<< 14-12-2017 >>
00:00:08*querfisch is now known as hochkantfisch
00:00:23FromGitter<mratsim> @Varriount nedmalloc, jemalloc, hoard ... It's quite competitive even for Cuda there are a lot of allocation schemes: part of my research on Cuda allocation (and how to avoid it) is here: https://github.com/mratsim/Arraymancer/issues/112
00:01:10FromGitter<Varriount> @mratsim How about we just use a bump allocator and let paging do the rest. :P
00:02:21FromGitter<mratsim> From edubart "I've done some benchs against jemalloc, tcmalloc, mkl_alloc, rpmalloc and nim allocator does outperforms all of them."
00:02:30FromGitter<Varriount> @mratsim "It is not nice to outright reserve 70~90% of the GPU memory to run MNIST, what if people are actually using a screen?"
00:03:10FromGitter<Varriount> @mratsim When was that posted?
00:03:17FromGitter<Varriount> (the comment from edubart)
00:03:32FromGitter<mratsim> 2 months ago
00:04:15FromGitter<Varriount> But the new allocator is less than a month old
00:04:39FromGitter<mratsim> By the way before I forget. Arraymancer v0.3 is out. Lots of improvements but super breaking change (ref semantics).
00:05:11FromGitter<Varriount> @mratsim But can I implement Excel with it? :D
00:05:22FromGitter<mratsim> Oh, it's flag-gated I guess? --newallocator ?
00:06:00FromGitter<Varriount> Probably. Or just on the latest development branch.
00:06:09FromGitter<mratsim> @Varriount http://www.deepexcel.net/
00:08:26FromGitter<mratsim> (note: this is an elaborate joke for the sigbovik joke paper conference ;) )
00:10:28FromGitter<Varriount> SIGBOVIK: The Association for Computational Heresy
00:14:46*skrylar joined #nim
00:20:51*hochkantfisch is now known as qwertfisch
00:21:37radagastif stdout.writeLn has been depracated, wha'ts the alternative to echo without flushing stdout?
00:28:19radagastnvm
00:32:00*gokr quit (Ping timeout: 272 seconds)
01:17:53*yglukhov joined #nim
01:22:08*yglukhov quit (Ping timeout: 240 seconds)
01:30:55Tanger@Araq, Hey mate, thanks for answering my question yesterday. I do have an expansion on the original question regarding how result handles properties of objects when it auto-intializes.
01:31:42TangerIf my proc returns an object with a string property, my code tends to segfault unless I specifically call new result
01:32:34FromGitter<zacharycarter> can you share code Tanger?
01:33:21*marenz_ quit (Ping timeout: 248 seconds)
01:34:03FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a31d50b0163b0281033ec3b]
01:34:07FromGitter<zacharycarter> works fine for me
01:38:43Tanger@zacharycarter Whoops, I mistyped. My object is a ref of object, so a ref of Foo in this case
01:56:15FromGitter<zacharycarter> use new then
01:57:16FromGitter<zacharycarter> or
01:57:27FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a31da86cc1d527f6b2c0c47]
01:57:29FromGitter<zacharycarter> Tanger ^
02:00:38TangerNeat, thanks for that @zacharycarter. I like that better than new calls littered around the place
02:01:58FromGitter<zacharycarter> No problem :)
02:16:18FromGitter<zacharycarter> I somehow missed - https://github.com/transfuturist/plugins
02:16:21FromGitter<zacharycarter> that's pretty damn cool
02:16:55FromGitter<zacharycarter> but there's still the whole matter of state :/
02:24:53skrylarzacharycarter: how do things like that work when macros, templates and methods are involved
02:25:23skrylari mean i would have solved it usign binary images smalltalk/lisp style, with a mini rdbms and a jit, but livecoding and general plugisn are very different
02:27:17FromGitter<zacharycarter> good question
02:27:40*endrago__ joined #nim
02:32:41*qih quit (*.net *.split)
02:32:41*Gertm quit (*.net *.split)
02:32:43*endragor quit (*.net *.split)
02:32:43*endragor_ quit (*.net *.split)
02:32:44*pwntus quit (*.net *.split)
02:34:21*yglukhov joined #nim
02:35:00*chemist69 quit (Ping timeout: 246 seconds)
02:36:17*qih joined #nim
02:36:17*Gertm joined #nim
02:36:17*endragor joined #nim
02:36:17*endragor_ joined #nim
02:36:17*pwntus joined #nim
02:39:04*yglukhov quit (Ping timeout: 255 seconds)
02:40:34*vlad1777d quit (Ping timeout: 272 seconds)
02:41:15radagastHi, bit confused about map here. Basically I have a whitespace-seperated string. I want to split, convert to int and find the mininum(just an example). So I am doing echo text.splitWhitespace().map(parseInt).min() but it seems that I only am getting the first int.
02:45:58FromGitter<Quelklef> parseInt from strutils?
02:46:15radagastYes
02:47:33radagasthttps://play.nim-lang.org/?gist=9136be3045bfd95a15dfa63c1fea8ac3
02:47:53FromGitter<Quelklef> oh, there's an online compiler! cool
02:48:31FromGitter<Quelklef> `@[12, 3465, 234, 283, 45].min()` is giving 12
02:48:46FromGitter<Quelklef> So presumably an issue with `min` or seqs rather than map
02:49:14FromGitter<Quelklef> wait
02:49:24FromGitter<Quelklef> that is the minimum number
02:49:27FromGitter<Quelklef> what's the issue
02:49:29*chemist69 joined #nim
02:49:40FromGitter<Quelklef> 12 < 45 < 234 < 283 < 3456
02:49:45FromGitter<Quelklef> what's your desired behaviour?
02:50:01radagastOh boy. hahaha, you are right.
02:50:05radagastSleep depravation
02:50:13skrylarmratsim: coded up winnow learning :b
02:50:21FromGitter<Quelklef> Oops, lol
03:04:40*vivus quit (Quit: Leaving)
03:34:33*d10n-work quit (Quit: Connection closed for inactivity)
03:49:58*yglukhov joined #nim
03:54:07*yglukhov quit (Ping timeout: 248 seconds)
03:58:30*dddddd quit (Remote host closed the connection)
04:23:53*arthurz joined #nim
04:43:05*arthurz quit (Quit: Leaving)
05:37:59*radagast_ joined #nim
05:39:41*radagast quit (Ping timeout: 260 seconds)
05:39:42*radagast_ is now known as radagast
05:49:12*Elronnd- joined #nim
05:54:36*Elronnd- left #nim (#nim)
06:03:13*yglukhov joined #nim
06:04:17*yglukhov quit (Remote host closed the connection)
06:04:30*yglukhov joined #nim
06:10:56*yglukhov quit ()
06:11:56*gokr joined #nim
06:25:44*nsf joined #nim
06:26:28*yglukhov joined #nim
06:42:51*xkapastel quit (Quit: Connection closed for inactivity)
07:14:46*yglukhov quit (Remote host closed the connection)
07:17:10*yglukhov_ joined #nim
07:18:46*PMunch joined #nim
07:19:24*solitudesf joined #nim
07:21:46*yglukhov_ quit (Ping timeout: 272 seconds)
07:45:28*chemist69 quit (Ping timeout: 240 seconds)
07:45:51*chemist69 joined #nim
07:56:10FromGitter<mratsim> @skrylar didn't know about it, interesting
07:57:26skrylarseems to have been used for spam filtering in the past, but i was reading about decision trees and the C5.0 algo uses it
07:57:46skrylarI thought making decision trees was going to be a lot of binary searching to find ideal split points, and then find out they just split on means :|
07:59:51*kunev quit (Ping timeout: 240 seconds)
08:02:23*vlad1777d joined #nim
08:02:29*kunev joined #nim
08:12:27*yglukhov joined #nim
08:26:19FromGitter<mratsim> For spam, I thought Naive Bayes was the antispam 101.
08:26:55FromGitter<mratsim> Oh, maybe I can sell Bayesian neural net to Gmail :D.
08:27:28skrylarit depends
08:28:40skrylari read the crm104 manual front to back, and paul graham's post on the matter, and naive bayes is great if you train it yourself; then they have some extra thing (Markovian, OSB) which stores word features in a short sequence as well, so it can learn the difference between "heil hitler" and "hitler sucks"
08:29:08skrylaror "nim ? ? great ?" and all those fun permutations
08:29:45*qih quit (Quit: leaving)
08:35:45*claudiuinberlin joined #nim
09:02:55*gmpreussner quit (Ping timeout: 248 seconds)
09:05:29*gmpreussner joined #nim
09:13:55skrylarapparently winnow with sparse bigrams outperforms naive bayes
09:14:08skrylarwhich i suppose makes sense, since sparse bigrams add context
09:15:39*chemist69 quit (Quit: WeeChat 1.9.1)
09:21:54*floppydh joined #nim
09:27:10*solitudesf quit (Ping timeout: 272 seconds)
09:39:21*vlad1777d quit (Ping timeout: 264 seconds)
10:03:23FromGitter<mratsim> @dom96 guibar post is not visible here https://forum.nim-lang.org/t/3415
10:31:32*jjido joined #nim
10:36:35*dddddd joined #nim
10:54:29FromGitter<tim-st> When I have two procs that take a range of same min but different max and I chain them so that the outer proc has bigger max I dont get a compiler warning, I think that would be good, wouldn't it?
11:06:30*chemist69 joined #nim
11:12:55Araqtim-st: maybe but it fails at runtime at least
11:16:57*marenz_ joined #nim
11:16:58*skelett joined #nim
11:25:15*Vladar joined #nim
11:33:54*jjido quit (Quit: http://www.kiwiirc.com/ - A hand crafted IRC client)
11:44:12*SenasOzys quit (Remote host closed the connection)
11:53:47*Arrrr joined #nim
11:53:47*Arrrr quit (Changing host)
11:53:47*Arrrr joined #nim
11:54:54ArrrrHello nimroders
12:07:18Araqhi
12:12:46yglukhovcan i walkdir in a nimble file?
12:15:20yglukhovok, found it. import oswalkdir + walkDir()
12:21:39FromGitter<Varriount> There's really an oswalkdir module?
12:30:39*nc-x joined #nim
12:32:46nc-xI always have this problem when i open the nim docs https://imgur.com/a/ZJBEJ , it goes away on ctrl+F5 though.
12:35:18yglukhovVarriount: yup =)
12:36:11FromGitter<Varriount> yglukhov: Erg. I really wish we could just use the procedures in the os module
12:36:29FromGitter<Varriount> I dislike all these "special" modules
12:36:41yglukhovVarriount: +1
12:37:24FromGitter<Varriount> nc-x: Same thing happens to me on the main forum page. I think it's something with the css
12:41:01nc-x@Araq @dom96 - Another issue - https://imgur.com/YBnLOlD (in vscode, goto definition is **sometimes** wrong), Issue with nimsuggest or vscode plugin, or expected? Should i open issue on github/nim-lang/nim?
12:41:02*endrago__ quit (Quit: Leaving...)
12:41:13*endrago__ joined #nim
12:43:36yglukhovAraq: this processPendingCallbacks should be hidden, imo. in fact, it should be gone altogether.
12:43:37yglukhovimo
12:44:06yglukhovAraq: pls consider this https://github.com/nim-lang/Nim/pull/6614
12:46:30*endragor_ quit (Quit: Leaving)
12:51:49Araqyglukhov: yeah was a misunderstanding on my part, will revert
12:53:04planetis[m]anyone played hackmud? :p
12:55:40*Arrrr quit (Read error: Connection reset by peer)
13:00:16Araqyglukhov: what do you think about 'drain'?
13:01:21yglukhovAraq: sorry, I haven't read your changes carefully, and got no idea what you're doing. Just noticed you were messing with processPendingCallbacks
13:04:50*solitudesf joined #nim
13:05:50*SenasOzys joined #nim
13:09:28*endrago__ quit (Quit: Leaving...)
13:09:40*endragor_ joined #nim
13:09:58FromGitter<mratsim> @dom96: you replied to an invisible OP: https://forum.nim-lang.org/t/3418
13:11:22dom96ugh, the forum needs to be fixed
13:11:23dom96thanks
13:13:12dom96Araq: Can you explain what the use case for this 'drain' is?
13:15:23FromGitter<alehander42> I guess for cases when errTypeExpected comes from a nkCall, maybe the compiler can append a "possible hint" ? like ⏎ `maybe you wanted a named argument? use f(arg=value) instead `
13:16:48*endragor quit (Quit: Leaving)
13:17:25FromGitter<mratsim> Shadows of Elm, coming soon in your favorite compiler.
13:17:37dom96yes, I've wanted hints in compiler's messages since forever.
13:17:47dom96I've added them to choosenim
13:17:55FromGitter<alehander42> hints are amazing in some cases
13:18:06FromGitter<alehander42> also the error points to the beginning of the call
13:18:11FromGitter<alehander42> so even column info doesn't help
13:18:20FromGitter<alehander42> for unexperienced users
13:18:21*endragor_ quit (Quit: Leaving...)
13:18:21dom96that's a bug then
13:18:31*endragor joined #nim
13:18:52FromGitter<alehander42> well at least for a simple example like ⏎ ⏎ ```proc s(e: int) = ⏎ echo e ⏎ ⏎ echo s(e: 2)``` [https://gitter.im/nim-lang/Nim?at=5a327a3bba39a53f1a5d5aa4]
13:24:27Araqdom96, see the feature request that I linked to
13:25:16Araqspend too much time discussing this, I don't understand anything anymore and prefer to work on things I consider more important
13:27:17*BitPuffin|osx joined #nim
13:28:18*nc-x quit (Quit: Page closed)
13:28:25dom96Araq: I don't see any feature request you linked to.
13:28:59Araq; implements #6523
13:29:07Araqhttps://github.com/nim-lang/Nim/issues/6523
13:30:34Araqalehander42, be my guest
13:31:04dom96Araq: That doesn't explain the use case.
13:31:26dom96In what situation is that a problem?
13:31:40FromGitter<alehander42> @Araq about the error hint?
13:31:50Araqsure
13:33:31Araqdom96, when you need to call poll() but want it to perform the possible IOs, like making httpclient process all HTTP headers instead of one line in the buffer. at least that's how I understand it.
13:35:42yglukhovAraq: doesn't my PR fix it?
13:36:17yglukhovmy pr makes the futures complete immediately
13:36:26yglukhovdom96: ping as well
13:36:36Araqyglukhov, IO completion ports currently require a call to poll() to GC_unref() the PCustomOverlapped
13:36:58Araqso your PR introduces leaks. tons of them.
13:37:36Araq(I think.)
13:37:40yglukhovno it doesn't
13:38:52yglukhovmy pr doesn't imply that poll() is not needed anymore.
13:38:52dom96Yeah, I don't see how yglukhov's PR introduces leaks
13:39:03dom96and if it solves #6523 then I prefer it
13:39:29dom96Araq: I've never needed to have this behaviour for poll(), can you elaborate?
13:39:43dom96Why do you want to process everything with a single poll()?
13:40:19AraqI think it's just about efficiency
13:41:27yglukhovas far as i understand, there endragor wants a way to make sure there's nothing waiting for completion
13:41:29Araqhttps://stackoverflow.com/questions/29782377/is-it-possible-to-run-only-a-single-step-of-the-asyncio-event-loop
13:41:45yglukhovto e.g. run some side event source
13:43:18yglukhovhey guys, please check out my new toy project: https://github.com/yglukhov/python
13:51:42planetis[m]wow nice!
13:51:50dom96yglukhov: awesome
13:52:55dom96but you'll need to rename the package :)
13:54:19dom96oh and btw, convention is `nimble test` not the plural :)
14:01:21*solitudesf quit (Ping timeout: 260 seconds)
14:01:31FromGitter<mratsim> Oh @yglukhov that's good. Nim-pymod was already not bad
14:02:23FromGitter<alehander42> very interesting @yglukhov
14:02:33yglukhovmratsim: yeah, unlike nim-pymod i tried to do everything in a single compile time
14:02:48yglukhovno scripts required
14:02:57yglukhovno outputs except the single .so
14:03:51*xkapastel joined #nim
14:04:02yglukhovyou don't even need python installed to make the extension
14:09:48*xkapastel quit (Ping timeout: 240 seconds)
14:11:08*xkapastel joined #nim
14:14:20*solitudesf joined #nim
14:18:38FromGitter<alehander42> @yglukhov does it work for python2 ?
14:18:50yglukhovalehander42: works for me =)
14:18:57yglukhov2.7 at least
14:20:22FromGitter<alehander42> I have to play a bit with it
14:21:47FromGitter<alehander42> very nice, I guess almost the same approach can work for yarv?
14:24:31*gokr quit (Ping timeout: 248 seconds)
14:25:26*natrys joined #nim
14:28:19FromGitter<mratsim> Have to explore how I can use it in Arraymancer, I think it would be very good to have a Python wrapper.
14:35:29*PMunch quit (Quit: Leaving)
14:38:07yglukhovmratsim: i want to make types/classes exportable as well. eventually... =)
14:38:30yglukhovalehander42: whats yarv?
14:43:46FromGitter<alehander42> the ruby vm
14:44:32yglukhovalehander42: right. sure, it will work as long as yarv has some c extension api
14:46:20yglukhovalehander42: there's also jnim (java - nim bridge), which leverages nim powers in a similar way, if you need any inspiration sources =)
14:50:13*claudiuinberlin quit (Quit: Textual IRC Client: www.textualapp.com)
14:51:09FromGitter<alehander42> thanks
14:52:15FromGitter<alehander42> I have started something related a month ago, but maybe I'll have time to work on it after holidays
14:53:57*natrys quit (Ping timeout: 256 seconds)
14:57:01*yglukhov quit (Remote host closed the connection)
14:57:35*yglukhov joined #nim
15:00:00*skrylar quit (Remote host closed the connection)
15:01:53*yglukhov quit (Ping timeout: 256 seconds)
15:03:21*sendell joined #nim
15:17:08*endragor quit (Remote host closed the connection)
15:32:28*gokr joined #nim
15:33:14chemist69yglukhov: awesome project!
15:56:44*natrys joined #nim
16:05:32*gokr quit (Ping timeout: 272 seconds)
16:11:04*fvs joined #nim
16:13:11*yglukhov joined #nim
16:14:04*yglukhov quit (Remote host closed the connection)
16:14:21*yglukhov joined #nim
16:15:14*yglukhov quit (Read error: Connection reset by peer)
16:15:48*yglukhov joined #nim
16:18:52FromGitter<tim-st> Is there a reverse function from ord() for enums? instead of ord(Enum) I want enum(int)?
16:20:33FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a32a4d187680e62301079f3]
16:21:32FromGitter<tim-st> @zacharycarter perfect thanks, didn't see this in the docs.
16:22:01FromGitter<zacharycarter> no problem
16:23:49*nsf quit (Quit: WeeChat 1.9.1)
16:40:34*floppydh quit (Quit: WeeChat 1.9.1)
16:43:16*Arrrr joined #nim
17:00:18*Trustable joined #nim
17:02:44FromGitter<mratsim> @alehander42 this made hacker news: https://github.com/EntilZha/PyFunctional
17:03:48*yglukhov quit (Read error: Connection reset by peer)
17:04:20*yglukhov joined #nim
17:08:01FromGitter<Greglefox> Hello everyone. I am trying to get the boehm GC working but at the execution I have a message saying boehmgc.dll required
17:08:28FromGitter<Greglefox> Where can I find the DLL already compiled for windows 7 64 bits?
17:10:03*miran joined #nim
17:16:42FromGitter<Yardanico> @Greglefox why are you trying to use boehm gc in the first place?
17:16:47FromGitter<Yardanico> it's not required for normal nim usage
17:16:59FromGitter<Greglefox> shared heap
17:17:23FromGitter<Greglefox> between thread
17:17:24FromGitter<zetashift> I can't find a dll, but I am on w64 I could try compiling it
17:18:50FromGitter<Yardanico> @Greglefox why you don't want to compile it yourself?
17:18:56FromGitter<Yardanico> because you already have C compiler 100%
17:19:48FromGitter<Greglefox> i am a bit a newbee on C compiler and makefile for Windows
17:20:39FromGitter<Yardanico> well I can't find any precompiled boehm dlls on google
17:20:50*xkapastel quit (Quit: Connection closed for inactivity)
17:22:08FromGitter<Greglefox> same here. I was asking in case of someone already successfully compiled the bohemgc 7.7.0 for w64
17:22:16FromGitter<Yardanico> and there are some build instructions - http://www.hboehm.info/gc/
17:22:20FromGitter<alehander42> cool @mratsim
17:35:10*drs joined #nim
17:44:08*sendell quit (Remote host closed the connection)
17:56:40FromGitter<Yardanico> am I right here btw? https://github.com/nim-lang/Nim/issues/6924#issuecomment-351787252
17:56:53FromGitter<Yardanico> i think the issue is that he uses `arm.linux` instead of `arm.android`
18:01:48*lastjedi joined #nim
18:11:08*lastjedi quit (Ping timeout: 240 seconds)
18:16:15*Jesin joined #nim
18:18:06*d10n-work joined #nim
18:45:11*yglukhov quit (Remote host closed the connection)
18:45:46*yglukhov joined #nim
18:47:50*Jesin quit (Quit: Leaving)
18:48:03*PMunch joined #nim
18:50:28*yglukhov quit (Ping timeout: 268 seconds)
18:57:32*yglukhov joined #nim
18:59:47*PMunch quit (Quit: leaving)
19:00:02*Jesin joined #nim
19:00:45*yglukhov quit (Read error: Connection reset by peer)
19:01:20*yglukhov joined #nim
19:21:00*nsf joined #nim
19:21:24*Vladar quit (Quit: Leaving)
19:29:21*BitPuffin|osx quit (Ping timeout: 256 seconds)
19:31:41FromGitter<cabhishek> is it possible to defined a table type and instantiate it later on? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a32d19d1a4e6c82230cd429]
19:39:49*PMunch joined #nim
19:40:50FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a32d3c2ffa3e379193a29ee]
19:51:42subsetparkzacharycarter: I think he's asking if he can avoid repeating the string, seq[int] part
20:03:45*xkapastel joined #nim
20:14:12FromGitter<mratsim> With TableRef and new?
20:23:20fvsWhat happens to declared but not used variables - do they end up in the compiled binary?
20:23:50*miran quit (Quit: Konversation terminated!)
20:28:37Araqno
20:35:00*vlad1777d joined #nim
20:38:07FromGitter<cabhishek> was trying to find a way to avoid `string, seq[int]` repetition
20:39:08Araqthere is none. sad, I know.
20:39:30Arrrrsets module has `init`
20:43:59*gokr joined #nim
20:58:20*natrys quit (Quit: natrys)
21:03:23*Trustable quit (Remote host closed the connection)
21:08:27*solitudesf quit (Read error: Connection reset by peer)
21:09:55*solitudesf joined #nim
21:12:47FromGitter<mratsim> If I need to read a stream of uint8/unsigned byte in a binary file, as there is no readByte, I should use readChar from the streams module and then cast it to uint8 right?
21:14:32ArrrrYes, and you can use readData too
21:17:44Araqplease use readBuffer/Data and read more than one byte at a time
21:18:08*Arrrr quit (Read error: Connection reset by peer)
21:19:54FromGitter<mratsim> I don’t think I can, this is the format:
21:19:58FromGitter<mratsim> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5a32eafe5355812e57d752c7]
21:20:19FromGitter<mratsim> And it’s big endian ….
21:20:58FromGitter<mratsim> It’s 60000 28x28 images of handwritten digits
21:21:29FromGitter<mratsim> So I have to store them in a Tensor[uint8]
21:23:12FromGitter<mratsim> Or maybe I can just copy into a proper sized seq
21:24:49Araqyay a 10 line codegen fix leads to linker errors for bootstrapping
21:26:56FromGitter<mratsim> Yeah I think copyMem into a seq[uint8] should work
21:27:17FromGitter<mratsim> readData*
21:34:53*claudiuinberlin joined #nim
22:02:30*PMunch quit (Quit: leaving)
22:07:13FromGitter<mratsim> What’s the nim usual location to store data in the user homedir, in ~/.packagename ?
22:10:03*vlad1777d quit (Remote host closed the connection)
22:11:50*vlad1777d joined #nim
22:13:06*nsf quit (Quit: WeeChat 1.9.1)
22:16:41*claudiuinberlin quit (Quit: Textual IRC Client: www.textualapp.com)
22:25:45FromGitter<alehander42> `~/.local/share/<packagename>` ? just a guess
22:33:43*solitudesf quit (Quit: solitudesf)
22:33:45FromGitter<mratsim> In this case it’s to load example datasets like this one https://en.wikipedia.org/wiki/MNIST_database
22:34:45*d10n-work quit (Quit: Connection closed for inactivity)
22:35:36*yglukhov quit (Read error: Connection reset by peer)
22:36:12*yglukhov joined #nim
22:47:59*yglukhov quit (Read error: Connection reset by peer)
22:48:02*geocar joined #nim
22:48:31*yglukhov joined #nim
23:16:33*gokr quit (Ping timeout: 264 seconds)
23:19:20*yglukhov quit (Remote host closed the connection)
23:19:54*yglukhov joined #nim
23:24:08*yglukhov quit (Ping timeout: 240 seconds)
23:24:34GitDisc<awr> I'm generating docs on a module that has *a lot* of macro-generated public procedures, how do I get nim to not generate documentation for these procs?
23:29:24Araqif it's exported you can't avoid producing docs for it
23:35:21radagast*sigh* The manual could be nicer
23:37:48FromGitter<raydf> @zacharycarter i been experimenting with the ui for a mobile app in cordova, i'm still using a second javascript file for some integrations with karax (callbacks), my main problem is ffi and some bindings that i don't understand in the nim js backend, but that's more lazyness from my part :). For the virtual dom it looks amazing, i'll try to contribute with additional examples but for the moment, the ones that @Araq
23:37:48FromGitter... provides are sufficient. I uploaded the todo list app to my android phone and the iphone 6s and it works perfectly with cordova, instant feedback by the virtual dom in the below 20ms range. So i could recommend the framework as a better alternative than angular and vuejs for nim programmers. Maybe later we could see a ... [https://gitter.im/nim-lang/Nim?at=5a330b4cffa3e379193b6d77]
23:39:29FromGitter<zacharycarter> radagast - I'm sure PR's are welcome :)
23:41:16FromGitter<zacharycarter> @raydf awesome! I've never played with cordova before. I have to run for a bit - and I haven't looked at react native's code. But I will send more thoughts when I get back.
23:43:05FromGitter<Varriount> @Araq Was it really necessary to create duplicates of all the random procedures? Why not just replace the existing ones?
23:43:30Araqbehaviour changed
23:46:54FromGitter<Varriount> But it looks ugly now. :(
23:48:34FromGitter<Varriount> I think this is something that deserves breaking backwards compatibility.