This is a fork. Upstream GPUI is developed by Zed Industries at zed-industries/zed; this repository adds pinch/magnify gestures and stylus pressure on top of the published 0.2.2 release. See UPSTREAM.md for provenance and Pinch gestures for the addition.

Welcome to GPUI!

GPUI is a hybrid immediate and retained mode, GPU accelerated, UI framework for Rust, designed to support a wide variety of applications.

Getting Started

GPUI is still in active development as we work on the Zed code editor, and is still pre-1.0. There will often be breaking changes between versions. You'll also need to use the latest version of stable Rust and be on macOS or Linux. Add the following to your Cargo.toml:

gpui = { version = "*" }

• Ownership and data flow

Everything in GPUI starts with an Application. You can create one with Application::new(), and kick off your application by passing a callback to Application::run(). Inside this callback, you can create a new window with App::open_window(), and register your first root view. See gpui.rs for a complete example.

Dependencies

GPUI has various system dependencies that it needs in order to work.

macOS

On macOS, GPUI uses Metal for rendering. In order to use Metal, you need to do the following:

• Install Xcode from the macOS App Store, or from the Apple Developer website. Note this requires a developer account.

Ensure you launch Xcode after installing, and install the macOS components, which is the default option.

• Install Xcode command line tools

  xcode-select --install

• Ensure that the Xcode command line tools are using your newly installed copy of Xcode:

  sudo xcode-select --switch /Applications/Xcode.app/Contents/Developer

The Big Picture

GPUI offers three different registers depending on your needs:

• State management and communication with Entity's. Whenever you need to store application state that communicates between different parts of your application, you'll want to use GPUI's entities. Entities are owned by GPUI and are only accessible through an owned smart pointer similar to an Rc. See the app::context module for more information.

• High level, declarative UI with views. All UI in GPUI starts with a view. A view is simply an Entity that can be rendered, by implementing the Render trait. At the start of each frame, GPUI will call this render method on the root view of a given window. Views build a tree of elements, lay them out and style them with a tailwind-style API, and then give them to GPUI to turn into pixels. See the div element for an all purpose swiss-army knife of rendering.

• Low level, imperative UI with Elements. Elements are the building blocks of UI in GPUI, and they provide a nice wrapper around an imperative API that provides as much flexibility and control as you need. Elements have total control over how they and their child elements are rendered and can be used for making efficient views into large lists, implement custom layouting for a code editor, and anything else you can think of. See the element module for more information.

Each of these registers has one or more corresponding contexts that can be accessed from all GPUI services. This context is your main interface to GPUI, and is used extensively throughout the framework.

Other Resources

In addition to the systems above, GPUI provides a range of smaller services that are useful for building complex applications:

• Actions are user-defined structs that are used for converting keystrokes into logical operations in your UI. Use this for implementing keyboard shortcuts, such as cmd-q. See the action module for more information.

• Platform services, such as quit the app or open a URL are available as methods on the app::App.

• An async executor that is integrated with the platform's event loop. See the executor module for more information.,

• The [gpui::test] macro provides a convenient way to write tests for your GPUI applications. Tests also have their own kind of context, a TestAppContext which provides ways of simulating common platform input. See app::test_context and test modules for more details.

Currently, the best way to learn about these APIs is to read the Zed source code, ask us about it at a fireside hack, or drop a question in the Zed Discord. We're working on improving the documentation, creating more examples, and will be publishing more guides to GPUI on our blog.

Running in the browser

This fork compiles to wasm32-unknown-unknown and runs gpui applications in the browser: windows are canvases rendered through WebGPU, text is shaped by the same cosmic-text stack as Linux (bring your own fonts -- browsers expose none), and DOM events become gpui input, including pinch gestures, IME composition and clipboard paste. Applications keep the normal gpui programming model; the platform differences that leak through (a non-blocking Application::run, no BackgroundExecutor::block) are small and documented. See docs/web.md for the status table, build steps, and three browser examples (hello_web, input_web, multi_window_web).

Pinch gestures

This fork adds a PinchEvent alongside ScrollWheelEvent, so trackpad pinch-to-zoom can be handled directly instead of being approximated with modifier+scroll:

div().on_pinch(|event: &PinchEvent, _window, _cx| {
    // `delta` is the multiplicative change since the previous event of this
    // gesture, so this is all a zoomable view needs:
    zoom *= event.delta;

    // `scale` is the cumulative change since the gesture began (1.0 at
    // `TouchPhase::Started`), for views that would rather snapshot their
    // starting zoom and multiply once.
    // `position` is the gesture centroid, for zooming about the fingers.
});

Platforms report magnification in different terms — macOS sends a per-event increment, Wayland and X11 an absolute scale relative to the start of the gesture, Windows an absolute finger distance — so all of them are normalised into delta/scale before dispatch. Gestures arrive as a Started event, zero or more Moved events, and an Ended event; delta is 1.0 for the first and last of those. Like scroll events, pinches are routed to the element under the centroid.

Platform support

Platform	Status	Mechanism
macOS	Supported	magnifyWithEvent: / NSEventTypeMagnify
Linux/Wayland	Supported	zwp_pointer_gestures_v1 pinch, when the compositor advertises it
Linux/X11	Supported	XI 2.4 gesture events (xorg-server 21.1+, libinput); older servers deliver nothing
Windows	Touchscreen; touchpad opt-in	WM_GESTURE / GID_ZOOM; precision touchpads need Direct Manipulation, see below

Pre-21.1 X11 servers cannot deliver a pinch at all, and on Windows a precision touchpad delivers one only with the opt-in below, so applications should keep a modifier+scroll zoom path as a fallback rather than relying on on_pinch alone. (On Windows a Ctrl+scroll fallback is precisely the form touchpad pinches otherwise arrive in.)

Precision touchpads on Windows

WM_GESTURE covers touchscreens only. A precision touchpad's contacts go to the pointer input stack instead, and a window that does nothing with them gets the legacy fallback: WM_MOUSEWHEEL for a two-finger pan, Ctrl+ WM_MOUSEWHEEL for a pinch. Direct Manipulation is the only interface that hands over the real gesture.

Claiming it is all-or-nothing. DM_POINTERHITTEST arrives before the gesture has been classified, so a window that takes the contact to get pinches takes the pans with it and stops receiving WM_MOUSEWHEEL for them. This fork therefore replaces both: the manipulation's scale becomes a PinchEvent and its translation a pixel-precise ScrollWheelEvent, with inertia on each.

Because that puts new code in the path of ordinary scrolling, it is off by default. Set GPUI_ENABLE_DIRECT_MANIPULATION to 1 or true to turn it on; anything else, or a failure to set Direct Manipulation up, leaves the window on the Ctrl+scroll fallback. It has not been exercised on real hardware -- see UPSTREAM.md.

Stylus pressure

MouseDownEvent, MouseUpEvent and MouseMoveEvent carry a pressure field, 0.0..=1.0. It is 1.0 for an ordinary mouse and on platforms whose tablet input is not wired up, so a caller can multiply by it unconditionally -- a brush that scaled its opacity by pressure would otherwise paint nothing at all on a mouse.

Platform	Status	Mechanism
macOS	Supported	NSEvent.pressure
Linux/Wayland	Supported	zwp_tablet_v2, when the compositor advertises it
Linux/X11	Supported	XInput2 "Abs Pressure" valuator
Windows	Supported	WM_POINTER pen pressure, carried onto the synthesised mouse messages

Two platforms cannot tell "no tablet" from "zero pressure" and report full pressure for both: AppKit reports 0 for an ordinary mouse click, which is indistinguishable from a stylus barely touching the tablet, so anything at or below zero is reported as full pressure, and Windows does the same for a zero reading, which is what a hovering pen reports. X11 identifies tablets by their pressure valuator and Wayland by the tool's advertised pressure capability, so on both a hovering stylus reports its true (zero) pressure and only devices without the axis report 1.0.

Wayland delivers tablet input on its own protocol rather than through wl_pointer, and a compositor stops emulating pointer events for a tool as soon as a client binds that protocol, so this fork synthesises the mouse event stream from it: the tip is the left button, the two barrel buttons are the right and middle buttons (matching what the X11 wacom driver and Windows pen input report), and a tool that leaves proximity produces a mouse exit. Tablet pads -- the ring of express keys on the tablet body -- are not mapped to anything. A tool gets its own cursor, so set_cursor_style reaches it only when the compositor supports cursor-shape-v1; otherwise the compositor's default cursor stays under the stylus.

On Windows the pen system gestures (press-and-hold for right-click, tap feedback, flicks) are disabled on GPUI windows, since they delay or swallow the pen events a drawing surface needs to receive immediately.