Ah, missed the line in the article about secondary tariffs. Thanks for pointing it out to me. A 100% tariff on China and India until they stop importing Russian oil would indeed be very disruptive

What I’m missing from the articles is putting the potential consequences of the threatened tariffs into context. In 2024, Russian imports into the USA were 3.27 billion USD, whereas the GDP of the Russian economy in 2024 was 2.17 trillion. Even if there is fakery in the reported GDP, exports to the USA are likely less than 1% of the Russian economy

German soldiers can refuse orders under certain conditions (e.g. if they against human dignity)

I prefer KeePass over Bitwarden because it is just a simple database file, less that can go wrong (no server component).

I am the original author of the Rust library for decrypting and modifying KeePass databases.. The current best implementation of KeePass, KeePassXC, is written in C++, so there could theoretically be security-relevant memory corruption bugs in it (though the developers of the project are excellent and I don’t think it is super likely). Rust is a language that does not have that class of issues by design, so I thought it would be interesting to see how far I could get. So far, I am still having fun and adding features bit by bit, and it is quite cool to me to be able to write one codebase that deploys to Windows, Linux, MacOS, Android (potentially iOS), and any modern web browser.

Our son is fortunately very relaxed, he eats and sleeps a lot so I can get some coding done while he is sleeping. Germany has decent parental leave, so my partner and I are both not working the first two months of his life.

Thanks! Our son is a bit less than a month old. The wife, our son, and https://omnikee.github.io/ are three different projects 😂

I’m taking care of a newborn and doing some FOSS work, so that project has been deprioritized for now 😅

This is exciting. My only request here is: whenever it works please release a standalone wasm file somewhere (anywhere). So many projects either require building the wasm themselves, or instead of releasing a .wasm, they release a JS wrapper that auto-loads the wasm/wasm-imports. Its a pain to try to extract the wasm out of those projects.

What I am doing is to create a omnikee-lib crate within the project that will get compiled to WASM, not just plain keepass, because I need additional adapter methods to interface with the web part of the application. I don’t have the bandwidth to turn keepass into a general WASM package that could be npm installed at the moment. As I am dogfooding the crate, I might get to a point where I know what a good JS interface for it would be, though, and the omnikee-lib crate could become the official WASM interface for keepass.

sweet! I sent you the invite.

Currently, SSH key management is not supported, but it would probably be possible to implement the SSH agent protocol in the Rust part of the application. I see that russh has a SSH agent server implementation. Let me know if you are interested in contributing such a feature - I am currently working on exposing all the custom entry fields in the UI, so the project would almost be ready. edit: would be ready to add that feature now

thanks for your interest! I have sent you a response with an invite link.

I am administering several other docker servers and a k8s cluster from the command line, so I’m well aware what I’m missing 😀 - in this case, I was hoping for a higher wife approval factor, which is at least partially there.

Thanks for the portainer on unraid tip. I set up portainer itself yesterday but will have to get around to migrating the 30 or so already deployed containers to it.

Since it doesn’t come installed by default on a fresh system, my guess would be that you won’t break anything fundamental, but this is pure speculation.

Yeah, I’m currently running unraid on it because I wanted a hands-off maintenance experience.

While it’s nice to get started, I’m really missing even intermediate Docker features such as support for compose files (so that there is some grouping of main services with the database instance that supports it, etc). Still, it’s been working reliably for the year that I’ve had it.

Edit: I have tried the Docker Compose Manger plugin but didn’t find the experience an improvement because of the way the YAML editing works

Is keeping the servers where they currently are (or with a friend) an option? Then you could just VPN into it from abroad.

If that isn’t an option, I’m currently running a homebuilt NAS off an Intel N100 Mini-ITX mainboard and I’m impressed with how many services it can run simultaneously, including Quick Sync Video for hardware transcoding.

I think the Google as an identity provider example is misleading. The more common use case will be medium to small companies where several admins/developers need to login to various servers and where manually adding and revoking keys across these servers will be cumbersome.

As the other commenter said, in those cases, the organization would also deploy its own IDP.

The mirrorlist is a configuration file listing servers that updates can get pulled in from.

When a package update is installed that contains a configuration file, it will not overwrite the old file but be installed with a pacnew extension so that you can merge the files (like you did). It will keep complaining at you until you remove the pacnew file, which is fine to do after you have merged successfully.

The graphical issues are probably due to something else that happened during the update.

Thanks for the comment. I have had exposure to similar claims, but wasn’t seeing anyone using AMD GPUs for AI unless they were somehow incentivized by AMD, which made me suspicious.

In principle, more competition in the AI hardware market would be amazing, and Nvidia GPUs do feel overpriced, but I personally don’t want to deal with the struggles of early adoption.

For inference (running previously-trained models that need lots of RAM), the desktop could be useful, but I would be surprised if training anything bigger than toy examples on this hardware would make sense because I expect compute performance to be limited.

Does anyone here have practical recent experience with ROCm and how it compares with the far-more-dominant CUDA? I would imagine that compatibility is much better now that most models are using PyTorch and that is supported, but what is the performance compared to a dedicated Nvidia GPU?

The pre mixed spices list their contents and it’s not that hard to come up with something similar by just using the individual ingredients.

OK, so cases where you control both ends of the communication. Thanks for the clarification.

I’m a developer and would appreciate you going into more specifics about which certificates you suggest pinning.