OCaml Weekly News

Previous Week Up Next Week


Here is the latest OCaml Weekly News, for the week of December 26, 2023 to January 02, 2024.

Table of Contents

castore 0.0.1 – a portable CA Store with no dependencies

ostera announced

Hi folks! :wave: happy to announce castore 0.0.1.

tl;dr I’m taking a page from the Elixir community’s playbook here, the Ca_store module includes an up-to-date public certificate chain from a generally trustworthy source (eg. Mozilla) that we’ll update via CI and publish automatically to opam periodically.

How to use it Easy, just opam install castore and in your dune-project make sure to use (castore (>= "0.0.0")) so you automatically upgrade to the latest certificate.

Now you can use it with ocaml-tls and when you need that .pem file you can feed it the contents of Ca_store.pem.

Why we did this I was building an HTTP client for Riot and realized that to support TLS I’d need to either have custom certificates or bring in ca-store and let it resolve them from the system.

The Elixir community’s approach to this is a lot simpler.

What’s missing/next The latest .pem file was updated on Dec 12th and I need to build the scheduled CI workflow that’ll update it / publish the lib, so if you’re into crypto (maybe i can nerdsnipe @hannes?) or ci (@ulrikstrid?) then ping me :)

Happy hacking! :camel:

/ Leandro

ostera later added

I spent a little time consolidating that preprocessing code into castore, and 0.0.2 is on its way on opam.

It’ll let you write this:

let decode_pem ca =
  let ca = Cstruct.of_string ca in
  let cert = X509.Certificate.decode_pem ca in
  Result.get_ok cert
let cas = List.map decode_pem Ca_store.certificates in
let authenticator = X509.Authenticator.chain_of_trust ~time cas in
let tls_config = Tls.Config.client ~authenticator () in
(* ... *)

fixgen 0.1.0 - a CLI tool for generating fixtures

Marc Coquand announced

Heya everyone!

This weekend I had some fun and built a fixture generator in OCaml! It is my very first DSL.

I found it very time consuming to set up test data, especially when dealing with relational data, where some data generated needs to reference some other data. Therefore, I wanted a slim DSL that can be used to quickly generate test data which I can import it into SQL or NoSQL.

So to give an example:

~: fixgen 'user (2): id uuidv4, name name; order (5): id uuidv4, buyer user.id, amount int(1,40), currency


FILE NAME: order

It can also export to JSON!

~: fixgen 'user (2): id uuidv4, name name; purchase (5): id uuidv4, buyer user.id, amount int(1,40), currency
("USD","MXN")' -f json

[{"name": "Ivan", "id": "f6c90f96-faad-4741-a52b-6d53b94d48f3"}, {"name": "Judy", "id":

FILE NAME: purchase
[{"currency": "USD", "amount": 14, "buyer": "f6c90f96-faad-4741-a52b-6d53b94d48f3", "id":
"8e59075c-1a6a-4093-86ca-e1135afd34f3"}, {"currency": "MXN", "amount": 19, "buyer":
"57eccc5c-ad09-4229-b0c7-a828280dba1b", "id": "fd89ad30-af07-4c9d-9ec4-ade7583e9cf0"}, {"currency": "MXN",
"amount": 17, "buyer": "f6c90f96-faad-4741-a52b-6d53b94d48f3", "id": "b15cd3d7-ede1-4612-b0f2-a36cc1a73f1c"},
{"currency": "MXN", "amount": 32, "buyer": "f6c90f96-faad-4741-a52b-6d53b94d48f3", "id":
"295a85b8-42d9-49a4-8ce6-e4d7b1a94ea9"}, {"currency": "USD", "amount": 29, "buyer":
"f6c90f96-faad-4741-a52b-6d53b94d48f3", "id": "925eff95-b7a2-4cbe-93b9-b519372a16cf"}]

For now it doesn’t have many generators (I plan to add more), the docs are a bit sparse and I don’t know if I shared the compiled binary correctly. However, I wanted to share it early to gather feedback!

8 months of OCaml after 8 years of Haskell in production

Dmitrii Kovanikov announced

Hi everyone :wave:

I’ve been using Haskell in production for the previous 8 years. But for the last 8 months, I used OCaml at Bloomberg.

I wrote a blog post comparing two languages from my POV:

I tried to provide a reasonable and balanced overview of these two languages, although my experience in both is significantly different. I hope this blog post can give a rough idea of both languages :relieved:

Enjoy! And let me know any feedback you have!

riot 0.0.7 – an actor-model multi-core scheduler for OCaml 5

ostera announced

Hi folks :wave: we skipped over 0.0.6 because of a CI bug, and ended up releasing Riot 0.0.7 instead :slight_smile:

Here’s a summary of the changes.

Single public package

The public surface of the Riot package should be limited now to the top-level `Riot` module.

Introduce dedicated I/O Scheduler

To improve long-tail latency when doing I/O, we’re splitting out the poll loop out of the main schedulers and using a dedicated thread for polling I/O.

Improved I/O primitives

The runtime now ships with an IO and a Buffer module that includes more low-level functions:

  • to do vectorized reads/writes,
  • await for file descriptor availability (for reading, writing, or both),
  • copy data between buffers
  • write buffers directly to writers

I/O Readers & Writers

We’re introducing Reader/Writer to build Read/Write streaming pipelines like you’d do in Rust with the Read/Write traits.

This one is particularly exciting because we can now write super clean transformations of data that can be made super efficient, and are tracked at the type-level, like this:

let file = File.open_read "test" in (* [ `r ] File.t *)
let read = File.to_reader file in (*  [ `r ] File.t Reader.t  *)
let decrypt = Decrypt.of_reader read in (* [ `r ] File.t Decrypt.t Reader.t *)
let unzip = Unzip.of_reader read in (* [ `r ] File.t Decrypt.t Unzip.t Reader.t *)
let data (* IO.Buffer.t *) =
  let buf = IO.Buffer.with_capacity 1024 in
  let* len = IO.Reader.read unzip ~buf in
  IO.Buffer.sub buf ~len
(* ... *)

This will pull 1024 bytes out of this stream, lazily reading backwards through the chain. And if you want it buffered, you just gotta call. IO.Reader.buffered reader and you’re good to go.

These new interfaces are available for the new File module as well as for the Net.Socket module.

Spatial Shell · A windows switcher for i3 and sway implemented in OCaml

Thomas Letan announced

Hi folks :wave: I have been willing to share Spatial Shell here for a while now, mostly because it is implemented in OCaml so it’s a good way for me to find potential contributors ahah (unfortunately, the codebase could be more commented tbh). Spatial Shell implements a spatial model inspired by Material Shell for i3 and sway. I have very recently tagged its 6th release, and I believe now is a good time to try and put it in the hands of other human beings. What might go wrong? :sweat_smile:

The README should be comprehensive enough to get you started, and I’ve also published a short blogpost giving a bit more background on my Spatial Shell journey, and has a nice video you can witch if you are curious.

If you run into problems while trying to test, do not hesitate to reach out (GitHub issues are probably the best way to do so). I do hope Spatial Shell becomes a usable programs for anyone interested in the workflow it enables on i3/sway deskop.

Other OCaml News

From the ocaml.org blog

Here are links from many OCaml blogs aggregated at the ocaml.org blog.


If you happen to miss a CWN, you can send me a message and I’ll mail it to you, or go take a look at the archive or the RSS feed of the archives.

If you also wish to receive it every week by mail, you may subscribe to the caml-list.