<< 28-11-2019 >>
00:01:29*thomasross is now known as Guest50548
00:01:29*thomasross_ joined #nim
00:01:29*Guest50548 quit (Killed (card.freenode.net (Nickname regained by services)))
00:01:29*thomasross_ is now known as thomasross
00:06:03xaceif ext not in [ ".jpg", ".png" ]: # is there a way to use `not` in this scenario?
00:06:18disruptektry notin
00:06:33xacethanks, that worked
00:16:07*sentreen_ quit (Ping timeout: 268 seconds)
00:29:43*sentreen_ joined #nim
00:30:18*krux02_ joined #nim
00:34:17*krux02 quit (Ping timeout: 265 seconds)
00:44:02*krux02_ quit (Remote host closed the connection)
00:48:59*clyybber quit (Quit: WeeChat 2.6)
01:09:29*uu91 quit (Ping timeout: 250 seconds)
01:10:14*uu91 joined #nim
01:24:55FromDiscord_<Winton> who is encouraged to create a Emulator Server
01:28:21disrupteknot me.
01:28:58madpropsis there a needed web service missing on mim? Like what the playground, docs, etc are. I could help on that
01:29:03madpropsnim*
01:30:04zedeusmaybe you could help improve the nimble directory? https://github.com/FedericoCeratto/nim-package-directory
01:45:32disruptekif my name were martin, i'd be in the olympics.
01:45:39disruptekjust think about that.
02:04:29*ponyrider joined #nim
02:04:36*ponyrider quit (Client Quit)
02:25:04*sealmove quit (Quit: WeeChat 2.6)
02:31:47FromDiscord_<Xydium> I'm trying to learn Nim by writing a state-search library, but I'm not sure if my design (https://pastebin.com/rxg28B9n) is simple/efficient. Anyone have any advice?
02:32:45*waterthese joined #nim
02:33:26*waterthese left #nim (#nim)
02:34:26disruptekxydium: https://github.com/Xe/mapatei/blob/master/src/mapatei/fsm.nim
02:48:59*uu91 quit (Read error: Connection reset by peer)
02:49:13*uu91 joined #nim
03:11:36FromDiscord_<Rika> madprops, no I don't think so
03:39:18*endragor joined #nim
04:08:53*tklohna quit (Ping timeout: 276 seconds)
04:09:16*tklohna joined #nim
04:14:16*chemist69 quit (Ping timeout: 252 seconds)
04:14:19*uu91 quit (Read error: Connection reset by peer)
04:14:35*uu91 joined #nim
04:16:09*chemist69 joined #nim
04:18:25*tklohna quit (Ping timeout: 250 seconds)
04:19:56*tklohna joined #nim
04:31:25*rockcavera quit (Remote host closed the connection)
04:38:21*nsf joined #nim
04:38:51*tklohna quit (Ping timeout: 265 seconds)
04:39:23FromGitter<zacharycarter> @mratsim ooh good tip - thanks!
04:40:29*tklohna joined #nim
04:43:57*dddddd quit (Remote host closed the connection)
04:55:59FromGitter<s0kil> How could I tell `proc debug[T](x: T) {.importc: "console.log", varargs.}` to use cstring, so I don't get output like this when trying to print a string: ⏎ `[82, 101, 113, 117, 101, 115, 116, 105, 110, 103, 32, 82, 111, 117, 116, 101, 58, 32] "about"`
04:58:55FromGitter<s0kil> I'm using it as: `debug("Requesting Route: ", href)`
05:14:57*njoseph quit (Ping timeout: 240 seconds)
05:18:26*tklohna quit (Ping timeout: 276 seconds)
05:18:52*tklohna joined #nim
05:23:50*tklohna quit (Ping timeout: 268 seconds)
05:24:17*uu91 quit (Ping timeout: 276 seconds)
05:24:38*uu91 joined #nim
05:26:40*tklohna joined #nim
05:29:25*njoseph joined #nim
05:31:58*tklohna quit (Ping timeout: 245 seconds)
05:32:34*cherrypie joined #nim
05:32:44*cherrypie is now known as blueberrypie
05:34:35*bacterio joined #nim
05:37:24*tklohna joined #nim
05:45:57*tklohna quit (Ping timeout: 240 seconds)
05:46:23*tklohna joined #nim
05:48:58*Araq quit (*.net *.split)
05:48:59*acidx quit (*.net *.split)
05:49:00*gsingh93 quit (*.net *.split)
05:54:54*Araq joined #nim
05:54:54*acidx joined #nim
05:54:54*gsingh93 joined #nim
06:08:22*nif quit (Quit: ...)
06:08:29*ltriant quit (Quit: leaving)
06:08:31*nif joined #nim
06:12:28*narimiran joined #nim
06:14:49*nif quit (Quit: ...)
06:14:58*nif joined #nim
06:31:03*njoseph quit (Ping timeout: 268 seconds)
06:32:05*bacterio quit (Read error: Connection reset by peer)
06:37:05*njoseph joined #nim
06:37:30*bacterio joined #nim
06:42:27madpropstrying to figure out how to get an instance of Ipc_perm from the posix module
06:46:56*njoseph quit (Ping timeout: 265 seconds)
06:47:05*tklohna quit (Ping timeout: 268 seconds)
06:47:29*tklohna joined #nim
06:52:50*njoseph joined #nim
06:58:11*njoseph quit (Ping timeout: 268 seconds)
07:02:18*njoseph joined #nim
07:09:19*LargeEpsilon joined #nim
07:09:24*solitudesf- joined #nim
07:21:17*njoseph quit (Ping timeout: 276 seconds)
07:28:30*njoseph joined #nim
07:28:58*clyybber joined #nim
07:40:37*njoseph quit (Ping timeout: 240 seconds)
07:41:56*njoseph joined #nim
07:47:56*njoseph quit (Ping timeout: 276 seconds)
07:52:14*PMunch joined #nim
08:00:00*gmpreussner quit (Quit: kthxbye)
08:00:40*njoseph joined #nim
08:01:50FromDiscord_<itmuckel> Hey guys! When I use `glGetString(GL_VERSION)` I get a `ptr GLubyte`. How do I turn that into a string? 🤔
08:03:04*nif quit (Quit: ...)
08:03:14*nif joined #nim
08:05:03*gmpreussner joined #nim
08:06:23clyybbercast it to a cstring
08:08:43lqdev[m]var str = newString(openGlStringLen); copyMem(str[0].unsafeAddr, openGlString[0].unsafeAddr, openGlStringLen)
08:09:19lqdev[m]if the string is null-terminated, you can simply $cast[cstring](openGlString)
08:13:24lqdev[m]@itmuckel
08:17:03*LargeEpsilon quit (Remote host closed the connection)
08:17:23*LargeEpsilon joined #nim
08:20:52clyybberglGetString does return a null terminated string
08:21:10clyybberAlso tagging doesn't work from IRC to discord
08:21:41FromDiscord_<Clyybber> have to do it in discord itself
08:21:45FromDiscord_<Clyybber> @itmuckel
08:22:10FromDiscord_<itmuckel> 😲
08:22:49*njoseph quit (Ping timeout: 250 seconds)
08:23:36FromDiscord_<itmuckel> Thanks! Directly casting to cstring did the job. 👌
08:23:38Araqfix the wrapper already so that it returns a 'cstring'
08:29:41PMunchHmm, is there a way to make Nim generate code like "const unsigned char PROGMEM background[] = {1, 2, 3, 4}" as a top level statement?
08:32:45PMunchI have tried "let background {.codegenDecl: "const $# PROGMEM $#".}: array[10, cuchar] = [1.cuchar, 2.cuchar, 3.cuchar, 4.cuchar]" but that doesn't work as it generates the x variable as a top level statement with the correct type, and then another top level statement with "NIM_CONST tyArray TM = {1, 2, 3, 4}" and then in NimMainModule it does a nimCopyMem to copy one over to the other
08:32:52PMunchWhich of course is completely pointless
08:43:19Araqit's an important feature request, I'll do a live coding session about it
08:47:34ZevvPMunch: right, that is what I was suggesting yesterday.
08:48:02Zevvbut you would still need all the nonportable hassle to do memcpys from progmem to ram
08:48:11*ng0 joined #nim
08:48:40Zevvit would be pretty cool if nim could abstract that away, make some data type that allows mapping const objects on progmem and use the objects just as if the were in ram
08:51:38PMunchWait, you suggested this yesterday?
08:52:16PMunchFor the specific thing I wanted to do right now it's not a huge issue. I'm writing a macro that allows you to write "sprites" directly in your code for the Arduboy
08:52:41PMunchSo since it's a macro I can just manually generate the emit statement
08:55:24Zevvgood enough
08:55:43clyybberPMunch, Araq: How should it look like? `const {.addressable.}` or just a global let?
08:57:37Araqglobal let
08:58:42clyybberBut that means that global lets wont work at runtime anymore, would it?
09:00:26FromDiscord_<mratsim> @Xydium: https://nimble.directory/pkg/negamax
09:01:52FromDiscord_<mratsim> seems like it's really simple
09:06:37*floppydh joined #nim
09:07:58FromGitter<zacharycarter> mratsim: probably going to try out your union approach this evening - although it will be interesting to see how the union makes it to the shader from the vertex buffer
09:08:29FromGitter<zacharycarter> the whole problem is a bit weird - I'm setting up the vertex buffer attribute as four float values
09:09:03FromGitter<zacharycarter> I need the values to not be interpolated, so on the shader side of things I'm using vec4s with the flat attribute so no interpolation is performed per fragment / vertex
09:09:46FromGitter<zacharycarter> originally I had the varyings in the vertex / fragment shaders as ivec4s - but metal doesn't like the fact that the vertex buffer attribute is four floats, and that the varyings in the shader are ivec4s
09:10:08FromGitter<zacharycarter> so I tried changing the ivec4 to a vec4, and then manually constructing an ivec4 in the vertex shader using the values from the vec4 varying
09:10:18FromGitter<zacharycarter> but in the fragment shader, the values all come through as zeros for the ivec4
09:10:29FromGitter<zacharycarter> weird shaiba
09:13:51FromGitter<zacharycarter> and all this hooplah because bgfx doesn't support int32/uint32 vertex buffer attributes
09:15:20*fredrik92 quit (Quit: Client disconnecting)
09:18:44clyybberWhy not just open an issue on bgfx?
09:22:18FromGitter<zacharycarter> well - I'm in their gitter as well
09:22:39FromGitter<zacharycarter> I might just be doing something stupid - not sure if there's any actual issue or not, most likely it's me doing something stupid
09:23:07FromGitter<zacharycarter> how's the vulkan stuff going?
09:25:07clyybberWorking on something else ATM, but Kaynato has done some GPU compute stuff with my bindings lately
09:26:25FromGitter<zacharycarter> nice
09:27:01*luis_ joined #nim
09:31:58*bacterio quit (Read error: Connection reset by peer)
09:32:31PMunchHmm, how can I importc the variable I created with "emit"?
09:33:59PMunchSo my macro spits out: {.emit: "/*VARSECTION*/\nconst unsigned char PROGMEM mySprite[] = {4, 5, 233, 158, 144};".}
09:34:23PMunchAnd I want to be able to pass that as a ptr uint8 in my program
09:35:40FromDiscord_<mratsim> @Clyybber but now he is stuck in making Arraymancer or Arrayfire or any tensor library work on Cuda windows :/
09:36:40clyybbermratsim: That was the motivation to use vulkan compute
09:37:29*chemist69 quit (Quit: WeeChat 2.6)
09:38:24*chemist69 joined #nim
09:43:56*Vladar joined #nim
09:45:04*njoseph joined #nim
09:45:44PMunchHmm, is this not possible?
09:50:12FromGitter<zacharycarter> PMunch: https://play.nim-lang.org/#ix=231Z
09:50:30FromGitter<zacharycarter> apparently this works too: https://play.nim-lang.org/#ix=2320
09:50:59FromGitter<zacharycarter> https://nim-lang.org/docs/manual.html#implementation-specific-pragmas-emit-pragma - is where `NIM_EXTERNC` is described
09:52:11PMunchAhh
09:52:20PMunchHmm, it creates a new symbol though..
09:52:40PMunchWhich is why it was failing for me
09:53:18PMunchIt creates a "extern NU8* mySprite;" when I give it "var mySprite: ptr uint8"
09:53:30PMunchOops, with {.importc.} of course
09:53:49*njoseph quit (Ping timeout: 250 seconds)
09:54:16*jeyr joined #nim
09:55:33*luis_ quit (Ping timeout: 250 seconds)
09:55:52PMunchHmm, so I guess it's because I'm unable to create the correct type..
09:57:43*njoseph joined #nim
10:03:04PMunchvar mySprite {.codegenDecl: "#define $2 mySprite //".}: ptr uint8
10:03:22PMunchI this is my ugliest hack to date :P
10:03:52PMunch(By the way the // is to ignore the ; that Nim appends to these strings)
10:04:32PMunchIs it pretty? Hell no! Does it work? Surprisingly yes :)
10:05:59*jeyr quit (Quit: jeyr)
10:07:30*njoseph quit (Ping timeout: 268 seconds)
10:09:10*njoseph joined #nim
10:13:11FromDiscord_<mratsim> Did you feel the force of the SQL injection when you did this hack?
10:14:39*luis_ joined #nim
10:18:59madpropshow much does the GC affects performance realistically?
10:20:11PMunchmretsim, haha this should actually be fairly safe
10:20:42PMunchThe $2 is the only thing that is injected, and it is a Nim-obfuscated variable name
10:20:46FromDiscord_<itmuckel> Hey guys! Using `opengl 1.2.3` I get a black screen on Linux, but not on Windows. I got the latest mesa driver and context creation works and all. Anyone had the problem too? I use sdl2 for context creation, but everything works fine, like I said.
10:20:47FromDiscord_<itmuckel>
10:20:47FromDiscord_<itmuckel> Are there alternatives? I tried generating gl bindings with the `glad` webservice, but that `gl.nim` file gives me compiler errors for example for `GL_TIMEOUT_IGNORED*: uint64 = uint64(0xFFFFFFFFFFFFFFFF)`. 🤔
10:22:53Araqmadprops, it's almost always an issue and why I'm looking forward to --gc:arc so much
10:22:54FromDiscord_<itmuckel> Oh, the context is OpenGL 4.5
10:23:50Araqnot because --gc:arc is faster. because it gives us much of the control back that we need to
10:24:19Araqhave in order to optimize the code to the levels we want to
10:24:35madpropsnot sure what arc is
10:25:26madpropsbut im guessing some why to establish intelligent algorithms to make gc kick in at good times
10:25:35madpropsway*
10:25:54Araqno, it's refcounting optimized via Nim's move semantics
10:26:41madpropsidk but that kinda sounds like rust's borrowing system?
10:27:07madpropsthat would be interesting
10:27:09Araqno, Rust's move semantics are orthogonal to Rust's borrowing system
10:27:34Araqand C++ has them too
10:27:51AraqNim's design is actually closer to C++'s, I think
10:28:03Araqexcept that we do it better ;-)
10:31:30FromDiscord_<00imvj00> Hello, just wanted to say hello to nim community. new user here.
10:31:55FromDiscord_<00imvj00> How is the async story for Nim? i really want to explore for communication services.
10:34:12Araqwell we have async in the core and the next "generation" via 'Chronos'
10:34:29*thomasross_ joined #nim
10:34:29*thomasross quit (Killed (wolfe.freenode.net (Nickname regained by services)))
10:34:29*thomasross_ is now known as thomasross
10:37:10Araqit works well.
10:51:08*dv` joined #nim
10:52:30dv`Is there an easy way to run a command at compile time and store the output in a string? I tried eg. const version = execProcess "git describe --tags" but I get a compile error saying that importc can't be used at compile time
10:55:29*Perkol joined #nim
10:56:37*solitudesf- quit (Ping timeout: 240 seconds)
10:57:40FromDiscord_<mratsim> staticExec / gorge
10:58:07FromDiscord_<mratsim> not sure you can store the output directly, you would have to put it in a file and staticRead/slurp it
10:58:42FromDiscord_<mratsim> for compiletime FFI there is a hidden flag, look for compiletime ffi on the tracker, it was added by timotheecour
10:58:52dv`Oh great, that looks like what I want. Thanks
10:59:29Araq const gitHash = gorge("git log -n 1 --format=%H").strip
10:59:32FromDiscord_<mratsim> so what's the status of araqsgc vs gc:arc?
10:59:36Araqis what the Nim compiler itself uses
11:01:03Araqmratsim: gc:arc is easier to develop as it's not a hybrid solution
11:04:25FromDiscord_<mratsim> btw, I used a mix of mimalloc and snmalloc design (message-passing based Microsoft allocators) for my new thread-safe memory pools
11:04:28madpropswondering if there's a built in way to walk inside a directory recursively, to check total size
11:04:58*couven92 joined #nim
11:04:58madpropsnot sure if walkdirs does it
11:05:08madpropsbut maybe /some/dir/* ?
11:05:16clyybberAraq: Why are extra opCodes needed for addr(a.x) ? Does the VM not have the concept of Lvalues?
11:05:36FromDiscord_<mratsim> available here: https://github.com/mratsim/weave/blob/memory-pool/weave/memory/memory_pools.nim
11:08:53FromDiscord_<Lantos> Hey guys trying to use nim at work. Stuck behind a proxy.
11:08:53FromDiscord_<Lantos> I can access npm through the proxy with
11:08:53FromDiscord_<Lantos> http_proxy=http://domain%5cusername:password@proxy:port
11:08:54FromDiscord_<Lantos>
11:08:54FromDiscord_<Lantos> Trying nimble with the following in the nimble.ini
11:08:54FromDiscord_<Lantos> Httpproxy= "http://domain\\username: password@proxy:port"
11:08:55FromDiscord_<Lantos>
11:08:58FromDiscord_<Lantos> Is failing the domain\\user is doing some odd encoding see with npm I am running %5c to represent \
11:08:59FromDiscord_<Lantos>
11:09:00FromDiscord_<Lantos> Any ideas?
11:09:26*FromGitter quit (Read error: Connection reset by peer)
11:09:44*FromGitter joined #nim
11:10:27FromDiscord_<mratsim> you managed to get through the proxy for npm?
11:14:09FromDiscord_<Lantos> Yeah
11:14:12*luis_ quit (Quit: luis_)
11:14:15Araqclyybber, think about it
11:14:33FromDiscord_<Lantos> It was a pain. The windows domain is the issue
11:14:40Araqhere is what you cannot do for 'addr(obj.field)':
11:14:58Araq- evaluate obj.field, store it as 'tmp'
11:15:05Araq- take the address of 'tmp'.
11:15:24FromDiscord_<Lantos> The windows name domain\username
11:15:24FromDiscord_<Lantos> Is causing an issue withe the backslash I think
11:15:45Araq'addr' is a very unusual function
11:16:04Araqso it needs special opcodes
11:17:56Araqbtw clyybber we don't destroy nkObjDownConv(f(...)) properly
11:18:01Araq:-(
11:18:13Araqbecause of that non-sense p vs pArg split
11:18:23clyybberdamn
11:18:43clyybberdo you want to unify it or shall I try?
11:19:23Araqthere is really only 2 different cases: we construct a thing and know where it will end up (moveOrCopy) or we need to bind it to a temporary (p/pArg)
11:20:32Araqclyybber, please give it a try, test program:
11:20:35FromGitter<alehander92> Araq how would arc help with async
11:21:27FromGitter<alehander92> i was just looking at some of its internals and tried to see where it fits
11:22:19ehmryFWIW I started writing a networking frontend to replace sockets https://git.sr.ht/~ehmry/nim_taps, the abstract interface is coming from the IETF and there should be an RFC for it next year
11:22:33FromGitter<alehander92> also, is there any possibility to optimize out closures
11:22:43FromGitter<alehander92> if e.g. one can detect the future cant outlive
11:22:47Araqhttps://github.com/nim-lang/Nim/issues/12758
11:22:49disbot➥ --gc:arc wrong refcounting @11Araq; snippet at https://play.nim-lang.org/#ix=232p 😏
11:22:50FromGitter<alehander92> the data it depends on
11:23:45Araqclyybber, why can't I assign you to the issue?
11:23:48clyybberAraq: Ok, will do. Btw, since you marked 12282 as easy, do you have a fix?
11:23:51clyybberAraq: Dunno
11:24:03Araqwanna be a core deve?
11:24:14clyybbero/ yes
11:24:45narimiranwoohoo, welcome clyybber! :)
11:25:09FromGitter<alehander92> finally !
11:25:17clyybber:D
11:25:45Araqis your nick the same on github? two y and two b?
11:25:50clyybberYeah
11:25:50FromDiscord_<mratsim> @alehander42 for my multithreading runtime I use lazyFutures, they are stack allocated via alloca https://github.com/mratsim/weave/blob/memory-pool/weave/datatypes/flowvars.nim#L61-L68
11:25:50FromDiscord_<mratsim> And when I detect that they will escape because they will be stolen by another thread, I heap allocate them lazily: https://github.com/mratsim/weave/blob/memory-pool/weave/victims.nim#L164-L201
11:26:00FromGitter<alehander92> btw "# TODO: Once https://github.com/nim-lang/Nim/issues/5617 is fixed. ⏎ ⏎ ```# TODO: Add line info to the complete() call!" in asyncdispatch``` [https://gitter.im/nim-lang/Nim?at=5ddfaec8114bb60b7870a26b]
11:26:01disbot➥ Ability to set a NimNode's lineinfo @11dom96
11:26:15FromGitter<alehander92> this issue is fixed, is this a problem anyone experienced
11:26:21FromGitter<alehander92> because then it can be fixed i guess
11:26:23*Vladar quit (Quit: Leaving)
11:26:51Araqalehander92: closures can be optimized much like other "partial object constructions"
11:27:19FromGitter<alehander92> mratsim nice
11:27:22Araqbut in Nim most closures do escape when they don't, Nim offers the 'mapIt' etc constructs
11:27:26FromGitter<alehander92> but is this applicable to single-threaded async
11:27:45Araqand even worse, closures in Nim introduce memory cycles
11:27:48FromGitter<alehander92> Araq yeah i wondered specifically about async
11:27:58FromGitter<alehander92> because it seemed harder to me to detect those cases
11:28:16Araqasync is all about moving stuff from the stack to the heap and to nuke the stack
11:28:29FromGitter<alehander92> which doesnt sound great to me :D
11:28:52*rockcavera joined #nim
11:28:53clyybberAraq: Nice, it worked. Assigned myself
11:29:12Araqit's the point of async, otherwise we would be talking about "coroutines"
11:30:32Araqbut here is what --gc:arc does: it avoids an entire subsystem of Nim, the conversative stack tracing
11:30:36FromGitter<alehander92> hm, i wonder if one can implement async sugar on top of coroutines
11:30:45FromGitter<alehander92> i am not sure how other "partial object constructions" are optimized
11:30:58Araqso you can have as many stacks as you want to
11:31:29Araqand so coroutines become much easier and than you can build async on top of coroutines
11:31:41FromGitter<alehander92> does this mean you can have a bigger stack
11:31:45FromGitter<alehander92> or just many smaller stacks
11:32:09Araqthe "GC" doesn't need to understand/know about the involved stacks
11:32:20Araqso you can do the usual setcontext hacks
11:32:46FromDiscord_<mratsim> @alehander42 I don't know, if it can be adapted. Maybe, but what makes it possible is that I package everything into a task and a Flowvar is just a channel
11:33:35FromGitter<alehander92> so does this mean
11:34:09FromGitter<alehander92> sorry i dont understand most of it
11:34:12FromGitter<alehander92> but i'll learn
11:34:38FromGitter<alehander92> ok, so those stacks are the existing stacks of previous calls and you jump between them
11:34:58FromDiscord_<mratsim> regarding other super low-level closures and stack management. This repo is very impressive, the overhead is very very low, and "closures" are saved and restored from any thread: https://github.com/chaoran/fibril/blob/master/src/fibrili.c#L29-L73
11:35:04*LargeEpsilon quit (Ping timeout: 268 seconds)
11:35:09FromDiscord_<mratsim> also the code is minimal
11:35:54FromDiscord_<mratsim> I still can't reach it's performance on silly fib(40) (40% more overhead), but I have 25% less overhead that the next best
11:36:46FromDiscord_<mratsim> Paper should be applicatble to coroutines as well: https://github.com/chaoran/fibril/blob/master/src/fibrili.c#L29-L73
11:36:56FromDiscord_<mratsim> sorry: http://chaoran.me/assets/pdf/ws-spaa16.pdf
11:37:46FromDiscord_<mratsim> and I explain the challenges of "cactus stacks" here: https://github.com/mratsim/weave/blob/memory-pool/weave/memory/multithreaded_memory_management.md#the-cactus-stack
11:40:48*Perkol quit (Remote host closed the connection)
11:41:36FromDiscord_<mratsim> A basic closure stack that you can save/restore would probably be awesome for both async and threading
11:43:20*uu91 quit (Read error: Connection reset by peer)
11:43:38*uu91 joined #nim
11:51:41*thomasross quit (Ping timeout: 250 seconds)
11:51:59*thomasross joined #nim
11:52:52Araqclyybber, the *real* question is why the VM doesn't need a special opcode for 'addr obj.field.field' ;-)
11:52:53*uu91 quit (Read error: Connection reset by peer)
11:53:28Araqand it's because 'let tmp = obj.field; addr(tmp.field)' is good enough due to the PNode representation in the VM
11:53:41clyybberah, I was wondering about that
11:54:02*uu91 joined #nim
11:54:35clyybberAraq: Does the VM have a different representation of var as opposed to a hidden pointer?
11:56:47clyybberI was wondering, because your PR says its `var T` improvements too
11:57:05Araqvar T is mapped to 'addr' in the frontend
11:57:11Araqquite early
11:57:20Araq(bad idea but it's a legacy)
11:57:27clyybberAgh, so we need extra op codes for #12282 too?
11:57:29disbothttps://github.com/nim-lang/Nim/issues/12282 -- 3var param with type conversion doesn't work in VM @11Clyybber; snippet at 12https://play.nim-lang.org/#ix=22XY 😏
11:57:48Araqclyybber, no why, as I said 'var T' is mapped to 'addr'
11:58:37clyybberyeah, so its essentially `addr SomeTaintedString.string`
11:58:40Araqthe issue is "easy" because inside the VM we must not lose lvalue-ness accross type conversions
11:58:48clyybberAh
11:58:51clyybberRight.
12:00:31Araqin the vmgen you can add a check for "trivial distinctness type conversions"
12:01:36Araqthe C codegen does it too
12:03:09clyybbernice, might want to tell that to krux02 because I think he attempted to solve that issue too, didn't he?
12:03:25clyybberI will try to fix it after the gc:arc thingy
12:03:34Araqno, he worked on something else
12:05:03clyybberah, right, that was the thing that was fixed with the new opcodes.
12:10:53Araqunfortunately it doesn't fix "mitems at compile-time"
12:11:03Araqwe'll get there
12:20:38madpropsis there a lib for iterating inside directory's content recursively? I just want to get the total size of directories
12:20:48madpropslike 4mb or whatever
12:22:52madpropsoh wait im getting somewhere with */*
12:22:56madpropsbut i wonder if that's just 1 level
12:23:44*LargeEpsilon joined #nim
12:23:57madpropshmm yeah doesn't work
12:26:07PMunchHmm, can't methods be used on compile-time?
12:26:38PMunchI'm trying to read PNG files on compile-time and convert them to an internal representation byte array
12:27:19FromGitter<alehander92> when you debug your code
12:27:23FromGitter<alehander92> trying to understand what is missing
12:27:36FromGitter<alehander92> you fix some stuff which actually might be a bug as well and you get to
12:27:54FromGitter<alehander92> <actualmethod which does the important stuff>: discard
12:28:32*LargeEpsilon quit (Ping timeout: 252 seconds)
12:29:04madpropsok got it. i can do it with walkDirRec
12:32:25*LargeEpsilon joined #nim
12:42:04FromDiscord_<mratsim> @PMunch methods are for dynamic types
12:42:16FromDiscord_<mratsim> types at compile-time are staic
12:42:43FromGitter<alehander92> but they can still be "dynamic" in theory
12:43:01FromGitter<alehander92> you can implement inheritance/methods in the vm
12:43:10FromGitter<alehander92> i am not sure if it would make sense
12:43:15PMunchHmm, so there is no way to run nimPNG on compile-time then..
12:43:43FromGitter<alehander92> think vm, not comptime imo
12:43:57PMunchSame difference
12:44:16FromGitter<alehander92> it even seems that there should be inheritance support in vm ?
12:45:13PMunchUgh, I really didn't feel like creating my own PNG parser atm..
12:46:36FromDiscord_<mratsim> @alehander42: https://danielkorn.io/images/debugging-xkcd.png (can't find the original xkcd)
12:47:19PMunchhttps://xkcd.com/1722/
12:51:06*eys joined #nim
12:52:08PMunchDamn it nimBMP uses addr so that doesn't work either..
12:59:14FromGitter<alehander92> mratsim sorry
12:59:22FromGitter<alehander92> debugging is bizarre
12:59:26FromGitter<alehander92> yeahh
13:00:11FromGitter<alehander92> and i dont even usually debug directly low level asm stuff as some of you guys :D
13:00:33FromGitter<alehander92> what was the hardest bug to debug for any of you
13:03:22PMunchOh, streams don't work on compile-time..
13:04:45clyybberThey do, parts of them
13:05:22PMunchWell, "newStringStream" and readStr(2) didn't work..
13:06:07clyybbernewStringStream *should* work
13:06:29clyybbersee its in the tests: https://github.com/nim-lang/Nim/pull/12284/files#diff-93574743bf0f766ad08398709499ef6aR15
13:06:31disbot➥ More of StringStream now works at compile time @11Clyybber
13:08:48*lritter joined #nim
13:10:13*krux02 joined #nim
13:14:54*LEdoian[m] quit (Read error: Connection reset by peer)
13:14:56*Manny8888 quit (Read error: Connection reset by peer)
13:14:56*joachimschmidt55 quit (Read error: Connection reset by peer)
13:14:56*isaac[m]1 quit (Read error: Connection reset by peer)
13:15:00*d-nice2[m] quit (Write error: Connection reset by peer)
13:15:03*shashlick[m] quit (Remote host closed the connection)
13:15:04*GitterIntegratio quit (Write error: Connection reset by peer)
13:15:05*Connor[m] quit (Read error: Connection reset by peer)
13:15:08*zielmicha[m]1 quit (Remote host closed the connection)
13:15:11*spymasterd[m] quit (Read error: Connection reset by peer)
13:15:12*macsek1911[m] quit (Write error: Connection reset by peer)
13:15:15*Balu[m] quit (Remote host closed the connection)
13:15:17*TheManiac[m] quit (Write error: Connection reset by peer)
13:15:18*Mike[m]2 quit (Read error: Connection reset by peer)
13:15:18*skrylar[m] quit (Read error: Connection reset by peer)
13:15:18*nc-x[m] quit (Read error: Connection reset by peer)
13:15:21*lqdev[m] quit (Remote host closed the connection)
13:15:21*salotz[m] quit (Remote host closed the connection)
13:15:22*xomachine[m] quit (Remote host closed the connection)
13:15:24*k0mpjut0r quit (Read error: Connection reset by peer)
13:15:24*planetis[m] quit (Remote host closed the connection)
13:15:24*Asbrn[m] quit (Remote host closed the connection)
13:15:24*zacharycarter[m] quit (Remote host closed the connection)
13:15:27*yglukhov[m] quit (Remote host closed the connection)
13:15:27*M948e5[m] quit (Remote host closed the connection)
13:15:27*leorize[m] quit (Remote host closed the connection)
13:15:28*Demos[m] quit (Remote host closed the connection)
13:15:28*BitPuffin quit (Remote host closed the connection)
13:15:29*Miguelngel[m] quit (Remote host closed the connection)
13:15:29*pigmej quit (Write error: Connection reset by peer)
13:15:30*meff[m] quit (Read error: Connection reset by peer)
13:15:32*nergal[m]1 quit (Remote host closed the connection)
13:15:32*lasso[m] quit (Remote host closed the connection)
13:18:31*endragor quit (Remote host closed the connection)
13:27:18*eys quit (Quit: eys)
13:29:20PMunchclybber, weird I just get the `addr` has limited support in VM
13:29:44PMunchWhen I try to do var myStr = staticRead("somefile"); var myStream = newStringStream(myStr)
13:31:42PMunchhttps://play.nim-lang.org/#ix=232D
13:32:13FromGitter<alehander92> can you call
13:32:15FromGitter<alehander92> staticExec
13:32:19FromGitter<alehander92> and serialize the output
13:32:34PMunchHuh?
13:32:45FromGitter<alehander92> for example
13:32:47FromGitter<alehander92> call a program
13:32:48*Hideki_ joined #nim
13:32:52FromGitter<alehander92> which does the nim image thing
13:32:57FromGitter<alehander92> and returns a serialized data structure
13:33:02FromGitter<alehander92> and deserialize it on compile time
13:33:06FromGitter<alehander92> this is sometimes a workaround
13:33:09FromGitter<alehander92> for ct code
13:33:19PMunchSure..
13:33:20FromGitter<alehander92> basically wrapping nim runtime code to run on ct
13:33:27PMunchBut then I need to write a little program and stuff
13:33:33FromGitter<alehander92> imo one can write it as a lib
13:33:38FromGitter<alehander92> and reuse it for many cases
13:33:41FromGitter<alehander92> something like
13:33:51FromGitter<alehander92> staticRunMyType ("module")
13:34:30FromGitter<alehander92> `staticRun[MyType]("module")`
13:34:37PMunchBetter yet, staticRun(myProc) which takes the implementation and writes it to a file, then compile and run it, and deserelises the output
13:35:00FromGitter<alehander92> yes thats very smart :P
13:35:21PMunchHmm, can I echo without newline on compile-time?
13:35:26PMunchstdout.write doesn't work..
13:35:42FromGitter<alehander92> but it needs to cache the compiled helper preferrably somehow
13:35:49PMunchWould need to grab all imports and such as well..
13:35:49FromGitter<alehander92> but otherwise i love it :D
13:35:50*tklohna quit (Ping timeout: 265 seconds)
13:36:08PMunchAnd some kind of marshaller for the type
13:36:16FromGitter<alehander92> well you should be able to reuse
13:36:22FromGitter<alehander92> some of the json libs
13:36:26FromGitter<alehander92> if any of them works on CT
13:36:38FromGitter<alehander92> it has serious limitations of course
13:36:56PMunchYeah, that one I'm less worried about
13:37:14FromGitter<alehander92> you can have a pragma staticRun
13:37:25FromGitter<alehander92> which lists the import deps
13:37:35PMunchHmm, that could work
13:37:37FromGitter<alehander92> and then transforms the call to a function which calls staticExec
13:37:45FromGitter<alehander92> like a decorator
13:37:51PMunchIs it possible to grab all the imports from a file? :P
13:38:21FromGitter<alehander92> i guess the problem is there might be (??) imports after your expansion
13:38:27FromGitter<alehander92> or as a result of other macro expansion
13:38:31PMunchNim should know where each procedure comes from, so theoretically it should be possible to implement a "grab all needed imports" thing
13:38:45PMunchYeah that's true
13:39:15FromGitter<alehander92> but the top imports sure
13:39:20FromGitter<alehander92> i guess this should be possible in theory
13:39:27FromGitter<alehander92> if someone exposes it
13:39:51FromGitter<alehander92> of course it is possible PMunch
13:39:53FromGitter<alehander92> even know
13:40:01PMunchEverything is possible :)
13:40:04FromGitter<alehander92> you just load the nim file using currentSourcePath
13:40:07FromGitter<alehander92> and parse it
13:40:12FromGitter<alehander92> using parseStmt or similar
13:40:16FromGitter<alehander92> so its doable
13:40:21PMunchI once wrote a macro that would read the code of the program it was in to grab some context :P
13:40:29FromGitter<alehander92> yeah the same !
13:40:31FromGitter<alehander92> wow
13:40:31PMunchOh yeah :P
13:40:40FromGitter<alehander92> so indeed this is true i wondereda bout it
13:40:48FromGitter<alehander92> we can write macros aware of their expansion context now
13:40:54FromGitter<alehander92> Araq would be not happy
13:40:59PMunchHaha, I mean it is possible, but it's not pretty...
13:41:38FromGitter<alehander92> but yeah this way i think staticRun can work
13:41:41FromGitter<alehander92> without anything special
13:41:59FromGitter<alehander92> and not on javascript backend of course
13:42:07clyybberPMunch: Try it in a static: block
13:42:38clyybberoh, nevermind
13:42:40FromGitter<alehander92> back to jumps
13:43:57PMunchalehander92, want to give it a go?
13:44:50FromDiscord_<Lantos> Hey Araq who is in charge of the discord?
13:44:50FromDiscord_<Lantos> I'd like to make some suggestions with the discord server
13:45:39PMunch"fun" fact, BMPs are stored upside down..
13:49:19*tklohna joined #nim
13:52:49FromGitter<alehander92> PMunch well no time :(
13:52:54FromGitter<alehander92> but its a fun idea
13:53:40PMunchHaha, that's why I didn't want to give it a go :P
13:57:02*solitudesf- joined #nim
13:57:09FromGitter<alehander92> i think its relatively easy to make a prototype
13:57:15FromGitter<alehander92> but dont lose your time with it
13:59:05FromDiscord_<python273> I have 2 threads:
13:59:05FromDiscord_<python273> 1. async websockets server (`ws` module)
13:59:05FromDiscord_<python273> 2. another thread that produces messages
13:59:06FromDiscord_<python273>
13:59:06FromDiscord_<python273> How do I pass messages from 2 thread to websocket connections in thread 1?
13:59:06FromDiscord_<python273>
13:59:09FromDiscord_<python273> For passing data from 1 to 2 I can use `Channel`, but I'm not sure how to listen to channel messages and pass it to websocket connections, since it's async
13:59:09FromDiscord_<python273>
13:59:11FromDiscord_<python273> code: https://pastebin.com/raw/YxGPb6YK
14:06:48*tklohna quit (Ping timeout: 246 seconds)
14:07:11*tklohna joined #nim
14:10:28AraqLantos: I don't know...
14:20:06FromDiscord_<python273> figured out https://pastebin.com/raw/jRwg3mC1
14:21:32PMunchHmm, I checked out the v1.0.4 branch from GitHub and but trying to build with koch or build_all.sh just gives me this: http://ix.io/232P/
14:22:40narimiranPMunch: you sure that's the correct link?
14:22:43*couven92 quit (Quit: Client disconnecting)
14:23:26PMunchOh woops, wasn't meat to be a link: Error: implementation of 'system.*(x: int64, y: int64) [declared in /home/peter/Projects/nim/nim/lib/system.nim(1304, 8)]' expected
14:24:03PMunchThat is an example of my little Arduboy macro that writes images as the internal Sprite representation on compile-time
14:24:19narimirani think you would see that if you try to build older nim (e.g. 1.0.4) with a newer one (e.g. devel)
14:24:33PMunchSo you don't have to convert the sprites with an online tool into C code and paste that into your editor
14:25:18narimiranbut building from csources should work.... :/
14:25:21PMunchTrying to build 1.0.4 with 1.0.4: http://ix.io/232R/
14:27:59narimiranyou sure choosenim isn't interfering somehow and `nim` used is not the one you expect?
14:28:28PMunchI think the issue was that build_all checks if things exists..
14:28:30narimirani know i had some similar problems until i decided to remove choosenim and just use git repo
14:30:04disruptekyeah, before build_all it's a good idea to blow away your binaries. i remove the sources as well, 'cause i don't think there's a `make clean` that works.
14:31:48*dddddd joined #nim
14:31:57*lqdev[m] joined #nim
14:31:58*M948e5[m] joined #nim
14:31:58*GitterIntegratio joined #nim
14:31:58*BitPuffin joined #nim
14:31:58*salotz[m] joined #nim
14:31:58*Demos[m] joined #nim
14:31:58*leorize[m] joined #nim
14:31:59*k0mpjut0r joined #nim
14:31:59*LEdoian[m] joined #nim
14:31:59*isaac[m]1 joined #nim
14:32:00*nergal[m]1 joined #nim
14:32:03*joachimschmidt55 joined #nim
14:32:04*Connor[m] joined #nim
14:32:04*lasso[m] joined #nim
14:32:04*Mike[m]2 joined #nim
14:32:04*planetis[m] joined #nim
14:32:04*skrylar[m] joined #nim
14:32:04*zacharycarter[m] joined #nim
14:32:04*zielmicha[m]1 joined #nim
14:32:04*yglukhov[m] joined #nim
14:32:04*spymasterd[m] joined #nim
14:32:05*xomachine[m] joined #nim
14:32:05*Asbrn[m] joined #nim
14:32:05*TheManiac[m] joined #nim
14:32:05*nc-x[m] joined #nim
14:32:06*pigmej joined #nim
14:32:06*Miguelngel[m] joined #nim
14:32:06*Manny8888 joined #nim
14:32:06*shashlick[m] joined #nim
14:32:06*macsek1911[m] joined #nim
14:32:06*meff[m] joined #nim
14:32:06*d-nice2[m] joined #nim
14:32:07*Balu[m] joined #nim
14:32:13*LargeEpsilon quit (Remote host closed the connection)
14:32:32*LargeEpsilon joined #nim
14:39:23*thomasross quit (Ping timeout: 276 seconds)
14:42:58*thomasross joined #nim
14:54:37*ng0_ joined #nim
14:56:36*ng0 quit (Ping timeout: 260 seconds)
14:56:54*ng0_ is now known as ng0
14:59:28*thomasross quit (Ping timeout: 245 seconds)
15:00:23*thomasross joined #nim
15:05:57*PMunch quit (Quit: Leaving)
15:13:37*thomasross quit (Ping timeout: 240 seconds)
15:23:15disruptekshashlick, Araq: how do you expect pre-1.1 nim to handle $nimbleDir when it finds it in a nim.cfg path statement?
15:23:33*NimBot joined #nim
15:27:38FromDiscord_<Lantos> Okay r.i.p. I'll try do a post on the forums then later tonight
15:33:56*Vladar joined #nim
15:54:37*Hideki_ quit (Remote host closed the connection)
16:06:37*M948e5[m] left #nim ("Kicked by @appservice-irc:matrix.org : User has been idle for 30+ days.")
16:08:57*Asbrn[m] quit (Quit: User has been idle for 30+ days.)
16:09:34*nif quit (Quit: ...)
16:09:38*blueberrypie quit (Quit: leaving)
16:09:44*nif joined #nim
16:09:56*cherrypie joined #nim
16:10:00*cherrypie quit (Client Quit)
16:10:32*blueberrypie joined #nim
16:30:21Araqdisruptek, I expect /nimbledeps to win and then all the --paths are relative anyway
16:31:17disrupteki'm saying it needs to be backported, else we'll end up with nim.cfg in repos that earlier nim pukes on.
16:32:18disrupteknimph won't produce $nimbleDir paths because they are ambiguous, but it'll consume them, angrily.
16:33:01disrupteknimbledeps shouldn't win -- having multiple nimblePath statements is useful.
16:33:08disruptekeven if one of them is nimbledeps.
16:34:45Araqmaybe, we'll see. Now that I think of it maybe a more general $variable mechanism for --path would have been desirable
16:35:35Araqwhere $variable is not a list of dirs, but a single one
16:35:44disruptekit doesn't take much imagination to see how multiple --nimblePath statements are useful.
16:36:06disruptekespecially before the compiler supports following .nimble-link files outside --nimblePath.
16:37:26*nsf quit (Quit: WeeChat 2.6)
16:40:56*LargeEpsilon quit (Ping timeout: 268 seconds)
16:43:38*Trustable joined #nim
16:46:36*floppydh quit (Quit: WeeChat 2.6)
16:51:28*leorize joined #nim
16:53:47Araqdisruptek, all I'm saying is that it's too early to backport it IMO
16:54:07Araqonce we tried it a little in the real world we can backport it
16:54:30disruptekwell, it cannot be tried in the real world without causing earlier compilers to puke on it.
16:54:45disruptekbut, whatever. i can handle it.
16:55:28AraqI'm talking about days here, the feature got merged today
16:55:55AraqI merged it in order to unblock shashlick_'s work on Nimble
16:56:03disrupteki understand.
16:56:36Araqfor example, I think we can merge lazyPaths and nimblePaths internally in the compiler as you said
16:56:51Araqafter a little bit of refactoring
16:58:04disruptekwhat does that buy us? we kinda need the separation between .nimblePaths and .lazyPaths because they mean different things.
16:58:35disruptekwe can't even remove .nimblePaths from .lazyPaths because sometimes you need to --nimblePath a single directory, like that of the compiler when installed via `nimble develop`.
16:58:48Araqthey do? how so?
16:59:11AraqlazyPaths was created for --nimblePath
16:59:23disruptek.nimblePaths holds the inputs, .lazyPaths holds the targets. what the compiler found when walking the inputs.
17:00:33disruptekif we remove .lazyPaths from the compiler, package managers will just have to walk .nimblePaths and build their own list. i don't mind that, but i'm not sure what it buys us. it's handy to share the compiler's knowledge; it prevents bugs.
17:05:08AraqIMO that's pretty much the package manager's job. we can add a stdlib API for getting the path walking right
17:05:33disruptekagree.
17:05:40*bacterio joined #nim
17:06:23AraqI don't mind adding stuff to the stdlib for easier interop, that's pretty much the stdlib's job. but I do mind my compiler knowing about version handling
17:06:29disruptek(as long as the compiler uses the same api)
17:06:51disrupteki agree that the compiler trying to manage versions is a liability, not an asset.
17:08:44FromGitter<deech> Does the magic pragma here ( `type ... int {.magic: Int.}` ) in `system.nim` mean that I can make a completely separate type like ( `type ... foo {.magic: Int.}` ) that has the same semantics as `int`?
17:11:29disruptekwell, pathSubs needs to support $nimbleDir and currently does not.
17:18:40Yardanicohttps://github.com/nim-lang/Nim/pull/12760 does Nim actually support 16-bit CPUs? I was surprised to see that
17:18:42disbot➥ Fix for 16 bit platforms @11PMunch
17:21:01*thomasross joined #nim
17:24:22AraqYardanico, it does support them
17:24:26Yardanicowoah
17:24:47Araqdeech: you can but only for some magics
17:24:56AraqInt should be among them
17:25:05AraqI wouldn't rely on it though :P
17:30:41*krux02 quit (Remote host closed the connection)
17:32:52*bacterio quit (Read error: Connection reset by peer)
17:41:32clyybberAraq: Is the expected output for #12758 "no leak: 0" ?
17:41:34disbothttps://github.com/nim-lang/Nim/issues/12758 -- 3--gc:arc wrong refcounting @11Araq; snippet at 12https://play.nim-lang.org/#ix=232p 😏
17:42:04clyybberI've gotten it to "no leak: 7872" but I think thats still a leak, no?
17:43:20*ng0 quit (Quit: Alexa, when is the end of world?)
17:48:09clyybbervalgrind says no leak, but I don't see where those 7872bytes would come from
17:57:49*rockcavera quit (Remote host closed the connection)
18:00:37*abm joined #nim
18:04:43FromDiscord_<yewpad> is the arc gc based on referencing counting_
18:04:53FromDiscord_<yewpad> *is the arc gc based on reference counting
18:05:00FromDiscord_<yewpad> ?
18:10:51leorizeyep
18:11:11leorizethe current gc is also refcount based
18:11:20leorizebut afaik arc simplify that
18:17:04njosephIs there a no-newline-at-the-end version of "echo" ?
18:17:15*ng0 joined #nim
18:20:15*Vladar quit (Quit: Leaving)
18:20:25narimirannjoseph: stdout.write
18:23:28njosephnarimiran: Thanks. That's similar to Python's sys.stdout.write. I think I'll remember it.
18:24:59FromGitter<deech> Are there any examples of mixing ref counting gc and destructors? I want some small resources to be refcounted but use destructors for the rest of it.
18:27:26*nsf joined #nim
18:28:12clyybberdeech: destructors for ref do refcount with --gc:destructors/arc
18:28:29clyybberAlso in --newruntime for unowned refs
18:30:38clyybberdeech: You can also implement your own ref counting using destructors
18:31:36FromGitter<deech> Nice thanks!
18:34:01*Trustable quit (Remote host closed the connection)
18:45:52Araqclyybber, there is some memory because of os.paramStr caching
18:46:02Araqso to check for leaks use this idiom:
18:46:08Araqlet mem = getOccupiedMem()
18:46:11Araqmain()
18:46:26Araqecho "MEM ", getOccupiedMem() - mem # should be 0
18:46:47clyybberI put a echo getOccupiedMem() at the start of the file
18:46:51clyybberit output 0
18:47:18Araqanyway I'm fixing the bug too but your refactoring will be welcome
18:47:28AraqI'm finding more bugs
18:47:42clyybberI too just found one :p
18:48:39clyybberthe optimization for f([...]) was never working correctly for `sink array[]` and the like
18:49:01clyybbermeaning it just didn't optimize anything.
18:50:32Araqyeah I know
18:50:41Araqbut i found different bugs :P
18:54:55clyybbernice
19:09:12*uu91 quit (Ping timeout: 246 seconds)
19:10:35*uu91 joined #nim
19:39:05*solitudesf-- joined #nim
19:41:58*solitudesf- quit (Ping timeout: 245 seconds)
19:43:53*LargeEpsilon joined #nim
19:48:46*rockcavera joined #nim
19:50:09*LargeEpsilon quit (Ping timeout: 246 seconds)
19:52:37*clyybber quit (Ping timeout: 240 seconds)
20:01:09*Ven`` joined #nim
20:07:33*solitudesf-- quit (Quit: Leaving)
20:07:49FromGitter<zacharycarter> mratsim: the unions worked perfectly - I guess the only issue is on the shader side I don't know whether they were passed as big or little endian
20:08:00*solitudesf joined #nim
20:08:03FromGitter<zacharycarter> so which component to apply the shift to is somewhat dicey
20:09:28FromGitter<zacharycarter> I'll worry about it when / if it actually becomes a problem
20:09:35FromGitter<zacharycarter> for now this should solve my metal woes
20:10:54*njoseph quit (Ping timeout: 252 seconds)
20:13:50*Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
20:19:15FromDiscord_<Lantos> Could an forum admin please un moderate me, I've just made a forum account :_)
20:19:18FromDiscord_<Lantos> Could an forum admin please un moderate me, I've just made a forum account 🙂
20:20:07*njoseph joined #nim
20:28:51FromDiscord_<IanIAnIAN> finally something in stb_image worked for me
20:29:36disruptekshowoff.
20:31:59FromGitter<zacharycarter> time to do animated skeletal meshes
20:32:03FromGitter<zacharycarter> tomorrow...
20:32:37FromGitter<zacharycarter> I'm trying to find a game to play but every game I look at, I don't want to play
20:32:39FromGitter<zacharycarter> I think I'm getting old
20:34:34FromDiscord_<IanIAnIAN> play Go then
20:35:59FromGitter<zacharycarter> Go?
20:36:08FromGitter<zacharycarter> I think I'm going to install TF2 and play that
20:36:20Yardanicohe probably meant https://en.wikipedia.org/wiki/Go_(game) :P
20:36:57FromGitter<zacharycarter> seems like a popular game amongst Nim users / programmers
20:38:10FromGitter<zacharycarter> too much thinking for me though right now
20:45:10FromDiscord_<IanIAnIAN> if you use an AI you don't have to think
20:45:17FromDiscord_<IanIAnIAN> you also win a lot more
20:45:38FromGitter<zacharycarter> this sounds less fun than TF2
20:45:44FromGitter<zacharycarter> and TF2 barely sounds fun right now
20:46:33FromDiscord_<IanIAnIAN> careful though, the man who made Idris plays Go
20:47:58*nsf quit (Quit: WeeChat 2.6)
20:51:16FromGitter<zacharycarter> well good thing I've never heard of Idris before
20:51:47FromDiscord_<IanIAnIAN> Time you watched Ivor the Engine then
20:57:15*thomasross quit (Ping timeout: 250 seconds)
21:02:14*thomasross joined #nim
21:04:11*narimiran quit (Ping timeout: 276 seconds)
21:11:54*clyybber joined #nim
21:18:17*uu91 quit (Read error: Connection reset by peer)
21:18:36*uu91 joined #nim
21:39:32*lritter quit (Quit: Leaving)
21:42:33*uu91 quit (Ping timeout: 265 seconds)
21:43:31*uu91 joined #nim
21:55:17*MnlPhlp joined #nim
22:11:19*uu91 quit (Read error: Connection reset by peer)
22:11:34*uu91 joined #nim
22:27:30*solitudesf quit (Ping timeout: 265 seconds)
22:44:12disruptek!repo msgpack4nim
22:44:13disbothttps://github.com/jangko/msgpack4nim -- 11 MessagePack serializer/deserializer implementation for Nim / msgpack.org[Nim] 8 64⭐15 10🍴 7& 2 more...
22:58:08disruptekyep, devel is completely broken wrt path handling now.
22:58:29*abm quit (Quit: Leaving)
22:59:38disrupteki have --clearNimblePath and --nimblePath="$home/git/shashtree/deps/pkgs" and yet my searchPaths includes paths from .nimble, /opt/nimble, and shashtree. shashlick Araq
23:01:02disrupteknimph can't even fix it because there's nothing to fix -- no configuration to remove or adjust, and no way to determine where these paths are coming from.
23:01:44shashlick_are you printing out searchPaths? what does it show
23:02:09disruptek🐞 search: /home/adavidoff/git/shashtree/deps/pkgs/regex-#v0.13.0/src/
23:02:09disruptek🐞 search: /home/adavidoff/.nimble/pkgs/regex-#v0.13.0/src/
23:02:10disruptek🐞 search: /opt/nimble/pkgs/regex-#v0.13.0/src/
23:02:17*uu91 quit (Ping timeout: 240 seconds)
23:02:27*uu91 joined #nim
23:02:47disruptekie. this for every single path in my --nimblePath that points to the tree where i keep all your projects.
23:04:27disruptekhmm, what else is different between 1.0 and devel?
23:09:29disruptekit must be my fault but somehow the inputs are differing. maybe it's the nim.cfg reorder.
23:11:45shashlick_i just tested and this doesn't show up in searchPaths
23:11:50shashlick_are you printing lazyPaths?
23:12:43disruptekomg it's nimbledir.
23:12:59disrupteknevermind, i forgot i turned it on to test it. my bad.
23:13:08disrupteki guess it's working perfectly. perfectly stupidly.
23:13:22shashlick_okay glad, not the day to debug anything
23:14:03disruptekhave a good one, shashlick. sorry to interrupt.
23:15:04shashlick_no problem
23:16:53*krux02 joined #nim
23:17:37*shashlick_ quit (Quit: quit)
23:18:28*shashlick joined #nim
23:32:11*MnlPhlp quit (Remote host closed the connection)
23:33:25*ng0 quit (Quit: Alexa, when is the end of world?)
23:48:13*krux02 quit (Remote host closed the connection)