<< 22-04-2018 >>
00:15:11*kunev quit (Ping timeout: 256 seconds)
00:15:54*kunev joined #nim
01:32:40*smt quit (Read error: Connection reset by peer)
01:32:46*Trustable quit (Remote host closed the connection)
01:33:01*smt joined #nim
01:34:09*Vantage_ quit (Quit: Leaving)
01:52:32*cspar quit (Ping timeout: 256 seconds)
01:53:03*cspar joined #nim
02:39:32*vantage quit (Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org)
03:07:07FromGitter<Varriount> Yay, NimLime now displays errors in the status bar when you click on the error regions.
03:31:46*yglukhov joined #nim
03:36:22*yglukhov quit (Ping timeout: 264 seconds)
03:41:17*dddddd quit (Remote host closed the connection)
04:35:01*DarkArctic quit (Read error: Connection reset by peer)
04:36:53*Senketsu quit (Quit: WeeChat 2.1)
05:13:25*r3d9u11 joined #nim
05:13:59*miran joined #nim
05:19:38*ftsf joined #nim
05:20:17ftsfhi, getting an error in js: "Uncaught ReferenceError: colonanonymous__223267 is not defined", no error in C version
05:20:26ftsfis that a known problem?
05:26:40Araqno.
05:28:36*r3d9u11 quit (Remote host closed the connection)
05:28:43ftsfhmm worked around by changing the anon function to a named one
05:29:16ftsfin the middle of a gamejam now but i'll make a proper test case after it's over.
05:31:08shashlickso if #4057 and #4103 are duplicates, can we close one?
05:36:06ftsfhttps://gist.github.com/ftsf/40ba9cbb8aa6c58bcafa8c221751244e
05:43:24*Vantage joined #nim
06:27:57*yglukhov joined #nim
06:38:09*nsf joined #nim
06:38:23*Vantage quit (Quit: Leaving)
06:46:14*miran quit (Ping timeout: 268 seconds)
06:58:33*Lord_Nightmare quit (Excess Flood)
06:59:08*Lord_Nightmare joined #nim
07:09:00*xet7 joined #nim
07:11:57*rokups joined #nim
07:14:04*yglukhov quit (Remote host closed the connection)
07:19:16*vantage joined #nim
07:31:16*yglukhov joined #nim
07:33:27*nc-x joined #nim
07:35:16nc-xmaybe https://github.com/nim-lang/Nim/issues/6975 is also duplicate of https://github.com/nim-lang/Nim/issues/4057 (though it does not uses --threads:on)
07:35:27*cspar quit (Ping timeout: 240 seconds)
07:39:49*nc-x quit (Ping timeout: 260 seconds)
07:40:51*cspar joined #nim
07:46:09*yglukhov quit (Remote host closed the connection)
07:54:56*yglukhov joined #nim
08:01:33*yglukhov quit (Remote host closed the connection)
08:10:03*cspar quit (Ping timeout: 256 seconds)
08:24:41*cspar joined #nim
08:29:39*ShalokShalom2 joined #nim
08:29:43ShalokShalom2https://en.m.wikipedia.org/wiki/Nim_(programming_language)
08:29:53ShalokShalom2Is Nimfix still in beta?
08:30:59*Senketsu joined #nim
08:42:38FromGitter<GULPF> I'm not sure it even exists anymore, it should probably be removed from the article
08:46:29*yglukhov joined #nim
08:46:31FromGitter<mratsim> spry-lang domain expired :/ (it’s a Rebol/smalltalk lang written in Nim and one of the best examples of Nim metaprogrmming …)
08:48:56FromGitter<alehander42> haha py2nim is on wikipedia, bizarre
08:49:12*yglukhov quit (Remote host closed the connection)
08:49:25*yglukhov joined #nim
08:49:39FromGitter<alehander42> yeah I remember spry-lang had some coverage on prog sites
08:50:44FromGitter<alehander42> I don't know why, but "new prog lang written in X" is 90% of the time receipt for upvotes and comments
08:57:58*xkapastel quit (Quit: Connection closed for inactivity)
09:01:43*rauss quit (Read error: Connection reset by peer)
09:03:45*rauss joined #nim
09:06:12FromGitter<alehander42> i am sure there was a reason, but why again can't I produce `await` from a macro :?
09:06:31FromGitter<alehander42> e.g. ⏎ ⏎ ```a = quote: ⏎ await f()``` [https://gitter.im/nim-lang/Nim?at=5adc50975d7286b43a54eaa5]
09:06:55FromGitter<alehander42> ah probably the async macro runs first
09:07:01FromGitter<alehander42> and then when my macro expands, it's too late?
09:08:10FromGitter<alehander42> ok workaround: i am not sure why, but even if it my macro returns "echo stuff; async call"
09:08:22FromGitter<alehander42> await macroInvocation() seems to work
09:24:48ShalokShalom2well, somebody with deep awareness about Nim can update that outdated article
09:43:07FromGitter<arnetheduck> travis CI runtimes with vs without #7669: 35 min 58 sec vs 42 min 51 sec
09:44:54*ofelas quit (Quit: shutdown -h now)
09:49:03FromGitter<alehander42> I am still not convinced: why not just make `--deadCodeElim:on` the default, what is the harm in having the switch?
09:49:56*BitPuffin joined #nim
09:50:17FromGitter<arnetheduck> what's the use of having the switch? every switch means more code to maintain - code that the test suite rarely passes over
09:51:18FromGitter<alehander42> well at the very least, having that switch makes it very easy to compare what does dead code elimination actually remove in your source code
09:51:38FromGitter<arnetheduck> so does the compiler warning about unused code
09:54:15FromGitter<mratsim> I think remove options from Nim is good, the language has so much features that it probably feels ovewhelming if introduced the wrong way to non programming lang theorists (you can do X, Y, Z, and the coffee too)
09:56:19FromGitter<alehander42> I agree, just I didn't imagine this switch to be such a feature, but I don't know how much code is related to it
09:56:21dom96alehander42: You can generate 'yield'
09:56:45*cspar quit (Ping timeout: 264 seconds)
09:58:33FromGitter<arnetheduck> and how does it make it simple? can't diff at the `c` files anyway because DCE changes the traversal order therefore code ordering, and since a lot of people say nim is not a transpiler, you're not supposed to look in there anyway :)
09:58:49*cspar joined #nim
10:00:20FromGitter<alehander42> well you can still find the otherwise ungenerated functions quite easily
10:00:25dom96if dead code elimination has a bug, there will at least be a way to disable it
10:00:32FromGitter<alehander42> yeah, that too
10:00:54FromGitter<alehander42> I agree it would be useful in very few cases, but I still it could be there at least for compiler developer purposes
10:01:15FromGitter<alehander42> but again that's IMHO, not too familiar with all compiler internals
10:02:17FromGitter<alehander42> btw @mratsim do you have any examples of features of Nim that can be removed
10:03:00FromGitter<mratsim> GC ? :D
10:04:16FromGitter<arnetheduck> er.. the existence of an option makes it more likely that it's bugged - less code = less bugs, it's that simple.. by that argument, you should intro an option for every fixed issue in the tracker, in case it breaks
10:04:17FromGitter<alehander42> you do need optional gc :D
10:04:21FromGitter<mratsim> no feature jumps at me, I always have some use for features, especially the hidden ones at the bottom of the manual
10:05:24FromGitter<alehander42> @arnetheduck that's a strawman, dead code elimination is an important component of a compiler, not just a random fix, it makes some sense to have a flag for it
10:06:11FromGitter<alehander42> it's simply a cost/benefit thing: if the code for enabling such a flag is indeed too complicated, drop it
10:06:40FromGitter<alehander42> but if there are enough valid usecases for it, don't
10:07:12dom96Dead code elimination has been off by default for a long time. To suddenly make it default *and* remove the option to disable it is a bit reckless.
10:07:24dom96At least keep the flag for a couple of versions as a deprecation path
10:07:29FromGitter<arnetheduck> @alehander42 not really.. it's a useless leftover from when compilers did that kind of stuff because noone imagined a better way..
10:07:59dom96Also I can't imagine this flag adding that much more complexity
10:08:13dom96if deadCodeElim: performCodeElim() vs. performCodeElim()
10:10:56FromGitter<arnetheduck> @dom96 er, you're being ironic about complexity, right?
10:11:00FromGitter<alehander42> @arnetheduck ok, so how much code does it cost indeed? is the compiler full with many `when defined(deadCodeElim) .. else` stuff
10:11:03*leorize quit (Quit: WeeChat 2.1)
10:11:49FromGitter<alehander42> I would imagine you should have one place where you write " if this is not unused or deadCodeElim is off, include it in the final tree"
10:12:05FromGitter<alehander42> I don't get why would that add to the complexity of a compiler
10:12:29FromGitter<alehander42> if you already have "detect if this is unused" logic
10:13:50dom96yeah, I'm not being ironic, that's my thinking too
10:13:56FromGitter<arnetheduck> every feature choice adds complexity... in lines of code, in open bugs, in cognitive effort needed to decide between them.. look at it this way, the js backend (and nlvm) never had that feature to begin with (it's a dead option) - you would *never ever* request it as a feature if it didn't exist already - you wouldn't be able to come up with a sane use case
10:15:59FromGitter<arnetheduck> in this case, it's easy to see how many lines of code at least: a little more than #7669 - but they're *all useless* - the best reason you've come up with so far is that they exist in case there's a bug in them
10:19:33FromGitter<alehander42> yes, I looked at the PR: it removes two functions and changes two more conditions
10:19:49FromGitter<alehander42> everything else is processing switch / docs about switch / formatting
10:20:44FromGitter<alehander42> and removing `{.deadCodeElim: on.}` which is a completely orthogonal issue related to making deadCodeElim on by default
10:20:48*nsf quit (Quit: WeeChat 2.0.1)
10:22:31FromGitter<arnetheduck> @alehander42 yeah, which is why I find it quite hard to believe that you're arguing for something that adds 25% compile time and is useless at the same time :) why on earth should devs have to think about adding `{.deadCodeElim: on.}`?
10:22:46FromGitter<alehander42> but I am not arguing about that :D
10:22:57FromGitter<alehander42> you see, your idea is 2 ideas:
10:23:03FromGitter<alehander42> making dce on by default
10:23:06FromGitter<alehander42> and removing the switch
10:23:10FromGitter<alehander42> I agree with the first part
10:23:35FromGitter<alehander42> and agree with @dom96 for the second: at least for some time, the switch should remain there
10:24:11FromGitter<alehander42> this is what happens with all removed features in nim anyway: they go through deprecation period
10:24:30FromGitter<arnetheduck> no, it's just one idea - DCE is not a feature in 2018 - it's just the way compilers should work.. I'd even remove it from the promo material, it's embarrassing that it's there :)
10:25:35FromGitter<arnetheduck> the idea is to remove the option, the rest is a coincidental side effect :)
10:26:04FromGitter<alehander42> no, the second idea depends on realizing the first idea
10:26:05FromGitter<alehander42> :D
10:29:37*r3d9u11 joined #nim
10:30:27FromGitter<alehander42> the middleground is obviously the right decision: you get :on by default, the compile times are improved, people don't need to add that manually, and if they add `:off` they know this is going to disappear as an option soon ⏎ ⏎ and if really there is absolutely no usecase for it, nobody will complain, and the option would be totally removed
10:30:40FromGitter<alehander42> `:on` *
10:35:55FromGitter<arnetheduck> again, the branch is not about improving compile times, it's about removing cruft from the compiler that doesn't serve any purpose - what you're proposing is not a middle ground, it's picking the smallest cherry among many large cherries - less code, less risk of bugs, less cognitive effort for developers, fewer "put-this-magic-string-to-make-it-compile-faster-with-no-downside"
10:37:44FromGitter<alehander42> maybe I can't express myself correctly
10:38:18FromGitter<arnetheduck> if you're worried the test suite passing is not enough to verify the sanity of the code, sure, feel free to provide a patch to flip the default.. feels like a waste of time though - for example, since there's a whole bunch of `{dce:off`} in the code, the feature is not completely untested either
10:38:24FromGitter<alehander42> what I propose fulfills "less risk of bugs", "less cognitive effort..", "fewer put-this-magic"
10:38:41FromGitter<alehander42> and it will fullfill the "less code" if it's indeed free of bugs in a short time
10:40:55FromGitter<arnetheduck> isn't that what `devel` is for? to flush out unexpected bugs.. or are you proposing to keep the code over a release?
10:41:02FromGitter<alehander42> the only thing different from your proposal is that somewhere there will be this *deprecated* option for a little longer: the same thing that happens with almost any other default/feature being removed
10:42:20vantagesometimes you might not want dead code elimination, like when compiling to a library
10:43:01FromGitter<mratsim> exportc proc are not eliminated iirc
10:43:17vantageok nvm then
10:44:40FromGitter<arnetheduck> I still don't get it - you would never do that for any other feature addition or fix - ie add or keep a broken branch of code just to have the option to disable it.. the branch keeps backwards compatibility by still parsing it both on command line and in pragma
10:45:33FromGitter<alehander42> which branch is broken?
10:45:47FromGitter<arnetheduck> dce is not new, it gets used all the time even when compiling without it on the command line, it's the only mode for the JS backend and passes the test suite
10:46:17FromGitter<alehander42> all I am saying is that you can't possibly know every usecase people might have for a feature, so keeping the option to turn it `:off` for a little while can't hurt: after such a period you remove it completely
10:46:27FromGitter<alehander42> it's the basic concept of "deprecation period"
10:46:30FromGitter<arnetheduck> imagine you're fixing a random bug - there's some code that's broken - you don't add an option and a branch to be able to disable the fix
10:47:36FromGitter<arnetheduck> just like you can't know if people rely on a bug - they might have written a workaround
10:48:14FromGitter<alehander42> the lack of an optimization is not a bug
10:48:40FromGitter<alehander42> you might like to call it that, but using the same logic you can forbid disabling any of the other optimizations
10:48:56vantagearne you could always just change nim.cfg?
10:50:09vantagei think make dead code elim automatic would have more losses than gais
10:50:19vantagegains*
10:52:21FromGitter<alehander42> even if you see it as a bug, that fact is that it has been a "feature" until now, so people have the right to rely on it for some weird purpose and have the right of a deprecation period for that
10:52:50*Vantage_ joined #nim
10:52:57FromGitter<arnetheduck> vantage, what would those gains be?
10:53:43vantagenim goes in c code comes out. dce removes that one to one relationship
10:54:01vantagemaking it more difficult to test the nim compiler
10:55:01vantageso you might add dce to always be on , but then your going to need to switch it off again
10:55:38vantageand that would need a large refactoring i think
10:56:05vantagethat might not be an issue but i imagine it could be.
10:56:50vantageat the end of the day "premature optimization is evil" i think thats why dce is on by default
10:56:59FromGitter<arnetheduck> if you're digging around in the c files, the only thing that changes is the order in which things are added it, and this order already does not correspond to what's in the nim file
10:57:41*vantage quit (Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org)
10:58:00Vantage_You could still have your line traces
10:58:35dom96arnetheduck: I'm surprised you're so against keeping this flag, is it because you fear you will have to support it in nlvm?
10:58:55FromGitter<alehander42> yeah, one can argue that in a visual debugger you might want to be able to see what kind of code would be generated for an unused function
10:59:23Vantage_btw arnetheduck, I've seen you vlvm project, its interesting, I wish there was cling repl for it
10:59:25FromGitter<alehander42> I am not saying I agree with that, just giving an example of a usecase you can't anticipate
10:59:32Vantage_nlvm*
11:00:26Vantage_I'm just wondering why its there to begin with
11:01:46Vantage_at the end of the day I think the total number of people that actually use dce is small. but probably most of them are those with finished products that are released. Does Dead code make debugging slower? whats so bad about it?
11:01:48FromGitter<arnetheduck> no, `nlvm` doesn't support that flag and won't - no one would ask for it, just like no one noticed it was missing for js :)
11:01:52FromGitter<mratsim> @Vantage "Premature optimization is evil” is often cited without the surrounding context: "We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%.”. Most of the code I write is performance critical and I always optimize down to the Assembly generatted for Nim code.
11:02:15FromGitter<alehander42> @mratsim nobody is arguing against making the option :on by default
11:02:22FromGitter<alehander42> `:on` , oh gitter ....
11:02:46Vantage_Well there's the saying, blind optimization is just stupid.
11:03:27FromGitter<mratsim> blind trust in sayings is stupid too ;)
11:03:40FromGitter<alehander42> no @Vantage_ that's not the case here, dce is important, my point is more general about the way to phase off a feature
11:03:55Vantage_I think that blind optimization is a stupid idea, it makes logical sense
11:03:58FromGitter<mratsim> I agree that it should follow the normal deprecation cycle.
11:04:00FromGitter<arnetheduck> Vantage, dce is neither a feature nor an optimization - it's just a way to traverse the AST - one is lazy, the other one eager.. there are already other ways in nim to make the traversal eager
11:04:53Vantage_what about... dum dum dum... a repl inside your program... you would need dead code then or no?
11:05:18FromGitter<mratsim> anyway, I think we are bikeshedding here, this is such a small feature with 0 impact, I’d rather talk about say Exceptions.
11:05:24FromGitter<arnetheduck> for it to be reachable from within your program, you'd have to export it
11:06:16FromGitter<alehander42> @mratsim btw did you have the chance to test with quicktest, or did you find a better way to randomize tests
11:06:21FromGitter<arnetheduck> haha, @mratsim of course we're bikeshedding - it's sunday evening and no pressure to do useful things :)
11:06:51FromGitter<mratsim> @Vantage_ All nim code currently should be compilable with or without dead-code elim. -d:release implies dead code slim, a code working on non-release but crashing on release is a bug.
11:07:31FromGitter<mratsim> @alehander42 didn’t try yet, I was busy squeezing the last drops of performance of mpint yesterday. Got it within 10% of ttmath without any assembly :D
11:07:43Vantage_mratsim what about Exceptions?
11:07:56FromGitter<alehander42> @mratsim haha awesome
11:08:08FromGitter<mratsim> @arnetheduck wants to remove exceptions too :thinking_face:
11:08:18Vantage_nooo!
11:08:24Vantage_why :'(
11:08:57FromGitter<mratsim> He said that they were dinosaurs leftover from a mystical amber originating from the Jurassic era (or something like that ;) )
11:09:16FromGitter<arnetheduck> so if we really wanna close this topic, maybe someone in favor of adding a deprecation period would be so kind as to create that branch that flips the switch? then my branch can stay purely useful and be merged whenever the deprecation gods have had their fill :)
11:09:41FromGitter<alehander42> the tools which would autotranslate Nim1 to Nim2 would make 2to3 look like a child toy :D
11:10:05FromGitter<mratsim> I think Go1to2 will give us a run for our money
11:10:16FromGitter<mratsim> they don’t have macros.
11:11:08FromGitter<alehander42> oh I don't think macros will be enough :D
11:11:44*r3d9u11 quit (Ping timeout: 260 seconds)
11:11:47FromGitter<mratsim> you can put the old compiler AST parsing as a macro and to AST to AST transformation?
11:11:52FromGitter<alehander42> ok, I'll add a flip branch
11:12:46FromGitter<arnetheduck> @alehander42 check with Araq first maybe, he seemed keen on getting rid of it altogether as well which is why I did the branch
11:12:53FromGitter<alehander42> ok
11:13:48FromGitter<alehander42> well that's what you would do anyway: you'll translate the old AST to newer one
11:14:32FromGitter<alehander42> which can reuse macro's helpers, but I don't think you can qualify it as macro itself :D (or I am missing something)
11:15:29FromGitter<alehander42> anyway converting exception-based code to error-based code will have a lot of gotchas :D
11:15:34Vantage_arnetheduck, araq should know more about the compiler than anyone, if he said it was good then why waste time arguing?
11:16:05Vantage_alehander, counts as a transpiler?
11:16:29FromGitter<alehander42> oh you used the forbidden word :D
11:16:34Vantage_is there an article or a forum post I'm missing here?
11:16:53Vantage_sometimes I wish nim copy-pasted for me
11:17:21FromGitter<alehander42> about what?
11:17:46FromGitter<arnetheduck> @alehander42 apparently that's impossible or at least infeasible - ie translate exceptions to non-exceptions without "special" support - special meaning that `c` is not enough
11:19:10dom96Araq would like to remove exceptions now, but I won't let him :P
11:19:12FromGitter<arnetheduck> bjarne & co tried that with cfront and gave up, per comments on https://stackoverflow.com/a/15731985
11:20:05Vantage_are errors and exceptions the same thing?
11:20:37FromGitter<matrixbot> `ehmry` I'm calling a private procedure from a private template in a module and I get `attempting to call undeclared routine`, how is this possible?
11:20:38FromGitter<arnetheduck> Vantage, to tease out principles and background, and possibly to come up with a practical use case
11:21:39FromGitter<arnetheduck> bikeshedding is great that way.. you don't learn anything about the subject at hand, but everything about what drives people and where they come from
11:21:58Vantage_arne, I think the reason it was still there was because "code dust"
11:22:19*PMunch joined #nim
11:22:28Vantage_"bikeshedding" is a new term for me
11:22:42FromGitter<arnetheduck> "code dust" is new to me :)
11:23:04PMunchbikeshedding is such a nice term
11:23:24Vantage_Its old and had a use but no it doesn't I made the term up
11:25:31FromGitter<arnetheduck> so what does it mean? maybe we can be trend-setters here :)
11:25:41FromGitter<mratsim> here you go: https://en.wikipedia.org/wiki/Law_of_triviality
11:26:28Vantage_mratsim, https://phinze.github.io/2014/05/24/useful-tech-terms-part-1.html
11:27:08FromGitter<matrixbot> `ehmry` vantage_: the source document for bikeshedding is worth reading
11:27:15Yardanicoehmry: pro tip - you can use matrix-irc bridge directly (without gitter)
11:27:15FromGitter<mratsim> @ehmry, template are inserted at the call point, so if the private proc is not visible at that position it cannot be called.
11:27:28FromGitter<matrixbot> `ehmry` I don't know what its called, parkinsons law maybe
11:27:32FromGitter<alehander42> @arnetheduck yeah probably you wouldn't be able to do in general
11:27:40Yardanicoehmry: https://gist.github.com/fstab/ce805d3001600ac147b79d413668770d
11:27:43FromGitter<mratsim> yes Parkinson law of triviality == bikeshedding
11:28:05FromGitter<alehander42> but the problem is there is no easy way to make so big backward-incompatible changes to a language
11:28:31FromGitter<alehander42> the cost is too great for such a benefit :D
11:28:40FromGitter<matrixbot> `ehmry` yardanico: the matrix IRC bridge sucks, thats why I use the gitter. its probably also written in JS though
11:28:50FromGitter<arnetheduck> @mratsim getting rid of exceptions felt useful enough to warrant a forum post and not just a few lines in a chat :)
11:29:14FromGitter<alehander42> the same thing as gc: nogc mode working well and compatible with most of stdlib etc would be amazing
11:29:20FromGitter<mratsim> I think error codes are the dinosaurs :P
11:29:32FromGitter<matrixbot> `ehmry` mamy: but the template should be hidden inside a procedure that is not exported?
11:29:42FromGitter<mratsim> gc: no-gc should work at one point with the —newruntime
11:29:43FromGitter<alehander42> but still I wouldn't want the mental burden of lifetimes/borrows for a lot of stuff
11:29:55FromGitter<matrixbot> `ehmry` er, the procedure is exported that calls the template, not the template
11:29:57*ShalokShalom2 quit (Ping timeout: 240 seconds)
11:30:09Vantage_alehander, I don't know how to "safely" write code with NOGC
11:30:18FromGitter<mratsim> one other thing, template calls must not use the method call syntax
11:30:30*ehmry joined #nim
11:30:32FromGitter<mratsim> you have to do yourtemplate(foo)
11:30:38FromGitter<alehander42> that's one of the reasons I use Nim after all: a huge amount of code doesn't need super-optimized memory handling and performs almost the same way
11:30:46Vantage_is valgrind working with nim.
11:31:00FromGitter<mratsim> no it doesn’t know about Nim allocator
11:31:11Vantage_because often in games, memory-leaks are inevitable
11:31:20FromGitter<arnetheduck> I'm not proposing error codes - I'm all for results however.. only sample code in `rust` is full of `unwrap`, for production code, there are better options and thanks to nim macros you could make it even more transparent
11:31:59FromGitter<mratsim> Ah, good :), if we can have something ergonomic, that is not error code without the issues of Exceptions that would be awesome
11:32:21FromGitter<alehander42> @arnetheduck don't they have `?` now
11:32:46Vantage_but I'd still like to treat Errors like objects :( (Might just be my python speaking)
11:33:34FromGitter<mratsim> @Vantage_ Nim GC shouldn’t leak, compared to C++. Especially if you don’t introduce reference cycles, it wouldn’t have to add the slow cycle detector (mark and sweep like in Java)
11:34:42Vantage_mratsim, sometimes you use external libraries. The way in which you use it might lead to leaks
11:34:46FromGitter<mratsim> I would like a memory pool module in the stdlib though.
11:35:08FromGitter<mratsim> or being able to mix memory regions with normal GC data structures
11:35:13FromGitter<arnetheduck> @alehander42 yeah, that's used for propagating errors up a call stack, and `Nim` could have a `try` equivalent as a block implemented using a macro that would do the same, but be explicit about two things: errors being part of the api of functions, and errors being a little visible in code that calls that API even if it's not handling it
11:35:38FromGitter<alehander42> I agree a code formatter would be nice, still I hate the gofmt "one convention for every possible project" thing, other langs do great with comfigurable formatters(ruby, python etc)
11:35:55Vantage_I think to learn nim first you have to learn how the GC works, then learn all the stuff nim automates
11:36:33*cspar quit (Ping timeout: 248 seconds)
11:36:48FromGitter<mratsim> I think it really depends from where you’re coming from. If you come from Python, you can be happy without think about that. If you come from C/C++ well I suppose you are tremendously curious about that.
11:37:04FromGitter<mratsim> and also what you want to do.
11:37:42FromGitter<alehander42> @arnetheduck yeah, that's what I would imagine, actually I remember developing a similar syntax for `mylanguage compiling to go` before finding out other languages already do that
11:37:43FromGitter<mratsim> Also coming from Python the first thing you have to earn are types :P
11:38:03FromGitter<mratsim> Ok, I’ll try to do something useful now.
11:38:06FromGitter<mratsim> brb
11:38:15FromGitter<arnetheduck> @alehander42 so nim without exceptions but with results would be even more ergonomic than rust with `?`, but still have the advantages
11:38:18FromGitter<alehander42> I also had a weird `escalate <stuff>` keyword that you could use to explicitly bubble some exceptions
11:38:25FromGitter<alehander42> errors*
11:39:22FromGitter<arnetheduck> Vantage, https://forum.nim-lang.org/t/3772
11:40:58FromGitter<alehander42> yeah, basically your thing is something like ⏎ ⏎ ```try: ⏎ a = f(2).g?(4) ⏎ b = a + 2 ⏎ except: ⏎ # do stuff with g's error``` ⏎ ⏎ ? [https://gitter.im/nim-lang/Nim?at=5adc74ca270d7d3708f2753f]
11:41:12FromGitter<mratsim> I think Rust just copy pasted Haskell for their error management: https://wiki.haskell.org/Handling_errors_in_Haskell. But this is much more elegant in Haskell due to the pattern matching and `>>=` (bind) operator.
11:43:05FromGitter<arnetheduck> yeah, there are those that insist that `result` and `either` are monads.. haskell people are so easy to identify from afar :)
11:48:46Vantage_mratsim, GC refs ptrs still affect code, so you still need to learn about it.
12:00:01*SenasOzys quit (Ping timeout: 248 seconds)
12:07:27*arecaceae quit (Remote host closed the connection)
12:07:50*arecaceae joined #nim
12:15:02*SenasOzys joined #nim
12:16:01*Lord_Nightmare quit (Ping timeout: 248 seconds)
12:18:38*dddddd_ joined #nim
12:19:40*yglukhov quit (Remote host closed the connection)
12:20:29*yglukhov joined #nim
12:20:36*dddddd_ is now known as dddddd
12:21:03*Lord_Nightmare joined #nim
12:22:57*Vantage_ is now known as CodeDoes
12:23:35*CodeDoes is now known as vantage
12:24:58*yglukhov quit (Ping timeout: 264 seconds)
12:35:07*nsf joined #nim
12:37:21*simonsolnes quit (Ping timeout: 248 seconds)
12:40:13*jjido joined #nim
12:52:55FromGitter<tim-st> how can I update my nim repo fork to lastest dev version? I have github desktop app
12:56:11vantagegit pull
12:56:25vantagegit checkout origin/devel
12:56:28vantageI THINK
12:56:59FromGitter<tim-st> not possible from gui version?
12:57:09vantageSure
12:57:22FromGitter<ephja> that would make it a terrible application :p
12:57:36vantageopen the git-gui
12:57:54FromGitter<tim-st> I have
12:57:56vantageremote->fetch from->origin
12:58:54FromGitter<tim-st> there is no button `remote`
12:58:56vantagebranch->checkout
12:59:04*miran joined #nim
12:59:12vantageyou have to be inside the repo when you open the git-gui
12:59:24vantagegit for windows right?
13:00:13FromGitter<tim-st> "Github Desktop" v1.1.0
13:00:24vantagelol can't help
13:00:30FromGitter<tim-st> ok, thx
13:00:34vantageI have only the git for windows
13:02:47*cheeseburger joined #nim
13:04:02cheeseburgerHey #nim is it possible to have an optional proc as an argument? Using `customProc: proc=nil` doesn't seem to work
13:04:26FromGitter<zacharycarter> I don't think you can assign nil to a proc
13:04:32cheeseburgerIn the lib I'm writing I want users to be able to optionally pass a different filter proc
13:04:47FromGitter<tim-st> maybe you should default to an no operation proc
13:04:47FromGitter<zacharycarter> probably provide an overload
13:05:17FromGitter<zacharycarter> or that
13:05:23FromGitter<zacharycarter> just a proc that discards the result
13:05:31FromGitter<zacharycarter> so
13:05:39FromGitter<tim-st> Or you can use static declartion an then with `when` you can create this optional param
13:05:39FromGitter<zacharycarter> customProc: proc() = discard;
13:05:41cheeseburgerThanks @zacharycarter
13:05:46FromGitter<zacharycarter> well it was Tim's idea :P
13:05:47FromGitter<zacharycarter> but np
13:05:54Araqvantage: valgrind does work with nim
13:06:06FromGitter<zacharycarter> but only for windows / linux?
13:06:11Araq'koch' knows how to do it, I have forgotten, need to document it
13:06:23FromGitter<zacharycarter> osx valgrind support is stinky
13:06:27cheeseburgerThanks @zacharycarter and @tim-st, I'll try that! :)
13:06:44Araqcheeseburger: customProc: proc() = nil
13:07:09vantagecheeseburger, make a proc that raises an error if its called
13:07:29FromGitter<tim-st> btw @Araq proc param=-1 still doesnt work :( only param = -1
13:07:45Araqwhy would it?
13:08:02vantagethanks Araq
13:08:02Araq=- is an operator on its own.
13:08:37FromGitter<ephja> the standard DOM API is so tedious. gotta add some abstractions
13:09:01Araqephja: use karax?
13:09:02FromGitter<ephja> the JS DOM API that is
13:09:09FromGitter<ephja> yeah maybe
13:11:52FromGitter<tim-st> @Araq because it shouldnt be possible to use this operator inside a proc param declaration
13:12:12FromGitter<tim-st> and then the only logical thing is = -1
13:12:29Araqthat's not how Nim or most programming languages work
13:12:55Araqtokens are not retroactively split into tokens that match "the only logical thing"
13:13:17Araqit's a layered design, tokenization, parsing, semantic checking
13:13:33FromGitter<tim-st> Python: `def test(a=-1): print a`
13:13:43Araqyou cannot divide by a pointer deref in C: a/*b
13:13:54Araqbecause /* starts a comment.
13:14:40FromGitter<ephja> I guess you can't really parse the JS API that well using official documents
13:14:58miranwow, we're under 1300 open issues! when did this happen? :)
13:15:55cheeseburgerI've tried to create a minimal example of what I want to do, if anyone can take a look it's much appreciated: https://ghostbin.com/paste/amgns
13:16:06cheeseburger(Is there a better place to paste code?)
13:16:41FromGitter<ephja> I think the IDL spec is interleaved with html tables and other crap
13:19:54Araqtim-st: Python works the same way, it's just more inconsistent in its operators, =- is not a single token, but -= is.
13:20:34FromGitter<tim-st> ok thanks, just thought it would be more developer friendly
13:20:39Araqevery tool renders it as 'def test(a = -1)' anyway, the spaces won't kill you.
13:20:58FromGitter<tim-st> wasnt sure anymore because nimpretty isnt integrated yet
13:21:11FromGitter<tim-st> and I didnt used pycharm for months
13:21:16Araqcomments are nasty beasts :-)
13:25:41federico3any way to import procs that have no asterisk?
13:33:45*simonsolnes joined #nim
13:34:01FromGitter<ephja> https://github.com/Microsoft/TypeScript/wiki/Using-the-Compiler-API that might be the best way to generate complete bindings
13:36:07Araqfederico3: include instead of import? :-)
13:36:32*user1101 joined #nim
13:37:20FromGitter<arnetheduck> wooohoo, thanks Araq, there's another feature to kill, `include`!
13:38:00Araqyou can't be serious.
13:40:11FromGitter<ephja> 🐤
13:41:36FromGitter<arnetheduck> totally, that's a nasty beast if there ever was one... incomplete and order-dependent snippets of code?
13:41:53*yglukhov joined #nim
13:45:12Araqsure let's write parseStmt(staticRead"file.nim") instead
13:45:47Araqoh I know, these should be removed too then because they can emulate 'include'
13:46:48FromGitter<arnetheduck> well, at least it's complex enough that no one in their right mind will use it, except as a last possible resort.. bad ideas should be hard
13:48:52Araqand it's bad because you say so, got it. :P
13:49:01FromGitter<arnetheduck> of coz ;)
13:49:18AraqI find it tremendously useful and it doesn't add much complexity to the compiler.
13:49:50Araqyou have my sympathy about removing exceptions though.
13:49:55FromGitter<arnetheduck> you know, sometimes there's a reason certain things disappear from newer languages that existed in old ones ;)
13:50:36Araqlike C# had to add partial classes for code generators.
13:50:40Araq;-)
13:51:48Araq"old" is not a criterion for anything.
13:52:10FromGitter<arnetheduck> well, looking at the compiler cgen, that's not how includes are used..
13:52:54Araqyeah, they are mostly a workaround for cyclic module dependencies
13:52:58Araqin the compiler.
13:53:13Araqin the stdlib they are used to split interfaces and implementation
13:53:28Araqand in fact should be used *more*, not *less* in the stdlib.
13:54:11FromGitter<arnetheduck> oh dear. more forward declarations? I was just about to add that to the list 🔥
13:54:27*yglukhov quit (Remote host closed the connection)
13:54:33Araqforward decls are one thing, recursive module dependencies another.
13:55:18FromGitter<arnetheduck> recursive module deps, that's a code smell if there ever was one 💩 and repeating the same thing twice? even vim does code folding these days..
13:55:48Araqrequired code folding is a code smell :P
13:56:02AraqI never use it, it always folds what I'm interested in.
13:56:22Araqthough that can probably be solved with smarter editors
13:58:04vantagehow do I include a folder in dll searching?
13:58:33vantage`--passL:"/LIBPATH:~/libs"` doesn't work for me
13:58:45FromGitter<zacharycarter> should just be
13:58:54FromGitter<zacharycarter> --passL: "~/libs" I think
13:58:59FromGitter<zacharycarter> you don't need LIBPATH:
13:59:32Araqfwiw, I want to remove exceptions, .tags, iterators, typed vs untyped, methods, closures, the GC and converters.
13:59:41Araq:P
13:59:57FromGitter<zacharycarter> I like all those things
13:59:57vantageTHank ya
14:00:04vantageforgot the space
14:00:04FromGitter<zacharycarter> I don't like exceptions
14:00:07FromGitter<arnetheduck> put that in the thread for posterity :)
14:00:08FromGitter<mratsim> includes are useful to test private proc. Also I use them when I directly {.emit.} C code because I get code duplication or whatever otherwise
14:00:09FromGitter<zacharycarter> I don't know what .tags are
14:00:26FromGitter<mratsim> tags are part of the Effect system
14:00:26vantagedoes that work with windows though
14:00:29FromGitter<zacharycarter> what is going to replace the GC?
14:00:38FromGitter<mratsim> Terminators
14:00:42FromGitter<zacharycarter> oooo
14:00:43Araqthe "new runtime".
14:00:54cheeseburgerWhen I pass "proc(item1, item2: T): T = nil" as a parameter type, what is the "nil" on the end doing? If I change "nil" to "item1" the compiler says it's an "undeclared identifier"
14:01:00vantageAraq, Want to make RUST MINUS?
14:01:21FromGitter<tim-st> Do you want to replace gc like rust does it? I think rust is quite difficult with the lifetime params
14:01:35FromGitter<zacharycarter> I don't think he's talking about an ownership system
14:01:36FromGitter<mratsim> @cheeseburg, nil assign a nil pointer to a ref type
14:01:44FromGitter<zacharycarter> but I could be totally wrong
14:01:49FromGitter<tim-st> how then?
14:01:51FromGitter<mratsim> @zacharycarter you didn’t read about lent type.
14:01:58FromGitter<mratsim> that’s like inverse borrowing
14:02:03FromGitter<zacharycarter> ah
14:02:05AraqI'm only half-serious.
14:02:11FromGitter<zacharycarter> I saw this last night
14:02:14FromGitter<zacharycarter> don't know if it's intersting at all or not
14:02:20FromGitter<zacharycarter> http://www.cs.columbia.edu/~simha/preprint_isca18_REST_memory_safety.pdf
14:02:29Araqthe features will all replaced by macros
14:02:30FromGitter<zacharycarter> or if everyone already knows about this and I'm just slow
14:02:32cheeseburger@mratsim what if the return type isn't known to be a ref type or not?
14:02:33FromGitter<zacharycarter> which is probably the case
14:02:38Araq*would all be replaced by macro solutions.
14:02:59FromGitter<mratsim> It is, for the escape analysis,/optimization opportunities that can come from it. I don’t like how seq can copy under your feet.
14:03:11FromGitter<mratsim> @cheeseburger: compile-time error
14:03:47cheeseburger@mratsim ok makes sense. Why can it be "nil" but not any other identifier?
14:03:52FromGitter<mratsim> you proably should put `T: not(ref|ptr|seq|tring)` in your signature though
14:04:20FromGitter<mratsim> nil is the keyword for null in Nim
14:04:24FromGitter<ephja> Araq: remove iterators from the language and replace with concepts?
14:04:31*yglukhov joined #nim
14:04:43FromGitter<mratsim> iterators are one of the best thing in Nim
14:04:53Araqthat's impossible, but iterators can be based on macros.
14:05:23FromGitter<tim-st> as long as the syntax will be the same. In rust many people didnt like their iterators and want python like iterators
14:05:47Araqanyway, calm down, relax. removing/replacing a feature is hard work
14:05:53*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
14:06:03FromGitter<mratsim> Rust iterators are fast: https://github.com/alehander42/zero-functional/tree/master/benchmarks
14:06:30Araqand every single item on my list is more a research project than a "solution" waiting to be implemented.
14:07:04FromGitter<mratsim> The “hidden” research is done there :D https://github.com/Araq/nox
14:07:22FromGitter<ephja> impossible because of the coroutine stuff or because of some other reason?
14:07:28miran"This repository is empty." ?
14:07:31FromGitter<ephja> actually, my POC defined a foreach macro
14:07:32Araqyes, even removing exceptions, you will be surprised to learn Nim's iterators do require exceptions
14:07:57FromGitter<mratsim> @miran it had “This is research for next features in Nim” at one point
14:08:01Araqor a comparable "on scope exit cleanup" feature.
14:08:30miranmratsim: i have seen that description, but - tha doesn't change that it is empty :)
14:08:41FromGitter<ephja> replacing exceptions with something else actually seems like a decent idea when the language in question actually has modern features
14:09:08cheeseburgerI can't get the optional proc argument thing to work. Is there another way of doing it I can look at?
14:09:21*yglukhov quit (Ping timeout: 264 seconds)
14:10:08Araqproc foo(opt: proc () = nil) =
14:10:16Araq if opt != nil: opt()
14:10:24FromGitter<ephja> it would be a good idea to upload a test case and the relevant output to gist or some other service
14:11:23cheeseburgerCool I've updated the paste with the output: https://ghostbin.com/paste/amgns
14:11:47Araqbtw you can map exceptions to 'if's, it's not pretty but it works and the result is what otherwise would have been hand coded.
14:12:00cheeseburger@Araq I'm using your example but no luck integrating it
14:12:20Araqexcept that it's slower and has much worse interop with JS and C++.
14:12:42*yglukhov joined #nim
14:13:44*vantage quit (Quit: Leaving)
14:14:46FromGitter<ephja> yes that or a default proc. btw, the earlier example was incomplete. you need to do `foo: proc() = (proc() = discard)`. discard alone won't make it a proc literal. the argument type can be omitted (inferred) in this case since a default value is provided
14:15:35*vantage joined #nim
14:16:02cheeseburgerHmm @ephja that gives me a "type mismatch"
14:16:20cheeseburgerBut that helps me understand what you were saying, thanks :)
14:16:47cheeseburgerI think I misunderstood the proc type argument
14:18:00FromGitter<mratsim> @arnetheduck thoughts about th `+%`, `-%` … operators?
14:18:17cheeseburgerOk it works now! Thanks for the help everyone. I am still learning nim
14:18:21cheeseburgerI've updated the paste
14:18:24FromGitter<mratsim> maybe those should be in a pointer arithmetic package
14:18:38cheeseburgerNow I have another question, which was the bug that I fixed
14:19:06mirancheeseburger: what was updated? where was the error?
14:19:08cheeseburgerThe proc definition was "proc doSomeStuffWithOptionalFilter[T] (items: seq[T], customProc: proc(item1, item2: T): T = nil) ="
14:19:15*Vantage_ joined #nim
14:19:32cheeseburgerAnd I was calling it with this: "items.doSomeStuffWithOptionalFilter[CustomItem]()"
14:19:38cheeseburgerBut that doesn't work
14:19:49cheeseburgerI had to remove the "[CustomItem]" bit
14:19:58cheeseburgerSo it's just "items.doSomeStuffWithOptionalFilter()"
14:20:01*xkapastel joined #nim
14:20:16cheeseburgerI suppose it is getting "T" from the type contained in the seq that was passed
14:20:47cheeseburgerBut why would specifying the type cause this?: "Error: expression 'doSomeStuffWithOptionalFilter(items, nil)' has no type (or is ambiguous)"
14:21:00cheeseburgerAs I am specifying the type
14:23:58FromGitter<arnetheduck> @mratsim no opinion there, I've worked with both and it's fine.. except that having both unsigned types and unsigned operators for signed types is.. well you've heard the rest of the story, useless options -> useless cognitive and other costs
14:25:24FromGitter<arnetheduck> or rather, not enough information to have an opinion.. I could think it's mostly a matter of habit, which way you prefer. unsigned/wrapping arithmetic is generally useful for more than just pointers btw
14:26:56Vantage_Is half the language gonna be macros now ???
14:27:41FromGitter<ephja> should have thought about all this in 2008 :p
14:29:58FromGitter<mratsim> @Vantage_ it’s better than having everything in the compiler actually. I really don’t like to look into the Nim code and see {.magic: “FooYouCannotReplicateTooBad”.}
14:31:27FromGitter<mratsim> Also that would mean that if you need a very advanced feature, you can implement it yourself, without compiler plugins or whatnot, distribute it yourself as well and you don’t have to wait for the core devs
14:31:30Vantage_mratsim I'd like less magic
14:31:34*yglukhov quit (Remote host closed the connection)
14:31:35Vantage_I think we all would
14:31:43FromGitter<mratsim> also it’s less surface to maintain for the core devs
14:32:22AraqVantage_: no, not "now", it's a long term vision.
14:33:10Vantage_what would define 1.0.0
14:33:29Araq1. bugfixes.
14:33:41Araq2. features behind a more fine-grained .experimental switch.
14:34:03PMunchShould probably resolve some of the large RFCs as well no?
14:34:23Araqso that .experimental is not an all-or-nothing switch.
14:34:48AraqRFCs .. which ones?
14:36:17PMunchI haven't looked through all of them, but for example the conv/to discussion and generic `new` procs are quite a big, but simple to implement, change to the stdlib
14:37:11Araqnot a fan of these RFCs.
14:37:29PMunchWhy not?
14:38:01Araqit's just more bikeshedding.
14:38:26PMunchNot really, the generics offer a completely different/new set of features
14:38:43PMunchWhich Nim currently doesn't have, and not really for a good reason
14:38:49Araqno.
14:38:59Araq1. it's entirely possible macros to get the same.
14:39:02PMunchIMHO it would also make the stdlib look more cohesive and thought through..
14:39:13Araq2. the underlying analysis is weak.
14:39:16PMunchYeah, but not procs or templates..
14:39:21PMunchWeak how?
14:39:40PMunchAnd what analysis?
14:39:53Araqyeah, exactly. what analysis.
14:40:16PMunch-_- I meant what do you want analysed?
14:40:58FromGitter<mratsim> conv/to? I think I missed this one
14:41:02Araqcosts vs benefits?
14:42:20Araqinit[Table[string, int]]() vs initTable[string ,int]()
14:42:49PMunchOops, got to go
14:42:49*PMunch quit (Quit: leaving)
14:42:55Araqyou claim it makes generic code easier to write, I claim it makes the common thing more cumbersome and generic code is possible already.
14:56:49FromGitter<ephja> the only thing that would add more convenience is return type inference like with @[...] imo, as long as you can't override the default constructor
14:57:30FromGitter<ephja> I don't mind having most things as "regular" parameters though. it's fairly obvious when a type is passed in rather than an object
14:58:48Vantage_http://wiki.c2.com/?EconomicsOfRefactoring
15:00:09FromGitter<ephja> maybe the contract stuff you considered implementing could be used for implementing user range types, and then it would be easy to support float ranges as well
15:00:50FromGitter<ephja> I guess you could do that now but with only runtime checks
15:02:01FromGitter<mratsim> I don’t see how you can make range type with compile-time range checks
15:04:43FromGitter<ephja> contracts could be checked at compile time if possible
15:04:56Araqis this a joke, Vantage_ ?
15:05:34Vantage_?
15:05:58Araq"Economics Of Refactoring" -- where is the economics? it's just same hand waving about how great XP is.
15:09:04Vantage_Best I can suggest is try to find a way to do it a little at a time. This way the customer still gets new functionality and the code gets better slowly.
15:09:04Vantage_Perhaps there are times when it makes sense to do it all in one big, focused effort. I don't know.
15:09:07Vantage_I guess the point of the article is it's better to make small intentional changes than big changes
15:09:52Vantage_http://wiki.c2.com/?ExtremeProgramming
15:33:29*SenasOzys quit (Ping timeout: 268 seconds)
15:33:30*cheeseburger quit (Ping timeout: 260 seconds)
15:40:02*edcragg quit (Quit: ZNC - http://znc.in)
15:41:00*edcragg joined #nim
15:48:40*SenasOzys joined #nim
15:49:55FromDiscord<geckojsc> is it possible to define a macro of this form:
15:49:55FromDiscord<geckojsc> mymacro foo:
15:49:55FromDiscord<geckojsc> a = 10
15:49:56FromDiscord<geckojsc> bar()
15:49:56FromDiscord<geckojsc> other expressions...
15:50:15FromDiscord<geckojsc> where foo is any valid identifier
15:52:15Yardanicoyes, why not? :)
15:52:44Vantage_geckojsc show us how you'd want to use it
15:53:35FromDiscord<geckojsc> I want to do something like
15:53:35FromDiscord<geckojsc> with p:
15:53:35FromDiscord<geckojsc> x = 10
15:53:35FromDiscord<geckojsc> y = 20
15:53:45FromDiscord<geckojsc> instead of
15:53:45FromDiscord<geckojsc> p.x = 10
15:53:45FromDiscord<geckojsc> p.y = 20
15:54:06FromDiscord<geckojsc> I'm totally new to this macro thing though, don't really know where to start xD
15:54:28YardanicoDo you want to do that inside of a proc which takes an object as a first argument?
15:58:31FromDiscord<geckojsc> I'm not sure xD would that affect the way I define it? Ideally it would work on any object regardless of where it's defined
15:58:34*JStoker quit (Killed (wolfe.freenode.net (Nickname regained by services)))
15:58:40*JStoker joined #nim
15:58:42FromGitter<alehander42> Araq what is the initTable example?
15:58:46FromGitter<alehander42> it doesnt render well in gitter
15:59:40*petersjt014[m] quit (*.net *.split)
15:59:40*Demos[m] quit (*.net *.split)
15:59:57FromGitter<alehander42> anyway iirc your argument for the rfc was that you want the stlib to not depend on all language advanced concepts , but generics can't be possibly included in those
16:01:26*dddddd quit (Remote host closed the connection)
16:02:49*beatmox quit (Ping timeout: 260 seconds)
16:03:11*beatmox joined #nim
16:03:31FromDiscord<claude> there's also the `{.this: p.}` pragma which works for the rest of the module
16:03:42FromGitter<alehander42> https://scontent.fsof2-1.fna.fbcdn.net/v/t1.0-9/31124539_1931351253543941_2586509720323883008_o.jpg?_nc_cat=0&oh=128cd57fccbceeb9a73a1b5b04fbaa83&oe=5B4FD1BB :D :D
16:05:13Vantage_doesn't the {.this: self.} pragma only work inside procs? claude
16:06:57FromDiscord<claude> yes
16:07:33*petersjt014[m] joined #nim
16:11:10FromGitter<mratsim> @geckojsc, the easiest solution is with templates: `template x() = p.x` though you might get recursive template replacmeent of x: example (that gets recursive replacement :/: https://github.com/numforge/number-theory/blob/e11ea81ae8aad7833468f1ffad953277f40c329b/src/modular_arithmetic.nim#L130-L142
16:11:30Vantage_alehander42 lol. "ONly a mother could love it"
16:12:41FromGitter<alehander42> srsly its not so bad, weird quote
16:13:12*Yardanico quit (Ping timeout: 260 seconds)
16:13:43FromGitter<mratsim> and otherwise @geckojsc, here is a macro that can replace a list of identifiers by a list of other NimNodes: https://github.com/status-im/nim-mpint/blob/d5794ffbe1717f4eba7c92794a916ff0d614b619/src/private/as_words.nim#L33-L54
16:15:10FromGitter<mratsim> what you’re looking for is probably to replace ident”x” by DotExpr.newTree(ident”p”, ident”x”)
16:15:22FromGitter<mratsim> nnkDotExpr* sorry
16:17:42*Trustable joined #nim
16:17:47FromGitter<mratsim> or just use: let replacement = quote do: `p`.`ident($x)` (you need to convert the input x, to its string symbol and recreate an ident from this I think.
16:18:45FromGitter<mratsim> but try templates first ;)
16:19:25FromDiscord<geckojsc> the `this` pragma sounds interesting, is it documented anywhere?
16:21:04FromGitter<mratsim> in the manual ;)
16:21:19FromGitter<mratsim> it doesn’t work with Generics and static though :/.
16:24:21*Vantage_ quit (Ping timeout: 264 seconds)
16:25:03*leorize joined #nim
16:25:30*Demos[m] joined #nim
16:26:38FromDiscord<geckojsc> ahh i think that's just what I was looking for tbh :)
16:28:01*nc-x joined #nim
16:28:48*solitudesf quit (Ping timeout: 255 seconds)
16:28:50*TheManiac quit (Ping timeout: 240 seconds)
16:28:51*libman[m] quit (Ping timeout: 240 seconds)
16:28:51*Abnegation quit (Ping timeout: 240 seconds)
16:28:51*unclechu quit (Ping timeout: 240 seconds)
16:28:52*notdekka[m] quit (Ping timeout: 240 seconds)
16:28:57*petersjt014[m] quit (Ping timeout: 245 seconds)
16:28:57*Miguelngel[m] quit (Ping timeout: 245 seconds)
16:28:58*planetis[m] quit (Ping timeout: 245 seconds)
16:28:59*narimiran[m] quit (Ping timeout: 245 seconds)
16:28:59*macsek1911[m] quit (Ping timeout: 245 seconds)
16:29:05*hohlerde quit (Ping timeout: 258 seconds)
16:29:05*stisa quit (Ping timeout: 258 seconds)
16:29:07*dyce[m] quit (Ping timeout: 246 seconds)
16:29:15*survivorm[m] quit (Ping timeout: 255 seconds)
16:29:15*zielmicha[m] quit (Ping timeout: 255 seconds)
16:29:21*byteflame quit (Ping timeout: 240 seconds)
16:29:21*mgdelacroix[m] quit (Ping timeout: 240 seconds)
16:29:30*Demos[m] quit (Ping timeout: 256 seconds)
16:29:35*amp907[m] quit (Ping timeout: 276 seconds)
16:29:42*pqflx3[m] quit (Ping timeout: 255 seconds)
16:29:46*krux02[m] quit (Ping timeout: 265 seconds)
16:30:37nc-xWhat is the reasoning behind " the stlib to not depend on all language advanced concepts ". Is it for ease of external contributors? I mean if a person has the skills and knowledge to for eg. write a json parser, I don't understand why he won't be able to understand generics or typedescs or concepts or any other such feature.
16:31:04nc-xOn the other hand reduction of copy paste code could lead to lesser bugs.
16:31:30nc-xAnd ease of maintability and contribution even for external contributors.
16:33:19*Vantage_ joined #nim
16:33:36FromGitter<mratsim> are you sure about the stdlib and not just system.nim
16:34:10FromGitter<mratsim> I think restrictions was more for gc:none or os:standalone compilation.
16:34:13FromGitter<mratsim> were*
16:34:38*nc-x_ joined #nim
16:34:59FromGitter<mratsim> and “advanced concepts” are macros i think, not generics.
16:35:55*nc-x quit (Ping timeout: 260 seconds)
16:36:42nc-x_mratsim: @Alehander42 and I are talking about the RFC's regarding the Stdlib API's to provide a generic interface for which Araq seems to be against.
16:37:15FromGitter<mratsim> ah
16:37:30*nc-x_ is now known as nc-x
16:47:19nc-xAs for the recent talks regarding removing Exceptions, IMHO it would be really great if that happens (I don't like exceptions, heck I would even prefer Go's error handling to exceptions (but Rust's is preferable :) ). But I guess it's too late now :(
16:50:35*Vantage_ quit (Read error: Connection reset by peer)
16:51:07vantagewhats the differency between exceptions and errors
16:51:39*nc-x quit (Ping timeout: 260 seconds)
16:54:35*Vantage_ joined #nim
16:59:12FromGitter<mratsim> you have your code littered with `let err = yourcode(foo, bar, bar); case err of FirstErr: … of SecondErr: …` breaking reading and program flow
16:59:54miranin python you can slice a container like `a[3:50:7]`, but in nim slices can only continous, with step=1?
17:00:05miran*only be
17:00:39FromGitter<mratsim> if it’s for an iterator use countup or countdown
17:01:12Vantage_mratsim isn't that an error?
17:01:24Vantage_What would python exceptions be then
17:01:28mirani know for those two, but i cannot do `a[countup(3, 50, 7)]`
17:01:28Vantage_an error or an exception
17:01:41FromGitter<mratsim> Python exceptions are exceptions
17:02:29FromGitter<mratsim> basically when something isn’t suppose to happen, it breaks the program immediately, but this allows you to chain functions, ie. Error codes don’t monopolize the function return values
17:03:10FromGitter<mratsim> otherwise you always have to do `var foo: Foo; let err = doSomething(foo, param)`
17:03:39FromGitter<mratsim> instead of `let foo = doSomethin(param)`
17:04:02FromGitter<mratsim> the main issue with exceptions is when you want to do a recoverable library or application
17:04:24Vantage_I would rather want to avoid always having to do let err=
17:05:43Vantage_if its just an implementation detail then I don't care. But if it actually changes the way we need to code then I'd have to reevaluate
17:06:03FromGitter<mratsim> It’s pervasive. That impacts everything.
17:06:33Vantage_well that's gonna be something controversial
17:06:49miranyou speak like this is going to happen? is it?
17:06:53Vantage_araq said he wants to use the power of macros to ease it
17:07:02FromGitter<mratsim> I see error code as `let err = div(a, 0); if err == 1: {.error: “You divided by 0”.}` ugh ...
17:07:14Vantage_Maybe "IN the future"
17:07:39Vantage_I like My try except or try catch
17:08:11FromGitter<mratsim> Yeah but, if you don’t exhaustively try you might not be able to recover properly
17:08:21Vantage_An Error "indicates serious problems that a reasonable application should not try to catch." An Exception "indicates conditions that a reasonable application might want to catch." Error along with RuntimeException & their subclasses are unchecked exceptions. All other Exception classes are checked exceptions. SAYS GOOGLE
17:08:28FromGitter<mratsim> when interacting with say databases, you might leave it in inconsistent state
17:08:52*rauss quit (Read error: Connection reset by peer)
17:08:57FromGitter<mratsim> Google designed Go … so I don’t believe them when talking about language designs :P
17:09:25Vantage_No I googled "error vs exception"
17:09:32Vantage_that was the result
17:10:04*rauss joined #nim
17:10:39FromGitter<Vindaar> @miran: in case you're fine with a copy of the stepped slice, this is what I use: http://ix.io/18sf
17:11:42miranyeah, i thought there might be some "automatic" "sugary" way :)
17:12:13FromGitter<Vindaar> yep, I know :)
17:12:51Vantage_Can you overload the operator `..` ?
17:13:48Vantage_https://rosettacode.org/wiki/Loops/For_with_a_specified_step#Nim
17:13:56Vantage_for x in countdown(10,0,3): echo(x)
17:14:06Vantage_isn't that what you wanted?
17:14:33miranthis is what Araq has posted before on a similar topic: https://gist.github.com/Araq/701565f110abd65801fbaea03143c24d
17:14:52mirani would really like to see this `a..b|c` syntax
17:15:11Vantage_thanx miran
17:22:12*Yardanico joined #nim
17:23:23FromGitter<mratsim> @miran this is what I use in Arraymancer: https://github.com/mratsim/Arraymancer/blob/master/src/tensor/accessors_macros_syntax.nim#L59
17:25:36miranmratsim: "step can be negative" is much more logical than `countdown(50, 3, 7)`, IMO
17:26:34FromGitter<mratsim> I’m for having this explicit. This creates silent bugs in Python code
17:26:51FromGitter<mratsim> That’s the whole point of having a statically typed language I think
17:27:41*r3d9u11 joined #nim
17:28:17FromGitter<mratsim> I mean, likere there: https://github.com/mratsim/Arraymancer/blob/master/src/tensor/accessors_macros_syntax.nim#L209
17:28:21FromGitter<mratsim> like*
17:28:43FromGitter<mratsim> if there is no -stepping I don’t “countdown”.
17:28:54FromGitter<mratsim> but in Nim there is no step so the question doesn’t arise ;)
17:29:15FromGitter<mratsim> -stepping ==> negative stepping.
17:29:46miranbasically: `5..0` -> nothing; `5..0|-1` -> six elements
17:29:57FromGitter<mratsim> and note that it checks `|-` it doesn’t check if the step is positive or negative (which is intended but might be a gotcha)
17:30:12FromGitter<mratsim> yes
17:30:25FromGitter<mratsim> actually it’s an error
17:31:19FromGitter<mratsim> https://github.com/mratsim/Arraymancer/blob/d774881059b1598a3073fede1b28a45731674e6f/src/tensor/private/p_checks.nim#L58-L73
17:35:48miranthis should be part of nim syntax, much nicer and more elegant than countdown/countup
17:37:05Vantage_He could always do `i%3 == 0`
17:46:28FromGitter<mratsim> @miran, I’m sure Araq wants more RFCs in the tracker :D
17:46:41miran...and mentions on IRC :P
17:47:02FromGitter<mratsim> I made sure to put an uppercase, it’s not the same identifier
17:49:30*yglukhov joined #nim
17:50:53*jjido joined #nim
17:51:15*SenasOzys__ joined #nim
17:51:26*r3d9u11 quit (Remote host closed the connection)
17:52:33*SenasOzys quit (Read error: Connection reset by peer)
17:52:46Vantage_https://gist.github.com/CodeDoes/8d071ade52c431d14327d21f4cb5f5f7
17:52:51FromGitter<alehander42> Iraq
17:53:20Vantage_Are RFC refering to Request for Comment or something else
17:53:55*yglukhov quit (Ping timeout: 256 seconds)
17:54:04FromGitter<alehander42> oh we really have to make json's `%` work with variant objects
17:58:03FromGitter<alehander42> oh tables too
17:58:04FromGitter<alehander42> why
17:59:14Vantage_why what?
18:00:17FromGitter<alehander42> ok, it seemed easy, weird
18:00:33FromGitter<alehander42> @dom96 are you ok with a PR adding default `%` for Table[K, V] ?
18:00:59FromGitter<alehander42> variant objects seem to work, I was mistaken initially
18:01:20dom96sure
18:05:21*simonsolnes quit (Ping timeout: 248 seconds)
18:13:08vegax87do you know if is possible to gather values from a txt just like numpy.loadtxt() with Arraymancer?
18:13:29*yglukhov joined #nim
18:14:28FromGitter<GULPF> @alehander42 I have a PR for that, #7203
18:16:24FromGitter<alehander42> @GULPF awesome!
18:18:48*yglukhov quit (Read error: Connection reset by peer)
18:19:23*yglukhov joined #nim
18:22:37*simonsolnes joined #nim
18:31:51FromGitter<tim-st> When I try to load a dll and the error is `could not load: dllname.dll` does it mean the file is found but somehow of wrong format?
18:35:02Yardanicofirstly check for 32/64bit
18:36:02FromGitter<tim-st> I only found a 64bit version^^
18:36:34YardanicoI mean, your program and your dll both should have same architecture
18:36:49FromGitter<tim-st> they have
18:37:20FromGitter<mratsim> @vegax87, there is a read_csv: https://github.com/mratsim/Arraymancer/blob/master/src/io/io_csv.nim
18:37:53Vantage_tim-st what dll?
18:37:57FromGitter<tim-st> @mratsim can I use this file with your wrapper?https://github.com/facebook/rocksdb/issues/2531
18:38:16Vantage_I once tried sdl_gfx.dll and it was invalid or the wrong version
18:38:44Vantage_with a similar message
18:38:47FromGitter<tim-st> I try to use the rocksdb nim wrapper with the dll I linked
18:39:00FromGitter<tim-st> I already got leveldb working without problems
18:39:24FromGitter<mratsim> @tim-st, I only tried on OSX, and it’s a 1-to-1 wrapper so if the original lib doesn’t work, the wrapper won't
18:39:24Vantage_if you totally take away the dll it shouldn't even be able to compile
18:39:42Vantage_and if you get to that message it should be that the dll might be invalid
18:40:34FromGitter<tim-st> @mratsim ok thanks, will try other dll versions
18:41:09FromGitter<tim-st> The dll is compiled by facebook, I think it should be ok
18:42:34Vantage_right os?
18:43:04FromGitter<tim-st> I am on windows 10 x64 I would say it's the correct version: https://repo1.maven.org/maven2/org/rocksdb/rocksdbjni/5.11.3/
18:43:19FromGitter<tim-st> I renamed it to `librocksdb.dll`
18:53:39*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
18:57:13*elrood joined #nim
18:58:38*jjido joined #nim
18:59:39*Vantage_ quit (Quit: Leaving)
19:02:49*Vantage_ joined #nim
19:04:47*jjido quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:07:24*rokups quit (Quit: Connection closed for inactivity)
19:14:09*jjido joined #nim
19:16:51*jjido quit (Client Quit)
19:22:00*CodeVantage joined #nim
19:22:06*vantage quit (Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org)
19:22:19*Vantage_ quit (Ping timeout: 256 seconds)
19:22:40*CodeVantage quit (Client Quit)
19:22:53*Vantage joined #nim
19:24:43*CodeVantage joined #nim
19:24:44*Vantage quit (Client Quit)
19:24:44*CodeVantage quit (Client Quit)
19:24:57*CodeVantage joined #nim
19:33:39*pqflx3[m] joined #nim
19:38:21*miran quit (Quit: Konversation terminated!)
19:41:52FromGitter<mratsim> what is this nimnomagic64 define in system.nim?
19:48:05*byteflame joined #nim
19:48:05*Abnegation joined #nim
19:48:05FromGitter<binary132> can anyone comment on a) the usable portability of nim between major OS, b) the same but particularly to do with debugging?
19:48:06*dyce[m] joined #nim
19:48:06*Miguelngel[m] joined #nim
19:48:06*unclechu joined #nim
19:48:06*solitudesf joined #nim
19:48:06*Demos[m] joined #nim
19:48:07*mgdelacroix[m] joined #nim
19:48:12*macsek1911[m] joined #nim
19:48:12*libman[m] joined #nim
19:48:12*TheManiac joined #nim
19:48:13*narimiran[m] joined #nim
19:48:13*petersjt014[m] joined #nim
19:48:14*zielmicha[m]1 joined #nim
19:48:14*amp907[m] joined #nim
19:48:14*notdekka[m] joined #nim
19:48:14*survivorm[m] joined #nim
19:48:14*planetis[m] joined #nim
19:48:15*stisa joined #nim
19:48:15*krux02[m] joined #nim
19:48:16*hohlerde joined #nim
19:48:28FromGitter<binary132> one of the major drawbacks for me to get into Rust is that its debugging and cross-platform support are not that good
19:50:43FromGitter<mratsim> @binary132 Cross platform support is top notch (unless you are looking for GUI programming but most languages just give up on that)
19:51:20FromGitter<binary132> gui doesn't seem impossible with some of the decent C simple gui solutions
19:51:45FromGitter<alehander42> @binary132 you can use most c-lang debuggers for Nim as it generates line directives
19:51:48FromGitter<mratsim> For debugging, if you stay in pure Nim code, once you get used to the terse error messages it’s not bad
19:52:00dom96huh, I'm curious, in what way is Rust not cross-platform?
19:52:03FromGitter<alehander42> there are also js sourcemaps in the works for the js backend
19:52:03FromGitter<mratsim> also yes, you can use GDB/LLDB for debugging
19:52:16FromGitter<alehander42> and there is a time-travelling visual nim debugger in the works
19:52:36FromGitter<mratsim> here is a tuto for GDB/LLDB with a view on Nim stack: https://nim-lang.org/blog/2017/10/02/documenting-profiling-and-debugging-nim-code.html
19:53:05FromGitter<mratsim> Also the stack traces are quite good
19:53:16FromGitter<mratsim> it’s not Elm but we’re getting there :D
19:55:00FromGitter<binary132> well I don't want Elm :P
19:55:11FromGitter<binary132> thx @mratsim , that's handy
19:55:22dom96I know Elm's compile-time errors are good, but are its stack traces that good too?
19:55:24FromGitter<binary132> wow, cool @alehander42 :)
19:55:39FromGitter<mratsim> ah no, I don’t know about Elm stack traces.
19:56:23FromGitter<mratsim> I like Nim ones
19:56:47dom96Yeah, hard to get better than Nim/Python's stack traces
19:57:38FromGitter<alehander42> yeah, I'd like to see those elm traces :D
19:57:57FromGitter<mratsim> Python’s take a lot of screen space sometimes. And for reason unknown (static, generic, templates?) I’ve seen some cases where Nim’s skips one call or template
20:00:45FromGitter<binary132> Go stack traces, duh
20:01:04FromGitter<binary132> who doesn't love picking apart a maze of coroutine references :)
20:01:16FromGitter<binary132> oh, on that note, does nim have coroutines?
20:04:52dom96Kind of. :)
20:05:08dom96We do have proper coroutines but judging by their usage they will be removed from the stdlib.
20:05:28dom96We also have "closure iterators" which come close and are used by Nim's async await functionality.
20:05:34*xet7 quit (Remote host closed the connection)
20:05:47FromGitter<binary132> ah, that's a bit like Rust's experimental concurrency api, no?
20:08:11dom96It's exactly like Rust's concurrency API :P
20:08:23dom96Nim's async await is also implemented using macros
20:08:38dom96The difference is: it's already usable and has been for years
20:08:47dom96Used in production too
20:11:09*Lord_Nightmare quit (Quit: ZNC - http://znc.in)
20:11:39*Lord_Nightmare joined #nim
20:16:28*smt quit (Read error: Connection reset by peer)
20:17:40*smt joined #nim
20:28:25*adeohluwa joined #nim
20:28:45adeohluwahttps://usercontent.irccloud-cdn.com/file/BEcSP6OT/Screenshot_20180422-212756.png
20:30:00adeohluwatrying to install nimlang on termux
20:30:07adeohluwaLmao!
20:31:16dom96Is that devel?
20:31:31dom96Somebody did get it running on termux
20:31:42dom96Wonder if they just didn't make a PR for the changes they had to make
20:31:46FromGitter<alehander42> hahaha wow
20:31:49*nim joined #nim
20:32:01nimhi
20:32:10adeohluwaso far
20:32:17dom96nim: creative nickname :)
20:32:19FromGitter<alehander42> oh it became sentient !
20:32:21adeohluwa$pkg install clang
20:32:26adeohluwato add gcc
20:32:31nimwhy nim not popular yet?
20:32:45adeohluwadom96: what's devel?
20:32:55nimand its really perf like c/rust - in real app?
20:33:26dom96adeohluwa: The git HEAD
20:33:27adeohluwanim: I would say because its new & not yet backed by super large organizations
20:33:48nimits from 2008 not ?
20:34:02nim10 year age..
20:34:56adeohluwadom96: yes it is
20:35:02*Trustable quit (Remote host closed the connection)
20:35:34*nim quit (Client Quit)
20:35:45dom96Pretty sure that's our troll.
20:36:54FromGitter<mratsim> So Ubuntu 14.04 support expires in July but Travis CI is still stuck on this version. And no updates in sight: https://github.com/travis-ci/travis-ci/issues/7260
20:37:21dom96Isn't there a flag you can set to get a newer version?
20:37:38FromGitter<mratsim> basically you’re forced to use docker
20:37:49FromGitter<mratsim> the flag allows you 12.04 or 14.04
20:40:12FromGitter<mratsim> @alehander42 is that the kind of use case for breeze? https://gist.github.com/mratsim/9d176b7b366fae9b990ff82932756ead
20:40:57FromGitter<mratsim> Having some kind of declarative configuration and generating types/proc from it?
20:41:16adeohluwaalready got it working...
20:44:28FromGitter<mratsim> ah, I see, that would be super useful
20:44:46adeohluwanew error
20:44:59adeohluwahttps://usercontent.irccloud-cdn.com/file/82DppA0T/Screenshot_20180422-214425.png
20:51:47FromGitter<mratsim> did you perhaps compile to C++ just before and now are compiling to C? If so try clearing your nimcache
20:52:58FromGitter<mratsim> I have this kind of symbol(.text+0xaf0) with C++ stuff in my cache
20:55:37*BitPuffin quit (Remote host closed the connection)
20:55:56*nsf quit (Quit: WeeChat 2.0.1)
21:10:54*yglukhov quit (Read error: Connection reset by peer)
21:11:29*yglukhov joined #nim
21:14:25*elrood quit (Quit: Leaving)
21:24:50*athenot joined #nim
21:27:51FromGitter<honewatson> O
21:29:28*xkapastel quit (Quit: Connection closed for inactivity)
21:35:53FromGitter<zacharycarter> Northern VA Nim meetup 2018 was a success :D
21:43:55*athenot_ joined #nim
21:44:44*athenot quit (Ping timeout: 276 seconds)
21:49:24FromGitter<mratsim> awesome!
21:50:08FromGitter<mratsim> The lineInfo proc in the macros module is great.
21:50:30FromGitter<zacharycarter> sooooo quick question
21:50:41FromGitter<zacharycarter> let's say I have a concept named Foo
21:51:24FromGitter<zacharycarter> if I have a proc that accepts a parameter of Foo and I try to use the name proc from typetraits on the value passed into the proc, it returns the name of the concept
21:51:33FromGitter<zacharycarter> not the type that was passed in
21:51:39FromGitter<zacharycarter> should it be that way?
21:52:29FromDiscord<claude> whats the result if you do proc[T: Foo](param: T)
21:52:33FromGitter<zacharycarter> and if it should - is there a way to do some sort of refinement?
21:52:45FromGitter<zacharycarter> well let me share a little code
21:53:02FromGitter<mratsim> I didn’t try concepts yet. I only raised ~40 static bugs :P
21:53:20FromGitter<zacharycarter> ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5add044f270d7d3708f47f3f]
21:54:20FromGitter<zacharycarter> the result is - `"Shape":{`
21:54:47FromDiscord<claude> just try `proc marshallShape[T: Shape](shape: T)`
21:54:57FromGitter<zacharycarter> okay thanks
21:55:16FromGitter<zacharycarter> you are awesome!
21:55:19FromGitter<zacharycarter> thank you!
21:55:37FromGitter<zacharycarter> this is for the aws lib @Varriount and I are working on
21:55:48FromGitter<zacharycarter> well he's been working on it - but that's what we met up about today
21:58:10*NimBot joined #nim
22:10:37*SenasOzys__ quit (Ping timeout: 256 seconds)
22:16:03*cspar joined #nim
22:17:14*athenot joined #nim
22:17:53*athenot_ quit (Ping timeout: 276 seconds)
22:21:29*SenasOzys joined #nim
22:25:04*CodeVantage quit (Ping timeout: 246 seconds)
22:46:40*leorize quit (Quit: WeeChat 2.1)
22:54:03*adeohluwa quit (Quit: Connection closed for inactivity)
22:55:52*xkapastel joined #nim
22:57:30*CodeVantage joined #nim
22:59:55*CodeVantage quit (Client Quit)
23:38:32*krux02 joined #nim
23:41:36krux02does anybody here uses emacs with nim?
23:41:49krux02Currently It seems to be broken: Error from syntax checker nim-nimsuggest: (wrong-type-argument flycheck-error-p [cl-struct-flycheck-error scratch.nim nim-nimsuggest ??? -1 0 system [Processing] info nil])
23:43:21*athenot_ joined #nim
23:44:20*athenot quit (Ping timeout: 276 seconds)
23:53:37*Lord_Nightmare quit (Ping timeout: 248 seconds)
23:56:26*Lord_Nightmare joined #nim