ping timeout: 276 seconds) |
20:30:49 | * | epichero joined #nim |
20:45:57 | * | BitPuffin|osx quit (Ping timeout: 250 seconds) |
20:49:00 | * | Mat4 joined #nim |
20:57:08 | * | Kingsquee joined #nim |
21:06:22 | * | wan joined #nim |
21:07:06 | * | taesoo quit (Ping timeout: 252 seconds) |
21:18:59 | * | intra quit (Ping timeout: 258 seconds) |
21:28:17 | * | taesoo joined #nim |
21:47:37 | federico3 | is any of you in San Francisco? |
21:47:51 | Xe | I'm in mountain view |
21:51:29 | * | Mat4 quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) |
21:56:48 | lll | is there a better way to stop float from infecting my float32s than this? proc `+`(x: float, y: float32): float = raise newException(OSError, "get off my lawn mister float64") |
21:57:15 | lll | repeated on all operation between float / float32 |
21:59:10 | lll | otherwise i have to make sure to qualify all my identifiers with : float32 + 'f32, and it get tedious fast + error prone |
22:02:30 | * | epichero quit (Remote host closed the connection) |
22:03:57 | * | coffeepot quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client) |
22:04:54 | * | BlaXpirit quit (Quit: Quit Konversation) |
22:07:49 | boop | so let's say i have loads of procs with this signature.. |
22:07:51 | boop | proc ndmInit(argc: cint, argv: ptr cstring): cstring {.cdecl, exportc, dynlib.} = |
22:07:57 | boop | (args, return type and pragmas) |
22:08:13 | boop | is it possible/easy to write a template/macro to clean that up? |
22:10:42 | lll | for the pragmas you can .push |
22:11:25 | lll | for the macro part you might ask mister macroman |
22:11:28 | lll | aka fowl |
22:14:38 | * | epichero joined #nim |
22:16:27 | * | MyMind quit (Ping timeout: 265 seconds) |
22:26:19 | * | epichero quit (Remote host closed the connection) |
22:27:04 | Araq | lll: you can use the .error pragma to catch it compile time |
22:27:18 | Araq | proc `+`(x: float, y: float32): float {.error.} |
22:27:30 | lll | o nice |
22:28:04 | lll | is there a way to just forbid any transformation from float32 to float? |
22:28:14 | lll | instead of overriding all the operations like that |
22:28:15 | flaviu | lll: Nice error message :) |
22:28:27 | lll | flaviu: thx ^^ |
22:29:49 | * | kumul quit (Ping timeout: 245 seconds) |
22:31:01 | * | sepisoad_ quit (Ping timeout: 264 seconds) |
22:32:22 | * | epichero joined #nim |
22:34:43 | * | BitPuffin|osx joined #nim |
22:35:25 | Araq | why are conversions from float32 to float a problem for you? |
22:50:39 | * | epichero_ joined #nim |
22:51:52 | * | epichero quit (Ping timeout: 256 seconds) |
22:53:21 | * | taesoo quit (Ping timeout: 264 seconds) |
23:10:37 | * | stefantalpalaru quit (Quit: Leaving) |
23:16:44 | * | epichero_ quit () |
23:17:03 | * | epichero joined #nim |
23:24:44 | * | brson joined #nim |
23:30:11 | * | brson quit (Ping timeout: 250 seconds) |
23:39:32 | * | boop is now known as boopisaway |
23:42:38 | lll | Araq: i know its early optimization, but im working with opengl, and i will always be sending float32 to it, so i dont really see the point of working with float64 |
23:43:37 | lll | i dont know the performance cost of all those conversions, plus it will double the memory used if i use float |
23:43:40 | lll | 64* |
23:45:30 | * | nande joined #nim |
23:47:57 | lll | i might just end up using float64 after all, its just to much of a pain to write 'f32 everyyyyyywhere |
23:48:58 | lll | and let whatever conversion that needs to happen happen |
23:49:42 | flaviu | lll: I'm pretty sure that intel CPUs at least do all FLOPs with 80 bit precision. |
23:52:40 | lll | flaviu: sure, but it still takes twice as much memory, and it get truncated anyway when it store it in some float32 array anyway, I guess ill have to setup some benchmark to compared 100% float32 operation, to mixed operations. |
23:53:02 | lll | compare* |
23:53:11 | flaviu | lll: Yep, the memory savings are not minor. |
23:53:31 | flaviu | Although as long as it stays on the stack (local variables), it makes no difference. |
23:54:03 | lll | for sure ill convert them before sending them to opengl, twice the bandwith used if not, and float64 arent even supported on many cards, and most are 2 to 24 times slower with float64 |
23:55:22 | flaviu | unrelated to the current conversation, I'm looking at astalgo, and there appear to be at least 3 different table implementations. I don't think that's an ideal situation to be in. |
23:57:06 | ekarlso | flaviu: what ya working on ? :p |
23:57:30 | * | Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
23:57:49 | * | Ven joined #nim |
23:58:20 | * | kumul joined #nim |
23:58:50 | flaviu | ekarlso: I'm looking into modernizing the compiler. |
23:59:30 | flaviu | See https://github.com/flaviut/Nim/commit/4ad2c33e144b008eed677ba77943ebf28a53f621 for example |
23:59:34 | * | Ven quit (Client Quit) |