<< 16-04-2018 >>
00:11:16FromGitter<Varriount> Anyone here know some fast/efficient xml libraries, possibly in C, that I could use?
00:11:32krux02tinyxml
00:11:40FromGitter<Varriount> Right now I've plans to use Nim's xmlparser, however I would like to compare alternatives.
00:11:43krux02but I don't know it, i just heared about it
00:12:27krux02well when you want a fast one it should not translate the entire file into a tree structure with xml node objects on parsing
00:12:52FromGitter<Varriount> Sorry, I meant parsexml: https://nim-lang.org/docs/parsexml.html
00:18:13FromDiscord<claude> oh i see, the pragma for allowing unused is `used`
00:57:46*user1101_ joined #nim
00:59:46*user1101 quit (Ping timeout: 264 seconds)
00:59:46*user1101_ is now known as user1101
01:45:08FromGitter<gogolxdong> How to let the proc spawn but return something first ?
01:45:45FromDiscord<claude> defer: spawn?
01:46:44FromGitter<gogolxdong> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5ad400845f188ccc156b4cd3]
01:47:16FromGitter<gogolxdong> want it return "done" and execute the spawn proc in background.
01:48:56*rockcavera joined #nim
01:53:37FromGitter<Varriount> You need to use flowvars
01:59:00FromGitter<gogolxdong> what's that?
02:08:05FromGitter<gogolxdong> @Varriount can you show me an exmaple code of how to new and operate this type?
02:24:17FromGitter<gogolxdong> I want to return something first than the spawning proc.
02:26:22FromGitter<gogolxdong> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5ad409ce7c3a01610de6f97b]
02:27:24FromGitter<gogolxdong> Learned from Nim source code of how to use FlowVar type but didn't find out how to achieve my goal.
02:38:54*dddddd quit (Remote host closed the connection)
03:02:37petersjt014[m]sometimes I run a comamnd with `exexCmdEx` and get a single newline back. Is there a common way to replace it with the null string to make error handling less obnoxious?
03:09:00*krux02 quit (Remote host closed the connection)
03:15:28petersjt014[m]the main thing I'm doing is validating the length later, so I was planning on using `.replace(re"^\s+$", "")` if there is nothing more appropriate
03:16:12FromGitter<Varriount> Why not just use setlen?
03:17:27FromGitter<Varriount> petersjt014[m]: You might also think about using the strip() procedure from strutils to clip whitespace off the ends of a string
03:17:48*tiorock joined #nim
03:17:48*tiorock quit (Changing host)
03:17:48*tiorock joined #nim
03:17:48*rockcavera is now known as Guest8634
03:17:48*Guest8634 quit (Killed (karatkievich.freenode.net (Nickname regained by services)))
03:17:48*tiorock is now known as rockcavera
03:17:49FromGitter<Varriount> I also have a version I can post which modifies the string in-place.
03:26:38petersjt014[m]Can I end up doing something like `var (r0, _) = execCmdEx interp"ping $idklol[0].strip"` then?
03:27:18FromGitter<Varriount> Remind me again, what values does execCmdEx return?
03:30:25FromGitter<Varriount> petersjt014: You can do that, if you want to discard the exit code (which is usually bad practice)
03:30:47petersjt014[m]A string (or TaintedString I'm told) which I (I think) am then turning into a seq of strings with split().
03:34:24petersjt014[m]and it becomes just " " if nothing is returned, which has length > 0 which is throwing me off.
03:36:00petersjt014[m]Would a case statement be reasonable here?
03:36:01FromGitter<Varriount> why not use a `if` to check then? or if you're doing it multiple times, use a template?
03:36:18petersjt014[m]or that
03:36:28FromGitter<Varriount> ``````
03:37:04FromGitter<Varriount> ```var (r0, _) = execCmdEx ... ⏎ if r0 == "": ... ⏎ else: ...``` [https://gitter.im/nim-lang/Nim?at=5ad41a6015c9b03114fe76f2]
03:38:08petersjt014[m]Yeah, I'll try that
03:39:23FromGitter<Varriount> Hm, now to fix doc-comment continuation in NimLime
03:40:35petersjt014[m]sounds fun
03:54:58FromDiscord<claude> i'd like an option for nim check arguments in NimLime, for example i need `-d:ssl` for my project to work, so i have to do `nim check -d:ssl` manually
04:06:38FromGitter<gogolxdong> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5ad4214d109bb04332c75fe3]
04:08:50FromGitter<gogolxdong> type mismatch.
04:15:00FromGitter<gogolxdong> Is that FlowVar type doesn't support table?
05:10:21*yglukhov quit (Remote host closed the connection)
06:12:04*PMunch joined #nim
06:17:00*nsf joined #nim
06:37:25FromGitter<Varriount> I don't believe so.
06:38:03FromGitter<Varriount> claude: Right now I'm in the process of a major refactor. If you need extra arguments though, you should probably put it into a project file
06:39:13FromGitter<Varriount> claude: https://nim-lang.org/docs/nims.html
06:40:37FromGitter<Varriount> My priorities right now are: fixing/refactoring the "nim check" feature, fixing doc-comment completion, fixing the project settings code, then adding project creation and build support.
06:41:14FromGitter<Varriount> Also, I'm removing Sublime Text 2 support, so if you're still using that, I recommend switching to Sublime Text 3
06:58:27*yglukhov joined #nim
07:01:12yglukhovtreeform: aye, pure nim ttf support would be awesome! But lately I was more thinking about using platform native api for bitmap generation.
07:11:25*yglukhov quit (Remote host closed the connection)
07:13:35*rokups joined #nim
07:43:35*floppydh joined #nim
07:45:01FromGitter<alehander42> @krux02 I don't like this def, non-turing completeness is a thing, also stuff like powerpoint is turing complete and it's not a proglang
07:46:47FromGitter<gogolxdong> `ambiguous call; both tables.[]=(t: OrderedTableRef[[]=.A, []=.B], key: A, val: B)[declared in lib\pure\collections\tables.nim(770, 5)] and tables.[]=(t: TableRef[[]=.A, []=.B], key: A, val: B)[declared in lib\pure\collections\tables.nim(428, 5)] match for: (Error Type, string, Future[json.JsonNode])`
07:50:43FromGitter<gogolxdong> It's ambiguous even specify to OrderedTableRef.
07:59:55FromGitter<abijahm> @gogolxdong example using tableand flowvar ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5ad457fb270d7d3708cff71d]
08:05:10FromGitter<gogolxdong> could it be newTablestring,FlowVar[Future[JsonNode () ?
08:09:59*sendell joined #nim
08:15:24FromGitter<gogolxdong> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5ad45b9c109bb04332c84a51]
08:15:30*Vladar joined #nim
08:15:45FromGitter<gogolxdong> type mismatch: got <JsonNode> ⏎ ⏎ but expected one of: ⏎ template `^`(x: int): BackwardsIndex ⏎ proc `^`T (fv: FlowVar[T]): T ... [https://gitter.im/nim-lang/Nim?at=5ad45bb12b9dfdbc3a956b4c]
08:24:55*xkapastel quit (Quit: Connection closed for inactivity)
08:27:12FromGitter<gogolxdong> Does FlowVar conflict with Future?
08:27:23FromGitter<abijahm> @gogolxdong using future ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5ad45e6b27c509a77417af34]
08:29:35FromGitter<mratsim> I need a set/hashset with the following properties: ⏎ ⏎ 1) Can store up to `~500 elements` at most (covered by `set` and `Hashset`) ⏎ 2) `Incl` and `Excl` as fast as possible (in hot path) (covered by `set` and `Hashset`) ⏎ 3) Can take the `length` without iterating (in hot path) (covered by `Hashset`, can use a `int16 length field` with `set`) ... [https://gitter.im/nim-lang/Nim?at=5ad45eef2dfcfffd2b2c7148]
08:30:54FromGitter<mratsim> It can be on stack or heap, it’s only allocated once per thread and can be reused for the lifetime of the program.
08:31:12FromGitter<mratsim> This is for my Go playing bot.
08:46:07*yglukhov joined #nim
08:46:33Araqmratsim: use an array
08:48:01FromGitter<narimiran> btw, Araq, why we don't have `pop` for sets?
08:48:17FromGitter<mratsim> you can’t pop an unordered dat structure
08:48:27FromGitter<narimiran> python can :P
08:48:39FromGitter<mratsim> it’s probably slow then :D
08:49:01FromGitter<narimiran> `pop` == "take a random value" ;)
08:49:13FromGitter<ChristianWitts> nah, it removes and returns an arbitrary set element
08:49:22FromGitter<ChristianWitts> so it’s O(1)
08:49:35FromGitter<mratsim> oh, I’ll check out their implementations then.
08:50:46*yglukhov quit (Ping timeout: 264 seconds)
08:51:44Araqnarimiran: because nobody wrote it
08:52:06FromGitter<mratsim> I see how I can do it with 2 arrays but with a single array I don’t see how to take a random value in the set without allocating a temporary with only the set values.
08:52:09Araqfeatures have that nasty tendency to start in the "not implemented" state
08:52:56Araqlol that PR queue will never get below 50 anymore
08:53:06FromGitter<narimiran> ok, i might give it a shot and try to implement it, but i'm not sure i understand how it should be done....
08:53:09Araqor the issues below 1000
08:53:49FromGitter<ChristianWitts> @mratsim https://github.com/python/cpython/blob/master/Objects/setobject.c#L705 ⏎ for the set.pop implementation
08:54:00FromGitter<narimiran> issues below 1200 seems impossible too
08:54:22FromGitter<mratsim> you beat me to it 👍
08:57:54FromGitter<mratsim> @ChristianWitts @miranPython is using a HashSet, that can be added to Nim Hashsets.
08:58:36FromGitter<mratsim> However nim “sets” are bitset, I don’t think you can pop at random efficiently on them
08:59:25FromGitter<narimiran> adding it only to HashSets would be a nice step forward!
09:01:55FromGitter<gogolxdong> @abijahm thanks very much.
09:45:43*user1101_ joined #nim
09:46:59*user1101 quit (Ping timeout: 256 seconds)
09:46:59*user1101_ is now known as user1101
09:47:17*rockcavera quit (Ping timeout: 260 seconds)
09:55:22Araqugh, I want to remove overloaded dot operators, they don't work.
09:56:40Araqmyjson.key.getStr # myjson["key"]["getStr"] or myjson["key"].getStr ?
09:57:44Araqwe can instead introduce a binary operator that binds like '.' such as '.?'
09:57:54Araqmyjson.?key.getStr
09:59:20Araqor myjson.@key or whatever
09:59:48Araqoverloaded dot operators turn the static guarantees Nim gives into a loophole fest
10:02:20FromGitter<alehander42> `.?` looks like a safe navigation operator
10:04:29FromGitter<alehander42> also, I don't think the json module overloads `.` ?
10:04:45Araqno my coworker did :P
10:05:50FromGitter<alehander42> well, one should just use op overloading with care :D each operator can be overloaded in a very confusing way
10:06:18FromGitter<alehander42> a lot of the existing js backend code depends on some kinf of `.` overloading somewhere deep :D
10:06:24Araqit's overloaded in the way the manual suggests is a good idea
10:06:36dom96This is why all those people saying "But what if someone does this..." might have a point :P
10:07:20Araqdom96: no they have a point when my nose tells me the same :P
10:07:40dom96Yes, but if it wasn't for this you wouldn't have known
10:07:45AraqI never liked overloading '.' to begin with.
10:07:49FromGitter<narimiran> `myjson@key`? (without the dot)
10:09:14Araqthe underlying design problem here is the goal to find a "consistent" interpretation for everything that otherwise has none.
10:09:23FromGitter<alehander42> I don't think this is such a huge problem for the js overload `.`
10:09:45FromGitter<alehander42> because it can happen when you have `.x` and `proc x(s: js)`
10:10:02FromGitter<alehander42> but you shouldn't really write most of your functions to accept `js`: they should expect normal nim types
10:10:22Araqthese interpretations that otherwise have no meaning are really important. These are the bug catchers.
10:10:28FromGitter<alehander42> so naturally your code should "close" the dynamic `.` usage in a cast or something like that
10:14:31Araqwhen I read Nim code I want to read Nim code and not some shit that is as fragile as JavaScript...
10:15:06dom96Actually, when I think about it
10:15:16dom96Shouldn't jsgen be a macro?
10:15:30dom96If your weird `.` overloading is in a macro then at least you know it's limited to that macro invocation
10:16:15Araqdom96: I'm not following. 'jsgen' ?
10:17:09dom96er, jsffi is what I mean
10:17:41FromGitter<alehander42> so `jsffi(a.b.c)` ?
10:17:48Araqoh look I found a meaning for "foo" "bar". it should be an implicit concatenation. handy, huh? no, it sucks. ("abc" "def", "xyz")
10:18:04*xet7 joined #nim
10:19:26FromGitter<alehander42> honestly I think this whole `.` thing can be resolved by a good editor plugin/IDE which highlights in a special way overloaded `.`-s
10:20:16Araqno, we can use the Unicode mid-dot instead of the dot and let the editor render it as a colored '.'
10:21:00FromGitter<alehander42> still better than `@` :D
10:21:15Araqor something I happen to be able to reach on my keyboard, foo∞bar
10:21:18dom96no, semantic highlighting is difficult
10:21:51Araqfoo…bar
10:22:10FromGitter<alehander42> that just looks like the range operator
10:22:26Araqfoo/bar
10:22:40Araqfoo\bar
10:22:56*yglukhov joined #nim
10:23:06Araqso many options, but no, we have to use the freaking same character.
10:23:13*couven92 joined #nim
10:23:31FromGitter<alehander42> `/` can be fine, also incredibly surprising for everybody who has used `/` as divide
10:23:47FromGitter<narimiran> `|`
10:23:48*yglukhov quit (Read error: Connection reset by peer)
10:23:55AraqI use os.`/` much more often than devide…
10:24:02FromGitter<alehander42> the php people use `\` for namespaces indeed, that could work
10:24:37*yglukhov joined #nim
10:24:40FromGitter<alehander42> you use os.`/` when the args are strings/paths
10:24:58FromGitter<alehander42> but the problem is more complicated
10:25:01Araqyeah my code usually uses paths more often than divisions
10:25:12FromGitter<alehander42> jsffi also needs to overload `.=` and `.()`
10:25:39Araqwhich is why tying operators to math is academic.
10:26:09Araqwell jsffi can use a macro once my 'semityped' macros are a thing
10:26:37FromGitter<alehander42> :D do you have an example semityped macro somewhere
10:26:56AraqI've been wanting to write an RFC for quite some time now
10:28:33FromGitter<alehander42> still, I can hardly find enough benefit in that, the manual should just outline the dangers of overloading `.` and to motivate people to wrap /convert most of their js "dynamic" code/values to typed stuff
10:30:04FromGitter<alehander42> even if you use a macro `.` you can easily cast it in the wrong way and overally all that casting/conversion should be centered in as few places as possible
10:30:52*yglukhov quit (Read error: Connection reset by peer)
10:31:03FromGitter<alehander42> maybe if the js `.` wasn't existing until now, adding a diff operator would be great, but I am afraid this can break a lot of code
10:31:33Araqwell there will be a deprecation period
10:32:01*yglukhov joined #nim
10:32:50Araqbut Nim in a nutshell is/should be "simple core for systems programming" + "macro system".
10:33:43dom96Araq: Why does jsffi need semityped?
10:33:45FromGitter<alehander42> obvsly + "type safety" :D
10:34:35Araqdom96: it doesn't but 'semityped' allows overloading of '.' within a macro implementation. unfortunately, but unavoidably.
10:34:58Araqit would be like
10:35:24Araqproc foo() {.jsdots.} = j.foo.bar
10:36:17FromGitter<mratsim> I’m all for removing ‘.’ overloading, but remember the poor guys remoting and only having “vi” without nimsuggest hints (i.e. remoting in a raspberry Pi or a server)
10:36:39FromGitter<mratsim> we can’t always rely on a “modern editor"
10:36:58FromGitter<mratsim> especially when those always come with 200MB of chrome dependencies ...
10:37:07dom96yep
10:37:09Araqthat's a different problem, but yes, I remember these poor souls.
10:38:14dom96Araq: Isn't it just a case of rewriting the expression?
10:38:43dom96jsffi(j): j.foo.bar -> j["foo"]["bar"]
10:39:23Araqdom96: sure. depending on how we do it.
10:39:32Araqbut it's a good point, it doesn't even need semityped.
10:40:19Araqjsffi(j.foo.bar.())
10:40:25Araqjsffi(j.foo.bar())
10:50:33FromGitter<zacharycarter> https://www.youtube.com/watch?v=_ahvzDzKdB0 - interesting talk
10:56:00*arecaceae quit (Remote host closed the connection)
10:56:23*arecaceae joined #nim
11:06:14*Torro joined #nim
11:17:51dom96Thanks for the invite niv :D
11:17:53*BitPuffin joined #nim
11:19:32FromGitter<Bennyelg> @dom96 whats wrong with the hierarchy of the project ? looks like nimble not recognise my new version ⏎ https://github.com/Bennyelg/nimPresto
11:20:23dom96Bennyelg: nimble will download the latest tagged version by default
11:20:25dom96not HEAD
11:20:44FromGitter<Bennyelg> but my tag refered to the master which is updated
11:20:44dom96if you want to install HEAD then use `nimble install db_presto@#head`
11:20:46FromGitter<Bennyelg> but still
11:20:55FromGitter<Bennyelg> version 1.0.0 is downloaded somehow
11:21:05*Vladar quit (Ping timeout: 240 seconds)
11:21:31FromGitter<Bennyelg> I refreshed my tag
11:21:34FromGitter<Bennyelg> now its working
11:21:37FromGitter<Bennyelg> thanks.
11:22:06dom96Here is a handy command for you: nimble search --ver db_presto
11:22:35FromGitter<Bennyelg> Fantastic, thanks.
11:23:39*Vladar joined #nim
11:23:56couven92Shouldn't Nim also look for nim.cfg in %ALLUSERSPROFILE% (i.e. C:\ProgramData) which basically is C:\Users\username\AppData\Roaming for all users on Windows?
11:24:03couven92I'd say yes! :P
11:24:56couven92It would be like having nim.cfg in /etc/nim.cfg on Linux...
11:25:09nivdom96: araq told me to
11:26:19dom96great :)
11:27:21couven92The benefit would be that it would present a machine-wide Nim configuration location that is not overwritten by a Nim install... My easy Nim install procedure on Windows is: Delete C:\Program Files\Nim download binaries from nim-lang.org and unzip into C:\Program Files and then rename nim-<version> to Nim
11:28:04couven92Point is: The download from nim-lang.org contains the default nim.cfg, and I do not want to change that every time with a new install...
11:29:59couven92On Windows other programs (e.g. Git, Java JRE, Docker, ...) also place their machine-wide configuration into C:\ProgramData (if they do not use the registry)
11:30:40dom96couven92: sounds good to me
11:31:20couven92PR it is then... :P
11:32:43dom96Araq: https://www.youtube.com/watch?v=_ahvzDzKdB0
11:34:15dom96couven92: Note: Araq might disagree
11:35:37couven92dom96, nah! Maybe he'll just be happy to get one of my crazy no-one-ever-uses-this PRs again! :P
11:35:54couven92it's been a while... :O
11:38:15Araqdom96: already know it
11:39:09dom96Araq: Oh good, cause I got bored of it much quicker than I thought I would
11:44:21Araqit's the old "I invented Java and regret it's not a lisp" point
11:50:08Araqcouven92: use a nim.cfg in your $HOME ?
11:51:47couven92Araq, not machine-wide and I definitely do not want to redploy a magic nim.cfg for every new user!
11:53:30Araqhuh?
11:56:21Araqmratsim: tried the for-loop macros?
11:57:05*dddddd_ joined #nim
11:57:16FromGitter<arnetheduck> http://www.informit.com/articles/article.aspx?p=2425867
11:58:10*dddddd_ is now known as dddddd
12:07:20*Vladar quit (Remote host closed the connection)
12:07:41*Vladar joined #nim
12:15:36Araq"Looking back at my top 10 list, more than half of these annoyances are a direct result of including a feature primarily because it would be familiar to users of other languages."
12:16:07AraqCopying bad design is not good design.
12:16:32FromGitter<mratsim> @Araq not yet
12:17:49Araqcome on, an arraymancer that works with an officially released Nim version is boring.
12:23:53nivquestion: is there a way to have "do" block syntax on a proc work for both with and without return type?
12:24:29nivexample: proc whatever[T](cb: proc(): T). i can do "discard whatever() do -> int: ...", but i cannot do "whatever() do:" [assuming void]
12:24:50nivbasically, i want a generic proc that works for both void and return types
12:26:12FromGitter<tim-st> I already asked this two times but it's still not clear: if I have two procs that are exactly the same but on one the param is declared `static`eg. `mult(x,y: int):int` and `mult(x,y: static[int]): int` is it necessary to overload it if I call it sometimes with static values? or does the compiler "sees" it? My example in release mode generated binaries that differ in only one byte
12:29:44FromGitter<tim-st> (regarding performance)
12:29:45FromDiscord<treeform> Araq, familiarity could be important, main reason I chose nim is that I wanted python but better.
12:30:08Araqsorry you got the best PL out there instead.
12:30:46Araq;-)
12:31:12Araqnah, I'm not serious. but we're getting there, slowly but steadily.
12:31:19FromDiscord<treeform> 😃
12:32:10Araqtim-st: static[T] is supposed to match better than T
12:32:34FromGitter<tim-st> ok, so it's not a performance optimization?
12:32:36Araqif you only sometimes provide static values, don't bother and look at the produced asm.
12:32:48FromGitter<tim-st> ok, thanks for the answer!
12:33:18FromGitter<tim-st> I thought it's because of performance, for example bool flags could be completely removed in loops
12:33:34FromGitter<tim-st> when static
12:33:48FromGitter<tim-st> strutils often doesnt use static bool flags
12:33:55Araqyeah but optimizers have a specialization pass
12:34:09Araqalways look at the produced machine code
12:34:20FromGitter<tim-st> ok, thanks!
12:35:03*Torro left #nim ("-bye")
12:36:09Araqniv: write whatever() do -> void: ?
12:37:04FromGitter<krux02> dom96: that talk is quite interesting, even though his definition of a cathedral is wrong.
12:39:05FromGitter<krux02> he basically explains how important operator overloading and metaprogramming is.
12:39:11FromGitter<krux02> And I totally agree.
12:40:17FromGitter<krux02> if he would have had more saying, Java might have been a better language today.
12:44:09*Lord_Nightmare quit (Ping timeout: 264 seconds)
12:45:05nivAraq: sorry, can't figure it out: https://gist.github.com/niv/d6807322d4296ccc4cde3d9758771280
12:45:17*krux02 joined #nim
12:48:37*Lord_Nightmare joined #nim
12:56:01FromGitter<krux02> well the problem is that void is not a real type
12:56:38FromGitter<krux02> your interface wants you to return someting of type T, but void is a fake type that says, hey I am not a type
12:56:52FromGitter<krux02> so just return an int
13:01:50nivyeah, that's what im doing. but the syntax is more fugly than it needs to be, imo.
13:05:01FromGitter<allochi_twitter> Hi guys, I have a question about strange behaviour with `awaitAny()`, on my MacBook Pro i7, I never pass 4 results and then my code just freeze, I tried to read docs and `awaitAny()` source code to see if I’m missing something, but I reached a dead end. any insights? Thanks!
13:06:23FromGitter<krux02> @allochi_twitter what do you mean you never pass 4 results?
13:07:13FromGitter<krux02> I never used `awaitAny()` so I can't give you specific help, but I would try to dubug it with gdb and break into the application and see, where it hangs
13:07:39FromGitter<krux02> fyi: if you want to use gdb, you need the compiler flag --debugger:native
13:08:13FromGitter<allochi_twitter> If I spawn 4 or less procs, it’s fine I get the results back, and the code terminate, but if I add one extra and pass 5 tasks I 4 results and then the code freezes.
13:09:05FromGitter<allochi_twitter> thanks I will try this, I’m trying to understand threading in Nim, so far so good except at this point XD
13:10:02FromGitter<allochi_twitter> @krux02 could I paste my sample code here and you try it out? maybe its an OS or machine problem.
13:10:55FromGitter<krux02> I can try, but it should be single file
13:11:14FromGitter<krux02> use ix.io or github gist for code pastings
13:11:27FromGitter<allochi_twitter> Ok, will do
13:11:30FromGitter<allochi_twitter> thanks
13:12:18euantoror use play.nim-lang.org
13:13:19FromGitter<allochi_twitter> Ok, trying play.nim-lang.org first, thanks!
13:13:24FromGitter<krux02> well I prefer to run it locally
13:14:06Araqwell it's a thread pool.
13:14:23Araqif you block all of its thread it's not gonna create more
13:14:39FromGitter<allochi_twitter> Yep, the Araq is here!!! WOOOWWW!!!
13:15:10FromGitter<krux02> Praise the Araq
13:15:24Araqhey. thanks! :-)
13:15:27FromGitter<allochi_twitter> Haha!
13:15:32FromGitter<allochi_twitter> https://gist.github.com/allochi/b8450b6e0599c15ff0b455ccec1f09a7
13:15:41FromGitter<allochi_twitter> this is my code
13:16:03FromGitter<krux02> what does the `^` operator do?
13:16:29FromGitter<allochi_twitter> return the result of the FlowVar
13:17:04FromGitter<allochi_twitter> maybe I’m not using the right terminology here :)
13:17:21FromGitter<krux02> well it works
13:17:39FromGitter<krux02> no problem to run it at all
13:17:43FromGitter<allochi_twitter> AHA!, this is good news, not for me, but good news
13:18:07FromGitter<allochi_twitter> so you get an array with 5 elements and the code terminate properly
13:18:10FromGitter<allochi_twitter> ?
13:18:27FromGitter<krux02> yes
13:18:37FromGitter<krux02> http://ix.io/17Vj
13:19:03FromGitter<allochi_twitter> I may try to install Nim on a linux machine and try out the code then, since it’s not working on my MacBook and I have been trying this for like a month now.
13:19:22FromGitter<allochi_twitter> this is great man, thanks a lot!
13:19:45Yardanicodid you try choosenim?
13:20:25FromGitter<krux02> I am on the devel branch and I compile from the git repository.
13:20:30FromGitter<allochi_twitter> Now I need to figure out why it’s not working on the Mac, any thought?
13:20:58FromGitter<allochi_twitter> are you using a Mac?
13:21:29FromGitter<allochi_twitter> nope, you are not :D
13:21:37FromGitter<krux02> well when I was a child I played frogger on the old discarded Apple ][ of my father.
13:21:46PMunchIs there a simple example of Nim GC control?
13:21:47FromGitter<krux02> but not recently
13:22:01FromGitter<allochi_twitter> Yeah, I think I’m old enough to be the age of you dad too ;)
13:22:06Yardanicoallochi_twitter: as I said, did you try choosenim? :)
13:22:25FromGitter<krux02> my dad is over 70
13:22:28FromGitter<allochi_twitter> Nope, but I will try it out, sorry for not getting back to you
13:22:41FromGitter<allochi_twitter> Ok, no then, I’m 44
13:22:51FromGitter<krux02> I am 30
13:23:03FromGitter<allochi_twitter> you look younger on your avatar XD
13:23:16FromGitter<krux02> thanks
13:23:33FromGitter<allochi_twitter> Anyway, I will try choosenim, hopefully it’s something resolved in the dev branch
13:23:37FromGitter<krux02> but that pitcuture is also not the newest
13:23:41FromGitter<krux02> I think 4 yers old
13:23:59FromGitter<allochi_twitter> well, this is why I said so, since you look 20ish in it
13:24:10Yardanicoallochi_twitter: you can always try to install Nim using instructions from readme in github repo
13:24:26Yardanicobut choosenim helps you to automate all that stuff :)
13:24:28FromGitter<krux02> I am from 1987, the year that introduced a lot of great game series
13:25:26FromGitter<allochi_twitter> Ok, I don’t know how to reply directly to Yardanico in Gitter, but I’m looking now into Choosenim
13:25:50YardanicoYou just need to write my nickname, most IRC clients automatically highlight your nickname
13:26:00Yardanicoand you did highlight me in your previous message :)
13:26:12FromGitter<data-man> @PMunch: tests/gc ?
13:26:42FromGitter<allochi_twitter> @Yardanico, cool thanks :)
13:27:05FromGitter<narimiran> @allochi_twitter finally i'm not the oldest here :)
13:27:49FromGitter<allochi_twitter> @narimiran don’t worry I’m ok to be everybody’s grandpa XD
13:27:56FromGitter<allochi_twitter> I’m the oldest here
13:28:18FromGitter<krux02> 1987 the year that gave us: Turrican, Mega Man, Zelda (outside japan), Great Giana Sisters, Castlevania, Metal Gear, Dungeon Master, and me
13:28:59FromGitter<krux02> @narimiran how old are you?
13:29:15FromGitter<data-man> I’m 45 and 9 months :-D
13:29:27FromGitter<allochi_twitter> Man, I just want to go back home today ruins someones day with Dota
13:29:44FromGitter<krux02> interesting, I thought that new programming languages would be more for young people
13:29:49FromGitter<allochi_twitter> @data-man you win, you are our grandpa now
13:29:53FromGitter<narimiran> (soon to be) 32 :) not that much older than you, but last time there was age-discussion here, most of the guys were in 17-25 range
13:30:52FromGitter<allochi_twitter> I had my eye on Nim since early stage, but then got involved in Go for the passed 5-6 years, now I fall in love with Nim
13:30:56FromGitter<allochi_twitter> so no, Nim is for everybody
13:31:19FromGitter<krux02> I also had my share of Go.
13:31:25FromGitter<narimiran> > interesting, I thought that new programming languages would be more for young people ⏎ ⏎ we're still young!!! right? right?
13:31:32FromGitter<krux02> I like that language, too, but by far not everything.
13:31:49FromGitter<allochi_twitter> @narimiran you are young, youth is in the heart
13:31:55FromGitter<krux02> I would even say Go is the best introductory language for beginner programmers right now on the market.
13:32:16FromGitter<krux02> we are all as old as we feel we are :P
13:32:22FromGitter<allochi_twitter> I actually don’t like Go anymore, and I build production systems in it
13:32:59FromGitter<allochi_twitter> the way that Go is Designed is, we wait for the problem to fade way instead of addressing it
13:33:07FromGitter<krux02> @allochi_twitter I don't like Go anymore either, but that doesn't change my opinion about it as an introductory language.
13:34:20FromGitter<allochi_twitter> I think Nim can also do that, I find that Nim users need to be more active on documentation and tutorials, I would be doing some when I feel comfertable with the language, although I do have production scripts written in Nim now
13:34:30shashlickya I too feel old in this gang and i'm 39
13:35:03FromGitter<krux02> a majority of people stick to the language they learn first. And I am ok when that language is Go. It's not a terrible language. At least it is much better than Java, and still has this minimal core to it.
13:35:13subsetparkolder is better, folks
13:35:16FromGitter<krux02> is there anybody under 30 here?
13:35:23FromGitter<allochi_twitter> Hey think about it this way, if you consider your age run by a thread pool, with 4 threads, then you are about 10 years old
13:35:56FromGitter<allochi_twitter> I’m 17 in soul
13:36:10FromGitter<krux02> I am 3 years old on my ass
13:36:18FromGitter<allochi_twitter> XD
13:36:59*PMunch quit (Quit: Leaving)
13:37:25FromGitter<alehander42> I am 16 actually
13:37:36FromGitter<krux02> interesting
13:37:40FromGitter<krux02> that is young
13:38:15FromGitter<krux02> When I was 16 I learned to program Blitz Basic.
13:38:17FromGitter<allochi_twitter> you should not be here among the old one XD
13:38:18FromGitter<alehander42> never trust random people on the internet @krux02
13:38:22FromGitter<allochi_twitter> Kidding ;)
13:38:22FromGitter<alehander42> I am 25
13:38:26FromGitter<krux02> And I was productive in it, I made playable games.
13:38:53FromGitter<allochi_twitter> productive in what? Go?
13:38:56FromGitter<krux02> :/
13:39:06FromGitter<krux02> productive in Blitz Basic
13:39:19FromGitter<allochi_twitter> oh
13:39:56FromGitter<alehander42> when I was 16 I didn't even know I'll become a programmer, and I think I didn't have internet
13:40:23FromGitter<allochi_twitter> for be it was 12, I used to write code on paper
13:40:28FromGitter<allochi_twitter> I didn’t have a computer
13:40:43FromGitter<allochi_twitter> until 17
13:42:33shashlickI was neck deep in the demoscene when I was 16
13:42:42shashlickbut didn't code back then
13:42:58FromGitter<krux02> https://www.blitzforum.de/forum/viewtopic.php?t=9865&highlight=
13:43:12shashlickwell maybe some basic and x86 assembly in school
13:43:24FromGitter<krux02> my game, source only
13:43:29FromGitter<krux02> no resources required
13:43:37FromGitter<krux02> everything wtih the draw line tool
13:44:03FromGitter<krux02> I love it, because it is minimal, and if you have blitz basic somewhere you could just copy paste the code to run it.
13:44:42FromGitter<krux02> the bad thing, blitz basic is windows only
13:44:48FromGitter<krux02> and I am now on Linux only
13:45:13FromGitter<krux02> hel people my recommendation, swich to linux as early as possible, so that this is not that much of a problem
13:46:10FromGitter<allochi_twitter> Hmm, a Linux advocate :)
13:47:09FromGitter<narimiran> i switched only recently, and i'm now disappointed that i didn't do that earlier....
13:48:37FromGitter<krux02> In the past I thought windows was bad, now I think it is even toxic to have it.
13:49:25FromGitter<krux02> I honestly think that companies that deal with customer data, should not be allowed to use this data on Windows systems, because it cannot be guranteed that this data is not secretly sent to Microsoft.
13:50:17FromGitter<krux02> Well year windows has these hidden options to disable sending data, but yea I know how these things work. There is an update and secretly those switches are reset and you think you are safe but you are not.
13:50:30FromGitter<allochi_twitter> Windows we agree on, but come on OSX with it’s proprietary software system.
13:50:46FromGitter<krux02> And then there is this problem that it has integrated advertisement in the start menu. When I first saw that, I was honestly "WTF"
13:51:12FromGitter<allochi_twitter> I though Ubuntu wanted to do the same thing too
13:51:29FromGitter<allochi_twitter> didn’t they?
13:51:34FromGitter<krux02> I can't judge that much on OSX, I never used it longer than a few minutes and instantly just wanted go back to linux.
13:51:57FromGitter<krux02> well yea. But before that I left ubuntu already for other reasons.
13:52:37FromGitter<allochi_twitter> let’s be reasonable, people who write software need to pay the rent
13:53:01FromGitter<allochi_twitter> and we have the right to choose what is acceptable to us for them to pay the rent
13:53:09FromGitter<allochi_twitter> if this makes sense :)
13:54:15*couven92 quit (Quit: Client disconnecting)
13:55:01FromGitter<data-man> Linux for programmers, Windows for users :)
13:55:16FromGitter<allochi_twitter> Maybe NimOS!
13:55:47FromGitter<allochi_twitter> Come on Araq, NimOS this month :)
13:56:15FromGitter<krux02> I started with Suse. Horrible gpu support. constantly destroyed my x server config and I hand to fix it from a tty terminal. then kubuntu, because it promised an easier way to install proprietary gpu drivers. at some point ubuntu fucked up that system and replaced it with an inferior system. and kde3 was replaced by kde4 which did not work on my computer, it was just too slow. Then I switch to ubuntu with gnome. And then
13:56:15FromGitter... gnome3 came and ubuntu also made their own DE. All horrible, I thought Xubuntu was my last resort. But the lack of updates on LTS versions of ubuntu and the constant destruction of the system on system upgrades made me just want to leave the ubuntu ship. I tested out arch. Much frustration in the installation pro ... [https://gitter.im/nim-lang/Nim?at=5ad4ab7f1130fe3d36e1d4a1]
13:56:24FromGitter<krux02> this forced on DE switching, I hate it.
13:56:59FromGitter<krux02> Linux for programmers, Windows for victims
13:57:13FromGitter<allochi_twitter> I use Debian on the server side, OSX as a desktop
13:57:47FromGitter<krux02> well when you have to use Linux anyway, you can simlify your life when you also use Linux as a desktop
13:58:08FromGitter<krux02> well year you can't use Adobe products. But they are a ripoff anyway.
13:58:18FromGitter<allochi_twitter> not really, I tried, software is not stable
13:58:26FromGitter<allochi_twitter> not the OS, the software
13:58:38FromGitter<krux02> what software does hook you to OSX?
13:58:45FromGitter<allochi_twitter> Chrome
13:58:48FromGitter<allochi_twitter> XD
13:58:52FromGitter<krux02> really?
13:58:55FromGitter<data-man> Win-users don't understand that they are victims
13:59:07FromGitter<krux02> I am on chromium here on this system.
13:59:12FromGitter<allochi_twitter> Kidding, I have a lot, beside I need to compile also for iOS
13:59:17FromGitter<krux02> chromium is by the way the superior version of chrome.
13:59:21FromGitter<allochi_twitter> and graphical packages
13:59:50FromGitter<krux02> it is chrome minus the harmful parts.
13:59:59FromGitter<krux02> what graphical packages?
14:00:04FromGitter<allochi_twitter> So let’s talk Nim, what are you doing with Nim? Any planned games?
14:00:41FromGitter<allochi_twitter> Oh, I have software that I need to use on OSX, some of them for work, and they are not available on Linux.
14:01:17FromGitter<krux02> well I am currently writing my master theses (yes I am an almost eteral student) about embedding shader programming in the outer environment
14:01:35FromGitter<allochi_twitter> And I’m not going to relearn how to do things on other software, so I just enjoy my ride, hopfully the younger generation will be full Linux powered ;)
14:01:39FromGitter<krux02> so I am until september locked to not write a game, but improve just the tooling about it.
14:02:22FromGitter<allochi_twitter> so what games you wrote, because I’m starting something in that direction.
14:02:34FromGitter<allochi_twitter> this was a question, what games you wrote?
14:02:41FromGitter<allochi_twitter> I would love to see a sample
14:02:49FromGitter<krux02> @allochi_twitter well the young generation will learn from the old generation and ask for advice. and when the old generation all can't give advice on Linux, or doesn't improve the way to actually use Linux, it is also fault of the old generation that the young generation still uses Windows.
14:03:11FromGitter<krux02> well I wrote tetris
14:03:14FromGitter<allochi_twitter> Haha, I agree
14:03:15FromGitter<krux02> no commercial games
14:03:22FromGitter<allochi_twitter> in Go? Nim?
14:03:30FromGitter<allochi_twitter> Basic?
14:03:33FromGitter<krux02> In Nim
14:03:49FromGitter<allochi_twitter> Cool, could you share it with me? Like is it on a GitHub?
14:04:04FromGitter<krux02> https://github.com/krux02/opengl-sandbox/blob/master/examples/tetris.nim
14:04:45FromGitter<allochi_twitter> This is great! thanks a lot, I will have a look at it as part of learning Nim
14:04:48FromGitter<allochi_twitter> Thanks!
14:05:35FromGitter<krux02> just look at the part "Render Code"
14:05:44FromGitter<krux02> all the rendering logic is tied together
14:06:07FromGitter<krux02> first I gather all data that is important for rendering, for examples which objects need to be rendered at what position
14:06:38FromGitter<krux02> and then I just render it all.
14:06:46FromGitter<krux02> the problem is that it doesn't run on OSX
14:08:43FromDiscord<treeform> Made some progress on font rendering in pure nim, wanted to share:
14:08:45FromDiscord<treeform>
14:08:45FromDiscord<treeform> https://cdn.discordapp.com/attachments/371759389889003532/435441420434669593/unknown.png
14:10:13FromGitter<krux02> @allochi_twitter https://www.g-truc.net/doc/OpenGL%204%20Hardware%20Matrix.pdf this is the reason it doesn't work on OSX
14:10:38FromGitter<allochi_twitter> Oh, OpenGL stuff
14:11:04FromGitter<krux02> just look at that document and see yourself how well OpenGL works on OSX
14:11:53FromGitter<allochi_twitter> Noooooo!
14:12:37FromGitter<allochi_twitter> So use OpenGL 4?
14:13:26FromGitter<krux02> well I use a few OpenGL 4 features, not very complicated ones. Like direct state access that makes the development much smoother
14:13:50FromGitter<krux02> but that breaks MacOS compatibility already
14:14:11FromGitter<krux02> and just look at when opengl 4.1 (latest supported opengl version on mac) was released
14:14:28FromGitter<allochi_twitter> Ok, I see
14:14:38Araqif only there were an OS with an alternative to OpenGL, widely used for gaming
14:15:02Araqwhere can deploy executables easily and they don't break after an OS upgrade
14:15:07FromGitter<allochi_twitter> Like NimOS?
14:15:20Araqenabling something like a video game industry
14:15:40FromGitter<allochi_twitter> Come on, we already established the fact that windows is evil
14:15:54FromGitter<krux02> Araq: well I hope WebAssembly will help to improve backwards compatibility by a lot.
14:16:30FromGitter<krux02> when the majority of applications is in WebAssembly, then people can easily switch the OS without loosing any of their software
14:16:34Araqwhere the OS is so good it does *not* come out of the box with outdated versions of 5 half assed scripting/dynamic languages
14:16:37FromGitter<allochi_twitter> whtat about the new thing, Volcan?
14:16:42FromGitter<krux02> they can even change the CPU
14:16:47FromGitter<krux02> Vulkan
14:16:54Araqwith good support for GPU programming.
14:16:59FromGitter<allochi_twitter> Vulkan, yah
14:17:21Araqand even an Unix environment for those who need it.
14:17:25FromGitter<krux02> Vulkan is just the continuation of OpenGl with broken backwards compatiblility
14:18:15FromGitter<krux02> well at the moment for webassembly you need a 10 GB development toolchain, and it doesn't even support threads. It is still young.
14:18:21FromGitter<krux02> but I think it is a great thing
14:18:21FromGitter<allochi_twitter> Come on man, we can’t win with you XD
14:18:43planetis[m]hey sup, how do I fix the ```cannot export: "dir" / module``` Error?
14:19:35FromGitter<krux02> the idea behind the opengl versions is that for each new major version you need now hardware
14:19:51planetis[m]I have an example if you want to take a look
14:19:52FromGitter<allochi_twitter> Ok, I have a meeting, but it was nice chatting with you guys, I will check in again when done.
14:20:06planetis[m]https://github.com/notTito/morpheus
14:20:17FromGitter<krux02> opengl 2 gives you shaders, so you need shader hardware, opengl3 gives you geometry shaders, so you need new hardware for that, and opengl4 gives you compute shaders I think.
14:20:44FromGitter<krux02> but you won't any hardware change for opengl4.1 to opengl4.6 it is purely a driver thing.
14:21:15FromGitter<krux02> And I really hope that we won't need any backwards compatibility breaking hardware in the near future.
14:21:30FromGitter<tim-st> is there an other keyword in nim beside `defer` that directly assumes a `:` ?
14:22:03FromGitter<krux02> else
14:22:51FromGitter<tim-st> ok, but I think it's a bit strange having a `:` behind `defer` but dont have one behind `assert`. I think it would be better to have defer without `:`
14:25:10FromGitter<tim-st> also discard `name` works without `:`
14:25:32FromGitter<tim-st> and import and so on
14:26:33Araqafter a colon a list of statements follow. (or a type description)
14:27:06Araq'assert' does not take a list of statements. the end.
14:27:42FromGitter<tim-st> ok
14:28:57Araqplanetis[m]: export modulename maybe.
14:29:57FromGitter<krux02> Araq: maybe it would be a good idea to allow functions with a single argument have a stmtList as argument
14:30:15planetis[m]hey Araq, thanks I will try that
14:30:45FromGitter<krux02> sorry, I just realized it works that way
14:30:48FromGitter<krux02> did not expect that
14:30:50planetis[m]great
14:31:36FromGitter<krux02> this just compiles, I like it: http://ix.io/17VK
14:32:30FromGitter<krux02> is it in the documentation that the last argument of a call can always be written as a code block that evaluatios to the last argument?
14:32:47FromGitter<krux02> (last statement/expression)
14:35:21Araqprobably not :-)
14:35:31FromGitter<tim-st> so assert allows both ways with `:` and without; but defer dont?
14:36:22planetis[m]can a module have the same name as a procedure?
14:36:44FromGitter<krux02> Araq: that is a very powerful feature of the language and it might better be documented. I instantly like it.
14:37:22FromGitter<tim-st> ok, I see it only allows one stmt
14:37:29FromGitter<krux02> lisp people will love it, because it is very common in lisp, too.
14:37:55Araqit's just the general "everything is an expression" thing
14:38:12Araqwhich we got rather late
14:38:39Araqand doesn't really work well with the indentation based syntax, but I love indentation based syntax too much to do anything about it
14:38:40FromGitter<krux02> @tim-st: dever is processed by the parser as a keyworld, and sadly it is not just a `template` in system that is implemented with a magic.
14:39:16FromGitter<tim-st> ok, thanks
14:39:30FromGitter<krux02> Araq: Will I don't like indentation based syntax, but I can live with it when the rest of the language is so great as in Nim.
14:39:54FromGitter<krux02> s/Will/Well
14:41:04subsetparkI was definitely thrown by defer: as well - given how block syntax works for so many other macros and keywords, I was typing defer:\n\tdoThing() and very confused when it wasn't working :/
14:41:15subsetparkBecause it's not a syntax error
14:42:42FromGitter<krux02> Araq: I think I mentioned it before, but it comes up before, `defer` as a keyword should disappear from the language, it should be just a magic proc/template in system.nim.
14:43:25AraqI tried, did not work well. I would rather remove 'defer' again.
14:43:45FromGitter<krux02> but what instead of defer?
14:44:46Araqtry finally, destructors
14:45:05FromGitter<krux02> I don't like try finally at all.
14:45:11FromGitter<krux02> I think is is very ugly
14:46:23AraqI think 'defer' is much uglier.
14:46:47Araqthe important thing is what is 'tried' and 'defer' doesn't even mention that
14:46:58FromGitter<krux02> just for example when you have three initalizations in a row that all might fail. Writing a try finally for all of them that covers all possible cases is just ugly.
14:47:06subsetparkI have to agree with Araq. If it obeyed block syntax it wouldn't be ugly, but `defer: foo()` is quite odd.
14:48:02FromGitter<krux02> what is so odd about defer: foo()?
14:48:35Araqfor a start, it doesn't work as a toplevel statement.
14:48:59FromGitter<krux02> that is true.
14:49:35*yglukhov quit (Ping timeout: 256 seconds)
14:49:38FromGitter<krux02> but apart from that?
14:50:26Araqwell that's it, it's not clear where the 'try' should be injected. that's enough to dislike it.
14:51:09Araqin procs it only sort-of-works because other constructs like 'if' and 'while' do not hide where they begin
14:51:14FromGitter<krux02> well I would compile: "defere: foo(), bar()" at top level to ``proc foobar(): void = foo(); bar();\n atexit(foobar)``
14:54:45subsetparkkrux02: what's so odd is that it's the only keyword i can think of that means one thing as `defer: foo()` but means something else, which is never what you meant, as `defer:\n\tfoo()` - but `defer foo()` is a syntax error
14:56:42FromGitter<krux02> defer:\n\tfoo() and defer: foo() do the same thing
14:56:47FromGitter<krux02> what is so weird about it?
15:04:57FromGitter<allochi_twitter> Hi, I’m back, and I also have bad news
15:06:24Araqyou're injured and couldn't join the marathon?
15:06:36FromGitter<allochi_twitter> @Araq it would be lovely if you can help, I re-installed using choosenim, and tried my previous code with both stable and devel, and they both don’t work. could you please give me a hint as what to do next?
15:06:49*miran joined #nim
15:07:28FromGitter<allochi_twitter> Nah, I don’t do Marathons, but I would love if you could help with the thread pool problem I’m having, I just don’t know what to do next.
15:07:49*nsf quit (Quit: WeeChat 2.0.1)
15:10:56FromGitter<data-man> @Araq: #5898 ready for review
15:11:31FromGitter<data-man> And #7623 :)
15:12:48Araqallochi_twitter: looking at it
15:14:02FromGitter<allochi_twitter> Thanks, I know you have a lot on your plate if you give the a direction I can try myself, also I’m trying this now on a linux machine, will check how it works.
15:14:35FromGitter<krux02> for everybody who prefers try over defer, here is a counter example: http://ix.io/17VT
15:14:52FromGitter<krux02> it's just an ugly mess with try catch
15:14:55shashlickallochi_twitter: is this the gist you shared two hours ago on await?
15:16:09FromGitter<allochi_twitter> yes
15:18:43shashlickdang - need to add threadpool => --threads:on detection in snip
15:19:31FromGitter<allochi_twitter> I have it in a .nims file
15:20:02shashlickya, i have a snippet tester I made to run such things quickly
15:21:24shashlickhttps://github.com/genotrance/snip
15:23:31*max3 joined #nim
15:25:20FromGitter<allochi_twitter> I just finished installing nim on my linux server, and tested the code, it works fine there
15:26:23FromGitter<allochi_twitter> it’s an OSX problem, not sure if it’s my machine, I will test this on another OSX machine tomorrow.
15:26:47FromGitter<allochi_twitter> the good news is that I wrote a working code XD, it’s not me it’s OSX
15:27:01FromGitter<krux02> well OSX tries more and more to abandon things that also would work on Linux
15:27:30FromGitter<allochi_twitter> You had to comment on that, didn’t you XD
15:27:45FromGitter<krux02> yea I had to
15:27:46FromGitter<allochi_twitter> I’m not switching man, I’m 44
15:28:04Yardanicoso what is the problem you're having on OSX?
15:28:22YardanicoYou can share install logs with us, we can help
15:28:40Yardanicobut try installing with instructions from repository README if choosenim doesn't work for you :)
15:28:48FromGitter<allochi_twitter> install of nim? I just re-installed it with choosenim
15:29:02Yardanicoso it works?
15:29:02FromGitter<allochi_twitter> it works
15:29:05Yardanicoah, ok
15:29:10FromGitter<allochi_twitter> choosenim yes
15:29:23FromGitter<allochi_twitter> and I installed both stable and devel
15:29:30euantorI take it the mac only has 4 logical cores @allochi_twitter?
15:29:39euantorHoe many logical cores does the Linux server have?
15:30:06FromGitter<krux02> I have 4 logical cores, and the coded did work
15:30:10FromGitter<allochi_twitter> I don’t know let me check the activity monitor
15:30:24FromGitter<allochi_twitter> this is true also :)
15:30:52FromGitter<krux02> lscpu
15:31:02FromGitter<krux02> that is the command to list the cpu s
15:31:13FromGitter<allochi_twitter> but, what even if 4 logical cores? this is not about cores
15:31:54FromGitter<allochi_twitter> I mean when I spawn, I’m creating a process right? I can have one core, and still run a thread pool
15:32:00FromGitter<allochi_twitter> am I correct?
15:32:13FromGitter<krux02> well it was a thread pool and when the thread pool is empty it doesn't have anyth more threads
15:32:19FromGitter<allochi_twitter> of course I’m correct I’m 44
15:32:22euantorIt may be something interesting to do with the Mac's scheduler if 4 tasks works, but 5 tasks doesn't
15:32:35FromGitter<allochi_twitter> my point
15:32:39FromGitter<krux02> and maybe the threadpool initialized with the amount of logical cores on the system
15:32:47euantorthe thread pool module creates threads == number of logical CPUs if I remember correctly
15:32:51FromGitter<allochi_twitter> I wish if I have the skills and time to find it myself and contribute
15:33:26euantorhttps://nim-lang.org/docs/threadpool.html#setMinPoolSize,range[]
15:33:26FromGitter<allochi_twitter> Intersting, give me a sec, I will test something
15:33:46euantorSo by default the minimum thread pool size is 4, so it starts with 4 threads when using the pool
15:34:25euantorhttps://github.com/nim-lang/Nim/blob/master/lib/pure/concurrency/threadpool.nim#L399
15:34:41euantorI haven't played with the thread pool module at all, so I'm just guessing at how it works
15:35:43dom96yglukhov created a PR to change this IIRC
15:35:49dom96might already be in devel?
15:35:56enthus1astis it just me or does "net.sendTo" does not support sending to ipv6?
15:36:27FromGitter<allochi_twitter> You are correct
15:36:42FromGitter<allochi_twitter> setMinPoolSize(12) solved the problem
15:36:49*xkapastel joined #nim
15:37:29FromGitter<allochi_twitter> @euantor it is about the threadpool size
15:37:48euantorGuessed it might be :)
15:38:21FromGitter<allochi_twitter> I don’t know if @Araq is following, I hope he doesn’t waist time on this anymore, but I think it worth documenting
15:39:13FromGitter<allochi_twitter> thanks a lot @euantor
15:39:17euantorNot sure why it exhibits the behaviour you are seeing though. I'd have thought that if the pool was exhausted a new thread would be created (up to max pool size, and if max size is reached new tasks would be queued) and there wouldn't be a problem
15:39:46FromGitter<allochi_twitter> It just hang
15:40:00FromGitter<allochi_twitter> doesn’t even throw an exception or anything
15:41:20FromGitter<allochi_twitter> you know it’s still a problem, since in my case this is just a sample example for me to learn nim, but if this is a production code, and I don’t know how many task I’m going to be dealing with ahead of time, it could be a problem.
15:42:13dom96So I'm playing around with the the Nim playground frontend code
15:42:18FromGitter<exts> Can macros make my `get` and `getSeq` into a single function? I don't remember much about macros the last time I played with nim and I don't know where to even get started. https://github.com/exts/monoworkflow/blob/master/monobuild.nim
15:42:35dom96Either Karax doesn't offer any Reactive features or the playground doesn't use them, which is a huge shame
15:43:16enthus1astit is reactive
15:43:21FromGitter<krux02> @exts well your question is wrong, but probably is the answer to your question yes
15:45:43FromGitter<krux02> macros transform code at compile time. So if you can describe what you want as a code transformation, then it can be implemented as a macro.
15:46:41FromGitter<exts> whats the question I should be asking
15:49:33*Trustable joined #nim
15:50:08*yglukhov joined #nim
15:51:45*couven92 joined #nim
15:52:33couven92!eval echo 42
15:53:34NimBotCompile failed: <no output>
15:58:38Araqallochi_twitter: the flow var completes too early and doesn't trigger the awaitAny which is issued too late
15:58:51Araqlooks hard to fix.
15:59:06Yardanicocouven92, umm
15:59:09Araqincreasing the thread pool size is merely a workaround. concurrency is hard :-)
15:59:11Yardanicopoor nim playground :(
16:00:58FromGitter<allochi_twitter> @Araq it seems to be only a problem with OSX, and even if I increase the delay it’s the same thing, I did think of this.
16:02:05dom96couven92: playground is down
16:02:08FromGitter<allochi_twitter> Thanks for giving it the time, I will try to debug from myside, I may have to nag you about it in the future, I want to replace some Go stuff with Nim and I don’t want something like this to bite me.
16:02:31*rockcavera joined #nim
16:02:55couven92sorry...
16:03:02Araqunderstandable.
16:03:54FromGitter<allochi_twitter> @Araq I can’t thank you enought for all the good work you put in Nim and the other too.
16:08:38FromGitter<data-man> @allochi_twitter : Try to compile with -d:nimPinToCpu ⏎ But I'm not sure, it's a hidden magical definition. :)
16:12:37FromGitter<exts> @krux02 can you clarify what you meant by "your question is wrong"
16:14:11FromGitter<exts> because I googled and return type overloading isn't a thing in nim, only thing I seen on this topic was this https://forum.nim-lang.org/t/3025 so I didn't ask about it. So what is the question I was supposed to be asking again?
16:22:12*SenasOzys quit (Ping timeout: 260 seconds)
16:26:43*SenasOzys joined #nim
16:31:37*nsf joined #nim
16:36:18*sendell quit (Remote host closed the connection)
16:39:18*flyx quit (Quit: ZNC - http://znc.in)
16:46:23FromGitter<exts> hmm opted for generics (kind of forgot about generics) and I don't think i'm writing nim generics correctly, but it compiles so win/win I guess https://gist.github.com/exts/eb81ea1d04a964742a85813944e1b602
16:49:15YardanicoAraq, what is the preferred count of newlines between end of the type section and start of any other section (proc, const, etc)?
16:50:26Yardanicofor stdlib
16:53:06FromGitter<krux02> @exts you should probably stick to all settings being seq[string], just some are seq of length 1
16:53:37FromGitter<krux02> or make everything a string
16:54:06FromGitter<krux02> you approach requires the programmer to know the type of a setting before looking it up. And that is not good.
16:54:27FromGitter<exts> i was going to do that, but it doesn't accomplish what im trying to do. would be pointless to have a separate container just for an array of options
16:54:35FromGitter<exts> so this does what I want, and does it well
16:55:49FromGitter<krux02> well you can also have two settings tables
16:56:33FromGitter<krux02> ``stringSettings: Table[string,string]`` and ``seqSettings: Table[string, seq[string]]``
16:56:45FromGitter<exts> this is true, that's why programming is awesome. you can do the same thing a million different ways
16:56:52FromGitter<krux02> in that way you don't need to cast or throw exceptions
16:56:54FromGitter<alehander42> I think the simplest would be to just have SettingsTable = Table[string, seq[string]] and save the string settings internally as seq[string] but implement the logic only in `get[seq[string]]` and make get[string] call getseq[string (settings, key)[0]
16:57:44FromGitter<krux02> @alehander42 that doesn't work exactly that way
16:57:52FromGitter<alehander42> this way you have a simple internal impl, you write the logic once and have the same api
16:57:53FromGitter<alehander42> why?
16:58:20FromGitter<krux02> well you need to know if you want to look up a string setting or a seq setting
16:58:50FromGitter<alehander42> well you do anyway
16:59:02FromGitter<alehander42> you need to know if you need stringSettings or seqSettings
16:59:11FromGitter<alehander42> that's the nature of a setting, you need to know what it is
16:59:16FromGitter<alehander42> usually you even have bools, ints etc
16:59:36FromGitter<krux02> well, long story short, if you follow that route you will realize that just putting the ``[0]`` outside of the function when looking up a string setting is the shortest and best implementation
17:00:11FromGitter<alehander42> basically yeah, but having that as getString or get[string] might be more clear
17:00:15FromGitter<krux02> and then you realize you can remove the entire getSetting fuction and just operate on the table directly
17:00:27FromGitter<krux02> @alehander42 nope
17:00:32FromGitter<alehander42> nope
17:00:40FromGitter<alehander42> if I see `[0]` i don't know if this is a string setting
17:00:47FromGitter<alehander42> or I just needed only the first one
17:00:47YardanicoAraq, also, should I remove modules which were deprecated in 0.16.0 or earlier?
17:00:50*athenot joined #nim
17:00:50FromGitter<krux02> using the table directly is more clear, because nothing is more clear than inlined code
17:00:58Yardanicolike asyncio/ftpclient
17:01:37FromGitter<alehander42> well you can just write the table get function inline if nothing is more clear than inline code
17:01:42FromGitter<krux02> @alehander42 well it would be best if nim had pattern matching then it would look like this:
17:01:42FromGitter<alehander42> you need a boundary
17:01:51miranYardanico: if you ask me (and you don't), <=0.16 (inclusive) should be removed
17:02:01FromGitter<krux02> let @[`mystrngsetting`] = getsetting("...")
17:02:38Yardanicomiran, Araq already said that everything which was deprecated in 0.16.0 or earlier should be removed, but I wanted to clarify about modules :)
17:02:54FromGitter<alehander42> ok, that would be good too, but I think the "just expose a good API" thing is simpler
17:03:16FromGitter<krux02> @alehander42 what boundary, he throws an exception when the setting type doesn't match, indexing into an empty array, well throws an exception
17:03:19FromGitter<alehander42> and you would need get[stuff] anyway if you have different types
17:03:31FromGitter<krux02> sor "raise" is the keyword in nim
17:04:34FromGitter<alehander42> no, the problem is, as I said without pattern matching "let a = settings.get(.)[0]" tells me nothing: I have no idea if he just needs the first setting or it is *accidentally* a way to get the only string one
17:04:46FromGitter<alehander42> it's just bad api design
17:05:48FromGitter<alehander42> having "let a = settings.getString(name) or get(string, name)" or something is just more obvious for the reader and you can do it without perf penalty (at least the first)
17:06:52FromGitter<alehander42> and later if you change all settings and make them have some weird special containers for each type, you don't need to change the api
17:08:47FromGitter<krux02> well I wouldn't
17:15:12*r3d9u11 joined #nim
17:16:15dom96I love that I can google "jester static dir" and get the exact syntax I need to set it :)
17:19:04Yardanicodom96, how much newlines there should be between the type section and const/var/proc or any other sections (in stdlib)?
17:19:09FromGitter<alehander42> https://stars.medv.io/nim-lang/nim
17:19:15FromGitter<alehander42> i wonder what's that bump from in th emiddle :D
17:19:20dom96Yardanico: At least one
17:19:57dom96alehander42: that's cool. I bet it's slashdot
17:20:25FromGitter<Varriount> Yardanico: I prefer 2
17:20:49FromGitter<alehander42> I started using 2 lines between all top level procs and sections
17:20:53*themagician joined #nim
17:21:37YardanicoI'm just removing a lot of T-prefixed types deprecations, so previously there was <type section><newline>{.deprecated....}<newline><other sections>
17:21:54Yardanicoand I'm removing deprecated section and doing <type section><newline><other sections>
17:23:18dom96Well, add 2
17:23:30Yardanicodom96, ok
17:23:33dom96but 1 is enough so don't kill people if they do not use 2 :)
17:25:10Yardanicodom96, also, should I remove modules which were deprecated in 0.9-0.11?
17:25:19dom96I'd say so
17:26:38Yardanicowe can then remove `deprecated` directory altogether :D
17:26:59Yardanico(but I think it should be kept)
17:27:17Yardanicowell, IDK, future.nim was deprecated but it wasn't moved into this directory..
17:29:23Yardanicodom96, and we'll finally remove those implicit (without client) get/post procs from httpclient!
17:31:04FromGitter<data-man> @Varriount: You've found xml-library?
17:32:06FromGitter<data-man> https://github.com/aleksandar-todorovic/awesome-c#xml
17:32:32*jxy quit (Quit: leaving)
17:35:09*jxy joined #nim
17:40:20FromGitter<krux02> XML came from html which came from sml which was originally designed to structure text
17:40:41FromGitter<krux02> and it text it might be harder to see matching braces, so the end tag was a thing
17:41:38FromGitter<krux02> but when you use xml for data tables, you end up with 90 overhead for tags and stuff. it's not good for data.
17:47:09FromGitter<data-man> sml or SGML?
17:52:16FromGitter<krux02> yes might have been sgml.
18:02:36*couven92 quit (Quit: Client disconnecting)
18:09:33YardanicoAraq, should I remove deprecated stuff from system.nim (which was of course deprecated before or in 0.16.0)
18:09:41Yardanicolike exceptions with E prefix
18:14:22*NoOp quit (Quit: Connection closed for inactivity)
18:32:02FromGitter<amscotti> Hi, I'm looking for a way to get an index of an item in an array. In Python I would just used `index` on a list. What would be the best way to do this in Nim?
18:32:36FromGitter<zetashift> @amscotti you coculd use find: https://nim-lang.org/docs/system.html#find,T,S
18:32:49FromGitter<zetashift> could*
18:33:09FromGitter<amscotti> Ah, nice! Thanks!
18:33:45FromGitter<krux02> @amscotti make sure to handle the -1 case, when you don't find the item
18:34:48FromGitter<amscotti> Right, will do! Thanks 👍
18:42:25*couven92 joined #nim
18:50:19FromDiscord<geckojsc> it feels really awkward that 'nim-lang/sdl2' and 'Vladar4/sdl2_nim' both keep their modules in a directory called 'sdl2', but the inner files coincidentially have different names so they can still coexist on a single machine
18:50:39*nsf quit (Quit: WeeChat 2.0.1)
18:50:52FromDiscord<geckojsc> unless there's something more to it that I'm misunderstanding?
18:55:01couven92why is importc not allowed in nimscript?
18:55:12*LyndsySimon joined #nim
18:56:05couven92you could get the VM to be able to that, right? Call GetProcAddress and push the args to the stack before invoking...?
19:00:07dom96geckojsc: yes, that's why sdl2_nim gives Nimble warnings when it's installed
19:01:42jxyyou could use dlopen that way, or just use libffi, which is pretty much how other interpreters interface with shared libraries anyway
19:02:21couven92jxy, yeah dlopen on Linux, GetProcAddress is the Windows equivalent
19:03:13couven92basically allow dynlib for importc
19:03:24couven92err nimscript I mean, sry
19:11:13FromGitter<exts> I hope in the future the docs will stack overloaded procs on the sidebar to make things cleaner and group those overloaded procs that are similar together for ease of finding
19:11:43FromGitter<exts> still my biggest complaint about nim since '15
19:18:10*jxy quit (Quit: leaving)
19:18:49*jxy joined #nim
19:21:30FromGitter<mratsim> having to scroll up because the sidebar does not float is also annoying
19:22:13FromGitter<exts> I created an issue https://github.com/nim-lang/Nim/issues/7628 not sure if it's well formatted, but I put an example to make my suggestion clear
19:23:06FromGitter<exts> @mratsim I wouldn't even mind a "back to the top" button floating if that's too much
19:24:33FromGitter<zacharycarter> I'm working on the re-write of the playground - I have the whole docker in docker thing working from what I can tell
19:24:38FromGitter<exts> i think the documentation is the weakest part of nim, that's probably the biggest compliment I can give a language :P - makes sense though as nim is still an infant in the language world and documentation is hard.
19:24:40FromGitter<zacharycarter> I need to try deploying it to a VPS and see if it works
19:25:09FromGitter<mratsim> Beyond hard, people that know enough to document are probably busy building something in Nim :P
19:25:31FromGitter<exts> that and working on other stuff outside of nim, documentation is a bit tedious :P
19:25:36FromGitter<exts> but needed unfortunately
19:26:09FromGitter<exts> i think crystal has my favorite documentation/api generation atm
19:26:16FromGitter<exts> elixir's up there to
19:26:37*r3d9u11 quit (Read error: Connection reset by peer)
19:26:47*r3d9u11_ joined #nim
19:27:24*r3d9u11_ quit (Client Quit)
19:28:22dom96exts: start adding issues suggesting how the docs could be improved :)
19:28:36FromGitter<exts> i created one above
19:28:49dom96oh great
19:28:58dom96Now we just need somebody that implements it :)
19:29:37FromGitter<exts> does anyone know what api generation template system this uses? <https://crystal-lang.org/api/0.24.2/> I see it everywhere nowadays
19:29:43FromGitter<exts> same template, different color
19:30:09FromGitter<exts> i like it because the sidebar doesn't move while the content still updates, makes looking at the api more enjoyable
19:35:55*yglukhov_ joined #nim
19:36:55FromGitter<exts> @mratsim that could probably solve your side menu doesn't scroll issue
19:37:20*yglukhov quit (Ping timeout: 255 seconds)
19:40:10FromGitter<zetashift> @exts doesn't look like anything I know, isn't it just markdown with some fancy CSS?
19:40:10*rokups quit (Quit: Connection closed for inactivity)
19:48:59*flyx joined #nim
20:05:42*athenot quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
20:09:13*Sembei joined #nim
20:10:28planetis[m]but... the anchored sidebar won't work on mobile :P
20:22:08*yglukhov_ quit (Remote host closed the connection)
20:22:52*yglukhov joined #nim
20:27:21*yglukhov quit (Ping timeout: 264 seconds)
20:35:19*miran quit (Quit: Konversation terminated!)
20:40:38*nsf joined #nim
20:42:00FromGitter<data-man> @exts: I closed your issue as duplicate #2775
20:44:25FromGitter<Varriount> @exts I know that it doesn't always work correctly on mobile browsers. :P
20:46:15*Trustable quit (Remote host closed the connection)
20:56:10*beatmox quit (Remote host closed the connection)
20:56:41*beatmox joined #nim
21:11:35dom96How nice of LLVM devs to add an enum value in the middle of an enum and break my wrapper :(
21:18:28FromGitter<krux02> dom96: well yea that sucks.
21:19:08FromGitter<krux02> I wish C/C++ had a better module system that would provide the entire API in a simple program readable way.
21:20:04FromGitter<krux02> Just for example OpenGL has the entire API specification in an XML document AFAIK, therefore people can just generate wrappers for any language, and wheneven there is new stuff to OpenGL added, it is just a matter of "regenerate"
21:21:08dom96Now wishing Nim allowed a 'when' in a enum declaration
21:22:36dom96https://github.com/nim-lang/Nim/blob/devel/tests/parser/twhen_in_enum.nim :'(
21:32:21*couven92 quit (Ping timeout: 240 seconds)
21:33:52FromGitter<exts> man that's funny, my post got deleted for duplication and the post it references has a post by me 3 years ago on my other account 😂that amuses me
21:33:57*xet7 quit (Quit: Leaving)
21:34:01FromGitter<exts> not deleted closed* excuse me
21:34:43FromGitter<exts> which also references a post by me 😄 from reddit aww man :D
21:36:50*nsf quit (Quit: WeeChat 2.0.1)
21:38:27FromGitter<mratsim> @exts did that ever happen to you: https://imgs.xkcd.com/comics/wisdom_of_the_ancients.png
21:38:49FromGitter<mratsim> Sometimes I find my own post on Nim forum, I read I look at the poster, and oh it’s me ...
21:41:55FromGitter<mratsim> Only 6 bugs away from 100 bugs on the Nim tracker!
21:43:35FromGitter<exts> let me read it lol
21:43:51FromGitter<exts> oh yeah lol
21:44:04FromGitter<exts> it reminds me of my biggest stackoverflow question i posted
21:44:08FromGitter<exts> then answered myself lol
21:44:33FromGitter<mratsim> @Araq, are https://github.com/nim-lang/Nim/issues/5053 and https://github.com/nim-lang/Nim/issues/7632 related. They sound awfully similar.
21:45:07FromGitter<exts> @mratsim i hate when I search for a problem and it's my post :(
21:46:18FromGitter<zacharycarter> I have one of those - https://social.msdn.microsoft.com/Forums/sqlserver/en-US/c2551609-5b54-46e1-a31c-05736ef9bc90/programmatically-create-a-powerpivot-data-refresh-job-in-sharepoint?forum=sqlkjpowerpointforsharepoint
21:46:38FromGitter<zacharycarter> years after I posted it I'd still get people replying and asking me questions about it
21:46:54FromGitter<zacharycarter> even though I no longer worked on .NET
21:47:17FromGitter<Varriount> I have something in a similar vein: https://forum.sublimetext.com/t/using-generators-for-fun-and-profit-utility-for-developers/14618
21:47:29FromGitter<zacharycarter> I even had to involve MS support in the question - and they told me it couldn't be done
21:47:49FromGitter<zacharycarter> or at least they didn't have a way to do it
21:49:02*themagician quit ()
21:52:16*user1101 quit (Ping timeout: 245 seconds)
21:54:42*SenasOzys quit (Ping timeout: 260 seconds)
21:55:53*Jesin joined #nim
22:02:08*SenasOzys joined #nim
22:02:18FromGitter<zacharycarter> gd - https://www.quora.com/Was-object-oriented-programming-a-failure/answer/Michael-O-Church and https://www.quora.com/Why-do-many-software-engineers-not-like-Java depress me
22:02:56FromGitter<zacharycarter> because everyone I work with seems to disregard these truths
22:18:37*rockcavera quit (Ping timeout: 260 seconds)
22:21:19*Jesin quit (Quit: Leaving)
22:21:27*Vladar quit (Quit: Leaving)
22:32:52dom96Man do I hate HTML base URL rules
22:34:45dom96!eval echo 42
22:35:06dom96zacharycarter: why is the playground down? :(
22:35:31dom96!ping
22:35:31NimBotpong
22:35:45NimBotCompile failed: <no output>
22:56:00*rockcavera joined #nim
22:58:50dom96We really need to get this playground onto one of the nim-lang.org VPS'
23:05:58*jjido joined #nim
23:11:21FromDiscord<jos> hey pals
23:11:45FromDiscord<jos> i'm trying to make my project multiplat
23:11:51FromDiscord<jos> there's a native C backend
23:11:57FromDiscord<jos> how can i do this elegantly with nim?
23:12:02FromGitter<zacharycarter> dom96: I don't know :/ something seems to have died with docker on the VPS
23:12:09FromGitter<zacharycarter> I'm working on a new version that's completely dockerized
23:12:27FromGitter<zacharycarter> so that as long as docker is running on the host - you should just have to launch the image w/ docker compose
23:12:39FromGitter<zacharycarter> I think I'll have it ready later this week / weekend at the latest
23:13:27FromGitter<zacharycarter> I haven't been able to get docker running again on the current VPS it's hosted on :/ I tried uninstalling and reinstalling docker but something is borked
23:14:14FromGitter<zacharycarter> sorry - I didn't mean to say VPS - I meant to say server | droplet
23:14:38FromGitter<zacharycarter> whatever the app is currently running on
23:18:55*floppydh quit (Remote host closed the connection)
23:20:05*floppydh joined #nim
23:26:49*jjido quit (Ping timeout: 265 seconds)
23:27:35FromGitter<zacharycarter> jos: use compiler arguments to change what cross compile toolchain you're using
23:27:56FromGitter<zacharycarter> https://nim-lang.org/docs/nimc.html#cross-compilation
23:29:20FromDiscord<jos> i also need to select an implementation
23:29:35FromDiscord<jos> is there a way to do that out of the box? or am i going to ifdef hell
23:29:41FromGitter<zacharycarter> you can do
23:29:46FromGitter<zacharycarter> well yeah - basically that
23:30:07FromDiscord<jos> 😃
23:30:18FromDiscord<jos> i'd prefer to have totally separate files
23:30:23FromDiscord<jos> if that's possible
23:30:31FromDiscord<jos> a common declaration and then two separate c files
23:30:36FromDiscord<jos> without wrapping them in an ifdef
23:31:50FromGitter<zacharycarter> Not sure about that one - I've faced this issue before, but I stopped working on the project shortly after encountering it
23:31:51*krux02 quit (Remote host closed the connection)
23:32:32FromGitter<zacharycarter> I'm not sure if there's some built in functionality that allows you to do this sort of thing or what
23:33:00FromGitter<zacharycarter> a macro - which Araq prescribes for most things - doesn't seem like the right way to handle this IMO
23:35:04FromGitter<zacharycarter> well I mean a user defined one
23:35:50FromGitter<zacharycarter> I now certain existing libraries suffer from this too - so I think the best thing to do is wait for someone smarter than me to chime in ;)
23:35:53*krux02 joined #nim
23:36:10FromDiscord<jos> okey dokey
23:36:20FromDiscord<jos> thanks for the help so far, that's a reasonable temporary solution
23:36:29FromGitter<zacharycarter> sure thing!
23:57:49Araqstop poking me.