Skip to content

DISTRO_COMPATIBILITY.md

Latest commit

 

History

History
557 lines (428 loc) · 21.1 KB

DISTRO_COMPATIBILITY.md

File metadata and controls

557 lines (428 loc) · 21.1 KB
Raw

Linux Distribution Compatibility

Implementation Status

The cross-distribution compatibility improvements have been implemented in phases:

Phase Status Description
Phase 1 ✅ Complete Dynamic GI_TYPELIB_PATH detection using pkg-config, multi-arch support, ALSA library fallbacks
Phase 2 ✅ Complete Enhanced distro detection (Gentoo, Alpine, Void, Solus, Mageia)
Phase 3 ✅ Complete System dependency checker, improved error messages
Phase 4 🟡 Partial pkg-config as a core dependency ✅. The CI distro matrix covers six containers, but it checks detection logic, shell syntax and --help — it never runs the installer, and the container jobs are continue-on-error. Real install runs are tracked as phase 2.4 of #701
Phase 5 ✅ Complete Fixed remaining hardcoded GI_TYPELIB_PATH values in install.sh and CI workflow
Phase 6 ✅ Complete Added wrapper script verification tests, updated documentation
Phase 7 ✅ Complete Flatpak packaging (whisper.cpp engine) for universal distribution support — see packaging/flatpak/. Not on Flathub: submission flathub#9368 closed 2026-07-23 on policy grounds; channel tracked in #167.
Phase 8 🚧 In progress Snap recipe in-repo + Store listing live (snapcraft.io/vocalinux); install --edge and pack/upload steps in docs/INSTALL.md

Technical Implementation

Dynamic GI_TYPELIB_PATH Detection

The installer now uses a robust multi-step approach to detect the correct GI_TYPELIB_PATH across different distributions:

  1. Primary Method: Uses pkg-config --variable=typelibdir gobject-introspection-1.0 (most reliable)

  2. Fallback Paths: Checks common distribution-specific paths in priority order:

    • Ubuntu/Debian multi-arch: /usr/lib/x86_64-linux-gnu/girepository-1.0
    • ARM64: /usr/lib/aarch64-linux-gnu/girepository-1.0
    • Fedora/RHEL: /usr/lib64/girepository-1.0
    • Arch/Generic: /usr/lib/girepository-1.0
    • Local installs: /usr/local/lib/girepository-1.0

  3. Architecture Support: Detected paths for x86_64, ARM64, ARMHF, RISC-V, POWER, and s390x

Wrapper Scripts

The installer creates wrapper scripts (~/.local/bin/vocalinux and ~/.local/bin/vocalinux-gui) that:

  • Set the correct GI_TYPELIB_PATH environment variable
  • Handle input group permissions for Wayland keyboard shortcuts
  • Provide seamless cross-distro compatibility

Desktop Entry

The .desktop entry is automatically configured with the detected GI_TYPELIB_PATH, ensuring the application launches correctly from the application menu regardless of distribution.

Officially Supported Distributions

These distributions are tested and known to work well with Vocalinux. A derivative is judged by the Python it ships rather than by its own version number, which rarely tracks the base release, and install.sh checks the interpreter rather than the release label.

Distribution Version Status Notes
Ubuntu 24.04+ ✅ Full Support 22.04 ships Python 3.10, below the 3.11 floor
Debian 12+, Testing ✅ Full Support Primary target; use Ayatana AppIndicator
Fedora 42+ ✅ Good Support Uses lib64 paths; 42 is the oldest Fedora upstream still supports, and the one in the CI matrix
Arch Linux Rolling ✅ Good Support Community tested, rolling release compatible
openSUSE Tumbleweed ✅ Good Support Good support with zypper package manager
Linux Mint On Ubuntu 24.04+ ✅ Good Support Mint 22 qualifies; Mint 21 is on Ubuntu 22.04, below the floor
Pop!_OS 24.04+ ✅ Good Support Uses Ubuntu's own numbering
elementary OS On Ubuntu 24.04+ ✅ Good Support elementary 8 qualifies; 7 is on Ubuntu 22.04
Zorin OS On Ubuntu 24.04+ ✅ Good Support Zorin 17 is on Ubuntu 22.04, below the floor
Manjaro Rolling ✅ Good Support Arch base, inherits Arch compatibility
EndeavourOS Rolling ✅ Good Support Arch base, inherits Arch compatibility

Experimental Support

These distributions may work but have known limitations or are less tested:

Distribution Status Known Issues
Gentoo ⚠️ Experimental No package manager support, manual install required. Packages are compiled from source which takes longer.
Alpine Linux ⚠️ Experimental Uses musl libc (not glibc). Some Python packages may not have pre-built wheels and may need to be compiled.
Void Linux ⚠️ Experimental Untested, manual install required. Uses xbps package manager.
Solus ⚠️ Experimental Untested, manual install required. Uses eopkg package manager.
Mageia ⚠️ Experimental Untested, uses dnf/urpmi package managers.

Not Supported

These distributions are known to be incompatible:

Distribution Status Reason
NixOS ❌ Not Supported Completely different filesystem layout (/nix/store), incompatible with standard installer

Requirements by Distribution

Ubuntu/Debian-based

sudo apt install -y python3-gi gir1.2-gtk-3.0 gir1.2-gdkpixbuf-2.0 \
  portaudio19-dev python3-dev python3-venv pkg-config

Note for Debian 12+: Use gir1.2-ayatanaappindicator3-0.1 instead of gir1.2-appindicator3-0.1

Debian-specific notes

Debian's standard repositories differ from Ubuntu in a few ways that matter for Vocalinux:

pywhispercpp build prerequisites — On a clean Debian install the following packages are not pulled in transitively (unlike Ubuntu) but are required when pywhispercpp must be compiled from source (e.g. for GPU support):

sudo apt install -y libssl-dev autoconf automake libtool patchelf

The installer now installs these automatically, but if you hit a CMake error like Could not find OpenSSL during a manual reinstall, add the above first.

ydotool — ydotool is not packaged in Debian's standard repos. The installer falls back gracefully to IBus or wtype for most Wayland compositors. KDE Plasma Wayland users should first select IBus Wayland in System Settings -> Keyboard -> Virtual Keyboard. If you specifically need ydotool as a fallback, compile it from source:

sudo apt install -y git cmake libevdev-dev
git clone https://github.com/ReimuNotMoe/ydotool.git /tmp/ydotool
cmake -S /tmp/ydotool -B /tmp/ydotool/build
sudo cmake --build /tmp/ydotool/build --target install
sudo systemctl enable --now ydotoold

Scoped source builds — If you need to force pywhispercpp to rebuild from source, use the package-scoped flag to avoid compiling unrelated deps like NumPy from source (which takes a very long time and can fail):

source ~/.local/share/vocalinux/venv/bin/activate
PYWHISPERCPP_CLEAN=1 pip install --force-reinstall --no-binary=pywhispercpp pywhispercpp
deactivate

Verifying libwhisper.so resolution — If Vocalinux starts with libwhisper.so.1: cannot open shared object file, check for unresolved symbols:

ldd $(find ~/.local/share/vocalinux/venv -name '*.so' -path '*/pywhispercpp*' 2>/dev/null | head -1) | grep 'not found'

If any libraries are listed, re-run the installer with --rebuild-whispercpp or switch to the VOSK engine (--engine=vosk) as a fully-packaged alternative.

Fedora/RHEL-based

sudo dnf install -y python3-gobject gtk3-devel gobject-introspection-devel \
  portaudio-devel python3-devel python3-virtualenv pkg-config

Arch Linux-based

sudo pacman -S --needed python-gobject gtk3 gobject-introspection \
  portaudio python pkg-config

openSUSE

PYVER=$(python3 -c 'import sys; print(f"python{sys.version_info.major}{sys.version_info.minor}")')

sudo zypper install -y \
  "${PYVER}-pip" "${PYVER}-gobject" "${PYVER}-gobject-cairo" \
  "${PYVER}-devel" "${PYVER}-virtualenv" \
  gtk3 typelib-1_0-AyatanaAppIndicator3-0_1 libayatana-appindicator3-1 \
  typelib-1_0-Notify-0_7 libnotify4 \
  gobject-introspection-devel portaudio-devel pkg-config cmake wget curl unzip \
  xdotool wtype

# Optional: only needed for whisper.cpp Vulkan GPU builds
sudo zypper install -y vulkan-tools vulkan-devel shaderc

Note: openSUSE Tumbleweed often uses versioned Python package names such as python313-devel. The -devel suffix means development headers for compiling native dependencies; it does not mean beta or unstable packages. If ${PYVER}-virtualenv is unavailable on your snapshot, try ${PYVER}-venv.

Gentoo

sudo emerge dev-python/pygobject:3 x11-libs/gtk+:3 dev-libs/libayatana-appindicator \
  media-libs/portaudio dev-lang/python:3.11

Note: Gentoo compiles packages from source, which will take longer.

Alpine Linux

sudo apk add py3-gobject3 py3-pip gtk+3.0 py3-cairo portaudio-dev \
  py3-virtualenv pkgconf

Note: Alpine uses musl libc. Some Python packages may need to be compiled from source.

Void Linux

sudo xbps-install -Sy python3-pip python3-gobject gtk+3 libappindicator \
  gobject-introspection portaudio-devel python3-devel pkg-config

Solus

sudo eopkg install python3-pip python3-gobject gtk3 libappindicator \
  gobject-introspection-devel portaudio-devel python3-virtualenv pkg-config

Mageia

sudo dnf install -y python3-gobject gtk3-devel gobject-introspection-devel \
  portaudio-devel python3-devel python3-virtualenv pkg-config
# or
sudo urpmi -y python3-gobject gtk3-devel gobject-introspection-devel \
  portaudio-devel python3-devel python3-virtualenv pkg-config

PyPI, pip, and pipx Installation

The PyPI package is useful when you want a standard Python package install, or when you want to manage Vocalinux with pipx. However, Vocalinux requires system desktop packages that cannot be installed by pip or pipx.

Use the official installer when possible. It installs the required system packages, creates the virtual environment, installs Vocalinux, sets up desktop integration, and downloads the default speech model:

curl -fsSL https://raw.githubusercontent.com/VocaHQ/vocalinux/main/install.sh | bash

Important: Install System Packages First

Before running pip install vocalinux, pip install "vocalinux[vosk]", or pipx install vocalinux, you must install the required system packages:

Ubuntu/Debian

sudo apt install python3-gi gir1.2-gtk-3.0 gir1.2-ayatanaappindicator3-0.1 \
  libgirepository1.0-dev libcairo2-dev portaudio19-dev python3-dev \
  python3-venv pkg-config xdotool wtype wl-clipboard xclip xsel

On Ubuntu 24.04+ or Pop!_OS, install libgirepository-2.0-dev if libgirepository1.0-dev is not available.

Fedora

sudo dnf install python3-gobject gtk3 gtk3-devel libayatana-appindicator-gtk3 \
  gobject-introspection-devel portaudio-devel python3-devel \
  python3-virtualenv pkg-config xdotool wtype wl-clipboard xclip xsel

Arch Linux

sudo pacman -S python-gobject gtk3 libayatana-appindicator gobject-introspection \
  python-cairo portaudio python-virtualenv pkg-config xdotool wtype wl-clipboard xclip xsel

openSUSE

PYVER=$(python3 -c 'import sys; print(f"python{sys.version_info.major}{sys.version_info.minor}")')

sudo zypper install "${PYVER}-gobject" "${PYVER}-gobject-cairo" gtk3 \
  typelib-1_0-AyatanaAppIndicator3-0_1 libayatana-appindicator3-1 \
  typelib-1_0-Notify-0_7 libnotify4 \
  portaudio-devel "${PYVER}-devel" "${PYVER}-virtualenv" pkg-config xdotool wtype wl-clipboard xclip xsel

pip Installation Steps

  1. Install system packages (see above for your distribution)

  2. Create a virtual environment with system site-packages enabled:

    python3 -m venv ~/.local/share/vocalinux-pypi/venv --system-site-packages
source ~/.local/share/vocalinux-pypi/venv/bin/activate
pip install --upgrade pip setuptools wheel

    --system-site-packages lets the venv use distro-provided GTK bindings such as python3-gi, which are often more reliable than building PyGObject from source in a venv.

  3. Install Vocalinux from PyPI:

    pip install vocalinux
# vosk engine: pip install "vocalinux[vosk]"

  4. Run Vocalinux:

    vocalinux

pipx Installation Steps

  1. Install system packages (see above for your distribution)

  2. Install pipx (if not already installed):

    # Ubuntu/Debian
sudo apt install pipx
pipx ensurepath

# Fedora
sudo dnf install pipx
pipx ensurepath

# Arch Linux
sudo pacman -S pipx
pipx ensurepath

  3. Install Vocalinux:

    pipx install vocalinux
# vosk engine: pipx install "vocalinux[vosk]"

  4. Run Vocalinux:

    vocalinux

Why System Packages Are Required

Vocalinux uses GTK3 for its GUI, AppIndicator/Ayatana for the system tray icon, PortAudio for microphone capture, and tools such as xdotool, wtype, and clipboard utilities (wl-clipboard, xclip, xsel) for text injection and layout-independent Wayland paste fallback. These are system libraries and desktop tools that must be installed with your distribution's package manager. Python packages from PyPI can provide Python bindings, but they do not provide the underlying GTK typelibs, tray support, audio development libraries, or text injection binaries.

Troubleshooting pip/pipx Installation

If vocalinux fails to start:

  1. Check for missing system packages - The error message will indicate which packages are missing
  2. Verify GI_TYPELIB_PATH - In rare cases, you may need to set this environment variable: 
    export GI_TYPELIB_PATH=/usr/lib/x86_64-linux-gnu/girepository-1.0
vocalinux
  3. Check package metadata - Run pip show vocalinux or pipx runpip vocalinux show vocalinux to confirm which version was installed.
  4. Check Python dependencies - Run pip check inside the venv, or pipx runpip vocalinux check for pipx.

If Vocalinux reports that a speech recognition model is missing, open Settings and download a model, or use the official installer to download the default model during setup.

Recommended Alternative

For the best experience with automatic dependency handling, use the official installer:

curl -fsSL https://raw.githubusercontent.com/VocaHQ/vocalinux/main/install.sh | bash

This installer automatically detects your distribution and installs all required system packages.

Text Injection Tools

Vocalinux requires text injection tools to work with X11 or Wayland:

For X11

  • xdotool: Required for X11 sessions

For Wayland

  • IBus: Recommended for KDE Plasma Wayland and generally the most reliable direct text input path
  • wtype: Recommended for wlroots-style Wayland compositors such as sway
  • ydotool: Universal alternative that works with both X11 and Wayland. Current distro packages use the per-user ydotool.service; source builds can install the system-level ydotoold.service. Vocalinux uses clipboard paste with wl-clipboard, xclip, or xsel to avoid non-US layout scrambling.
  • xdotool: May work via XWayland fallback

Installation by Distribution

# Ubuntu/Debian/Fedora/Arch/openSUSE
sudo apt install xdotool wtype wl-clipboard xclip xsel     # Ubuntu/Debian
sudo dnf install xdotool wtype wl-clipboard xclip xsel     # Fedora
sudo pacman -S xdotool wtype wl-clipboard xclip xsel       # Arch
sudo zypper install xdotool wtype wl-clipboard xclip xsel  # openSUSE

Ubuntu releases which package ydotool install a user unit named ydotool.service. Enable it only for the account running Vocalinux:

sudo apt install ydotool
sudo usermod -aG input $USER  # then log out and back in
systemctl --user enable --now ydotool.service

Do not use sudo systemctl --global enable ydotool.service. Global user-unit enablement also starts the input-injection daemon for the display-manager greeter and other accounts. If it was previously enabled globally, repair the setup with:

sudo systemctl --global disable ydotool.service
systemctl --user enable --now ydotool.service

The command is different for the Debian source-build path documented above: the upstream install supplies the system unit ydotoold.service, so sudo systemctl enable --now ydotoold.service is correct only for that route.

Manual Installation

For unsupported or experimental distributions:

  1. Run the dependency checker:

    bash scripts/check-system-deps.sh

  2. Install missing dependencies using your package manager based on the checker output

  3. Run the installer with --skip-system-deps:

    ./install.sh --skip-system-deps

  4. Or install from source in a virtual environment:

    python3 -m venv venv
source venv/bin/activate
pip install -e .[whisper]

Desktop Environment Compatibility

Desktop Environment Status Notes
GNOME ✅ Full Support Primary target, uses AppIndicator
KDE Plasma ✅ Full Support On Wayland, select IBus Wayland in System Settings -> Keyboard -> Virtual Keyboard
Xfce ✅ Full Support Works with AppIndicator
Cinnamon ✅ Full Support Works with AppIndicator
MATE ✅ Full Support Works with AppIndicator
LXQt ✅ Full Support Works with AppIndicator
sway ✅ Good Support Wayland native, uses wtype
i3 ✅ Good Support X11, uses xdotool
Hyprland ⚠️ Experimental Wayland, may need wtype configuration

Autostart Compatibility

Vocalinux uses XDG autostart desktop entries for "Start on Login".

  • Desktop-session autostart file location:
    • $XDG_CONFIG_HOME/autostart/vocalinux.desktop, or
    • ~/.config/autostart/vocalinux.desktop (fallback)
  • This launches Vocalinux as a user GUI app after login.
  • Vocalinux does not create a systemd service/unit for this feature.

Practical caveat:

  • Mainstream desktop environments generally honor XDG autostart.
  • Minimal/custom WM sessions may need an explicit autostart manager or WM startup hook.

Display Server

Display Server Status Notes
X11 ✅ Full Support Uses xdotool for text injection
Wayland ✅ Full Support Uses IBus, wtype, ydotool, or xdotool fallback depending on compositor
XWayland ✅ Full Support Automatically detected, falls back to xdotool if needed

Architecture Support

Architecture Status Notes
x86_64 (AMD64) ✅ Full Support Primary target
arm64 (AArch64) ✅ Good Support Tested on Raspberry Pi, ARM64 servers
armhf ⚠️ Experimental Limited testing
riscv64 ⚠️ Experimental Limited testing, may need to compile dependencies

Reporting Issues

When reporting issues, please include:

  1. Distribution information:

    cat /etc/os-release

  2. Desktop environment:

    echo $XDG_CURRENT_DESKTOP
echo $SESSION_TYPE

  3. Dependency checker output:

    bash scripts/check-system-deps.sh

  4. Python version:

    python3 --version

  5. Error messages or logs:

    • Any error messages from the installer
    • Output from running vocalinux-gui from terminal
    • Relevant journal logs if applicable

  6. Steps to reproduce:

    • What you were trying to do
    • What happened instead
    • Expected behavior

Testing Checklist

Use this checklist when testing on a new distribution:

Pre-Install

  • Run dependency checker: bash scripts/check-system-deps.sh
  • Note missing packages
  • Check desktop environment
  • Check display server (X11/Wayland)

Installation

  • Run installer: ./install.sh
  • Check for errors during installation
  • Verify virtual environment was created
  • Check wrapper scripts exist

Post-Install

  • Launch from terminal: ~/.local/bin/vocalinux-gui
  • Launch from application menu
  • Verify tray icon appears
  • Test keyboard shortcuts (Hold Right Alt by default)
  • Test audio recording
  • Test text injection in different applications
  • Test Settings dialog opens correctly
  • Test model download if using Whisper

Known Issues to Watch For

  • AppIndicator not showing in tray
  • Audio device not detected
  • Text injection not working
  • Model download failures
  • Permission errors with udev/input

Contributing Compatibility Fixes

If you successfully get Vocalinux working on an unsupported or experimental distribution:

  1. Test thoroughly using the checklist above
  2. Document any distribution-specific requirements
  3. Submit a PR with:
    • Distribution detection logic (if needed)
    • Package installation commands
    • Any patches or workarounds required
    • Update to this compatibility document

See Also