Reason Seoul @ReasonSeoul

@ReasonML Seoul/S.Korea community. Sharing knowledge & Introducing Reason, ReScript and OCaml related stuff to the Korean tech industry. meetup.com/Reason-Seoul Seoul Joined June 2020

Tweets

512
Followers

160
Following

134
Likes

802

ReScript @rescriptlang

2 years ago

Another new feature in v12 - variant type spreads in pattern matching! This allows you to refine variant types and easily match on a whole subset type without having to enumerate it. This was always possible with polyvariants, and now also with regular variants.

4 10 57 3K 4

View Details

Dmitrii Kovanikov @ChShersh

2 years ago

I'm happy to see such huge interest in OCaml and Functional Programming! As promised, here's the OCaml version of this code. But first of all, despite my questions about this Go code, I'd like to acknowledge all the amazing stuff Go devs have built. My favourite is @charmcli: charm.sh You can build such great free OSS tools only if you're passionate about the programming craft, and this is the one thing I deeply respect. As for the particular code, my main beef is the following two points: 1. Error handling is a boilerplate that makes reading the actual business logic harder. 2. Error handling is not composable and reusable. I can't abstract the handling of the error into a separate function. What I want: Throw error only once in the entire app Handle error only once in the entire app This way I achieve the decoupling of error handling from the actual business domain and can solve each problem independently. Go did one thing great: errors as value. This is very functional, exceptions throwing is not functional (even though, my favourite language, OCaml, has exceptions 🤬) FP languages go one step further: they apply monads to error handling to compose errors easily. Monads are not used just for error handling, they're quite versatile. But it just happens that error handling is one of the most popular use cases. Anyway, the proper idiomatic OCaml code for this contrived Go example, if translated one-to-one but with a sprinkle of monadic error handling, looks like this: This code makes a few assumptions but overall it's pretty accurate. I can go one step further and move the printing inside helper functions to refactor things even further and get a nice pipeline-looking of function applications. Or, remember I mentioned that monads are quite versatile. I can easily switch 'Result.Syntax' with something like 'Lwt_result.Syntax' and use 'and*' in addition to 'let*' to get async parallel execution. The sky is the limit! The golden rule of SWE: If the problem is difficult, don't solve it. Make someone else solve it. Error handling is difficult. Instead of handling the error in each case, the caller needs to handle it once at the very end (or, they can delegate it to their callers!).

Dmitrii Kovanikov @ChShersh

2 years ago

To err is human To let* is divine If this post gets 100 likes, I’ll post an OCaml version of this code to show what I mean when I say that a higher-level declarative FP language not just removes boilerplate but offers a different programming paradigm.

24 11 433 68K 62

17 8 172 46K 97

View Details

ReScript @rescriptlang

2 years ago

It's also accompanied by a new release of ReScript Core, the standard library that'll be the default from v12 going forward: github.com/rescript-assoc…

0 2 16 1K 0

View Details

ReScript @rescriptlang

2 years ago

Read more in our release blog post: rescript-lang.org/blog/release-1… And a big thank you to all contributors who's helped make this release possible! Onwards toward v12 🫡

1 2 14 2K 2

View Details

ReScript @rescriptlang

2 years ago

It's our great pleasure to announce the release of ReScript 11.1! This release brings a ton of new features and improvements: - Tagged template literals - JSX for more than React - Native BigInt support - Array spread syntax - %todo annotation for incomplete code ...and more🎉

8 39 169 17K 9

View Details

Antonio Monteiro @_anmonteiro

2 years ago

blogged about our work in Dune to allow sharing code between Melange and native OCaml melange.re/blog/posts/dun…

3 14 45 5K 7

View Details

🦋 hyeseong.kim @cometkim

2 years ago

Use @rescriptlang and no worry

Tzvetan Mikov @tmikov

2 years ago

The perf impact of this is curious. Arguments are very cheap in JS, unlike objects. Using the wrong order will kill the hidden class. A lot of optimizations need to happen to eliminate the object allocation and the property access cost. The function better be hot and inlineable.

31 20 310 170K 135

0 4 7 1K 0

View Details

Yaron (Ron) Minsky @yminsky

3 years ago

Another OCaml Unboxed, talking about why locals matter for performance. It walks through an example of what you need to do to take a bit of ordinary OCaml code, and modify it so that it doesn't need to do any heap allocation. youtu.be/UqBk5DgqxRg?si…

2 14 106 9K 37

View Details

Florian @fham_r

3 years ago

With v11.1 ReScript will finally support tagged template literals! Thank you @tsnobip, for implementing them.

3 13 74 6K 6

View Details

Patrick Ecker @ryyppy

3 years ago

Thanks to the ReScript folks, the @rescriptlang api docs are now generated from source. Took a while to get from my original doc extraction utility prototype to the final product, but the team did a great job getting it into shape and make it actually useful! Looks great too.

2 11 81 5K 5

View Details

🦋 hyeseong.kim @cometkim

3 years ago

JS 타겟 컴파일 언어인 ReScript v11 버전이 출시됐습니다. rescript-lang.org/blog/release-1… 조상격인 언어에서 파생되어 ReScript로 새단장한 이후 언어로써는 가장 급진적인 변화가 있는 버전입니다. 오랫동안 지켜봐왔지만 나날히 더 쓸만한 언어/컴파일러가 되어가고 있습니다. JavaScript 개발하시면서 강타입, 함수형 프로그래밍에 관심있으신 분들라면 꼭 한 번 배워보시길 권합니다.

1 36 73 8K 18

View Details

ReScript @rescriptlang

3 years ago

The day is finally here - ReScript v11 is out! This release is packed with features, including massively improved JS/TS interop capabilities for variants, uncurried by default, dynamic imports, and more. Read about it all in our announcement post: rescript-lang.org/blog/release-1…

4 56 146 29K 10

View Details

sabine 🐫🦀 / 'use ocaml'; @sabine_s_

3 years ago

discuss.ocaml.org/t/the-ocaml-pl… The OCaml Platform Roadmap is Adopted 🔥🐫

7 12 60 15K 8

View Details

Felipe O. Carvalho @_Felipe

3 years ago

I now understand why @jimmykoppel says only ML and OCaml have real "module types" (as defined by Robert Harper – separate compilation, static dispatching). pathsensitive.com/2023/03/module…

2 5 15 2K 7

View Details

David Sancho @davesnx

3 years ago

I finished a fun weekend project that can be useful for rendering HTML Write JSX, render HTML Most features from JSX and a few improvements (lowercase components and no annotations). No React idioms, and type-safe. It works with Reason syntax and MLX. github.com/davesnx/html_o…

2 5 27 5K 4

View Details

Shakthi Kannan @shakthimaan

3 years ago

OCaml Platform Newsletter: September 2023 discuss.ocaml.org/t/ocaml-platfo… #Multicore #OCaml #Compiler @tarides_

1 3 9 964 0

View Details

Harshal Patil @mistyHarsh

3 years ago

Customizable variants is a game changer for @rescriptlang rescript-lang.org/blog/improving… The DU - Discriminated Unions are absolute delight and 100% zero runtime cost. It puts it very close to TypeScript.

1 6 25 3K 3

View Details

Texas Toland @texastoland

3 years ago

I'm excited about OCaml again as an alternative to TypeScript. For me it all started with a project called @reasonml by @jordwalke who also created React. But there's a separate almost identical project called @rescriptlang. Why? How are they different? What about tools? A 🧵

David Sancho @davesnx

3 years ago

The rebrand was BuckleScript to ReScript. Reason was affected, because it was integrated with BuckleScript. Really hard to calculate, but the harm made by a poor announcement and a long story of bad explanations of the tools are insane.

6 1 53 22K 0

1 11 50 11K 18

View Details

Gabriel Nordeborn @_zth_

3 years ago

Another place where unboxed variants in ReScript really enhances the developer experience. This Bun API returns the TypeScript type in the picture. We can model that zero cost in ReScript as in the second picture, and then pattern match on it naturally as in the third picture.