| 00:16:12 | * | boop is now known as boopisaway |
| 00:25:43 | vikaton | anyone here good with the Windows API ? |
| 00:25:56 | Varriount | vikaton: I have a decent amount of experience. |
| 00:25:58 | vikaton | because I need a better alternative than GetAsyncKeyState |
| 00:26:16 | vikaton | Varriount: I need to user a low-level keyboard hook |
| 00:26:22 | Varriount | vikaton: In fact, I'm currently writing up a web scraper to collect the Windows API from MSDN |
| 00:27:02 | vikaton | nice |
| 00:27:21 | vikaton | anyhow, this is the only thing I can find related to my issue in the windows module, http://nim-lang.org/docs/windows.html#WH_KEYBOARD |
| 00:27:29 | vikaton | not rly that helpful |
| 00:27:36 | Varriount | vikaton: You've tried looking throught the list of functions here: https://msdn.microsoft.com/en-us/library/windows/desktop/ff468859.aspx ? |
| 00:27:45 | Varriount | vikaton: What's wrong with getAsyncKeyState? |
| 00:27:48 | * | rgv151 joined #nim |
| 00:28:17 | vikaton | Varriount: have you ever ran a program "as an adminstrator" ? |
| 00:28:26 | Varriount | Yes... |
| 00:29:00 | vikaton | well GetAsyncKeyState does not GetAsyncKeyState in that state of confirming to run as admin |
| 00:29:02 | Varriount | vikaton: You won't be able to intercept ctrl+alt+delete sequences without writing a device driver. |
| 00:30:39 | * | johnsoft quit (Ping timeout: 245 seconds) |
| 00:30:49 | vikaton | Varriount: It doesnt log keys when this shows up: https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTQNPYHGi-k3JrgopGpI2FKykFWQ_si0Izx0Su7cE6FwWLbgsMMVw |
| 00:31:27 | * | johnsoft joined #nim |
| 00:31:36 | Varriount | vikaton: Uh, that's a security measure. |
| 00:31:50 | vikaton | exactly |
| 00:32:10 | Varriount | vikaton: This page may be of some explanation: http://stackoverflow.com/questions/15898769/how-to-receive-keyboard-input-from-application-with-admin-privileges-to-a-non-ad?rq=1 |
| 00:32:21 | * | rgv151 quit (Ping timeout: 250 seconds) |
| 00:33:47 | vikaton | for some reason idk if we are on the same page |
| 00:34:07 | vikaton | Varriount: the program executes and works fine without admin privledges |
| 00:34:38 | Varriount | vikaton: Your program is unable to get input from 'elevated' programs, such as the User Account Control prompt, the Ctrl+Alt+Delete prompt, etc. |
| 00:34:56 | vikaton | ok yeah we are on the same page |
| 00:35:10 | vikaton | Varriount: would low level hooks work or nah? |
| 00:35:16 | Varriount | vikaton: As I said, that's a security measure, and one that is non-trivial to circumvent. |
| 00:35:50 | Varriount | "As for interacting with elevated programs, I think the only solution, and I believe the one TeamSpeak uses, is to embed an application manifest that sets uiAccess=true. This allows you to bypass UIPI without needing to run the application as administrator." |
| 00:35:53 | vikaton | oh damn |
| 00:37:43 | Varriount | vikaton: See the 'uiAccess Values' section here - https://msdn.microsoft.com/en-us/library/bb756929.aspx |
| 00:38:53 | Varriount | vikaton: Despite what *nix advocates might want to believe, the Windows development team isn't *entirely* stupid |
| 00:39:06 | vikaton | lol |
| 00:40:07 | Varriount | vikaton: As to how to go about authenticode signing and such... It would probably be easier to use visual studio and it's bundled tools to do that. |
| 00:54:35 | * | gsingh93 joined #nim |
| 01:11:41 | * | Xtopherus` is now known as Xtopherus |
| 01:16:34 | * | Xtopherus is now known as banister |
| 01:29:25 | * | rgv151 joined #nim |
| 01:34:01 | * | rgv151 quit (Ping timeout: 264 seconds) |
| 01:37:05 | IronY | I dont think the windows develops are stupid |
| 01:37:13 | IronY | but with having to cater so much to legacy software |
| 01:37:23 | IronY | and not being able to make cruicial breaking changes |
| 01:37:25 | IronY | well |
| 01:37:26 | IronY | windows |
| 01:41:18 | onionhammer | Varriount checking in.. whats up? |
| 01:41:41 | Varriount | onionhammer: I'm trying to get Sublime Text 3 working with the plugin... and failing T_T |
| 01:41:55 | onionhammer | :\\ |
| 01:41:59 | Varriount | It will only load the settings file if the plugins are in a zipped package. |
| 01:42:25 | Varriount | And reloading settings on setting file change won't work, for some unknown reason. |
| 01:43:05 | onionhammer | huh.. i havent run into that using st3 |
| 01:43:19 | Varriount | onionhammer: What OS are you using? |
| 01:43:24 | onionhammer | os x |
| 01:43:34 | Varriount | Hm. I'm using Windows. |
| 01:43:56 | Varriount | onionhammer: Anyway, it just needs testing and fixing. |
| 01:43:56 | onionhammer | i have st2 on my win box, and st3 on my work laptop(windows) |
| 02:00:09 | * | vendethiel quit (Ping timeout: 265 seconds) |
| 02:01:38 | ozra | Yo, Varriount - I pushed a "convert..." of the syntax file, I had missed that on the call syntax. |
| 02:01:54 | ozra | pushed as in PRed ;) |
| 02:04:20 | Varriount | ozra: Thanks. I'll look over it in a bit. |
| 02:04:30 | Varriount | I'm hands deep in data exporting code. |
| 02:05:21 | ozra | Cool :) |
| 02:05:48 | Varriount | ozra: I wonder how large msdn is... |
| 02:06:02 | * | ddl_smurf joined #nim |
| 02:06:06 | ozra | Wooooa, don't even dare to think that I don't wanna know, haha |
| 02:06:37 | ozra | Luckily yeeears since I had to reference it :) |
| 02:06:58 | ozra | I saw you were working on a scraper for it? |
| 02:09:33 | * | vendethiel joined #nim |
| 02:18:09 | Varriount | ozra: Yeah. |
| 02:18:39 | Varriount | ozra: The current windows modules is both incomplete and unreliable. c2nim is a good tool, but it would be a pain to feed the entire windows sdk through it. |
| 02:18:41 | ozra | What's the intention? Just collecting it in an orderly fashion for easier reference? |
| 02:18:53 | ozra | Aha, auto gen interfaces? |
| 02:18:59 | Varriount | Something like that. |
| 02:19:07 | ozra | Super cool! |
| 02:19:14 | Varriount | If nothing else, it's a good way for me to learn xpath |
| 02:19:25 | ozra | (Not that I use windows - but I hear some people do ;-) |
| 02:19:41 | Varriount | ozra: Someone has to develop software for it. |
| 02:19:44 | ozra | xpath - nim-mod? |
| 02:20:15 | Varriount | ozra: It's a query language for xml/html documents. I'm using a python libary to scrape things. |
| 02:20:57 | ozra | I was aware of that - just wondering if there was a nim for it. Alright. |
| 02:47:26 | reactormonk | ozra, no need to reference other issues, github inserts if you mention them in others - kinda |
| 02:50:34 | * | ozra quit (Ping timeout: 246 seconds) |
| 02:55:58 | * | gsingh93 quit (Ping timeout: 272 seconds) |
| 02:56:51 | * | darkf joined #nim |
| 03:01:08 | * | dtscode is now known as hit1er |
| 03:01:17 | * | hit1er is now known as dtscode |
| 03:05:22 | * | dtscode is now known as charmander |
| 03:08:26 | * | charmander is now known as cheraphy |
| 03:09:03 | * | cheraphy quit (Disconnected by services) |
| 03:09:29 | * | dtscode joined #nim |
| 03:21:45 | * | gsingh93 joined #nim |
| 03:29:41 | * | Jesin quit (Quit: Leaving) |
| 03:31:08 | * | rgv151 joined #nim |
| 03:36:07 | * | rgv151 quit (Ping timeout: 264 seconds) |
| 03:40:32 | * | vendethiel quit (Ping timeout: 258 seconds) |
| 03:49:42 | * | vendethiel joined #nim |
| 03:53:34 | * | kumul quit (Ping timeout: 258 seconds) |
| 04:06:15 | * | gsingh93 quit (Ping timeout: 252 seconds) |
| 04:22:18 | * | endragor joined #nim |
| 04:24:13 | * | pregressive joined #nim |
| 04:28:13 | * | darkf quit (Read error: Connection reset by peer) |
| 04:29:01 | * | darkf joined #nim |
| 04:32:08 | * | rgv151 joined #nim |
| 04:32:45 | * | endragor quit (Remote host closed the connection) |
| 04:35:08 | * | endragor joined #nim |
| 04:35:59 | * | endragor quit (Remote host closed the connection) |
| 04:36:22 | * | endragor joined #nim |
| 04:36:30 | * | rgv151 quit (Ping timeout: 258 seconds) |
| 04:36:46 | * | Kingsqueee joined #nim |
| 04:40:07 | * | endragor quit (Remote host closed the connection) |
| 04:42:32 | * | kumul joined #nim |
| 04:42:59 | * | endragor joined #nim |
| 04:43:51 | * | endragor quit (Remote host closed the connection) |
| 04:51:40 | * | endragor joined #nim |
| 04:52:01 | * | BitPuffin|osx quit (Ping timeout: 272 seconds) |
| 04:52:30 | * | endragor quit (Remote host closed the connection) |
| 04:57:14 | * | ddl_smurf quit (Quit: ddl_smurf) |
| 04:58:34 | * | endragor joined #nim |
| 04:59:26 | * | endragor quit (Remote host closed the connection) |
| 05:03:11 | * | Kingsqueee quit (Quit: Konversation terminated!) |
| 05:03:59 | * | endragor joined #nim |
| 05:04:53 | * | endragor quit (Remote host closed the connection) |
| 05:20:57 | * | vendethiel quit (Ping timeout: 250 seconds) |
| 05:41:25 | * | kumul quit (Ping timeout: 272 seconds) |
| 06:03:50 | * | Demon_Fox joined #nim |
| 06:25:29 | * | solidsnack joined #nim |
| 06:26:53 | * | saml_ quit (Quit: Leaving) |
| 06:41:35 | * | strcmp1 quit (Ping timeout: 272 seconds) |
| 06:43:35 | * | Mat4 joined #nim |
| 06:44:09 | * | Mat4 quit (Client Quit) |
| 06:53:27 | * | Jesin joined #nim |
| 06:59:52 | * | ddl_smurf joined #nim |
| 07:00:01 | * | ddl_smurf quit (Client Quit) |
| 07:04:26 | * | intra joined #nim |
| 07:15:24 | * | vendethiel joined #nim |
| 07:16:32 | * | endragor joined #nim |
| 07:17:27 | * | endragor quit (Remote host closed the connection) |
| 07:19:14 | * | jubalh joined #nim |
| 07:28:04 | * | BlaXpirit joined #nim |
| 07:43:37 | * | Demon_Fox quit (Quit: Leaving) |
| 07:46:51 | * | mischanix left #nim (#nim) |
| 07:48:23 | * | mAch17 joined #nim |
| 07:54:11 | * | strcmp1 joined #nim |
| 07:55:05 | * | mAch17 quit (Quit: Leaving) |
| 07:56:10 | * | solidsnack quit (Ping timeout: 272 seconds) |
| 08:01:45 | * | gmpreussner|work quit (Read error: Connection reset by peer) |
| 08:08:05 | * | jubalh quit (Quit: Leaving) |
| 08:09:15 | * | dtscode quit (Quit: what a weird night) |
| 08:11:27 | * | dashed joined #nim |
| 08:15:27 | * | sepisoad joined #nim |
| 08:21:45 | * | ozra joined #nim |
| 08:26:25 | * | solidsnack joined #nim |
| 08:46:21 | * | vendethiel quit (Ping timeout: 250 seconds) |
| 08:50:40 | * | johnsoft quit (Ping timeout: 255 seconds) |
| 08:52:44 | * | johnsoft joined #nim |
| 08:57:58 | * | banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 09:02:04 | * | milosn quit (Read error: Connection reset by peer) |
| 09:02:16 | * | milosn joined #nim |
| 09:04:10 | * | sepisoad quit (Quit: Leaving) |
| 09:05:51 | Varriount | Araq: The crawler is working. The files it outputs are (nearly) c2nim compatible - I'm leaving in the annotations. |
| 09:07:54 | Varriount | They're even sorted by what dll's they reside in. |
| 09:24:46 | * | TEttinger quit (Ping timeout: 258 seconds) |
| 09:27:46 | ozra | Varriount: cool to here. :) |
| 09:27:52 | ozra | *hear* |
| 09:29:51 | * | ddl_smurf joined #nim |
| 09:30:01 | * | ddl_smurf quit (Client Quit) |
| 09:33:24 | Varriount | ozra: I'm particularly proud of the enum translator |
| 09:34:01 | ozra | May I ask, where do you live Varriount ? |
| 09:34:13 | Varriount | ozra: Virginia. |
| 09:34:19 | Varriount | (In the US) |
| 09:35:02 | ozra | Suspected the US version (don't know any other ;-) Yeah, I was surprised to see you "so soon" so I figured you had other hours :) |
| 09:35:31 | Varriount | ozra: My sleep schedule got wrecked yesterday. I spent all night/morning working on this crawler. |
| 09:36:05 | Varriount | So I woke up at something like 5PM |
| 09:36:43 | * | solidsnack quit (Ping timeout: 264 seconds) |
| 09:37:48 | ozra | Ah, ok. I've had horrid sleep cycles (since I was in the womb according to dear mom). Tonight I slept 3h. Sometimes I don't sleep for three days. But now I'm on a good streak - it's been three weeks without missing a nights sleep :-) Slept aroun 4h every night, at least :-) |
| 09:38:32 | ozra | Ah, well, except for a few with only three, now that I think about it.. |
| 09:38:38 | Varriount | O_o |
| 09:40:24 | * | Arrrr joined #nim |
| 09:40:26 | ozra | Nice to hear it progressing so nicely man. Do you happen to know the right practise of "peeking" on a token inthe parser? |
| 09:40:42 | Varriount | ozra: You mean, in the Nim parser? |
| 09:40:52 | ozra | Varriount: Ah, yeah |
| 09:41:00 | Varriount | No idea. |
| 09:41:10 | Varriount | What are you trying to do? |
| 09:41:12 | ozra | Alright, I'll keep hopping around the source |
| 09:41:51 | Varriount | ozra: The parser is pretty self contained. |
| 09:42:43 | ozra | Yeah, its clean, haven't really figured out the 'responisibility distribution', it's a bit ad hoc, so, it's best to just read it all I guess :-) |
| 09:43:17 | Varriount | ozra: Well, the parser is clean. The rest of the compiler code... needs work, at least as far as documentation goes. |
| 09:44:16 | ozra | I'm working my way outwards. Began with some fixes in the lexer. Now I'm on to the parser. I guess it's at ast-mangling / rewriting it gets nasty? |
| 09:44:36 | Varriount | ozra: Yes. And Lambda Lifting. |
| 09:44:55 | Varriount | Lambda lifting... if like staring into the void. |
| 09:44:59 | Varriount | *is like |
| 09:45:09 | ozra | I imagine. Will be interesting... |
| 09:46:42 | ozra | Hey, btw, the "goto definition" in NimLime. It worked for a while, but now it's grey. Anything to it? Or maybe it's some problem with me having a symlink to my local repo I cloned from you (before, it was installed via pkg mgr) |
| 09:47:07 | Varriount | ozra: No idea. I didn't write that code, nor have I touched it. |
| 09:47:41 | Varriount | ozra: I wrote the check, document comment, and language files, mainly. |
| 09:47:59 | ozra | Ah, ok. Where did you find it? |
| 09:48:05 | Varriount | ozra: If you're using the unstable version from the repository, I've also rewritten the Nimble interface. |
| 09:48:25 | Varriount | ozra: onionhammer wrote the documentation and nimsuggest code. |
| 09:49:24 | Varriount | To be honest, it's a huge work-in-progress right now. I've been making large refactorings, and have done about 1/2 of the existing code base so far. |
| 09:49:27 | ozra | Varriount: Aha, it does use nimsuggest, ok cool. Was just wondering if it was some ad-hoc solution or using that, more 'forward compatible' solution.. |
| 09:50:17 | ozra | Varriount: Alright. Great job. Love it. And sublime. I wouldn't have found that editor if I hadn't looked for nim syntax highligthing and found NimLime! :-D |
| 09:50:23 | Varriount | :D |
| 09:53:10 | * | vendethiel joined #nim |
| 09:53:14 | Varriount | ozra: My favorite feature is multi-cursor support. |
| 09:54:10 | * | Matthias247 joined #nim |
| 10:05:17 | ozra | Varriount: was out smoking and taking the dog for a piss, had a streak of sun shine in a otherwise cloudy day. Yeah, that multi cursor thing - I thought it was a bug first, haha. I haven't looked at its purpose. What for? |
| 10:06:03 | Varriount | ozra: General usefulness. You can select multiple occurences of a word, edit multiple lines of similar structure, etc. |
| 10:06:57 | ozra | There are lots of great features yet to learn. I like that. It's already enormously useful :-) |
| 10:07:43 | ozra | Reminds me. This "magic quotes and brackets insertion" is killing me. Didn't find a setting for it? |
| 10:08:43 | Varriount | ozra: You mean, selecting something and typing a quote/brace/bracket, and having it put them on either side of the selection? |
| 10:09:49 | ozra | yeah. but I googled now. "auto_match_enabled": false. great. didn't find yesterday. bad google day I guess ;) |
| 10:10:28 | fowl | Morning all |
| 10:10:36 | ozra | Morning! |
| 10:11:04 | ozra | fowl: do you know the "right way" of peeking a token in the parser? |
| 10:12:36 | fowl | Hmm no |
| 10:13:21 | ozra | alright |
| 10:18:45 | Varriount | ozra: If you give me a min, I can try looking |
| 10:19:05 | Varriount | I found a bug in my xpath code which made the crawler skip all the structs. :< |
| 10:19:17 | ozra | Ok. Ayay. |
| 10:19:20 | * | bluenote_ joined #nim |
| 10:20:38 | ozra | I'm making a call to `exprList()`, and it calls getTok first thing. The thing is, before that I need to see if an _optional_ token is there, which I need to call getTok for. And if it's not there - a token ends up being skipped... |
| 10:21:09 | ozra | adding a 'tokenFetched' bool to params is last resort.. |
| 10:22:28 | ozra | Maybe I should refactor it. It breaks the pattern of the parser. Most procs call getTok() after their handling, leaving a fresh tok for the next bugger. exprList() reverses the pattern :-/ |
| 10:23:37 | * | ddl_smurf joined #nim |
| 10:24:46 | ozra | Nah, it's not that consistant.. I'll have to print it out and study it over coffee. Focus on your crawler Varriount :-) |
| 10:24:55 | * | johnsoft quit (Ping timeout: 256 seconds) |
| 10:25:52 | * | johnsoft joined #nim |
| 10:31:10 | * | banister joined #nim |
| 10:41:46 | * | endragor joined #nim |
| 10:42:32 | * | jubalh joined #nim |
| 10:48:30 | * | endragor quit (Remote host closed the connection) |
| 10:49:49 | * | endragor joined #nim |
| 10:53:06 | * | Trustable joined #nim |
| 10:53:32 | * | Trustable quit (Remote host closed the connection) |
| 10:54:15 | * | Trustable joined #nim |
| 10:56:30 | * | endragor quit (Remote host closed the connection) |
| 11:12:46 | * | jaco60 joined #nim |
| 11:13:46 | Varriount | ozra: If nothing else, additional documentation on the compiler internals would be much appreciated. |
| 11:19:11 | jaco60 | hi... is there a way to ask for the CPU "native" operand size for int ? |
| 11:19:53 | Varriount | jaco60: Could you rephrase your question? I'm not sure I understand. |
| 11:20:14 | def- | jaco60: sizeof(int)? |
| 11:21:06 | Varriount | Nim's ints are always the size of a pointer for the architecture the program is compiled on. |
| 11:22:05 | jaco60 | Ok : afaik, the size of a int is the size of a pointer (64 bits on my Core i7). But, on the i7, operand size are 32 bits. If i'm using "int" for my computing tasks, the code will use 64 bits instead of 32 bits |
| 11:22:19 | jaco60 | so, i'm using int32... |
| 11:22:25 | Varriount | Uh.. where did you hear that? |
| 11:22:55 | Varriount | jaco60: At least in the context of assembly, there is no such thing as a 'native' operand size. |
| 11:23:06 | Varriount | (assembly language) |
| 11:23:37 | jaco60 | for the very same code, switching for int to int32 divide the execution time by 2 |
| 11:24:14 | Varriount | jaco60: Depends on the code. That doesn't mean that the native operand size (whatever that means) is 32-bit. |
| 11:24:28 | jaco60 | http://en.wikipedia.org/wiki/X86-64#Intel_64 |
| 11:24:39 | * | dashed quit (Quit: Connection closed for inactivity) |
| 11:26:09 | Varriount | jaco60: I see nothing on there stating that operations on 32-bit data types are inherently 'native' or faster. |
| 11:27:14 | jaco60 | if the operand size is 32 bits, a 64 bit operation (say, add 2 int) will be slower, isn't it ? |
| 11:28:51 | Varriount | jaco60: Not necessarily. As I said, there are a great number of factors involved. |
| 11:29:45 | Varriount | Is the operand in a register? Has the operation been done speculatively? |
| 11:31:03 | jaco60 | well, that's not rocket science but compare time exec of this trivial code with both int and int32...http://sprunge.us/iTTA |
| 11:32:27 | jaco60 | http://sprunge.us/XAia, sorry |
| 11:33:57 | Varriount | jaco60: How are you actually testing speed? |
| 11:34:20 | Varriount | I don't see an import to the times module. |
| 11:34:40 | jaco60 | i'm on Unix, so i do "time ./amis" |
| 11:35:38 | jaco60 | 3.09sec for the int32 version, 9s for the int version... |
| 11:36:03 | Varriount | jaco60: Did you run with -d:release? |
| 11:36:06 | jaco60 | yes |
| 11:36:30 | jaco60 | and the same goes with gcc or clang |
| 11:36:56 | jaco60 | with i32 version, i get the same times than in D or Rust |
| 11:38:45 | Varriount | jaco60: Unfortunately, the benchmark doesn't measure *just* integer optimizations. I see float conversion there, as well as loops. |
| 11:39:49 | * | Arrrr quit (Ping timeout: 245 seconds) |
| 11:39:50 | Varriount | In this case, a guess (and only a guess) as to why the version using 32-bit integers is faster is because they cause less cache misses. |
| 11:40:56 | jaco60 | but if register are 32 bits, you're agree that using 64 operands would be slower, no ? |
| 11:41:14 | Varriount | Yes, however on a 64-bit system, the registers are 64-bits. |
| 11:41:26 | jaco60 | Varriount, i'm not so sure |
| 11:41:48 | def- | jaco60: I'd guess that 64 bit int division is much slower, but haven't checked |
| 11:42:40 | def- | and there can also be cache effects which would help 32bits, but don't see that here |
| 11:42:49 | ozra | jaco60: I raised this issue here in the IRC a couple a days ago but got no response.. |
| 11:43:44 | Varriount | ozra: What day? |
| 11:44:03 | ozra | Varriount: jaco60: It's a complex scenario to say the least. But often, i32 is faster. i64 _can_ be faster in specific scenarios. But I'm wondering if "using int size of pointer size of platform" is really the smartest decision. |
| 11:44:32 | ozra | Varriount: I can't remember, sorry. Got no replies what soever, so there's not much to read. |
| 11:45:10 | jaco60 | ozra, hence my initial question :) |
| 11:45:14 | Varriount | ozra: I really wouldn't make any assumptions until a good series of benchmarks is made. |
| 11:45:31 | ozra | Also in SIMD ops, more ops can be fit in registers, etc. |
| 11:45:58 | def- | if you know you want int32s, then use int32 explicitly |
| 11:46:25 | ozra | Varriount: jaco60: True. Measurements are important! It could be good to have, like in C/C++, a fastint type.. (fastest available int, all things considered) |
| 11:46:54 | Varriount | ozra: There's such a type? |
| 11:47:32 | ozra | def-: _If_ there's a measurable great improvement on a bunch of different code I think int should be re-specified... Otherwise. Use int32 in perf. code.. |
| 11:47:59 | ozra | Varriount: Yes, in C/C++? In stdlib, or somewhere, don't remember where it's defined. |
| 11:48:18 | Varriount | ozra: At this point, changing the size of an integer would probably break most of the C/C++ wrappers |
| 11:49:03 | ozra | I can't imagine those being based on a datatype that has a different specification from any base datatype in C++?? Sounds unlikely!? |
| 11:49:12 | * | ingsoc joined #nim |
| 11:49:20 | Varriount | ozra: Huh? |
| 11:50:19 | ozra | I imagine (haven't looked) that any inter mingling with C* must be done with specific width datatypes? Most C++ implementations on 64bit has 32bit int, for instance. The spec is not the same for C* int and Nim int. |
| 11:50:24 | Varriount | ozra: The fact that sizeof(int) == sizeof(pointer) is used in some places. I know it's used in the Windows pointer. |
| 11:50:34 | Varriount | *Windows module |
| 11:51:02 | ozra | Wooot? Sounds weird to me. Like it's begging to break. What am I missing here? |
| 11:51:13 | Varriount | When doing things like declaring types. |
| 11:51:47 | ozra | Shit. That doesn't sound good to me. I was hoping there wasn't such a relationship behind the int spec! :-( |
| 11:51:58 | fowl | when we use dlls we rely on ABI compatibility |
| 11:52:11 | fowl | when use headers we rely on source level compatibility |
| 11:52:26 | ozra | fowl: elaborate? |
| 11:52:29 | Varriount | Anyway, if we're going to go with anecdotal evidence, I'd like to point out that Nim compiles/runs faster when compiled for 64-bit, vs 32-bit, on my machine. |
| 11:53:16 | fowl | ozra, i got a pointer to some struct from a dll, to read info from it i better use the right offsets into the struct |
| 11:53:20 | jaco60 | Varriount, but on my machine, i'm compiling in 64 bits... The fact is the i32 operands are much more fast |
| 11:53:37 | jaco60 | the exec is a 64 bits exec... |
| 11:53:40 | Varriount | jaco60: *In your situation* |
| 11:53:41 | flaviu | Have you run it in a profiler? |
| 11:53:53 | ozra | Varriount: I don't doubt it runs faster using 64-bit instruction set. We're talking size of a specific datatype - that doesn't necessarily translate to being the width of the architecture. The main idea behind 64-bit was to be able to adress more memory. |
| 11:53:56 | def- | ozra: I like the simplicity of an int being pointer size |
| 11:54:11 | jaco60 | Varriount, on your machine, my code run faster with int rather than int32 ? |
| 11:54:14 | fowl | ozra, on the other hand, if my struct is imported from some header and the real int type is long and i use int32, probably be fine because the c compiler knows that in the header it should be type X |
| 11:54:26 | Varriount | jaco60: I can't time it on my machine at the moment. |
| 11:54:49 | jaco60 | it's mesurable just by running it :) |
| 11:54:54 | jaco60 | no need for any tool |
| 11:55:06 | jaco60 | 9 sec compared to 3 sec is sensible |
| 11:55:29 | ozra | def-: I don't see the simplicity in it. Other than that _if_ you cast ints and pointers back and forth their compatible. But that's a no no. And _when_ you do evil, there should be an ptrWidthInt type or the like - it should not be seen as intrinsic. Madness! |
| 11:56:15 | fowl | eh? |
| 11:56:18 | ozra | You posted code? I'll try it.. |
| 11:56:25 | fowl | simply put "int" and "float" dont belong in wrappers |
| 11:56:34 | jaco60 | ozra, http://sprunge.us/XAia |
| 11:56:49 | jaco60 | just swap from int to int32 for MyInt |
| 11:56:55 | ozra | Got it. Looks very micro-benchmarky though ;-) |
| 11:57:14 | ozra | fowl: Agree. |
| 11:57:27 | * | ARCADIVS quit (Quit: ARCADIVS) |
| 11:58:16 | jaco60 | ozra, my intention was not to benchmark but to learn Nim by implementing some algos to compare with other languages |
| 11:59:17 | ozra | jaco60: I understand, just saying, in a more complex program, the bench could turn out quite differently.. |
| 12:00:08 | * | Varriount goes and installs KCacheGrind for Windows |
| 12:00:20 | * | MyMind joined #nim |
| 12:00:43 | Varriount | ozra, jaco60: If I sound irritated, it's because I don't like not having a clear answer. |
| 12:01:42 | flaviu | hm, this is interesting. It's possible to see pipelining in the profiler |
| 12:01:49 | ozra | Varriount: Didn't perceive that. Clear answer to what? |
| 12:01:56 | * | Sembei quit (Ping timeout: 258 seconds) |
| 12:02:39 | Varriount | ozra: Why the two types behave so differently. |
| 12:02:46 | def- | jaco60: ok, i checked and it's just the div/mod that's much faster with smaller int types. everything else is the same speed on my machine |
| 12:03:28 | Varriount | def-: How did you find out? (I haven't used any C/C++ profilers before...) |
| 12:03:43 | Varriount | I'd like to see how Windows compares. |
| 12:03:52 | * | Arrrr joined #nim |
| 12:04:29 | def- | Varriount: i just converted the int64s to int32 before doing mod and div |
| 12:04:39 | ozra | Varriount: 1. How the processor is constructed, if they decide to optimize for int32 because it's most common, then that will be faster, for instance (and "re-writing" hardware is a bit slower iteration ;-). 2. cache-line hits/misses. 3. amount fittable in registers. etc. etc. |
| 12:05:05 | flaviu | The expensive part is definitely the division. |
| 12:07:35 | ozra | Ok. I ran that micro-bench a couple of times on my 64bit Ubuntu. Took the geometric mean (fairest mean for bench times imo), and the results are: int32: 6s, in64: 12.5s - _more_ than twice as slow. |
| 12:07:40 | Varriount | flaviu: A look at this is interesting.. http://www.agner.org/optimize/instruction_tables.pdf |
| 12:07:42 | flaviu | The sign extension also doesn't seem to be cheap. |
| 12:08:44 | ozra | flaviu: unsigned is often slower in real world code in my experience. C compilers has spec that they should do overflow checks on unsigned. But not on signed. |
| 12:09:07 | flaviu | ozra: unsigned is faster here. |
| 12:09:21 | ozra | flaviu: cool. Good to know. |
| 12:10:11 | ozra | I'll see if it's possible to re-compile the compiler with int32 - just for kicks. To get a more "real world" scenario benchmark... Might not be possible though.. |
| 12:11:25 | Varriount | ozra: Iterestingly enough, that table states that the number of clock cycles the i7 takes for div/idiv instructions using 64-bit registers is variable... and depends on the number. |
| 12:13:05 | ozra | Yes, ofc. And when we're down at the clock cycle level, other things are more important: pipe lining and cached vs not cached data |
| 12:14:10 | ozra | Varriount: a bsic optimization in hardware is to have fast machinery for counting leading zero bits in mem cells, this helps it reduce bit fiddling iterations in calculations.. |
| 12:14:31 | flaviu | The unsigned version emits a div instead of idiv. I wonder what the difference is. |
| 12:14:39 | Varriount | ozra: Huh. I'm learning all sorts of new things today. :D |
| 12:14:58 | ozra | The grat thing about being awake ;-) |
| 12:15:04 | * | xcombelle joined #nim |
| 12:15:07 | flaviu | Ah, unsigned vs signed. |
| 12:15:13 | jaco60 | if i change MyInt to an unsigned int, i get compile error :' |
| 12:15:13 | ozra | flaviu: LOL |
| 12:15:28 | def- | jaco60: import unsigned |
| 12:15:47 | flaviu | So unsigned division is approx. 50% faster than signed division :O |
| 12:15:52 | jaco60 | Ach ! |
| 12:16:23 | ozra | jaco60: unsigned is 'not recommended' by Nim.. hence the import. It recommends using "positive" datatype instead. I agree with much of the reasoning. |
| 12:16:57 | ozra | I'll modify the bench and re-run. It is sooo uncommon to do int divs.. |
| 12:17:23 | Varriount | ozra: Thankfully? |
| 12:17:49 | Varriount | ozra: I still don't like the idea of changing the size of the 'int' data type |
| 12:18:18 | Varriount | Although, I wonder what Araq's motivation for the decision was. |
| 12:19:11 | BlaXpirit | it's the only thing that makes sense -_- |
| 12:19:11 | ozra | Varriount: yes, thankfully. I mean, ints are basically used for iteration and indexing. Hence, _mostly_, add, sub and mul. That's what should be benched. I can imagine CPU designers don't give a shit about optimizing int-div... |
| 12:19:58 | ozra | And for Decimal types (where exact precision of for instance 0.1, 0.2 etc. is needed [those numbers do not exist in float]) |
| 12:19:59 | * | boopisaway is now known as boop |
| 12:20:58 | ozra | Varriount: I wonder too. I _hope not_ for casting ints and pointers easily. |
| 12:21:03 | * | gokr joined #nim |
| 12:21:24 | ozra | BlaXpirit: What do you mean? |
| 12:21:36 | fowl | Idk but when you ask him please write it in the FAQ |
| 12:21:47 | fowl | These discussions come up every few months |
| 12:22:05 | ozra | fowl: Yeah, that's a good idea, inevitably again, yes, haha |
| 12:23:48 | Varriount | ozra: Possibly for their range? Given that he's a proponent of signed types. |
| 12:24:20 | * | kumul joined #nim |
| 12:24:43 | Varriount | Given that 64-bit integer on 32-bit systems would be much slower, but 32-bit integers on 64-bit systems would have a smaller range. |
| 12:25:14 | ozra | But signed 32 bit is standard in most C compilers for instance. Its range is more than enough for most indexing and looping chores. When more is needed, one explicitly use a guaranteed minimum width datatype (like "atLeast64BitInt") |
| 12:26:00 | Varriount | *shrug* |
| 12:26:29 | ozra | The 32 vs 64 becomes especially interesting if someone creates an array of int. Then it will have a real impact on cache-fitability and performance, regardless of micro-timings like pipelining, registers and the like. Cache is king. |
| 12:27:16 | ozra | Varriount: and if you run Nim on a 32 bit system, int is still 32 bit, so... |
| 12:27:46 | Varriount | ozra: Feel free to put a case up for it. I don't really have too much of an opinion on the matter. |
| 12:27:55 | ozra | You wouldn't rely on that for range capability. The reasoning must be something else. |
| 12:28:26 | BlaXpirit | do you really not see it |
| 12:28:36 | BlaXpirit | it's just the processor's int |
| 12:28:41 | ozra | Varriount: The reason I haven't created an issue is that, before I've done real-word benchmarks, there's not a practical argument against it. |
| 12:29:00 | Varriount | BlaXpirit: The thing is, that argument doesn't make sense. |
| 12:29:11 | BlaXpirit | it's not an argument |
| 12:29:13 | BlaXpirit | it's the definition |
| 12:30:06 | ozra | BlaXpirit: There's no such thing as "the processors int". A 64 bit processor can adress with 64bit pointers, THAT is the point of 64 bit. That most ops and registers then are adapted and have 64bit sizes is a 'bonus' sort of. It doesn't make int64 "its datatype"!!! |
| 12:30:23 | Varriount | ^ |
| 12:30:29 | * | Jehan` joined #nim |
| 12:30:47 | Varriount | Oh yay, Jehan` is here. And gokr |
| 12:30:57 | Jehan` | What did I do? :) |
| 12:31:08 | ozra | Woke up! Congratulations! ;) |
| 12:31:27 | Varriount | Jehan`: Discussion for changing the size of Nim's 'int' data type to be always 32 bits |
| 12:31:32 | Jehan` | Well, I did that several hours ago … so it can't be that. |
| 12:31:33 | * | gokr quit (Remote host closed the connection) |
| 12:31:37 | * | gokr1 joined #nim |
| 12:31:44 | Jehan` | Varriount: Ugh. That's a complicated story. |
| 12:32:00 | Varriount | Jehan`: I know! Isn't it wonderful! |
| 12:32:07 | BlaXpirit | are you sure it was *changed*? |
| 12:32:11 | ozra | Varriount: Jehan` Not "always 32". But speced differently. _iff_ needed. We don't have benchmarks to support it. It's more a question of validity. |
| 12:32:12 | BlaXpirit | and wasn't so all along? |
| 12:32:12 | Jehan` | You can make arguments for and against, but note that first of all, this would break a lot of stuff. |
| 12:32:46 | Jehan` | Quite a bit of the library depends on sizeof(ptr T) == sizeof(int). |
| 12:33:09 | Jehan` | You can also make an argument that the default should be 64-bit. |
| 12:33:19 | ozra | Jehan`: I'm waiting for the bomb.. That it has to do with casting ptr <-> int. Please god no. |
| 12:33:25 | * | Varriount waits for ozra to have an apoplectic fit |
| 12:33:40 | BlaXpirit | that's actually a big part of it |
| 12:33:44 | Jehan` | This is a performance hit for 32-bit architectures, of course, but it would reduce the risk of inadvertent overflows. |
| 12:33:46 | ozra | Hahaha. All I want is what's best. |
| 12:34:20 | Jehan` | You can't do 100_000*100_000 in 32-bit, for example. |
| 12:34:34 | Varriount | Jehan`: Well, apparently division and module arithmatic for 64-bit integer, on 64-bit architectures, is twice as slow as the same for 32-bit integers. |
| 12:35:06 | ozra | Jehan`: The overflow issue is quite miniscule. When working on big data where big indexers and iteration vars are needed, one specifically use 64b< type. (like size_t in the "dreaded language") - remember, this is a systems language, for performance... |
| 12:36:24 | Varriount | Jehan`: To play devils advocate, one could argue that the standard library could/should be updated to use the 'SomeInteger' typeclass |
| 12:36:31 | Jehan` | Varriount: Yes, but it has the advantage that on 64-bit, 100_000^2 is correct, while on 32-bit architectures, 100_000^2 == 1410065408. |
| 12:36:41 | ozra | And once again, integer division is so uncommon, it should not be used for judging, imo. |
| 12:37:45 | Jehan` | The problem here is really that int means "we like to pretend that precision is infinite but that the actual precision is good enough for our particular application". |
| 12:38:57 | ozra | Since on my machine, int div, as you said also, was _more_ than twice as slow. But the most common uses for ints are indexing and looping. And then add, sub and mul are the important ones. For most arithmetic, like your example Jehan` , one would use a float*. Unless something like your example was for an array, seq, matrix like stuff. And then itn that case it's really real _big data_. |
| 12:38:57 | jaco60 | ozra, crypto make intensive use of integer division and modulo, afaik |
| 12:38:57 | Jehan` | In practice, we're subsuming several different concepts under the type "int". |
| 12:39:25 | Jehan` | From array indices to an approximation of natural numbers for simple numbertheoretic calculations. |
| 12:39:33 | fowl | Is there a module or package for benchmarking |
| 12:39:50 | ozra | jaco60: Yes, of course, there are specific cases of algos. I should clarify, I mean in _most_ common applications / solutions. |
| 12:39:54 | Jehan` | ozra: I wouldn't use float. Dear god, float is hell. |
| 12:40:23 | Jehan` | You use float when you (1) need it and (2) know exactly what you are doing. |
| 12:40:34 | Jehan` | Well, sometimes people use it for other stuff, but they deserve what they get. |
| 12:40:40 | ozra | Jehan`: with float*, I mean float32|float64|float128. Ofcourse it has its problems, but it's better than dealing with fixed point like in the old days, argg. |
| 12:40:42 | BlaXpirit | misunderstanding ensues |
| 12:41:18 | Jehan` | ozra: One of the problems is that you cannot check for overflow. |
| 12:41:45 | Jehan` | With fixed-size ints, you can at least get an error if things go wrong. |
| 12:41:50 | def- | fowl: maybe https://github.com/winksaville/nim-benchmark |
| 12:42:05 | Jehan` | But I'm probably the wrong person to ask here, anyway. |
| 12:42:17 | fowl | def-: thx |
| 12:42:44 | Jehan` | Please keep in mind that even though I may do a lot of stuff in Nim, its design goals are actually a mismatch for several of my needs. |
| 12:42:47 | ozra | Jehan`: what? do you mean? There's overflow, underflow, etc.. |
| 12:43:09 | ozra | But then oftenly, it's _not used_.. |
| 12:43:14 | Jehan` | ozra: FP overflow means your exponent got to big, not that you lost bits somewhere. |
| 12:43:45 | ozra | Jehan`: loss of resolution, sort of? |
| 12:44:25 | Jehan` | ozra: With float64, 2^54 == 2^54+1. |
| 12:45:04 | Jehan` | I.e. in floating point land, == means "approximately equal", not "equal". |
| 12:45:27 | Jehan` | There are plenty of application domains where this is just not acceptable. |
| 12:46:44 | ozra | Jehan`: Yeah I know (from coding Javascript and simulating int64's ;-) but as said, the alternative is fixed point. And then you'll have to use 128bits or so to be able to handle that number - if you want to handle a reasonable number of decimals. And you still cannot represent all numbers. I mean. It's just no way around it unless CPU's get a vaue-type made out of ratios.. |
| 12:47:02 | ozra | But it would still translate it to bit-operations somewhere along the line, and the res. would be lost.. |
| 12:48:12 | Jehan` | ozra: Yeah, but the problem is that 2^31-1 is a pretty small number when you're using it to represent real-world entities. 2^63-1 is not. |
| 12:48:21 | Jehan` | It's not safe, but it is safer. |
| 12:48:22 | ozra | Sure in f64 you can't represent 0.1 (or is it 0.2 [or both]) exactly. But you can't represent 1/3 correctly, and wouldn't think much about it. It's like a base10-is-god-habituation-fallacy.. |
| 12:48:37 | Jehan` | Again, I need to stress that my personal bias is colored by my personal needs. |
| 12:49:28 | * | johnsoft quit (Ping timeout: 252 seconds) |
| 12:49:44 | ozra | Jehan`: I still hold that integers are mostly used for indexing and looping. Arith in floaters... other uses would be exceptional cases. And _then_ you can be specific. I'm just saying the generic int should be for most common generic use. |
| 12:50:01 | * | johnsoft joined #nim |
| 12:50:16 | Jehan` | I am relatively unconcerned about the speed of the default settings, as long as you have the option to choose different ones for performance. |
| 12:50:24 | ozra | And, still, we don't know that int = ptr-size is bad. It's more of a question of "is it?". "For general performance"... |
| 12:51:29 | Jehan` | I'm sort of okay with int being the "native" size, because 64-bit ints on 32-bit machines can be a pretty nasty performance hit (especially on mobile). |
| 12:51:49 | ozra | Jehan`: Well if one writes for a case where you really have to milk the cycles out of the machinery, than having code in i32 and calling library procs with i64 will cause a hell of a lot of datatype promotions. So, you are basically forced to use the "native" one. |
| 12:51:57 | Jehan` | I'm not so sure that I want to trade down from 64-bit on architectures that can handle it with little performance loss. |
| 12:52:38 | * | banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 12:52:40 | fowl | ozra, same thing would happen in c but it would be implicit |
| 12:53:09 | ozra | Alt. is to reimplement dependencies. Ofc. you're completely right in your argument. Extreme performance apps are special cases too, and warrants extra care.. |
| 12:53:24 | ozra | fowl: definitely. |
| 12:53:38 | Jehan` | ozra: Not sure what you are arguing for right now, to be honest. :) |
| 12:55:01 | ozra | Simply put: To benchmark i32 vs. i64. _If_ general code benefits largely from i32, I think int spec should be re-considered. That's about it :-) So I guess we can drop this now, haha. |
| 12:55:34 | Jehan` | ozra: So, i32 as the default int on 64-bit architectures, too? |
| 12:55:45 | * | Varriount ponders how terrifying a programmer with experience on a debate-team would be |
| 12:56:06 | ozra | Jehan`: Exactly. Considering the big _iff_ |
| 12:56:31 | Jehan` | ozra: Ah. I'd be personally opposed to it. |
| 12:56:46 | * | bluenote_ quit (Ping timeout: 246 seconds) |
| 12:57:19 | ozra | Jehan`: The reasoning? Considering that you can't rely on int handling 100_000 * 100_100 for instance... (example) |
| 12:57:51 | Jehan` | ozra: Simple example would be that reading a file >= 2GB into a string would break. |
| 12:58:11 | Jehan` | And with -d:release, you'd get an immediate buffer overflow that may be exploitable. |
| 12:58:17 | ozra | Jehan`: Then that breaks on 32bit machine. Unless you specify it as int64! |
| 12:58:27 | Jehan` | On a 32-bit machine, you run out of memory. |
| 12:58:48 | ozra | So you're making dangareous assumptions. |
| 12:59:22 | ozra | Jehan`: Out of a practical stand point, probably, but not necessarily. But you mean that you basically assume int is 64 bit, and ignore 32 bit platforms? |
| 12:59:37 | Jehan` | No, I'm dealing in probabilities, not absolutes. Plus, that was an example. |
| 13:00:25 | * | Siecje joined #nim |
| 13:00:30 | Jehan` | I think that as a matter of practical usefulness, the default int size should be able to express the size of in-memory structures. |
| 13:00:47 | ozra | Yeah. Well, in any event, since I like the code to be deterministic, and you're good with probabilistic - we could round off now, and pick it up again, iff it actually becomes an issue ;-) |
| 13:01:03 | ozra | Jehan`: That is a good notion though. |
| 13:01:04 | Jehan` | That's not what I'm saying, sigh. |
| 13:01:22 | ozra | Jehan`: sorry. Your second statement made it clearer :) |
| 13:02:23 | Jehan` | Simply put, I *have* been bitten often enough by ints being 32-bit in C for this particular scenario (i.e. int being too short to handle large sizes) to not consider it a hypothetical. |
| 13:03:43 | ozra | Jehan`: Ok. I see some parallels with yesterdays type discussions here, haha. I've never been bitten by range limits, but by type less ness. I think it is very fair to say that different people attract different computing errors B-) |
| 13:04:06 | Varriount | ozra: You work with python at all? |
| 13:04:09 | Jehan` | ozra: I'm currently working on a machine with 500 GB of memory (64-core AMD Opteron). |
| 13:04:15 | Jehan` | 2 GB is *nothing*. |
| 13:04:49 | Jehan` | And I expect these numbers to go up over the next decade. |
| 13:05:30 | ozra | Varriount: no, have never actually.. Though I coded a transpiler to C/C++ with pythonish syntax around 99/2000, so I obviously like the syntax :) |
| 13:05:32 | * | skyfex_____ joined #nim |
| 13:05:56 | ozra | Varriount: (and has rewritten many times since then.. but..) anyway. Nim is here to save us, haha |
| 13:06:09 | Varriount | ozra: Python's integers automatically convert to bigints |
| 13:07:12 | Jehan` | Varriount: Same for Ruby and Smalltalk, by the way. |
| 13:07:39 | ozra | Jehan`: I think I do suffer from a problem: My first computer had 16KB RAM, my second had 64KB RAM, such things probably etch a bit of "must save bytes" wiring in to the brain ;-) Though, as said, I like benching. But all in all. I'm starting to sway in my opinion. So thanks for discussing this a bit. |
| 13:08:32 | Jehan` | ozra: Oh, I can relate to that. My first computer was a ZX 81. I did a lot of programming on a ZX Spectrum with 48k of memory. |
| 13:09:13 | ozra | Jehan`: It makes more sense taking all the different scenarios together. And: "future proofing" (64bit ops on the CPU is bound to be optimized more in the future) etc. |
| 13:09:30 | ozra | Jehan`: A friend had a Spectrum, coded a bit on that too :) |
| 13:09:45 | Jehan` | ozra: In general, I like to point to the 90/10 rule when it comes to performance. |
| 13:09:56 | Varriount | What's the 90/10 rule? |
| 13:10:00 | ozra | What happened to 80/20? |
| 13:10:01 | * | banister joined #nim |
| 13:10:02 | Jehan` | I.e. 10% of the code accounts for 90% of the runtime. |
| 13:10:19 | ozra | ;) |
| 13:10:38 | Jehan` | 80/20 is too high for most code. |
| 13:10:41 | Varriount | ozra: Huh. I know of 80/20 as a building material |
| 13:11:18 | Jehan` | So, there's the question whether you want to weigh down the majority of the code with design decisions made for optimization. |
| 13:11:29 | Jehan` | Or, pragmatically: Measure, THEN optimize. |
| 13:11:42 | * | johnsoft quit (Ping timeout: 272 seconds) |
| 13:12:20 | * | johnsoft joined #nim |
| 13:12:30 | ozra | Varriount: 80/20 is the paretos principle. Basically what said above, but regarding _anything_ (20% of customers gives 80% of income. (perhaps unrelated) 20% of customers takes 80% of the time.. etc) |
| 13:14:27 | ozra | Jehan`: Yeah. I mean, I only optimize in perhaps a few funcs in most apps, because of measurements. So I have to agree my "must use the optimal int-type for general increased perf of all apps compiled" is a bit ludicrous. So can we please drop this while I'm just feeling a little ashamed over my hard to shake off hard-wiring? ;) |
| 13:15:04 | Jehan` | ozra: I don't consider 32-bit ints as the default to be ludicrous. |
| 13:15:22 | ozra | Though, one of the better discussions in a long while. The ones where you form a new opinion :) |
| 13:15:31 | Jehan` | I think the disadvantages outweigh the benefits, but it's not as though it's clearcut. |
| 13:15:50 | ozra | No, that's for sure. |
| 13:18:15 | Jehan` | Incidentally, one of the things that I find attractive about Smalltalk and Dart is that they give you arbitrary-precision ints by default. |
| 13:18:33 | Jehan` | That's probably not the right choice for a system programming language, but it's nice when you can have it. |
| 13:18:56 | ozra | Yeah, I shiver a bit, haha |
| 13:19:15 | ozra | (the performance devil crawling out from behind the ear) |
| 13:19:44 | Jehan` | Actually, tracing compilers tend to deal with that pretty nicely. |
| 13:20:05 | Jehan` | You generally don't pay for the performance hit unless your numbers actually do become so big. |
| 13:20:53 | ozra | Aha, they basically just infer datatype dependig on expected "numerical size" of expressions? |
| 13:21:25 | ozra | Eh, but how? Or are they JIT with datatype usage caches, or what? |
| 13:22:13 | Jehan` | JIT, tracing. |
| 13:23:40 | Jehan` | Very simply put, the machine speculates that operations don't overflow and branch to the more expensive infinite-precision code when they do. |
| 13:24:16 | Jehan` | You do still pay the overhead for the overflow checks (usually pretty cheap), but even those can be optimized away in a lot of cases. |
| 13:27:13 | ozra | ah, ofc. hence the tracing.. Yes, I think JIT compiling has a lot of potential, where it _does_ have types also, it would save speculation, and if not, you still get the benefits mentioned. The bailouts can _usually_ be made pretty cheap too as you say. So I'm keeping eyes on this are of processing tech. So far I still have to use static compiling for my use cases though.. |
| 13:27:35 | ozra | ...this *area* of... |
| 13:28:16 | fowl | Varriount, is this from fowltek https://github.com/Varriount/DevIL ? |
| 13:28:44 | fowl | Varriount, more importantly is it complete (can i remove the DevIL in fowltek) |
| 13:30:16 | Jehan` | ozra: For example, these two pieces of code take about the same time: https://gist.github.com/rbehrends/2dec89e3ac5a0444490a |
| 13:30:35 | Jehan` | 6.3 seconds for Nim on my machine, 6.5 seconds for Dart. |
| 13:30:49 | Jehan` | And the Dart overhead is probably for running the JIT compiler mostly. :) |
| 13:31:56 | ozra | A bit tooo simplitic case ;-) Even just a little bit of arithmetic might test it harder, care to throw in? It'd be interesting to see.. |
| 13:32:04 | Jehan` | Actual optimization levels can be less than that (JIT compilers still have limitations in practice), but it's really impressive how close to native performance you can get. |
| 13:32:15 | fowl | Yous guys have inspired me to benchmark some of my code |
| 13:32:26 | fowl | Testing float v float32 |
| 13:32:39 | Jehan` | ozra: Arithmetic is unlikely to change things much; there are ways to really confuse JIT compilers if you want to. |
| 13:33:47 | Jehan` | I have some code that really hurts LuaJIT because it has so much polymorphic dispatch that is not amenable to tracing. |
| 13:34:04 | fowl | Jehan`: by overusing virtual functions? |
| 13:34:13 | fowl | Ah |
| 13:34:26 | ozra | Jehan`: I mostly have experience with V8 because of work involving iojs. It compiles quickly first, when parts become hot, it uses the "CrankShaft", and when it determines that datatype of certain vars has been stable for a time it inlines it as straight machine ops. And now there's turbofan, but so far it's alpha and not that impressing. The interesting part though is saving time on compilation, and optimizing the actually used parts.. |
| 13:35:11 | fowl | Smalltalk does similar with polymorphic inline caches |
| 13:35:12 | Jehan` | ozra: There's a lot of mindshare between the V8 and Dart compilers, as far as I know (since both teams work at Google). |
| 13:35:51 | ozra | Yeah, but in discussions I've been, it seems like even V8 devers hardly know what the other one does.. So, I dunno ;) |
| 13:35:59 | Varriount | fowl: I'm fairly sure I generated that directly from c2nim |
| 13:36:53 | ozra | Jehan`: I figured even simple arith would require them to make more checks on data size, while the loops are deterministic enough to just set the datatype at start - so it's not really a realistic test.. |
| 13:36:56 | fowl | Varriount: ok ^^ I'm trying to remove extra stuff from fowltek |
| 13:37:57 | ozra | Oy. Sunshine again - I better run out and take my Nim-parser print outs, coffee and tobacoo with me. bbl. |
| 13:39:42 | * | BlaXpirit_ joined #nim |
| 13:39:43 | * | BlaXpirit quit (Killed (orwell.freenode.net (Nickname regained by services))) |
| 13:39:43 | * | BlaXpirit_ is now known as BlaXpirit |
| 14:04:51 | Araq | Varriount: 'int' is most of the time used for *counting*. counting stuff that's in memory. hence the size reflects the pointer size. |
| 14:05:21 | Araq | that cast[int](p) is guaranteed to work is just a bonus, not the primary reason. |
| 14:09:12 | fowl | proc `/=`*[T: float|float32](x: var T, y: T) {.inline, noSideEffect.} doesnt seem to work for float32 |
| 14:09:59 | * | pregressive quit (Remote host closed the connection) |
| 14:14:53 | jaco60 | is there some stuff to do (conversion ?) to apply a sequtil proc to a nim string ? |
| 14:14:58 | * | yymoto2 joined #nim |
| 14:15:39 | BlaXpirit | jaco60, lol you're out of luck |
| 14:15:58 | BlaXpirit | one thing u can do is copy its source :| |
| 14:16:11 | * | Jehan` quit (Quit: Leaving) |
| 14:20:16 | fowl | most sequtils are templates |
| 14:20:24 | fowl | they should work with strings |
| 14:22:08 | jaco60 | fowl... i tried (intuitively, i admit) mess.filter(proc (c: char): bool = c > 'j') |
| 14:22:29 | fowl | jaco60, try filterIt (the template version) |
| 14:22:49 | fowl | mess.filterIt(it > 'j') |
| 14:23:33 | fowl | oh |
| 14:23:36 | fowl | hm |
| 14:23:59 | fowl | no that wont work with string either |
| 14:24:07 | jaco60 | nope... seems i have to do a toSeq(mess.items).filter... |
| 14:24:17 | jaco60 | (as seen on Stackoverflow) |
| 14:24:35 | fowl | bleh |
| 14:25:19 | * | yymoto2 quit (Quit: leaving) |
| 14:26:03 | * | bluenote_ joined #nim |
| 14:27:12 | * | gmpreussner|work joined #nim |
| 14:27:20 | fowl | jaco60, https://gist.github.com/fowlmouth/3041f72efa37bea27d1b |
| 14:29:01 | fowl | string can be casted to seq[char] >_> |
| 14:30:14 | jaco60 | if a write toSeq(mess.items).filterIt(it > 'j'), i get a seq of filtered char... so i have to reconvert this seq in string |
| 14:30:59 | fowl | jaco60, use the modified one that i just linked |
| 14:31:12 | fowl | that way is silly |
| 14:31:22 | jaco60 | fowl, yes, thanks... |
| 14:31:59 | fowl | toSeq() first makes a new seq, then filter makes a new one, then to convert it to a string thats a new object too |
| 14:32:18 | jaco60 | but if i have to wrote all this stuff for each functionnal proc (map, foldr, etc.), it could be tedious |
| 14:33:10 | fowl | jaco60, strings arent often used as seqs |
| 14:33:31 | jaco60 | in Nim... i see that :) |
| 14:34:50 | fowl | like i said thought you can cast string to seq[char] |
| 14:35:23 | jaco60 | i admit i struggle a bit with the nim doc :( |
| 14:39:09 | fowl | jaco60, have you seen this? https://nim-by-example.github.io/ |
| 14:40:32 | jaco60 | fowl: since two days, i've read most of blogs post about nim :)... (the one mentionned on nim-lang, at least) |
| 15:01:26 | ozra | sorry, just jumping in here.. jaco60, are you maniping strings as seq's to approach it functionally? |
| 15:02:28 | ozra | Ah, bit late. Need to walk the dog.. |
| 15:02:40 | jaco60 | ozra... i wanted to to so... but it seems it's not so easy in Nim... at best, i get a final seq[char] but, now, i have to reconvert it in string |
| 15:04:22 | BlaXpirit | well i have no idea how to do it nicely |
| 15:04:45 | Araq | jaco60: usually people use strutils and call it a day. no need to filter / map stuff when you have 'replace' |
| 15:05:27 | jaco60 | with future's lc, i have why i want... but i got a seq[char]... |
| 15:06:33 | jaco60 | Araq, i want to apply map to all chars of a string |
| 15:07:14 | Araq | why would you? :P |
| 15:07:34 | jaco60 | to do a rot13, for example |
| 15:08:16 | vikaton | jaco60: http://rosettacode.org/wiki/Rot-13#Nim |
| 15:08:57 | jaco60 | vikaton, except i want to obtain a string, not display each char as it comes |
| 15:09:34 | fowl | you can't figure it out from there? |
| 15:09:35 | jaco60 | basically, i try to wrote a rot13(mess: string): string proc |
| 15:09:37 | fowl | zz |
| 15:10:12 | jaco60 | fowl... i just have to recreate a string from the seq i've obtained with lc, yes |
| 15:10:34 | Araq | why is string not compatible to openArray[char]? fowl, can you make this work? |
| 15:10:49 | BlaXpirit | why is openArray a thing? |
| 15:10:51 | Araq | seems to be just an oversight of sigmatch.nim |
| 15:13:12 | fowl | Araq, in typeRel? |
| 15:13:22 | Araq | yes |
| 15:15:32 | * | darkf quit (Quit: Leaving) |
| 15:18:45 | vikaton | could windows.h theoretically be written completely in Nim without C FFI ? |
| 15:19:10 | xcombelle | cikaton theoricately ? |
| 15:19:29 | xcombelle | s/te// |
| 15:19:53 | vikaton | yes |
| 15:20:57 | Araq | is that another way of asking "can Nim *really* do everything that C can do?" |
| 15:21:26 | Araq | and yes, it *really* can do that. C is nothing special. |
| 15:22:20 | vikaton | gg |
| 15:25:36 | * | Matthias247 quit (Read error: Connection reset by peer) |
| 15:28:27 | Arrrr | vikaton, how is going on your keylogger? |
| 15:28:55 | vikaton | Arrrr: it's finished for the most part https://github.com/Vikaton/nim-keylogger |
| 15:35:02 | * | gokr joined #nim |
| 15:45:29 | fowl | Araq, wow that was easy |
| 15:46:14 | ozra | jaco60: Araq - isn't it possible to overload map etc. to handle strings like 'pseudo' arrays? |
| 15:47:36 | fowl | ozra, no thats the lame way |
| 15:47:48 | fowl | ozra, string should just be compatible with openarray[char] |
| 15:47:59 | ozra | fowl: haha, so eh, why isn't char seq's working? |
| 15:48:12 | ozra | (sorry haven't followed discussion) |
| 15:48:24 | ozra | sounds reasonable |
| 15:49:19 | federico3 | we all want the Nim book. |
| 15:50:56 | Araq | federico3: it's in the works! :D |
| 15:51:06 | jaco60 | ozra: a string.chars() to produce a seq[char] and a conversion from seq[char] to string would be enough, imo |
| 15:51:28 | ozra | should be piece of cake, no? |
| 15:51:41 | jaco60 | in the meanwhile, future's lc and a for loop do the trick |
| 15:51:48 | fowl | Araq, well it was easy to make string match for openarray[char], now i have to make it match for openarray[t] |
| 15:58:27 | * | rgv151 joined #nim |
| 15:59:06 | Arrrr | Will the book teach us how to become masters using macros ? |
| 16:10:11 | iznogoodd | Varriount: "The fact that sizeof(int) == sizeof(pointer) is used in some places. I know it's used in the Windows pointer." |
| 16:10:21 | iznogoodd | Im using this in opengl code too. (the standard nim opengl wrapper was wrong last time i checked) |
| 16:11:37 | * | rgv151 quit (Remote host closed the connection) |
| 16:14:00 | jaco60 | i have a question about case... Given this code (http://sprunge.us/bXhh), it's clear i would prefer to write 'return c' in the else branch but Nim complains |
| 16:14:37 | jaco60 | Error: value of type 'int literal(97)' has to be discarded |
| 16:15:03 | jaco60 | i understand why but is there some nim idiomatic to handle this kind of situations ? |
| 16:18:53 | * | rgv151 joined #nim |
| 16:21:09 | ozra | araq: We had a long discussion earlier about the reasoning around sizeof(int) == sizeof(ptr). Is there a faq, doc, or what's the rationale? We finally arrived at 'indexing as much memory as adressable is good' |
| 16:21:18 | ozra | *as is adr...* |
| 16:22:50 | vikaton | Anyone creating something interesting in Nim ? |
| 16:26:15 | ozra | Implementing new Nim features, does that count? ;-) |
| 16:26:46 | Arrrr | in which features are you working ozra ? |
| 16:27:07 | jaco60 | conversion from seq to string :) |
| 16:27:29 | vikaton | yeah that counts haha |
| 16:27:44 | ozra | Right now? I dislike the 'of' syntax of cases, unfortunately, but like '|' from LS etc. So I'm tinkering with the parser making an experimental implementation... |
| 16:28:25 | ozra | It makes the branches of a case-construct so super clear imo, no noise.. |
| 16:28:34 | fowl | yea |
| 16:28:45 | fowl | one character less will save man hours surely |
| 16:29:02 | ozra | It's about _readability_ - not typing. |
| 16:29:02 | vikaton | wait can I see an example syntax ? |
| 16:29:03 | fowl | but i have to hold shift to get | so its the same amount of keystrokes |
| 16:29:36 | fowl | readability? |
| 16:29:39 | iznogoodd | what we need is pattern matching, not different syntax for cases |
| 16:29:45 | iznogoodd | that will help readability |
| 16:29:47 | ozra | It's even worse on swedish layout ;-) typing is not a priblem |
| 16:29:53 | vikaton | iznogoodd: agreed |
| 16:30:23 | * | matkuki joined #nim |
| 16:33:09 | jaco60 | iznogoodd, plus one |
| 16:33:23 | ozra | vikaton: haven't done a gist before, hope I got it right. basically: https://gist.github.com/ozra/74b0c1afaea8c87040b3 |
| 16:34:23 | ozra | iznogoodd: example of pattern matching (there are so many variants). This implementation I'm doing doesn't replace 'of', just complements... |
| 16:34:43 | ozra | (question mark after pattern matching) |
| 16:34:54 | vikaton | ozra, I kinda like it :) |
| 16:34:59 | iznogoodd | T(x,y) = T(1,2) |
| 16:35:11 | iznogoodd | thats what pattern matching is |
| 16:35:37 | ozra | Alright. So. No clash. No harm. |
| 16:35:41 | iznogoodd | in a case --- T(x,y) => x + y |
| 16:35:41 | fowl | ozra, i wish you would fix bugs instead of implementing new features |
| 16:35:54 | iznogoodd | or whatever the syntax |
| 16:36:13 | iznogoodd | fowl: or needed feature |
| 16:36:14 | vikaton | oh does Nim have lambdas? |
| 16:36:16 | fowl | or syntax even |
| 16:36:19 | fowl | a non feature |
| 16:36:46 | ozra | vikaton: I also like that for if .. elif .. elif ... It can be muddy if a log if-else suddenly is a 'new' if-else chain. with `case if: | a == 47: bla\n | foo(a) ==bar(a): blorgh` - if you understand the one line code ;) |
| 16:37:16 | ozra | vikaton: lamdas are implemented as a macro (!) :) |
| 16:37:40 | fowl | er? |
| 16:37:44 | fowl | proc = .. is a lambda |
| 16:37:57 | jaco60 | for my own, the absolute missing part of Nim is a book :) |
| 16:37:59 | ozra | fowl: I fixed the number literals yesterday. Now I'm making a 'fun coding'. I much prefer them. So why should I not try it out? |
| 16:38:02 | * | kumul quit (Quit: Leaving) |
| 16:38:30 | vikaton | ozra, tell me when u do a pull request |
| 16:38:35 | * | kumul joined #nim |
| 16:38:47 | ozra | iznogoodd: , fowl - for me it _is_ a needed feature. Not 'deadly needed', but strongly wanted. |
| 16:39:16 | ozra | fowl: meant shorthand notation of lamdas |
| 16:39:27 | * | skyfex______ joined #nim |
| 16:39:52 | iznogoodd | of vs |, not much diff, same keystroke |
| 16:39:56 | ozra | vikaton: sure. It does seem like most syntactical things I want are rather unwanted though :-/ |
| 16:40:12 | ozra | iznogoodd: keystroke? |
| 16:40:18 | iznogoodd | key stroke |
| 16:40:27 | ozra | iznogoodd: key stroke? still? haha |
| 16:40:30 | iznogoodd | ... |
| 16:40:37 | iznogoodd | finger hitting a key |
| 16:40:55 | iznogoodd | shift + # vs o + f |
| 16:41:26 | jaco60 | or AltGR + 6 on french keyboard... |
| 16:41:33 | vikaton | ozra, always worth a try |
| 16:41:36 | * | skyfex_____ quit (Ping timeout: 256 seconds) |
| 16:41:37 | ozra | not you too? really? typing? One READS code. Types it once, reads it over and over. I find of's to be less readable than `|` by far, even less readable than `case` from C. |
| 16:41:46 | iznogoodd | http://www.thefreedictionary.com/keystroke |
| 16:41:51 | vikaton | ozra, and if its not accepted, use templates :D |
| 16:41:53 | ozra | altgr + '<' (next to left shift) on swedish... |
| 16:42:20 | Arrrr | i prefer of, and i doub | will be accepted. But i should not stop you from experimenting |
| 16:42:21 | fowl | ozra, maybe | is a template already |
| 16:42:28 | fowl | case .. of |1: ... |
| 16:42:30 | jaco60 | ozra, while speaking of readability, indentation of "of" or "|" would be better imho |
| 16:42:42 | ozra | Not sure this can be solved in templates? |
| 16:42:43 | fowl | now its case .. | |1: .. ? |
| 16:43:04 | fowl | no it cant be solved in templates because its not currently valid syntax |
| 16:43:18 | jaco60 | but i buy |... that the same as Haskell :) |
| 16:43:54 | * | johnsoft quit (Ping timeout: 244 seconds) |
| 16:44:06 | ozra | jaco60: It's there and LiveScript I've gotten used to it from. Yup. That's why I try it out in the parser. Have to mutate a token also, from the lexer. So probably very non-templatable.. |
| 16:44:10 | * | johnsoft joined #nim |
| 16:44:22 | * | pregressive joined #nim |
| 16:44:43 | * | fowl left #nim ("Leaving") |
| 16:45:18 | * | Jehan` joined #nim |
| 16:47:07 | ozra | vikaton: Yes, I will try a PR, but as seen, alternative syntaxes aren't to popular of what I've met here so far - even though they don't break anything and create no ambiguouties :-/ "My way or the high way - and it's got to be the old way", is the feeling I get sometimes. I mean, OK if it totally broke a feature, or had no precidense what soever. But that's not the case. |
| 16:47:42 | ozra | And the compile cost is zero (not used = no time consumed) |
| 16:50:06 | ozra | Arrrr: No, I mean that's the only way to learn the ropes of the compiler also, whether it's accepted or not, the next bug is squashed faster because of what one learns :) |
| 16:51:31 | vikaton | cool |
| 16:51:44 | vikaton | is this outdated? http://rosettacode.org/wiki/Dining_philosophers#Nim |
| 16:56:21 | vikaton | yeah thats pretty outdated |
| 16:56:28 | vikaton | we goota f1x |
| 16:59:16 | ozra | jaco60: BTW, indenting 'of' is already possible - but perhaps that's what you meant? |
| 16:59:26 | jaco60 | yes |
| 16:59:35 | ozra | ah. |
| 16:59:51 | iznogoodd | jaco60: let base = if c in 'a'..'z': ord('a') else: ord('A') |
| 17:00:00 | iznogoodd | after you are you its a letter |
| 17:00:37 | iznogoodd | if not (c in Letters): return c |
| 17:00:43 | iznogoodd | sure* |
| 17:00:59 | iznogoodd | (letters is from strutils |
| 17:02:57 | jaco60 | iznogoodd, thanks... |
| 17:03:45 | ozra | vikaton: well the implementation of '|' cases is "ready" - it works on my test-cases now. Gonna tinker a bit more :) |
| 17:03:59 | iznogoodd | or you could simply implement some helper functions to get the final base = ord(applyIf(isLowerCaseLetter(c),toLower, 'A')) |
| 17:04:07 | ozra | Mmm, gonna test so that they work indented too. Think so. |
| 17:04:32 | jaco60 | iznogoodd, too cryptic, imho |
| 17:04:36 | vikaton | sweet @ ozra |
| 17:04:46 | iznogoodd | lol not cryptic if you know functinal programming |
| 17:05:35 | jaco60 | iznogoodd, but as i begin with Nim, i prefer to wait to scramble FP paradygms with it ;) |
| 17:05:37 | Arrrr | But nim uses explicit sintaxis (if aCond and bCond/ case of, etc) |
| 17:06:36 | ozra | jaco60: '|' works with indentation too as expected, if you're interested when it's "done done".. |
| 17:06:57 | jaco60 | ozra, cool |
| 17:08:20 | jaco60 | so, my caesar code miss only a conversion from seq to string (i've done it via a for loop, which is rather ugly |
| 17:08:53 | vikaton | anyone wanna help? |
| 17:11:25 | * | jubalh quit (Quit: Leaving) |
| 17:12:00 | ozra | Next test after a smoke break: replace entire compiler syntax with '|' (just for testing - promise!) |
| 17:17:32 | Araq | ozra: already explained why 'int' is the way it is |
| 17:18:09 | Araq | and '|' for 'case' is rather inconsistent. |foo currently is prefix operator | followed by 'foo'... |
| 17:19:08 | BlaXpirit | looks like a bunch of bikeshedding |
| 17:19:16 | Araq | there is no other construct that uses operator symbols why is it essential for 'case'? |
| 17:34:39 | iznogoodd | and it steals the | operator |
| 17:35:07 | ozra | Araq: aha, today? I'll search. Regarding '|' it is fully unambioguous in case contructs. It does not steal the operator at all. I just compiled part of the compiler with '|' notation. No problem. |
| 17:35:32 | iznogoodd | ||x... |
| 17:35:38 | iznogoodd | it would be ugly |
| 17:35:51 | iznogoodd | and confusing |
| 17:35:52 | BlaXpirit | but WHY |
| 17:37:04 | ozra | Araq: I find it to be much clearer - it gets out of the way completely, the of's are small enough to create noise, 'case' actually is less "noisy" imo, but I rather have it like is, but optionaly being able to use the '|' notation, simply because I find it superior for readability. |
| 17:37:13 | iznogoodd | the | syntax could make sense or the ADT but not for cases imo |
| 17:37:22 | iznogoodd | for* |
| 17:37:56 | iznogoodd | of x | y |
| 17:38:05 | * | pregressive quit (Remote host closed the connection) |
| 17:38:18 | ozra | It's common i guard clauses and similar so has a relation imo. Also worked flawlessly in LS for years... But, as said, it's an experiment. But seems stable already.. |
| 17:38:47 | Jehan` | I think there should be corollary to Greenspun's Tenth Rule of Programming; namely that every language's syntax diverges towards being a dialect of Perl. :) |
| 17:40:08 | ozra | Jehan`: haha. That's a complex bastard syntax wise.. But as long as things don't clash and makes the compiler more complex, I prefer if anyone can choose _their_ preferred style. This mod is rather un-intrusive. It's one block of code in one own branch in the parser... |
| 17:40:45 | Jehan` | Well, the thing is having dozens of dialects baked into the same language. |
| 17:41:10 | Jehan` | It makes reading code a hell of an experience if everybody writes in a different style. |
| 17:42:01 | iznogoodd | exactly, alternative syntax is ok if its needed, but with this nothing would be gained |
| 17:42:19 | ozra | Jehan`: yes I see your point. When it comes to readability, everyones eyes works differently, the semantics aren't changed in any way. I don't like the style of the nim compiler source at all, but I have no problem coding in it and sticking to it. It's a different thing with the projects one spend day in and out on. That's of no others business than the team working on it imo. |
| 17:42:50 | * | dashed joined #nim |
| 17:44:01 | ozra | Jehan`: It is already possible to write code that looks like from mars in Nim, thanks to macros etc. That doesn't mean one _has_ too. And also, if modifying someone elses code, say a module you love, or like this language I fancy a lot, you can live with accepting the style of the authors code. But it's silly to lock down syntax style when the cost is basically zero, and some would love it. |
| 17:44:38 | Jehan` | ozra: Well, in the end, you'll have to convince Araq, not me. :) |
| 17:44:57 | Araq | ozra: one doesn't *have* to do it, but you surely *intend* to use it everywhere. |
| 17:45:35 | iznogoodd | lol, while your at it, implement a pragma that can change all syntax, aka {.syntax[of]: |.} lol /s |
| 17:45:41 | Araq | you might as well implement #! ozra's dialect. |
| 17:45:52 | * | gsingh93 joined #nim |
| 17:46:15 | ozra | Well, yes I do, in _my_ code in _my_ projects. And I'd be happy for PRs, but if someone can't even bare to look at my code because of say dashes or the pipes, so be it, it's ultimately I who will work on it formost. |
| 17:46:27 | ozra | Araq: hahaha. I had that one coming! |
| 17:47:08 | ozra | But seriously, this is the only other thing I'd really enjoy. Most everything else I can live without, postfix if's etc. That's not necessary. |
| 17:47:16 | Araq | ozra: the problem is that people will blame Nim for having #! ozra's dialect, not you for using it. |
| 17:48:45 | ozra | I don't think there will be any blame, since it doesn't clash with the existing one? It's perfectly simple to just use 'of', I think that's a great syntax if it's found to be clearer to the user. Rather it is a bonus, as I see it. |
| 17:49:46 | * | anthgur joined #nim |
| 17:50:47 | Araq | personally I think #!dialects are all well worth it. Better have an ozra coding Nim modules in #!dialect than an ozra coding in C++ which is much *more* incompatible with the rest of Nim, but that's an unproven assertion. We might as well lose brilliant programmers because "Nim is just a mess". |
| 17:51:37 | ozra | Here's a piece of the compiler source, tell me it's not fucking super clear: https://gist.github.com/ozra/74b0c1afaea8c87040b3 |
| 17:51:43 | iznogoodd | Araq: yes thats true, but look at haskell now, with all its damn extensions |
| 17:52:10 | flaviu | ozra: I'm not against `|`, I'm against different dialects. |
| 17:52:12 | iznogoodd | like the overloaded strings GADTS etc |
| 17:52:37 | flaviu | If someone says "this code is in Nim", it should look like Nim. |
| 17:52:45 | Araq | ozra: it's ugly and believe it or not |
| 17:53:03 | ozra | araq: I understand your reasoning. I just think, Nim as a potential to attract people like me, from C++ imperativism, with a functional coding interest, and people from python, from golang (when they realize Nim is the same but better ;-), people from lisp. I mean, ultimately, we all want a big module universe, right? :-) |
| 17:53:22 | iznogoodd | ozra: tbh i wouldnt mind one syntax or the other, but having both in a language divides everyone |
| 17:53:47 | iznogoodd | ozra implement Tagged union with | syntax instead + pattern matching |
| 17:53:51 | ozra | Haha, fair enough, araq, we do have different tastes to some syntactical elements, though we all here seem to have same tastes to the language and semantics.. |
| 17:53:55 | iznogoodd | that would be more productive |
| 17:54:32 | Araq | Oberon uses | too and I said WTF when I saw it. |
| 17:54:42 | Araq | cause it simply doesn't fit the language at all. |
| 17:54:46 | ozra | And again, I don't think it's really a divide, just a route to more contributions and modules. But of course, this is nothing I could possibly prove. |
| 17:55:20 | Araq | if - elif - else |
| 17:55:28 | Araq | try - except - finally |
| 17:55:37 | Araq | case - | - | - else |
| 17:55:38 | Araq | wtf? |
| 17:56:31 | * | milosn quit (Ping timeout: 255 seconds) |
| 17:56:35 | boop | are there any examples of concepts in use? |
| 17:57:32 | ozra | well, for what it's worth, what I use in my code for long if elif else are ("nimified"): case if: | a == b: doFoo(); | bar(a) == true: doMuckyPup(); | _: doDefault(); (with semicolons 'simulating newlines here') |
| 17:58:00 | BlaXpirit | ozra, please just stop :/ |
| 17:58:17 | ozra | In that example gist I posted - I find it looks like a bullet list of possible events. Just fucking super clear :) I don't care if it follows a pattern. |
| 17:58:30 | iznogoodd | i prefer && and || vs OR and AND, but the language isnt like that |
| 17:58:32 | ozra | BlaXpirit: Well, well. I'll quit then. |
| 17:59:03 | flaviu | ozra: "bar(a) == true" ;) |
| 17:59:33 | flaviu | But I might be the only other person that likes that syntax |
| 17:59:42 | ozra | flaviu: ? I missed the joke? :-/ |
| 17:59:48 | vikaton | iznogoodd: you could always use templates :) |
| 18:00:01 | * | synthmeat quit (Quit: WeeChat 1.3-dev) |
| 18:00:06 | flaviu | ozra: "bar(a) == true" is equivalent to "bar(a)" |
| 18:00:08 | iznogoodd | vikaton: lol i like being standard |
| 18:00:24 | vikaton | flaviu: as suggestion to nim by example, you should add arrows on the left and right |
| 18:00:30 | vikaton | for easy navigation |
| 18:00:31 | * | matkuki quit (Quit: ChatZilla 0.9.91.1 [Firefox 38.0.1/20150513174244]) |
| 18:00:31 | ozra | flaviu: haha, yeah - just wanted to clarify the context.. |
| 18:01:00 | flaviu | vikaton: I've been told, but I can't come up with a good way to implement it. |
| 18:01:19 | iznogoodd | flaviu: just copy rust by example |
| 18:01:25 | Arrrr | lol |
| 18:01:27 | ozra | btw (I will quit in a sec, ok?) araq, it is `case - | - | - | _` (not that it matters in the discussion much) |
| 18:01:38 | flaviu | iznogoodd: They use gitbook, I use nanoc. |
| 18:02:35 | ozra | Well. I'll finish my tests in my feature anyhow. I've learnt alot about the parser today in any event :) |
| 18:02:37 | * | synthmeat joined #nim |
| 18:02:46 | * | synthmeat quit (Client Quit) |
| 18:02:57 | iznogoodd | flaviu: it would be pretty ez to add with javascript |
| 18:03:10 | Araq | muhahaha |
| 18:03:11 | * | kumul quit (Ping timeout: 250 seconds) |
| 18:03:25 | flaviu | iznogoodd: There's an idea! Thanks! |
| 18:03:40 | iznogoodd | flaviu: the left menu has all what is needed |
| 18:04:00 | iznogoodd | (to find the next page etc) |
| 18:04:01 | * | BitPuffin|osx joined #nim |
| 18:04:06 | * | synthmeat joined #nim |
| 18:05:32 | Araq | ozra: so first {. .} is too ugly. then the world really needs Unicode dashes. now it's 'case of'. tomorrow it might be the * export marker. at which point will you enjoy Nim? |
| 18:06:28 | BlaXpirit | when u have these urges, go use LiveScript |
| 18:07:57 | * | Demon_Fox joined #nim |
| 18:08:40 | flaviu | ozra: Small imperfections are not a big deal. |
| 18:08:47 | ozra | Araq: Let's put it this way, I much prefer it over C++ already, I mean my thin transpile is an ugly hack admittedly, and yes I might be a little picky, I just see like, when it's _possible_, and _not in the way_ (non breaking, no additional compile cost, etc..), that's how I reason, "why not try and get it in"? I like the semantics and the meta features over C++. |
| 18:09:34 | flaviu | ozra: There is a cost, it's just not a cost that is immediately apparent. |
| 18:09:37 | BlaXpirit | you want changes personally for you which would harm everyone else |
| 18:09:55 | ozra | flaviu: Araq: No you're right. It's not the end of the world, but when one has tried basically everything, and think that one has built pretty enlightened choices as to some syntactical looks, (with one obviously being me), it's worth a bit of arguing ;) |
| 18:09:59 | BlaXpirit | reasons: there is more to learn |
| 18:10:15 | * | milosn joined #nim |
| 18:10:22 | BlaXpirit | if you say there is no need to learn it, then why add it in the first place |
| 18:10:23 | flaviu | ozra: You are welcome to keep a personal patch set. I do that myself (granted, my patches are pretty minor) |
| 18:10:25 | ozra | BlaXpirit: I don't agree. If I thought it hurt, then I would not argue for it at all. I don't. |
| 18:10:35 | BlaXpirit | if someone doesn't learn it, they can't understand code written using it |
| 18:10:50 | BlaXpirit | if nobody uses this, it's bloat and maintenance cost |
| 18:10:57 | ozra | Yes, flaviu, that is an alternate route I consider. But then my code won't work for anyone else.. *shrugs* |
| 18:11:16 | BlaXpirit | if someone new comes across this, they'll say "yuck" and leave nim |
| 18:11:47 | ozra | That's your opnion BlaXpirit. Based on what? |
| 18:11:54 | BlaXpirit | if you don't think it would hurt... then start thinking |
| 18:12:16 | jaco60 | anyway, i prefer addition of functionnalities like transparent str <-> seq (yes, i insist) than syntax addition |
| 18:12:36 | ozra | BlaXpirit: please elaborate. |
| 18:13:00 | BlaXpirit | i stated reasons it is hurtful and don't see a single reason why it's beneficial |
| 18:13:08 | Arrrr | +1 to @[j,a,c,o,6,0].toStr |
| 18:13:23 | BlaXpirit | well that's a pull request away |
| 18:13:37 | BlaXpirit | there may even be something like that already, just that we can't find it |
| 18:13:44 | flaviu | I think that the Go people have a good point, they just drew the line in the wrong spot. |
| 18:14:07 | ozra | BlaXpirit: Your reasons for it being hurtful is based on your personal opinion of what would scare you away..? |
| 18:14:21 | BlaXpirit | i didn't say it would scare me away |
| 18:14:42 | BlaXpirit | and you're disregarding the rest |
| 18:14:47 | BlaXpirit | of what i said |
| 18:15:43 | ozra | I'm sorry BlaXpirit, I might have missed something, being to occupied 'defending the feature'. If you'd be so kind and sum it up? I'd be happy (I can search) |
| 18:16:00 | BlaXpirit | I said it 5 minutes ago |
| 18:17:10 | ozra | I'm scanning :) |
| 18:17:15 | BlaXpirit | please don't let me discourage you, i'm just direct like this :| |
| 18:18:31 | BlaXpirit | the main and obvious point should be... alternative syntaxes are usually bad. |
| 18:18:46 | BlaXpirit | and no, i don't have anything to confirm this point |
| 18:20:45 | ozra | Wow, can't believe it took me so long to find it! I think it's fair to say that there are features in many languages that people never learn until they at some point dives into some 3rd part project and see it, and go, "hmm, what's that?", they then look it up and gets it. |
| 18:21:08 | BlaXpirit | features, sure |
| 18:21:18 | BlaXpirit | but we aren't really discussing a feature |
| 18:21:29 | ozra | features, constructs, sytax, what have you not.. |
| 18:22:22 | ozra | Yes, well, I will quit my ramblings about this here now. I really enjoy Nim, don't get me wrong on that! I'll keep this feature local to my machine for now. |
| 18:26:06 | ozra | Int reasoning confirmed. Jolly good :) |
| 18:33:49 | * | bluenote_ quit (Ping timeout: 246 seconds) |
| 18:37:27 | jh32 | hi |
| 18:38:23 | * | rgv151 quit (Quit: Leaving...) |
| 18:39:15 | Araq | jh32: hi |
| 18:39:21 | jh32 | I'm playing with the parallel example from the docs, is it expected that it doesn't compile as parallel.nim, but works when saved in pi.nim? |
| 18:39:44 | Araq | yes. |
| 18:39:49 | jh32 | or is this some problem in my setup? |
| 18:40:02 | Araq | I'm not saying it's good. but it is expected. |
| 18:40:05 | jh32 | ah, ok |
| 18:40:23 | Araq | parallel then is the module name, confusing the compiler |
| 18:41:02 | jh32 | .. was confusing me too :-) |
| 18:47:36 | boop | is there an easy way to collapse an iterator into a string? |
| 18:47:40 | boop | a la list(iter) in python |
| 18:48:13 | Jehan` | Into a seq, you mean? strutils.toSeq does that. |
| 18:48:41 | boop | thanks! |
| 18:48:59 | * | anthgur quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 18:50:05 | boop | hmm- i'm trying to use this http://nim-lang.org/docs/strtabs.html#keys.i,StringTableRef |
| 18:50:18 | boop | but it keeps spitting out "undeclared identifier: 'keys'" unless i use it in a for |
| 18:52:00 | iznogoodd | flavius: https://gist.github.com/HOLYCOWBATMAN/064621887e58c01ed6aa |
| 18:52:13 | * | Jesin quit (Quit: Leaving) |
| 18:52:31 | iznogoodd | flaviu:^ |
| 18:52:41 | iznogoodd | that would give you previous and next Href |
| 18:52:47 | flaviu | iznogoodd: thanks, I'll use that. |
| 18:53:11 | iznogoodd | i just tested it in my firebug live, but it should work everywhere ithink |
| 18:53:28 | * | Mat4 joined #nim |
| 18:54:01 | iznogoodd | flaviu: sorry for no break and the missing ; |
| 18:54:43 | flaviu | iznogoodd: No worries, I'll figure it out. |
| 18:56:36 | vikaton | flaviu: ur implementing it :D |
| 18:56:46 | flaviu | vikaton: iznogoodd did most the work :P |
| 18:57:06 | vikaton | anything works :P, gj iznogoodd |
| 18:57:13 | iznogoodd | =) |
| 18:57:38 | vikaton | guys we need to update this rosseta code for Nim |
| 18:57:54 | vikaton | http://rosettacode.org/wiki/Dining_philosophers#Nim |
| 18:58:48 | * | sepisoad joined #nim |
| 18:59:14 | def- | vikaton: I think my version is updated here: https://github.com/def-/nim-unsorted/blob/master/diningphilosophers.nim |
| 19:00:14 | * | Siecje quit (Ping timeout: 245 seconds) |
| 19:00:25 | vikaton | def-, what is acquire and release? |
| 19:01:17 | Araq | verbs. |
| 19:01:26 | iznogoodd | #rekt |
| 19:01:45 | * | Siecje joined #nim |
| 19:02:09 | * | vikaton cries |
| 19:02:45 | Araq | "good day sir. I'd like to acquire and then release the lock." |
| 19:04:06 | sepisoad | i just compiled some nim code on raspberrypi, It works... :) |
| 19:04:36 | Araq | not self-explanatory? |
| 19:04:43 | vikaton | sepisoad: did u follow a tutorial ? |
| 19:04:49 | flaviu | sepisoad: yep, ARM is also in the test matrix: http://buildbot.nim-lang.org/waterfall |
| 19:04:55 | vikaton | Araq,I need to read up on threads, but makes sense |
| 19:04:56 | iznogoodd | the tutorial of life |
| 19:05:17 | sepisoad | actually i used the arch linux arm build |
| 19:05:19 | Arrrr | computer science is not included in that tutorial. |
| 19:05:26 | sepisoad | there is already nim packages there |
| 19:08:45 | * | dtscode joined #nim |
| 19:15:02 | vikaton | def-, I think your code goes on forever |
| 19:15:23 | vikaton | dtscode: the keylogger is finished btw |
| 19:15:26 | def- | vikaton: i think that was the intention |
| 19:16:04 | * | gokr quit (Ping timeout: 245 seconds) |
| 19:16:34 | dtscode | vikaton: nice |
| 19:16:38 | vikaton | Oh, then the Rust version is wrong :/ |
| 19:18:41 | * | onionhammer quit (Read error: Connection reset by peer) |
| 19:19:23 | * | onionhammer joined #nim |
| 19:23:57 | Arrrr | Why is koch necessary to debug certain crashes? |
| 19:24:40 | * | onionhammer quit (Read error: Connection reset by peer) |
| 19:24:42 | Araq | Arrrr: it isn't. but it's convenient. |
| 19:24:42 | * | Matthias247 joined #nim |
| 19:25:10 | * | onionhammer joined #nim |
| 19:25:10 | onionhammer | ok |
| 19:25:37 | Araq | onionhammer: are you Arrrr ? o.O |
| 19:25:54 | Arrrr | Yes, you discovered me |
| 19:28:21 | Arrrr | Could someone explain me what im doing wrong here? http://pastebin.com/uWsCxZeV |
| 19:30:56 | * | gsingh93 quit (Ping timeout: 272 seconds) |
| 19:31:45 | flaviu | Arrrr: but your IP is from Spain and onionhammer's is from Texas. |
| 19:32:13 | flaviu | err, Minnesota |
| 19:32:30 | Arrrr | That's because i use proxies |
| 19:33:24 | flaviu | Of course ;) |
| 19:43:16 | * | sepisoad quit (Quit: Leaving) |
| 19:43:19 | * | TEttinger joined #nim |
| 19:45:12 | * | Jehan` quit (Quit: Leaving) |
| 19:46:40 | * | solidsnack joined #nim |
| 19:49:18 | * | gokr1 is now known as gokr |
| 19:49:31 | dom96 | Arrrr: why the two identities? |
| 19:50:26 | BlaXpirit | dom96, in case you're not playing up for a joke, that was a... joke |
| 19:51:03 | dom96 | oh. |
| 19:51:25 | BlaXpirit | at least i'm almost sure |
| 19:56:25 | * | ozra quit (Ping timeout: 246 seconds) |
| 19:57:13 | * | sepisoad_ joined #nim |
| 20:04:39 | * | sepisoad_ quit (Quit: Lost terminal) |
| 20:07:53 | * | Jesin joined #nim |
| 20:11:03 | flaviu | vikaton, iznogoodd: Pushed, you should see the change in a few minutes. |
| 20:11:25 | vikaton | nice! |
| 20:14:33 | * | pregressive joined #nim |
| 20:14:55 | * | banister quit (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 20:15:29 | iznogoodd | lolol https://please-enable-js/ |
| 20:17:29 | flaviu | iznogoodd: I figure that that's the best way to do it. I mean, I can't do alert("please enable js"), can I? |
| 20:18:03 | BlaXpirit | D: |
| 20:19:02 | iznogoodd | but i got javascript enabled :D |
| 20:19:15 | Araq | what? so now I need to enable JS for "nim by example"?! |
| 20:19:31 | iznogoodd | <a id="arrow-prev" class="text-icon disabled" href="↽">https://please-enable-js/">↽</a> |
| 20:19:46 | flaviu | Araq: It works fine without JS, but JS makes it a little bit more convenient. |
| 20:20:24 | * | Trustable quit (Remote host closed the connection) |
| 20:20:24 | Araq | ah good. cause I really like to browse the web with lynx on my toaster |
| 20:21:13 | flaviu | Araq: That's not a problem at all. Nim-by-example works fine in lynx. |
| 20:21:47 | vikaton | flaviu: I have a suggestion for the syntax highliting |
| 20:22:12 | flaviu | vikaton: Go ahead and post it |
| 20:23:14 | Araq | cool. now the only problem is that the download is too slow with my 1 baud modem. |
| 20:23:59 | vikaton | flaviu. some thing like http://radiux.io/articles/code-syntax/ or http://siciarz.net/24-days-of-rust-hyper/ |
| 20:24:20 | flaviu | Araq: Minimum requirements are 115000 baud, sorry. |
| 20:24:20 | * | Jessin joined #nim |
| 20:24:35 | dom96 | Is it just me or is nim-by-example incredibly slow now? |
| 20:24:39 | iznogoodd | flaviu: allright its working, it doesnt work at root tho |
| 20:24:51 | dom96 | in fact, it's still loading... |
| 20:25:05 | Mat4 | Araq: You can still get the download probably per e-mail ;) |
| 20:25:12 | vikaton | loads fine for me |
| 20:25:18 | * | Jesin quit (Disconnected by services) |
| 20:25:20 | * | Jessin is now known as Jesin |
| 20:25:53 | dom96 | It seems better now |
| 20:26:15 | dom96 | still an annoying delay sometimes |
| 20:26:33 | dom96 | I'm guessing it's the time it takes for the JS to load |
| 20:26:38 | dom96 | because the ticks take a while to load |
| 20:26:47 | dom96 | as well as the previous and next arrows |
| 20:26:53 | * | ingsoc quit (Quit: Leaving.) |
| 20:27:23 | * | Arrrr quit (Quit: lynx2-8-8) |
| 20:27:39 | flaviu | dom96: While I attempt to provide a good experience on every device, you might experience some lag on toasters :P |
| 20:28:05 | dom96 | toasters? |
| 20:28:18 | strcmp1 | a reference to netbsd, i think |
| 20:28:21 | flaviu | Literal toasters. |
| 20:28:55 | * | banister joined #nim |
| 20:29:07 | cazov | do you support PIC10? I only have 384 words available to me |
| 20:29:20 | vikaton | flaviu: thoughts? |
| 20:29:33 | flaviu | vikaton: You mean a dark theme for code highlighting? |
| 20:29:43 | * | iznogoodd is now known as HOLYCOWBATMAN |
| 20:30:23 | dom96 | flaviu: http://i.imgur.com/Tkz6To8.png |
| 20:30:44 | flaviu | cazov: Sure, but the content might be truncated a bit. |
| 20:31:22 | flaviu | dom96: That's your network's fault, I can't help you there. |
| 20:31:52 | dom96 | Everything else loads fine. |
| 20:32:00 | flaviu | It's mostly blue, "waiting". |
| 20:32:25 | dom96 | yeah. It's strange. |
| 20:32:42 | dom96 | oh well |
| 20:34:16 | flaviu | Although it does load in 500ms for me, which is a little slow. |
| 20:35:53 | vikaton | flaviu: yeah |
| 20:36:36 | * | ozra joined #nim |
| 20:39:22 | * | Kingsquee joined #nim |
| 20:40:20 | * | xcombelle quit (Quit: Leaving) |
| 20:41:29 | ozra | Kadaam. Just had a power outage caused by... beavers! |
| 20:42:25 | BlaXpirit | D: |
| 20:44:41 | dom96 | ozra: Beavers? For real? |
| 20:46:01 | ozra | dom96: haha. yeah, I've been waiting for it to happen. And finally it did. Still came as a surprise. They've been goin rather hard at the trees down by the stream. And it's been windy like hell today. So. Baaam. |
| 20:46:13 | Varriount | ozra: And thus, my plan to take over the world with an army of beavers proceeds... |
| 20:46:16 | ozra | (air borne wires...) |
| 20:46:25 | ozra | Varriount: haha |
| 20:46:37 | * | synthmeat quit (Quit: WeeChat 1.3-dev) |
| 20:47:17 | ozra | Luckily the wire didn't break, so they just cut power to chop up the hanger.. |
| 20:47:40 | ozra | Varriount: So, you've got to train your beavers harder! |
| 20:50:12 | * | synthmeat joined #nim |
| 20:51:54 | vikaton | wow ._. |
| 20:52:16 | vikaton | I just now realized that its Consonant not Constant ( ._.) |
| 20:52:28 | Varriount | vikaton: Huh? |
| 20:52:54 | vikaton | in the English language |
| 20:54:19 | dtscode | vikaton: You mean the non vowels? |
| 20:54:25 | vikaton | yes |
| 20:54:43 | Araq | vikaton: that's cool. create a bug report for it. Our docs should have mentioned that. |
| 20:55:05 | vikaton | Araq: in the middle of it right now |
| 20:55:34 | vikaton | it was found in nim-by-example |
| 20:55:55 | vikaton | flaviu: any news and possibly changing it? |
| 20:56:35 | flaviu | vikaton: Sorry, I've been focusing on optimizing load time to absurdity :) |
| 20:56:40 | flaviu | it should be easy enough to change. |
| 20:56:47 | flaviu | err, to add an option to change it. |
| 20:56:59 | vikaton | Yes that would be nice |
| 20:57:38 | vikaton | flaviu: the menu button doesnt work |
| 20:57:56 | * | Mat4 left #nim ("http://quassel-irc.org - Chat comfortably. Anywhere.") |
| 20:57:59 | pigmej | hmm, guys what would be way to port lib from non async to async ? |
| 20:58:07 | HOLYCOWBATMAN | flaviu: https://github.com/flaviut/nim-by-example/pull/34/ |
| 20:58:14 | HOLYCOWBATMAN | fixed a bug with the navigation |
| 20:58:38 | HOLYCOWBATMAN | if the current url cant be found in the list (as in the root url) it will act as if it were on hello world |
| 20:58:49 | dom96 | pigmej: what lib? |
| 20:59:25 | HOLYCOWBATMAN | i havent tested it tho, i edited it live on github |
| 20:59:33 | flaviu | HOLYCOWBATMAN: I think I like the current behavior better. |
| 20:59:40 | flaviu | I'll just add the root URL to the toc. |
| 20:59:47 | HOLYCOWBATMAN | ok sure |
| 20:59:48 | pigmej | dom96: nim-nanomsg |
| 20:59:58 | pigmej | so it's more like porting C |
| 21:00:14 | HOLYCOWBATMAN | just look at your console, currently it break everything since you do not return NAV at all if its not found |
| 21:00:54 | HOLYCOWBATMAN | if you would have more javascript, it wouldnt run |
| 21:01:38 | dtscode | <pigmej> so it's more like porting C |
| 21:01:40 | dtscode | I'm so sorry |
| 21:01:40 | flaviu | HOLYCOWBATMAN: Good point, I'll fix that. |
| 21:01:51 | dom96 | pigmej: Do you know how async works in nanomsg? |
| 21:02:07 | dom96 | I see a poll proc but that's about it. |
| 21:02:55 | * | HOLYCOWBATMAN is iznogood |
| 21:03:05 | * | filwit joined #nim |
| 21:03:16 | Varriount | Hello filwit |
| 21:03:23 | filwit | hi Varriount |
| 21:03:54 | * | BitPuffin|osx is now known as BitPuffin |
| 21:03:55 | pigmej | dom96: well not very precise but there is a pool, and that's it |
| 21:04:25 | pigmej | epoll on linux systems, etc, nothing very special I suppose |
| 21:04:53 | pigmej | afair nanomsg doesn't have any weird / hacks as zmq have. |
| 21:04:54 | HOLYCOWBATMAN | flaviu: o sorry i thought the getting started page was the same as root |
| 21:05:19 | HOLYCOWBATMAN | i just saw that its not |
| 21:05:23 | dom96 | pigmej: Why not just use Nim's sockets? |
| 21:06:11 | pigmej | nano have quite cool survey protocol |
| 21:06:21 | dtscode | Nim actually has a very nice socket lib :D |
| 21:06:24 | pigmej | it's bus is also quite fun to use. |
| 21:06:39 | pigmej | dtscode: definitely agree ;-) |
| 21:06:55 | dom96 | dtscode: That's nice to hear :) |
| 21:08:10 | pigmej | btw dom96 how is your experiment with async + threads going ? |
| 21:08:17 | dtscode | or rather, compared to the other ones I've used. Reminds me a lot of pythons lib, but it isn't a toy :D |
| 21:08:26 | dom96 | pigmej: don't really have time for that right now. |
| 21:08:37 | dom96 | pigmej: It's kind of in Araq's hands, but he doesn't have time for it either heh |
| 21:08:54 | pigmej | typical deadlock... ;D |
| 21:09:08 | filwit | hey Araq, whenever you're around, can you explain to me why the lineF/linefmt commands wont call my custom system.chcks.raiseNilError via '#raiseNilError' ? The compiler keeps saying it can't find that procedure, and all my greps aren't helping. |
| 21:09:47 | * | BitPuffin is now known as BitPuffin|osx |
| 21:10:24 | filwit | Araq: err.. lineftm in the cgen, if that wasn't clear |
| 21:10:51 | dom96 | pigmej: I wonder if nanomsg's fds are just normal OS fds. |
| 21:11:16 | pigmej | dom96: kinda |
| 21:11:33 | dom96 | questions is whether you can pass them to epoll |
| 21:11:37 | dom96 | *question |
| 21:12:13 | dom96 | If you can then you can just call newAsyncSocket(nanoMsgFd) |
| 21:12:14 | pigmej | maybe the simple ones, but I'm not that familiar with nanomsg internals |
| 21:12:20 | pigmej | hmm |
| 21:12:36 | dtscode | pigmej: link to nanomsg? |
| 21:12:52 | pigmej | dtscode: https://github.com/nanomsg/nanomsg ? |
| 21:12:54 | pigmej | or nanomsg.org |
| 21:12:59 | dom96 | well, newAsyncSocket(nanoMsgFd, true) # or false if you want unbuffered |
| 21:13:02 | * | gsingh93 joined #nim |
| 21:13:06 | dtscode | thanks |
| 21:13:24 | pigmej | hmm, i will play then with it in this way... maybe it will work somehow ;) |
| 21:13:46 | pigmej | but I think the advanced protocols are a bit more complicated but who knows ;-) |
| 21:13:58 | * | Jesin quit (Quit: Leaving) |
| 21:14:12 | dom96 | pigmej: If that won't work then you will need to write a dispatcher. |
| 21:14:22 | dom96 | pigmej: probably, there are many ways to do this. |
| 21:14:36 | dtscode | afaict its normal fd's |
| 21:14:47 | dom96 | pigmej: Nobody in the history of Nim has done it yet, sooo... you're going to have to experiment :) |
| 21:16:06 | pigmej | dtscode: I'm not sure, https://github.com/nanomsg/nanomsg/blob/master/src/protocols/survey/surveyor.c this doesn't look like fd |
| 21:16:38 | dom96 | I wonder if this also works on Windows: http://nanomsg.org/v0.5/nn_poll.3.html |
| 21:16:56 | dtscode | hmmm |
| 21:17:12 | dom96 | and if so, how. |
| 21:17:20 | onionhammer | Araq what? no im not Arrrr :P |
| 21:17:41 | Varriount | dom96: What are SP sockets? |
| 21:17:56 | dom96 | we should also get Nim on here: http://nanomsg.org/documentation.html |
| 21:18:10 | pigmej | dom96: windows... who cares;p |
| 21:18:29 | Varriount | pigmej: I do! |
| 21:18:49 | * | Varriount is the kind of person who reads 'Windows Internals' |
| 21:18:57 | dom96 | Varriount: good question |
| 21:18:59 | pigmej | Varriount: ouch |
| 21:19:09 | dom96 | I care about Windows too |
| 21:19:52 | dom96 | Nim developers are very Windows friendly :P |
| 21:20:13 | Varriount | dom96: SP = Serial Port |
| 21:21:01 | dom96 | Varriount: You sure? |
| 21:21:36 | Varriount | dom96: I can't think of anything else that would make sense. |
| 21:21:41 | pigmej | ehs,,, windows ;) |
| 21:21:46 | dtscode | special purpose? |
| 21:23:11 | dom96 | Well I just went into #nanomsg and asked. |
| 21:23:15 | dom96 | Because I am really curious heh |
| 21:24:12 | dom96 | oh, I think it's Socket Pair |
| 21:24:13 | pigmej | it's special purpose |
| 21:24:24 | pigmej | dom96: pair is just one of types |
| 21:24:33 | dom96 | hrm |
| 21:24:40 | dom96 | Special Purpose does make sense. |
| 21:24:44 | pigmej | hah |
| 21:24:45 | pigmej | nop |
| 21:24:46 | pigmej | Scalable Protocols |
| 21:24:48 | pigmej | "Scalable Protocols" |
| 21:25:07 | pigmej | one of GO packages have it;D |
| 21:25:53 | dom96 | omg lol |
| 21:26:44 | pigmej | but special purpose sounds better;d |
| 21:26:47 | * | phira quit (Ping timeout: 256 seconds) |
| 21:26:58 | dtscode | :D |
| 21:29:38 | reactormonk | pigmej, muh, I'm stuck somewhere at a social event, wanna take a look at nim-mode? |
| 21:29:50 | reactormonk | I'd suggest TS, but it's a bit too noisy here |
| 21:30:04 | pigmej | reactormonk: 23:30 there ;/ |
| 21:30:14 | reactormonk | sounds like a good time to code |
| 21:30:22 | pigmej | my brain is currently in mode 'tomorrow is f*** monday' |
| 21:30:38 | pigmej | reactormonk: but what's wrong? company issues / |
| 21:30:39 | pigmej | ? |
| 21:30:53 | * | Matthias247 quit (Read error: Connection reset by peer) |
| 21:31:19 | reactormonk | pigmej, I can't poke it to start anything |
| 21:31:37 | reactormonk | since nimsuggest now works, could you give it a try? |
| 21:32:38 | pigmej | k, but tomorrow, I had to prepare for tomorrow ;/ new week, new month ... |
| 21:32:47 | pigmej | reactormonk: do I still need your fork / |
| 21:32:49 | pigmej | |
| 21:32:50 | pigmej | ? |
| 21:32:55 | reactormonk | pigmej, nope |
| 21:32:59 | reactormonk | it's been fixed. |
| 21:33:02 | pigmej | ok so all merged :) |
| 21:33:06 | reactormonk | yup |
| 21:33:08 | pigmej | only nim-mode from epc branch? |
| 21:33:40 | HOLYCOWBATMAN | hey, i care about windows! |
| 21:33:52 | reactormonk | pigmej, yup, exactly. |
| 21:33:58 | dtscode | ^ Windows lag at its finest |
| 21:34:06 | HOLYCOWBATMAN | lol |
| 21:34:22 | reactormonk | ouch. |
| 21:34:38 | pigmej | dtscode: rotfl |
| 21:34:41 | * | HOLYCOWBATMAN is now known as Windows |
| 21:34:42 | reactormonk | wow, we're nearing #3k in nim |
| 21:35:14 | Windows | at least i got better games than pong |
| 21:35:24 | Windows | #rekt |
| 21:36:06 | * | Windows is now known as HOLYCOWBATMAN |
| 21:36:15 | dtscode | Who needs games better than pong? |
| 21:36:37 | pigmej | dtscode: a gamer who uses windows ... probably ;] |
| 21:36:38 | HOLYCOWBATMAN | pong per email |
| 21:36:43 | * | dom96 considers making NimBot kick whoever says #rekt :P |
| 21:37:03 | reactormonk | #shrekt |
| 21:37:04 | * | dtscode kicks dom96 |
| 21:37:09 | dtscode | #r_ekt |
| 21:37:17 | dom96 | #r-ekt |
| 21:38:18 | HOLYCOWBATMAN | r = ek/t |
| 21:38:33 | pigmej | ;d |
| 21:38:34 | dtscode | #star-trekt |
| 21:38:58 | pigmej | reactormonk: https://github.com/reactormonk/nim-mode/blob/epc/company-nim.el this makes sense or not at all? |
| 21:39:11 | dom96 | #t-rext |
| 21:40:21 | vikaton | #r #e #c #t |
| 21:40:33 | vikaton | oh wow i misspelled |
| 21:40:49 | dtscode | oh no |
| 21:40:56 | * | dom96 joined #nim |
| 21:40:56 | * | dtscode joined #nim |
| 21:40:58 | flaviu | :P |
| 21:41:01 | dtscode | listen here fucker |
| 21:41:07 | dom96 | oh dear |
| 21:41:15 | dom96 | Somebody will get their +o powers taken away. |
| 21:41:24 | * | flaviu1inch joined #nim |
| 21:41:27 | dtscode | I'll have you know I am a part of 30+ succefull nimperations |
| 21:41:44 | flaviu1inch | T_T |
| 21:41:59 | flaviu | I pasted in 5 kicks at the same time, I guess freenode didn't like that. |
| 21:42:24 | dom96 | how many people did you actually kick? |
| 21:42:27 | dtscode | > flaviu1inch |
| 21:42:35 | dtscode | If you know what I mean |
| 21:42:38 | * | vikaton joined #nim |
| 21:42:38 | * | flaviu1inch is now known as HOLYCOWBATMAN |
| 21:42:48 | dom96 | somebody needs to patch NimBot so that it logs kicks |
| 21:42:55 | flaviu | HOLYCOWBATMAN, dom96, dtscode, vikaton |
| 21:42:57 | dom96 | anybody? |
| 21:42:58 | dtscode | I'm on it |
| 21:43:01 | dom96 | anybody at all |
| 21:43:01 | vikaton | what |
| 21:43:02 | dom96 | yay |
| 21:43:19 | vikaton | are u going to apologize? |
| 21:43:25 | vikaton | because i wont accept it :[ |
| 21:43:31 | flaviu | Nope :B |
| 21:43:35 | dom96 | vikaton: aww, i'm sorry for flaviu |
| 21:44:02 | dtscode | wheres the nim bot repo again/ |
| 21:44:16 | pigmej | dtscode: github ;p |
| 21:44:29 | dom96 | dtscode: the internet |
| 21:44:33 | pigmej | dtscode: dom puts all his stuf on github ;p |
| 21:44:40 | dom96 | sorry, I couldn't resist. |
| 21:44:42 | dtscode | Did you know the internet is on computers now? |
| 21:44:51 | dom96 | dtscode: https://github.com/nim-lang/nimbot |
| 21:45:04 | vikaton | dtscode: its on them? |
| 21:45:11 | vikaton | that sounds ugly |
| 21:45:25 | dtscode | vikaton: The internet (stacks) typically are |
| 21:45:40 | HOLYCOWBATMAN | #53454B54 |
| 21:45:49 | * | dtscode wonders if he can just edit it straight in github |
| 21:46:00 | dom96 | dtscode: yeah... no. |
| 21:46:05 | dtscode | nmd I need to test it |
| 21:46:41 | dom96 | yeah, please make it join #nim-offtopic for testing |
| 21:46:50 | dom96 | or better yet a channel you own |
| 21:47:06 | vikaton | flaviu: u know when ur gonna start the syntax highlighting thingamejig? |
| 21:47:08 | pigmej | reactormonk: btw what was the issue ? |
| 21:47:32 | flaviu | vikaton: Some time in the next 10 years. |
| 21:48:04 | vikaton | flaviu: i may be able to halp |
| 21:48:11 | dtscode | I have far too much of an unhealthy interest in IRC bots |
| 21:48:49 | flaviu | vikaton: Just be patient, I'll get around to it. |
| 21:48:51 | dom96 | dtscode: I do too. Let's form a club. |
| 21:49:17 | vikaton | Ok |
| 21:49:18 | dtscode | :D |
| 21:49:43 | dom96 | whoa |
| 21:49:51 | * | strcmp1 quit (Quit: Leaving) |
| 21:49:51 | dom96 | This PR implement's Go's GC in Nim? https://github.com/Araq/Nim/pull/2851 |
| 21:50:38 | dtscode | I need some music then I'll get started. What format should the kicks look like? Just "foo has kicked bar from #baz (Because he is not dtscode)" ? |
| 21:50:58 | Araq | holy shit |
| 21:51:07 | dom96 | Araq: I know right? |
| 21:51:09 | Araq | so we have Go's GC for Nim?! |
| 21:51:24 | dtscode | Does Go have a good GC? |
| 21:52:12 | Araq | they claim it's getting excellent |
| 21:52:20 | * | BitPuffin|osx quit (Ping timeout: 264 seconds) |
| 21:52:26 | pigmej | Araq: they claim a lot of things... |
| 21:52:32 | Araq | lol |
| 21:52:34 | ozra | pigmej: is nanomsg in a usable state now? I followed it for a while, having used ZMQ alot... You've used it a bunch? |
| 21:52:35 | Varriount | O_o |
| 21:52:48 | pigmej | ozra: depends what do you mean by 'usable' |
| 21:53:03 | pigmej | for me zmq was not usable at all (specific needs) |
| 21:53:05 | ozra | stable... |
| 21:53:10 | Araq | dtscode: I honestly don't know. It used to be a joke. |
| 21:53:23 | dom96 | This implementation is quite short. |
| 21:53:28 | * | BitPuffin|osx joined #nim |
| 21:53:29 | dtscode | Araq: Most of Go is a joke |
| 21:53:31 | dtscode | :D |
| 21:53:38 | ozra | pigmej: ok, if zmq did "pretty good" for my needs.. |
| 21:54:08 | pigmej | ozra: the thing is nanomsg is really simple (implemenatation / idea) |
| 21:54:12 | dom96 | oh this uses a DLL |
| 21:54:19 | dom96 | cool! |
| 21:54:34 | Varriount | dtscode: Now now, lets not be mean. |
| 21:54:39 | dtscode | :D |
| 21:54:41 | pigmej | ozra: there are some gotchas, but I haven't used it in very big project yet |
| 21:55:01 | pigmej | probably the current one with python transmist like 10M events /day |
| 21:55:13 | ozra | pigmej: yeah it was gonna support pluginable arch.. did that land? |
| 21:55:22 | ozra | pigmej: I'll get up to date via the web, save you the trouble :) thanks for reminder.. |
| 21:55:38 | pigmej | ozra: hard to say for me, I'm not that good in C |
| 21:55:49 | Araq | Varriount: I think we are overdoing this PC nonsense. calling something a joke must remain possible. |
| 21:56:06 | * | intra quit (Ping timeout: 258 seconds) |
| 21:56:27 | pigmej | but protocol is quite simple |
| 21:56:38 | ozra | Go = fucking joke + 1 hehe |
| 21:56:38 | pigmej | the protocol protocol ;) |
| 21:56:55 | pigmej | well, go has a lot of tension... a lot of hype |
| 21:57:02 | pigmej | a lot of buzzwords ,a lot of marketing |
| 21:57:04 | dom96 | Guys, let's keep the language bashing down low. |
| 21:57:11 | pigmej | and a lot of money ;-) |
| 21:57:13 | reactormonk | pigmej, I can't get company to suggest me anything |
| 21:57:27 | ozra | a lot of money can get you alot of crap |
| 21:57:32 | dtscode | I should really fork nim-lang/irc one of these days |
| 21:57:37 | pigmej | reactormonk: ok, I once succededd with some of my backends, so I will try then tomorrow ;) |
| 21:57:38 | flaviu | It's not fair to classify something as a joke. There are several interesting aspects of Go, even if they've made poor decisions in other areas. |
| 21:57:48 | dom96 | dtscode: why? It's perfect? :P |
| 21:57:50 | pigmej | ozra: sure, but 99% will not notice it |
| 21:58:02 | dtscode | dom96: To have my own copy for reference |
| 21:58:03 | reactormonk | pigmej, ok. The epc code should work as is. |
| 21:58:12 | dom96 | precisely what flaviu said. |
| 21:58:16 | dtscode | It is perfect :D. Ruined IRC for me in other languages |
| 21:58:24 | dtscode | I'm trying to write a c++ lib similar to it |
| 21:58:35 | dom96 | I've written an IRC lib in many languages now. |
| 21:58:42 | dtscode | <flaviu> It's not fair to classify something as a joke. There are several interesting aspects of Go, even if they've made poor decisions in other areas. |
| 21:58:46 | dom96 | and over the years I have become quite efficient at it ;) |
| 21:58:47 | ozra | pigmej: they probably spent 99% of the money of marketing ;) |
| 21:58:50 | dtscode | Sorry :3 I just really dislike go |
| 21:59:12 | dtscode | dom96: I wasn't suggesting I want to change it :D. Although I might write some documenation for it |
| 21:59:14 | Varriount | dtscode: Like me and Java? |
| 21:59:38 | dtscode | Varriount: probably :D. Java, php, and go I really don't like |
| 21:59:42 | reactormonk | dtscode, I think the idea of the go GC in nim is awesome. I just don't know about more information on seqs |
| 21:59:51 | dtscode | I bash ruby though just because its not python |
| 21:59:55 | ozra | Gwooaa. Java is _the_ worst! I'd snuggle with go anytime before touching that.. |
| 22:00:13 | reactormonk | ozra, guess why I didn't major in computer science |
| 22:00:24 | * | Varriount throw an AbstractFactoryBeanFactory at ozra |
| 22:00:28 | Varriount | *throws |
| 22:00:34 | dom96 | reactormonk: what did you major in? |
| 22:00:46 | ozra | reactormonk: Congrats to not doing so! haha |
| 22:01:07 | ozra | Varriount: hahaha. ozra-mode shields me and re throws |
| 22:01:13 | reactormonk | dom96, computational linguistics |
| 22:01:31 | Varriount | Eh! It buuurrrns! |
| 22:01:33 | dom96 | reactormonk: My uni doesn't even do that. |
| 22:01:36 | dom96 | AFAIK |
| 22:02:08 | dtscode | Does a kick register as an EvMsg dom96 ? |
| 22:02:10 | ozra | Varriount: but it burns in a very verbose way! |
| 22:02:14 | * | anthgur joined #nim |
| 22:03:17 | dom96 | dtscode: yep |
| 22:03:34 | dom96 | dtscode: There is only EvMsg, EvDisconnect IIRC |
| 22:03:40 | dom96 | EvMsg has all the messages. |
| 22:03:46 | dtscode | EvReconnect :p |
| 22:03:55 | HOLYCOWBATMAN | hey java is nice |
| 22:03:57 | HOLYCOWBATMAN | https://github.com/EnterpriseQualityCoding/FizzBuzzEnterpriseEdition |
| 22:04:04 | dtscode | I figured it was. I just didn't want to go digging again to see |
| 22:04:22 | HOLYCOWBATMAN | a no-nonsense implementation of FizzBuzz |
| 22:04:23 | * | solidsnack quit (Read error: Connection reset by peer) |
| 22:04:55 | Varriount | HOLYCOWBATMAN: The problem with Java is that it tends to damage the brain. People I've met who work with it tend to look at everything as an object, and get confused when they see a language like Nim |
| 22:05:23 | HOLYCOWBATMAN | Varriount: lol just look at the code, you might see that it was sarcasm =) |
| 22:05:30 | Varriount | HOLYCOWBATMAN: I know. |
| 22:05:33 | HOLYCOWBATMAN | i cant stand java |
| 22:09:27 | * | anthgur quit (Quit: My Mac has gone to sleep. ZZZzzz…) |
| 22:10:17 | flaviu | 85ms load time :D |
| 22:12:02 | Varriount | flaviu: What protocol are you using? HTTPS1 or 2? |
| 22:12:43 | flaviu | Varriount: Looks like github pages uses HTTP1 |
| 22:13:19 | Varriount | flaviu: Hm. I remember reading somewhere that loading time decrease is non-linear with http1 |
| 22:13:20 | * | vendethiel quit (Ping timeout: 264 seconds) |
| 22:16:55 | dtscode | message = i.msg.nick & " has kicked " & i.msg.params[1] & " from " & i.msg.params[0] & " (" & i.msg.params[2] & ")" |
| 22:17:02 | dtscode | is that format fine dom96 ? |
| 22:17:32 | dtscode | <kicker> has kicked <kickee> from <channel> ([reason]) |
| 22:17:43 | dom96 | sure |
| 22:17:56 | Varriount | dtscode: What are you building? |
| 22:18:12 | * | vendethiel joined #nim |
| 22:18:52 | dtscode | Varriount: A dooms day device |
| 22:19:15 | Varriount | O_o |
| 22:21:59 | dtscode | o.O is jester not in nimble anymore? |
| 22:22:05 | dom96 | it is |
| 22:22:12 | * | gsingh93 quit (Ping timeout: 252 seconds) |
| 22:22:38 | dtscode | I can $ nimble install jester succesfully, but then $ nimble build says it can't open jestr |
| 22:23:39 | dtscode | Does nim c not look in .nimble? |
| 22:25:21 | dom96 | where are you trying to `nimble build` in? |
| 22:25:48 | dtscode | nimbot/ |
| 22:25:59 | dom96 | did you add jester to your .nimble? |
| 22:26:10 | dtscode | What do you mean? |
| 22:27:12 | dom96 | as a dependency |
| 22:27:34 | dom96 | although if you're building nimbot then it should already have it |
| 22:27:54 | dom96 | but it does not https://github.com/nim-lang/nimbot/blob/master/nimbot.nimble#L13 |
| 22:28:04 | dom96 | my bad |
| 22:28:09 | dom96 | add it in there |
| 22:28:13 | * | BlaXpirit quit (Quit: Quit Konversation) |
| 22:28:14 | dom96 | jester#head |
| 22:28:47 | dtscode | Thanls |
| 22:28:54 | dtscode | Not thanks, but Thanls |
| 22:29:22 | dom96 | ne prolamo |
| 22:30:10 | flaviu | Google doesn't really like nim-lang.org: https://developers.google.com/speed/pagespeed/insights/?url=http%3A%2F%2Fnim-lang.org%2F&tab=mobile |
| 22:30:59 | dtscode | Err can message[8 .. -2] be changed to message[8 .. ^2] and be semantically the same |
| 22:31:39 | Varriount | dtscode: What type is message? A string? |
| 22:31:50 | pigmej | anyway guys, cu! |
| 22:31:54 | dtscode | Varriount: I believe so |
| 22:31:57 | filwit | can anyone know if this is a bug with the cgen or a bug with my 'new' macro: https://gist.github.com/PhilipWitte/2d167c888f95e67f0292#file-concept-nim-L109 |
| 22:32:04 | vikaton | google doesnt like twitter |
| 22:32:17 | def- | vikaton: me neither |
| 22:32:33 | filwit | Ship.new(...) is invoking a 'new' macro which returns a block statement |
| 22:33:04 | flaviu | vikaton: choose one: https://richleland.github.io/pygments-css/ |
| 22:33:14 | flaviu | make sure to turn on mixed content on that page |
| 22:33:18 | filwit | i'm not sure if 'seq[T].add <nnkBlockStmt>` is valid AST or not |
| 22:33:31 | dtscode | oh well I'll just tias |
| 22:33:48 | vikaton | flaviu: monokai |
| 22:34:02 | * | boop is now known as boopisaway |
| 22:34:56 | dtscode | Great. nimbot doens't compile on 11 |
| 22:35:22 | Varriount | filwit: Can you actually make a seq[T].add <nnkBlockStmt> in non-synthesized code? |
| 22:35:36 | dom96 | dtscode: yes |
| 22:35:41 | dtscode | dom96: It does? |
| 22:35:46 | Varriount | dtscode: Nor does Nimble. |
| 22:35:49 | dom96 | you can |
| 22:35:57 | dom96 | change it to ^2 |
| 22:35:59 | dom96 | I mean |
| 22:36:21 | dtscode | Ah |
| 22:36:32 | dtscode | Varriount: Those damn nim devs :/ |
| 22:36:38 | filwit | Varriount: i'm not sure.. what I really want is to return an encapsulated expression.. but without the block node, i get 'item redefined' errors |
| 22:37:07 | dtscode | lib/core/macros.nim(300, 39) Error: ambiguous identifier: 'Response' -- use a qualifier |
| 22:37:16 | dtscode | Please tell I'm not going to have to fix my stdlib? |
| 22:38:25 | filwit | I'm not sure how to generate a hygenic expression without the blockStmt node |
| 22:40:46 | Varriount | filwit: Delegate to a template? |
| 22:41:39 | Varriount | filwit: What about moving the block outside the add operation? |
| 22:41:56 | filwit | yeah you're probably right about delegating to a template |
| 22:42:07 | dom96 | dtscode: httpclient.response vs. jester.response perhaps |
| 22:42:17 | dom96 | The compiler really needs a nicer error message here. |
| 22:42:19 | * | vendethiel quit (Ping timeout: 265 seconds) |
| 22:43:13 | filwit | Varrount: not sure what you mean by "move block outside of add".. if you notice, I'm calling Ship.new(..) a bunch above the line I highlighted and it's working just fine. Either way, I think delegating to a template might work... |
| 22:44:47 | dtscode | dom96: Thanks |
| 22:45:33 | dtscode | err where would the response be? |
| 22:45:48 | Varriount | Windows has a header called 'Unknwn.h' |
| 22:48:16 | flaviu | vikaton: Uh, adding a button to the code boxes is going to be almost impossible with the way krandowm works. |
| 22:48:39 | vikaton | oh no D: |
| 22:48:54 | vikaton | flaviu: what about inside the boxes? |
| 22:49:49 | flaviu | specifically, injecting code like https://i.imgur.com/0PUuPe1.png is impossible. |
| 22:50:17 | vikaton | o |
| 22:50:38 | flaviu | Adding a button to the footer should be possible, if a little less obvious. |
| 22:50:48 | jaco60 | grrrr |
| 22:50:51 | vikaton | better than nothing |
| 22:51:06 | * | vikaton hides behind flaviu from jaco60 |
| 22:51:21 | * | Jesin joined #nim |
| 22:52:33 | jaco60 | why this code doesn't compile ? http://sprunge.us/PVdJ (i've imported sequtils) |
| 22:52:53 | dtscode | because a seq isn't an int :p |
| 22:53:13 | jaco60 | foldr don't return a seq |
| 22:53:28 | jaco60 | but, wait... |
| 22:53:30 | dtscode | Oh sorry missed that |
| 22:53:36 | dtscode | Can you give us your error? |
| 22:54:37 | jaco60 | Error: execution of an external program failed |
| 22:54:47 | dtscode | odd |
| 22:55:15 | Varriount | Sounds like the C compiler is barfing. |
| 22:55:47 | jaco60 | the exact message is : Error: execution of an external program failed; rerun with --parallelBuild:1 to see the error message |
| 22:56:00 | vikaton | I need to listen to some frank sinatra right about now ._. |
| 22:56:06 | HOLYCOWBATMAN | foldr(toSeq(2..n),a * b) |
| 22:56:26 | HOLYCOWBATMAN | instead of toSeq(2..n).foldr(a * b) |
| 22:56:28 | Araq | filwit: sorry i'm too busy to help you |
| 22:56:42 | HOLYCOWBATMAN | jaco60:^ |
| 22:56:44 | jaco60 | HOLYCOWBATMAN, but a simple let fact = toSeq(2..12).foldr(a * b) works perfectly fine |
| 22:57:13 | HOLYCOWBATMAN | probably something to do with the return |
| 22:57:14 | Varriount | jaco60: It's a C codegen error. |
| 22:57:25 | Varriount | jaco60: Please report it. |
| 22:57:30 | HOLYCOWBATMAN | return foldr(toSeq(2..n),a * b) works |
| 22:58:24 | jaco60 | HOLYCOWBATMAN, yes... odd |
| 22:59:42 | HOLYCOWBATMAN | redeclaration of 'result_91009' with no linkage |
| 22:59:50 | filwit | Araq: no problem, i'm doing other things anyways.. will put that off for later |
| 22:59:51 | Varriount | jaco60: https://gist.github.com/Varriount/02507a25e72484ebdfb4 |
| 22:59:51 | jaco60 | yep |
| 23:00:09 | jaco60 | Varriount, same here |
| 23:00:23 | filwit | Varriount: I fixed it by using genSym and not returning a BlockStmt: https://gist.github.com/PhilipWitte/2d167c888f95e67f0292#file-concept-nim-L15 |
| 23:00:38 | filwit | Varrount: just showing you for completeness |
| 23:00:40 | Varriount | filwit: Yay! |
| 23:05:04 | * | solidsnack joined #nim |
| 23:08:27 | dom96 | http://forum.nim-lang.org/t/1279 |
| 23:09:58 | dtscode | What is a screencast? |
| 23:10:22 | dtscode | Also, where do I need to swap out HTTPclient.repsonse with Jester.response dom96 ? |
| 23:10:48 | dom96 | dtscode: dunno |
| 23:11:09 | dtscode | damn |
| 23:11:28 | * | dtscode wonders if he can checkout a v0.10 release |
| 23:12:37 | vikaton | wait |
| 23:12:45 | vikaton | which of those photos is Araq ? |
| 23:13:46 | dom96 | dtscode: Do you get any other messages from the compiler? |
| 23:14:01 | dtscode | dom96: yeah one sec |
| 23:15:12 | dtscode | dom96: https://bpaste.net/show/80587daeb6bf |
| 23:15:55 | dtscode | vikaton: http://www.oreilly.com/pub/au/6501 |
| 23:17:45 | dom96 | dtscode: Did you change the code? |
| 23:18:01 | dtscode | dom96: of Nimbot? sure |
| 23:18:09 | dom96 | how? |
| 23:18:16 | dom96 | what did you change I mean |
| 23:18:48 | dtscode | I added that MKick stuff to irclogrender.nim |
| 23:19:02 | dom96 | was it compiling before you added it? |
| 23:19:11 | * | dtscode shrugs |
| 23:19:14 | dtscode | I didn't try that |
| 23:19:17 | dom96 | ok... |
| 23:19:24 | dom96 | try removing 'httpclient' from the import line |
| 23:19:30 | dom96 | and then write from httpclient import nil |
| 23:19:41 | dom96 | then fix errors |
| 23:19:48 | dtscode | ok |
| 23:19:54 | Araq | import httpclient except foo, bar |
| 23:19:58 | dom96 | or that |
| 23:20:04 | dom96 | import httpclient except Response |
| 23:20:12 | dtscode | Ok thanks |
| 23:20:53 | * | vendethiel joined #nim |
| 23:21:02 | dtscode | Oh goody it compiled |
| 23:21:09 | dtscode | Does nimble have a -y option? |
| 23:23:12 | flaviu | vikaton: Pushed! |
| 23:23:35 | vikaton | :O |
| 23:23:52 | dom96 | dtscode: yes |
| 23:24:19 | vikaton | flaviu: Nice job ! |
| 23:24:22 | Varriount | Araq: MSDN... is huge. |
| 23:24:23 | dtscode | Odd now nimbot doesn't want to run |
| 23:24:37 | * | dashed quit (Quit: Connection closed for inactivity) |
| 23:25:32 | vikaton | dtscode: does it just not run, or compile error? |
| 23:25:39 | dtscode | vikaton: Not run :p |
| 23:25:57 | vikaton | dtscode: Yeah I got that a few times with my Nimbot |
| 23:26:14 | dtscode | err the nim-lang nimbot? |
| 23:26:16 | vikaton | I would logout then try again |
| 23:26:28 | dtscode | ? |
| 23:26:29 | vikaton | No just ysing the irc module in general |
| 23:26:36 | vikaton | using* |
| 23:26:43 | dtscode | That works fine. I use that all the time |
| 23:26:47 | dtscode | I'm talking about dom's bot |
| 23:30:32 | dtscode | dom96: Why does your bot not want to run with src/nimbot --il logs/ |
| 23:36:48 | dom96 | dtscode: gimme a sec |
| 23:37:05 | dtscode | one mississippi |
| 23:37:07 | dtscode | :D |
| 23:38:06 | * | kumul joined #nim |
| 23:39:38 | * | gsingh93 joined #nim |
| 23:42:15 | * | vendethiel quit (Ping timeout: 265 seconds) |
| 23:44:09 | * | dtscode quit (Remote host closed the connection) |
| 23:48:51 | * | solidsnack quit (Ping timeout: 272 seconds) |
| 23:49:20 | * | dtscode joined #nim |
| 23:53:45 | ozra | anyone here know off the bat: './koch tests' - does it compile the tests with '--lib:./lib'? Or does it use libs as found by paths? |
| 23:54:14 | dom96 | dtscode: ok |
| 23:54:23 | dtscode | You figured out the issue? |
| 23:54:31 | dom96 | nope |
| 23:54:36 | dom96 | but I have time to help you now |
| 23:54:36 | dtscode | damn |
| 23:55:06 | ozra | dom96: Do you now about above? ^ |
| 23:55:20 | dtscode | Ok. So the only things i have changed is IRClogrender.nim (to make it render kicks) and the channel that it joins in nimbot.nim |
| 23:55:34 | dom96 | ozra: not sure unfortunately |
| 23:55:46 | dom96 | dtscode: try adding some echos in open() |
| 23:55:48 | ozra | ok. |
| 23:55:53 | dom96 | see where it blocks |
| 23:55:54 | dtscode | I then build it successfully, and try src/nimbot --il logs/ but it just keeps spitting out that syntax call |
| 23:56:06 | dom96 | what does it spit out? |
| 23:57:15 | dtscode | dtscode@dtscode-305E4A-305E5A-305E7A:~/Desktop/projects/nimbot$ src/nimbot |
| 23:57:15 | dtscode | No IRC logs filename specified. |
| 23:57:54 | dom96 | specify the logs filename? |
| 23:58:02 | dtscode | I did, with --il |
| 23:58:38 | cazov | you specified a directory, not a filename. you could try --il logs/log.txt ? |
| 23:59:01 | dtscode | I tried a filename too |
| 23:59:09 | dtscode | (alhtough a directory is a file :p) |
| 23:59:40 | cazov | (fair point :p) |
| 23:59:44 | dom96 | you didn't in what you pasted |
