kitty is great, but I stopped using it because iTerm's terminal search feature - which I use almost daily - is much more ergonomic. The existing solution in kitty (pipe scrollback buffer to something like less/vim and search with /) is a lot worse IMO which is unfortunately a dealbreaker for me

I have a repository where I make classic programs (game of life, sudoku solver, brainfuck interpreter, etc) but entirely within Typescript's type system

https://github.com/ruyili/cursed-typescript/

> First of all, there are no otters here. I'm utterly sorry about that.

incredibly important suggestion: replace "utterly" with "otterly"

did you miss the part there are no otter?

Awesome work on this, it looks like a game changer!

From my (admittedly limited) knowledge on this space, it seems like this is a straight upgrade over WASM, with the only limiting factor being that your stack is limited to Bun + C. Are there any downsides of this feature when compared with the alternatives mentioned at the start of the article? There are some tradeoffs listed at the bottom but I'm not sure how napi/WASM perform in those aspects either.

NAPI and WASM let you use optimizing compilers, so they're suitable for use with larger C and C++ codebases

but that'll be at runtime

Starting off with "In the realm of..." instantly gave it away

Maybe they're gamers :) https://en.wikipedia.org/wiki/Forgotten_Realms

the combat is intense, it really felt like i was slowly getting pickaxed to death

this gives me an idea for a new typescript type level project

This is hilarious, I'm changing all my websites to have "Made with cold indifference by $NAME" in the footer now

I think I’ll use “Made with slight annoyance”.

Hand crafted with mild arrogance

I think I'll try "Made with bellowing rage".

I write quite a few things purely fueled by rage

Would unikernels be feasible for online judge platforms (e.g. codeforces, leetcode)? I was thinking something along the lines of: spawning a unikernel for each submission -> running a single file program -> streaming the output back to a server, all in <3s and potentially with multiple submissions running in parallel. I'm fairly new to this, so I'm not sure if that would be overkill.

Yes, it would be overkill.

Online judge platform do a very small subset of problem. You can sandbox it to no network or filesystem accesses, and no syscalls except the few like read/write/select.

Yes, if your goal is to use unikernels. But if your goal is to pick the best solution for this application, there isn't an obvious advantage for unikernels.

Why wouldn’t you use a standard container based approach like Docker?

The last time I tried, it was pretty slow to fire up, especially when I tried to run multiple submissions at once. It's also even more overkill for my use case; I currently use isolate (https://github.com/ioi/isolate) which is just a wrapper around cgroups/namespaces, and it's been a lot faster.

Sidenote: I'm not really looking to replace it, I was just asking out of curiosity since this is my first time hearing of unikernels

> I currently use isolate (https://github.com/ioi/isolate) which is just a wrapper around cgroups/namespaces, and it's been a lot faster.

Yes. This is the fastest you can get.

If you want safer, add pr_set_seccomp _in addition_ to it. but that would be a custom solution.