<< 03-08-2013 >>
00:45:08*rubino123 quit (Quit: Leaving)
01:17:27*DAddYE quit (Remote host closed the connection)
01:17:47*DAddYE joined #nimrod
01:18:58*DAddYE quit (Remote host closed the connection)
01:19:32*DAddYE joined #nimrod
01:23:53*DAddYE quit (Ping timeout: 245 seconds)
01:38:57*Associat0r quit (Read error: Connection reset by peer)
02:08:49*q66 quit (Quit: Leaving)
02:20:04*DAddYE joined #nimrod
02:26:14*DAddYE quit (Ping timeout: 240 seconds)
03:23:13*DAddYE joined #nimrod
03:29:35*DAddYE quit (Ping timeout: 245 seconds)
03:37:31*Raynes quit (Ping timeout: 240 seconds)
03:38:29*Raynes joined #nimrod
04:26:22*DAddYE joined #nimrod
04:32:38*DAddYE quit (Ping timeout: 240 seconds)
04:43:14*OrionPK quit (Quit: Leaving)
05:29:28*DAddYE joined #nimrod
05:36:11*DAddYE quit (Ping timeout: 264 seconds)
05:41:37*BitPuffin joined #nimrod
05:54:04BitPuffinAraq: Did you figure out what the problem was with the matrices? Why I couldn't do C.low etc
05:59:58*DAddYE joined #nimrod
06:29:41*wan_ quit (Quit: ChatZilla 0.9.90.1 [Firefox 22.0/20130625173050])
07:16:18*fowl quit (Ping timeout: 256 seconds)
07:17:30*Endy joined #nimrod
07:22:35*BitPuffin quit (Quit: No Ping reply in 180 seconds.)
07:22:53*BitPuffin joined #nimrod
07:27:44*fowl joined #nimrod
07:50:52*BitPuffin quit (Quit: No Ping reply in 180 seconds.)
07:54:10*Associat0r joined #nimrod
07:54:10*Associat0r quit (Changing host)
07:54:10*Associat0r joined #nimrod
07:55:38*BitPuffin joined #nimrod
08:16:07*Endeg quit (Ping timeout: 264 seconds)
09:08:13dom96hello
09:08:22Araqhi
09:09:06*DAddYE quit (Remote host closed the connection)
09:09:38*DAddYE joined #nimrod
09:14:49*DAddYE quit (Ping timeout: 276 seconds)
09:19:32BitPuffinhey dom96 and Araq
09:19:38dom96hey BitPuffin
09:19:45BitPuffinwhaddup?
09:20:35BitPuffinWhat should I do if the return type is none, one or more of many, many different things?
09:20:43BitPuffin(for the collada loader)
09:21:16Araqreturn a Json object
09:21:33BitPuffinAraq: You sure? isn't there a better nimrod native type?
09:22:15dom96Create your own custom type?
09:22:42BitPuffinwell
09:22:51BitPuffinI was thinking about creating a type for every thing
09:22:56BitPuffinTColladaMesh etc
09:23:14BitPuffinHmm
09:23:21BitPuffininheritance perhaps?
09:23:36BitPuffinthey could all be an object of TColladaResource
09:23:52BitPuffinso the return type could be an array of TColladaResources
09:23:54dom96yeah, or you could use an object variant.
09:24:20BitPuffinYeah that's true
09:24:41BitPuffindom96: I guess make an enum that is resourcetype
09:24:45dom96yeah
09:25:00BitPuffinyeah that should work out pretty nicely
09:25:04BitPuffinthanks
09:25:19BitPuffinanyways Araq I don't know if you saw what I wrote earlier
09:25:33BitPuffinactually maybe you could be of help here too dom96 :)
09:25:45BitPuffinhttps://bitbucket.org/TheLonelyByte/linagl/src/25fa332cee272a662115f94e60ed87cd81c20041/src/linagl/matrix.nim?at=tip
09:25:58BitPuffinI can't call high and low
09:26:03BitPuffinon the ranges
09:29:14AraqBitPuffin: it's not easy to fix
09:29:33Araqotherwise I would already have done it, I told you I looked at that code
09:30:10Araqtry TMatrix[T, R, C] = array [R, array [C, T]]
09:30:45Araqyou'd need to multiply the dimensions btw
09:30:57Araqunless I missed something
09:31:09BitPuffinAraq: awh :(, well I actually don't even want to do it that way, I guess it's better to wait for the generic values perhaps?
09:31:39BitPuffinAraq: and I don't know you well enough to know that just because you didn't say something it means that it's too hard to fix :)
09:31:42BitPuffinYeah
09:31:45BitPuffinmy bad
09:32:00BitPuffinrows + columns -.-
09:32:09BitPuffinsometimes you do idiotic things :D
09:32:15Araq0.. rows*columns-1
09:32:48BitPuffinAraq: the thing is that I don't want to use a multidimensional way because OpenGL is better at handling single dimenion
09:33:16BitPuffin:%s/way/array/g
09:37:36Araqopengl doesn't care
09:37:43BitPuffinI read that it did
09:37:44*[1]Endy joined #nimrod
09:37:47BitPuffinsomewhere
09:37:47Araqas long as it's linearly laid out in memory
09:37:55BitPuffinmaybe it was in the superbible
09:37:56Araqand that's what it is in nimrod
09:40:19BitPuffinhmm
09:40:24BitPuffinwell then maybe it's juuust fine
09:40:27BitPuffinWe'll see
09:40:57*Endy quit (Ping timeout: 264 seconds)
09:40:58*[1]Endy is now known as Endy
09:41:31BitPuffinthat's also a bit easier to work with
09:43:19BitPuffinAraq: the only problem is that when you are writing the matrix literal, it will look wrong in code, since opengl traverses row major rather than column major. So you have to write the matrix transposed, and I'd like to have that happen in the background
09:43:40BitPuffinI guess just make a constructor or something
09:53:01*q66 joined #nimrod
10:09:38BitPuffinis there a way to restrict a field to only be possible some of the variants
10:09:51BitPuffinlike if I have the variants a, b and c
10:10:08BitPuffinand in my object I want a field that can only be of kind b and c
10:10:13*DAddYE joined #nimrod
10:11:37dom96of b,c: field: Type, is this what you have in mind?
10:15:01BitPuffindom96: no I don't think so, unless I can also have a case inside the variant
10:15:14BitPuffinor actually that will probably not work either
10:16:54dom96object variants do get a bit annoying when you have a field which you want to be shared for two variants.
10:17:18*DAddYE quit (Ping timeout: 264 seconds)
10:18:49dom96in fact I would like for that to be improved, but Araq is busy enough as it is.
10:19:57*BitPuffin quit (Ping timeout: 264 seconds)
10:22:20*BitPuffin joined #nimrod
10:23:03*Endy quit (Ping timeout: 245 seconds)
10:27:50*BitPuffin quit (Ping timeout: 240 seconds)
10:35:53*BitPuffin joined #nimrod
10:37:07*Associat0r quit (Read error: Connection reset by peer)
11:11:22BitPuffindom96: Hmm, well I wonder what I should do
11:13:24BitPuffindom96: perhaps I'm worrying too much
11:14:19dom96Could you explain in detail how you want it to behave? Maybe I can help you better.
11:15:03BitPuffindom96: well so I'm creating a variant for all kinds of collada resources: mesh, animation and so on
11:15:20BitPuffindom96: and these tags can all contain some of the other tags available (probably at least)
11:15:43BitPuffindom96: but not all of the other tags, so I would like to restrict the kind of variant allowed in another variant
11:15:52BitPuffinif you grok what I mean
11:16:39dom96hrm.
11:16:46dom96So each resource can contain children.
11:17:19dom96But each type of resource's children can only contain certain types of resources...
11:17:53BitPuffindom96: yup, at least I think that's the way it is
11:18:26dom96hrm, that's a toughie
11:20:14BitPuffinyep
11:20:54dom96how would it be done in other languages?
11:21:50dom96perhaps it's simply impossible to do at compile-time.
11:21:57dom96or rather, with the type system.
11:22:12dom96And you will have to check it at runtime.
11:22:18BitPuffindom96: well I'm not sure how it would be done in a statically typed language, that's the problem, the only thing I can think of is casting
11:22:20BitPuffinbut that's fugly
11:22:48BitPuffineither that or some odd template voodo probably :P
11:22:59BitPuffinI wonder how assimp does it
11:23:11dom96yeah, I think template voodoo is your best bet.
11:23:41BitPuffinnimrod templates?
11:23:51*Endy joined #nimrod
11:23:59BitPuffinI can't seem to figure out how they would help though
11:23:59dom96yeah, perhaps some 'add' template which checks the types for you
11:24:07BitPuffinmaybe they can build a datastructure or something
11:24:41dom96i'm not sure if you can check the field of your object variant which determines the type of your resource at compile time though
11:25:48BitPuffinno you can't
11:26:30BitPuffinargh
11:26:36AraqBitPuffin: you need a component system, both fowl and zahary worked on one but both disappeared
11:26:55Araqthey'll come back though
11:26:57BitPuffinAraq: a component system you say
11:27:36dom96Well this works: https://gist.github.com/dom96/aeabe322d4ee42b56d9b
11:28:43Araqdom96: that's a nice solution
11:29:40BitPuffinTBleh :D
11:29:46dom96:)
11:32:29BitPuffindom96: Hmm, so you think I should do when blah != bleh {.fatal.} for example?
11:32:43dom96yeah
11:32:58BitPuffinBut hmm
11:33:04BitPuffinit still can't be known at compile time
11:33:18BitPuffinmaybe it will work though
11:34:02BitPuffinI'll give it a shot
11:51:05BitPuffinbut child.kind is not known at compile time
11:51:11BitPuffinwe'll see what happens
11:59:55*Endy quit (Ping timeout: 276 seconds)
12:16:11*Boscop joined #nimrod
12:36:23*Endy joined #nimrod
12:48:42dom96Aww. I tried connecting 10K clients to my local server and got a 'System error [EOS]'. On the bright side it's the first i've seen this error, so achievement unlocked! :D
12:53:28BitPuffindom96: dawh
13:15:00*DAddYE joined #nimrod
13:21:40*DAddYE quit (Ping timeout: 268 seconds)
13:22:39NimBotAraq/Nimrod asyncmacro 725ea68 Dominik Picheta [+2 ±1 -2]: Implemented 'await connect' fully. Fixed a buffering and 'reqReg' bug.... 9 more lines
13:23:35dom96whew
13:23:57BitPuffinwoho!
13:31:44*rubino123 joined #nimrod
14:00:06*Endy quit (Read error: Operation timed out)
14:11:54*BitPuffin quit (Ping timeout: 264 seconds)
14:39:46*Endy joined #nimrod
15:07:05*___ joined #nimrod
15:07:23*___ quit (Client Quit)
15:42:30*Endy quit (Ping timeout: 264 seconds)
16:19:04*DAddYE joined #nimrod
16:23:49*DAddYE quit (Ping timeout: 276 seconds)
16:27:39*OrionPK joined #nimrod
16:29:40*Endy joined #nimrod
16:31:35*Associat0r joined #nimrod
16:31:35*Associat0r quit (Changing host)
16:31:35*Associat0r joined #nimrod
16:32:18*XAMPP joined #nimrod
16:43:39*Endy quit (Quit: HydraIRC -> http://www.hydrairc.com <- Nine out of ten l33t h4x0rz prefer it)
17:06:25*Trix[a]r_za is now known as Trixar_za
17:14:24*Mat2 joined #nimrod
17:14:26Mat2hello
17:15:25dom96hey Mat2
17:19:46*DAddYE joined #nimrod
17:25:11Mat2hi dom96
17:26:16*DAddYE quit (Ping timeout: 264 seconds)
17:34:54*silven quit (Ping timeout: 264 seconds)
17:40:16*silven joined #nimrod
17:52:23*rubino123 quit (Ping timeout: 268 seconds)
18:00:22*Trixar_za is now known as Trix[a]r_za
18:04:53*rubino123 joined #nimrod
19:14:06dom96Mat2: So did you test Aporia on your RPI?
19:15:56NimBotAraq/Nimrod master 853965f Araq [+0 ±7 -0]: new VM: next steps
19:17:16Mat2dom96: no, at moment building a suitable os configuration. After that's finished I can compile the sources (this will take some time, I espect 2-4 hours). I thin somewhere in 4 hours I can tell you the results
19:17:29Mat2^think
19:17:51dom96alright. Thanks.
19:20:53*NewGuy joined #nimrod
19:21:08NewGuyHowdy, fellas.
19:21:53Mat2hi NewGuy
19:22:27dom96hello NewGuy :)
19:22:56Mat2ehm, what mean fellas ?
19:23:15dom96Same as 'guys'
19:23:24Mat2ah
19:24:16dom96NewGuy: What brings you here?
19:31:22NewGuySorry, about that, back.
19:32:26NewGuyAnywho - I'm new to Nimrod (thus the name), and I'm going to go grab the compiler now. But I was wondering if you recommend using the 0.9.2 that has the easy installer, or just go ahead and recompile from the git, say, weekly?
19:33:38dom96best to get it from git to get all the newest bug fixes and features.
19:34:02dom96Pretty much everyone targets the compiler from git, including me.
19:34:21NewGuyOh okay, well that answers that!
19:34:55NewGuyThanks for the quick reply, Dom.
19:35:01dom96np
19:37:05*Mat2 debian running, next step: configuration, dev -packs installation (+ nimrod, ide)
19:37:17dom96:D
19:44:12*exhu joined #nimrod
19:45:32exhuhey Araq, what's the intent behind enforcing new code style and case-sensitivity? Is it that many people have whined about pascal-style naming?
19:47:50dom96I believe that is it indeed. People are scared by the T/P prefixes.
19:55:46Mat2well, I can still not understand the reason for this
20:01:43Mat2however, all development dependencies seem to be quite large for this little board. I would prefer crosscompiling for the future
20:09:17Araqexhu: meh ... actually the new way is pretty much the same as the old way really
20:09:50Mat2hi Araq
20:09:53Araqbut I hope it's much better for marketing
20:10:10Araqhowever, it breaks pretty much every single line of code
20:12:58Araqbut don't worry
20:13:32Araqeither it stays as it is or we're gonna get close to 100% perfect translation to the new style
20:17:09Araqexhu: indeed the primary reason is that people consider the prefixes archaic
20:18:00Araqthe problem is that the T/P also really help, so I'm torn
20:18:14exhuAraq, so should i write my lib and new bindings according to the new style right now?
20:18:25Araqhell no
20:19:18exhubut how old code can be run if std lib changes?
20:19:56*Mat2 now compiling nimrod
20:20:31Araqok, I'll reply on the forum ;-)
20:20:36Araqbut here is the short version:
20:21:16Araqcurrently Nimrod is SI so it can be translated to any style we pick
20:21:33Araqthis is only a cosmetic change so nothing breaks
20:22:49Araqafter everything adheres to the style, we can make the language full CS or - like I prefer - only CS for the first character which keeps most of the original flexibility but allows 'var a: A'
20:23:48Araqafterwards we can introduce aliases like 'type TServer {.deprecated.} = Server' and get rid of T/P
20:24:01*DAddYE joined #nimrod
20:24:45*NewGuy quit (Ping timeout: 250 seconds)
20:25:30Araqthat said, 'var a: A' already works with the current language so we could also simply decide to ditch T/P for new code
20:31:04*DAddYE quit (Ping timeout: 264 seconds)
20:31:20Mat2dom96: I espect compiling nimrod takes ~2 hours
20:32:13dom96Perhaps. The RPI is kinda slow. I don't think it took that long when I compiled it on the RPI though.
20:33:08Mat2I hope so :)
20:33:57Araqexhu: however the proposed style is more frictional wrt the original C names, so I would be interested how it works out for you with X11 and the WinAPI etc.
20:34:44exhuAraq, i'm afraid it'll be a hell to support bindings if they adhere to the new style
20:35:54Araq proc FindFirstFileW*(lpFileName: widecstring,
20:35:55Araq lpFindFileData: var TWIN32_FIND_DATA): THANDLE {.
20:35:57Araq stdcall, dynlib: "kernel32", importc: "FindFirstFileW".}
20:36:00Araqvs.
20:36:34Araq proc findFirstFileW*(lpFileName: Widecstring,
20:36:35Araq lpFindFileData: var Win32FindData): Handle {.
20:36:37Araq stdcall, dynlib: "kernel32", importc: "FindFirstFileW".}
20:38:32Araqyears of different programming languages and styles made me dyslexic ... I'm pretty bad at deciding these things :-)
20:38:57exhuAraq, btw Widecstring or WideCstring ? -)
20:39:18Araqindeed
20:39:35Araqand what's worst: you have both and they mean different things
20:40:15exhuAraq, i prefer to use original style for bindings, because of easy search through files. But not sure if it will be a problem
20:40:45exhuAraq, also i wonder how GTK bindings will look like -)
20:40:52dom96I think it makes sense to go for WideCString in that case.
20:40:57exhuAraq, c2nim will need conversion too.
20:41:04dom96'C' is separate word in that case
20:41:20exhudom96, so goes for TCP -)
20:41:47dom96Araq already said that it should be Tcp :P
20:42:13exhudom96, sorry, TCP is abbreviation and C is not, my fault
20:42:23dom96indeed
20:42:49exhuGL_STENCIL_BUFFER_BIT -> GlStencilBufferBit ?
20:43:02Araqof course now we all know WideCstring should be an exported identifier whereas wideCstring is a private one
20:43:18dom96Araq: huh?
20:43:39Araqthat's very useful to know because obviously bugs are never in public symbols and private/public is a totally stable criterium ...
20:43:55*Araq is making fun of Go again
20:45:44dom96i see...
20:46:30Araqexhu: true c2nim needs to know about the One True Style
20:46:50Mat2"One True Style" *g*
20:46:59exhuAraq, it's so much work
20:47:00dom96As sanctioned by the BDFL.
20:50:44Araqexhu: as I said the GL_STENCIL_BUFFER_BIT -> GlStencilBufferBit would be automated
20:51:27Araqand maybe it should be StencilBufferBit
20:51:27*Mat2 nimrod compiled without problems, now compiling the IDE
20:51:58dom96good, it didn't take 2 hours :)
20:53:53Mat2hmm, I think my old 8-bit commodore system compiles faster (LUnix) but however, it works
20:54:12exhuAraq, on my home LCD GL_STENCIL_BUFFER_BIT is more readable than StencilBufferBit -)
20:54:47exhui use small fonts -)
20:56:31Mat2(^little joke)
20:57:20AraqI don't think so, exhu
20:57:43AraqOTHERWISE_YOU_WOULD_HAVE_CODE_LIKE_THIS_EVERYWHERE
20:57:52Araqbut you don't, only for constants
20:58:08Araqwhat you mean it's more readable because you're used to this name
20:58:55exhuAraq, yes, i got used to it, it's the same for all opengl books for the last ten years :)
20:59:43Araqyeah so the argument "it doesn't matter there is a rule to translate from the C name to the Nimrod name" is not valid
21:00:37Araqthat's what I feared
21:01:11Araqso enforcing things to look the same everywhere only works if we don't consider external libraries
21:01:48exhuAraq, exactly, that's why i feel uncomfortable with the new style enforced on bindings as well.
21:02:57Mat2in addition, there exist multible style conventions dependent on the library (and the preferences of there programmers)
21:03:27exhuAraq, however, given a rule, there will be some time to adopt and then the bindings users/writers will get accustomed to.
21:04:45exhuAraq, there're already some exceptions in bindings, like functions to access unions or field names which are keywords in nimrod etc.
21:05:50Araqthe Golang OpenGl wrapper uses ALL_CAPS for the constants too
21:06:21Araqfortunately they are all exported anyway ;-)
21:06:45Araqit strips the GL_ prefix though
21:06:56Mat2*wah*
21:07:22AraqT = 0x2001 # yeah, nice ...
21:07:27Araqwhat's GL_T ?
21:13:56exhuAraq, GL_T maybe about texture wrapping
21:14:58Araqor about matrix transpose?
21:15:57*Mat2 now cloning Aporia
21:16:21exhu"Here is how to specify texture coordinate generation :
21:16:21exhu coordinate is the component of the texture coordinate ie GL_R, GL_S, GL_T or GL_Q."
21:18:09*exhu quit (Quit: Ex-Chat)
21:18:38AraqI wish people would at least say "bye"
21:21:51dom96agreed
21:30:09Mat2dom96: compiling Aporia resulting in a SIGSEGV exception, the nimrod compiler itself works fine
21:31:27dom96Can you bootstrap in debug mode and show us the traceback?
21:31:34Araqso the compiler crashes? or aporia at startup?
21:32:09Mat2nether the compiler nor aporia but debian is the problem here
21:32:56Mat2I think the memory usage is just too high (and this is possibly a bug of the memory-manager)
21:34:08dom96create a swap file
21:34:32Mat2already done
21:35:15Mat2I will try compiling from my external hd
21:37:03*NewGuy joined #nimrod
21:37:21NewGuySo I just finished building the compiler from git.
21:37:55NewGuyWhen I go about updating it again, what would be the steps?
21:37:55dom96cool
21:37:58NewGuyShould I just rinse and repeat?
21:38:16dom96yeah, git pull && ./koch boot
21:38:41NewGuySimple and sweet, thanks my friend.
21:39:31Araqwell ... sometimes you have to build from C
21:39:32Mat2the exception is caused compiling line 120 of search.nim
21:42:04AraqMat2: that doesn't help
21:44:00Mat2I stor the traceback
21:44:04Mat2^store
21:46:33AraqMat2: what's the CPU of the rasp?
21:47:53Mat2an ARMv6 SoC from broadcom (BCM2835)
21:48:07Mat2the traceback is quite short:
21:48:17Araqwhat does 'nimrod' say?
21:49:29Mat2nothing, only one warning about deprecated OSWError and two hints
21:49:37Mat2^OSError
21:50:00Araqno sorry
21:50:10AraqMine says:
21:50:12AraqNimrod Compiler Version 0.9.3 (2013-08-01) [Linux: amd64]
21:50:23Araqwhat does yours say?
21:51:30dom96Mat2: Execute: nimrod -v
21:51:31Mat2Version 0.9.3
21:51:51Araqgah no
21:52:06Araq[Linux: amd64] is the interesting part
21:52:15Araqwhat CPU does it claim to be on?
21:52:42Mat2[linux: arm]
21:52:59Mat2(of course, I'm running it on the board)
21:54:46Mat2back to the traceback, one moment (I must typing from display):
21:56:08Mat2> Program received SIGSEGV, Segmention fault.
21:56:57Mat2> 0x0005e920 in typetostring_196020 ()
21:57:22Mat2have it, its a buffer overflow
22:06:36Mat2sadly I can not figure out what's the source of these typetostring function
22:18:33Mat2ok, I save a coredump
22:24:52Mat2get some sleep, ciao
22:25:14*Mat2 quit (Quit: Verlassend)
22:29:08*DAddYE joined #nimrod
22:34:21*DAddYE quit (Ping timeout: 264 seconds)
23:02:48*Trix[a]r_za is now known as Trixar_za
23:22:51*NewGuy quit (Ping timeout: 250 seconds)
23:28:34*NewGuy joined #nimrod
23:30:58*DAddYE joined #nimrod
23:31:36NewGuyAnyone know why Aporia keeps giving me "could not load: libglib-2.0-0.dll."?
23:32:20NewGuyCompiled it exactly as the git page said, install all dependencies - I have double checked and the GTK bin folder is definitely on the path, and the file exists.
23:32:46NewGuy(with the correct filename)
23:37:30*DAddYE quit (Ping timeout: 245 seconds)
23:39:12NewGuyAh, yes, and I'm sporting Windows 7.