Building native packages is complicated | Packaging Anubis as native packages
Building native packages is complicated | Packaging Anubis as native packages
Especially if you want a UX as good as Docker images give you.
Anubis provides protection against bots scraping websites and DDoSing projects.
This blog post is about Xe's reasoning for originally only providing docker packages and their work to provide native packages.
Now I get why most server apps are docker only
There must be a tool that allows you to build packages for multiple systems in multiple formats (deb, rpm, nix, flatpak, snap, etc.). Does that not exist? After 20 years of these systems existing, somebody must've tried...
Also, it's clear that once again, open source needs some kind of funding model, because it's a little crazy that a project like this can get so popular so fast, the dev flooded with praise, thanks, and issues but not money to maintain and develop it.
Long time ago I've tried making a makefile which packages as tarball, deb, rpm and appimage.
It's just a goofy attempt.Problem is that distro1 has req-lib2.5.3 while distro2 has req-lib2.7.8, but your project was developed on distro3 with req-lib2.9.5 so you have to deal with every distro having different lib versions and compatibility issues that come with it, not just different packaging formats.
And different names too. Debian and Fedora have different packages names, so req-lib on debian becomes req-devel on Fedora.
Suse's open build system does this. It's just very enterprisy to me, so I haven't really used it myself
I have detected bloat, you YAML merchant you! Reject complexity! Return to native packages!
Lmao
Anubis without JavaScript is what I'm waiting for. I know that the Darknet forum Dread has a PoW system that doesn't use JS (or maybe it does something else entirely and I completely missed it)
How would that work? And how easy would it be to circumvent? Anubis probably forces spinning up a browser or something that supports a JS runtime (again probably a browser), so it's not as easily scriptable as just callling an HTTP endpoint. I'm curious how you would implement a system without JS.
Honestly I have no idea how PoW works at all. I'm just happy to see the innovation in this space. With that said, people have raised the idea of using TOR's PoW mechanism, or in my case, Dread's PoW. I believe both of them work without JavaScript
It might be pretty difficult to implement the work part of proof of work without JS in a practical way. Of the three languages available on the web, HTML, CSS & JS (+ WebAssembly, which requires a bit of JS IIRC & would probably not be available) JS is the only one that allows you to perform the work in a sane way. (It might be possible to use CSS magic with remote resources, but that has its own problems if it's even possible.)
It would be possible to use a dedicated program or another website to perform the work, but it would be far from seamless to users.
I don't know how PoW works but do you think that TOR and Dread's PoW can be reused?