<< 30-03-2020 >>
00:01:45*Trustable quit (Remote host closed the connection)
00:03:33*dadada__ joined #nim
00:15:23*couven92 quit (Ping timeout: 260 seconds)
00:26:47*dadada__ quit (Ping timeout: 252 seconds)
00:33:32*dadada joined #nim
00:33:55*dadada is now known as Guest27287
00:56:32*rnrwashere joined #nim
00:57:14*Guest27287 quit (Ping timeout: 246 seconds)
00:59:36FromGitter<sealmove> How to do int -> enum?
01:01:50FromDiscord<Rika> intVar.Enum perhaps?
01:02:38*dwdv quit (Ping timeout: 256 seconds)
01:03:44*dadada__ joined #nim
01:05:15FromGitter<sealmove> almost... you have to do Enum(intVar), that's why I was confused
01:06:07FromDiscord<Rika> you dont????
01:06:21FromDiscord<Rika> UFCS????
01:06:38FromGitter<sealmove> for some reason UFCS doesn't work in this case
01:07:01Yardanicoif it's like 5+3.enum then of course
01:07:06FromGitter<sealmove> or perhaps I used it wrongly, because of complex expression
01:07:07Yardanicoyou need to do (5+3).enum
01:07:16FromGitter<sealmove> yeah, that was my pitfall probably
01:09:59*rnrwashere quit (Remote host closed the connection)
01:20:34*lritter quit (Quit: Leaving)
01:22:07*rnrwashere joined #nim
01:27:01*rnrwashere quit ()
01:27:48*dadada__ quit (Ping timeout: 256 seconds)
01:28:26*chemist69 quit (Ping timeout: 240 seconds)
01:30:54*chemist69 joined #nim
01:33:34*dadada joined #nim
01:33:58*dadada is now known as Guest23452
01:35:49*flaviu quit (Ping timeout: 258 seconds)
01:56:32*endragor joined #nim
01:57:19*Guest23452 quit (Ping timeout: 250 seconds)
02:03:31*dadada__ joined #nim
02:08:35*endragor quit (Remote host closed the connection)
02:14:17*endragor joined #nim
02:18:37*endragor quit (Ping timeout: 252 seconds)
02:24:11*rnrwashere joined #nim
02:27:59*dadada__ quit (Ping timeout: 265 seconds)
02:33:29*dadada joined #nim
02:33:52*dadada is now known as Guest91738
02:41:05*muffindrake quit (Ping timeout: 240 seconds)
02:43:28*muffindrake joined #nim
02:57:14*Guest91738 quit (Ping timeout: 240 seconds)
03:02:52*rnrwashere quit (Remote host closed the connection)
03:03:33*dadada__ joined #nim
03:17:48FromDiscord<Rika> is it possible to "dynamically" generate test cases in
03:17:51FromDiscord<Rika> in unittest*
03:23:29FromGitter<sealmove> what do you mean by "dymanically"?
03:23:42*endragor joined #nim
03:27:14*dadada__ quit (Ping timeout: 240 seconds)
03:33:32*dadada joined #nim
03:33:55*dadada is now known as Guest3872
03:38:56*sz0 quit (Quit: Connection closed for inactivity)
03:39:44*waleee-cl quit (Quit: Connection closed for inactivity)
03:53:21*silvernode joined #nim
03:57:24*Guest3872 quit (Ping timeout: 265 seconds)
03:58:43*chemist69 quit (Ping timeout: 252 seconds)
03:59:40*chemist69 joined #nim
04:00:22FromDiscord<Rika> uh
04:00:55FromDiscord<Rika> converting a for loop of a file walk into test cases per file
04:00:59FromDiscord<Rika> or something like that
04:03:28leorizeuse testament
04:03:30leorizeor
04:03:33leorize!repo testutils
04:03:34*dadada__ joined #nim
04:03:34disbothttps://github.com/status-im/nim-testutils -- 9nim-testutils: 11testrunner et al 15 6⭐ 2🍴
04:04:09FromDiscord<Rika> are there docs for testament?...
04:04:36leorizein some contributing documents iirc
04:04:38FromDiscord<Benumbed> "docs"
04:05:08FromDiscord<Rika> """docs"""
04:07:38FromDiscord<Rika> do you recommend testament or testutils
04:07:47FromDiscord<Benumbed> The main thing I don't like about Status' test runner is that the files have to be named .test
04:08:00FromDiscord<Benumbed> Like y tho
04:08:00*silvernode quit (Ping timeout: 256 seconds)
04:08:22FromDiscord<Rika> ok so testament it is
04:08:36FromDiscord<Benumbed> Orrr write your own 😉
04:09:32FromDiscord<Rika> lolno
04:09:59leorizetestament is not too hard to use tbh
04:09:59disruptekuse testament. i will probably fix it to run async tests before i return to working on testutils.
04:12:52shashlickdisruptek: how are things in nimph land
04:21:03*opal quit (Ping timeout: 240 seconds)
04:25:21*sleepyqt joined #nim
04:27:26*dadada__ quit (Ping timeout: 256 seconds)
04:31:37disruptekshashlick: i dunno, i guess there might be a bug in how it handles choosenim because maybe choosenim puts packages under $nim, which nimph mistakes for compiler-supplied imports like, say, compiler.
04:32:03disruptekkungtotte has some sort of problem with mustache and i haven't had time to look at it properly.
04:32:54disruptekany idea where my extra output came from?
04:33:32*dadada joined #nim
04:33:55*dadada is now known as Guest2015
04:36:52*rockcavera quit (Remote host closed the connection)
04:46:48*opal joined #nim
04:54:02*leorize quit (Remote host closed the connection)
04:54:02*opal quit (Remote host closed the connection)
04:54:15*opal joined #nim
04:55:27*leorize joined #nim
04:57:23*Guest2015 quit (Ping timeout: 252 seconds)
05:03:41*dadada__ joined #nim
05:09:13*thomasross quit (Ping timeout: 264 seconds)
05:16:31FromGitter<sealmove> Rika I had similar needs and made https://github.com/sealmove/testify
05:18:45leorizedo you feel like making a tool that output junit from testament test results
05:21:10FromGitter<sealmove> would be awesome
05:22:06FromGitter<sealmove> but testament is hard to understand for people who have done no work on nim compiler
05:22:14FromGitter<sealmove> i don't understand it at all
05:27:47*dadada__ quit (Ping timeout: 265 seconds)
05:32:17*silvernode joined #nim
05:33:32*dadada joined #nim
05:33:55*dadada is now known as Guest2810
05:41:06*Asgaroth_ quit (Read error: Connection reset by peer)
05:42:08*Asgaroth_ joined #nim
05:43:59FromDiscord<Rika> i think i understand it now
05:47:04FromDiscord<Rika> okay, so i'm pretty dumb
05:47:15FromDiscord<Rika> i could actually do what i described in unittest with a template
05:47:41FromDiscord<Rika> is this what being in the house for weeks does to you
05:51:43FromGitter<sealmove> unittest's main limitation is that it only works with 1 file. there is no process isolation or anything.
05:51:56*silvernode quit (Ping timeout: 246 seconds)
05:56:03*opal quit (Ping timeout: 240 seconds)
05:57:09*Guest2810 quit (Ping timeout: 252 seconds)
05:58:14FromDiscord<Rika> i dont think i have an issue with that
05:59:00*narimiran joined #nim
05:59:22*opal joined #nim
05:59:32*crem quit (Ping timeout: 260 seconds)
05:59:49*crem joined #nim
06:03:38*dadada__ joined #nim
06:04:59FromGitter<sealmove> then unittest should be perfect :>
06:09:39*solitudesf joined #nim
06:27:13*dadada__ quit (Ping timeout: 252 seconds)
06:33:35*dadada joined #nim
06:33:59*dadada is now known as Guest18345
06:52:48*tefter joined #nim
06:57:26*Guest18345 quit (Ping timeout: 256 seconds)
07:00:00*gmpreussner quit (Quit: kthxbye)
07:03:38*dadada__ joined #nim
07:04:35*gmpreussner joined #nim
07:12:03*PMunch joined #nim
07:13:14axionOk I can sleep now. My first piece of nim code is up on nimble.directory. :)
07:15:41PMunchLink?
07:16:25axionhttps://nimble.directory/pkg/origin I'll read the newbie code hate when I wake up
07:27:05*dadada__ quit (Ping timeout: 250 seconds)
07:29:40leorize[m]axion: it's full of lisp code?
07:31:00axionleorize[m]: I sent a PR. It was a typo for the web link. you can click the actual repo link to go to the correct one.
07:31:38leorizelol
07:32:20axionI am tired and should have done this when I'm more awake. :(
07:33:42*dadada joined #nim
07:34:05*dadada is now known as Guest86397
07:57:25*Guest86397 quit (Ping timeout: 252 seconds)
08:03:33*dadada__ joined #nim
08:12:07*Asgaroth_ quit (Read error: Connection reset by peer)
08:12:32*Asgaroth_ joined #nim
08:20:01AraqI have a minor patch for testament in the pipeline
08:22:51*couven92 joined #nim
08:53:24*silvernode joined #nim
09:08:17*silvernode quit (Ping timeout: 246 seconds)
09:13:39*NimBot joined #nim
09:16:44*dwdv joined #nim
09:18:14Araqmost C libraries use pointers for everything and the sizes are irrelevant
09:19:13AraqI don't see how a "dummy C file" helps with anything
09:20:02*kenran joined #nim
09:26:51shashlickwhich lib is this
09:29:03PMunchIt's Unbound
09:29:26PMunchNot really a library, but I create a DLL that it interfaces with
09:29:45PMunchSo I get structures that are defined in the program
09:30:29PMunchAraq, well Nim is probably smarter about the sizes if I define the objects as proper C structs and tell Nim to use that rather than telling Nim the sizes of things.
09:31:57Araqnarimiran, I don't understand https://github.com/nim-lang/Nim/pull/13803/files
09:31:58disbot➥ faster CIs
09:37:01narimiranAraq: i've just answered your question about nightlies. is there something else?
09:37:23shashlickPMunch: we already spoke about unbound before 🙂 https://irclogs.nim-lang.org/21-10-2019.html#13:46:03
09:37:56PMunchYeah I know :P
09:38:27*uvegbot joined #nim
09:38:56Araqnarimiran, which tests were removed?
09:39:03Araqwhat was the criterion?
09:39:11shashlicknarimiran: only comment on csources, we use it in nightlies to build csources into release
09:40:04Araqnightlies should really use 'koch nightlies' (which doesn't exist yet)
09:40:11PMunchshashlick, this is pretty much the stuff I have now: http://ix.io/2fYu
09:40:14PMunchAnd it's not great..
09:40:24shashlicknarimiran: okay saw Araq's comment and your response
09:40:52shashlicknightlies does all sorts of docker stuff, some portions could go into koch though, yes
09:41:02shashlickleorize was looking into github actions tho
09:41:12narimiranshashlick: yeah, i've spoke to him about it
09:41:41shashlickcool
09:43:35narimiranAraq: in most of the tests i just lowered the number of iterations and/or slightly reduced sleep time. after that, some tests were merged into one larger test. and the lowest amount of commits are test removals as follows:
09:44:50narimiran- in 'assert' category, `testhelper.nim` file was removed and the tests that test that file; `tassert2.nim` was modified to work without that helper
09:46:15narimiran- `tindexerrorformatbounds.nim` was removed as i felt it was not very useful
09:46:56Araqwhy was tests/parallel/tdisjoint_slice2.nim removed?
09:47:02*filcuc joined #nim
09:47:19Araqah merged into the first test, got it
09:47:19narimiranit was merged into `tdisjoint_slice1.nim`
09:47:20Araqok
09:47:25*dadada__ quit (Ping timeout: 252 seconds)
09:47:36*ganytheodes joined #nim
09:48:10ganytheodesHi guys how can I do something like if foo() and bar(): echo "both ok?"
09:48:16narimiransame for tbitops2, tdeepcopy2, dcollections_to_string, tstreams2, tstreams3
09:48:39*dadada joined #nim
09:48:45narimiranganytheodes: like you have written it if those procs return bool
09:49:01*dadada is now known as Guest43612
09:49:25ganytheodesboth return bool but it does not seem to work
09:49:44narimiranganytheodes: can you post your code on https://play.nim-lang.org/ ?
09:50:20shashlickPMunch: check out http://ix.io/2fYw
09:50:40shashlicksome types are stubbed out but might work for you
09:50:57narimiranAraq: 'rodfiles' category was removed as i thought "rod" stuff is long gone - if not true, i'll revert that commit
09:51:25ganytheodesnarimiran: i am just checking paths so not sure what's there to post
09:51:38narimiran...but the tests there seem to be very basic and (probably) covered by other tests
09:53:31PMunchshashlick, what's that?
09:54:03shashlickWrapper for module.h
09:54:52Araqnarimiran, nah the rest is fine, the question really was about tdisjoint_slice
10:03:24*krux02 joined #nim
10:05:50*Guest43612 is now known as dadada
10:16:01PMunchHmm
10:16:28PMunchI won't be able to access any of the fields though will I?
10:17:58AraqPMunch, your wrapping code looks fine to me
10:19:39PMunchYeah, but I struggle with it crashing
10:19:46PMunchSIGSEGV
10:20:06PMunchTraced it back earlier to an issue with alignment
10:20:32Araqwe use C's default alignment rules
10:20:52Araqwrap all the fields inside the struct in the right order and there should be few problems
10:21:04PMunchAnd yet there is..
10:23:20AraqPMunch, well you can import from the .header
10:23:40Araqin fact, I think you're confused
10:23:41Araq{.emit: fmt"""
10:23:41Araq#include "{here}/config.h"
10:23:41Araq#include "{here}/../util/module.h"
10:23:41Araq"""}
10:23:58Araqmeans that you really want to use .header instead
10:29:36*ganytheodes quit (Remote host closed the connection)
10:30:09PMunchHmm, just tried to remove that and it builds fine
10:30:17PMunchSo that's probably left over from something
10:38:43*Trustable joined #nim
10:39:54FromGitter<alehander92> morning
10:42:25*PMunch quit (Read error: Connection reset by peer)
10:45:00*liblq-dev joined #nim
10:45:43*PMunch joined #nim
10:47:45*solitudesf quit (Remote host closed the connection)
10:49:28*solitudesf joined #nim
10:57:54*abm joined #nim
11:01:51*natrys joined #nim
11:09:33*sz0 joined #nim
11:10:22*kenran quit (Quit: leaving)
11:17:11*Vladar joined #nim
11:22:11FromDiscord<clyybber> morning
11:22:35FromDiscord<Recruit_main707> what time is it there?
11:22:50Araqalehander92: https://github.com/nim-lang/Nim/pull/13743/
11:22:51disbot➥ DrNim (Nim compiler with Z3 integration)
11:23:14Araqmost features have been implemented, now "only" the testing needs to be done
11:23:37*filcuc quit (Ping timeout: 264 seconds)
11:32:17*uvegbot quit (Ping timeout: 252 seconds)
11:35:01*uvegbot joined #nim
11:35:07*clyybber joined #nim
11:40:22FromGitter<timotheecour> @araq any other changes need for stacktrace PR https://github.com/nim-lang/Nim/pull/13351 ? I’d really like to have it merged because it helps when debugging code for other PRS
11:40:24disbot➥ stacktraces can now show custom runtime msgs per frame ; snippet at 12https://play.nim-lang.org/#ix=2e0u
11:45:58Araqok and thank you for pinging me on open PRs
11:46:27clyybberhey, we got nim music: https://www.youtube.com/watch?v=rJwnWKFb-m8
11:46:27FromGitter<timotheecour> yay! thanks
11:46:34Araqit remains to be the case that I don't get every update
11:46:39Araqnot sure why
11:49:21clyybbertimotheecour: Why are those required in that PR? https://github.com/nim-lang/Nim/pull/13351/files#diff-5cfdfeab79c2d2b7bbc78af7060c38c3R58-R73
11:49:23disbot➥ stacktraces can now show custom runtime msgs per frame ; snippet at 12https://play.nim-lang.org/#ix=2e0u
11:49:41clyybberTo be added to system.nim that is
11:50:05FromDiscord<Rika> can i use `when` instead of `case` to make a variant object?
11:51:48Araqclyybber, he only moved the decls around
11:51:53clyybberoh
11:51:57clyybberI see
11:52:43*tane joined #nim
11:54:05*solitudesf- joined #nim
11:56:17*solitudesf quit (Ping timeout: 246 seconds)
12:03:00FromGitter<timotheecour> ya
12:05:02FromGitter<alehander92> Araq
12:05:06FromGitter<alehander92> this is good
12:05:13FromGitter<alehander92> i will like to check it out
12:05:47FromGitter<alehander92> two questions: cant it be a flag to the compiler: or is it too much to bundle optionally z3 with it (maybe the flag can just invoke the drnim binary)
12:05:58FromGitter<alehander92> and how do you plan to solve nil checking with it
12:06:20FromGitter<alehander92> also, i want to make a simple rfc about tracking a single blocking effect, but i want to ask first here
12:06:40FromGitter<alehander92> for some clarifications (e.g. can i track just a single tag/effect? without having to annotate the other ones)
12:07:46Araqit's ok as a separate tool IMHO, the dependency is bad for many
12:08:09Araqnil checking doesn't really benefit from it though and shouldn't be part of my todo
12:08:45FromGitter<geotre> How do you run a macro on the output of another macro?
12:09:25AraqgetAst(macroCall(...)) iirc
12:09:43Araqit happens rarely, usually macros only call into .compiletime proc helpers
12:09:45FromGitter<alehander92> Araq ah ok
12:10:01Araqbut we got .requires and .ensures
12:10:12FromGitter<alehander92> but the setLen thing
12:10:15FromGitter<alehander92> seems pretty nice
12:10:19FromGitter<alehander92> overally good work
12:10:31FromGitter<alehander92> does it have a lot of overhead?
12:10:46FromGitter<geotre> Thanks @Araq!
12:10:47Araqno idea, for now it's opt-in via .push staticBoundChecks
12:11:05FromGitter<alehander92> and do you guys plan to annotate the stdlib eventually, as one needs that to be able to use the feature well imo
12:11:30Araqthat's why the pragmas are builtin, they should always exist
12:11:42FromGitter<alehander92> (yeah one can just assert a variable >= n for a cast so maybe not too fatal)
12:12:33Araqbut the stdlib isn't as crucial to annotate as you would think because currently there is little available
12:12:53FromGitter<alehander92> hm, and can one annotate e.g. requires: a > 0, other > a
12:13:00FromGitter<alehander92> like args depending on each other
12:13:00Araqfor instance, you cannot write down that e.g. socket.read works on opened sockets
12:13:22FromGitter<alehander92> ah something like typestate
12:13:37Araqmy goal was to solve range and index checking and so far it supports that really well
12:13:42*rockcavera joined #nim
12:13:47FromGitter<alehander92> we imagined a mechanism similar to a generalized version of notnil checking for that once (with zah i think)
12:14:03FromGitter<alehander92> but it was just a dream :D
12:14:06Araqonce that matured, I will extend it to Nim's "disjoint checking" for threads
12:14:41FromGitter<alehander92> i havent used that i think
12:15:02*filcuc joined #nim
12:15:30FromGitter<alehander92> but the typestate thing: how would you write down expected transitions (e.g. pre: Open, post: Running)
12:15:43FromGitter<alehander92> would they be based on enums
12:16:38Araqhaven't thought about it but {.requires: f.state == opened.} comes to mind
12:16:59FromGitter<Clyybber> when you have typestates depend on other typestates
12:17:00Araqproc open(): File {.ensures: f.state == opened.}
12:17:09FromGitter<Clyybber> we have type attached proofs
12:17:18Araqproc close(f: File) {.ensures: f.state == closed.}
12:17:36FromGitter<alehander92> ha yeah, you might want to ensure the state is just previous toggle of a state or something
12:17:39Araqpretty easy to do once all the other machinery is in place
12:17:46FromGitter<alehander92> and probably more complicated case
12:18:01clyybberATS does it wonderfully
12:18:15FromGitter<alehander92> but thats what i wondered before: can we write a > 0, other > a
12:18:24Araqtry it.
12:18:32FromGitter<alehander92> :D ok, later
12:19:09FromGitter<alehander92> otherwise, my problem is : i want to annotate `a()` with `Blocking` (but you can imagine it for `IO`): can i somehow check only for that effect without writing down all the others
12:19:35FromGitter<alehander92> like, having a separate "group" for each effect independently of each other
12:19:53FromGitter<alehander92> the same way i can track *types* but not *raises* etc
12:20:19FromGitter<alehander92> i imagined that maybe thats what tags are
12:20:25FromGitter<alehander92> but it seems tags and effects are the same
12:20:43Araqthey are, yes.
12:21:54FromGitter<alehander92> so i cant do like `{.exportc, Blocking.}` or something like that
12:22:44FromGitter<alehander92> or like `{.tag[Blocking].}`
12:22:52*dadada quit (Ping timeout: 256 seconds)
12:22:52AraqI still don't care about "blocking", it's stupid :-)
12:23:08FromGitter<alehander92> and tell the compiler: check only for that in the set of all tags in this function
12:23:32FromGitter<alehander92> well, i had a case where it would've been possible to tell me that my async function is blocking
12:23:42FromGitter<alehander92> and i realized after a strange behavior problem
12:24:05FromGitter<alehander92> like, i agree its not always super useful, but still cool to have
12:24:23Araq.tags: [not Blocking]
12:24:27*dadada joined #nim
12:24:34Araqseems to be what you're asking for
12:24:50*dadada is now known as Guest28055
12:24:56Araqwe don't have 'not' for tags though
12:25:04FromGitter<alehander92> especially if one can easily write something like "await this from a long-running thread" etc
12:25:14FromGitter<alehander92> well yes
12:25:55Araqclyybber, how does ATS do it?
12:26:57*kenran joined #nim
12:27:18clyybberThey have proofs as explicit CT only objects
12:28:02clyybberI guess we could attach those proofs to the type by default
12:28:11clyybbers/type/symbol
12:28:46clyybberFor example open() would return a file and a proof that the file is open
12:29:05Araqwhat's the difference to .ensures?
12:29:29clyybberNot a lot, or nothing possibly
12:30:07Araqhttps://github.com/nim-lang/Nim/pull/13743/files#diff-f0f01a54ace4713173aa55429f9a7a7dR16-R17
12:30:09disbot➥ DrNim (Nim compiler with Z3 integration)
12:31:29Araqwe need to handle join points in a smart way though
12:31:58clyybberYeah, so we have ensures and requires. When f = open {.ensures: f.state == opened.}; ... close(f) {.requires: f.state == openend.} work then there should be no difference
12:32:01Araqprobably replace the sempass2 patches with a new CFG based pass
12:32:48clyybberAraq: We could attach those ensures (axioms/proofs) to the symbol, here f or f.state
12:32:53clyybberLike in the ast
12:33:01clyybberThats what I meant
12:33:12Araqthey are attached to the proc type
12:33:19Araqso that even indirect calls work properly
12:33:40clyybberI mean in the caller scope
12:33:47clyybberWhen you do f = open
12:34:03clyybberthen the symbol f has a proof for f.state == opened attached to it
12:34:04Araqin the caller scope we track the "knowledge", currently via a stack
12:34:08clyybberYeah
12:34:41Araqthe stack is good for when you only allow structured programming (no 'break', return etc)
12:34:46Araq*good enough
12:34:52clyybberI see
12:35:02FromGitter<Vindaar> thanks to whoever wrote this little paragraph: https://nim-lang.github.io/Nim/manual.html#closures-creating-closures-in-loops <3 ⏎ I realized what the problem was, but had no idea how to fix it.
12:35:18clyybberI think it was foldl
12:36:27clyybberAraq: It would be amazing to give "us" access to the proofs in our code, so that we could manually construct new proofs when needed
12:37:26Araqbut you can, via .assume
12:37:33clyybberYeah
12:37:38clyybberRight
12:38:26clyybberReally cool
12:38:58clyybberAraq: Can you do ifs in assume?
12:39:25clyybberLike: if ensured x > 0: assume x > 1 ?
12:39:29clyybberstupid example
12:40:01Araqyou can "assume" an implication
12:40:09clyybberAh nice
12:40:18clyybberso we would do {.assume x > 0 -> x > 1
12:40:22clyybberright?
12:40:23Araq.assume: (x >= 0) -> (y == 4)
12:40:27clyybbercool
12:43:25clyybberAraq: I think for open; ...; close; to be provable we need to consume proofs by default
12:43:29clyybberThats what ATS does too
12:43:41Araqwhat does that mean?
12:44:40clyybberIt means that for example in f = open(); close(f); close "consumes" the proof created by open: f.state == opened, and returns a new proof: f.state == closed
12:46:01clyybberSimilarily any call like someProc(f) needs to consume f.state == opened by default
12:46:45clyybberAnd when we can prove that someProc(f) doesn't close f, someProc can return a "new" f.state == opened
12:46:49Araqseems to be covered by .requires and .ensures
12:47:33Araqthe state transitions need to be path sensitive though and we need to document/clean up the rules about locations
12:47:35FromGitter<alehander92> Araq
12:47:59clyybberAraq: Yeah, so that we don't get f.state == opened and f.state == closed at the same "time"
12:48:42Araqwell that's the "invalidateFacts" that assignment does
12:48:51AraqI'm not worried about it, it's simple
12:49:01AraqI'm worried about things like:
12:49:06Araqif cond: result = 3
12:49:12Araqelse: result = 4
12:49:17Araq{.ensures: result > 0.}
12:49:37Araqwe need to 'or' together the facts after an 'if'
12:50:26clyybberMaybe we should give z3 the cond
12:50:30clyybberand the ifs
12:50:45Araqwe do that
12:50:50clyybberOh
12:50:50Araqbut incompletely
12:51:12Araqbut yeah, that's quite do-able when you use structured programming
12:54:09Araqif a[i] < a[j]: swap(a[i], a[j]) # how to tell Z3 that afterwards a[i] >= a[j] holds?
12:55:04*dddddd joined #nim
12:55:38*Kaivo joined #nim
12:58:53clyybberAraq: I think you must construct a new a
12:59:16clyybberThat has a[i] = oldA[j] and a[j] = oldA[i]
12:59:59clyybberAnd have those as implications of the condition
13:01:39clyybberSo you would have a new copy of a that is the old one with (oldA[i] < oldA[j]) -> a[i] == oldA[j] and a[j] == oldA[i]
13:02:18clyybberBasically invalidatingFacts about parts of a
13:03:43Araqbbl
13:03:57Araqclyybber, keep in mind that Z3 has native array support
13:04:20Araqwe can essentially map a[i] to Z3's a[i] and let it deal with it
13:10:10clyybberah, cool!
13:14:40*CcxWrk_ joined #nim
13:14:40*CcxWrk quit (Killed (weber.freenode.net (Nickname regained by services)))
13:14:40*CcxWrk_ is now known as CcxWrk
13:15:48*k0mpjut0r quit (Ping timeout: 240 seconds)
13:15:48*Cadey quit (Ping timeout: 240 seconds)
13:15:49*golechwi quit (Ping timeout: 240 seconds)
13:15:49*gangstacat quit (Ping timeout: 240 seconds)
13:15:49*drewr quit (Ping timeout: 240 seconds)
13:15:49*unclechu quit (Ping timeout: 240 seconds)
13:15:49*lqdev[m] quit (Ping timeout: 240 seconds)
13:15:49*FromGitter quit (Ping timeout: 240 seconds)
13:15:52*golechwi1 joined #nim
13:16:17*FromGitter joined #nim
13:16:29*waleee-cl joined #nim
13:16:47*gangstacat joined #nim
13:17:09*Cadey joined #nim
13:20:27*k0mpjut0r joined #nim
13:20:55*lqdev[m] joined #nim
13:21:54*unclechu joined #nim
13:29:00clyybberAraq: I think skipConv must be a little smarter. https://github.com/nim-lang/Nim/issues/13780 is caused by this: nkObjDownConv(nkObjUpConv(g.vm))
13:29:02disbot➥ arc: codegen or inject destructor error ; snippet at 12https://play.nim-lang.org/#ix=2fZC
13:29:14clyybberWhere g.vm is a RootRef
13:29:34clyybberAnd nkObjUpConv(g.vm) is a PCtx
13:30:02clyybberAnd nkObjDownConv(nkObjUpConv(g.vm)) is a PPassContext
13:30:32clyybberSo ideally skipConv should skip both
13:30:59*drewr joined #nim
13:31:31clyybberBut because it is "afraid" that the nkObjUpConv would loose too much information it doesn't skip it
13:32:06*lritter joined #nim
13:39:53*Guest28055 quit (Ping timeout: 252 seconds)
13:45:25*dadada joined #nim
13:46:13*dadada is now known as Guest40319
13:46:49clyybberHmm, or maybe thats not the issue. Somehow a manual (PPassContext(PCtx(RootRef(g.vm)))) = nil works fine
13:54:21clyybberHa, think I found the culprit. skipConv skips nkObjDownConv and nkObjUpConv's too hastily
14:17:04*oswin[m] quit (Quit: Idle for 30+ days)
14:17:24clyybberAraq: I wrote down my thoughts here if you want to take a look: https://github.com/nim-lang/Nim/issues/13780#issuecomment-606024178
14:17:26disbot➥ arc: codegen or inject destructor error ; snippet at 12https://play.nim-lang.org/#ix=2fZC
14:18:16*couven92 quit (Ping timeout: 258 seconds)
14:23:01*Guest40319 quit (Ping timeout: 264 seconds)
14:29:38*jxy quit (Ping timeout: 240 seconds)
14:33:34*dadada joined #nim
14:33:56*dadada is now known as Guest46925
14:47:14*endragor quit (Remote host closed the connection)
14:50:04*jxy joined #nim
14:52:58dom96a virtual Nim conference huh
14:52:59dom96cool
14:53:53dom96Araq, what month/date were you thinking about?
14:55:28*Jesin quit (Quit: Leaving)
15:01:57Araqdom96, end of April?
15:02:10Araqhmm no, too early
15:02:22Araqtalkers must have a chance to get prepared
15:02:27Araq*speakers
15:02:42Araqso ... I dunno, my birthday is in June...
15:03:17Araqclyybber, I think I figured it out
15:03:49clyybberAraq: The DrNim issue?
15:03:53Araqif cond: a else: b can be translated to 'facts(a) or facts(b)'
15:04:23Araqbut it's much better to translate it into: cond -> facts(a) and (not cond) -> facts(b)
15:04:29clyybberBut we must tell it the condition somehow, right, otherwise we are too pessimistic
15:04:36clyybberYeah
15:04:47Araqand then we should be able to prove the sorting operation
15:05:09Araqlikewise for 'case' statements
15:05:41Araqwe model the machine state completely by propositions
15:06:05clyybberYeah, thats the correct approach I'm sure
15:06:10Araqfor example, a = 3; b = 4 ==> a == 3 and b == 4
15:06:26Araqyeah, the state explosion could be bad though
15:07:08FromGitter<alehander92> Araq, is there a way
15:07:14Araqif 'if' becomes an 'implies' I really begin to question our DFA with join points
15:07:15FromGitter<alehander92> to annotate a referece to a const object
15:07:49Araqas we clearly want the "domination" information to be as easy to see as reasonable
15:08:05clyybberYou mean the DFA that we use in injectdestructors?
15:08:09Araqyes
15:08:23Araqit's fine for our 'wasMoved' analysis ofc
15:08:32Araqbut it might be the wrong data structure for DrNim
15:08:45clyybberOh, you use it there too?
15:08:49clyybberI wasn't aware
15:08:58*oz quit (Quit: EOF)
15:09:08AraqI don't but I considered it
15:09:12clyybberah
15:09:45*Jesin joined #nim
15:10:00*kenran quit (Quit: leaving)
15:10:03clyybberI think its fine. Since its a list when you go through it you can build up your info about the "current state" yourself
15:10:13clyybberI think its fine as an underlying data structure
15:10:23Araqmaybe
15:10:27*literal quit (Ping timeout: 260 seconds)
15:10:32FromGitter<alehander92> it might be like `a: const A`: it would require for one to just pass a `const`or a parameter which was already passed as "const"
15:10:43*oz joined #nim
15:10:50FromGitter<alehander92> which seem enough to me, but i havent used enough c++ to realize all the probable problems :D
15:10:52*literal joined #nim
15:10:53Araqwould be nice if it worked because then we can continue to use 'return' and 'break' :P
15:11:18clyybberYeah, return and break must be usable
15:11:37FromGitter<alehander92> but this way it might look like a "const reference to non-const value"
15:11:41FromGitter<alehander92> which is not what i mean hmm
15:11:58*planetis[m] quit (Ping timeout: 240 seconds)
15:12:00*skrylar[m] quit (Ping timeout: 240 seconds)
15:12:01*khanate[m]0 quit (Ping timeout: 240 seconds)
15:14:14*NimBot joined #nim
15:14:18dom96Araq, I was thinking May
15:14:22FromGitter<alehander92> because my bug was location = process.location; await stuff(location) ... later another function actually works with changed location.path
15:14:40*planetis[m] joined #nim
15:14:41FromGitter<alehander92> i admit, with some code discipline this is probably obvious but after all we are into type checking
15:14:47FromGitter<alehander92> write tracking might be useful for more things like multiple refs i guess
15:15:17FromGitter<alehander92> but can it work without interfunction analysis
15:16:04*skrylar[m] joined #nim
15:16:40FromGitter<alehander92> (and yeah in this case one might be able to use a non-ref object, but often this might not be the case)
15:19:24FromGitter<cooldome> Guys, hasDestructor(for tyRef) in compiler returns false in `--gc:arc` mode is this intentional?
15:19:28*rockcavera quit (Remote host closed the connection)
15:20:08clyybbercooldome: Hmm, I don't think so
15:21:18FromGitter<Clyybber> @cooldome Btw, I've investigated #13780 and wrote down the results there: https://github.com/nim-lang/Nim/issues/13780#issuecomment-606024178 . I think the core issue is a backend one
15:21:19disbot➥ arc: codegen or inject destructor error ; snippet at 12https://play.nim-lang.org/#ix=2fZC
15:23:52*rockcavera joined #nim
15:25:24AraqClyybber: how can it be a codegen issue? shouldn't injectdestructors produce correct code instead
15:25:27*khanate[m]0 joined #nim
15:25:42Araqbackends are always allowed to introduce temporaries
15:26:04*oz quit (Quit: EOF)
15:26:09FromGitter<cooldome> @ clyybber. Thanks @clybber, confirmed if I define ref in type section hasDestructor() gives true, but if write "ref MyObj" in place I get hasDestructor() false. I will fix it
15:26:34Araqcooldome: there are bugs left...
15:27:24*oz joined #nim
15:35:43*filcuc quit (Quit: Konversation terminated!)
15:38:25narimirandom96: start of may is only a month away (yep, a rhyme); we (both organizers and speakers) will need more time than that to prepare, IMO
15:39:33Araqwe need to schedule a sound&hardware check 2 days before the conference
15:40:47*zama quit (Ping timeout: 260 seconds)
15:41:19dom96narimiran, yeah, I was thinking late May
15:41:36dom96Maybe early May we could have a practice run
15:41:40*zama joined #nim
15:42:32Araqfirst week of June
15:42:37Araqmy suggestion
15:43:53dom96sure, I'm happy with that
15:45:09*matti quit (*.net *.split)
15:45:10*d-nice2[m] quit (*.net *.split)
15:45:10*oprypin quit (*.net *.split)
15:45:10*qwertfisch quit (*.net *.split)
15:45:11*WilhelmVonWeiner quit (*.net *.split)
15:45:35*matti joined #nim
15:45:35*d-nice2[m] joined #nim
15:45:35*oprypin joined #nim
15:45:35*qwertfisch joined #nim
15:45:35*WilhelmVonWeiner joined #nim
15:47:37*Jjp137 quit (Ping timeout: 264 seconds)
15:47:38*mwbrown quit (Ping timeout: 264 seconds)
15:48:06*BitPuffin quit (Ping timeout: 240 seconds)
15:48:26*unclechu quit (Ping timeout: 240 seconds)
15:48:30*d-nice2[m] quit (Ping timeout: 240 seconds)
15:48:35*planetis[m] quit (Ping timeout: 252 seconds)
15:48:39*leorize[m] quit (Ping timeout: 246 seconds)
15:48:39*salotz[m] quit (Ping timeout: 246 seconds)
15:48:46*lqdev[m] quit (Ping timeout: 240 seconds)
15:48:49*k0mpjut0r quit (Ping timeout: 246 seconds)
15:48:57*vycb[m] quit (Ping timeout: 246 seconds)
15:48:57*sendell[m] quit (Ping timeout: 246 seconds)
15:48:57*grantmwilliams quit (Ping timeout: 246 seconds)
15:49:14*khanate[m]0 quit (Ping timeout: 256 seconds)
15:49:14*Demos[m] quit (Ping timeout: 256 seconds)
15:49:14*pqflx3[m] quit (Ping timeout: 256 seconds)
15:49:17*skrylar[m] quit (Ping timeout: 260 seconds)
15:49:18*reversem3 quit (Ping timeout: 260 seconds)
15:49:18*chris2020[m] quit (Ping timeout: 260 seconds)
15:49:18*s87651[m] quit (Ping timeout: 260 seconds)
15:49:21*watzon quit (Ping timeout: 272 seconds)
15:49:47*matti quit (Ping timeout: 240 seconds)
15:50:00*mwbrown joined #nim
15:52:54clyybberAraq: Its a codegen issue because the backend shouldn't be allowed to introduce temporaries here IMO
15:53:03clyybberSince we are talking about lvalues here
15:53:21clyybberAnother option of course would be to disallow PCtx(g.vm) = nil
15:53:26clyybberAnd patch injectdestructors
15:54:05*matti joined #nim
15:54:27*oprypin quit (Ping timeout: 240 seconds)
15:55:04*oprypin joined #nim
15:56:30Araqmaybe you're right
15:59:14Araqlvalues are important, yes
16:22:26*thomasross joined #nim
16:27:24*solitudesf joined #nim
16:30:30*solitudesf- quit (Ping timeout: 256 seconds)
16:36:13*leth` quit (Read error: Connection reset by peer)
16:42:24PMunchWait, are we having a virtual NimConf?
16:43:50Yardanicoyeah I saw that post too :P
16:44:07*qwertfisch quit (Ping timeout: 240 seconds)
16:44:21Yardanicosad that I don't really have anything to talk about, but I'll at least enjoy all other talks :P
16:44:30Yardanicoand I hope it all gets recorded in good quality and put on youtube
16:44:51PMunchPost?
16:44:56PMunchI just saw someone talking about it here?
16:45:26PMunchAh, found it on the forum
16:45:33Yardanicohttps://forum.nim-lang.org/t/6132
16:45:35Yardanicoye
16:45:36*s87651[m] joined #nim
16:45:45*qwertfisch joined #nim
16:45:47Yardanicowe really need a feature to pin important threads on top of the forum :P
16:46:29Yardanicooh wait https://github.com/nim-lang/nimforum/pull/229 got merged, so I hope we'll also have categories soon
16:50:24FromGitter<alehander92> Yardanico does gdb work for you now
16:50:30Yardanicojust fyi the forum frontend is still running the version compiled by nim 0.18.1
16:50:39Yardanico(https://forum.nim-lang.org/js/forum.js)
16:50:54Yardanico@alehander92 I didn't really test it since I don't really use GDB anyway, I just was curious back then
16:50:58FromGitter<alehander92> btw guys what about async lifetimes: something that async functions can check cooperatively to know if they need to drain themselves
16:51:00YardanicoI might test it now though
16:51:06FromGitter<alehander92> i guess thats like cancellation
16:54:18*Guest46925 quit (Ping timeout: 256 seconds)
16:54:41*dadada joined #nim
16:55:03*dadada is now known as Guest80430
17:03:32PMunchalehander42, you can pass in a distinct boolean as a cancellation token
17:05:45shashlickare `typedef struct X` or `typedef struct X {}` valid C?
17:07:18Yardanicohttps://en.wikipedia.org/wiki/Typedef#Examples
17:07:28Yardanicooh wrong link
17:08:18*Demos[m] joined #nim
17:08:18*reversem3 joined #nim
17:08:18*watzon joined #nim
17:08:18*k0mpjut0r joined #nim
17:08:18*unclechu joined #nim
17:08:19*khanate[m]0 joined #nim
17:08:19*leorize[m] joined #nim
17:08:19*salotz[m] joined #nim
17:08:19*lqdev[m] joined #nim
17:08:19*GitterIntegratio joined #nim
17:08:19*planetis[m] joined #nim
17:08:20*BitPuffin joined #nim
17:08:24*chris2020[m] joined #nim
17:08:25*grantmwilliams joined #nim
17:08:25*sendell[m] joined #nim
17:08:25*vycb[m] joined #nim
17:08:26*d-nice2[m] joined #nim
17:08:28*skrylar[m] joined #nim
17:08:32*pqflx3[m] joined #nim
17:10:40*jjido joined #nim
17:11:49FromGitter<sealmove> wow, for inheritance order of declaration in type section is significant
17:12:24FromDiscord<clyybber> yeah, and for sets of enums too, but thats a bug
17:12:44FromDiscord<clyybber> sealmove: Do you mean you cant inherit from a type that you havent yet defined?
17:14:15FromGitter<sealmove> yes, but I thought order under the same type section wasn't important. for example you can define a type that has another type (say X) as field and define X after it
17:26:34*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
17:28:36FromGitter<sealmove> Is is possible to make a destructor for a type that uses inheritance?
17:29:35FromGitter<sealmove> https://play.nim-lang.org/#ix=2g17 or https://play.nim-lang.org/#ix=2g18 don't work.
17:30:17FromGitter<sealmove> sorry https://play.nim-lang.org/#ix=2g1a *
17:30:40YardanicoI don't think you need "var" for a "ref object"
17:31:00Yardanicobut I don't really know about using destructors with ref objs :)
17:32:15FromGitter<sealmove> `=destroy` only accepts var parameter
17:32:27leorizehttps://play.nim-lang.org/#ix=2g1c
17:33:57leorize@sealmove: ^ you gotta test it though :P
17:33:59FromGitter<sealmove> wow leorize nice! what's the difference between `A = ref AObj` and `A = ref object of AObj`?
17:34:38leorizeit's `ref T` vs `a ref object inherited from AObj` :)
17:36:13leorize@sealmove: I'm not sure if `=destroy` is inherited, so you gotta verify if that's the case
17:36:23leorizeI'd recommend using devel because it has more fixes
17:37:01*clyybber quit (Quit: WeeChat 2.8)
17:37:40FromGitter<sealmove> I'm using devel anyway
17:39:50FromGitter<sealmove> this should work, but doesn't? what am I missing? https://play.nim-lang.org/#ix=2g1e
17:41:56*a_b_m joined #nim
17:42:10leorize@sealmove: it does work if you use arc
17:42:17leorizelooks like an stdlib bug
17:42:26leorizeI mean compiler bug?
17:42:28*a_b_m quit (Client Quit)
17:42:43FromGitter<sealmove> I see
17:42:51leorizeit won't work if you try it with B though, so maybe a bug report for that too
17:43:32FromGitter<sealmove> I am sure Araq knows that ref objs don't work with destructors
17:43:45leorizea bug report never hurts
17:44:02FromDiscord<clyybber> hey
17:44:13FromDiscord<clyybber> ref objects should work with destructors
17:44:30FromDiscord<clyybber> but you need to define =destroy on the "dereffed" type
17:44:32*abm quit (Ping timeout: 246 seconds)
17:44:50*Guest80430 quit (Ping timeout: 256 seconds)
17:44:59leorizeclyybber: they don't appear to be invoked for the --gc:refc
17:45:12FromDiscord<clyybber> ah
17:45:13FromDiscord<clyybber> yeah
17:47:24*Jjp137 joined #nim
17:47:46FromGitter<sealmove> which gc then?
17:48:14FromDiscord<clyybber> arc
17:48:23leorizeit works if you use arc, however not across inheritance, so a bug should be filed for that
17:48:32*dadada joined #nim
17:48:55*dadada is now known as Guest21968
17:49:00FromGitter<sealmove> btw where can I see all the --gc: options?
17:49:33FromGitter<sealmove> https://nim-lang.org/docs/gc.html seems outdated
17:49:33leorize--fullhelp
17:49:49leorizehttps://nim-lang.github.io/Nim/gc
17:50:14FromGitter<sealmove> oh thanks, both very useful!
17:58:01*jjido joined #nim
17:58:37FromGitter<alehander92> PMunch yeah i guess i am doing something like that
17:58:48FromGitter<alehander92> but inside an object to have info like name
17:59:20*pbb quit (Remote host closed the connection)
18:00:41*pbb joined #nim
18:02:29FromGitter<alehander92> hey guys: if i decide to try to candidate for a talk for this online nim conf :D what would you want? 1) playing with a toy os with async primitives and "type checked" function-like processes 2) some kind of thing about our debugger env project 3) gara/pattern matching
18:05:05*pbb quit (Excess Flood)
18:05:55*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
18:08:02FromGitter<alehander92> btw why isnt rawWriteStackTrace public
18:09:18*pbb joined #nim
18:10:38*Yardanico quit (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.)
18:10:55*Yardanico joined #nim
18:11:54*abm joined #nim
18:11:57*Yardanico quit (Client Quit)
18:12:30*Guest21968 quit (Ping timeout: 256 seconds)
18:14:41*Yardanico joined #nim
18:18:23*Yardanico quit (Client Quit)
18:18:40*dadada__ joined #nim
18:19:31*Yardanico joined #nim
18:32:44*Kaivo quit (Quit: turning off vpn)
18:33:26*tefter quit (Quit: WeeChat 2.7.1)
18:35:37*Kaivo joined #nim
18:37:12Araqalehander92: 1 or 2 or both :-)
18:41:38*couven92 joined #nim
18:42:38*ldlework joined #nim
18:42:47*dadada__ quit (Ping timeout: 260 seconds)
18:44:51FromDiscord<ksandvik> If all this is recorded and published to YouTube, the more the better.
18:45:31FromGitter<alehander92> love it!
18:47:06FromGitter<alehander92> btw async cancellation: how hard is it to add this
18:48:02FromGitter<alehander92> i am interested in just being able to `.cancel()` a future and recursively cancel children but maybe this is somehow related to destructors
18:48:21FromGitter<alehander92> maybe thats what i heard last time, sorry for not remembering
18:48:30*jjido joined #nim
18:48:42*dadada joined #nim
18:49:06*dadada is now known as Guest55387
18:52:30narimiran@alehander92: all three, pretty please
18:58:15liblq-devis there *any* way besides procCall that a base method could be called instead of an inherited method?
18:58:34Araqhardly, what's wrong with proccall
18:58:38*mwbrown quit (Quit: Exiting)
18:59:01liblq-devnothing, it's just that for some strange reason a base method is called where a child method should be
18:59:31Araqthen you failed to "override" the base method
18:59:44liblq-devright *here* in fact: https://github.com/liquid600pgm/rdgui/blob/master/src/rdgui/control.nim#L49
18:59:50liblq-devit works everywhere except in this proc
19:00:04liblq-devthe method is ctrl.width
19:00:27liblq-dev(and ctrl.height too, but let's keep it down to one method because the issue is likely the same)
19:02:11Araqstrange, are you sure the Control is a Box?
19:02:19Araqyou can check via the 'of' operator
19:03:52*mwbrown joined #nim
19:04:43liblq-devit's a Button, which is correct.
19:05:17liblq-devI checked via `ctrl of Button`, as you suggested
19:07:41Araqwell I only looked at the Box's width declaration
19:07:45Araqwhere is Button's?
19:08:08liblq-devit's in a different file, https://github.com/liquid600pgm/rdgui/blob/master/src/rdgui/button.nim#L17
19:09:57Araqliblq-dev, you used 'proc' and not 'method' ;-)
19:10:08liblq-devoh crap
19:10:13liblq-devthanks
19:10:25*luis_ joined #nim
19:10:40liblq-devman, I wonder why bugs like this happen to me. every. single. time.
19:15:15*luis_ quit (Remote host closed the connection)
19:15:49*luis_ joined #nim
19:15:51*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:19:53Araqliblq-dev, ask for linting support
19:20:01Araqthe compiler could issue a warning
19:22:21rayman22201alehander92 Chronos has promises with cancel. Maybe investige switching async frameworks?
19:27:06*luis_ quit (Ping timeout: 246 seconds)
19:30:31leorizeso is chronos the to-go framework now?
19:44:48FromGitter<alehander92> rayman22201 but i dont want a split
19:47:08*rockcavera quit (Remote host closed the connection)
19:55:18FromGitter<zetashift> @dom96 are you still using SFML for your Nim games?
19:56:01*rockcavera joined #nim
19:56:16FromGitter<alehander92> narimiran no, thats a bit much :D
20:05:20dom96zetashift: only for Ludum Dare, currently using sdl for my new game
20:18:05*narimiran quit (Ping timeout: 265 seconds)
20:19:52FromGitter<zetashift> ah okay
20:31:04*luis_ joined #nim
20:31:26*jjido joined #nim
20:34:19*qbradley quit (Quit: WeeChat 2.6)
20:34:27*qbradley joined #nim
20:37:55*filcuc joined #nim
20:45:40*PMunch quit (Quit: leaving)
20:47:10*s87651[m] quit (Quit: Idle for 30+ days)
20:47:15*liblq-dev quit (Quit: WeeChat 2.7.1)
21:02:09*abm quit (Quit: Leaving)
21:16:25*filcuc quit (Ping timeout: 264 seconds)
21:18:59*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
21:21:49*Vladar quit (Quit: Leaving)
21:23:34*moerm joined #nim
21:24:16moermHello. I'm trying my irc client for the first time and hope it works/you can see me
21:24:26Yardanicomoerm: yeah, I can see you :P
21:24:36moermThanks.
21:25:15moermI'm stupid with those messenger/chat thingies. discord fails to send the verif. email so I had to try irc
21:26:09moermI see the same content/discussion as on discord here, right?
21:26:16FromDiscord<ksandvik> Yes.
21:26:18leorizeyes
21:26:18*ksandvik joined #nim
21:26:33leorizewith a perk of being able to ping people with tab-completion :P
21:27:31moermThanks (whatever "pinging" means on irc ...). I'm grateful for everybodies hand holding/help
21:28:27leorizepinging on irc is just typing out their name :)
21:28:51moermSo, it's like tab completion in the shell? Nice.
21:29:15moermafk (some minutes)
21:29:43ksandvikAFAIK you can't see all the previous conversations in irc unlike Discord but you could use https://irclogs.nim-lang.org
21:32:27*rockcavera quit (Remote host closed the connection)
21:34:32*moerm_ joined #nim
21:35:22*chris2020[m] left #nim ("Kicked by @appservice-irc:matrix.org : Idle for 30+ days")
21:35:26*pqflx3[m] left #nim ("Kicked by @appservice-irc:matrix.org : Idle for 30+ days")
21:35:27*d-nice2[m] left #nim ("Kicked by @appservice-irc:matrix.org : Idle for 30+ days")
21:35:27*vycb[m] quit (Quit: Idle for 30+ days)
21:35:28*grantmwilliams quit (Quit: Idle for 30+ days)
21:35:50*ksandvik quit (Quit: Leaving.)
21:36:12FromGitter<deech> Is this correct behavior for NimScript's `thisDir`? https://play.nim-lang.org/#ix=2g2V
21:36:51*moerm quit (Ping timeout: 250 seconds)
21:36:59FromGitter<deech> In the above example even though `proc test ...` is in `/tmp/a` because it is imported by `/tmp/b/b.nims` `thisDir` prints `/tmp/b`.
21:37:26leorizesounds like a bug
21:37:55FromGitter<deech> Great. Thanks. And is that the right way to import a nims file?
21:38:08FromGitter<deech> It seems to do that same as `include`.
21:39:00leorizeusually we just write nim code then import it as usual
21:39:14leorize(ie. put them in .nim files)
21:40:09FromGitter<deech> How would you import a nim module in a relative directory? `import ../a/a`?
21:40:17Araqyeah
21:40:19FromDiscord<Rika> yeah
21:40:51Araqleorize: hmmm not sure if I consider it a bug but tbh I don't know its documentation
21:41:52FromGitter<deech> Changed `/tmp/a/a.nims` to `/tmp/a/a.nim` and `import "../a/a.nims"` to `import ../a/a`. Same result.
21:43:52leorizeAraq: looks like not a bug? https://github.com/nim-lang/Nim/blob/1d665adecde3b3bf16e64068e83c0b3cb0171856/lib/system/nimscript.nim#L312
21:43:58Araq"Retrieves the directory of the current ``nims`` script file"
21:44:06leorizeor maybe :p
21:44:29*tane quit (Quit: Leaving)
21:44:52Araqworks as documented, you cannot really have multiple .nims files
21:45:05Araqthere is only the .nims file that NimScript executes
21:45:32FromGitter<deech> Where does it say you can't have multiple nims files?
21:45:33leorizewell they mentioned renaming the imported module to .nim, still same result
21:47:03FromGitter<deech> And if that's true why is `import` supported in `nims` files?
21:47:33Araqyou named your other nim modules as .nims but that doesn't mean anything, you can also name them module.txt and 'import "module.txt" '
21:47:36FromGitter<deech> Oh wait, so the idea is that you have a single nims file that import multiple `nim` files?
21:47:49Araqexactly
21:48:39FromGitter<deech> Ah, ok makes sense. Then the second case where I changed a.nims -> a.nim and it still did the same thing is a bug, yes?
21:49:36AraqI doubt it
21:50:22Araqthe compiler knows best what "the current nims script file" is
21:50:43Araqcould not import: X509_check_host
21:50:48Araqgah
21:50:56leorizemore openssl issues :)
21:52:09*silvernode joined #nim
21:53:46Araqalready solved it
21:55:07FromGitter<deech> Suggest adding to the no-multiple Nims files somewhere in the docs.
21:56:06Araqnoted
22:03:07*rockcavera joined #nim
22:04:57shashlickhas anyone made an easy azure pipelines or github actions setup for testing nim libraries across multiple versions?
22:05:58*fredrik92 joined #nim
22:07:13*rnrwashere joined #nim
22:09:15*couven92 quit (Ping timeout: 265 seconds)
22:09:33*rnrwashere quit (Remote host closed the connection)
22:10:23*Trustable quit (Remote host closed the connection)
22:18:46*natrys quit (Quit: natrys)
22:19:17*solitudesf quit (Ping timeout: 246 seconds)
22:26:05*fredrik92 is now known as couven92
22:28:39*Jesin quit (Quit: Leaving)
22:30:21*rnrwashere joined #nim
22:32:55*Jesin joined #nim
22:37:45blackbeard420is db_postgres threadsafe? i assume it is since DbConn just wraps a PGconn*
22:37:57Araqyeah
22:40:27*rnrwashere quit ()
22:47:57*Guest55387 is now known as dadada
22:51:50*bunbunbunbunny joined #nim
22:54:34moerm_Araq How can I get and build drnim to have a closer look and play with it (and hopefully can make some sensible remarks/suggestions)?
22:59:46*luis_ quit (Ping timeout: 240 seconds)
23:08:49moerm_Araq - still here?
23:12:50leorizehe is rather busy with his work, so I won't expect answer to come out immediately :P
23:13:03*dadada quit (Ping timeout: 265 seconds)
23:13:17moerm_Thx
23:14:07leorizewhat I do know however, is that you can test this out by cloning the araq-doctornim branch
23:14:28leorizethe tool can be built using `koch drnim`
23:16:25*silvernode quit (Ping timeout: 264 seconds)
23:16:30moerm_I don't know git well. Could you give me a link to that branch, please?
23:16:49moerm_(that I can put behind "git clone")
23:17:26leorizeif you already cloned Nim, then just get in the folder, run: git pull; git checkout araq-doctornim
23:18:26moerm_Thanks, I'll try that. And when I did that I use 'koch drnim'?
23:18:38moerm_(sorry but we use only fossil)
23:18:42*dadada joined #nim
23:18:57leorizeafter you have checked out the branch, build koch
23:19:00leorizerun nim c koch
23:19:05*dadada is now known as Guest29536
23:19:11leorizethen run ./koch drnim
23:19:21moerm_OK, thx.
23:21:55*bunbunbunbunny quit (Quit: leaving)
23:23:35moerm_Yuck, it clones and builds Z3 again (I already have it and it's not a small build). Oh well, git ...
23:26:00moerm_- afk for a while - (but I will look later for new messages)
23:33:18moerm_... /Nim/compiler/vmops.nim(105, 10) Error: cannot open file: std/compilesettings
23:33:53leorizetry bootstraping the compiler first
23:33:57leorize./koch boot -d:release
23:34:01leorizethen do drnim
23:37:41*moerm__ joined #nim
23:37:42moerm__OK. Thx
23:38:08moerm__Same error
23:39:10moerm__(I'm in ... git/Nim created by cloning Nim)
23:41:10*moerm_ quit (Ping timeout: 256 seconds)
23:42:38*Guest29536 quit (Ping timeout: 256 seconds)
23:48:38*dadada joined #nim
23:49:03*dadada is now known as Guest2172
23:51:06*krux02_ joined #nim
23:53:41*krux02 quit (Ping timeout: 252 seconds)
23:55:20*dwdv quit (Ping timeout: 256 seconds)