惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
V
Visual Studio Blog
IT之家
IT之家
博客园 - 聂微东
The Cloudflare Blog
月光博客
月光博客
阮一峰的网络日志
阮一峰的网络日志
S
SegmentFault 最新的问题
Apple Machine Learning Research
Apple Machine Learning Research
酷 壳 – CoolShell
酷 壳 – CoolShell
爱范儿
爱范儿
H
Help Net Security
博客园 - 叶小钗
V
V2EX
WordPress大学
WordPress大学
J
Java Code Geeks
Hugging Face - Blog
Hugging Face - Blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
博客园_首页
C
Check Point Blog
B
Blog
D
DataBreaches.Net
美团技术团队
罗磊的独立博客

Hacker News: Show HN

PurrrrrFocus: Pomodoro Timer App - App Store Workflow Engine — Multi-Step Orchestration for Bun RapidPhoto: Pro Photo Editor App - App Store GitHub - DheerG/swarms: Achieve extraordinary results with claude code across a variety of tasks SPICE simulation → oscilloscope → verification with Claude Code — Lucas Gerads Show HN: VCoding – A 5 MB native Windows IDE with no dynamic dependencies Show HN: LLMs don't hallucinate because they're bad at math, it's the format GitHub - Agent-FM/agentfm-core: AgentFM is a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. AgentFM lets you run massive AI workloads directly across a global mesh of idle CPUs and GPUs. Show HN: Tracking Top US Science Olympiad Alumni over Last 25 Years GitHub - Potarix/agent-hub: One place to talk to all your agents Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration) GitHub - dubeyKartikay/lazyspotify: Terminal Spotify client for macOS and Linux GitHub - the-banana-tool/king-louie: Easy to use GUI Personal AI Assistant. Win/Linux/Mac. Show HN I made my vacation rental bookable by AI agents–no Airbnb, 0% commission GitHub - basteez/jsf-autoreload: maven plugin to enable hot reload on jsf projects uvm32/hosts/host-gdbstub at main · ringtailsoftware/uvm32 GitHub - labsai/EDDI: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 12+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus. GitHub - glitchnsec/fortyone-oss: AI Executive Assistant Platform Quickstart | Alien GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. GitHub - ocrbase-hq/ocrbase: 📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable. GitHub - impactjo/home-memory: MCP server that lets your AI assistant remember everything about your home. GitHub - Sets88/dbcls: DbCls is a powerful terminal database client that supports various databases GitHub - neptun2000/heor-agent-mcp GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh RollQuation: Math Puzzles - Apps on Google Play GitHub - dropbox/witchcraft Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis GitHub - opentalon/opentalon: OpenTalon is an open-source platform built from the ground up in Go as a robust alternative to OpenClaw LinkedIn™ 职位抓取工具 - Chrome 应用商店
GitHub - tschk/alpenglow: hyperlightweight minimal diskle...
undivisible · 2026-06-17 · via Hacker News: Show HN

General-purpose musl+LLVM Linux distribution. dinit init, Oil packages. Boots to login in <1s on native virt (x86_64 KVM or aarch64 HVF).

Two deployment modes:

  • Diskless/immutable — initramfs-only, RAM root, GlowFS overlay (appliance mode)
  • Rootfs — normal root-on-disk (ext4/btrfs/zfs), package-managed
scripts/boot-native.sh           # build + boot initramfs (QEMU)
system/backends/appliance/scripts/qemu.sh   # boot existing build

Platform support:

  • x86_64 — main branch (primary target)
  • aarch64 — arch/aarch64 branch — QEMU virt, Apple Silicon HVF
  • riscv64 — arch/riscv64 branch — QEMU virt, OpenSBI
  • Rockchip RK3566 — board/rk3566 branch — PINE64 Quartz64

Quick Start

# Initramfs (diskless) mode — build + boot in QEMU
./scripts/boot-native.sh

# Rootfs mode — install to disk
curl -L alpenglow.example.org/installer | sh   # or:
oil install alpenglow-core

# Custom kernel
KERNEL_BUILD=1 ./scripts/boot-native.sh

Design

Layer Choice Notes
Init dinit Parallel dependency graph, both modes
Userland toybox (838KB), oksh Static musl, no glibc
Package mgr Oil APK-compatible, standalone binary
Kernel Linux 7.0 + Rust modules CONFIG_RUST=y, alpenglow_core.ko
Kernel ctrl kernelctl (Zig, 89KB) Static, µs-scale startup
Network netd (Rust), udhcpc, iwd Zero-external-deps netd
Root FS Diskless: GlowFS/erofs/squashfs in RAM. Rootfs: ext4 over LUKS
Compositor Wayland + cage + foot Optional, not in base
Security AppArmor, read-only root (optional) Hardened by default
Audio ALSA + PipeWire
Kernel Linux 7.0+ with CONFIG_RUST=y, GlowFS in-tree

Project Layout

system/
  backends/
    appliance/          Primary profile (kernel configs, dinit services, scripts)
  kernelctl-zig/        Cgroup + kernel policy (Zig, 89KB static)
  netd/                 Network state daemon (Rust, zero deps)
  glowfsctl-zig/        GlowFS image tooling (Zig, 164KB)
  oil/                  Package manager (Rust, APK-compatible)
  glowfs/               GlowFS kernel module source (C+Rust)
  kernel-modules/       Rust kernel modules (alpenglow_core, alpenglow_bootstat)
  init/                 Zig init (4.8KB static, initramfs fallback)
scripts/                Build, CI, benchmark scripts
docs/                   Architecture, build, install docs

Kernel configs live at system/backends/appliance/kernel/.

Performance

Boot to login (QEMU KVM, quiet)

OS Boot Initramfs Kernel Idle RAM
Alpenglow min 0.6s 1.4K 4.4MB ~17MB
Alpenglow std 1.3s 1.7MB 4.4MB ~26MB
Alpine Linux virt 1.3s 8.7MB 6.5MB ~58MB
Void Linux 2.5s 12MB 7MB ~80MB
Ubuntu Server 15s 40MB 12MB ~200MB

Alpenglow minimal (Zig init, 4.8KB) boots in 0.6s on x86_64 KVM. The standard build (dinit + toybox + getty) is 1.3s. Alpine matches boot speed but has 6000x larger initramfs and 3x the RAM. Both modes use the same toolchain — the difference is just initramfs contents.

Binary size (static musl, x86_64)

Tool Lang Size vs alternative
kernelctl Zig 89KB 501KB (Rust)
glowfsctl Zig 164KB 501KB (Rust)
init Zig 4.8KB 937KB (toybox+sh)
dinit C++ 1.6MB 20MB+ (systemd)
toybox C 838KB 10MB+ (coreutils)
alpenglow_core.ko Rust 9.2K kernel built-in

Modes

Alpenglow runs in two deployment modes sharing the same codebase:

Diskless/Appliance — boot from initramfs, root in RAM (tmpfs), state on persistent partition. Uses GlowFS squashfs for verified immutable root. Target: embedded, edge, kiosk, containers.

Rootfs/Desktop — install to disk (ext4 over LUKS), normal r/w root. dinit manages services, Oil installs packages. Target: workstation, server, development.

The init script auto-detects mode: if /dev/disk/by-label/alpenglow-root exists, it switches root; otherwise runs diskless. Both modes use the same kernel, toolchain, and package format.

Services

Service Status Appliance Desktop Managed by
SSH (dropbear) dinit
NTP (chronyd) dinit
DNS cache (dnsmasq) dinit
Logging (syslogd) dinit
DHCP networking dinit
WiFi (iwd) optional dinit
Wayland + sway optional dinit
Audio (PipeWire) optional dinit
Package manager (Oil) dinit
Kernel policy (kernelctl) dinit
GlowFS mount optional dinit

Status

21/22 milestones complete. Last milestone: real hardware boot (QEMU only for now).

See AGENTS.md for full milestone table and docs/ for architecture docs.