In an era where Linux packaging systems proliferate and fragment; every other day there's a new packaging format or a new package manager. Hobby Distros & Mainstream distros alike continue to keep reinventing the wheel that only addresses their problems. Existing solutions like flatpaks, homebrew & snaps etc continue to play favourites, ignoring alternative LIBC & only supporting a handful of the big distros. They have become gatekeepers while not addressing any of their core issues. Even if one of these existing solution is adopted by everyone, it still will not solve the problem of pulling in a zillion dependencies, bloating everything or requiring root access just to install applications that don't even need root. Meanwhile solutions like NixOs (NixPkgs) are so bloated that they end up recreating a distro within a distro.
Soar stands as a beacon of simplicity, portability, and accessibility. We envision a world where software packaging transcends the boundaries of distributions, where users don't have to waste their time waiting for:
- Use 100 different package managers to manage 100 different kind of packaging formats, only to watch promising projects get abandoned months later
- Packages to finish compiling from source when binary options should exist
- Upstream maintainers to package the latest versions, forcing users to stick with broken, outdated & insecure packages
- Be forced to use specific distributions solely because they're the only ones that package certain software
Our mission is to make Linux packaging truly portable, simple, and distro-independent.
We aim to:
- Dismantle the barriers imposed by incompatible packaging systems
- Writing packaging recipes should be simple, easy & avoid unneccessary complexity (SBUILD)
- Packages once compiled should work universally across ALL distributions, ALL LIBC implementations & require minimal to zero additional dependencies to
just work
- Have the most recent, the latest & greatest versions of applications whilst still offering stable versions
- No more using bleeding edge rolling distros that break your system just because you wanted newer version of apps
- Maintain both cutting-edge and stable versions & let the user choose what they prefer
- Not reinvent things that don't need reinventing
- Let distro's official package managers do what they do best i.e handle core system tools/libraries
- Avoid depending/linking on core system tools/libraries by providing truly standalone packages that
just work- Coexist with existing package managers by avoiding conflicts, being completely functional in userspace & using XDG Specifications.
- Provide a unified, streamlined approach that benefits the entire Linux ecosystem
- Become the true Linux User Repository by becoming Distro Agnostic; Users should be free to use whatever distro/system they like to use & get the same packages to
just work
- Minimalism where it makes sense
- Compile packages with sensible profiles: MUSL for lightweightness, mimalloc & LTO for performance, ASLR/PIE for security
- Use container formats for dependency-heavy packages; while this increases application size, the tradeoff is worthwhile in an era where storage is abundant but time is precious
- Speed
- Provide prebuilt binary cache & also make this as fast as possible, so installing a package is only limited by your bandwidth & disk IO, not by our packaging tooling
- Transparency & Security at Core
- Keep an unforgeable record of any & all changes for all our source code, tooling & more
- Provide reproducible (or nearly reproducible) builds & artifacts
- Provide CI logs, checksums, provenance, signing & more for each & every package
If any of these principles resonates with your experience and vision for Linux packaging, join us in this revolution by becoming a part of the Soar community.
- Discord: https://docs.pkgforge.dev/contact/chat
- Discussions: https://github.com/pkgforge/soarpkgs/discussions
- Docs: https://docs.pkgforge.dev/repositories/soarpkgs
Together, let's make Linux packaging soar.