<< 24-08-2019 >>
00:12:47*PrimHelios joined #nim
00:18:32*wildtrees quit (Quit: Leaving)
00:29:20*btop quit (Ping timeout: 272 seconds)
01:07:00*btop joined #nim
01:13:17FromDiscord_<treeform> $
01:13:36FromDiscord_<DeltaPHC> I was about to ask if Nim as the concept of zero-sized types, but then I went on the playground and verified that it does indeed, heh
01:13:41FromDiscord_<DeltaPHC> *has
01:22:49FromDiscord_<treeform> why would you need that?
01:27:40*owl_000 joined #nim
01:29:18FromDiscord_<DeltaPHC> Well, I'm unsure yet how useful it is in Nim specifically, but in another language I use it as a type system handle for generics. If you have an object like `StateMachine[T]`, that T can be a zero-size type of, for example, `InitialState`, `SomeOtherState`, etc. This way, you can store a form of state at the type level while keeping the same data.
01:29:18FromDiscord_<DeltaPHC>
01:29:18FromDiscord_<DeltaPHC> Then, you can have procedures that specifically take a `StateMachine[SomeOtherState]`. If the user can provide this type, this means that the data is guaranteed to be in a certain state, and so you can reduce runtime checks that way
01:32:07FromDiscord_<DeltaPHC> It works better in a language like Rust, for the reason that you can have functions there that "consume" (take ownership of) a type where the compiler guarantees that it'll never be used again. I know Nim is headed in that direction, though
01:44:49*ng0_ joined #nim
01:46:08FromDiscord_<DeltaPHC> Something like this: https://play.nim-lang.org/#ix=1Teh
01:46:38*ng0_ quit (Client Quit)
01:47:44*ng0 quit (Ping timeout: 260 seconds)
02:05:05*darithorn joined #nim
02:21:36*vlad1777d joined #nim
02:42:58*btop quit (Ping timeout: 272 seconds)
02:54:40*btop joined #nim
02:54:53*laaron quit (Remote host closed the connection)
03:00:39*NimBot joined #nim
03:05:02*laaron joined #nim
03:31:10shashlick@leorize - trying neovim on ubuntu, v0.3.8, i just get "nimsuggest is not running for this project"
03:36:02*btop quit (Quit: leaving)
03:48:39*darithorn quit (Quit: Leaving)
03:50:27*chemist69 quit (Ping timeout: 264 seconds)
03:52:06*chemist69 joined #nim
04:07:40*skrylar[m] scrolls the backlog and sees the display id one
04:07:50skrylar[m]wikipedias hatred of primary sources has always amused me
04:08:01skrylar[m]"don't listen to people who know what they are talking about" good policy.
04:12:52*vlad1777d quit (Ping timeout: 248 seconds)
04:43:22FromGitter<Obround> How do you forward declare a proc with its return type set to `auto`? This code des not work:
04:47:37*owl_000 quit (Ping timeout: 250 seconds)
04:50:29skrylar[m]how would that even work
04:51:43FromGitter<Obround> How do you forward declare a proc with its return type set to `auto`? This code does not work: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5d60c25f3c1aba311bc03977]
05:15:29*owl_000 joined #nim
05:15:35*owl_000 left #nim (#nim)
05:19:52*nif_ quit (Quit: ...)
05:20:01*nif joined #nim
05:32:38*owl_000 joined #nim
05:53:07*narimiran joined #nim
06:16:55*laaron quit (Remote host closed the connection)
06:18:39*solitudesf joined #nim
06:20:33*dddddd quit (Remote host closed the connection)
06:25:13*laaron joined #nim
06:32:35FromGitter<zacharycarter> in an attempt to wrap C++ code - I'm getting these strange errors
06:32:48FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5d60da10dbf5ae2d42b5eb04]
06:33:14FromGitter<zacharycarter> Not sure where the `Marker_` is coming from
06:33:38FromGitter<zacharycarter> One of the objects I'm wrapping is an interface declared like - ⏎ ```struct __declspec(novtable) AllocatorI```
06:33:54FromGitter<zacharycarter> and then I have another class that inherits from that
06:34:51FromGitter<zacharycarter> `Marker_bx::DefaultAllocator` keeps ending up in my generated C++ code though...
06:36:44FromGitter<awr1> @Obround make an issue, maybe? or use code reordering
06:46:38leorizeshashlick: you have nimsuggest installed, right?
06:55:49*narimiran quit (Remote host closed the connection)
07:00:00*gmpreussner quit (Quit: kthxbye)
07:00:18*nsf joined #nim
07:02:59owl_000how to use nimpretty, in cmd it is not recognizable. on the other hand in nim/bin directory there is no nimpretty.exe
07:03:00FromGitter<gogolxdong> Anyone met the following nimble error iof recent version? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5d60e124be3a8a311ad9d65f]
07:04:48*gmpreussner joined #nim
07:05:05*laaron quit (Quit: ZNC 1.7.1 - https://znc.in)
07:05:35*laaron joined #nim
07:22:46lqdev[m]owl_000: nimpretty does not come installed when you use choosenim, for some reason. you have to compile it manually
07:22:59lqdev[m]same goes for finish.exe
07:30:00FromGitter<zacharycarter> if I use the compile pragma to compile some C++ code - I should need to do any further linking steps to importcpp symbols defined in the C++ source I'm compiling - is that correct?
07:30:04FromGitter<zacharycarter> I should just be able to importcpp them?
07:32:31*gangstacat quit (Quit: Ĝis!)
07:37:40*nif quit (Quit: ...)
07:37:50*nif joined #nim
07:40:41FromGitter<zacharycarter> ah - figured it out, I wasn't compiling the cpp file that was defined in, yay unblocked on this a bitr
07:44:16*terps joined #nim
07:47:07*ng0 joined #nim
08:48:31*aq60 quit (Ping timeout: 260 seconds)
08:49:11*gangstacat joined #nim
08:50:32FromGitter<zacharycarter> sweet - I got bimg working :D
08:52:48FromGitter<zacharycarter> https://gist.github.com/zacharycarter/6481a1e67c4d17b4a6114f18150be8ff
08:53:51FromGitter<zacharycarter> now I should be able to continue on with this terrain example: https://bkaradzic.github.io/bgfx/examples.html#tess
08:54:07*terps quit (Ping timeout: 276 seconds)
08:58:12*Minimisthupper joined #nim
09:01:35*Minimisthupper quit (Client Quit)
09:06:05*Vladar joined #nim
09:09:38*Ven`` joined #nim
09:11:16*krux02 joined #nim
09:28:19*Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
09:28:25*shomodj joined #nim
09:31:34*nsf quit (Quit: WeeChat 2.5)
09:34:05*shomodj quit (Read error: Connection reset by peer)
09:34:35*shomodj joined #nim
09:36:55*terps joined #nim
09:56:31*vegai joined #nim
10:00:50*krux02 quit (Remote host closed the connection)
10:17:10*terps quit (Ping timeout: 246 seconds)
10:25:53*terps joined #nim
10:39:00*terps quit (Ping timeout: 248 seconds)
10:46:14*vegai quit (Disconnected by services)
10:47:53*shomodj quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
10:48:45*shomodj joined #nim
10:55:18*shomodj quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
11:00:38*solitudesf- joined #nim
11:02:19*solitudesf- quit (Client Quit)
11:02:38*solitudesf- joined #nim
11:02:57*solitudesf quit (Ping timeout: 245 seconds)
11:12:33*Trustable joined #nim
11:26:06*Trustable quit (Remote host closed the connection)
11:46:06*nsf joined #nim
11:48:15*owl_000 left #nim ("Leaving")
11:48:40*owl_000 joined #nim
12:06:24*theelous3 joined #nim
13:03:12*laaron quit (Remote host closed the connection)
13:04:45*laaron joined #nim
13:05:36*endragor joined #nim
13:21:54*Ven`` joined #nim
13:21:58*aq60 joined #nim
13:37:41*livcd quit (Remote host closed the connection)
14:14:47*laaron quit (Remote host closed the connection)
14:16:59*nif quit (Quit: ...)
14:17:08*nif joined #nim
14:19:07*endragor quit (Remote host closed the connection)
14:22:20*laaron joined #nim
14:22:43shashlick@leorize yes I was using latest devel
14:25:49*endragor joined #nim
14:29:42Zevvis it in your path, can you just run "nimsuggest" on the command line?
14:34:52*laaron quit (Remote host closed the connection)
14:37:21*vlad1777d joined #nim
14:38:33*laaron joined #nim
14:53:50*owl_000 quit (Ping timeout: 272 seconds)
14:56:28*Ven`` quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:58:02*laaron quit (Quit: ZNC 1.7.1 - https://znc.in)
14:58:39*laaron joined #nim
14:59:48*Ven`` joined #nim
15:03:25*solitudesf- quit (Quit: Leaving)
15:03:44*solitudesf joined #nim
15:06:37*Ven`` quit (Ping timeout: 246 seconds)
15:13:56*laaron quit (Quit: ZNC 1.7.1 - https://znc.in)
15:14:51*laaron joined #nim
15:15:36*laaron quit (Remote host closed the connection)
15:19:47*laaron joined #nim
15:23:06*laaron quit (Client Quit)
15:23:49*laaron joined #nim
15:26:06*nsf quit (Quit: WeeChat 2.5)
15:42:07*endragor quit (Remote host closed the connection)
15:45:27shashlickYa
15:51:04*vlad1777d quit (Remote host closed the connection)
15:53:20*vlad1777d joined #nim
15:53:32*owl_000 joined #nim
15:53:51*owl_000 left #nim (#nim)
15:54:04*owl_000 joined #nim
15:55:55*owl_000 quit (Client Quit)
15:59:04*vlad1777d quit (Ping timeout: 272 seconds)
16:07:13*aq60 quit (Remote host closed the connection)
16:10:02Zevvand do you see anything when checking for :messages?
16:11:04*vlad1777d joined #nim
16:15:47shashlickI'll check it out
16:16:44shashlickSometimes it says nimsuggest exited with 0 on exit but doesn't do anything for the project and gives the message that not running for this project
16:16:54*darithorn joined #nim
16:17:01shashlickWas hoping I could use it at least for Linux
16:20:36shashlickNew nimterop feature in test - https://github.com/nimterop/nimterop/blob/v020/nimterop/build.nim#L498
16:20:41shashlickLooking for feedback
16:21:43shashlickHere's a sample header - https://github.com/nimterop/nimterop/blob/v020/tests/lzma.nim#L11
16:22:18shashlickAnd how to use - https://github.com/nimterop/nimterop/blob/v020/tests/getheader.nims#L23
16:22:29leorizeshashlick: what file did you test it on?
16:23:05shashlickI tried a few different nims and a nim files
16:23:21leorizenimscript doesn't work with nimsuggest
16:28:27shashlickOk I'll check messages with a Nim file and get back
16:28:39shashlickAny feedback on the nimterop feature?
16:29:52*dddddd joined #nim
16:30:59*krux02 joined #nim
16:31:26leorizeit's a nice feature
16:32:00leorizenimterop is on the verge of having a package builder integrated though :P
16:33:09*narimiran joined #nim
16:33:23leorizeI think you can build an object/tuple instead though
16:33:50leorizeinstead of creating `const xxxPath` and `const xxxLDir`
16:34:18leorizealso, are you sure that static params are safe to use now?
16:34:45leorizethe last time I used them they got corrupted at a random position within the macro
16:42:50shashlickThey have worked fine so far, but I'm still debugging to get it working on Travis across Nim versions
16:42:59shashlickStill some ways from osx as well
16:43:09shashlickBut probably not related to static
16:43:46shashlickI'm calling builders, is that a builder too? :)
16:43:59shashlickIt detects the builder and deals with it
16:44:27leorizewell that's pretty much how people packages software :P
16:44:53shashlickNot sure I get your object / tuple feedback
16:45:12leorizehttps://github.com/nimterop/nimterop/blob/2c1d8d3bd8e25e331cfd58e17354cd1dc306e5a4/nimterop/build.nim#L518-L520
16:45:23shashlickYou need the const to be used in the following cimport call
16:45:37shashlickConsidering it is a macro, I cannot return values
16:45:52leorizeyou can?
16:46:14leorizewell I'm pretty sure Zevv's managed to do that
16:46:42shashlickI'm creating defines with the header name so need to generate ast for that
16:47:03shashlickAnd that means I have to return a nimnode
16:47:48shashlickSee line 557
16:49:53shashlickAny feedback on the api signature or feature in general
16:50:37leorizethe feature is great
16:50:58leorizepersonally I'd like it if you can generate {.compile.} calls :P
16:52:59leorizeshashlick: since everything runs at compile-time, do you really need to generate code that do "when defined" checks?
17:03:35shashlickWell how do I check for defines without hard coding
17:04:28shashlickThe nimterop code has to be generic but I don't want the wrapper writer to have to specify too much
17:06:08shashlickOnly way I could generate the defines to check was this way
17:06:39shashlickIn fact, this way, the wrapper can support a variety of ways but end user can pick a specific one with defines
17:06:39*krux02 quit (Ping timeout: 264 seconds)
17:06:56shashlickBut that's why I'm asking cause I wonder if there's a better or simpler way
17:08:32leorizeI'm pretty sure defined() takes an indent nimnode
17:09:01leorizeident*
17:10:07leorizehmm, it doesn't :/
17:11:58*elrood joined #nim
17:14:06shashlickI prefer quote since it is readable
17:14:56ZevvI didnt return values but stored them in .compiletime. vars to save state over macro calls
17:15:22Zevvthat took me three weeks because I did not know of the existence of .compileTime. vars :/
17:15:51shashlickSo I can make them var instead of const
17:17:54ZevvI assumed that - because normal procs just work at compile time also - normal vars should work at compile time as well. but it seems they dont unless you mark them so epxlicitly
17:18:15Zevvwhich in hindsight kind of makes sense, but was not really obvious to me
17:19:01Zevvso make them var {.compileTime.}
17:38:09shashlick@leorize - regargin {.compile.} calls
17:38:11shashlickregarding
17:38:24shashlickthat's what i was doing for all other wrappers i've worked on
17:38:38shashlickand it was hard and tedious but was cool to make a static binary with no dependencies
17:38:45shashlickbut the effort does not scale
17:39:02shashlickand with nimarchive, i hit a wall where it compiled but crashed
17:39:11shashlickand i didn't see the crash with a static link to a .a file
17:42:57*laaron quit (Remote host closed the connection)
17:49:57*laaron joined #nim
18:06:57*livcd joined #nim
18:07:23*krux02 joined #nim
18:21:36*laaron quit (Quit: ZNC 1.7.1 - https://znc.in)
18:22:15*laaron joined #nim
18:35:54*elrood quit (Remote host closed the connection)
18:48:33*nsf joined #nim
19:08:54*snooptek quit (Remote host closed the connection)
19:09:18*Vladar quit (Remote host closed the connection)
19:18:55*actuallybatman joined #nim
19:23:22*endragor joined #nim
19:25:14*wildtrees joined #nim
19:52:48Cadeydom96: i'm trying to make a jester app on a custom port, but i'm getting this error: ... /home/cadey/code/wat/src/wat.nim(175, 3) Error: 'port' cannot be assigned to
19:52:51Cadeyuploading code for you, sec
19:53:30Cadeyhttps://github.com/Xe/wat/blob/master/src/wat.nim#L174-L175
19:54:55dom96try putting it above `routes`
19:56:35Cadeysame error dom96
19:57:34Cadeyusing jester 0.4.3
19:58:16*endragor quit (Remote host closed the connection)
19:58:41dom96Cadey, can you check if this works? https://github.com/dom96/jester/blob/master/tests/alltest.nim#L17
19:59:40Cadeynope
19:59:44Cadeyit doesn't work
20:01:45Cadeyport = Port(5939)
20:01:50Cadeystill get the same error
20:01:56Cadeyshould this before routers too?
20:02:26Cadeyokay it works before the routers
20:03:02Cadeyi don't know if this is a bug in jester or not
20:03:54dom96routers creates its own settings block if it doesn't detect one AFAIK
20:04:01dom96and it probably generates it as a `let`
20:04:49Cadeyah
20:05:04CadeyAraq: how would i use ormin across multiple nim modules?
20:05:22Cadeydom96: it works now, thanks
20:05:38dom96Cadey, please make an issue for this
20:05:43Cadeywull do
20:05:46Cadeyi'm working on the jester stuff while making a fairly complicated exmaple
20:09:03Cadeydom96: please take a look: https://github.com/dom96/jester/issues/209
20:09:22dom96looks good, thanks
20:10:34Cadeyno problem
20:23:13FromGitter<Varriount> Well... This is certainly interesting: https://www.reddit.com/r/programming/comments/cus0zu/a_3mil_downloads_per_month_javascript_library/
20:29:20FromGitter<awr1> has anyone ever asked if we could get the `not nil` annotation on `distinct pointer` types?
20:30:21FromGitter<mratsim> not nil is a bit tricky to work with ... there is always one step where the compiler cannot prove not nil ...
20:30:22FromGitter<awr1> i've got a library i'm binding too with a lot of opaque handles so I'm making them `distinct pointer` to avoid any wrongful type messiness
20:31:08FromGitter<awr1> but `not nil` doesn't work on distinct so I'm wondering if I should just make them `ptr` to some...hidden type thing?
20:31:20FromGitter<awr1> *binding to
20:32:07*narimiran quit (Ping timeout: 246 seconds)
20:32:11FromGitter<awr1> alternatively I could use `Option`, but `Option` probably doesn't do the nullptr optimization for distinct pointers
20:34:05FromGitter<mratsim> oh
20:34:38FromGitter<mratsim> I think not nil should work on distinct and Option should also do the null pointer optimisation for distinct pointer
20:35:37FromGitter<awr1> should I make an issue I guess?
20:36:33FromGitter<mratsim> Yes, note it might probably be impossible to implement without reworking distinct type vs bind many due to reusing the same keyword for 2 different concepts: https://github.com/nim-lang/RFCs/issues/153
20:36:40FromGitter<mratsim> note that*
20:37:15FromGitter<awr1> the `Option` thing can probably be easily fixed, I think @timotheecour or someone else made that `distinctBase()` macro (or something along the lines of that) that can be used to acquire the root type
20:37:48Araqtype Foo = ptr object
20:38:02Araqis how you wrap opaque pointers, not with 'distinct'
20:38:40FromGitter<awr1> hmm! didn't know you could do that
20:38:40FromGitter<awr1> thanks
20:41:17FromGitter<mratsim> btw, regarding https://github.com/nim-lang/Nim/pull/11816, thoughts on adding the enhancements in different phases?
20:41:49FromGitter<awr1> oh crap i need to work on that lol. i've been busy as of recent
20:42:13FromGitter<awr1> yeah that's a good idea
20:43:50FromGitter<mratsim> just the x86 stuff would help me not having to submodule cpuinfo in all my low-level repo :P https://github.com/mratsim/weave/tree/master/benchmarks
20:44:09FromGitter<awr1> you probably want the x86 features soon. ARM is trickier because `/proc/cpuinfo` needs to be probed, which doesn't have a perfect Windows analogue (i.e. IsProcessorFeaturePresent())
20:44:58Cadeydom96: should before filters in subroutes be bubbling up to the parent?
20:45:10Cadeysubrouters*
20:45:29FromGitter<awr1> and I mention Windows on ARM since supposedly WIndows on ARM laptops are potentially going to be a thing in the next couple of years
20:45:29FromGitter<mratsim> I think Windows ARM is not really a priority :P
20:45:37FromGitter<awr1> not a priority right now, i agree
20:45:45FromGitter<mratsim> Windows Phones were on ARM?
20:45:49FromGitter<awr1> but it could in the future
20:45:51FromGitter<awr1> they were
20:46:17FromGitter<awr1> microsoft did this new windows 10 on ARM thing where it's not the cut down store-only .NET stuff
20:46:35FromGitter<mratsim> Facebook's CPUInfo only considers Android, Iphone, Linux: https://github.com/pytorch/cpuinfo/tree/master/src/arm
20:46:46FromGitter<awr1> and it is full native Win10, plus i386 emulation built in WoW64
20:47:29FromGitter<awr1> i mean it's not relevant right now, but the future could be weird
20:47:35FromGitter<awr1> idk
20:48:27AraqWindows on ARM is old stuff
20:49:19Cadeyhttps://gist.github.com/Xe/8f5ca6baa68186247433a7fcfff71b50 dom96 like this
20:49:20FromGitter<awr1> @Araq not anymore https://mspoweruser.com/at-least-4-snapdragon-8cx-windows-10-on-arm-laptops-coming-to-europe-next-year/
20:50:17FromGitter<awr1> microsoft scrapped the old windows sandboxed .NET-only ARM model from a few years back and came out with this rather recently
20:50:38FromGitter<mratsim> I expect they will introduce some hardware info abstraction
20:51:04FromGitter<mratsim> parsing proc/cpuinfo is a bit meh
20:51:31FromGitter<awr1> and considering apple wants to move their laptops to ARM too, the industry being shaked up by this is certainly a possibility
20:51:37FromGitter<mratsim> 1) ideally you don't want to allocate for such a low-level feature and working with strings is not ideal
20:51:44FromGitter<mratsim> +
20:52:08FromGitter<awr1> i think there is a way you can get cpu features on linux without proc cpuinfo
20:53:55Araqiofnctl
20:54:21FromGitter<awr1> ARM makes this annoying because all their "cpu feature" stuff is hidden behind supervisor-level instructions
20:56:02dom96Cadey, that's a good question
20:56:14Cadeyi found it surprising personally
20:56:17dom96I think they should only apply for /api
20:56:24Cadeyi'll go file another bug
20:56:52FromGitter<awr1> i'll see if I can get the cpuinfo x86 features ready sometime today or tomorrow. i don't know if i can do tests for it since its hardware sensitive
20:56:58*shomodj joined #nim
20:57:33FromGitter<mratsim> I also had a look into hwloc from the OpenMPI project (https://github.com/open-mpi/hwloc/tree/master/hwloc) because it's the only one that supports NUMA and also detect hyperthreading siblings (which are very annoying in high-performance computing)
20:57:41FromGitter<mratsim> but it's quite complex
20:57:53Cadeyhttps://github.com/dom96/jester/issues/210
20:58:25*shomodj quit (Client Quit)
20:58:32FromGitter<awr1> you can check for HTT/thread topology with cpuid in x86, but i think the "preferred way" is to just use the OS
20:59:09FromGitter<awr1> since the OS can turn off HW threads in certain configurations if it really wants to
21:00:39*al_ joined #nim
21:00:54*shomodj joined #nim
21:01:20FromGitter<mratsim> AFAIK it's a mess with cpuid :p Intel had various changes between pentium 4, Nehalem and Sandy bridge+
21:02:36FromGitter<mratsim> but my main need is knowing on a 4 cores machine if cores {0, 1} and {2, 3} are siblings or if its cores {0, 2} and {1, 3}
21:03:06Araqwrite some startup logic to detect it at runtime :P
21:03:47FromGitter<awr1> i remember a video IBM had demonstrating SMT performance on POWER on some linux system and they swtiched from 2 SMT threads per core to 8 SMT threads to show off the performance increase, so i assume that's just a feature that OSs can do apparently. although exactly *why* one would need to turn off hardware threads while the system is up is beyond me
21:04:27FromGitter<mratsim> Hyperthreading means that the logical cores share L1/L2 cache, so with HT those are 2x smaller on a workload that can make use of all the cache. It also means that memory bandwidth is shared. And obviously HPC workloads are almost all memory bandwidth bound ...
21:04:55FromGitter<mratsim> Xeon Phi have 4 Hyperthread siblings per physical core
21:05:44Cadeymratsim: this smells like premature optimization
21:05:58FromGitter<mratsim> It's not
21:06:36FromGitter<mratsim> It's state-of-the art optimization and making regular Nim code faster than C/FOrtran/C++
21:06:51FromGitter<mratsim> and assembly
21:07:51FromGitter<mratsim> You can read the papers and the comments interleave with my code here: https://github.com/numforge/laser/blob/master/laser/primitives/matrix_multiplication/gemm_tiling.nim#L12-L30
21:08:13FromGitter<awr1> are "hot-pluggable CPUs" a thing in the HPC space?
21:08:23FromGitter<mratsim> I need to take into account L1 cache, L2 cache, TLB sizes and number of registers to get the max performance
21:08:52Cadeydom96: based on what i'm reading, couldn't this be fixed by having the before and after filters respect the prefix in the extend command handler? https://github.com/dom96/jester/blob/master/jester.nim#L1076-L1107
21:09:51FromGitter<mratsim> @awr1, not sure, I didn't try "true HPC" with a cluster of thousands of CPUs :P
21:10:21FromGitter<mratsim> I know that in CERN you reserve a timeslice and some CPU so you basically have a VM
21:20:50*endragor joined #nim
21:26:18*endragor quit (Ping timeout: 245 seconds)
21:37:51*al_ quit (Quit: al_)
21:49:49*solitudesf quit (Ping timeout: 246 seconds)
22:00:03*actuallybatman quit (Ping timeout: 245 seconds)
22:07:05*actuallybatman joined #nim
22:07:36*actuallybatman quit (Client Quit)
22:13:31*nsf quit (Quit: WeeChat 2.5)
22:34:48*actuallybatman joined #nim
23:03:57*krux02 quit (Remote host closed the connection)
23:04:57*actuallybatman quit (Remote host closed the connection)
23:10:22*actuallybatman joined #nim
23:19:48*shomodj quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
23:20:30*couven92 quit (Read error: Connection reset by peer)
23:21:37*couven92 joined #nim
23:37:27*couven92 quit (Read error: Connection reset by peer)
23:38:34*couven92 joined #nim
23:51:30*couven92 quit (Read error: Connection reset by peer)
23:52:29*couven92 joined #nim