<< 02-12-2016 >>
00:01:59*rynsin quit (Quit: rynsin)
00:17:23*rynsin joined #nim
00:18:02*rynsin quit (Client Quit)
00:21:40*Matthias247 quit (Read error: Connection reset by peer)
00:36:21*yglukhov joined #nim
00:40:48*yglukhov quit (Ping timeout: 260 seconds)
00:43:53*themagician quit ()
00:47:16*chemist69 quit (Ping timeout: 258 seconds)
00:47:42*cheatfate quit (Read error: Connection reset by peer)
00:48:43*cheatfate joined #nim
01:05:09*SianaGea1z quit (Ping timeout: 248 seconds)
01:11:01*Sentreen quit (Ping timeout: 248 seconds)
01:14:17*chemist69 joined #nim
01:18:55*yglukhov joined #nim
01:21:40*Sentreen joined #nim
01:23:39*yglukhov quit (Ping timeout: 260 seconds)
01:36:17*SianaGearz joined #nim
01:52:36*libman quit (Remote host closed the connection)
01:53:09*ibk joined #nim
01:55:01*Varriount|Mobile joined #nim
01:55:12Varriount|Mobilevktec: Yes
01:55:43Varriount|MobileHowever procedures defined for B can't be used for A
01:55:46*Varriount|Mobile quit (Client Quit)
02:01:22*yglukhov joined #nim
02:05:37*yglukhov quit (Ping timeout: 240 seconds)
02:13:54*brechtm_ quit (Read error: Connection reset by peer)
02:14:32*brechtm joined #nim
02:17:12*boop quit (Ping timeout: 265 seconds)
02:17:37*boop joined #nim
02:21:34*bjz quit (Ping timeout: 244 seconds)
02:22:08*bjz joined #nim
02:43:41*yglukhov joined #nim
02:45:03*myp joined #nim
02:47:58*yglukhov quit (Ping timeout: 245 seconds)
03:00:37*chemist69 quit (Ping timeout: 260 seconds)
03:03:34*couven92 quit (Quit: Client disconnecting)
03:14:18*chemist69 joined #nim
03:19:39*bjz_ joined #nim
03:20:49*bjz quit (Ping timeout: 268 seconds)
03:26:02*yglukhov joined #nim
03:30:24*yglukhov quit (Ping timeout: 246 seconds)
03:34:08*vlad1777d quit (Remote host closed the connection)
03:39:24*byte512 quit (Ping timeout: 250 seconds)
03:40:50*brson quit (Quit: leaving)
03:43:07FromGitter<ftmamud_twitter> Hello everyone! :smile:
03:52:30*krux02 quit (Remote host closed the connection)
04:08:22*yglukhov joined #nim
04:09:15*rynsin joined #nim
04:12:04*rynsin_ joined #nim
04:12:04*rynsin_ quit (Client Quit)
04:12:18*rynsin quit (Read error: Connection reset by peer)
04:13:02*yglukhov quit (Ping timeout: 250 seconds)
05:03:30*space-wizard quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
05:03:58*space-wizard joined #nim
05:04:15*space-wizard quit (Client Quit)
05:04:43*space-wizard joined #nim
05:05:03*space-wizard quit (Client Quit)
05:05:32*space-wizard joined #nim
05:05:51*space-wizard quit (Client Quit)
05:06:20*space-wizard joined #nim
05:06:39*space-wizard quit (Client Quit)
05:07:08*space-wizard joined #nim
05:07:27*space-wizard quit (Client Quit)
05:07:59*space-wizard joined #nim
05:08:15*space-wizard quit (Client Quit)
05:08:18*rynsin joined #nim
05:08:48*space-wizard joined #nim
05:09:03*rynsin quit (Client Quit)
05:09:03*space-wizard quit (Client Quit)
05:56:26*nsf joined #nim
06:08:34*rynsin joined #nim
06:15:42*yglukhov joined #nim
06:20:10*yglukhov quit (Ping timeout: 250 seconds)
06:30:10*rynsin quit (Quit: rynsin)
06:35:41*vendethiel quit (Ping timeout: 268 seconds)
06:35:54*vendethiel joined #nim
06:49:36*Guest80905 is now known as wuehlmaus
06:57:43*yglukhov joined #nim
07:00:37*myp quit (Ping timeout: 240 seconds)
07:01:06FromGitter<vegansk> What's wrong with dash in the module's name? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=58411c3216207f7b0ed1d2b0]
07:02:02*yglukhov quit (Ping timeout: 250 seconds)
07:02:30*myp joined #nim
07:26:04*rokups joined #nim
07:40:13*yglukhov joined #nim
07:44:30*yglukhov quit (Ping timeout: 250 seconds)
08:00:19*Arrrr joined #nim
08:11:53*vendethiel quit (Ping timeout: 268 seconds)
08:13:05*vendethiel joined #nim
08:13:57ArrrrI guess there isn't a way to import several modules from the same folder at the same time, isnt it? Like "import tests/[Test1,Test2]" or something ugly like that
08:18:07ibkHi all, is there a kind of todo list in http(client, server, anything related) modules? I can't promise anything, but i will contribute if i can, my company has interests on it, and i can see that it still lack of many features
08:22:36Arrrrhttps://github.com/nim-lang/Nim/issues?q=is%3Aopen+is%3Aissue+label%3Aasync
08:33:28*vendethiel quit (Ping timeout: 256 seconds)
08:44:54*brechtm_ joined #nim
08:48:16*brechtm quit (Ping timeout: 268 seconds)
08:53:44*Andris_zbx joined #nim
09:31:41*yglukhov joined #nim
09:32:01*yglukhov quit (Read error: Connection reset by peer)
09:35:03*cheatfate_ joined #nim
09:37:57*cheatfate quit (Ping timeout: 240 seconds)
09:38:08*BlaXpirit_ joined #nim
09:38:15*nim-buildbot quit (Ping timeout: 246 seconds)
09:38:18*vlad1777d joined #nim
09:38:23*BlaXpirit quit (Ping timeout: 245 seconds)
09:38:44*BlaXpirit_ is now known as BlaXpirit
09:46:51*vlad1777d quit (Remote host closed the connection)
10:29:05*arnetheduck joined #nim
10:58:58*themagician joined #nim
11:12:05*vlad1777d joined #nim
11:18:51*kier quit (Remote host closed the connection)
11:23:13*dmi0 quit (Ping timeout: 260 seconds)
11:23:58*yglukhov joined #nim
11:28:41*yglukhov quit (Ping timeout: 265 seconds)
11:47:48*yglukhov joined #nim
11:50:56*elrood joined #nim
11:51:44*nsf quit (Quit: WeeChat 1.6)
11:56:56vktec<vktec> If I do type A = B, can I use procs defined for B on A?
11:56:56vktec<Varriount|Mobile> vktec: Yes
11:56:56vktec<Varriount|Mobile> However procedures defined for B can't be used for A
11:57:09vktecUmmm... so can they or can't they? 0_o
11:58:47*shodan45 quit (Ping timeout: 260 seconds)
12:00:35vktecFrom playing around, it seems they can't, so I guess I should type A = ref B and then deref?
12:03:22*byte512 joined #nim
12:14:10*myp quit (Read error: Connection reset by peer)
12:14:23flyxvktec: if you want procs that work on A, why don't you just define them for A?
12:15:29flyxvktec: ah wait, your naming is confusing. B ist the base type, A is the new type
12:15:54federico3Nim benchmarks https://github.com/costajob/app-servers
12:16:03flyxI guess what Varriount wanted to say is that you cannot use procs defined for A on B
12:19:05*dmi0 joined #nim
12:19:32*myp joined #nim
12:19:39vktecAh, okay
12:23:50*kier joined #nim
12:30:02*couven92 joined #nim
13:01:29*azur_kind joined #nim
13:52:12*brechtm joined #nim
13:55:00*brechtm_ quit (Ping timeout: 250 seconds)
13:58:33*Arrrr quit (Ping timeout: 244 seconds)
14:01:58*Arrrr joined #nim
14:20:17*brechtm quit (Ping timeout: 240 seconds)
14:28:06*brechtm joined #nim
14:28:32*wuehlmaus quit (Ping timeout: 250 seconds)
14:32:06yglukhovAraq: hi, where does --genMapping put the mapping?
14:58:14*wuehlmaus joined #nim
14:58:38*wuehlmaus is now known as Guest46490
15:34:29vktecHmmm... I'm getting 'Error: module names need to be unique per Nimble package' and I'm curious as to why this is. Surely modules under different subdirectories can have the same name?
15:37:08*Guest46490 quit (Ping timeout: 245 seconds)
15:43:12vktecIs there any way to forward all symbols in an imported module? I've tried `export foo` and `export foo.*` but neither seem to work as expected
15:57:56Araqvktec: Nim wants package.module.symbol to be universally unique
15:58:08vktecAh, okay
15:58:56*shodan45 joined #nim
15:58:58vktecSo it's not package.subdir.module.symbol?
16:00:08Araqno it's not, and 'subdir' is hard to recover, you could have it in your --path
16:00:13Araqyglukhov: I don't know :-)
16:00:36Araqin fact, that's reason number 2. same module names are simply confusing
16:00:53vktecAraq: Fair enough. I guess I'm familiar with the Python way of doing things :)
16:02:36AraqArrrr: you could write a macro that does that, but I agree with a syntax for this
16:04:37vktecAraq: Is there something like `from foo export nil` (ie. the opposite of `from foo import nil`)?
16:06:35*wuehlmaus joined #nim
16:06:52Araqwhat does that mean?
16:06:58*wuehlmaus is now known as Guest94556
16:07:00Araqto export nothing, don't write an export :P
16:07:19vktecWell from foo import nil imports all symbols, enforcing full qualification, right?
16:07:29vktecI'd like to export all symbols from a module
16:12:50*gangstacat quit (Quit: Ĝis)
16:15:55*cheatfate_ is now known as cheatfate
16:17:41*gangstacat joined #nim
16:29:21*chemist69 quit (Ping timeout: 260 seconds)
16:29:27*Trustable joined #nim
16:33:55*chemist69 joined #nim
16:37:39*Andris_zbx quit (Remote host closed the connection)
16:44:17*Arrrr quit (Quit: WeeChat 1.5)
16:47:22dom96vktec: module-with-dashes doesn't work because `-` is an operator in Nim,
16:47:38dom96and the operator does not require spaces around it
16:47:59vktecdom96: Umm... wrong person :P
16:48:15dom96oh, sorry, *vegansk
16:48:19*vktec redirects to @vegansk
16:48:46dom96as for your question around type A = B, I don't think Varriount's answer is correct unless I'm misunderstanding
16:48:50dom96That is simply an alias
16:49:08dom96so proc foo(x: A) is the same as proc foo(x: B)
16:49:59vktecOkay, cool
16:50:20vktecWhat about reexporting all symbols in a module?
16:50:42dom96federico3: Regarding the benchmarks, I'm actually impressed by asynchttpserver's performance.
16:51:11vktecbrb, rebooting to try and fix my browser D:
16:51:14dom96vktec: don't think there is a way to do that.
16:51:26vktecOh, that's a shame
16:51:35vktecI guess I'll have to do it manually then :-/
16:52:10dom96It's better to use the asterisk
16:52:16dom96For everything that you want to import
16:52:29dom96That way you can easily make something private
16:52:29federico3dom96: the threaded vs unthreaded comparison was a bit unfair
16:52:46dom96federico3: indeed. In addition, Crystal appears to be using libevent
16:52:57dom96I'm curious how Nim will perform on this https://blog.pusher.com/golangs-real-time-gc-in-theory-and-practice/
16:53:13dom96I might port their benchmark to Nim tonight
16:53:26federico3please do ;)
16:53:44*shodan45 quit (Quit: Konversation terminated!)
16:55:08vktecdom96: Oh, I may have been unclear. Say I have modules A, B and C. Module A imports module B, which then imports module C. I want module A to be able to access all symbols in module C, without explicitly importing it
16:55:50Xedom96: i've been wondering about making nim's runtime output data compatible with the go profiling tools
16:56:17dom96Xe: That would be awesome, do you think it's easily do-able?
16:56:50Xedom96: i haven't been able to find out much about how the nim runtime profiling stuff works
16:56:55*nsf joined #nim
16:58:40dom96Xe: Afraid I can't help much here, except suggest that you look at the nimprof module.
17:11:04dom96Damn, somebody beat me to it! http://forum.nim-lang.org/t/2646
17:11:14dom96Not that I'm complaining :)
17:12:53*yglukhov_ joined #nim
17:13:16*rokups quit (Quit: Connection closed for inactivity)
17:16:03*yglukhov quit (Ping timeout: 246 seconds)
17:17:39*yglukhov_ quit (Ping timeout: 265 seconds)
17:23:29*azur_kind quit (Remote host closed the connection)
17:28:51jh32hi
17:29:24vktecHi
17:30:04jh32i still get errors sometimes with parallelBuild on DragonFlyBSD it is supposed to be fixed some time ago for FreeBSD. Is anyone else getting linker errors (files missing)?
17:30:37jh32and then if you check manually the object files are there
17:46:40jh32it's no big deal i'm mostly annoyed that i can't find out what's going on...
17:50:44dom96cheatfate: do you know anything about jh32's issues?
17:50:50*space-wizard joined #nim
17:51:23*libman joined #nim
17:52:14*brechtm_ joined #nim
17:54:12*nsf quit (Quit: WeeChat 1.6)
17:55:15*brechtm quit (Ping timeout: 246 seconds)
17:55:23cheatfatedom96, nope
17:55:47cheatfatebut i'm not using parallelBuild - even dont know what it is
18:00:00*Matthias247 joined #nim
18:03:26*yglukhov joined #nim
18:08:05*yglukhov quit (Ping timeout: 248 seconds)
18:10:54Araqcheatfate: you do use it, it's the default
18:11:41Araqhttps://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#Common-Function-Attributes # when somebody comes up with how big Nim is because of pragmas, show him this
18:11:41cheatfatethen, i have no any problems on freebsd
18:11:58Araqis DragonFlyBSD still alive?
18:12:03AraqI thought it was dead.
18:14:56Araqhmm I need __declspec(selectany) for Gnu C and LLVM
18:14:59Araqany ideas?
18:20:47*Trustable quit (Remote host closed the connection)
18:25:20elroodAraq, probably __attribute__((weak))
18:26:36*yglukhov joined #nim
18:28:25*brechtm_ quit (Remote host closed the connection)
18:30:27jh32Araq: it's pretty alive - small but very nice community
18:31:27jh32cheatfate: good to know it works on freebsd
18:32:02elroodAraq, clang might directly support __declspec with -fms-extensions. gcc in theory should too, but the implementation is a little sketchy, afaik
18:33:26cheatfatejh32, what is default c compiler for dragon? gcc or clang? if gcc - what version?
18:34:42jh32cheatfate: in nim dragonfly is just mapped to freebsd - it's using clang here
18:35:55cheatfatedragonfly support clang?
18:36:27Araqelrood: well if I read the docs correctly it's only available on Windows
18:38:15jh32cheatfate: yes, it's even in base but i use the ports version - i could play with different compilers to see if it makes a difference
18:38:48cheatfatejh32, so with all compilers you got a problem? could you gist it?
18:43:04*rolha joined #nim
18:44:08jh32cheatfate: i will check it - so far i only tested with clang 3.8.1
18:46:56*yglukhov quit (Remote host closed the connection)
18:47:11elroodAraq, it being ms-extensions? i don't think so, but since i'm not sure about the backend implementation and implications, can't test now and don't know any better i'd say it's possible. all i can confidently say is that a snippet with __declspec compiles with -fms-extensions on linux
18:48:49*Senketsu joined #nim
18:50:34jh32cheatfate: hm, no problem with gcc so far, but it could be that it's just not hitting the race due to different timing
18:50:53*yglukhov joined #nim
18:51:22elroodAraq, with clang, that is. gcc complains
18:54:00Araqelrood: ah, ok.
18:54:19Araqso gcc doesn't really support it, gotcha.
19:01:06*rolha quit (Remote host closed the connection)
19:01:09*yglukhov quit ()
19:01:42*rolha joined #nim
19:02:59*[ui] joined #nim
19:04:04*nsf joined #nim
19:12:29*yglukhov joined #nim
19:19:29*yglukhov quit (Remote host closed the connection)
19:23:45*nsf quit (Quit: WeeChat 1.6)
19:26:47*yglukhov joined #nim
19:32:41*dddddd joined #nim
19:34:01*ibk quit (Remote host closed the connection)
19:34:01*subsetpark quit (Remote host closed the connection)
19:34:01*[ui] quit (Remote host closed the connection)
19:34:02*zielmicha_ quit (Remote host closed the connection)
19:34:02*NhanH quit (Remote host closed the connection)
19:34:02*r4vi quit (Remote host closed the connection)
19:34:35*dmi0 quit (Quit: ~)
19:35:58*yglukhov quit (Remote host closed the connection)
19:38:18*Matthias247 quit (Quit: Matthias247)
19:40:57*NhanH joined #nim
19:42:22*yglukhov joined #nim
19:46:26*[ui] joined #nim
19:47:26*zielmicha_ joined #nim
20:04:27dom96well, I wasn't able to get below 10ms without cheating
20:04:35dom96(10ms max pause time)
20:04:43dom96http://forum.nim-lang.org/t/2646
20:04:55dom96So Go beats us by 2ms
20:05:10dom96Just imagine how much we could accomplish if we had Google behind us ;)
20:06:53*subsetpark joined #nim
20:12:10*Trustable joined #nim
20:18:18*rolha quit (Quit: My Mac has gone to sleep. ZZZzzz…)
20:19:19*r4vi joined #nim
20:20:10*ibk joined #nim
20:27:48Araqwhat? 10ms? in my compiler tests I got it to 1ms pause times, or 0.5ms
20:30:45Araq"On my machine the typical worst push time is 10ms, both for my version of the code and yours."
20:30:50Araqwhat does that even mean?
20:31:20dom96Araq: Which part do you not understand?
20:31:23dom96Try running the code
20:32:03Araqbecause 10ms is the worst time you use the worst case all the time?
20:32:36*Araq sighs.
20:33:11Araqthis code is ridiculous to tune for Nim, I bet I can get it to 0.01ms
20:33:53Araqah, I misread the code :P
20:34:01Araqbut still. let me give it a shot.
20:34:41dom96please do
20:39:13*Matthias247 joined #nim
20:41:52Araq.\gc_latency.exe
20:41:54AraqWorst push time: 0.50ms
20:42:58*nsf joined #nim
20:43:01Araqcannot get it lower since epochTime()'s resolution sucks
20:43:46Araqmaybe def-'s benchmarking module helps here?
20:50:19*rolha joined #nim
20:52:32dom96how did you get it that low?
20:52:54Araqwasn't hard. I just followed my own docs.
20:53:43Araqnow I'm using better timing to more accurate results
20:55:23Araqinterestingly the value I pass to GC_step hardly matters
20:55:43Araqit's at 0.5ms with your value of 10_000 too.
20:56:33Araqbut I think I can get it down to 0.1ms with some tuning
20:56:51dom96well, can you enlighten me? :P
20:56:57dom96I followed your docs too
20:57:08dom96Seems they need improvement if I can't get it working
20:57:12Xei'd be curious as to how you got those numbers too
20:57:52AraqI need to run this code on a slower CPU.
20:58:23XeAraq: i have a few low powered atoms and a core m7 laying around as well as a few arm boards, how slow do you want?
20:59:42Xealso: what OS are you testing those on, i'd be curious to see if that matters much
20:59:43Araqthe art of benchmarking lies in convincing your readers that your setup is actually realistic.
21:00:45*rolha quit (Quit: My Mac has gone to sleep. ZZZzzz…)
21:02:26Araqhmm nano seconds into ms -- divide by 1 million, right?
21:03:30Xeyeah
21:03:30dom96Araq: can you pleaseeee tell me how you got it that fast?
21:03:34Araqhmm I cannot measure this latency.
21:04:56AraqWorst push time: 22994 nano seconds
21:06:35Araqwill write it on the forum for you guys to try
21:15:38Araqhttp://forum.nim-lang.org/t/2646/1#16347
21:15:55jh32found it. running() is setting p.exitStatus without checking return value of waitpid() - i'll make a PR
21:19:04Araqjh32: nice! :-)
21:20:16dom96Araq: the import "lib..." line should use "$lib"
21:20:27dom96Doesn't compile for me
21:20:35dom96jh32: Thanks for investigating!
21:20:55Araqyeah, somebody should have made a stdlib out of it
21:21:12AraqI'm running it on my mac book now to see if I can get somewhat similar values at all
21:21:27dom96Araq: I just ran it: 1272989 ns
21:21:35dom96i.e. 1.2ms
21:21:51AraqXe to answer your question. got the result on Win 10
21:22:00Araqbah. that's bad.
21:23:09dom96not too bad :)
21:29:05dom96You should test the Go version on your machine
21:29:21Araqyou must not run via 'nim c -r'
21:29:27Araqstart the process separately.
21:29:42Araqthen I get times around 0.6ms
21:29:42dom96I did
21:29:51Araqoh?
21:30:51*libman quit (Quit: Leaving.)
21:31:05*libman joined #nim
21:31:14Araqwell the benchmark is unclear about WCET
21:31:39Araq"latency", does that mean "hard realtime"?
21:34:01dom96WCET?
21:34:05*Trustable quit (Remote host closed the connection)
21:34:15Araq"worst case executing times"
21:34:28Araq*execution
21:36:15XeWorst push time: 1237000nano seconds
21:38:52Araqso 1.2ms
21:39:02Araqseems pretty consistent then.
21:41:43dom96it spikes to 9ms sometimes
21:42:43*Salewski joined #nim
21:43:54SalewskiWorst push time: 130000nano seconds # that is the largest value I can get on Linux, average is about halve that value.
21:45:57Araqgot 168850 nano seconds here
21:45:58SalewskiIt is for the tiny Intel NUC box NUC6i7KYK
21:46:13Araqon OSX, but pretty much an outlier
21:46:50AraqSalewski: that's more what I would have expected :-)
21:47:17Araqmaybe osx is just shitty at scheduling
21:47:29Araqcan we take the minimum?
21:47:40Araqthen a single super low result gets the mark
21:50:09Araqanybody running a realtime OS?
21:50:32dom96Definitely worth converting it to ms
21:50:50dom96Easier to see how fast it is
21:51:14dom96For me, 1.199 is the lowest
21:51:47Araqcan we ensure we don't measure the context switches?
21:52:37*chemist69 quit (Ping timeout: 240 seconds)
21:53:34SalewskiGot a few times: Worst push time: 43000nano seconds # which is 43us or 0.043ms
21:54:07dom96I just got 0.153ms randomly
21:54:21dom96the variance is quite large
21:55:27*chemist69 joined #nim
21:56:17*[ui] quit (Quit: Connection closed for inactivity)
21:57:02*libman quit (Remote host closed the connection)
22:02:43*ryanhowe joined #nim
22:09:26Araqdom96: ok, I'm sure we get >1ms due to a context switch
22:09:45Araqsince it stresses the CPU enough, we usually get a context switch
22:09:59Araqbut rarely we get none and then we get the much better times
22:10:13Araqcan't think of any different explanation.
22:10:38Araqmaybe we should use cpuTime() ?
22:27:17dom96perhaps
22:27:37dom96best to stick with the other benchmarks though
22:27:51dom96the go one uses real time
22:27:57dom96*Golang
22:30:02*ibk quit (Quit: Connection closed for inactivity)
22:30:52*elrood quit (Remote host closed the connection)
22:43:12*ryanhowe quit (Quit: WeeChat 1.4)
22:45:59*Salewski left #nim (#nim)
23:03:14*nsf quit (Quit: WeeChat 1.6)
23:08:31*PMunch joined #nim
23:09:10*PMunch quit (Client Quit)
23:09:34*PMunch joined #nim
23:16:55dom96This reminds me of the issue in our C sources repo regarding trust http://manishearth.github.io/blog/2016/12/02/reflections-on-rusting-trust/
23:46:20*Varriount|Mobile joined #nim
23:47:02*yglukhov_ joined #nim
23:47:29Varriount|Mobiledom96: I'm beginning to loathe that "On trusting trust" article.
23:47:44*yglukhov quit (Read error: Connection reset by peer)
23:48:27Varriount|MobileThe concerns are valid, but unlikely, especially for Nim
23:50:28Varriount|MobileIt's far more likely a poisoned binary of some commonly-sudo'd system utility will make its way into the systems package repository.
23:52:13Araqit's stupid.
23:52:26Araqdo you compile your Linux kernel on a Linux machine?
23:52:32Araqsee the problem?
23:53:56AraqLinux could have been infected so that when it runs GCC it makes GCC insert a backdoor when compiling Linux
23:54:16PMunchI think the concern here is that those are an obvious target. Which coincidentally makes them less of a target because the are more scrutinized. Attack a compiler and you might be able to sneak it past while the users of a very secure, critical system might pour over the source code ten times over.
23:54:33PMunchAnd it only takes one bad binary
23:54:41Araqdo we now run around and scream Linus should reimplement Linux from scratch in order to prevent this?
23:54:47vktecVarriount|Mobile: Which article is this?
23:54:59PMunchhttp://manishearth.github.io/blog/2016/12/02/reflections-on-rusting-trust/
23:55:35vktecThanks
23:55:38PMunchIt might not be very relevant for Nim but it is certainly a fair concern
23:56:39PMunchHow many people work on the Linux kernel? How many people work on Nim, or Rust, or some strange esoteric language which made a DLL that some program trusts?
23:57:18PMunchThe problem here is more of a problem with open source than anything
23:57:22Araqit doesn't matter. Linux is built on Linux machines.
23:57:25PMunchOr rather, trusting anyone
23:57:29Araqmostly.
23:57:57PMunchFor all we known Windows or OSX could compile it into their kernels as well
23:58:17Varriount|MobilePMunch: Don't forget hardware
23:58:57*themagician quit ()
23:59:47PMunchNSA shifting some money around and all off a sudden the Windows, OSX or Linux kernel has something shifty in it. Only difference with OS like Linux is that you can actually see the sources so it's slightly easier to spot if you are looking