| 00:00:55 | dom96 | All I can find in regards to the error I get is it happening to people who are trying to deploy it on heroku. |
| 00:07:09 | dom96 | good night |
| 00:07:15 | Araq | same here |
| 00:07:18 | Araq | good night |
| 00:12:15 | reactormonk | Araq, you know stdout isn't really defined in JS |
| 00:24:06 | * | q66 quit (Remote host closed the connection) |
| 01:01:11 | reactormonk | Araq, found a rather deep bug - isObj calls base_ID where base is usually defined without ID |
| 01:01:41 | reactormonk | wait, not so deep - let's see. |
| 01:07:17 | reactormonk | nah :-( |
| 07:16:19 | * | gour joined #nimrod |
| 08:34:30 | gour | if nimrod would fail for whatever reason, which language(s) would you recommend to use instead with the condition that their name does not start with C or J? |
| 09:16:26 | * | Trix[a]r_za is now known as Trixar_za |
| 11:30:32 | * | Trixar_za is now known as Trix[a]r_za |
| 11:51:17 | * | Araq_ joined #nimrod |
| 12:20:05 | Araq_ | gour: I haven't even started to really advertise it |
| 12:20:28 | Araq_ | and these questions are offensive in #nimrod :P |
| 12:26:13 | * | Araq_ quit (Quit: ChatZilla 0.9.90 [Firefox 19.0.2/20130307023931]) |
| 13:48:07 | gour | Araq: well, you yourself said that nimrod's busfactor dropped to 1.5 ;) |
| 13:51:23 | gour | i believe it's reasonable concern for starting non-toying project |
| 14:30:55 | * | Trix[a]r_za quit (*.net *.split) |
| 14:30:56 | * | Amrykid quit (*.net *.split) |
| 14:31:04 | * | Trix[a]r_za joined #nimrod |
| 14:31:05 | * | Amrykid joined #nimrod |
| 14:52:34 | * | q66 joined #nimrod |
| 15:57:10 | gour | Araq: are you going to advertise Nimrod before 1.0/2014? |
| 17:43:49 | * | avarus joined #nimrod |
| 17:43:57 | avarus | hi |
| 17:45:43 | dom96 | hey |
| 18:36:19 | * | avarus quit (Read error: Connection reset by peer) |
| 20:21:39 | * | exhu joined #nimrod |
| 20:21:56 | exhu | hi dom96 |
| 20:31:42 | Araq | hi exhu, maybe I can help you instead |
| 20:37:37 | * | Amrykid quit (Changing host) |
| 20:37:37 | * | Amrykid joined #nimrod |
| 20:42:12 | dom96 | hey exhu, what's up? |
| 20:42:55 | * | Zor_ joined #nimrod |
| 20:44:13 | * | Trix[a]r_za is now known as Trixar_za |
| 20:47:16 | * | Zor quit (Ping timeout: 246 seconds) |
| 20:50:19 | * | gradha joined #nimrod |
| 20:54:01 | gradha | dom96: I was going to search this in the source code but exhu beat me to asking it http://forum.nimrod-code.org/t/136 |
| 20:54:11 | exhu | dom96, have you seen a babel topic on the forum? |
| 20:55:33 | dom96 | It's possible that that functionality is not implemented yet |
| 20:56:05 | dom96 | In fact yeah, i'm quite certain it's not. |
| 20:57:20 | Araq | Babel is very bare bones and we can't agree on most design questions :P |
| 20:57:22 | exhu | i'm afraid of using babel yet because i can post an url to master branch and then push changes which will break. |
| 20:57:47 | gradha | that's a good motivation to improve babel |
| 20:57:49 | dom96 | Yeah, designing a package manager is quite hard. |
| 20:58:07 | Araq | and it doesn't help that I am quite opposed to package management in general ;-) |
| 20:59:59 | exhu | the idea is nice but i think you need a tar.gz etc to store snapshots/releases, not urls to git repositories. |
| 21:00:15 | gradha | can't you link to github's tagged tar.gz? |
| 21:00:37 | Araq | I don't dare to say it but ... |
| 21:00:38 | dom96 | exhu: Yes, that's what I think too. |
| 21:00:58 | Araq | somebody should really look at Go's package management and see if we should copy that |
| 21:01:23 | gradha | Araq: are they doing something <gasp> right? |
| 21:01:28 | dom96 | This packages.json and .babel file means the same data is found in two different places and it gets really messy. |
| 21:02:04 | dom96 | I think cabal is a better example. |
| 21:02:30 | dom96 | And my initial plan for babel was to base it off of cabal. |
| 21:02:51 | Araq | gradha: they *might* ;-) |
| 21:03:06 | dom96 | But then /someone/ suggested to make it work similar to Rust's package manager ... :P |
| 21:03:24 | Araq | the Rust people can only overengineer it |
| 21:03:58 | Araq | the Go people have a chance to arrive at the cheap solution |
| 21:04:03 | gradha | Araq: surely you jest https://gitorious.org/go-package-manager ? "A simple, but powerful package manager written in Bash and Python", just as nice as combining ice and fire |
| 21:04:59 | exhu | there's also an open question about possible packages that provide bindings together with C/c++ libs, whether to support that feature or not, like python c modules... |
| 21:05:53 | gradha | exhu: for simple C code I've ended up concatenating C files and embedding that in the nimrod code |
| 21:06:43 | exhu | gradha, no, i think about binding complex toolkits, like bullet physics etc. |
| 21:08:14 | exhu | unfortunately have to sleep now, bye |
| 21:08:27 | Araq | good night |
| 21:08:35 | gradha | wouldn't you take a similar approach then like gtk bindings? dynamic linking instead of static linking to allow easy nimrod compilation |
| 21:09:07 | * | exhu quit (Quit: Leaving) |
| 21:10:00 | dom96 | There are so many questions, and the answers to them are hard. |
| 21:10:46 | dom96 | I think our primary place of discussion for this should be the forums. |
| 21:11:53 | Araq | this takes too much time |
| 21:11:58 | gradha | is that go package manager official? can't seem to find references to it in the web page http://golang.org/search?q=package+manager |
| 21:12:14 | Araq | forum discussions are too slow |
| 21:12:37 | Araq | plus jokes always feel out of place on the forum ... |
| 21:13:34 | dom96 | Yes, but talking about it on IRC doesn't feel official. |
| 21:13:40 | dom96 | And the ideas always get lost. |
| 21:14:01 | dom96 | Well we can discuss it here, but please write the final agreed upon design on the forum. |
| 21:14:34 | gradha | finally, http://golang.org/cmd/go/#hdr-Download_and_install_packages_and_dependencies seems about the same as any other manager I've seen |
| 21:16:08 | * | Zerathul joined #nimrod |
| 21:16:49 | gradha | oh, just remembered |
| 21:17:09 | gradha | yesterday during sleep the divine hand of inspiration touched me briefly |
| 21:17:16 | gradha | I managed to note down an excellent idea for babel |
| 21:17:26 | gradha | you know, babel needs an icon too |
| 21:17:55 | gradha | so the icon would be a sort of translucent double helix which raises to reach a floating crown with a bright light |
| 21:18:13 | gradha | and this would be fine for a small icon version, but we need obviously a huge high resolution splash for retina display macs |
| 21:18:51 | gradha | because typing "babel" should display a splash image on macs for 10 secs |
| 21:18:51 | gradha | in the high resolution version you would be able to read small letters engraved in the double helix |
| 21:18:58 | gradha | one helix could have the names of the initial babel packages already available |
| 21:19:09 | gradha | the other could contain names of their developers, for bragging rights |
| 21:19:21 | gradha | and just to satisfy freaks, it could all be written in weird tolkien characters |
| 21:19:41 | gradha | there you go |
| 21:20:17 | gradha | I'm sure we could even fit a pink unicorn going up the helix |
| 21:21:15 | * | dom96 loves the idea |
| 21:21:39 | dom96 | the pink unicorn is obviously my favourite part |
| 21:22:10 | gradha | maybe the pink unicorn has already rised to the top and its horn is touching this brigh light emanating from the floating crown |
| 21:24:30 | gradha | plus the icon has hidden meanings to offend the greatest range of users: floating crown/light (religion), double helix (evolution), pink unicon/elfic runes (fantastic creatures |
| 21:25:31 | * | gour quit (Disconnected by services) |
| 21:25:31 | * | gour_ joined #nimrod |
| 21:31:03 | * | gour_ is now known as gour |
| 21:31:43 | gradha | Araq: something gour mentioned made me curious, what is your goal for Nimrod? do you have a measure of success at all for it? |
| 21:33:16 | Araq | I do but I don't want to discuss it here ;-) |
| 21:39:29 | gradha | I'm wondering if it would be possible to use something like http://www.txl.ca/nresources.html which has C/C++ grammars to generate nimrod bindings like swig |
| 21:42:12 | Araq | a C++ grammar doesn't help c2nim |
| 21:42:37 | Araq | you either write a swig-xml -> nimrod transformator |
| 21:42:46 | Araq | or improve c2nim to deal with C++ |
| 21:43:01 | Araq | anything else looks like much more work |
| 21:43:16 | gradha | isn't txl just like swig? found it reading on http://www.complang.org/colm/ |
| 21:44:33 | Araq | I don't think it matters |
| 21:44:46 | Araq | there are quite some libraries using swig already |
| 21:45:14 | Araq | so you can use and adapt existing swig files |
| 21:50:13 | Araq | gradha: one doesn't want to actually parse C++ for wrapper creation; you want to parse an adhoc language |
| 21:50:49 | Araq | so that #defines are kept |
| 21:51:10 | gradha | that's why txl or colm seem more suitable than swig, which seems specialized on C/C++ |
| 21:52:05 | Araq | specializing on C/C++ is the whole point as these are so common and yet so aweful |
| 21:55:45 | Araq | also you need to support adhoc rules like "pair of int length; T* data; means 'data' is an array really and not a pointer" |
| 21:56:14 | Araq | swig's type maps suppor that quite nicely iirc |
| 22:04:20 | gradha | this is possible https://gist.github.com/gradha/5138245, what was the loophole with object variants you are trying to solve through constructors? |
| 22:05:03 | gour | dom96: https://blogs.gnome.org/xjuan/2013/03/06/glade-drag-drop-support/ |
| 22:06:06 | Araq | gradha: the loophole is that any assignment to 'kind' that would trigger an object branch change is forbidden |
| 22:06:49 | Araq | but it cannot be forbidden for the assignment from TKind(0) because that's the default/start value |
| 22:07:10 | Araq | which creates the nasty loophole in the language |
| 22:07:40 | gradha | I don't understand, if I try to change the kind twice I get "assignment to discriminant changes object branch" |
| 22:08:29 | Araq | well that can also be a simple bug as I never tested nested variants |
| 22:09:22 | Araq | implementation bugs are unimportant in comparison to language design bugs :P |
| 22:09:58 | gradha | ok, so the language design bug is that you can change kind, and constructors are there to disallow this at compile time rather than runtime? |
| 22:10:26 | Araq | changing the kind is allowed and a feature |
| 22:10:34 | Araq | the compiler uses that feature heavily |
| 22:11:03 | Araq | changing the kind so that it would trigger a branch transition is what's forbidden |
| 22:12:54 | Araq | new(n); n.kind = tk1b; # forbidden! but then how can we ever create the tk1b branch? |
| 22:13:02 | Araq | --> constructors ftw: |
| 22:13:16 | Araq | n = TK1(kind: tk1b) |
| 22:15:09 | gradha | tried var test6: ref TK3; test6 = new(TK3); test6.kind = TK3B and it seems to work |
| 22:15:15 | gour | Araq: spent short time chatting with guys in #modula-2 about the status of R10..the main guy knows about nimrod but says "it's not my cup of coffee" and "the language is not as low-level as M2"...otherwise, they're quite 'interesting' personalities :-) |
| 22:16:46 | gradha | oh, the ref version allows me to change kind all the time, the non ref only once |
| 22:17:21 | Araq | gradha: it may work but it shouldn't :P |
| 22:17:35 | Araq | it's not safe in general |
| 22:18:08 | gradha | heh, gc crashed on me at nimGCunrefNoCycle when i tried to assign a variable of the second re-type |
| 22:18:57 | gradha | it's however nice I don't understand what I'm doing, so I'll just move to something else |
| 22:19:43 | Araq | gour: I know the modula-2 guy and guess what --- I know about nimrod than him and it's as low level as modula 2 in that you can write a GC in it |
| 22:19:48 | Araq | *know more |
| 22:20:14 | gradha | gour: are you looking for a lower level language to... cripple your productivity writing apps? |
| 22:22:46 | gradha | wow, that glade video is impressive, dragging and dropping is so '95 |
| 22:24:05 | gour | Araq: well, i replied to him there are untraced references in nimrod...and even those are too low-level for me :-) |
| 22:24:21 | gour | gradha: :-) |
| 22:26:32 | Araq | gradha: thanks for your test cases, should be helpful |
| 22:26:51 | gradha | test cases? what? |
| 22:27:00 | Araq | your variants.nim gist |
| 22:27:18 | gradha | good I made it public, it's not on my hd any more |
| 22:28:21 | * | gour --> sleep...bb (maybe) tomorrow evening |
| 22:28:33 | gour | 'night folks |
| 22:28:37 | * | gour quit (Quit: WeeChat 0.4.0) |
| 22:43:57 | gradha | so implementing the $ proc for a sequence of TRune, is it better if it was varargs/openarray? |
| 22:45:01 | Araq | varargs is completely out of the question |
| 22:45:35 | Araq | and openarray seems unnecessary |
| 22:57:46 | * | Zerathul quit (Quit: ChatZilla 0.9.90 [Firefox 19.0.2/20130307023931]) |
| 23:23:48 | * | gradha quit (Quit: bbl, have youtube videos to watch) |
| 23:53:23 | reactormonk | Araq, ping |
| 23:57:28 | Araq | reactormonk: pong |