<< 04-06-2018 >>
00:20:00*rockcavera joined #nim
00:31:58shashlickThey are nimgen global sections, hence the n. to distinguish
00:57:23*flaviu quit (Remote host closed the connection)
01:03:36shashlickLet me know if you have any feedback
01:04:51FromGitter<Varriount> shashlick: Is there a way to change the output name of a translated C file?
01:14:04*endragor joined #nim
01:18:55shashlickvarriount: not as of yet, so far, nimgen converts - to _ since - is not allowed in module names
01:19:18shashlickbut i need that for the 7z wrapper I was trying out since its first letter is a number which nim doesn't like again
01:26:29shashlickwhat's your specific use case?
01:30:50FromGitter<Varriount> shashlick: Well, I was thinking of something like "import nimreadline" and "import nimreadline_core", where the "_core" contains the actual wrapper. The other module would contain higher-level functions.
01:31:47shashlicki've been thinking about that
01:32:39FromGitter<Varriount> What I would really like to do is "import nimreadline" and "import nimreadline/core", but I suspect that the module system doesn't allow that.
01:33:22shashlickright now, all nimgen modules have an output dir so you import them as `import nimssl/sha`, I was wondering if you could have a `nimsha.nim` in the root - would nimble and nim accept that
01:34:13FromGitter<Varriount> The best way would either be to ask dom96 or araq, or test it yourself.
01:38:47shashlickvarriount: nope, nimble doesn't like it, complains about package structure
01:40:31shashlickwill have to have it as a sibling in the output directory, but nimgen deletes all .nim files because all are expected to be generated
01:40:49FromGitter<Varriount> shashlick: Man, c2nim has come a long way. I was actually able to translate the linenoise source file (not the header) with c2nim.
01:41:05shashlicknice
01:41:11FromGitter<Varriount> I had to make 2 edits, but that was it.
01:56:07*dddddd quit (Remote host closed the connection)
02:02:53*zszf6cefj5iz quit (Quit: WeeChat 2.1)
03:43:25*rauss joined #nim
04:39:09*sz0 quit (Quit: Connection closed for inactivity)
04:52:31*leorize quit (Quit: WeeChat 2.1)
04:52:53*leorize joined #nim
04:53:16*leorize quit (Client Quit)
05:00:52*lainon quit (Ping timeout: 260 seconds)
05:09:07*lainon joined #nim
05:09:31*lainon is now known as Guest27633
05:09:59*shashlick quit (Ping timeout: 276 seconds)
05:10:53*Guest27633 quit (Client Quit)
05:36:00*nsf joined #nim
05:53:31*pwntus joined #nim
05:54:06*pwntus quit (Changing host)
05:54:06*pwntus joined #nim
06:13:43*xkapastel joined #nim
06:41:37FromGitter<Paalon> Why the first code can run but the second becomes error? ⏎ ⏎ ```code paste, see link``` [https://gitter.im/nim-lang/Nim?at=5b14df21ba1a351a68ea019a]
06:49:13*lainon joined #nim
06:50:56*skrylar joined #nim
07:00:33*Vladar joined #nim
07:02:09FromGitter<Varriount> !eval let b = 1;let c = 0.1;let d = b + c;echo d
07:02:11NimBotCompile failed: in.nim(1, 33) Error: type mismatch: got <int, float64>
07:02:41Araqbecause integer literal are special and when put into a 'let' the type 'int literal' becomes 'int'
07:02:50*yglukhov[i] joined #nim
07:09:05FromGitter<Varriount> Huh. I would have thought that int+float would convert the int to a float automatically
07:14:29FromGitter<Paalon> Ok. I see.
07:16:50*PMunch joined #nim
07:17:49FromGitter<Paalon> I'm rewriting lib/pure/complex.nim, and realized that there are `+`(complex, float) and `+`(float, complex) but not `+`(int, float) and `+`(float, int).
07:19:04Araqimport lenientops
07:21:18FromGitter<Paalon> This seems inconsistent, so what should be the module complex?
07:21:38FromGitter<Paalon> Preparing lenientops for complex?
07:21:54Araqint is 64 bits
07:21:57Araqfloat is 64 bits
07:22:04Araqcan't convert without losing information.
07:22:22Araqcomplex is based on floats, complex+float loses no information, should be supported.
07:22:41FromGitter<Varriount> On 32-bit platforms though, it would work. Hm.
07:23:00FromGitter<Paalon> There are no int + complex
07:23:08Araqwe don't base language rules on what "would work on 32bits"
07:25:08Araqlikewise complex+int would lose information
07:25:46Araqit all makes sense and I when I hear that word "inconsistent" once more, I'll rage quit for a month
07:26:21Araq;-)
07:27:28FromGitter<Paalon> I understand.
07:27:44Araq"Inconsistency" comes into play after you understand the set of rules and you detect a violation of the rules. It is not a fancy word for "I am not sure I understand this and I'm a smart guy and in Python it's different, hence 'inconsistent'"
07:28:47FromGitter<Paalon> I think we should support int + complex like lenientops does.
07:37:23FromGitter<Paalon> I think it's not inconsistent
07:39:58Araqwell create a PR and let me see how you solve it
07:43:03FromGitter<Paalon> OK. Thanks.
07:46:00PMunchIs > and >= just a template over
07:46:03PMunch<
07:46:28PMunchI noticed that there's no > and >= in lenientops, only < and <=
07:46:44PMunchHi everyone by the way, been away for a while :P
08:02:57*gmpreussner quit (Ping timeout: 240 seconds)
08:05:46*gmpreussner_ joined #nim
08:08:21FromGitter<alehander42> hey : )
08:09:11PMunchHi
08:09:36FromGitter<gokr> PMunch: Saw your combparser thingy
08:09:54PMunchOh, thoughts?
08:10:23FromGitter<gokr> I haven't looked closely, but I always loved working with PetitParser in Smalltalk - which also is a parser combinator.
08:10:49PMunchIf you want a larger and more readable example you can check out https://github.com/PMunch/protobuf-nim/blob/master/src/protobuf/private/parse.nim
08:12:04FromGitter<gokr> Are they dynamic or are you also doing compile time stuff?
08:12:57FromGitter<gokr> Ah, could run on compile time it says.
08:12:59PMunchIt's all turned into procs, apart from the root parsers which are templates that creates procs to get instantiation information
08:14:45PMunchBasically a parser is a proc that takes an input and outputs a parser result which contains error messages and all that stuff
08:15:19PMunchI have a fix that I've yet to push that speeds it up a whole lot as well, using my strslice library
08:17:24PMunchBut I might also add in a table of fixed errors and map to that instead of copying all the error strings. Since errors can be removed by something like an "or" combinator a lot of the error messages are simply never seen, nor relevant. So having the overhead of copying them into memory doesn't make much sense.
08:17:42PMunchAnd I want to flesh out the onerror thing a bit more as well, would be nice to do both at the same time.
08:18:24PMunchWith onerror you can create more meaningful error messages than the massive stack-trace looking thing it currently spits out
08:19:12AraqPMunch: welcome back.
08:19:49Araqone thing I would like a parser generator to do is to produce an AST out of the box. and the AST should be based on my packedjson ideas...
08:22:28PMunchHmm, interesting
08:23:08*xkapastel quit (Quit: Connection closed for inactivity)
08:24:15PMunchSo packedjson identifies were in the string values are, and then converts them on demand to the type it should have?
08:26:23PMunchTBH the combparser library is a bit misguided for Nim. It is heavily influenced by the functional way of doing things, while in Nim it would probably be better to have a macro that returned a more optimized parser.
08:28:48PMunchHaven't looked at how glossolalia works, but that looks promising
08:31:27*lainon quit (Ping timeout: 240 seconds)
08:37:22*skrylar quit (Remote host closed the connection)
08:42:09*ftsf quit (Quit: Leaving)
08:43:05YardanicoGitLab right now :D https://i.imgur.com/dFMRm5a.png
08:50:12*floppydh joined #nim
08:51:48AraqPMunch: that's just an implementation detail but the core of the idea is that things stay in compressed form and then accessors uncompress.
08:52:34Araqso it's as convenient to work with as it always was, it's just that the accessors now have something to do and are not just cargo cult OO practice
08:57:06*Arrrrrrrrrr joined #nim
08:59:27*athenot joined #nim
09:01:38Arrrrrrrrrrhttps://github.com/nim-lang/Nim/issues/7917 I don't think is this a good idea. I'd rather initialize variables by assigning them a type than typing its default value every time.
09:04:15AraqArrrrrrrrrr: well but the existing Nim requires you to write 'var x = initTable' instead of 'var x: Table' and the assignment is conceptually much cleaner
09:04:56AraqI don't see how you can argue for 'var x: Table' tbh
09:07:31ArrrrrrrrrrYeah, maybe except for table and one or two more types. For the rest I like to either directly assign a value (var done = false) or rely in default assignment (var done: bool)
09:07:41ArrrrrrrrrrThat measure would imply doing both.
09:07:51YardanicoArrrrrrrrrr https://github.com/nim-lang/Nim/issues/7917#issuecomment-394182090
09:08:18Yardanicoso we'll be able to provide default values in type declarations it seems :)
09:09:17ArrrrrrrrrrBut it is more common to define primitive variables.
09:09:56Arrrrrrrrrr*variables with a primitive type
09:10:31Yardanicoyeah I know, and a lot of Nim code relies on it AFAIK
09:10:56Yardanicobut if it will happen, it should happen before 1.0 :)
09:11:19Yardanico(and I don't mind it if it would help to optimize Nim even more)
09:15:44ArrrrrrrrrrI don't know why it would make the code more optimized, what's the difference with forcing you to assign a 'false' vs the compiler whenever a value is not assigned? I don't see a problem here. For objects maybe it would avoid issues with nil, that's right. With objects, It is more the fact that null works different than mistakenly forgetting an assignment. For that reason I could sympathize with the 'not nil' rule
09:16:50Araqvar b: bool # please compiler, check that every code path sets it
09:16:58Araqis a valuable feature in its own right
09:17:07Araqvs.
09:17:18Araqvar b = false # start with the 'false' value.
09:17:29AraqI don't see why it would be annoying.
09:17:58Araqif you argued "omg, I need to update so much code then" then you would have my sympathy, but you don't...
09:18:05FromGitter<survivorm> btw, not been looking for some time. Is there a way to parse json Node to some complex type out of the box?
09:18:30Araqyeah, the json.nim module support that, the macro is called 'to' iirc
09:19:21yglukhovI've got lots of code with `var v: T` pattern without specialization. I Is it bad?
09:19:53FromGitter<survivorm> @Araq thanks!
09:20:28Araqyglukhov: well the Nim compiler is FULL of this.
09:20:43Araqexpect a long transition period with warnings instead of errors.
09:21:26yglukhovhmm.. but i kinda like this pattern... what's wrong about it?
09:21:32ArrrrrrrrrrI don't know much about the compiler, I thought it would be more obvious to tell if a variable was already assigned or not. I don't mind directly assigning a value, and probably I use it a lot. But I'm happy with the alternative for cases where I know what the def value but direct assignment looks ugly, for example: 'var shit = 0'u64' or 'var blah = uint8(45)'.
09:22:04Araqugly as opposed to what alternative?
09:22:06ArrrrrrrrrrWell, the last case doesn't count
09:22:24Araqthe 0'u64 syntax is there in ordinary assignments too.
09:23:04ArrrrrrrrrrI mean, It is more readable 'var myVar: int64' than 'var myVar = 0'i64/int64(0)'.
09:24:19ArrrrrrrrrrThe process goes like "This var is of type X. I can imagine the default value" vs "This var is of type X initialized as Y". B gives more obvious information.
09:25:02ArrrrrrrrrrExcept for some types where the initialization is way simpler (bools) or where you cannot avoid it (your table example)
09:25:56Araq'var myVar: int64' is implicit, an explicit zero makes more sense to me
09:26:49ArrrrrrrrrrYes, but it is so common and obvious you eventually internalize it, like in 24 hours. Consider the cpFile, newSeq, etc ... what is cp, what is seq?
09:27:09Araqyglukhov: how else would I get the "check every code path sets it" feature? 'var myVar: int64 = void' ?
09:27:47yglukhovthe cases where implicit init makes sense to me is conditional initialization. var a: T; if cond: a = 1 else: a = 2... Another case is by var init. var a: T; init(a)
09:28:26yglukhovAraq: well the problem is only not-nils, right?
09:28:43yglukhovEverything else is perfectly fine with implicit init
09:29:30yglukhovAlso enums with weird 0 meaning, i guess...
09:29:45FromGitter<alehander42> @Araq have you had problems with wrong "similar" EqResult in karax before?
09:30:17Araqalehander42: not really, I'm unaware of any remaining bugs with the DOM diffing
09:31:07Araqyglukhov: well historically default(T) was just there to ensure memory safety, and it quickly turned into an idiom.
09:31:43Araqbut now we have better ways to do it, do some simple minded control flow analysis. like Java did in the 90ies.
09:32:37Araqand given that it doesn't work well for nil, enums I found it fair to deprecate it altogether
09:32:48FromGitter<alehander42> yeah, diffing has been great & very stable for me for a long time
09:33:01FromGitter<alehander42> I'll try to reproduce my problem in a smaller program, but it would be very hard
09:36:16Araqwell you can start by adding echos into karax.nim
09:36:40Araq'nimble develop' is excellent.
09:37:28Araqmaybe you can fix it more easily than reproducing it in a smaller program :-)
09:38:21yglukhovAraq: i'm not sure i understand what the problem is. init verification should be done only for locations which can't be 0. that's either not nils or some enums. explicit initialization doesn't always make sense: `var someInts: array[10, int]; getSomeInts(someInts);`. I don't mind about not-nils init requirement, but even then it can often be verified through codeflow analysis.
09:40:59Araqlol, now I have to bring up the "consistency" word. treating all types the same creates a more consistent language.
09:41:34FromGitter<alehander42> ahhhh
09:41:35FromGitter<alehander42> https://github.com/pragmagic/karax/commit/87edc4cdb9ac9de6ec50b4cd19bb1ce6e503a3f5
09:41:40FromGitter<alehander42> I suspect this is my bug :D
09:43:43yglukhovwhat should i do with `var someInts: array[10, int];` then? all these patterns with be just found and replaced with `var someInts = default(array[10, int])`? except what do we do with `{.noInit.}`?
09:44:24ArrrrrrrrrrYou are going to destroy many lifes
09:44:47Araqok ok, fine, let's turn this into a vote
09:45:00AraqI personally don't mind much either way.
09:45:47Arrrrrrrrrryglukhov: you should post your points about conditional initialization and arrays in that issue
09:46:43Araqthe "init by var T" idiom is a big problem, I have to admit
09:47:14Araqand might turn into an 'init by out T' which we lack.
09:47:37PMunchHow does changing values work in packedjson by the way? Does it modify the string? That has to be costly
09:47:54*yglukhov[i] quit (Remote host closed the connection)
09:48:25Araqit modifies the string in place, it is costly.
09:49:15Araqwhen that causes problems, use the traditional json.nim :P
09:49:39Araqnote that for ASTs this is usually fine as you need to re-create a transformed AST anyway
09:49:51Araqfor different reasons
09:50:58PMunchOh yeah, I was just wondering
09:51:33PMunchTechnically you could keep a ledger of changes that only got applied when you needed to write the JSON back out
09:51:46Araqtoo complex.
09:54:06Araqalehander42: that should only have performance consequences
10:00:52*Arrrrrrrrrr quit (Quit: Page closed)
10:07:58FromGitter<alehander42> @Araq it's strange, with different instead of similar there, I get into a something that seems like an infinite loop in diff
10:08:47FromGitter<alehander42> well that's not true, e.g. updateStyles would be called only for similar
10:08:58FromGitter<alehander42> which is my original issue (it updates a node with the wrong class)
10:09:10Araqthe algorithm is nuts
10:09:23Araqwe should replace it back to its former, simpler variant
10:09:49Araqit is over-optimized to some benchmark IMO
10:10:06*nsf quit (Quit: WeeChat 2.1)
10:10:12Araqand faces exponential blowups in fringe cases that I never really got rid of
10:12:06FromGitter<alehander42> but I can't really understand how it can go into a loop (it seems to me, from my logs that it somehow repeats processing the same nodes, but I might be mistaken, and it might be just exponential combinations)
10:12:54FromGitter<alehander42> well, with the old "similar" value in that branch, it seems to almost work correctly here, otherwise I didn't have any other problems overally
10:13:37FromGitter<alehander42> maybe it can support the simpler one and a "optimized" diff strategy so users can choose one? or that would be too complicated
10:15:40*thomasross quit (Read error: Connection reset by peer)
10:17:36*nsf joined #nim
10:20:36*girvo quit (*.net *.split)
10:21:26*girvo joined #nim
10:22:17FromGitter<alehander42> actually the old behavior "similar" even looks 100% correct in that branch in my case, my problem comes from `if a.class != b.class or not eq(a.style, b.style) or not sameAttrs(a, b):` (but probably even before that, somehow the wrong nodes seem to be mapped to each other)
10:22:26FromGitter<alehander42> but I'll make a workaround for now
10:39:59dom96So let's have a vote: who thinks Microsoft will ruin GitHub?
10:40:12Yardanico-1 from me, I don't think it will
10:40:27dom96Yeah, I'm cautiously optimistic
10:40:45dom96you never know, but MS is different now
10:40:48PMunchI'm cautiously neutral..
10:41:08PMunchI have a feeling that they will mess it up, but I really really hope they won't
10:41:58PMunchdom96, that sounds like someone in an abusive relationship "oh but he has changed, he only requires Windows Live login when he's drunk"
10:43:43Yardanicowell, new Microsoft is much different from old Microsoft :)
10:43:52Yardanicofor example - they don't earn a lot of money on Windows anymore
10:44:25euantorNo, they earn money on services like GitHub. Hopefully that doesn't mean aggressive monetisation of GitHub
10:45:01Yardanicoyeah
10:49:40*yglukhov[i] joined #nim
10:50:12*xet7 quit (Remote host closed the connection)
10:51:44FromGitter<alehander42> I am more afraid of google, than ms these days
10:52:09FromGitter<alehander42> ms make some great developer tools & awesome system/language research
10:53:52*yglukhov[i] quit (Remote host closed the connection)
10:54:08*yglukhov[i] joined #nim
10:54:17FromGitter<alehander42> at least it seems that quality github code tooling aligns somehow with their interest, but I have no idea, people say skype & linkedin suck more now, is it true
10:56:01*athenot quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
10:56:03*yglukhov[i] quit (Read error: Connection reset by peer)
10:56:16FromGitter<mratsim> Google has a tendency to start tons of stuff and abandon them (Google+, Google Reader), it lacks coherency.
10:56:22*yglukhov[i] joined #nim
10:58:15*Kaynato quit (Read error: Connection reset by peer)
10:58:35*athenot joined #nim
10:58:42*Kaynato joined #nim
10:59:10PMunchSkype pretty much got gutted and turned into a glorified web-app
10:59:28PMunchNever really used LinkedIn so I'm not sure about that..
11:00:00*athenot quit (Client Quit)
11:05:38*Yardanico quit (Ping timeout: 260 seconds)
11:10:52FromGitter<alehander42> anyway, gitlab etc are always there, so I am not too concerned (I use gitlab a lot , honestly github's + is mostly the existing OS communities)
11:12:29Araquh yeah, Skype got worse with its ads
11:23:13*Yardanico joined #nim
11:52:33PMunchYardanico, I just updated binaryparse to fix your issue. Care to try it out?
11:52:37Yardanicoyeah
11:57:30YardanicoPMunch yeah, that works. also wanted to ask - do you plan to add support for little endian?
12:02:04*arecaceae quit (Remote host closed the connection)
12:02:28*arecaceae joined #nim
12:10:42*fvs joined #nim
12:13:15FromGitter<mratsim> LinkedIn is turning into Facebook (for the silly posts) + Twitter (for the bot accounts) but that’s not Microsoft fault.
12:13:50*Snircle joined #nim
12:15:02PMunchYardanico, I am considering it. But I'm not sure how I would do it
12:15:15YardanicoPMunch did you see NESM?
12:15:33PMunchI mean converting 32-bit or 64-bit integers to big or little endian is fair enough, but a 3 bit integer?
12:15:41Yardanicowell, yeah
12:15:56Yardanicohttps://github.com/xomachine/NESM
12:16:04PMunchHmm. no I haven't seen that
12:16:37Yardanicoit's very complex too and supports a lot of stuff
12:16:47Yardanicohttps://xomachine.github.io/NESM/
12:18:11Yardanicothe only downside to it - it doesn't actually support custom-sized integers
12:19:49PMunchYeah, I wanted binaryparse to be very little magic. Only specify your format and off you go
12:20:17PMunchI got inspired from stuff I was doing at work with CCSDS headers for spacecraft data
12:20:41PMunchAnd I actually ended up using it a couple months after I wrote it when I was tasked with creating some test data
12:27:51PMunchSo custom-sized integers were a must, as many of the headers use them for different things
12:35:47*endragor quit (Remote host closed the connection)
12:39:52*athenot joined #nim
12:41:20*athenot quit (Client Quit)
12:48:14FromGitter<data-man> @Araq: my system.nim contains everything from typetraits, various sugar (toOpenArray for strings, |=, &=, ^=, etc.) and honey (LoopUnrollingStep, prefetch* templates). Honey is experimental. ⏎ BTW #7944 is updated.
12:48:21*nsf quit (Quit: WeeChat 2.1)
12:50:05*endragor joined #nim
12:52:51FromGitter<mratsim> `^=` ?
12:53:09FromGitter<mratsim> `^` is pow btw, not `xor`
12:54:07FromGitter<data-man> ```a = a xor b```
12:54:25*endragor quit (Ping timeout: 248 seconds)
12:54:34FromGitter<data-man> As in C.
12:55:10FromGitter<mratsim> yeah but Nim is not C, and pronounceable words were preferred to symbols.
13:00:14FromGitter<data-man> Therefore I haven't made a PR. :)
13:00:50Araqgood boy. :-)
13:01:25Araqbtw I came to the conclusion that the implicit 'import system' was a mistake.
13:01:39Yardanicoreally?
13:01:44PMunchHow so?
13:01:46Araqyeah, pretty much.
13:02:06Araqit doesn't save anything in non-toy examples
13:02:21Araqwhich end up importing strutils or os or something else anyway
13:02:29Araqit makes
13:02:34Araqimport system except foo, bar
13:02:41Araqreally hard to support internally
13:03:01FromGitter<data-man> Can we split system to tiny modules?
13:03:02Araqand it means people push for their favorite shortcuts into system.nim
13:03:22Araqbecause system.nim is so special.
13:04:10PMunchHmm, all good points
13:04:18PMunchWhat is in systems.nim anyways?
13:04:34Araqintegers, strings, seqs, io routines
13:05:06Araqmagics like isMainModule, etc etc
13:05:16FromGitter<data-man> systemtypes.nim, systemio.nim, etc. ?
13:05:28PMunchWait, all the types are in system. What would you be able to do without it? :P
13:05:36AraqI wouldn't mind that, data-man
13:05:45Araqthe details need to be fleshed out though.
13:05:47PMunchI guess it's too late now though, that would break all of existing Nim code..
13:05:58FromGitter<Varriount> Araq: I'm in favor of the explicit initialization, especially if it gives us "uninitialized variable" detection.
13:06:30AraqPMunch: well I can envision a transition period, but we have more important things to work on
13:07:45FromGitter<data-man> We need RFC for this (split system.nim) :)
13:08:03Araqyup
13:08:23Araqprobably start with "move IO routines to io.nim"
13:08:35Araq"except echo because echo is everybody's darling"
13:09:30AraqPMunch: you would need to import it, but it would also be more obvious how to import alternative system.nim implementations
13:09:57Araqwhich wouldn't be all that crazy given Nim's .magic hooks
13:10:07Araqthat make sytem.nim tick.
13:10:39PMunchTrue
13:12:01Araqalso 'include prelude' exists and is probably under-developed
13:12:23Araqbut it solves the problem more elegantly IMO.
13:18:05dom96It's too late for this IMO
13:19:13FromGitter<data-man> I need help for memmapstream for Windows. Currently the same file can't be opened for writing and for reading in Windows. This would be useful for interprocess communication.
13:19:31FromGitter<mratsim> Good luck with that
13:19:38FromGitter<data-man> :)
13:19:52FromGitter<mratsim> Microsoft still doesn’t solve it properly for Excel files on shared drives
13:20:05FromGitter<mratsim> sometimes you lock yourself out of the file
13:20:41FromGitter<mratsim> open an Excel file, and try to upload it to an email for example, it will error out.
13:22:45FromGitter<data-man> AFAIR in Windows 10 something was improved for memmapping.
13:22:59FromGitter<Varriount> Of course, you could go the other direction, and make something like `var foo: ref Bar` automatically allocate.
13:29:34Araqvarriount: sounds terrible :-)
13:31:19Araqdata-man: unicode PR?
13:31:26*DarkArctic joined #nim
13:32:45FromGitter<data-man> Opinions about Unicode RFC? #7902
13:38:27Araqmmm? RFC?
13:38:29Yardaniconew github header
13:38:31Yardanicohttps://github.com/
13:38:45Yardanicohttps://blog.github.com/2018-06-04-github-microsoft/
13:43:50FromGitter<data-man> Currently genunitables can be used so: ```genunitables <Unicode version> <Emoji Version>```. I can set data's version or links to beta .txt files. ⏎ This script download data from unicode.org and generates unitables.inc. unicode.nim includes this file.
13:44:33PMunchYardanico, and so it begins
13:49:24FromGitter<data-man> I think genhtmlentities are also needed for htmlparser. We have very outdated data.
13:49:25Araqdata-man: sounds good, just make it point to the latest stable Unicode version (11?)
13:50:22Araqthese got updated recently iirc ?
13:51:07Araqanyway, that would be a separate PR. and HTML entities are a relict of pre-Unicode times IMO.
13:51:44FromGitter<data-man> Unicode 11.0 will released in this month, AFAIK.
13:52:20FromGitter<data-man> Official latest is 10.0
13:53:17FromGitter<data-man> So unitables.inc or unitables.nim?
13:55:24Yardanicoinc?
13:55:54FromGitter<data-man> for including :)
13:59:46Araq.nim
13:59:58Araqwe are not FPC, .inc files are not a thing here
14:04:16FromGitter<data-man> I doubt about emoji. According to https://www.unicode.org/Public/emoji/11.0/emoji-data.txt decimal digits are also emoji. :)
14:09:09Araqwho cares about emojis.
14:09:13Araq:P
14:10:43FromGitter<data-man> It's me. :)
14:11:54Araqalways remember: Bojack doesn't twitter emojis.
14:15:04dom96*tweet
14:17:14Araqoh yeah, thanks.
14:17:24Araqin German it's a verb :P
14:18:55FromGitter<data-man> BoJack Horseman? :) (I googled it)
14:20:05*yglukhov_ joined #nim
14:21:50Araqsure
14:23:32*yglukhov[i] quit (Ping timeout: 245 seconds)
14:45:31PMunchHmm, no one have been crazy enough to write a XLSX reader with styling support in Nim yet?
14:47:11*endragor joined #nim
14:53:10*PMunch quit (Quit: Leaving)
15:05:26FromGitter<data-man> @PMunch: wrapper for https://github.com/brechtsanders/xlsxio ?
15:10:29*Trustable joined #nim
15:14:40*donotturnoff joined #nim
15:18:02*lainon joined #nim
15:19:37FromGitter<alehander42> @Araq my gf still calls you bojack horseman, when I talk to her about Nim, I think she doesn't remember your real name
15:20:20FromGitter<alehander42> @PMunch isn't xlsx excel files? what kind of style do they include
15:21:06Araqalehander42: I don't mind
15:21:41FromGitter<alehander42> I really can't think of more breaking thing than requiring import system.nim, but it would make sense in a way
15:23:19AraqI won't do it but it would make sense :P
15:24:35*endragor_ joined #nim
15:25:47*endragor_ quit (Remote host closed the connection)
15:27:57*endragor quit (Ping timeout: 240 seconds)
15:28:35FromGitter<data-man> I don't remember the exact version of Delphi, system.pas was splitted into several modules.
15:28:54FromGitter<data-man> It's good example. :)
15:31:23FromGitter<alehander42> did pascal also have system.pas :O :D
15:31:39FromGitter<data-man> Yes! :)
15:36:12Araqit never was splitted.
15:36:18AraqI think.
15:36:30*yglukhov_ quit (Read error: Connection reset by peer)
15:37:02*yglukhov[i] joined #nim
15:44:15FromGitter<data-man> http://docwiki.embarcadero.com/Libraries/Tokyo/en/System :)
15:46:51Araqhttp://docwiki.embarcadero.com/Libraries/Tokyo/en/System.IEnumerable part of System.pas
15:46:58FromGitter<data-man> That I want http://docwiki.embarcadero.com/Libraries/Tokyo/en/System.Types
15:47:11AraqI don't see how it supports your claims :P
15:48:00AraqSystem.Types lacks Integer for all I can tell
15:48:06FromGitter<data-man> We need RFC :)
15:49:27Araqall they did is make System a directory too with plenty of modules in it
15:49:45*shashlick joined #nim
15:51:51FromGitter<data-man> If module A imports systemio.nim should systemio.nim exports systemtypes.nim?
15:58:55*donotturnoff quit (Quit: Leaving)
16:06:03*leru joined #nim
16:09:23*PMunch joined #nim
16:11:39*miran joined #nim
16:24:41*xet7 joined #nim
16:29:36leruGood evening! Are there any plans to quit GitHub?
16:30:17PMunchHaha, we talked a bit about it earlier
16:30:20FromGitter<data-man> https://github.com/craigbarnes/dte/issues/141 - Move to GitLab as primary repo/issue host :)
16:30:51PMunchI think the general consensus was to wait and see if Microsoft turns GitHub into the new Skype/LinkedIn
16:31:04PMunchOr possibly when..
16:33:46*xkapastel joined #nim
16:34:05ehmryi was hoping now that the nim forum is done nim hub is next
16:34:21mirani can't believe all the reactions about MS and github
16:34:59*floppydh quit (Quit: WeeChat 2.1)
16:36:23PMunchI honestly expected more
16:36:51mirani bet at least 1/3 of those people use windows
16:36:54FromGitter<data-man> Can DigitalOcean host a git server?
16:36:57miranand 1/5 use vs code
16:38:08FromGitter<Electrux> Wait… u guys are moving from GitHub?
16:38:34PMunchElectrux, nothing planned AFAIK
16:40:01FromGitter<data-man> My vote goes to Fossil :)
16:40:25PMunchThere's nothing wrong with Git though..
16:40:45ldleworkOr github. Chances are, nothing at all happens for years.
16:40:46Yardanico@data-man Git is better because more people know it
16:40:49Yardanicoyeah
16:41:02Yardanicoand don't forget that this agreement between github and microsoft should go through regulators
16:41:12Yardanicoso it will officially happen somewhere in the end of 2018
16:41:46FromGitter<Electrux> i am kinda skeptical about microsoft, i do trust github, but m too lazy to bother moving anything… guess which one has the highest score :P
16:42:03YardanicoI don't really think that Microsoft will do anything bad about GitHubn
16:42:16FromGitter<Electrux> i hope so :)
16:42:21Yardanicothey know that people will immediately migrate to other alternatives
16:42:42PMunchWith all their issues, contributors, and PRs on GitHub?
16:42:44Yardanico(well, some people are already doing it)
16:42:47ldleworkYeah they just spent 7 billion dollars on it
16:42:58PMunchThat will be a tough migration process
16:42:58FromGitter<Electrux> although i can’t imagine why would github sell at all. i imagine it would have been doing financially pretty well
16:43:20PMunchWell, as far as I can tell they did decently. But they were struggling with finding a new CEO
16:43:26PMunchOr something like that
16:43:34ldleworkNot 7 billion dollars well.
16:43:45PMunchOh no, that's for the users
16:43:55ldleworkRight but from the perspective of the company
16:44:05ldleworkHowever good the company was doing, it wasn't doing 7 billion dollars good.
16:44:17PMunchNo as I said, that's for the users
16:44:24ldleworkI'm addressing the reason why github would sell
16:44:25PMunchRemember, if the product is free, you are the product
16:44:37ldleworkthe assumption "they were doing pretty well" means they wouldn't sell for 7 billion
16:44:46FromGitter<Electrux> so… nim is free :O
16:44:48Yardanicoinitial rumors were about $5
16:44:49*ldlework rubs his forehead.
16:45:17PMunchAnd while GitHub has private paid repositories I can guarantee you that Microsoft will want to harvest a little bit of that sweet, sweet data and "integrate" it with LinkedIn or something..
16:45:50PMunchYardanico, I would buy GitHub for $5
16:45:58YardanicoPMunch well, I've meant $5 billion :D
16:46:01PMunchHaha :P
16:46:06FromGitter<Electrux> lol
16:46:13miranwill somebody buy nim? :P
16:46:24Yardanicomiran I hope, but not *buy*, but *support*
16:46:34miranpotato-potato :D
16:46:44Yardanicomiran like mozilla supporting rust :)
16:46:45Yardanicoor apple supporting llvm :)
16:46:54miranM$ supporting nim? :D
16:47:04Yardanicoyeah, that would be pretty good :D
16:47:11Yardanicobut they already have their C# and Typescript
16:47:17FromGitter<Electrux> if MS starts harvesting off github, m moving all the way to zip based version control xD
16:47:24FromGitter<data-man> I know one great candidate for this :)
16:47:25Yardanicowhy? just host your gitlab instance
16:47:33miran@Electrux why not gitlab?
16:47:35PMunchYeah, paid support/feature priority is probably the most common commercial model for programming languages and tools
16:47:44FromGitter<Electrux> nim > C# and Typescript… just saying...
16:47:45YardanicoPMunch and nim has that too :P
16:47:47ldleworkThere's no reason to change your VCS because of the acquisition.
16:48:11PMunchLet's start mailing Nim back and forth as a zip file!
16:48:28FromGitter<Electrux> Miran: well, because what i said was a joke and apparently, u destroyed my poor little joke :’(
16:48:30FromGitter<Electrux> xD
16:48:35miranhaha, sorry
16:48:50miranbeen reading too much of "the end of the world is coming" comments
16:48:53Yardanico@Electrux don't worry :)
16:49:13*Arrrrrrrrrr joined #nim
16:49:31ehmryI plan certainly plan to move the repos I control, but not until the dust settles
16:49:33ArrrrrrrrrrHi. I read somewhere that nim is going to migrate to sourceforge. Is this really true?
16:50:10FromGitter<Electrux> Eh, as far as github works fine for me without harvesting *too much* of my data, i m fine with whoever acquires it
16:50:12YardanicoArrrrrrrrrr no :)
16:50:26dom96Yardanico: I'd say that Mozilla definitely owns Rust
16:51:21FromGitter<Electrux> :O
16:51:41dom96Arrrrrrrrrr: Yep, I for one think wherever Nim is hosted should have an Internet Speed Test feature. SourceForge offers that amazing feature! /s
16:51:42FromGitter<Electrux> i liked rust… till i realized how much i had to use the error handling :(
16:51:58FromGitter<Electrux> lol
16:53:39FromGitter<zetashift> I wouldn't want Microsoft supporting Nim, mainly because of F# getting no love from MS
16:53:57Yardanicowell, I started using nim (as I said multiple times) because it looked like Python, and Rust looked too "wordy"
16:54:08FromGitter<Electrux> ^ exactly my reason
16:54:21ArrrrrrrrrrMicrosoft would hire Araq to work on C#. Then Nim would die.
16:54:31FromGitter<Electrux> lol and python is too popular for me to like it xD dont get me wrong, it’s great… just too popular :P
16:54:46miran@zetashift oh, it would be great if F# received more love....
16:58:02Yardanico@Electrux if you're using something which is not very popular, it somehow feels like you're taking a bigger part in the community :)
16:58:16FromGitter<Electrux> EXACTLY!!!!
16:58:24Yardanicoe.g. there are tons of python repositories on github, and only 2k+ nim repositories
16:58:28ldleworkI really enjoy F#
16:58:40miran+1
16:58:41FromGitter<Electrux> it feels like m not just another drop in the whole wide ocean
16:58:46PMunchHmm, git-dit looks interesting
17:02:04AraqArrrrrrrrrr: lol SourceForge, I've been tinkering with Fossil / Chisel fwiw
17:02:18Araqbut I did that before github's acquisition :P
17:04:01*Snircle quit (Quit: Textual IRC Client: www.textualapp.com)
17:05:24Arrrrrrrrrrwhat's the advantage?
17:08:32FromGitter<data-man> @Varriount why #7415 was closed? You have some benchmarks?
17:12:57YardanicoArrrrrrrrrr fossil says it's better for projects where there's not a lot of main developers :D
17:13:02ldleworkhttps://monitor.gitlab.net/dashboard/db/github-importer?orgId=1&panelId=5&fullscreen&from=now-2d&to=now
17:13:05Yardanico"SQLite uses cathedral-style development. 95% of the code in SQLite comes from just three programmers, 64% from just the lead developer. And all SQLite developers know each other well and interact daily. Fossil is designed for this development model."
17:13:14Yardanicosounds a bit similar to nim :D
17:13:38Arrrrrrrrrrlol yes, only araq and dom contributes, and they are probably the same person.
17:14:23Yardanicono, they have a lot of different opinions :)
17:14:34FromGitter<data-man> Yes, we can see it on videos from FOSDEM. :-D
17:14:44Yardanicodon't you see them argue fairly frequently in repositories?
17:14:45Yardanico*in repository
17:14:58ArrrrrrrrrrThey never argue at the same time, guess why.
17:15:21PMunchWho's to say it's just one person trying to drum up discussion?!
17:15:30AraqFossil is awesome because it's not Git, not because of this cathedral-style nonsense
17:15:46YardanicoAraq well, it's on their page about differences
17:15:53ldleworkAs far as I'm concerned, the universe is just a solipsistic projection of mental phenomena and you're all me.
17:16:03ldleworkProve me wrong.
17:16:42ldlework(that was a joke)
17:17:35Yardanicowell, ok, let's move into #nim-offtopic :P
17:17:38Araqanyway, Nim started on some Python based DVCS
17:17:47Araqdon't remember the name
17:18:02Araqdarcs?
17:18:09FromGitter<data-man> Mercurial? :)
17:18:45FromGitter<zetashift> https://pijul.org/ also looks promising
17:19:08Araqno, darcs was that Haskell thingie
17:19:38FromGitter<zetashift> however what github did well was ease of creating a community around a project
17:20:38Araq"Bazaar", I think.
17:20:40FromGitter<zetashift> for better or worse I guess
17:20:57Araqand then people demanded git/github and I moved it over there
17:21:32FromGitter<data-man> Pijul written on Rust :)
17:21:34ArrrrrrrrrrWould you change the nest again?
17:22:10AraqArrrrrrrrrr: yes.
17:22:41Araqbut not without taking the issues/PRs with us, of course
17:22:46FromGitter<zetashift> you could just mirror the nim repo to gitlab something similiar to satisfy the leavers?
17:22:59Araqand not without good reasons.
17:23:18Yardanico@zetashift it's already mirrored
17:23:29Yardanicobut mirroring failed 2 weeks ago for some reason, I said that to Araq :)
17:23:48Yardanicoand you couldn't make PRs to Nim repo from GitLab anyway :D
17:24:17Araq"do it because M$ sux bolls and Linux rulez" is not a good reason
17:24:22Araq;-)
17:24:26ArrrrrrrrrrWell github is still the biggest git hosting community, that's hard to ignore
17:25:30FromGitter<alehander42> sadly I think git is way too mainstream now, even mercurial doesn't seem close
17:26:35FromGitter<alehander42> otherwise I wondered if gitlab will try to mirror the github "stats" for projects? a lot of os projects wouldn't want to leave all their contribs/stars/watchers behind , but I guess such a feature would be too invasive
17:27:13ehmrygithub was too centralized, so I think what happend is a good thing anyway
17:27:42ArrrrrrrrrrI don't think anything happened. Most migrations come from personal/small projects.
17:28:27ArrrrrrrrrrNot yet the time from decentralization.
17:28:31Arrrrrrrrrr*for
17:29:38FromGitter<alehander42> @Araq is there a unifiying "core" reason why some comb explosions happen in diffing? or are they all different random issues
17:30:13FromGitter<alehander42> (I know I can try to debug it all, but I really don't want to go there if I can find a workaround)
17:30:15FromGitter<data-man> Why talk only about GitLab? There is also BitBucket.
17:30:43leruJust in case you haven't seen this yet: Many people migrating to GitLab alredy: https://monitor.gitlab.net/dashboard/db/github-importer?orgId=1
17:31:03Yardanicoyeah, I shared that today, but maybe not many people saw it
17:31:20Yardanicobut I don't think that's "a lot of people migrating to GitLab"
17:31:27Yardanicomaybe it's just a copy in case something goes very badly
17:31:34Yardanicoor they just wanted to try GitLab
17:32:41FromGitter<data-man> GitLab so slow for me. :(
17:33:53Yardanicowell, I think they're under a high load right now :D
17:34:17Yardanicousually gitlab is pretty fast
17:34:32ldleworkscaling. it is hard.
17:35:42FromGitter<data-man> There is http://repo.or.cz also. :)
17:35:48leruI don't believe M$ won't fuck up GitHub. See Windows, OneNote, Skype, Nokia, Minecraft. I have a very bad feeling about that acquisition
17:36:22Araqalehander42: yes, the reason being "the algorithm is wrong"
17:36:41Araqer, "Is bad"
17:36:52AraqI never really understood why though :-)
17:37:32Yardanicoleru how did they fuck up Minecraft?
17:37:55Yardanicoit's only better with Windows 10 C++ version (but of course it would be better if it was a standalone app)
17:38:20FromGitter<alehander42> @Araq I see :D
17:38:43FromGitter<alehander42> and the previous "simpler" algorithm, how far back should I go in the commit history to compare it ?
17:39:05Araqfar.
17:39:23Araqit's easier to remove the prefix/suffix detection logic
17:39:31Araqand see if it makes a difference
17:39:49AraqI think it causes the exponential explosions
17:40:28leruYardanico, they're advertising the c++ version a lot and making it harder and harder for the java version. Minecraft's popularity is coming from the modding scene though. There are no real mods to the C++ version
17:40:33*nsf joined #nim
17:45:36*endragor joined #nim
17:47:16*endragor_ joined #nim
17:49:29FromGitter<alehander42> @Araq indeed, I commented out some prefix/suffix logic and it doesn't explode now, but it "forgets" and doesn't update some stuff :D
17:49:45FromGitter<alehander42> I have to spend some more time on it this week
17:49:59*endragor quit (Ping timeout: 256 seconds)
17:52:04ArrrrrrrrrrGood night
17:52:06*Arrrrrrrrrr quit (Quit: Page closed)
17:57:33FromGitter<zetashift> nokia was dying already(as well as windows phone), but for example linkedin is also doing alright, no real involvement of MS just like you'll want with GH
18:01:42FromGitter<kinkinkijkin> > @kinkinkijkin Oh!, your code theme looks nice, What font-family and theme setup are you using in you editor? ⏎ I'm two days late @faustinoaq but it's the font cherry by someone I know on discord, and it's a custom colour theme in vim
18:02:05FromGitter<kinkinkijkin> oof quotes do the whole thing eh
18:02:12leruConspiracy maybe idk: Nokia was dying because it was infiltrated by microsoft. Look up Stephen Elop
18:02:20FromGitter<kinkinkijkin> sorry I only have this chat open when I'm doing programming in nim
18:10:15*leru quit (Quit: Leaving)
18:13:25*natrys joined #nim
18:15:32*endragor_ quit (Remote host closed the connection)
18:20:07FromGitter<kinkinkijkin> how exactly do I use the json parsing stuff? I want to fetch the name of an item which has a certain property
18:20:10*nsf quit (Quit: WeeChat 2.1)
18:20:21PMunchWhat da you mean?
18:20:38FromGitter<kinkinkijkin> like, I'm trying tto figure out how to use it at all
18:20:55FromGitter<kinkinkijkin> and the guide on it in the docs isn't the most helpful
18:21:11FromGitter<kinkinkijkin> specifically I'm running a tool that returns a json and I want to parse it
18:21:45FromGitter<kinkinkijkin> currently I just run jq on it in the tool's commandline but that's slower
18:21:50PMunchparseJson
18:22:21PMunchstdin.readAll.parseJson
18:22:35PMunchIf you pipe it to your program
18:23:06FromGitter<kinkinkijkin> currently using execProcess
18:23:21PMunchUgh, play.nim-lang.org should be fixed so gists works..
18:23:50PMunchOkay, so execProcess(<args>).parseJson
18:24:25PMunchOr maybe .string.parseJson as execProcess returns a TaintedString
18:24:25*gb00s joined #nim
18:25:21FromGitter<kinkinkijkin> I'm not using taintmode
18:25:27FromGitter<kinkinkijkin> so it doesn't really matter
18:26:57Yardanico@kinkinkijkin get a string from the process
18:27:02Yardanicoand parse it using parseJson ?
18:29:15FromGitter<kinkinkijkin> okay, that's done, how would I then get say, which items in the json have the value "urgent" ?
18:29:22FromGitter<kinkinkijkin> as true
18:30:01PMunchmyJson["urgent"]
18:30:14PMunchWhere myJson is the result of parseJson
18:30:18Yardanicoyeah, and if it's a string, you would call .getStr()
18:30:22PMunchOh yeah
18:30:25Yardanico(refer to the docs for more procs)
18:30:51Yardanico!eval import json; echo parseJson("""{"a": 1}""")["a"].getNum()
18:30:53FromGitter<kinkinkijkin> so would I just check each entry and see if its "urgent" value is true?
18:30:55NimBot1
18:31:15Yardanico@kinkinkijkin wait, so your json contains an array of elements?
18:31:18dom96kinkinkijkin: hrm, how can we improve the docs?
18:31:19FromGitter<kinkinkijkin> yes
18:31:27Yardanico@kinkinkijkin use getElems then
18:31:44Yardanicoand honestly json docs are not bad
18:32:03dom96You access arrays just like you would normally: jArray[0]
18:33:30Yardanicowell, he wants to check if "urgent" is true
18:33:38Yardanicofor every element
18:36:08dom96for item in jArray: echo item["urgent"].getBool()
18:42:52*yglukhov[i] quit (Read error: Connection reset by peer)
18:43:25*yglukhov[i] joined #nim
18:52:56livcddid anyone already make a joke about MS acquiring Nim ?
18:53:18Yardanicoyes
18:53:29Yardanicoat least in Telegram :)
18:53:32miranlivcd: that would be me...
18:53:45livcdoh ok! atleast i asked
18:54:24PMunchlivcd, yes :)
18:54:40PMunchhttps://i.imgur.com/4Be8wz1.png
18:55:49livcdwell anyway the new microsoft is a good microsoft. People should complain when they actually hit issues with MS/GH behaving differently than usual
18:57:39dom96The best thing about this acquisition is the new meme power
18:57:42PMunchIt's a better Microsoft
18:57:49PMunchHaha
18:58:04livcdwell the .net core step was a big deal :)
18:58:11livcdsorry for being off topic
18:59:17PMunchHaha, this channel has been off-topic for most of today :P
18:59:30PMunchAs long as no-one actually needs help I don't think anyone minds
18:59:51FromGitter<kinkinkijkin> I do need help but I don't mind this
19:01:11PMunchOh, I didn't realise you still needed help. Still struggling with JSON?
19:01:19FromGitter<kinkinkijkin> yeah
19:01:45livcdPMunch: i know but sometimes araq does not like it :D
19:02:23livcdanyone that used Vim/Emacs that switched to something else ?
19:02:40livcdkinda tired of bugs when updating the editor
19:03:39*yglukhov[i] quit (Read error: Connection reset by peer)
19:03:55PMunchWhat kind of bugs are you having
19:04:00PMunchI haven't had any trouble with Vim
19:04:08PMunchOr are you talking about Vim/Nim integration?
19:04:13FromGitter<kinkinkijkin> https://i.imgur.com/C5nn7qG.png I think this will work but I won't be able to compile for a while
19:04:25FromGitter<kinkinkijkin> admittedly this is quite a bit cleaner and faster than what I was doing before
19:04:36livcdI have updated emacs to 26.1 and I have a trouble opening an .org file. It opens it and start org-mode but it does not behave correctly
19:04:43*yglukhov[i] joined #nim
19:05:06livcdwell for Nim i use vscode because most of you guys are on vscode and i know it works by default :D
19:05:23Yardanicolivcd I'm using VSCode too
19:05:32Yardanicobecause there are no nim extensions better than one for vscode
19:05:52Yardanico@kinkinkijkin btw, you don't need to specify type of currfocpsace
19:05:53PMunchkinkinkijkin, have you seen https://github.com/FedericoCeratto/nim-i3ipc
19:06:04PMunchAnd please share code snippets as text..
19:06:12PMunchThat way I could actually try it
19:07:31livcdYardanico: tinkering with configurations is tiring :/
19:07:38Yardanicowhat configurations?
19:07:48livcdi mean usually..so i am glad vscode just works with Nim
19:09:14livcdeven gopath was bit of a nuisance
19:11:57FromGitter<kinkinkijkin> @pmunch I'm not a fan of importing nonstandard libraries
19:12:38PMunchWhat? Why?
19:12:49PMunchYou'll be doing a lot of wheel-reinventing
19:13:07FromGitter<kinkinkijkin> usually there's some licensing stuff to deal with and I get more stress from that than redoing things myself as I need them
19:13:27FromGitter<kinkinkijkin> I also generate smaller binaries by tailoring my implementation to exactly what I'm doing
19:13:44livcdeverything has its cons and pros
19:13:47PMunchEh, whatever floats your boat I guess
19:14:08PMunchBut yeah, I think your code should do about what you need
19:14:10Yardanico@kinkinkijkin why wouldn't you write your own OS for your needs then? :P
19:14:24*fvs left #nim ("ERC (IRC client for Emacs 25.3.1)")
19:14:27Yardanico(just a joke)
19:14:39FromGitter<kinkinkijkin> I've actually thought of doing that
19:16:30PMunchHaven't we all :)
19:16:55dom96kinkinkijkin: looks like that will work
19:17:04dom96But `== true` is really redundant
19:17:17FromGitter<kinkinkijkin> oops forgot about that
19:17:22FromGitter<kinkinkijkin> haven't done code in mooooooonths
19:18:44livcdyglukhov: are you working on some new game ? :)
19:21:33livcduh and i just realized yglukhov works for a company that gave us mackeeper :D
19:22:29Yardanicowait what?
19:22:50livcdthat's what internet says
19:25:04dom96what?
19:25:09dom96link?
19:28:54livcdZeo alliance founders created mackeeper
19:30:45FromGitter<kinkinkijkin> okay it works :)
19:30:51FromGitter<kinkinkijkin> no more help needed
19:32:10FromGitter<kinkinkijkin> yeah I just cut off most of the cpu usage from that
19:32:10FromGitter<kinkinkijkin> good
19:32:28FromGitter<kinkinkijkin> it was only 1% on this cpu but it was still 1%
19:39:54*lainon quit (Ping timeout: 260 seconds)
19:43:48Araqalso give packedjson a try
19:45:23*Trustable quit (Remote host closed the connection)
20:00:08*leorize joined #nim
20:02:11Yardanicois there any tutorial/example for pegs? I want to try to use it for parsing mathematical expressions (AFAIK that's possible)
20:02:29Yardanicoah, probably it won't work
20:10:40*nsf joined #nim
20:11:29FromGitter<Varriount> @dom96: Does nimforum handle CORS?
20:14:51FromGitter<Varriount> @data-man That's odd... I didn't click close on that ticket.
20:16:23*icebattl1 joined #nim
20:17:22PMunchProbably Microsoft deeming it done /s
20:17:36*PMunch puts on tinfoil hat
20:19:05*icebattle quit (Ping timeout: 240 seconds)
20:32:08FromGitter<faustinoaq> @kinkinkijkin Oh, pk 👍
20:34:43shashlickso who all have merge access on nim-lang/Nim?
20:37:19Yardanicoshashlick only Araq knows :)
20:37:48*leru joined #nim
20:38:00*leru quit (Client Quit)
20:39:52*Sembei quit (Ping timeout: 256 seconds)
20:40:50shashlickthat's one thing that's important for Nim to grow, multiple lieutenants with specific focus and access/freedom
20:41:19shashlickand it should be documented so you know who to work with in order to understand specific areas and to contribute PRs
20:43:03Araqnot sure I'd agree with that
20:44:16Araqit's not clear to me how many contributions a single person can be responsible for before becoming a bottleneck
20:44:47*nsf quit (Quit: WeeChat 2.1)
20:46:15*lainon joined #nim
20:46:25AraqI mean, sure, I would love to delegate more of it. but a single person could easily handle all of it
20:47:18Araqanyway, what does need to be merged?
20:48:34shashlicki'm just asking in general but that's the only way to scale
20:52:30*PMunch quit (Quit: leaving)
20:53:08shashlickin all honesty, I know my PRs sit around for a while until someone has time to look at them and merge, it's not quite discouraging but definitely boring
20:53:54*yglukhov_ joined #nim
20:54:35FromGitter<Varriount> shashlick: My policy is to only review PRs for the standard library. I don't have the necessary knowledge to adequately evaluate compiler-related PRs.
20:54:40shashlickit's also uncomfortable to keep nagging you and dom96 to look at them, I don't know who else can review, coach and merge. It's not like you guys have nothing better to do
20:54:53FromGitter<Varriount> shashlick: You can bug me.
20:55:05*yglukhov[i] quit (Ping timeout: 240 seconds)
20:55:43shashlickultimately, the way to expand our collaborators is to help coach and guide people till they get the Nim way and for that you need to know who to work with
20:55:50Araqshashlick: yeah, that's true, but we don't necessarily need multiple lieutenants, a single person would do
20:55:54FromGitter<Varriount> shashlick: What's your github username?
20:55:58shashlickvarriount: yes, I've seen you merging stuff in recently
20:55:59Araqif that person had enough time... :-)
20:56:05*miran quit (Ping timeout: 245 seconds)
20:56:29shashlickaraq: genotrance <= but no one has enough time, that's why i feel divison of labor is good
20:57:08Araqyeah and I feel a single person could improve the quality of the review process
20:57:10shashlickno doubt nim-lang/Nim is a giant repo and I don't think github has fine grained enough access permissions to restrict what kind of PRs a person can merge
20:58:46shashlickyou really need someone who can help guide and coach and make the next batch of contributors and coaches
21:00:02shashlickalso applies to the many other repos like c2nim, nimble, choosenim, nimforum, etc.
21:02:25Araqwell it's hard :P
21:02:31Araqtake this one: https://github.com/nim-lang/Nim/pull/7907
21:02:44Araqit's a simple fix with a test case.
21:03:04Araqwhy haven't I merged it already? what the fuck is wrong with me?
21:03:41Araqproblem is: this fix makes no sense to me. I know the lexer's logic and this patch should make no difference whatsoever
21:03:56*Vladar quit (Quit: Leaving)
21:04:28AraqI'm probably wrong and it's a correct fix. but I need to understand it too, effectively I have to debug it on my own to be sure
21:05:17FromGitter<Varriount> Araq: You could ask for the change to be explained?
21:06:03Araqthat's bad for scaling but in the past I accepted PRs without this understanding and then I get back into the compiler's code and am I like "WTF? what is this crap..."
21:06:25Araqvarriount: I wouldn't buy the explanation.
21:08:29Araqnow I can see the fix is in fact correct.
21:09:36Araqbut I'm unhappy with the test :P
21:10:24Araqand it takes another roundtrip to GULPF...
21:10:39Araqit's frustrating but I don't see how to do better.
21:10:58Araqoften I merge and correct things later myself but that is not free either.
21:11:19shashlickthat's exactly how it should work but we need more araq's out there :)
21:11:28FromGitter<Varriount> If it makes you feel better, Python has the same problem.
21:11:46shashlickor at least pseudo araqs who can take care of lesser things like build, installer, CI, etc.
21:12:01shashlickgroom experts in specific areas
21:12:20shashlicknim is too big, let alone any of the extended repos
21:12:52shashlicksomeone should be the httpclient expert for example, so any PRs related to that, they can handle
21:13:35Araqwell other parts have different problems.
21:13:51Araqhttpclient has way too few tests afaict
21:14:16Araqand so every change is an exercise in playing with fire.
21:15:15shashlickanyway, I'm not saying I have a solution, but I feel more skilled hands = more progress
21:15:28Araqtests are really better at catching problems than a review process is.
21:15:48shashlickbut skilled hands have to be groomed and invested in to get there
21:16:18shashlickand more hands = more diverging opinions which also need to be managed effectively
21:16:41Araqwell I've been refactoring the compiler's internals a lot in order to help contributors
21:17:08Araqand I should probably invest more time in "live coding" the compiler.
21:17:26Araqbut for httpclient please write a test suite :P
21:17:30shashlick:D
21:17:52shashlickwho should I nag to merge that test case PR? ;)
21:18:09Araqwhich one?
21:20:03shashlickthis not yet done httpclient test suite PR that you want
21:21:20Araqwell you can nag me.
21:29:49shashlickthat's what I don't like, should I let you work on compiler bugs or 1.0 or other cool features or a lowly test case PR for httpclient
21:30:33shashlickshould I nag dom96 for my nimble and choosenim PRs or let him improve the forum
21:30:42shashlickthese simpler things should get delegated as much as possible
21:31:40shashlickevery PR you see can distract from bigger things
21:31:49shashlicktake up thought space which could be utilized better
21:40:53Araqok, so nag varriount about it :-)
21:42:07*lainon quit (Quit: Leaving)
21:45:35shashlickbrilliant :)
21:45:52dom96Yes, we do need more people. But not just any people, passionate people.
21:46:06shashlickvarriount: you in the US right? I can nag you even later in the day!
21:46:47shashlickdom96: we need to get more collaborators into your projects as well - nimble, choosenim, etc.
21:47:58dom96I wonder if adopting a Linux-based style system would work for us.
21:48:55dom96Hrm, I'm probably not clear: I mean how the Linux project handles contributions
21:49:05shashlickthat helps
21:49:06FromGitter<alehander42> what is the linux-based system
21:49:11shashlicknot sure how they do it
21:49:20FromGitter<alehander42> the hierarchical lieutenants?
21:49:41dom96I'm not 100% sure how it works to be honest heh
21:50:04dom96But AFAIK Linus gets a big diff at the end and he can either trust the people below him or review the diffs
21:50:06dom96or a mix of both
21:50:37dom96the people below him review smaller diffs for the corresponding subprojects (network, core, drivers, etc.)
21:51:10FromGitter<alehander42> I see, so he gets the most "promising"/improved PR-s
21:51:41dom96yeah, so it's like a progressive filter
21:53:11*xet7 quit (Remote host closed the connection)
21:55:10dom96In any case, regarding the Nimble PR I'm afraid you'll have to wait a bit longer
21:55:20dom96Jester is at the top of my priority list right now
21:55:33dom96and another top secret project that I'm working on for myself :)
21:56:28shashlickNo worries, I'm okay waiting and last thing I want to do is hold you guys up with smaller things
21:56:56shashlickBut it's something that needs to be figured out at a program level - you need to delegate
21:57:30shashlickFor agility and for giving people ownership
21:57:48shashlickThen they will be passionate and diligent
21:57:52dom96Yes, well, volunteers welcome
21:58:04dom96There is nothing stopping people from reviewing your PR
21:58:27FromGitter<xmonader> How are you guys :(
21:58:44shashlickBut will it still wait for you to merge
21:58:51dom96xmonader: why so sad?
21:59:07shashlickIt's about building trust and letting go of less important stuff
21:59:31shashlickWhat you do and don't deem important is your call
21:59:33FromGitter<xmonader> @dom96 because I don't do enough nim these days and my year goals aren't happening
21:59:59dom96xmonader: ahh, don't worry about it. There is still plenty of this year left :)
22:00:51FromGitter<xmonader> I hope so thanks
22:05:00dom96I'm really curious what this person is up to https://github.com/tbrand/which_is_the_fastest/issues/227
22:05:51*lainon joined #nim
22:07:09FromGitter<xmonader> @dom96 it's not clear why he needs to set the post body manually? but i quickly checked his python version https://github.com/tbrand/which_is_the_fastest/blob/master/python/flask/server.py
22:07:40FromGitter<xmonader> which is just wrong benchmarking anyways because ppl run servers with uwsgi anyways
22:09:24FromGitter<Varriount> shashlick: I'm available for nagging. :D
22:09:49FromGitter<Varriount> shashlick: You can ping me in gitter, preferably. I get way too many emails from github.
22:12:25FromGitter<Yardanico> Lol, what, Apple deprecated OpenGL and OpenGL in macOS 10.14
22:13:44FromGitter<Yardanico> https://developer.apple.com/macos/whats-new/
22:14:12shashlickvarriount: sounds good 👍
22:17:35dom96Yardanico: :'(
22:19:54*krux02 joined #nim
22:20:52*rockcavera is now known as Guest56914
22:20:52*tiorock joined #nim
22:20:52*Guest56914 quit (Killed (tolkien.freenode.net (Nickname regained by services)))
22:20:52*tiorock is now known as rockcavera
22:24:06dom96Network Framework, hrm, interesting
22:36:21FromDiscord<ethanpmorgan> Are there any libs for gif creation?
22:38:31dom96https://github.com/ftsf/gifenc https://github.com/rxi/gifwriter
22:39:07FromDiscord<ethanpmorgan> Ah thanks Dom
22:48:05*natrys quit (Quit: natrys)
22:58:09FromGitter<honewatson> Is there anything for generating travis configs/Readme/Changelogs for nimble projects?
23:04:27*lainon quit (Ping timeout: 240 seconds)
23:12:59*find0x90 joined #nim
23:25:49FromGitter<zetashift> I'm stumbling upon a proc defined as ``` proc `position = `*(self: Node2D, val: Vector2) ``` how the hell do I use it
23:26:27FromGitter<zetashift> I'm trying to use it like ```position = velocity * speed``` but then I get an error saying position can't be assigned to
23:27:34FromGitter<data-man> @zetashift: ```proc `position= `*(self: var Node2D, val: Vector2)``` ?
23:28:25FromGitter<zetashift> I meant using this: https://pragmagic.github.io/godot-nim/v0.5.3/godotapi/node_2d.html#=,Node2D,Vector2 proc in code
23:28:54FromGitter<zetashift> or calling that proc actually
23:31:04FromGitter<zetashift> So how I call a proc in those backticks(I've seen it before used as ``` `type` ``` but the ``` position = ``` confuses me a bit
23:31:40FromGitter<data-man> Why there is space between ```position``` and ```=```?
23:31:59FromGitter<zetashift> no idea :D
23:32:20FromGitter<data-man> Try remove it. :)
23:34:04FromGitter<zetashift> ah thanks
23:34:21FromGitter<data-man> Good night!
23:35:36FromGitter<zetashift> yep you too
23:42:36*lainon joined #nim
23:42:59*lainon is now known as Guest28788
23:43:12*find0x90 quit (Quit: find0x90)
23:45:57*yglukhov_ quit (Ping timeout: 240 seconds)
23:50:53*Guest28788 quit (Quit: Leaving)
23:55:14*leorize quit (Read error: Connection reset by peer)