<< 23-05-2017 >>
00:18:28*def-pri-pub joined #nim
00:24:45*JHonaker joined #nim
00:30:59*chrisheller quit (Remote host closed the connection)
00:54:31*JHonaker quit (Ping timeout: 260 seconds)
01:05:04*planhths quit (Quit: Konversation terminated!)
01:17:34*lockdown quit (Quit: leaving)
01:35:14*brson quit (Quit: leaving)
02:35:59*dddddd quit (Remote host closed the connection)
02:58:44*pilne quit (Quit: Quitting!)
03:10:08*vlad1777d joined #nim
03:10:27*chrisheller joined #nim
03:36:05*vlad1777d quit (Quit: Leaving)
03:53:40*daaf joined #nim
04:07:39FromGitter<zacharycarter> can finally make tiled maps with frag
04:08:01FromGitter<zacharycarter> (https://files.gitter.im/nim-lang/Nim/58l8/Screen-Shot-2017-05-23-at-12.03.22-AM.png)
04:12:37ftsf\o/
04:12:49ftsfawesome
04:14:16*literal quit (Ping timeout: 260 seconds)
04:14:31FromGitter<zacharycarter> :D thanks
04:14:58*literal joined #nim
04:16:17FromGitter<zacharycarter> I think frag is in for a big clean up / refactoring effort and then I need to work on android support a bit more, but it's getting damn close to ready for an alpha
04:16:51*SusWombat quit (Remote host closed the connection)
04:22:06*SusWombat joined #nim
04:23:38*smt joined #nim
04:23:41*vivus quit (Ping timeout: 246 seconds)
04:23:49def-pri-pubI recently discovered RayLib. It's made in Pure C and is XNA like. I'm kind of tempted to start bindings on it:
04:23:52def-pri-pubhttp://www.raylib.com/index.html
04:24:03*vivus joined #nim
04:24:12def-pri-pubOr at least see how well c2nim could generate some.
04:27:19*daaf quit (Remote host closed the connection)
04:28:14FromGitter<zacharycarter> interesting
04:31:41*daaf joined #nim
04:36:38FromGitter<zacharycarter> def-pri-pub: how does it handle html5?
04:37:47FromGitter<zacharycarter> ah webgl I see
04:40:58FromGitter<zacharycarter> or emscripten can't tell
04:49:14FromGitter<zacharycarter> okay definitely emscripten
04:49:29*daaf quit (Ping timeout: 255 seconds)
05:15:07*def-pri-pub quit (Quit: leaving)
05:33:19*vlad1777d joined #nim
06:01:44*skrylar joined #nim
06:02:22skrylarmorning
06:15:16*nsf joined #nim
06:15:48FromGitter<Varriount> Hey
06:17:34FromGitter<zacharycarter> o/
06:17:43FromGitter<zacharycarter> hrm I wish the new json macro supported optional fields
06:28:08FromGitter<Varriount> "Error: cannot prove 'newSeq(0)' is not nil"
06:30:22*Vladar joined #nim
06:35:35*BennyElg joined #nim
06:36:15*BennyElg_ joined #nim
06:36:15*BennyElg quit (Read error: Connection reset by peer)
06:37:11*BennyElg joined #nim
06:37:11*BennyElg_ quit (Read error: Connection reset by peer)
06:53:28*Trioxin quit (Ping timeout: 240 seconds)
06:53:51FromGitter<Varriount> Looks like system.nim needs to be annotated
07:14:53*Arrrr joined #nim
07:14:53*Arrrr quit (Changing host)
07:14:53*Arrrr joined #nim
07:26:14FromGitter<Varriount> @zacharycarter I've written the bulk of the type generation code for the AWS API in less than 300 lines of code.
07:30:48ftsfwhoa
07:31:30*chemist69 joined #nim
07:32:49*Tiberium joined #nim
07:34:22FromGitter<Varriount> ftsf: ?
07:34:29ftsfthat's awesome!
07:34:57FromGitter<Varriount> ftsf: I have yet to do structure generation though. That's probably going to be another 50-100
07:35:15ftsfstill sounds better than boto =)
07:35:21FromGitter<Varriount> ftsf: https://gist.github.com/Varriount/32fdd9b380246e9df605012276104891
07:35:32ftsfalthough i have not much need for AWS APIs any more since i quit my day job
07:35:53*chemist69 quit (Client Quit)
07:36:10FromGitter<Varriount> ftsf: Boto3 is... weird
07:36:15ftsfit is
07:36:22ftsfboth botos are pretty bad
07:36:28FromGitter<Varriount> I mean, I understand the type generation at runtime, it's just an odd paradigm
07:36:47ftsfit makes sense considering the moving target
07:36:54ftsfbut it's very unwieldy to use
07:37:28FromGitter<Varriount> `ec2 = boto3.client('ec2') # have to pass 'ec2' everywhere`
07:37:36*Andris_zbx joined #nim
07:38:56*chemist69 joined #nim
07:38:59FromGitter<Varriount> ftsf: Regarding the code I posted, do you think it would make more sense to move the 'quote do' sections to actual templates?
07:39:33ftsftbh i'm not familar with the quote do syntax
07:39:43ftsfnot much exp with macros/templates
07:40:16FromGitter<Varriount> quote do is kinda like inline templates
07:40:48FromGitter<Varriount> it turns to body into an AST, and substitutes the backticked idents
07:41:14ftsfoh nice
07:41:29FromGitter<Varriount> No manual ast construction. :D
07:41:35ftsfyeah, sounds much nicer
07:41:46ftsfi'd leave them as quote do, seems more readable
07:41:54ftsfand readability is king
07:42:24FromGitter<Varriount> ^ Exactly why I prefer Nim
07:43:03ftsfyeah, when in doubt, go for the more readable option
07:44:29*libman quit (Quit: Connection closed for inactivity)
07:46:15FromGitter<Varriount> ftsf: By the way, is your game going to be on GOG too?
07:46:45ftsfhmm maybe, seems the requirements for indie games on GOG are much higher
07:47:04ftsfthey're a lot more choosy
07:47:23FromGitter<Varriount> I ask, as I prefer to buy my games without DRM/Steam requirements
07:47:43ftsfmmm that's why it's on itch.io
07:47:50FromGitter<Varriount> Ah, ok.
07:47:51ftsfand don't think steam requires you to have DRM
07:49:50FromGitter<Varriount> No, but it makes it complicated when I want to play multiplayer games with my siblings.
07:50:22ftsfahh i mean, you can download a DRM free game on steam and copy it to another computer and run it and it should be fine
07:50:25*gokr joined #nim
07:50:44FromGitter<Varriount> Usually we can afford to buy copies, but when there are games that are $60+ , buying 4 copies is somewhat expensive
07:50:58ftsfyou should use the family sharing functionality
07:51:16ftsfyou can buy one copy and share it amongst your "family"
07:51:24ftsfalthough i don't think you can both play at the same time
07:51:57skrylaryou can't
07:52:09skrylarif someone tries to borrow a game in your library then the othe rguy can't play anything
07:53:52FromGitter<Varriount> Yeah. We use that to cut down on costs
08:06:47FromGitter<ephja> I've never used quote, but I do define templates and pass them to getAst whenever possible
08:07:30*BennyElg quit (Remote host closed the connection)
08:07:50*BennyElg joined #nim
08:08:41FromGitter<Varriount> ephja: How well does that technique handle passing in idents and literals?
08:09:37*vivus quit (Quit: Leaving)
08:15:54FromGitter<Varriount> Araq: Why would I be getting a "cannot prove not nil" error for a member that is an object?
08:16:24*xet7 quit (Quit: Leaving)
08:24:35FromGitter<ephja> @Varriount "template t(x: untyped) = x + x" does of course accept identifiers and literals, and the resulting AST might look like this ⏎ `StmtList(StmtList(Infix(OpenSymChoice(Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sy ⏎ m(+), Sym(+), Sym(+)), Ident(!"x"), Ident(!"x"))))` ⏎ or this ⏎ `StmtList(StmtList(Infix(OpenSymChoice(Sym(+), Sym(+), Sym(+),
08:24:35FromGitter... Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sym(+), Sy ... [https://gitter.im/nim-lang/Nim?at=5923f1c3f3001cd3424cada2]
08:25:01FromGitter<ephja> that's not the most concise example. anyway, did I understand you correctly?
08:25:31FromGitter<Varriount> Possible. If you look at the link to the gist I posted, there are several procedures that use `quote do`.
08:26:03FromGitter<Varriount> In those procedures, I have to manually turn certain members of the "Shape" type into literals or idents
08:26:32FromGitter<Varriount> How would I pass those variables into a template?
08:30:56FromGitter<ephja> I wonder if 'quote' is better though. it's very concise
08:31:04*couven92 joined #nim
08:36:39FromGitter<ephja> @Varriount this seems to work ⏎ ⏎ ```template t(x: NimIdent) = type x = ref object``` [https://gitter.im/nim-lang/Nim?at=5923f4979f4f4ab05bf07ad7]
08:40:06*m712 quit (Quit: bye-nyan!)
08:43:04FromGitter<ephja> but it does compile even if the parameter is of some other type, and the actual type ends up being whatever is passed in, which would be a NimIdent in my case. time for some more bug report hunting then
08:46:00*erwana[m] quit (Remote host closed the connection)
08:46:03*hohlerde quit (Remote host closed the connection)
08:46:04*dyce[m] quit (Read error: Connection reset by peer)
08:46:05*Jipok[m] quit (Read error: Connection reset by peer)
08:46:07*mindB quit (Read error: Connection reset by peer)
08:46:08*ehmry[m] quit (Remote host closed the connection)
08:46:08*jivank[m] quit (Remote host closed the connection)
08:46:08*TheManiac quit (Read error: Connection reset by peer)
08:46:09*MrAxilus[m] quit (Read error: Connection reset by peer)
08:46:09*Guest73656[m] quit (Read error: Connection reset by peer)
08:46:11*vopi181[m] quit (Read error: Connection reset by peer)
08:46:11*M-Quora quit (Remote host closed the connection)
08:46:12*stisa quit (Remote host closed the connection)
08:48:33*gmpreussner quit (Ping timeout: 240 seconds)
08:48:33*Araq quit (Ping timeout: 240 seconds)
08:49:03*dom96 quit (Ping timeout: 240 seconds)
08:49:05*Araq joined #nim
08:49:23*vqrs quit (Ping timeout: 260 seconds)
08:49:29*erwana[m] joined #nim
08:51:30*gmpreussner joined #nim
08:53:05*dom96 joined #nim
08:54:15*vqrs joined #nim
08:56:12*MrAxilus[m] joined #nim
08:56:12*dyce[m] joined #nim
08:56:13*M-Quora joined #nim
08:56:13*ehmry[m] joined #nim
08:56:19*Guest73656[m] joined #nim
08:56:19*vopi181[m] joined #nim
08:56:19*stisa joined #nim
08:56:19*jivank[m] joined #nim
08:56:20*TheManiac joined #nim
08:56:20*Jipok[m] joined #nim
08:56:20*mindB joined #nim
08:56:20*hohlerde joined #nim
09:18:00*adeohluwa joined #nim
09:21:34*bjz joined #nim
09:33:46FromGitter<Varriount> ephja: I think if I have to add more to the quote sections, I'll break them out into templates.
09:41:51*skrylar quit (Ping timeout: 260 seconds)
09:50:05*Neomex quit (Ping timeout: 255 seconds)
09:51:45*Neomex joined #nim
09:55:58*fr1end joined #nim
09:56:30fr1endHi. How do I logical-OR things? https://nim-lang.org/docs/manual.html#lexical-analysis-operators says | is an operator, but trying to use it gives Error: undeclared identifier: '|'
09:56:48fr1end...interesting, so does the 'or' keyword
09:56:49fr1end> echo(4 or 2)
09:56:49fr1endHint: nrpltmp [Processing]
09:56:49fr1endnrpltmp.nim(1, 3) Error: undeclared identifier: '|'
09:56:59fr1end> nim -v
09:56:59fr1endNim Compiler Version 0.16.0 (2017-03-12) [Linux: amd64]
09:57:44fr1endoh, sorry, that was just nrpl repeating the last error
09:58:59chemist69you need a boolean expression for `or`, like `4 == 4 or 4 == 2`
09:59:04fr1endOkay, seems like 'or' does work, but | doesn't by default do logical or
09:59:41fr1endI need to do bitwise OR, seems or does do it
09:59:46fr1end> echo(4 or 2)
09:59:46fr1endHint: nrpltmp [Processing]
09:59:46fr1end6
10:00:31fr1end(0b100 | 0b010 = 0b110)
10:01:46chemist69ok, so you answered your question. :P
10:02:21fr1endyeah, sorry
10:02:38fr1endfor the log, if anyone wants to get |, do:
10:02:39fr1endtemplate `|`(x, y: untyped): untyped = x or y
10:02:40fr1endecho(4 | 2)
10:02:46fr1end6
10:03:16fr1endblah. Pasting is cursed.
10:03:29FromGitter<coffeepots> hey guys, I updated to 0.17 last night and had some trouble with typed blocks in templates and injected variables https://gist.github.com/coffeepots/8884cdfebcf1f941e282da8b0bdd5802
10:04:33FromGitter<coffeepots> So I guess the first example 'make1' works, but I was surprised that the variable 'x' doesn't seem to get injected if the template is given a 'typed' block?
10:05:11FromGitter<coffeepots> if I pass a generic to the template I can't even get it to instantiate but I'm not sure why as the above worked before 0.17
10:15:17fr1endHow can I index into https://github.com/barcharcraz/nim-assimp/blob/master/assimp.nim#L55 ? I have a for loop that generates ints 0..mesh.vertexCount, and I'm trying to mesh.vertices[vertex] but I get "Error: cannot evaluate at compile time: vertex"
10:15:28fr1endwhich I guess means it's trying to do some kind of bounds checking, but it can't
10:16:11fr1endpossibly just cast to UncheckedArray, which seems to be what the other pointers are behind?
10:20:39FromGitter<Varriount> Hm
10:21:26FromGitter<Varriount> fr1end: Can you give a link to your code?
10:23:24FromGitter<Varriount> coffeepots: typed parameters must resolve to a type
10:24:06FromGitter<Varriount> Since `action` in `make2` is a typed parameter, it errors because `x` in the passed in body isn't declared.
10:24:29fr1endhttps://pastebin.com/C8TQes6H
10:26:37FromGitter<coffeepots> Varriount: That must have changed pretty recently then - are typed parameters now checked earlier? Because before if you inject 'x' you can use it in the block, just as you still can do with untyped
10:27:04FromGitter<Varriount> Apparently. There's been quite a bit of internal compiler development.
10:27:31FromGitter<coffeepots> it's weird that the compiler considers var x {.inject.}: T valid in an untyped block, but doesn't resolve in typed
10:28:39FromGitter<Varriount> As for the last example, ```proc `name`[T]: T =``` is being replaced with ```proc test3[int]: int =``` which doesn't make sense
10:28:41FromGitter<coffeepots> on top of that not sure why I can't pass in a generic parameter to make3 anymore - has they syntax changed? The second part of the error says 'missing parameter' for the generic yet it's there
10:28:43fr1endhttp://assimp.sourceforge.net/lib_html/structai_mesh.html this is what the C structure looks like
10:29:25FromGitter<Varriount> fr1end: Your error is on which line?
10:30:28fr1end42
10:32:12FromGitter<coffeepots> Varriount: Ah yes, taking the generic out of the proc declaration seems to work - but it still complains about 'x' being undeclared when the template parameter is typed, despite the inject
10:34:05FromGitter<coffeepots> The trouble is, in my real code I'm passing in a generic type to the proc, so that's why the example looks a bit weird. Trying to mock up something to see if it works
10:34:17FromGitter<Varriount> @coffeepots Can you paste your real code?
10:34:39FromGitter<Varriount> fr1end: The trouble is that for a TMesh, `vertices*: ptr TVector3d`
10:34:40FromGitter<coffeepots> sadly not as I don't have it here, I could tonight though - it's a quadtree implementation
10:35:21fr1endI just worked out how to do such a cast, and "let v = cast[UncheckedArray[TVector3d]](mesh.vertices)[vertex]
10:35:24FromGitter<Varriount> fr1end: You can try casting it to an array of a large length
10:35:30fr1end" appears to compile, but hopefully it does what it appears to
10:35:32fr1endyeah
10:35:54fr1endUncheckedArray (part of standard runtime) is just an array with 100_000_000 size (>_>)
10:36:07fr1endhttps://github.com/zah/grip-lang/blob/master/lib/pure/collections/rtarrays.nim#L22
10:36:17fr1endit's marked to not be used externally, but... it's there lol
10:36:35FromGitter<Varriount> fr1end: I suggest you don't try copying it to a local variable. :P
10:36:58fr1endI wanted to make this neater, is there a way to get references like C++'s?
10:37:06FromGitter<Varriount> fr1end: There's un updated way to do unchecked arrays: https://nim-lang.org/docs/manual.html#foreign-function-interface-unchecked-pragma
10:37:15FromGitter<Varriount> fr1end: Howso?
10:38:16fr1endlike, I want to be able to refer to v, n, t, u instead of doing that ridiculous cast and index for every single float I'm copying out
10:39:49fr1endin C++, that's aiVector3D& v = mesh->mVertices[vertex];
10:40:16fr1endand then only a pointer is on the stack (and the optimiser probably elides even that)
10:41:15FromGitter<Varriount> fr1end: Use a pointer type?
10:41:46fr1endWhat do you mean?
10:43:00FromGitter<Varriount> fr1end: Sorry, I don't think I understand the question.
10:44:36FromGitter<Varriount> Can you perhaps show me a code sample of what you think the Nim syntax might/should be (even if it's not technically valid)?
10:44:37fr1endSo, what I'm doing here right now, after casting to an UncheckedArray[TVector3d], and then accessing a TVector3d by [vertex], is copying whole TVector3d structures onto the stack
10:44:43fr1endI think.
10:44:53fr1endThat or it's only putting a reference to the original structure
10:45:25fr1endI want to only have a refernce to the original structure, just like the C++ would
10:45:49FromGitter<Varriount> fr1end: You would have to store a pointer to the structure on the stack then.
10:46:46FromGitter<Varriount> `let v = addr castUncheckedArray[TVector3d (mesh.vertices)[vertex]
10:46:55fr1endah, addr
10:47:12fr1endI was trying 'ptr'
10:47:51fr1endargh, mesh.nim(42, 64) Error: expression has no address
10:47:57FromGitter<Varriount> One moment...
10:49:36*bjz_ joined #nim
10:49:57FromGitter<Varriount> fr1end: `let v = addr castptr UncheckedArray[TVector3d (mesh.vertices)[vertex]`
10:50:41FromGitter<Varriount> I would probably either put that cast in a template, or create a new unchecked array type.
10:51:13FromGitter<Varriount> That wrapper is really odd.. It uses unchecked arrays in some parts, but not in others.
10:51:28*bjz quit (Ping timeout: 272 seconds)
10:51:51fr1endOK, looks like "let v = addr cast[ptr UncheckedArray[TVector3d]](mesh.vertices)[vertex]
10:51:54fr1end" works
10:55:09*PMunch joined #nim
10:55:46FromGitter<Varriount> fr1end: Interfacing with C is never easy (unless you're also using C)
10:58:13fr1endYeah. Part of the reason I'm trying Nim again is because Rust FFI is such a pain.
10:59:31*skrylar joined #nim
11:00:56FromGitter<Varriount> fr1end: Really?
11:01:38fr1endYeah. I think I might abandon this though.
11:02:09FromGitter<coffeepots> I updated my gist to hopefully highlight a bit more what I want https://gist.github.com/coffeepots/2dd94fdb17a9f83a7951341763bff45e
11:02:30fr1endThere's a lot I like about Nim but the same immaturities are there; they'll probably always be there because they're the kind that arise from not enough eyeballs
11:02:38FromGitter<Varriount> fr1end: Have you looked at what @zacharycarter is doing?
11:02:49*Snircle joined #nim
11:02:57FromGitter<coffeepots> has the syntax changed for passing generics to templates maybe?
11:03:15fr1enddo you mean their fork of this assimp wrapper?
11:03:40FromGitter<Varriount> He has a fork of the wrapper?
11:03:45fr1endbecause I just remembered, that's what I'm actually using (but the relevant section I highlighted is teh same)
11:03:51FromGitter<Varriount> No, I meant http://fragworks.io/
11:04:00fr1endhttps://github.com/zacharycarter/nimassimp
11:04:05FromGitter<coffeepots> btw fr1end: it sounds like you just need to get used to the low level workings with Nim. It took me a while to get used to it myself but once I did it makes sense
11:04:54FromGitter<Varriount> fr1end: The reason that 'addr' was needed was that an array represents a block of memory, but not a pointer to a block of memory
11:05:11FromGitter<weberschm_twitter> hi guys, I'm trying to run the opengl package example (nim c glut_example.nim) but having an error "Cannot open X"
11:05:51fr1endweberschm_twitter: does glxgears (C language GL demo) definitely work on your system?
11:06:27FromGitter<weberschm_twitter> yep. do i have to install this https://github.com/nim-lang/x11 "manually"?
11:06:59FromGitter<Varriount> @weberschm_twitter Is that the exact error message?
11:07:18fr1endVarriount: This issue was not my concern with nim. It's that the documentation is sparse, concepts are not yet documented well, Nim has the atmosphere of an experimental language still
11:07:35fr1endgiven its version number, none of this is surprising
11:07:39FromGitter<ephja> nim-glfw should also have configuration objects. it's currently very hard to decipher error messages involving calls with 5+ arguments atm
11:07:43FromGitter<weberschm_twitter> opengl/src/opengl.nim(26, 10) Error: cannot open 'X'
11:08:04FromGitter<weberschm_twitter> do i have to unpack https://github.com/nim-lang/opengl at a specific location ?
11:08:14FromGitter<Varriount> fr1end: The problem, as you said, is that we don't have enough contributors.
11:08:41FromGitter<coffeepots> fr1end: what would you say are the main areas that feel experimental? Concepts perhaps but the base language feels fairly complete to me at least
11:08:41fr1endabsolutely, and I hope that what I've said doesn't undermine the value of the effort the contributors nim DOES have have made
11:08:47FromGitter<Varriount> We only have 3 core developers.
11:09:12FromGitter<Varriount> @coffeepots I must admit that documentation (on the standard library) is quite sparse.
11:09:14FromGitter<ephja> @weberschm_twitter it's not listed as a dependency, so you have to install that package manually
11:09:33FromGitter<Varriount> Although I feel the language manual is fairly good.
11:09:55fr1endwell, say you take https://nim-lang.org/docs/manual.html#exception-handling
11:10:07fr1endit doesn't tell you how to create your own exception types
11:10:07FromGitter<Varriount> Though, documentation on how Nim interfaces with C could be improved in the manual.
11:10:14FromGitter<ephja> the question is, is it possible to include such a dependency conditionally now with metadata in the form of nim scripts? I'll investigate
11:10:17FromGitter<coffeepots> newException?
11:10:23fr1endit tells you about the newException function
11:10:44fr1endbut, you have to search for example code somewhere to find there's an Exception base class you inherit to do that
11:10:52fr1endwhich was obvious in retrospect, but...
11:11:02fr1endI didn't know whether the first argument was just an enum, or what
11:11:11fr1endthat's a random nitpick
11:11:21FromGitter<weberschm_twitter> @ephja i was just tripping.. not using nibmle
11:11:23FromGitter<coffeepots> fair enough
11:11:23fr1endI'm trying to find more concrete examples of problems
11:12:03FromGitter<coffeepots> I know when I was doing low level stuff i got tripped up by cast[TypeIWant] rather than cast[ptr TypeIWant]
11:12:11FromGitter<weberschm_twitter> example works nopw
11:12:34FromGitter<ephja> you have to install it manually anyway
11:12:43*sz0 joined #nim
11:12:57FromGitter<weberschm_twitter> i was missing a nimble install opengl
11:13:16FromGitter<weberschm_twitter> and then apt-get install freeglut3 freeglut3-dev for libglut.so
11:13:29fr1endthe proper thing to do is to, when finding a doc problem, correct it and submit a PR; but nobody does that when trying to solve their problem because of the hole-in-my-bucket effect
11:13:45FromGitter<Varriount> @coffeepots @fr1end: I just wish that people were more willing to contribute. The documentation is all there in the repository.
11:14:12FromGitter<ephja> my brain is too hyperactive
11:14:35FromGitter<Varriount> It's not even as if such changes have to wait for Araq's approval - since it's not compiler internals, dom96 or I could accept them
11:14:46FromGitter<ephja> it's actually a dependency, so it will simply be redundant on platforms that don't have X11
11:15:40fr1endcontributor critical mass is a frustrating thing to try to achieve
11:16:50FromGitter<coffeepots> I was considering writing a blog post, does that count? :)
11:17:08fr1endthe other thing that feels experimental is error messages
11:17:37fr1endthey are pretty obtuse, like, incorrect typing in a function call reports the error at the parenthesis
11:17:52fr1endas just 'type expected'
11:18:08fr1endI've written languages before and know that improving that is nontrivial
11:18:33FromGitter<coffeepots> See, error messages... they do make sense but sometimes you have to 'decipher' them
11:18:48fr1endthese are things that are minor to someone familiar with a language, but most likely to scare away newcomers
11:19:46fr1endI hope this comes across as constructive, because like I said I do like Nim and I understand why things are the way they are
11:20:01FromGitter<coffeepots> at least nowadays we get some hints like "did you mean to use a var parameter?" when you try to alter an immutable and "did you meant to use = " when declaring a type like "var thing: newseqint ()" and stuff
11:20:17FromGitter<coffeepots> we could do more of those type of hints I think
11:21:15FromGitter<coffeepots> also, when you have a parameter mismatch, it'd be great to have it point out exactly which parameters don't match to any existing signature but I bet that's a lot trickier than it sounds
11:21:40FromGitter<zacharycarter> morning
11:21:49FromGitter<zacharycarter> lots of good convo
11:21:51FromGitter<coffeepots> good morning :)
11:21:56*yglukhov joined #nim
11:22:15FromGitter<zacharycarter> yeah Nim is rough around the edges but you can definitely build quite impressive things with it as it stands today
11:22:35FromGitter<coffeepots> yes, for sure, although my current project is broken after 0.17 :/
11:23:07FromGitter<zacharycarter> what's wrong with it?
11:23:10FromGitter<coffeepots> having said that, the new vtref stuff looks exceedingly cool
11:23:32FromGitter<coffeepots> well I can't seem to pass a generic to a template in order to create a proc
11:23:38FromGitter<coffeepots> https://gist.github.com/coffeepots/2dd94fdb17a9f83a7951341763bff45e
11:24:26FromGitter<coffeepots> and I use this in my quadtree implementation to create custom tree walking procedures, and... without quadtrees, well that's kinda the crux of my whole project
11:24:51FromGitter<zacharycarter> well... this definitely works on glot.io which is a few versions back I believe
11:24:54PMunchIn jester, is there a way to specify a route that matches both "/chips" and "/chips/"? Currently I have a route for "/chips" and "/chips/@id".
11:25:01FromGitter<Varriount> By the way, fr1end, have you seen https://play.nim-lang.org/ ?
11:25:06FromGitter<zacharycarter> hrm
11:25:10FromGitter<coffeepots> yeah it was working in 0.16 devel but when i updated to 0.17, broke :(
11:25:18FromGitter<zacharycarter> yeah it's compiling on ^ too
11:25:47dom96PMunch: "/chips/?@id"
11:26:02dom96Possibly
11:26:03FromGitter<Varriount> @zacharycarter ^ == HEAD?
11:26:50dom96zacharycarter: any luck getting gist integrated with the playground? or a way to specify code as URL params
11:26:58FromGitter<Varriount> @coffeepots Why does the inner procedure in the make3 template have a generic parameter?
11:27:25PMunchHmm, and a nil check to see if it is there or not?
11:27:33FromGitter<coffeepots> the inner proc takes a quadtree[T], which is simulated with Obj[T]
11:28:15FromGitter<Varriount> @coffeepots Yes, but the template also takes a generic parameter.
11:28:17FromGitter<zacharycarter> dom96: I haven't tried yet I was working on frag all weekend but I'll give the playground some love today
11:28:25FromGitter<zacharycarter> @dom96 ^
11:28:34FromGitter<zacharycarter> @Varriount with ^ i meant play.nim-lang.org
11:28:41dom96(You don't need the @ to highlight me)
11:28:50FromGitter<zacharycarter> ah sweet
11:28:54FromGitter<zacharycarter> forgot you're in IRC
11:28:59FromGitter<coffeepots> @Varriount yes, how else would you declare what type you want make the proc `name` take?
11:29:23FromGitter<Varriount> @coffeepots Use the generic parameter the template is taking?
11:30:20FromGitter<coffeepots> if you take out [T] from proc `name` then it says it can't instiantiate T in Obj[T] in the proc
11:31:26FromGitter<coffeepots> actually... it works if you take out the [T] from the template...
11:32:07FromGitter<Varriount> I'm just confused why you have both. One or the other would work.
11:32:14FromGitter<coffeepots> This just makes the proc generic to anything and not constrained
11:32:42FromGitter<Varriount> Oh wait, are you not running HEAD?
11:33:26FromGitter<zacharycarter> dom96: I think I can probably get away with storing snippets in redis and generating GUIDs to share
11:33:26FromGitter<coffeepots> well the initial idea was to constrain the type, so you'd write make3[Entity] and it would create proc test3Entity (qt: Quadtree[Entity] ... )
11:34:08FromGitter<coffeepots> ok here I'm running version 0.17.0, at home I'm running devel
11:34:22dom96zacharycarter: Sure, redis will work for that.
11:35:01FromGitter<Varriount> Yes, which you would do like this: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=59241e6505e3326c67fe4f4c]
11:35:54FromGitter<coffeepots> @Varriount if you do that it says cannot instantiate T for me
11:36:14FromGitter<Varriount> Hm. It works on HEAD
11:36:24FromGitter<coffeepots> wicked, i'll update...
11:37:22FromGitter<coffeepots> to be honest it's probably better NOT to over specify the type and have the constructed procedures generic anyway (ie; remove [T] from template and let the compiler decide what type to use when you call test3)
11:37:49FromGitter<stisa> what about keeping T on the template and doing `proc name V:T (obj: Obj[V]): V` ? That should work
11:38:37FromGitter<coffeepots> @Silvio huh! Thanks that works!
11:39:17FromGitter<coffeepots> gonna try it on head tho too
11:41:59FromGitter<Varriount> @coffeepots You might like what I'm working on: https://gist.github.com/Varriount/32fdd9b380246e9df605012276104891
11:45:24PMunchHmm, dom96 get "/role/?@id" didn't match /role or /role/ only /role/<id>
11:45:58dom96PMunch: add a ? after the @id
11:46:41*adeohluwa quit (Quit: Connection closed for inactivity)
11:46:42PMunchHmm, didn't match /role
11:46:52PMunchBut matched /role/ and /role/<id>
11:47:11FromGitter<coffeepots> @Varriount What does the code do? I have trouble decoding what macros do sometimes
11:47:34FromGitter<coffeepots> pattern matching?
11:48:02PMunchYeah I could use the regex thing but it seems a bit overkill..
11:49:01FromGitter<Varriount> @coffeepots It creates types based on "shapes". It's going to be used to generate a library for interfacing with the AWS API.
11:49:28FromGitter<Varriount> There are JSON documents describing the API. I'm going to generate types and procedures from them.
11:49:43*Tiberium quit (Remote host closed the connection)
11:51:58FromGitter<coffeepots> whoa niiiice
11:53:16FromGitter<coffeepots> IMHO Nim just needs more libraries. Seriously thinking of creating a barebones 'gamedev' library to stand alone with common algorithms in
11:53:47PMunchcoffeepots, you are welcome to contribute to my gamelib :)
11:54:00FromGitter<coffeepots> on that note: @zacharycarter Frag.io is well good!
11:54:09PMunchhttps://github.com/PMunch/SDLGamelib
11:54:38PMunchIt's currently based on SDL but I'm considering to create a compile time flag to remove all SDL stuff
11:54:51FromGitter<Varriount> @dom96 https://github.com/nim-lang/Nim/pull/5872
11:54:56PMunchSince a lot of it is general game related stuff (like collisions and tweening)
11:54:56FromGitter<coffeepots> @PMunch hey I was looking at that just last night! Nice work on all the collision options you have in there
11:55:09PMunchThanks, took a while to implement them all :P
11:55:21FromGitter<coffeepots> I was thinking of basing on opengl as the GLvectorfX are not bad types to work with
11:55:25FromGitter<coffeepots> I bet!
11:55:34dom96PMunch: Just create another route for "/role" and redirect to "/role/" then :)
11:55:55PMunchdom96, yeah that would work I guess :P
11:56:14PMunchWhat's the default return value for a sequence by the way?
11:56:25PMunchA sequence of length 0, or nil?
11:56:38FromGitter<coffeepots> @PMunch Do you have a broadphase like quadtree in there? Couldn't see it at first glance
11:57:13FromGitter<Varriount> PMunch: You mean, what value does `var s: seq[int]` have by default?
11:57:58PMunchVarriount, yes. But this is a seq[seq[string]].
11:58:16FromGitter<Varriount> sequences are nil by default.
11:58:21FromGitter<coffeepots> We really need a really performant ECS too. I was looking at making one with slotmaps in the hope to avoid hash tables and keep stuff in cache
11:58:21PMunchcoffeepots, not at the moment no. I was thinking of adding it in there but haven't gotten around to it yet
11:58:43FromGitter<coffeepots> Well I could send you mine but it's probably badly written :D
11:58:59FromGitter<coffeepots> tbh the crux of the code is only a couple hundred lines at worst
11:59:13FromGitter<zacharycarter> Thank you @coffeepots
12:00:03*FromGitter * coffeepots is off to lunch
12:00:14PMunchWell I've taken a slightly different approach with the tick Ticker and Renderer modules in the gamelib
12:01:13*Tiberium joined #nim
12:04:20*Tiberium quit (Remote host closed the connection)
12:07:05*Tiberium joined #nim
12:07:58*arnetheduck_ joined #nim
12:09:14*MrAxilus[m] quit (Remote host closed the connection)
12:09:16*ehmry[m] quit (Remote host closed the connection)
12:09:17*dyce[m] quit (Remote host closed the connection)
12:09:18*vopi181[m] quit (Read error: Connection reset by peer)
12:09:18*stisa quit (Read error: Connection reset by peer)
12:09:20*TheManiac quit (Remote host closed the connection)
12:09:20*hohlerde quit (Read error: Connection reset by peer)
12:09:20*Jipok[m] quit (Remote host closed the connection)
12:09:21*M-Quora quit (Read error: Connection reset by peer)
12:09:21*Guest73656[m] quit (Remote host closed the connection)
12:09:21*erwana[m] quit (Read error: Connection reset by peer)
12:09:21*jivank[m] quit (Read error: Connection reset by peer)
12:09:22*mindB quit (Read error: Connection reset by peer)
12:13:22*krux02 joined #nim
12:14:00*erwana[m] joined #nim
12:16:55skrylarbleh. trying to implement hirise in nim
12:17:02skrylarbut i'm having trouble understanding the actual paper :|
12:19:01PMunchAah, dom96 apparently get "/role/?": works
12:19:21dom96cool
12:20:17*MrAxilus[m] joined #nim
12:20:17*dyce[m] joined #nim
12:20:18*M-Quora joined #nim
12:20:18*ehmry[m] joined #nim
12:20:24*Jipok[m] joined #nim
12:20:24*hohlerde joined #nim
12:20:24*jivank[m] joined #nim
12:20:24*TheManiac joined #nim
12:20:24*Guest73656[m] joined #nim
12:20:25*vopi181[m] joined #nim
12:20:25*stisa joined #nim
12:20:25*mindB joined #nim
12:31:26skrylarwat. there's no sort function in the lists module.
12:32:01skrylarbleh i'll post the mergesort i wrote later maybe dom96 or Araq can cannibalize it
12:35:27krux02skrylar, for what are you using lists?
12:35:51krux02seq is the type for appendable arrays
12:37:43PMunchkrux02 he said he's using the lists module
12:37:54PMunchhttps://nim-lang.org/docs/lists.html
12:38:09PMunchBasic implementation of linked lists
12:38:34skrylari had some code using them withe efl's eina and made a similar module in nim
12:38:43skrylarit had a sort and sorted insert, so i have functions for that
12:38:54skrylarif they're wanted for the stdlib i don't mind if they yoink them
12:39:09krux02PMunch: I know linked lists. I just wanted to say that seq is probably the better alternative even without knowing the exact use case
12:39:34skrylarseq implies a lot of shuffling
12:39:47PMunchkrux02, depends. If you need deletion then seqs are out as you can't keep a reference for deleting nodes easily.
12:40:09skrylari *do* use a vector/seq for storing "notes" in a midi sense. but i use doubly linked lists for storing the midi events.
12:40:21skrylarBecause when a note is moved, the note on and note off have to be popped and reinserted sorted
12:40:22krux02PMunch: I know that is the default argument. But for deletion you normally need to scan though the list to filter out the elements you want to delete
12:40:36krux02so it stays linear time
12:41:32krux02skrylar: is the midi event a pointer/ref type or a value type?
12:41:36PMunchkrux02, in a single list you would just return the node reference when you insert. Then to delete simply go to the last node (assuming doubly linked) and link that to the next and link the next to your last.
12:41:44PMunchThat's not linear time
12:41:46skrylarkrux02, i think right now its a pointer
12:42:23skrylardidn't optimize the note part but i did the events themselves
12:42:34skrylaralthough the nim port doesn't let you write notes in yet anyway :/
12:44:18krux02my point is, moving memory that is just linear and not scattered around is fast
12:44:30def-dom96, Araq: Isn't the nim-0.17.0.tar.xz compressed with XZ_OPT=-9? 10 times smaller when I compress it manually with XZ_OPT=-9 tar cfZ nim-0.17.0.tar.xz nim-0.17.0. The scripts I can find in nim repo look fine, so not sure why that's happening
12:44:32*Tiberium quit (Remote host closed the connection)
12:45:19dom96def-: no idea, it was created via 'koch xz'
12:45:19krux02so inserting into a seq at the front is very fast compared to iterationg a linked list where every node might be a cache miss
12:45:31krux02a cache miss can be up to 200 CPU cycles
12:46:45PMunchkrux02, you don't have to iterate a linked list to add to the front
12:47:03PMunchYou should make your new node point to the old head and then change the head to the new node
12:48:35krux02PMunch: I know how linked lists are supposed to be used. The problem is just that it is wrong in practice. The CPU cache eleminates all theoretical advantages of a linked list.
12:48:58PMunchThat depends on how you use them
12:49:05PMunchBut I agree that they are bad for cache purposes
12:49:27krux02They also have a lot of overhead
12:50:03PMunchRecently I tried in Nim to create a seq and a linked list of nodes into which I inserted a series of procs and then ran 100_000 loops through both to execute the procs. Linked list was faster for some reason
12:50:17def-dom96: ./koch csource -d:release && ./koch xz && ls -lha build/nim-0.17.0.tar.xz
12:50:26def-3.6 MB instead of ~20 MB from website
12:50:43krux02PMunch: do you have a link?
12:50:56dom96def-: you need to build docs too with './koch web0'
12:51:07*Tiberium joined #nim
12:51:18dom96https://github.com/nim-lang/Nim/wiki/Creating-a-release
12:51:52PMunchkrux02, not really. I was just testing it for myself since I needed it for the renderer in gamelib
12:52:27PMunchhttps://github.com/PMunch/SDLGamelib/blob/master/gamelib/scenegraph.nim
12:52:38PMunchThat's the current implementation with a double linked list
12:53:11PMunchI was adding RenderProcs to the seq and linked list
12:53:12def-dom96: ok, 3.9 MB, but point stands
12:53:42dom96Dunno then, I built on a Mac maybe that's why?
12:53:50PMunchMy guess is that the procs doing the rendering invalidates the cache
12:54:15def-dom96: possible. Should I make it tar and xz separately as a workaround?
12:54:35dom96def-: no idea, I didn't implement this.
12:54:37def-I guess on mac you use gtar. Maybe it's an old version and it doesn't know about XZ_OPT yet
12:55:55PMunchdom96, by the way. About dumpCode/codeRepr, would dumpAstGen and astGenRepr be better?
12:56:20PMunchI want to stick with the dumpX Xrepr pattern used for lisp and tree
12:56:20dom96yes
12:57:52*yglukhov quit (Remote host closed the connection)
13:01:56*yglukhov joined #nim
13:01:59*yglukhov quit (Remote host closed the connection)
13:02:47PMunchdom96, there now I renamed it
13:05:46*yglukhov joined #nim
13:09:25krux02PMunch: Well probably you are right the the procs invalidate the cache.
13:10:00PMunchBut I was still surprised to see that the linked list was faster and not only as fast
13:10:23PMunchIIRC it was a considerate amount as well
13:10:41krux02Well I can't explain that. But I am also not 100% sure what you mesured and how
13:11:59krux02At least you did the measurement so that is a very good point that you are right
13:12:14krux02It just conflicts with my view on the world
13:12:46PMunchMine as well
13:13:16PMunchDid a course on Green computing which was a lot about cache-locality and that stuff. So it was odd to see this behaviour
13:13:32*yglukhov quit (Remote host closed the connection)
13:13:55krux02Well I acutally saw the "seq" implementation and it was not so simple as I thought it to be
13:14:52krux02but I thought that at least iterating it should be fast
13:14:53FromGitter<ephja> so, how come zah's fork always show up first on google searches?
13:15:03PMunchBasically I created a seq, then in a loop I added 10_000 procs to that seq. The procs were all a simple call to render a texture. Then in the main loop I iterated through both and ran all the procedures. I timed the milliseconds between each frame and converted this to an FPS measurement.
13:15:32krux02ephja: google has personalized search results, maybe you were on that page from time to time
13:18:24*gokr quit (Ping timeout: 260 seconds)
13:18:57*gokr joined #nim
13:20:29FromGitter<ephja> I don't think so. nimjs popped up this time when I searched for the system module
13:21:05FromGitter<ephja> oh well
13:21:55*yglukhov joined #nim
13:26:49FromGitter<coffeepots> hey, so whats up with this in test2? https://gist.github.com/coffeepots/9cad1538b434a45cca684f2e99461ce0
13:27:03FromGitter<coffeepots> again, if you change actions to be untyped it works!
13:27:15FromGitter<coffeepots> as typed, injected variable is not visible
13:31:18*vivus joined #nim
13:31:38vivuswhat is the correct syntax for doing this: echo i + ")" + reg_display[i] ?
13:31:42FromGitter<stisa> I believe with `typed` actions is already type checked before entering the template, so it can't see `x` ( because it doesn't exist yet )
13:32:33FromGitter<ephja> & for concatenation
13:32:45vivusoh yes, not +
13:33:40FromGitter<ephja> echo will attempt to convert every argument to a string, so you might want to do "echo a, b, x"
13:33:41vivuswoops, still the same error: Error: type mismatch: got (range 0..3(int), string)
13:34:14*chemist69 quit (Ping timeout: 246 seconds)
13:35:17FromGitter<ephja> just replace + with , in other words
13:35:49vivusa, b, x ... worked :) thank you @ephja
13:39:48FromGitter<coffeepots> @stisa yeah I think you're right. That must have changed recently
13:40:08FromGitter<coffeepots> since you can apparently return values from untyped, i guess I'll just use that
13:41:03FromGitter<coffeepots> i kinda thought it'd make more sense to not allow return values from untyped as there's typed, but i guess untyped can be anything so why not
13:42:28FromGitter<coffeepots> thanks to everyone that's helped me, should be able to get my project back up and running tonight :)
13:46:01*Trioxin joined #nim
13:47:43*Lord_Nightmare quit (Ping timeout: 258 seconds)
13:56:13*Arrrr quit (Read error: Connection reset by peer)
13:56:48*Neomex quit (Ping timeout: 260 seconds)
14:01:26*chemist69 joined #nim
14:03:07FromGitter<ephja> as long as the output passes the semantic phase
14:12:45*rokups joined #nim
14:16:52krux02coffeepots: I think the fact that *untyped* can return values and *typed* can't doesn't make a lot of senso
14:17:06krux02I think the old stmt and expr did make more sense
14:17:41krux02but you can actually put the type there when your macro/template is supposed to return value of that type
14:18:07krux02you can't use return in that case though, you have to use result = ...
14:19:07chrishellerAnyone know of a Nim memory dumper? e.g. give it address/pointer and a size and it will pretty print the memory contents starting at that address?
14:19:31chrishellerDidn't see anything in nimble or standard library
14:20:55*yglukhov quit (Remote host closed the connection)
14:32:02FromGitter<ephja> krux02: I did try to return an actual literal once, but it didn't work. perhaps it does make more sense to construct a node instead
14:33:55FromGitter<Varriount> chrisheller: No, although it's a neat idea
14:34:11FromGitter<coffeepots> @krux02 I agree it is weird that untyped can return values, wait are you saying typed cannot return values? That is very strange! I thought typed == expr and untyped == stmt
14:34:16FromGitter<Varriount> Like something that would return a hex dump of a memory block?
14:34:33chrishellerExactly.
14:34:59FromGitter<ephja> Araq: "macro m(x: type(enum)): set[x]", "ordinal type expected". reported?
14:35:04chrishellerI can put something together. I was just surprised that no one else had done that yet
14:35:54*BennyElg quit (Remote host closed the connection)
14:36:16FromGitter<ephja> is github being slow for anyone else right now?
14:38:08*yglukhov joined #nim
14:38:41*sz0 quit (Quit: Connection closed for inactivity)
14:42:20*yglukhov quit (Remote host closed the connection)
14:42:24Trioxinman objective-c is terrible. was just looking on rosetta code
14:43:09demi-why do you think it is terrible?
14:43:35Trioxinwas just skimming this page and saw this: https://rosettacode.org/wiki/HTTP#Objective-C
14:44:22demi-that seems like a very outdated code sample, but what about it
14:47:40*yglukhov joined #nim
14:48:45Trioxinnot very intuitive and is very busy for what it's doing. obj-c is a relatively modern language right? I guess the c++ example isn't much better but I really don't think all that is needed in C++. The C one shows using curl which can be used from just about any language
14:49:09demi-it is like 30-45 years old
14:49:55Trioxinmaybe I'll edit some of these
14:50:05*bjz_ quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
14:50:13Calinouwell, it's Apple, what did you expect :P
14:50:14Trioxindamn look at cobol
14:50:47Trioxinthat's cobol + curl
14:51:17demi-and that is almost identical to the swift example; which is because they are both using the samne framework, which isn't a great example of the language at all tbh, since the language on its own is basically just the C stdlib plus some small-talk inspired classes
14:53:22vivuswhoa cobol
14:53:23*yglukhov quit (Remote host closed the connection)
14:53:59*yglukhov joined #nim
14:55:08Trioxinrandom Ada examples look neat. I would be a full-time Ada programmer right now if my USAF recruiter hadn't screwed me over.
14:55:24Trioxinwhen I was 18
14:56:49TrioxinI imagine they still use it a lot. this was 2002 I think
14:58:58*Lord_Nightmare joined #nim
15:00:32*Tiberium quit (Remote host closed the connection)
15:00:47FromGitter<coffeepots> I do not envy cobol developers
15:02:25*yglukhov quit (Remote host closed the connection)
15:03:12*yglukhov joined #nim
15:03:54*Arrrr joined #nim
15:03:54*Arrrr quit (Changing host)
15:03:54*Arrrr joined #nim
15:05:56*yglukhov quit (Remote host closed the connection)
15:06:26*yglukhov joined #nim
15:08:38*Tiberium joined #nim
15:10:24*m712 joined #nim
15:16:04*Andris_zbx quit (Remote host closed the connection)
15:16:14FromGitter<zacharycarter> Hrmm I have some questions about spawn / threadpools
15:16:39FromGitter<zacharycarter> if a FlowVar.isReady == true and I call the ^ proc on the flow var, shouldn't there be no blocking time?
15:16:41*adeohluwa joined #nim
15:18:18dom96zacharycarter: yep, there shouldn't be.
15:19:18FromGitter<zacharycarter> hrm that's not what I'm finding, I must be doing something wrong
15:24:32skrylarmeh, will fiddle with hirise again some other time
15:35:35*gokr quit (Ping timeout: 260 seconds)
15:37:27*chemist69 quit (Ping timeout: 245 seconds)
15:47:26FromGitter<zacharycarter> @dom96: not sure what about my code is screwed up, but as a quick test I wrote this: https://gist.github.com/zacharycarter/0c3afefe9594a04ce7dfeed78a80f476
15:47:32FromGitter<zacharycarter> might be useful as a simple spawn example
15:50:41*yglukhov quit (Remote host closed the connection)
15:54:20*nsf quit (Quit: WeeChat 1.7.1)
15:59:30*skrylar wonders if += should be overloaded for UIs
15:59:41skrylarsomewindow += somebutton
15:59:59skrylarI mean it is the logical interpretation of summing a container with a control, but it's also sort of cryptic
16:00:03FromGitter<ephja> add/&/&=?
16:00:05fr1endwhy not & like strings?
16:00:17skrylarbecause i haven't used nim in years and i don't know what silly voodoo you do
16:00:26skrylarother than these like 4 nim files i wrote lol
16:00:51fr1endsame boat lol, had to spend ~hr paging it all back in before I could do anything
16:01:55skrylari probably won't use &= tho, personally
16:02:05skrylari prefer readability a lot (i mean, smalltalk <3)
16:02:25fr1end'readability' ;)
16:02:40*yglukhov joined #nim
16:02:46fr1endI don't dislike smalltalk but readable isn't what I'd call it
16:03:22krux02``newLit(1'i8) returns an int literal, not ant int8 literal
16:04:26skrylarIRC throwSticksAt: fr1end
16:04:39*chemist69 joined #nim
16:06:01*Trustable joined #nim
16:06:20*gokr joined #nim
16:06:32fr1end(defend self :with lisp)
16:07:24Trioxinin this example, why does sock have to be defined before the procs? https://github.com/rgv151/libssh2.nim/blob/master/examples/ssh_exec.nim
16:07:36Trioxinshouldn't it just be before the procs are used?
16:09:54*mwbrown quit (Quit: Exiting)
16:14:10Trioxinguess the compiler needs it
16:14:43*mwbrown joined #nim
16:16:27skrylarah poo. varargs in templates don't seem to let you make curry templates
16:16:42skrylartemplate (derp) = if something != nil: something(...)
16:25:11krux02well that's weird
16:25:42krux02int int32 int64 are all distinct types, but float and float64 are not
16:29:53skrylarmight depend on the build
16:30:26skrylarblah.
16:30:57skrylarI made a concept type, changed some list functions over to that, and instead of it (correctly) realizing that my type was compatible it just borked that it wasn't the same
16:32:05FromGitter<ephja> I wonder if I reported that
16:42:50*brson joined #nim
16:47:51*BennyElg joined #nim
16:52:17*gokr quit (Quit: Leaving.)
16:58:35BennyElgHey all, how I "Unpack" SELECT results into pre defined type and fill is parameters
17:00:51*yglukhov quit (Remote host closed the connection)
17:01:48*couven92 quit (Quit: Client disconnecting)
17:02:27*rokups quit (Quit: Connection closed for inactivity)
17:03:40demi-from a DB query you mean?
17:06:54BennyElgyes
17:07:25BennyElgfor example in go I build struct type with the rows parameters and then assign it to the row with Scan(&myStructInstance)
17:08:33demi-i think you want this module: https://nim-lang.org/docs/marshal.html
17:10:30BennyElgThanks!
17:11:22demi-hmm, i seem to have found a 404 page in the official documentation, which repo is the correct place to report this?
17:12:00BennyElgI cant do this? var rows = db.fastRows(sql"SELECT * FROM table")
17:12:41demi-i think it depends on what your backing store is?
17:14:27BennyElgAn Iterator Object just with nice presentation
17:14:38demi-oh, you may want this: `sql"SELECT id,<your key> FROM <table name>"`
17:14:41BennyElginstead of saying for row in con.fastRows....
17:14:45BennyElgjust saying for row in rows
17:18:08*nsf joined #nim
17:18:10skrylarfr1end, y'know, i just got thinking. i think it might be entirely possible to write smalltalk syntax in nim macros. lol
17:19:08fr1endthat would be impressive
17:19:11skrylari can't think of a good reason to *do* that, but
17:19:27skrylarthat's a lie
17:20:04skrylarif you had a scripting interpreter that read smalltalk syntax (it's not that hard; you could transpile ST-80 to javascript and run duktape, or to lua and run luajit) then say "well you can copy/paste it to this nim macro if it needs to be faster."
17:20:33skrylari know some people liked using angelscript for that reason. similar enough syntax to C++ that if they needed to speed up a function it needed very little editing
17:20:44skrylarhmm
17:31:51*adeohluwa quit (Quit: Connection closed for inactivity)
17:35:40*pilne joined #nim
17:36:11*krux02 quit (Ping timeout: 272 seconds)
17:36:20*krux02 joined #nim
17:40:17skrylarYou know. I suppose `parent=` could logically be suggested as "attach to this"
17:40:27skrylarIt's a weird way of doing it but it does read correctly
17:46:41*yglukhov joined #nim
17:49:32krux02I am procrastinating, and because I did not want to do what I should do, I created a pull request for Nim
17:49:52krux02I hope that is not offending anyone
17:51:08*yglukhov quit (Ping timeout: 258 seconds)
17:58:44FromGitter<zacharycarter> hrm I can't figure out if this is a bug or not, my flowvar says it's ready but it's still blocking for over a second sometimes, depending what I'm doing in the flow var
17:58:49FromGitter<zacharycarter> err sorry in the spawned proc I mean
17:59:26FromGitter<zacharycarter> I'll try to make a self contained example
18:05:26*mwbrown quit (Quit: Exiting)
18:07:04FromGitter<zacharycarter> https://gist.github.com/zacharycarter/97d04f9d864f40ce22cb4a1844e39f3b
18:07:19FromGitter<zacharycarter> the thread blocks for 0.051s after the flow var claims to be ready
18:11:32FromGitter<Varriount> @zacharycarter How much data is being copied?
18:12:28FromGitter<zacharycarter> in the flow var?
18:12:38FromGitter<Varriount> Yes.
18:12:43FromGitter<zacharycarter> Hrm just the parsed json
18:12:55FromGitter<Varriount> Not the tileset information?
18:13:05FromGitter<zacharycarter> not in this example
18:13:09*mwbrown joined #nim
18:13:21FromGitter<zacharycarter> this example is just contrived, in my actual code the blocking is happening for much longer ~2 seconds sometimes
18:13:30FromGitter<zacharycarter> but the code is structurally similar
18:13:33FromGitter<Varriount> You might find it worthwhile to run some sort of C benchmarking tool
18:13:46FromGitter<zacharycarter> if anyone has any suggestions for one I'd appreciate it
18:14:00FromGitter<Varriount> You're on a Mac, right?
18:14:03FromGitter<zacharycarter> yessir
18:19:28krux02are here Nim maintainers?
18:21:57*vlad1777d quit (Quit: Leaving)
18:22:05*yglukhov joined #nim
18:22:42FromGitter<Varriount> @zacharycarter You can try instruments. It's part of xcode
18:22:56demi-instruments is very good forcodeprofiling
18:22:56FromGitter<Varriount> krux02: Araq, dom96, and zah are the maintainers
18:25:15FromGitter<zacharycarter> hrm yeah trying to get my head around it now
18:25:32FromGitter<zacharycarter> my app creates log files and every time I start up my app in instruments it crashes saying it can't find the file
18:25:55*chemist69 quit (Ping timeout: 246 seconds)
18:27:00demi-zacharycarter, how are you attaching/starting it up?
18:27:09*chemist69 joined #nim
18:27:20FromGitter<zacharycarter> I'm just directing time profiler to the binary
18:27:29FromGitter<zacharycarter> ah I see
18:27:32FromGitter<zacharycarter> I need to specify a working directory
18:28:47*dddddd joined #nim
18:41:25*yglukhov quit (Remote host closed the connection)
18:42:06*yglukhov joined #nim
18:49:12*Tiberium quit (Remote host closed the connection)
18:49:38krux02well I hate at the moment that the array size is not just an integebr but this weird range thing
18:49:54krux02how do I get the bounds of such a range?
18:52:00FromGitter<Varriount> krux02: You can specify the size of an array..
18:52:03vivushigh - low ?
18:52:33FromGitter<Varriount> But yes, high(arr) will give you the highest index, and low(arr) the lowest
18:56:42krux02well I am in generic code
18:56:50krux02I post an example
18:57:10krux02http://ix.io/uG2
18:57:35krux02how do I use N1 and N2?
18:57:56krux02if arrays would be just a length, I could simply add N1 and N2
18:59:01krux02scratch.nim(11, 81) Error: ambiguous call; both system.+(x: int, y: int) and system.+(x: float, y: float) match for: (N1.high, N2.high)
19:00:03*xet7 joined #nim
19:02:14FromGitter<zacharycarter> not sure what to make of this...
19:02:15FromGitter<zacharycarter> (https://files.gitter.im/nim-lang/Nim/T1GD/Screen-Shot-2017-05-23-at-3.01.49-PM.png)
19:02:23FromGitter<zacharycarter> that's the section during which the main thread seems to be blocked
19:03:12FromGitter<ephja> an addInt function?
19:03:59*yglukhov quit (Remote host closed the connection)
19:04:10FromGitter<zacharycarter> I just picked a random point in time during the programs execution where the main thread was blocked
19:04:22FromGitter<zacharycarter> I have no idea how to use this tool / figure out what to do about the thread blocking
19:04:53*yglukhov joined #nim
19:05:22FromGitter<Varriount> @zacharycarter If you look at the call stack, you can see it's part of a deep copy operation
19:05:25demi-zacharycarter, what analysis tool in instruments are you running this under?
19:08:38demi-zacharycarter: http://help.apple.com/instruments/mac/8.3/ this will probably be helpful if you have never used this before
19:08:38*yglukhov quit (Remote host closed the connection)
19:11:04FromGitter<Varriount> krux02: bing
19:11:05FromGitter<Varriount> https://gist.github.com/Varriount/7df2090adc52e9832647a3fb7b00e399
19:11:58*yglukhov joined #nim
19:12:20FromGitter<Varriount> @zacharycarter Looking at the call stack, it appears `^` (whatever that does) is copying data from one thread to another
19:12:58FromGitter<Varriount> You might try using a pointer to string or a cstring
19:13:32FromGitter<zacharycarter> @Varriount gotcha thank you I'll try that
19:13:41FromGitter<zacharycarter> demi- : thanks
19:14:49FromGitter<Varriount> @zacharycarter Could I see the code you're using?
19:15:23*krux02 quit (Remote host closed the connection)
19:15:29FromGitter<zacharycarter> sure
19:15:40FromGitter<zacharycarter> @Varriount let me commit really quickly
19:18:19FromGitter<zacharycarter> @Varriount there's quite a bit of code here but I"ll try to point out the relevant sections
19:19:12FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/examples/desktop/08-tiled-map/main.nim#L60
19:19:31FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/examples/desktop/08-tiled-map/main.nim#L109
19:19:50*yglukhov quit (Remote host closed the connection)
19:20:12FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/src/frag/modules/assets.nim#L165
19:20:30FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/src/frag/modules/assets.nim#L313
19:21:01FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/src/frag/maps/tiled_map.nim#L217
19:23:14FromGitter<Varriount> Is there a single section dealing with the spawn logic?
19:23:23FromGitter<ephja> those darn suffixes
19:24:55FromGitter<Varriount> @ephja Still haven't found the cause of that in the highlighting files. Sorry.
19:24:59*Matthias247 joined #nim
19:25:09FromGitter<zacharycarter> @Varriount yeah one sec
19:25:23FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/src/frag/modules/assets.nim#L313 issues the spawn call
19:26:04FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/src/frag/modules/assets.nim#L313
19:26:09FromGitter<zacharycarter> checks to see if the FlowVar is ready
19:26:29FromGitter<zacharycarter> https://github.com/fragworks/frag/blob/html5/src/frag/modules/assets.nim#L313 is called once it's ready
19:26:41FromGitter<Varriount> Why are you doing all those casts? You shouldn't need to cast anything
19:28:36FromGitter<zacharycarter> omg I just linked the same line 3 times lol my bad
19:28:58FromGitter<zacharycarter> I figured I needed to cast the FlowVarBase instances into their actual types
19:30:58FromGitter<zacharycarter> not sure how to do it without a cast
19:31:40FromGitter<Varriount> Wait, you're storing results from 'spawn' s of differing return types in one sequence?
19:32:13FromGitter<zacharycarter> yes
19:32:54FromGitter<zacharycarter> well kind of
19:33:23FromGitter<zacharycarter> I eventually end up populating the assets sequence
19:33:52FromGitter<zacharycarter> only some of the spawn calls return ref Assets, others return other ref types and then I create and insert a ref Asset after I've retrieved the result from the spawn call
19:36:07*Ven joined #nim
19:36:30*Ven is now known as Guest41355
19:38:13*Arrrr quit (Read error: Connection reset by peer)
19:43:59FromGitter<Varriount> @zacharycarter My best guess is that your asset data is being deep copied
19:44:06*Guest41355 quit (Ping timeout: 272 seconds)
19:44:41*yglukhov joined #nim
19:45:22FromGitter<zacharycarter> hrm
19:45:33FromGitter<zacharycarter> I can try to do a shallow copy instead
19:46:10*Ven_ joined #nim
19:46:25*Vladar quit (Remote host closed the connection)
19:49:35FromGitter<zacharycarter> not sure I have any control over that actually
19:50:33FromGitter<zacharycarter> I notice the sooner I get out of the spawned procedure, the less time things are locked up for
19:52:24FromGitter<zacharycarter> in fact @Varriount if I comment out this section of code: https://github.com/fragworks/frag/blob/html5/src/frag/maps/tiled_map.nim#L232-L258 there's basically no blocking
19:54:56FromGitter<Varriount> @zacharycarter https://nim-lang.org/docs/threadpool.html#^,FlowVar[ref.T]
19:55:06FromGitter<Varriount> blocks until the value is available and then returns this value.
19:56:02FromGitter<zacharycarter> yeah but I check to see if the flow var is ready first
19:56:04FromGitter<zacharycarter> before I call that
19:56:31FromGitter<zacharycarter> so the proc should have already completed prior to me calling ^ regardless
19:59:08*Sentreen quit (Read error: Connection reset by peer)
19:59:23*Neomex joined #nim
19:59:47FromGitter<ephja> @Varriount ok :smile:
20:01:14*Sentreen joined #nim
20:02:53*Ven_ quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
20:03:12*Sentreen quit (Client Quit)
20:03:22*Sentreen joined #nim
20:18:12*Trustable quit (Remote host closed the connection)
20:21:26skrylarhrm.
20:21:31skrylarits telling me that float.high is invalid
20:21:39skrylaris there no such thing as maxfloat?
20:23:54*bjz joined #nim
20:25:48*noethics quit (Remote host closed the connection)
20:26:08*noethics joined #nim
20:26:19FromGitter<Varriount> I don't think so
20:32:45FromGitter<Varriount> @zacharycarter I'd try refactoring/commenting your code as much as reasonable, then possibly file an issue
20:33:55PMunchskrylar, Varriount, of course there's a max
20:36:00PMunchThe highest representable in a float is about 3.402823 × 1038
20:37:36*PMunch quit (Quit: leaving)
20:41:31*vivus quit (Quit: Leaving)
20:45:39*gokr joined #nim
20:59:30*Sembei joined #nim
20:59:39*dddddd quit (Remote host closed the connection)
20:59:39*smt quit (Remote host closed the connection)
20:59:47*smt joined #nim
21:01:25demi-there should be float-max and float-min
21:03:51*Ven joined #nim
21:04:15*Ven is now known as Guest25118
21:12:50gokrI still think it would be awesome if Nimmers could work together on a good Nim wrapper of Urho3D. I do note that Miguel de Icaza and Microsoft/Xamarin also realized Urho3D is a good engine (they are actively maintaining a wrapper now for).
21:13:50FromGitter<Varriount> gokr: https://github.com/3dicc/Urhonimo ?
21:14:08gokrYeah, but it would need to get updated and improved.
21:15:01gokrI see so many here building their own game engines in Nim (which I of course applaud and can relate to, always cool to build your own) - but standing on top of Urho3D is a solid superb base.
21:15:35gokrAnd Urho3D relies on basically the same base libraries that I see mentioned here over and over (SDL2, Assimp, Bullet etc etc)
21:21:32*PMunch joined #nim
21:22:38*couven92 joined #nim
21:26:51PMunchdom96, there's something wrong with Jester
21:26:58PMunchresp doesn't terminate a route any longer
21:30:37FromGitter<zacharycarter> gokr: I built frag mostly as a learning experience, and also with the goal to build something that's usable to others. Using someone's engine is great but at a certain point you're not learning how to write an engine, you're learning how to write game code.
21:31:18FromGitter<zacharycarter> Or you're learning the specifics of that framework / engine
21:31:29*Guest25118 quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
21:32:00gokrYes, and do note my careful words - I just feel that... well a) building a really good game engine (as Urho3D is) is a LOT more work than people may think and b) if we had a stellar Nim wrapper of Urho3D that could very well be a "killer app" for Nim.
21:32:10FromGitter<zacharycarter> I don't disagree with you that bindings to something like Urho3d, or that engine that def-pri-pub mentioned last night would be nice - I might even use them - but it still wouldn't scratch the itch for learning that level of thigns yourself
21:32:38gokrI know! :)
21:33:05PMunchAnother benefit of having it in pure Nim is that the Nim compiler is able to not treat it as a black box
21:33:33PMunchI was working with wxWidgets (C++ bindings) recently and it is also cross-platform
21:33:48PMunchWhich means that it wraps everything into it's own types
21:34:10PMunchWhich is great if you use wxWidgets in C++, but in Nim it's a bit annoying..
21:34:40*skrylar quit (Remote host closed the connection)
21:35:01gokrOne idea would be to actually port Urho3D to Nim - but in order for that to not just turn into a "curiosity" it would probably mean you need a few of the core Urho3D devs convinced that Nim is better than C++.
21:35:19gokrAnd of course, it would feel (to many) as a lot of work for no big award.
21:35:40FromGitter<zacharycarter> might as well just write our own game engine in Nim then
21:35:49PMunchYeah, we came to that conclusion with wxWidgets as well
21:36:06PMunchIt would be better to have something Nim native, but it would be a lot of work
21:36:11dom96PMunch: bug report please
21:36:13gokrWell, as I said, you may be game engine wizards - but I just have a tremendous respect for the guys behind Urho3D.
21:37:03FromGitter<zacharycarter> Urho3d is a nice engine for sure, just a lot of work to build and maintain bindings for it
21:37:21FromGitter<zacharycarter> but if you want to make them I'd surely help test them
21:37:25couven92PMunch, though the interesting thing with native GUI is the cross-platform challanges... I actually liked our wxWidgets work, it was fun... :)
21:37:30demi-dom96: there is a 404 page on the official docs (https://nim-lang.org/docs/sexp.html) where should i be reporting this?
21:38:33PMunchAah, the Jester bug seems to be with templates
21:38:54gokrzacharycarter: You are aware of Urhonimo I guess?
21:40:28FromGitter<dom96> Dmi
21:40:36FromGitter<zacharycarter> gokr: I am
21:40:42FromGitter<dom96> Demi-: Nim repo
21:40:52demi-ok will do, thanks
21:41:02FromGitter<zacharycarter> I considered using it before writing my own engine and Araq suggested I do the latter instead of trying to bind to a C++ engine with Nim
21:41:15FromGitter<zacharycarter> he said he thought it'd be more worhtwhile
21:41:17FromGitter<zacharycarter> worthwhile*
21:41:45FromGitter<zacharycarter> since I was attempting to learn engine dev I agreed and started doing a bunch of OpenGl experiments with Nim and then eventually started frag
21:45:39PMunchHmm, I'm having a hard time actually figuring out what's wrong dom96
21:45:57PMunchI've got two templates that basically wrap some tests up in a try catch
21:47:17PMunchWait, I think I've got it
21:48:42gokrzacharycarter: Cool
21:49:41gokrWell, I just skimmed a bunch of discussions on the Urho3D forum and I quickly realized (not surprising) that they are (as all core game engine guys) C++ all the way.
21:50:00gokrAnd for good reasons, all the libraries they bring together are C++ AFAIK.
21:50:22subsetparkSo I have a Python that I've written in Nim - the way I've been installing it thus far is to run the nim compiler on the machine during install. But I have set things up so that I can distribute just the C source files and distribute them as a normal python extension. The only thing missing is nimbase.h.
21:50:30subsetparkShould I just include nimbase.h in my package?
21:51:34*xet7 quit (Remote host closed the connection)
21:52:00gokrzacharycarter: So I should forget the idea of a good Nim wrapper - since every time I bring it up I get basically zero response. :)
21:53:29FromGitter<Varriount> Someone might try one day.
21:53:40FromGitter<Varriount> The problem is that interfacing with C++ is such a pain.
21:54:02FromGitter<Varriount> Even if you get a binary interface, you have to deal with all the objects - vectors, etc.
21:54:58couven92guys... crazy alert: HOw about NimCUDA?
21:55:00PMunchdom96, https://github.com/dom96/jester/issues/114
21:55:17couven92do we have that?
21:55:46gokrVarriount: The Urhonimo showed it can be done IMHO. It even includes some fairly advanced demos.
21:56:33FromGitter<zacharycarter> gokr: my advice is if you want it you're probably going to have to start it yourself
21:56:44gokrzacharycarter: We did
21:56:45PMunchcouven92, well there is this: https://github.com/jcosborn/cudanim
21:56:59FromGitter<zacharycarter> yeah but it wasn't maintained
21:57:14*_orangejuice joined #nim
21:57:28gokrzacharycarter: I agree, due to various factors we had to put it on ice.
21:57:30FromGitter<zacharycarter> if you're speaking to UrhoNimo
21:57:43*_orangejuice left #nim (#nim)
21:57:54FromGitter<zacharycarter> I think there's interest but there also has to be a driving force behind it
21:58:00gokrI just mean - we did "start" it. But noone jumped in.
21:58:15FromGitter<zacharycarter> another interesting engine is this one
21:58:17FromGitter<zacharycarter> which is pure C
21:58:54FromGitter<zacharycarter> hrm need to look through irc logs
21:59:21FromGitter<zacharycarter> http://www.raylib.com/
21:59:23gokrWhen we elected Urho3D for our work at 3DICC I looked at tons of engines and Urho3D stood easily out.
22:04:51gokrWell, from a cursory look I don't think it's even close to Urho3D. Btw, to see some examples live (emscripten) look at: https://urho3d.github.io/samples/
22:05:21gokrFor example: https://urho3d.github.io/samples/19_VehicleDemo.html
22:05:38*brson quit (Ping timeout: 260 seconds)
22:06:21*brson joined #nim
22:07:10FromGitter<zacharycarter> I don't see any reason you couldn't build something similar to that with raylib
22:07:23FromGitter<zacharycarter> but I don't know I haven't used either engine
22:07:27FromGitter<zacharycarter> both seem fairly capable
22:08:01*Matthias247 quit (Read error: Connection reset by peer)
22:10:46FromGitter<zacharycarter> the nice thing about raylib again is it's pure C which is super easy to interface with
22:14:05gokrNo skeletal animation, no networking, no directx and tons more in Urho3D that I don't see in Raylib. BUT... having said that, of course, C is nice and simple.
22:14:20gokrBut I dare say it's not comparable.
22:16:31gokrI knew very little of game engines before - but I now have a different perspective of all the work going into them. Just stuff like all the shaders, advanced culling techniques, yaddayadda...
22:17:45*bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
22:19:42FromGitter<zacharycarter> @Varriount remember how you said to file an issue?
22:19:49FromGitter<zacharycarter> what do you think of this as an example ?
22:20:14FromGitter<zacharycarter> https://gist.github.com/zacharycarter/7ab2b31c1e1943241140ce6be7277fca
22:25:08*subsetpark quit (Ping timeout: 240 seconds)
22:25:29*subsetpark joined #nim
22:28:31FromGitter<zacharycarter> I made a better one
22:32:47*chemist69 quit (Ping timeout: 260 seconds)
22:32:49FromGitter<zacharycarter> https://github.com/nim-lang/Nim/issues/5878
22:34:56*chemist69 joined #nim
22:37:36*PMunch quit (Quit: leaving)
22:48:12*Jesin quit (Quit: Leaving)
22:54:27*libman joined #nim
22:57:17*Jesin joined #nim
23:01:03*daaf joined #nim
23:10:07Trioxinwhat would be awesome to have in Nim would be persistent, linked data collections
23:15:43*smt_ joined #nim
23:17:08*gokr quit (Ping timeout: 240 seconds)
23:17:36*nsf quit (Quit: WeeChat 1.7.1)
23:19:16*smt quit (Ping timeout: 255 seconds)
23:19:17*yglukhov quit (Remote host closed the connection)
23:21:02*couven92 quit (Quit: Client Disconnecting)
23:33:38*BennyElg_ joined #nim
23:33:40*BennyElg quit (Read error: Connection reset by peer)
23:35:12Trioxinwould it be safe to monitor many directories with walkfiles?
23:38:16Trioxingobs and gobs