* feat: split playback into separate monitoring and mic streams
AudioPlayer held a single rodio stream. WirePlumber auto-connected its node to
the default sink (that is the monitoring path) while PWSP explicitly linked the
very same node into pwsp-virtual-mic. Both paths were fed by one Player, so
Player::set_volume scaled them together and they could not be separated — hence
the report of "volume 2 is fine on the other side but deafening locally".
There are now two independent streams with independent gains:
mic stream --ensure_route()--> pwsp-virtual-mic
monitoring stream --ensure_route()--> selected Audio/Sink (or WirePlumber's
choice when no device is pinned)
PipeWire side:
- DeviceType::Sink, so sinks are discoverable like inputs already were, plus the
playback_* port mapping that Audio/Sink nodes need (monitor_* stays unmapped:
linking a stream there would be a feedback loop).
- Link globals are tracked in the registry listener; GetLinks/DestroyGlobal make
ensure_route() possible. It points a node at one target and prunes every other
link leaving it, creating before pruning so the node is never left unlinked —
otherwise WirePlumber's autoconnect would re-attach it behind our back.
Replaces link_player_to_virtual_mic().
Audio side:
- PlayerPair wraps the two rodio Players of a track so transport controls reach
both from one place instead of being fanned out at every call site.
- Streams are identified by diffing our own stream nodes before and after opening
one, not by indexing a sorted list: PipeWire reuses freed node ids, so the
stream opened second can end up with the lower id. Only nodes that look like
ours are ever considered — pruning a stranger's node would silence another
application. Opens are therefore sequential, and node discovery checks the
graph before its first sleep to stay off the play latency path.
- Streams are dropped once nothing is playing, as before: an open stream keeps
the audio device busy and stops laptops from suspending. The routing is
rebuilt on the next play.
- Each track decodes twice, once per path. rodio's Buffered is the only shareable
source and it cannot seek, which the position slider depends on.
- effective_gain() is the single home of the master * track * multiplier math and
collapses non-finite or negative gains to silence.
Protocol, config and GUI:
- get/set_monitoring_volume, get/set_mic_volume, get/set_output, get_outputs.
- DaemonConfig gains default_output_name, default_monitoring_volume and
default_mic_volume.
- Volumes arriving over IPC are validated: hotkeys store raw Request JSON, so a
bad value can reach the daemon without passing through the CLI.
- The footer carries a monitoring slider and a mic slider, both running to 200%,
plus an output combo box; with four widgets it no longer fits on one line and
is laid out as two rows.
- SliderLatch owns the "local value wins while the user is interacting" logic
every slider needs — the daemon is polled at 60 Hz, so without it a slider
fights the user mid-drag and snaps back on release. That was previously
copy-pasted per slider as a value/dragged/ignore-until triple.
Fixes along the way: setting the volume multiplier no longer overwrites the
master volume, and get_volume(Some(id)) reports the track's own volume instead of
the product of all three factors.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(cli): add monitoring-volume, mic-volume and output commands
get/set volume keeps its old meaning — without --id it moves both masters at
once, which is the "make everything quieter" shortcut. The new commands address
one path each, and values above 1.0 pass through on purpose: amplifying what
goes into the microphone without deafening yourself is the point.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(gui): move device selection into settings
The footer tried to hold two combo boxes, two sliders and two buttons on one
line. Their labels alone run ~300px, so at the 800px minimum window size the
sliders and buttons were pushed off the edge and simply vanished.
Device pickers are set-once controls and belong next to the theme selector, so
they move to the settings screen. The footer keeps the two volume sliders, which
are the ones worth reaching for mid-call, and now fits at the minimum width.
The right-edge spacer is also clamped at zero: computed negative, it used to
shove the buttons out of view rather than merely crowding them.
Verified with screenshots at 800px and 1200px.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(gui): align the footer volume icons and sliders
The speaker icon and its slider sat a couple of pixels above the microphone
pair. A horizontal layout centres each widget against the row height known when
that widget is placed, so a row that grows while being filled leaves whatever
was added first sitting too high.
Every footer element is now allocated the same box height, which makes the
centring independent of placement order, and the row is given that height up
front. Measured from screenshots at 800px: icon ink centres were 776.5 and
780.5, now both 776.5; slider rails were 2.5px apart, now 0.5px, which is
sub-pixel rounding.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(gui): let the output device be unpinned again
The output picker listed devices but no way back, so choosing one was a one-way
door. It now offers "system default" as the first entry, which sends set_output
with an empty name; the daemon takes that as "stop pinning".
Unpinning deliberately leaves the existing link alone rather than tearing it
down. WirePlumber's autoconnect only runs when a node first appears, so removing
the link would strand the monitoring stream with no output at all. Streams close
as soon as playback stops, so the next sound opens a fresh node that gets routed
like any other application's.
Picking the default target ourselves was tried and reverted: the global
default.audio.sink is only a fallback, and a per-stream target.node overrides it.
On this machine that meant monitoring would have gone straight to the hardware
sink, bypassing the user's EasyEffects chain that WirePlumber routes it through.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor: trim comments and drop an abstraction that earned nothing
Removes the RouteTarget trait and its two unit-struct implementations. They
existed to let route() resolve a target lazily, but both call sites resolve one
line earlier just as well, so a trait, two types and an impl block collapse into
one parameter.
Narrows should_sync, commit and prune_links_from to private: all three were
public but only ever called from within their own module.
Comment pass: drops the decorative section dividers that this file never had,
and rewrites the ones that leaned on a particular machine's setup or narrated a
past bug. What is left explains things the code cannot: why streams close when
idle, why they are opened one at a time, why a link is created before stale ones
are pruned, why playback_* maps to inputs while monitor_* stays unmapped, and
why the footer row height is pinned up front.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* use one global Arc<DaemonConfig>
* implement commands
* replace Arc<DaemonConfig> with Arc<Mutex<DaemonConfig>> and fix pwsp-gui
* add pwsp-cli support
* replace serde_json::to_string_pretty with to_string
* refactor
* implement UpdateDaemonConfigCommand and use it in pwsp-gui so it now uses real in-memory config, not the saved one
* implement utils::gui get_daemon_config and update_daemon_config to simplify code
* move `pwsp-daemon.service` from `pwsp-gui/assets` to pwsp-daemon/assets + autorestart service + better logging + wait for pipewire until start
* adapt all packages
* improve systemd service calling for .deb package
* Fix path traversal vulnerability in audio download logic
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* std::path::Path -> Path
* add random id to the filename if it's unknown
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Avoid unnecessary string cloning and allocation in `pwsp-gui/src/gui/mod.rs` when checking if a file matches the search query.
Specifically, we replaced `.to_string_lossy().to_string()` with `.to_string_lossy()` because `.to_lowercase()` directly works on `Cow<str>`, avoiding an intermediate String allocation.
Measured improvement was verified using an isolated benchmark demonstrating roughly 10-15% reduction in execution time.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* ci: refactor and optimize GitHub Actions workflows\n\n- **Consolidate build processes (`build.yml`, `release.yml`)**: Merged the standalone `.zip` (linux-build) and `.deb` (deb-build) jobs into a single matrix job (`build-and-package` / `build-release`).
This prevents compiling the Rust binaries twice per architecture, effectively cutting the build time in half and saving CI runner minutes.
- **Parallelize and fix Flatpak deployment (`flatter.yml`)**:
- Converted `x86_64` and `aarch64` flatpak jobs to run concurrently via a matrix strategy, removing the sequential bottleneck.
- Resolved an issue where GitHub Pages would only host the `aarch64` repository. Introduced a new `prepare-repo` job that downloads the `.flatpak` bundles from both architectures, merges them into a
single OSTree repository locally, signs it with GPG, and deploys the unified multi-arch repository to GitHub Pages.
- Configured `flatter.yml` to automatically attach the offline `.flatpak` bundles to GitHub Releases when triggered by a release event.
- **Toolchain and cleanup**: Switched hardcoded Rust toolchain version (`1.96.0`) to `stable` across all workflows to prevent manual bumps in the future. Simplified bash scripts and removed redundant
dependency installation steps.
* Extract Copr trigger into a separate job: Copr build now starts immediately in parallel with GitHub Actions builds, eliminating unnecessary waiting time.
* use ubuntu 22.04 to build binaries instead of ubuntu-latest
* fix cache poisoning
* another attemp to fix cache poisoning
* disable cache in setup-rust-toolchain
* use ubuntu-latest
* optimize dependencies in debug profile
* remove --release flag from build.yml and CARGO env vars
* fix flatter names collision
* better flatter key changing
* auto formatting
* add auto updating cargo-sources for dependabot pull requests
* add groups to dependabot.yml
* add github-actions to dependabot
* add ppa deb repo
* add concurrency: gh-pages-deploy
* ⚡ [performance] pre-filter directory contents by supported extension
💡 **What:** Moved the check for supported audio file extensions from the GUI rendering loop into the directory read/caching layer.
🎯 **Why:** The file extension string parsing and check was executing on every frame of the render loop for every file listed, causing unnecessary CPU overhead. By caching the pre-filtered items, we only execute the check once per directory load.
📊 **Measured Improvement:** In a micro-benchmark simulating 10k files (with 50% matching extensions), the unoptimized loop took ~14.2ms to execute, while the optimized loop takes ~5.8ms. This yields a ~59% speed improvement in the iteration logic over the baseline.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* ⚡ [performance] pre-filter directory contents by supported extension
💡 **What:** Moved the check for supported audio file extensions from the GUI rendering loop into the directory read/caching layer.
🎯 **Why:** The file extension string parsing and check was executing on every frame of the render loop for every file listed, causing unnecessary CPU overhead. By caching the pre-filtered items, we only execute the check once per directory load.
📊 **Measured Improvement:** In a micro-benchmark simulating 10k files (with 50% matching extensions), the unoptimized loop took ~14.2ms to execute, while the optimized loop takes ~5.8ms. This yields a ~59% speed improvement in the iteration logic over the baseline.
Flatpak cargo-sources.json regenerated to include criterion dependency to fix CI.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* chore: remove criterion bench dependency from gui app
This removes the `bench.rs` and `criterion` dependencies, which failed to compile in offline flatpak builds.
CI build is now fixed.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* cargo fmt
* deps: update cargo-sources.json
* chore: update cargo lock and flatpak sources
After removing criterion, the cargo lockfile and flatpak sources have been refreshed to properly remove the offline build dependencies issue.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* Revert "chore: update cargo lock and flatpak sources"
This reverts commit 27379d8e78.
[no ci]
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaces the previous approach of spawning new PipeWire main loops and
awaiting a 100ms timeout for every `get_all_devices` call. Instead,
this introduces a `PipeWireManager` which maintains a single, persistent
background thread with a PipeWire main loop and registry listener. The
actor caches `AudioDevice`s and `Port`s continuously in a `HashMap`,
providing instantaneous responses to device queries via a channel.
This architectural change replaces `Sender<Terminate>` with a simple
`PwTerminator` drop-guard to manage the lifetimes of PipeWire objects
(virtual mic and links) via `PwCommand::DestroyObject` without creating
additional OS threads or PipeWire contexts.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* Refactor project into a Cargo workspace with distinct packages
- Created a root `Cargo.toml` defining a workspace.
- Moved `src/types` and `src/utils` into a new `pwsp-lib` crate for shared logic.
- Split binaries into their own crates: `pwsp-daemon`, `pwsp-cli`, and `pwsp-gui`.
- Shifted all dependencies into `[workspace.dependencies]` for centralized version management.
- Updated import paths across all crates (e.g. from `pwsp::` to `pwsp_lib::`).
- Updated build scripts, GitHub actions, Flatpak manifest, and AUR PKGBUILD to support the new workspace structure.
- Ensured no core application logic was altered.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* Fix cargo-deb build process in GitHub actions for workspace architecture
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* Fix cargo-deb asset discovery by using exact target/release paths
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* refactor deps in Cargo.toml files
* fix incorrect assets path
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Split the long continuous block in `draw_footer` into smaller,
modular methods (`draw_mic_selection`, `draw_master_volume`,
`draw_hotkeys_button`, `draw_settings_button`) for better
readability and maintainability.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
- Extracted the main token processing loop body in `commands_loop` into `handle_connection` to resolve deep nesting and improve code readability.
- Improved request reading logic by using `(&mut stream).take(request_len as u64).read_to_end(&mut buffer)` to strictly bound allocation to `request_len` and prevent initialization overhead.
- Passed `cargo fmt` and `cargo clippy`.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Moved redundant struct initialization into `AudioDevice::new` and unified port mapping assignments in an `add_port` method. This removes nesting using early returns and eliminates an unnecessary clone on the hashmap conversion step.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
The `play` method in `src/types/audio_player.rs` previously used `.unwrap()`
directly on `self.stream_handle.as_ref()`. This posed a security/stability risk
where if `stream_handle` was uninitialized or became `None` unexpectedly
despite the prior `ensure_stream()` call, it would cause the thread to panic
and potentially crash the application.
This commit replaces the `.unwrap()` call with `.ok_or_else` to safely handle
the `None` case, returning an `anyhow` error instead of panicking, adhering to
the project's no-panic policy.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaced the monolithic `src/gui/draw.rs` with a new `src/gui/views` directory module.
The GUI drawing logic is now cleanly separated into distinct files:
- `body.rs`
- `footer.rs`
- `header.rs`
- `hotkey_capture.rs`
- `hotkeys.rs`
- `settings.rs`
- `waiting_for_daemon.rs`
This organization vastly improves readability and maintainability without altering functionality. All shared helpers are centralized in `src/gui/views/mod.rs` and imports are strictly managed.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* 🔒 Fix insecure fallback directory and secure file creation
The daemon's fallback runtime directory `get_runtime_dir()` was hardcoded to `/run/pwsp`, creating a risk of shared, insecure access in multi-user systems.
This commit secures the fallback logic by:
1. Creating a user-specific temporary directory (`/tmp/pwsp-$UID`).
2. Ensuring directory creation happens atomically with `0o700` permissions using `std::fs::DirBuilder`.
3. Validating the fallback directory strictly (checking UID, 0o700 permissions, and symlink status) if it already exists to mitigate symlink attacks.
4. Using `libc::geteuid()` for robust cross-platform UID extraction.
5. Fixing `is_daemon_running` and locking logic to use `fs::OpenOptions` instead of `fs::File::create` to prevent accidental file truncation on active lock files.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* 🔒 Fix insecure fallback directory and secure file creation
The daemon's fallback runtime directory `get_runtime_dir()` was hardcoded to `/run/pwsp`, creating a risk of shared, insecure access in multi-user systems.
This commit secures the fallback logic by:
1. Creating a user-specific temporary directory (`/tmp/pwsp-$UID`).
2. Ensuring directory creation happens atomically with `0o700` permissions using `std::fs::DirBuilder`.
3. Validating the fallback directory strictly (checking UID, 0o700 permissions, and symlink status) if it already exists to mitigate symlink attacks.
4. Using safe `rustix::process::geteuid()` for robust cross-platform UID extraction, avoiding `unsafe` blocks.
5. Fixing `is_daemon_running` and locking logic to use `fs::OpenOptions` instead of `fs::File::create` to prevent accidental file truncation on active lock files.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* small refactor
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
- Extracted search field rendering to `draw_files_search_field`
- Extracted list rendering to `draw_files_list`
- Split `draw_tree_node` file and directory branch logic to `draw_tree_node_file` and `draw_tree_node_dir`
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
This commit optimizes the GUI render loop in `src/gui/draw.rs` during the rendering of the drag and drop directory list. Previously, `self.app_state.dirs.clone()` was cloning the entire vector of `PathBuf`s on every frame, which caused unnecessary allocations.
Now, `std::mem::take` temporarily removes the list of directories from `app_state.dirs` inside `show_vec`, and items are passed by reference rather than being cloned (`let path = item;` instead of `item.clone()`). Finally, the original list is restored into `app_state.dirs`. To ensure the state doesn't mutate or invalidate when `self.open_dir(&path)` is clicked, this logic has been deferred to run after the `app_state` vector is restored.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
This optimization removes an unnecessary `.cloned()` call inside `draw_hotkeys_table`
which previously forced a clone of every filtered `HotkeySlot` on every frame render.
Instead, we now hold a `Vec<&HotkeySlot>` and only clone `slot.slot` exactly when
a user interaction requires ownership to dispatch a `HotkeyAction`.
This eliminates constant heap allocations of `String` and `Request` components
while scrolling or idling in the hotkeys view.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
- Break down the monolithic `draw_hotkeys` method into smaller,
focused component functions: `draw_hotkeys_header`,
`draw_hotkeys_search`, `draw_hotkeys_table`, and
`handle_hotkey_action`.
- Improve readability and maintainability of the `src/gui/draw.rs` file
while preserving identical behavior.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Extracted distinct UI sections (playback controls, position slider, volume controls, stop button) into their own well-scoped helper functions within `SoundpadGui`. This significantly improves the maintainability and readability of `draw_track_control` while preserving the existing layout structure and state mutation behavior.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* perf: optimize UI rendering loop by removing unnecessary Vec clone\n\n- Removed `clone()` on `self.audio_player_state.tracks` in `draw_header`\n- Iterated by reference instead of using an owned collection\n- Benchmarked and showed a significant performance improvement (7us -> 87ns)
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* build(flatpak): update cargo-sources.json to include criterion\n\nThe CI failed during the offline flatpak build because the newly added `criterion` dev-dependency was missing from `cargo-sources.json`. Regenerated `packages/flatpak/cargo-sources.json` to fix it.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* Delete benches/ui_benchmark.rs
* refactor: remove garbage
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
This commit addresses a security and stability vulnerability where failures in PipeWire context setup, connection, or registry acquisition would crash the entire thread (or daemon) via `.expect()` panics.
We now gracefully capture and propagate initialization errors up the call stack. A `sync_channel(0)` is used to signal the success or failure of the initial pipewire setup back to the calling functions (`get_all_devices`, `create_virtual_mic`, `create_link`). This prevents unexpected crashes and improves error resilience.
Also removed unneeded `pw_sender` panics on channel termination by simply dropping/ignoring the result.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Tarasov AleksandrGitHubgoogle-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
* chore(ci): Add flatter to host Flatpak repo on GitHub Pages
- Update release.yml to not upload .flatpak file to releases
- Create flatter.yml to automate building and hosting of Flatpak via GitHub pages using andyholmes/flatter
- Add nightly branch for main pushes and stable branch for releases
- Update README.md with the new Flatpak installation instructions
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* Potential fix for pull request finding 'CodeQL / Workflow does not contain permissions'
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
* fix: truncate file button text in draw function so footer is no clipped
* fix(gui): fix hotkeys table clipping with egui_extras::TableBuilder
fully reworked hotkeys page
* deps: update flatpak cargo-sources.json
Add unit tests for parse_command in src/utils/commands.rs to ensure
robust handling of set_volume edge cases including missing or
invalid volume and id arguments.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaced `.chars().next().unwrap()` with `.chars().next().is_some_and(...)` in `chord_from_event` and `parse_chord` functions in `src/gui/input.rs`. This ensures that even if the string is empty, the application will not panic, adhering to the project's safety guidelines and resolving a potential security vulnerability.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Extract duplicated `pipewire::init()`, `MainLoopRc::new()`, and `ContextRc::new()` setup code from `pw_get_global_objects_thread`, `create_virtual_mic`, and `create_link` into a shared `setup_pipewire_context` helper in `src/utils/pipewire.rs`. Also ran codebase-wide linters to improve code quality.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Extract sections from the long `handle_input` function into smaller,
context-specific helper methods such as `handle_hotkey_assignment`,
`handle_toggles`, `handle_playback_and_focus`, `handle_file_playback`,
`handle_navigation`, and `handle_hotkey_triggers`. This significantly
improves the maintainability and readability of `src/gui/input.rs`
while preserving original functionality.
In addition, ran `cargo clippy --fix` on the project to resolve a few
other minor health issues, like collapsing nested `if` statements and
reducing unnecessary allocations.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Changed `HotkeyConfig::find_conflicts` to return a `Vec<(&str, &str)>` rather than allocating owned `Strings`. In `src/gui/draw.rs`, the code now builds a `HashSet<&str>` directly from the borrowed strings using array-based flat-mapping, avoiding intermediate `Vec` allocations and redundant clones.
Benchmarked to be approximately 3.5x faster in scenarios involving many configured slots.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* chore(flatpak): prepare application for Flathub submission
- Change Flatpak filesystem permissions from `host` to `home` to comply with Flathub sandbox rules
- Remove `--share=network` build argument and implement offline Rust building using `flatpak-cargo-generator.py`
- Add generated `cargo-sources.json` to the Flatpak manifest to download dependencies
- Update `CARGO_HOME` environment variable to ensure vendored dependencies are found by `cargo build --offline`
- Update `.desktop` categories to meet Flathub specs (`AudioVideo` requirement)
- Add required `<releases>`, `bugtracker`, `vcs-browser` URLs, and valid `<developer>` tags to `metainfo.xml`
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* chore(flatpak): prepare application for Flathub submission
- Change Flatpak filesystem permissions from `host` to `home` to comply with Flathub sandbox rules
- Remove `--share=network` build argument and implement offline Rust building using `flatpak-cargo-generator.py`
- Add generated `cargo-sources.json` to the Flatpak manifest to download dependencies
- Explicitly set `CARGO_HOME=$PWD/cargo` in build-commands to ensure vendored dependencies are found by `cargo build --offline`
- Update `.desktop` categories to meet Flathub specs (`AudioVideo` requirement)
- Add required `<releases>`, `bugtracker`, `vcs-browser` URLs, and valid `<developer>` tags to `metainfo.xml`
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* chore(flatpak): add script to generate cargo sources
- Added `packages/flatpak/generate-sources.sh` to automate the generation of `cargo-sources.json`
- Script downloads the `flatpak-cargo-generator.py` tool from upstream, generates the offline sources map based on `Cargo.lock`, and cleans up after itself
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* Fix virtual mic audio linking by managing it in AudioPlayer lifecycle
- Moved `link_player_to_virtual_mic` to `src/utils/pipewire.rs` and updated it to return a termination sender.
- Added `player_link_sender` to `AudioPlayer` to manage the PipeWire link between the daemon and the virtual mic.
- Integrated linking logic into `AudioPlayer::play` and `AudioPlayer::update` to ensure the link is established when audio starts playing.
- Ensured the link is terminated in `AudioPlayer::drop_stream` when the audio sink is closed.
- Removed redundant and potentially failing startup linking loop from the daemon.
- Fixed log spam by ensuring `link_player` is only attempted when necessary and errors are handled gracefully.
- Maintained compatibility with stable Rust by avoiding unstable features.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* small refactor
* refactor
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaced the large match blocks in `chord_from_event` and `parse_chord`
with `egui::Key::name()` and `egui::Key::from_name()`. This drastically
reduces boilerplate code while maintaining the existing behavior that
strictly allows only single-character alphanumeric keys and 'F' keys for
chords.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Addresses a security vulnerability where the daemon or client could be
forced to allocate up to 10MB of memory per malformed socket message,
potentially leading to Out-Of-Memory (OOM) crashes.
Changes:
- Introduced a central `MAX_MESSAGE_SIZE` constant of 128KB in `src/types/socket.rs`.
- Enforced the 128KB limit on incoming requests in `src/bin/daemon.rs`.
- Enforced the 128KB limit on incoming responses in `src/utils/daemon.rs`.
- Preserved detailed `eprintln!` logging when messages are rejected.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* refactor: removed garbage
* change version to 1.7.1
* cargo fmt
* cargo update
* docs: add information about hotkeys to README
* docs: small refactor
Replaced an unsafe `.unwrap()` with `.unwrap_or_default()` in `src/gui/draw.rs`
when parsing file names. This prevents potential panics on invalid paths.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* cargo fmt
* deps: bump clap to 4.6.0
* deps: cargo update
* Fix daemon autostart issue caused by sync pipewire retry loop (#43) (#44)
At boot time, PipeWire takes some time to register the `pwsp-daemon` and `pwsp-virtual-mic` devices. Previously, the daemon's retry loop for `link_player_to_virtual_mic()` was synchronous and limited to 5 attempts (1.5 seconds total). This caused systemd autostarts to fail with a code 1 if the devices were not yet available.
This change replaces the synchronous wait with an asynchronous `tokio::spawn` task. It will retry the link attempt up to 60 times with a 1-second delay without blocking the startup of the rest of the daemon. This prevents it from exiting abruptly during autostart.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* deps: bump egui & eframe version to 0.34.1
* feat: replaced App::update with new App::logic and App::ui
* deps: bump egui_material_icons to 0.6.0
* deps: bump egui_dnd to 0.15.0
* fix: use .codepoint for icons
* refactor
* refactor: replaced deprecated CentralPanel::show with CentralPanel::show_inside
* refactor
* change version to 1.6.3
* update rust toolchain in github actions
* update freedesktop platform version to 25.08 for flaptak
* update github actions for flatpak builds
* add flatpak-builder installation inside actions
* add flathub configuration to actions
* add --user flag to flathub configuration
* remove sudo from flatpak actions
* Fix/dev flatpak actions 6082245116761610541 (#47)
* remove sudo
* remove steps
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
At boot time, PipeWire takes some time to register the `pwsp-daemon` and `pwsp-virtual-mic` devices. Previously, the daemon's retry loop for `link_player_to_virtual_mic()` was synchronous and limited to 5 attempts (1.5 seconds total). This caused systemd autostarts to fail with a code 1 if the devices were not yet available.
This change replaces the synchronous wait with an asynchronous `tokio::spawn` task. It will retry the link attempt up to 60 times with a 1-second delay without blocking the startup of the rest of the daemon. This prevents it from exiting abruptly during autostart.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
This commit replaces two instances of `contains_key` followed by
`get_mut().unwrap()` with the more idiomatic `if let Some(...)` pattern
in `src/utils/pipewire.rs`. This reduces redundant hash map lookups and
eliminates potential panics from `unwrap()`.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Refactored the navigation logic in `src/gui/input.rs` to use idiomatic Rust patterns.
Replaced clunky manual index calculations and type casting with `match` expressions and modular arithmetic on `usize`.
This improvement enhances readability and maintainability by eliminating nested `if/else` blocks and potential overflow issues from integer casts.
🎯 **What:** The code health issue addressed
- Refactored directory and file navigation logic to use `usize` and modular arithmetic.
- Replaced manual wrap-around logic with idiomatic `match` expressions.
💡 **Why:** How this improves maintainability
- Eliminates unnecessary and potentially risky type casting (e.g., `i8`, `i64`).
- Reduces code nesting and complexity, making it easier to read and extend.
- Standardizes the circular navigation pattern across the GUI.
✅ **Verification:** How you confirmed the change is safe
- Manually reviewed and verified the logic for all key combinations (ArrowUp, ArrowDown, both, or none).
- Confirmed correct behavior for both initial selection (None) and existing selection (Some) states.
✨ **Result:** The improvement achieved
- Cleaner, more idiomatic Rust code for list navigation.
- Reduced potential for index-related bugs.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaces manual vector extension and linear search with an iterator
chain. This avoids an unnecessary allocation and potential reallocation
of the `input_devices` vector and allows for short-circuiting if the
device is found early.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Directly attempt to remove the daemon socket file and handle NotFound errors
instead of checking for its existence first. This prevents a potential
race condition where the file could be replaced between the check and
the removal.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* 🧹 Refactor: Replace unsafe unwrap in get_audio_player
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* 🧹 Refactor: Replace unsafe unwrap in get_audio_player
Resolved GitHub CI failure where a syntax error was introduced due to a bad automated merge with main. Rebased cleanly to ensure only the get_audio_player code health changes are included.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* Delete tests/perf_play.rs
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
* perf(audio_player): offload synchronous I/O and decoder init to spawn_blocking
Moved synchronous file system operations (`fs::File::open` and `file_path.exists()`) and CPU-bound decoder initialization (`Decoder::try_from`) inside the async `AudioPlayer::play` method to `tokio::task::spawn_blocking`.
This prevents starving the Tokio worker threads during disk operations and significantly reduces event loop latency.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
* perf(audio_player): offload synchronous I/O and decoder init to spawn_blocking
Moved synchronous file system operations (`fs::File::open` and `file_path.exists()`) and CPU-bound decoder initialization (`Decoder::try_from`) inside the async `AudioPlayer::play` method to `tokio::task::spawn_blocking`.
This prevents starving the Tokio worker threads during disk operations and significantly reduces event loop latency.
Co-authored-by: arabianq <55220741+arabianq@users.noreply.github.com>
---------
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaced sequential unwraps on pipewire properties ("node.id", "port.id", "port.name")
with an `if let` and `and_then()` pattern in `src/utils/pipewire.rs`. This provides
safety against daemon crashes when properties are missing or malformed by silently
returning instead of panicking.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Offload synchronous `fs::File::open` and `Decoder::try_from` operations to `tokio::task::spawn_blocking` in `AudioPlayer::update`. This allows the Tokio runtime to process other asynchronous tasks concurrently without being blocked by file I/O operations and audio header decoding during looped playback.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
The daemon was allocating memory based on an unverified length prefix
sent over the unauthenticated Unix socket, potentially allowing a malicious
client to cause an Out-Of-Memory panic (DoS). A 10 MB size limit has been
introduced.
Note: The previously reported `unwrap()` panic on invalid JSON payloads
was already fixed and replaced with a safe `match` block in a prior commit.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Moves the check `if let Some(current_input_name) = &audio_player.input_device_name`
outside the loop over `input_devices` in `GetFullStateCommand::execute`.
By creating two separate loops (one for when an input device name is selected,
and one for when it is not), we eliminate the overhead of evaluating the `Option`
on every single iteration of the `input_devices` array. This effectively unswitches
the loop and avoids repeatedly accessing the `audio_player` struct field.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
In `draw_footer`, the `all_inputs` HashMap was being collected into a Vec
and sorted on every single UI frame (~60 FPS). This caused unnecessary
overhead and allocations.
This commit introduces a cached `all_inputs_sorted` vector in the
`AudioPlayerState` which is populated only when the inputs map changes
during the state sync. `draw_footer` now loops over this pre-sorted
vector directly.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
When the user's daemon.json or gui.json configuration files become corrupted or invalid (e.g. invalid JSON), the application panics as the load_from_file function previously bubbled up the error causing a panic. This fix modifies load_from_file for both DaemonConfig and GuiConfig to catch JSON parsing errors using a match statement and return a Default configuration instead.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Removed the unsafe `.unwrap()` call when attempting to get the parent directory of `config_path` in `DaemonConfig::save_to_file` and `GuiConfig::save_to_file`. Replaced it with an idiomatic `if let Some(config_dir)` check to improve code safety and maintainability.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Replaces `audio_player_state_shared.lock().unwrap()` with `.unwrap_or_else(|e| e.into_inner())` in `src/utils/gui.rs` to allow safe recovery from poisoned locks and avoid application panics.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
This commit addresses a code health issue in `src/gui/input.rs` where an `.is_some()` check was followed by an unsafe `.unwrap()` on `self.app_state.selected_file`.
The logic has been updated to use the idiomatic `if let Some(path) = self.app_state.selected_file.clone()` pattern. The `.clone()` is necessary because the subsequent methods (`self.play_file` and `self.stop`) require a mutable borrow (`&mut self`), which would conflict with an immutable borrow of `self.app_state.selected_file`. This change ensures the code is safe and panic-free while satisfying Rust's borrow checker rules. Behavior remains unchanged.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
- Replaced `.unwrap()` with proper error handling during JSON serialization in `GetStateCommand`, `GetTracksCommand`, and `GetFullStateCommand`.
- Added error handling for malformed client requests in the daemon's main loop.
- Ensured the daemon stays running even if serialization or deserialization fails.
- Handled potential errors from `get_all_devices()`.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Moved the access of `audio_player.input_device_name` outside the loop
in `GetFullStateCommand::execute` to avoid repeated field access and
Option checking during iteration.
Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
