The article ArkScript September 2025 update is the last one I wrote, covering all the changes I made on the language this summer.

Finally, I have released this huge set of breaking changes that makes ArkScript v4, and I'm pretty proud of it. I won't stop working on the language, however it's a big milestone for me: I've reach a point where the language is more than decent to use every day, errors are correctly reported, and the documentation is pretty good too (I might be biaised, I wrote it myself so I don't have an objective point of view): https://arkscript-lang.dev/

I've also written an article comparing ArkScript with other Lisps (which is still a WIP but is already good enough) for the curious ones here.

As many others here, I have a home lab at home, with various containers like FreshRSS, Ampache…

I also have a netdata dashboard to monitor CPU and temps, disk usage… that sometimes send me alerts without me having configured anything, eg too much CPU used for more than 15 minutes.

However it doesn’t seem to cover log monitoring, or at least not in the way I want. I have a job and can’t dedicate thousands of hours to building something myself, nor configuring deeply some software stack.

All I want is my services to be monitored log-wise, with a single docker where you could mount multiple log directories, and have a simple interface that filters through the logs (based on their type/name, eg nginx logs aren’t treated the same way as kernel or auth logs, but without me having to configure more than the source type), to tell me if something is weird or just bad (eg someone logged in).

Does it exist without installing grafana + Prometheus + this and that + doing a shit ton of configuration and crying?

I’ve been working on this (not so little anymore) project for some time now, and I’m finally happy with the branding, UX and docs state.

It’s a scripting language I made at first as a toy, to learn new parsing methods, explore compiler optimizations, and go back to VM land where everything is low level and amazing (at least for me) ; it’s now a fully fledged language that can be used as a scripting language like Python or Ruby, and can also be very easily embedded inside a project, as one would do with Lua.

Let me know your thoughts and opinions on the project!

I’ve been working on this (not so little anymore) project for some time now, and I’m finally happy with the branding, UX and docs state.

It’s a scripting language I made at first as a toy, to learn new parsing methods, explore compiler optimizations, and go back to VM land where everything is low level and amazing (at least for me) ; it’s now a fully fledged language that can be used as a scripting language like Python or Ruby, and can also be very easily embedded inside a project, as one would do with Lua.

Let me know your thoughts and opinions on the project!

ArkScript is an interpreted/compiled language since it runs on a VM. For a long time, runtime error messages looked like garbage, presenting the user with an error string like "type error: expected Number got Nil" and some internal VM info (instruction, page, and stack pointers). Then, you had to guess where the error occurred.

I have wondered for a long time how that could be improved, and I only started working on that a few weeks ago. This post is about how I added source tracking to the generated bytecode, to enhance my error messages.

I finally found a better memory layout to store variables in ArkScript, and I got a 76% performance boost on the binary tree benchmark, and a 21% perf boost on Ackermann(3, 7) Who knew using a contiguous storage buffer could be beneficial? 🤡

I retraced all the performance improvements I applied to ArkScript through the last five years, with updated benchmarks, AND DAMN what a journey.

Generating swaggers at compile time

Hi everyone!

I’m sharing with you a solution I designed for generating swaggers (http4s, tapir, open api) for apps.

At work we always had to remember to launch the app and all the databases containers, which was cumbersome and we would often forget to update the swaggers (which led to generated code for clients that wasn’t up to date).

I wanted to design a funny keyboard with an alternative to TRRS, so I made this floppy disk sized keyboard! (Perfect replica, under 10cm x 10cm)

I made a build guide for it too: https://lexp.lt/posts/floppy_keyboard/

I tried accessing https://programming.dev/c/programming_languages but it tells me that the community can not be found. Is that a lemmy bug?