02:01 <discocaml> <Ada> think i’ve found a funny bug in ocamlbuild

02:02 <discocaml> <Ada> on the diskuv windows distribution it appears to be replacing the \t in my username with a tab 😆

04:28 <discocaml> <bluddy5> Moonbit doesn't surprise me. The most important constraint on programming languages in the real world is social psychology. We all live with some compromises so that we can be part of a community that builds an ecosystem, and that takes a lot of time and effort. The authors of Rescript never accepted that, and decided to wreck the Reason community for the sake of having a language that was more to their liking. It's not surprising that they ma

04:55 <discocaml> <Hongbo> Rescript is in the good hands of the community. Moonbit is a completely new project focused on WASM

04:59 <discocaml> <Hongbo> Most important pieces of work on Reason/Rescript were built by me, and I gave it back to the community for free, I shift my interest into Wasm these days

05:07 <discocaml> <Hongbo> I have been working on rescript for almost 7 years, it is well maintained by the community, time to do something new for me

05:21 <discocaml> <Hongbo> When I started BuckleScript, my goal is to make FP mainstream which means 10-100x more users than OCaml, it's never my goal to make Reason/ReScript a part of the OCaml , we had a very ambitious goal. ReScript itself went pretty far and were working pretty good without me these days, I feel a bit relieved and decide to work on something new

08:00 <discocaml> <JM> It's your time, and you are right to spend however you like.

08:00 <discocaml> <JM> But if the goal is going mainstream, sticking with the community even if the language doesn't go as far as you like might be necessary. Most BDFL of mainstream languages had to have tough discussion with their community over the years about language development. Being able to negotiate these successfully is what enables them to be mainstream.

08:00 <discocaml> <JM> Anyway I wish you great success with Moonbit! 👍

12:27 <discocaml> <bluddy5> Rescript split off from Reason 3 years ago. That's *nothing* in language years. As far as I can tell, the split caused the Reason momentum (which was very beneficial to OCaml) to grind to a halt. Reason *may* survive thanks to Melange, but it will never thrive now.

12:27 <discocaml> <bluddy5> You are free to do whatever you want, but your choices have consequences. You *chose* to split off from Reason and create a very similar-sounding and looking language. You now *chose* to leave that new fork alone and move on to something else.

12:27 <discocaml> <bluddy5> Toy languages can work like this, but languages in the real world cannot. Ecosystems require buy-in and long-term stability. Any show of instability causes the countless people and companies involved to change their minds about the huge amount of money and effort they are and were going to invest.

12:27 <discocaml> <bluddy5> I can only say that I would personally stay far away from any language initiative you're involved with.

12:40 <discocaml> <Hongbo> I would be happy if Reason or Melange thrive since it is still my work. You are never in the inner circle or does not really what had happened

13:46 <companion_cube> Isn't Reason the work of Jordan initially?

13:46 <companion_cube> (ofc I'm not talking about BS)

14:40 <discocaml> <Hongbo> if you havs some experience in language tooling, you know that the syntax is the least complex part. The most valuable part of Reason or Melange is Bucklescript.

14:42 <discocaml> <Hongbo> Since we worked in the same company, they were called REASON to avoid confusion

14:43 <companion_cube> @bluddy5 was talking about the benefits of Reason to OCaml. Afaict that was mostly the syntax part.

14:44 <companion_cube> OCaml already has the rest of the language tooling.

14:45 <companion_cube> BS did help with popularizing it for frontend devs, but I don't think that was the only appeal of Reason

14:45 <companion_cube> (also Reason was started at FB and BS at bloomberg, no??)

14:46 <discocaml> <Hongbo> That part was intact, I didn't do anything to it. Most Reason users are actually using Bucklescript, otherwise why the Melange fork

14:47 <companion_cube> that's a logic error

14:47 <companion_cube> melange shows that some people were using BS and wanted a way forward

14:47 <companion_cube> doesn't mean other people aren't using reason

14:51 <discocaml> <Hongbo> I wish you love reason native and enjoy it. i did not do anything with that part

15:34 <discocaml> <vphantom> Yeah that's why I got curious about Melange: for what we did, moving away from OCaml syntax was not a good fit, so we abandoned BuckleScript when it moved to its own syntax and only recently re-discovered Melange as more or less a continuity where BS left off.

17:30 <discocaml> <yendor> Oh it’s a fork of the rescript compiler. Was curious how it compared to js_of_ocaml

17:42 <discocaml> <JM> It's a fork of bucklescript (an OCaml to JS compiler), the rescript compiler is also a fork of bucklescript but things have diverged since.

17:44 <discocaml> <yendor> Yeah I knew rescript was continuation of bucklescript but melange docs says it’s a fork of rescript.

17:44 <discocaml> <yendor>

17:44 <discocaml> <yendor> Ever since the rebranding to rescript and then reason things are really weird to google from a year or 2 ago

17:50 <discocaml> <vphantom> Doesn't help that they use common nounds.

17:51 <discocaml> <vphantom> Doesn't help that they use common nouns.

17:51 <discocaml> <vphantom> Doesn't help that they use common nouns as brands.

17:51 <discocaml> <JM> I guess if it was fork after the rename but before any major change it's about the same thing... but antonio would know these details better 🤷

17:51 <discocaml> <vphantom> They forked the very earliest ReScript, before the divergence from OCaml.

17:51 <discocaml> <vphantom> They forked the very earliest ReScript, before the divergence from OCaml. Then they went their separate way to get Dune/OPAM friendly.

17:52 <discocaml> <vphantom> They forked the very earliest ReScript, before the divergence from OCaml. Then they went their separate way to become Dune/OPAM friendly.

17:52 <discocaml> <JM> Yeah, `reason` is slightly better than `go` in term of keyword search, but not by much 😂

17:52 <companion_cube> (please don't edit messages on this channel, it messes up the IRC bridge)

17:52 <discocaml> <vphantom> There's a reason all these languages end up with a "lang" suffix like "GoLang", "RustLang", "ReasonML", "ReScriptLang" in order to be searchable again. Unfortunately the suffix is not mandatory so it's still hit & miss.

17:53 <discocaml> <JM> companion_cube: Are reply ok? I don't remember

17:53 <companion_cube> sure, it seems so

17:53 <discocaml> <JM> 👍

17:53 <companion_cube> we just don't see where they come from iirc

17:53 <discocaml> <vphantom> Random thought: the IRC bridge should ignore events that cause issues.

17:54 <companion_cube> I'm not reimplementing a bridge just for that feature

17:54 <discocaml> <JM> And you'll be left with the unedited message? probably not what the author intended

17:55 <discocaml> <companion_cube> test reply

17:55 <discocaml> <vphantom> Wait, why rewrite?

17:56 <companion_cube> yeah can't see what it replies to

17:56 <companion_cube> or contribute, whatever. I just run the thing.

17:56 <discocaml> <vphantom> Roger that.

17:56 <discocaml> <eduardorfs> Antonio still tries to cherry pick whenever ReScript does something interesting to the OCaml community tho

17:57 <discocaml> <vphantom> Oh I thought he was forbidden to do that. Good.

17:58 <companion_cube> yep

17:58 <companion_cube> woops

18:56 <greenbagels> companion_cube: what bridge do you use?

18:58 <companion_cube> matterbridge

19:07 <discocaml> <anmonteiro> It’s a fork of BuckleScript around summer 2020, and as Eduardo said we cherry pick interesting improvements (mostly around runtime / stdlib); there were some interesting code generation patches too. We haven’t picked up any changes since releasing 1.0

19:23 <greenbagels> What do you guys use for writing TUIs in ocaml? Notty?

20:49 <discocaml> <JM> Yeah, give that a try

20:49 <discocaml> <JM> Tarides has a nice intro https://tarides.com/blog/2020-09-24-building-portable-user-interfaces-with-nottui-and-lwd/

20:50 <discocaml> <Kali> i've used notty, but i've heard lambda-term is nicer (i've never used notty extensively and haven't used lambda-term at all, though)

20:57 <discocaml> <JM> Does lambda term has UI/widgets? I remember it being lower level than that

21:45 <discocaml> <lecondorduplateau> lambda term has a plenty

21:46 <discocaml> <lecondorduplateau> it's hard to just draw on a canvas

21:46 <discocaml> <lecondorduplateau> lambda term is also not declarative, unlike notty 😔