iiinova — where your digital self keeps evolving

Jonas Kruckenberg @jonaskruckenberg.de · 24/07/2026

Last year was a blast!! make sure to reserve your Impl Day ticket when you buy a conf ticket!

031

Jonas Kruckenberg @jonaskruckenberg.de · 21/06/2026

been a bit quiet recently, but back with a treat! k23 now has proper ISO images that you can boot at home!! #osdev

141

Jonas Kruckenberg @jonaskruckenberg.de · 18/06/2026

me and my airline standard meal, hanging in madrid! If youre traveling back from @webengineshackfest.org I will literally be here all day 🫠🫠

030

Jonas Kruckenberg @jonaskruckenberg.de · 14/05/2026

new record people, dont think your target folder can get much bigger than this

a rust target folder that is 158,51 GB large

140

Jonas Kruckenberg @jonaskruckenberg.de · 21/02/2026

I unfortunately think the best friend shaped computers is this, too bad its a harbinger of the panopticon :/

110

Jonas Kruckenberg @jonaskruckenberg.de · 17/02/2026

arrived for @rustnationuk.bsky.social !! looking forward to the workshop day tomorrow 👀

020

Jonas Kruckenberg @jonaskruckenberg.de · 26/12/2025

This years christmas challenge: getting this bad girl booting properly 👀

0130

Jonas Kruckenberg @jonaskruckenberg.de · 23/12/2025

This neighborhood belongs to us!! (the macos kernel)

A photo of a building with a large, weathered “XNU” graffiti

050

Jonas Kruckenberg @jonaskruckenberg.de · 12/12/2025

this is basically the setup: editor with remote connection (through tailscale) to my macmini. So I basically have the exact same experience as before but my terminal is switched out

220

Jonas Kruckenberg @jonaskruckenberg.de · 11/12/2025

Ladies and gentleman: The NixOS mac mini 🔥

2270

Jonas Kruckenberg @jonaskruckenberg.de · 22/11/2025

3 times???? One movie was bad enough but three? my lord

030

Jonas Kruckenberg @jonaskruckenberg.de · 05/11/2025

Mainmatter is looking for an experienced Rust developer to join one of our projects as a freelancer. We're migrating a database from C to Rust. The project is complex and performance-sensitive. That's what makes it fun! If you're interested, send an email to the address in the screenshot.

031

Jonas Kruckenberg @jonaskruckenberg.de · 01/11/2025

Film is back from the lab! Heres a few from the @eurorust.eu impl Day this year!

051

Jonas Kruckenberg @jonaskruckenberg.de · 19/08/2025

you‘ve heard of don’t drink & drive. Now get ready for

010

Jonas Kruckenberg @jonaskruckenberg.de · 14/08/2025

On my way to @rustforgeconf.com 🥳 I gave a good feeling about this one (you can still get tickets 👀)

020

Jonas Kruckenberg @jonaskruckenberg.de · 10/07/2025

@github.com don’t attribute automatic messages to me like that pls

Screenshot of GH mobile rendering the automatic “the base branch was changed” messages as a comment made by the PR author (me)

For comparison a screenshot of the website of the same PR showing a better version. still looks weird but at least distinguishable from a real comment

040

Jonas Kruckenberg @jonaskruckenberg.de · 06/07/2025

Still crazy to me, I live literally next door to the Seven Mountains yes of the seven dwarfs fame

picture of fields and in the background a number (7) of smallish hills, which are in fact hills and not mountains even though they are called “the seven mountains” (people living here back in the days probably never seen *actual* mountains so like)

210

Jonas Kruckenberg @jonaskruckenberg.de · 12/05/2025

her: oh noo are you not feeling well? :( me: yeah :/ I got a really terrible

030

Jonas Kruckenberg @jonaskruckenberg.de · 05/05/2025

@github.com … we’ve been through this! You cant just tell chatGPT „pls generate valid JSON 👉👈“ THATS NOT HOW THIS WORKS

screenshot of a GitHub copilot error that says: „Understood“ is not valid JSON

131

Jonas Kruckenberg @jonaskruckenberg.de · 28/04/2025

noticed a person mentioning my prosthetic and Im afraid its quite boring! It has been missing since birth and I get cool robot arms as replacement (thanks universal healthcare!)

130

Jonas Kruckenberg @jonaskruckenberg.de · 18/04/2025

dyeing eggs with the family 🐰

a pot of differently colored bags with eggs in them in a pot

020

Jonas Kruckenberg @jonaskruckenberg.de · 09/04/2025

hell yeeeaahhh k23 finally has working host functions!!!!!!! you have no idea how excited i am

131

Jonas Kruckenberg @jonaskruckenberg.de · 01/04/2025

great to see! thanks for being so proactive

031

Jonas Kruckenberg @jonaskruckenberg.de · 26/03/2025

crazy to think that this was me *yesterday* and tomorrow i’ll be talking about wasm a lot

foreground: a photo of me, hiking backpack on, carrying sleeping mat on a small hill
background: the mountains of madeira

050

Jonas Kruckenberg @jonaskruckenberg.de · 18/03/2025

she: so what do you do for a living? me: i have a

010

Jonas Kruckenberg @jonaskruckenberg.de · 09/03/2025

Logistic growth: the natural enemy of finance bros

000

Jonas Kruckenberg @jonaskruckenberg.de · 03/03/2025

Challenge for my #rust friends! Can you figure out what the issue is?

Rust code snippet thats hard to describe, but here’s the original source code https://github.com/JonasKruckenberg/k23/blob/99a88bdff29d30654ea2b9b714ea3298300ab466/kernel/src/wasm/compile/mod.rs#L300-L307

210

Jonas Kruckenberg @jonaskruckenberg.de · 10/12/2024

it’s similar to the previous where only part of the range is valid. With the difference that here our operation spans a hole. if we were to be maximally strict users would have to split their requests :/

a virtmem operation where the range spans two mapped regions, but also a hole in between the two

100

Jonas Kruckenberg @jonaskruckenberg.de · 10/12/2024

but its not so obvious in other cases, below for example part of the operation actually exists. Do we also fail in this scenario or do we silently ignore the non-existent region? ie are we maximally strict or as helpful as possible?

a virtmem operation where the first half is mapped, but the latter half of the range is unmapped

100

Jonas Kruckenberg @jonaskruckenberg.de · 10/12/2024

but consider this case where the user is attempting to operate on a non-existent range. What do we do? the obvious thing here is to reject the operation and fail with an error.

a virtmem operation where the entire range is unmapped

100

Jonas Kruckenberg @jonaskruckenberg.de · 10/12/2024

friends, help me with something: when working with virtual memory all operations work on contiguous regions. In the example deciding what to do is easy because the range is the same as the region. #osdev

A contiguous virtual memory region from 0x1000 to 0x3000 with a memory operation that spans exactly the same ramge

100

Jonas Kruckenberg @jonaskruckenberg.de · 28/11/2024

And here we are today! 63k weekly downloads and 42 contributors 🤯🤯 absolutely insane

170

Jonas Kruckenberg @jonaskruckenberg.de · 23/11/2024

how bad of an idea is it to have an iterator expose the underlying thing in rust? im thinking this but idkk

000

Jonas Kruckenberg @jonaskruckenberg.de · 21/11/2024

Good morning people! New profile picture to celebrate moving to this new app

010

Jonas Kruckenberg @jonaskruckenberg.de · 13/11/2024

Here’s a sneap peek at the #rustlang dortmund meetup on the 27th As the introduction I will be sharing stories from the trenches building an #os and #wasm runtime in #rust and after that we want to hear your stories! There will be food and drinks as well! www.meetup.com/rust-dortmund/

Talk title slide: k23 next generation webassembly operating system

021

Jonas Kruckenberg @jonaskruckenberg.de · 11/11/2024

This is what life is all about, horrifically cursed inline assembly to make a somewhat nice API :)

GH pr screenshot: In essence we use ugly non-local control transfer (setjmp/longjmp) to turn ugly non-local control transfers (hardware traps) back into nice, friendly (and easy to reason about) Rust results returned from the call function.

120

Jonas Kruckenberg @jonaskruckenberg.de · 08/11/2024

You can see a few interesting things, mostly that time per byte actually decreases dramatically with larger payloads which is a good sign (see screenshot below) You can also see that at the very end of the largest benchmark we get weird outliers which might be alloc related? more to investigate

010