0 Fork 0

Files

T

ddidderr ad76ae2301 feat(core): port App/Surface and wire up the runnable ghostty binary

This is the integration step that turns the ported crates into a working
terminal: `cargo run` now opens a GTK window running the user's shell.

ghostty-core ports App.zig, Surface.zig (split into keyboard, mouse,
binding, clipboard and config submodules), surface_mouse.zig and
global.zig, and implements the runtime traits. perform_binding_action
handles every binding action.

Cross-crate integration:
- The core<->runtime traits (CoreApp, CoreSurface, RtApp, RtSurface,
  Waker) move from apprt-gtk into ghostty_apprt::runtime_api so the core
  stays runtime-agnostic; Scrollbar/RendererHealth stand-ins are replaced
  by the real terminal/renderer types.
- termio now uses ghostty_renderer::state::{State, SharedState}; the
  renderer calls kitty_update; image-storage-limit is applied at init
  and on config change.
- The search thread locks the terminal through a TerminalLock trait so
  it can share the render state mutex; termio's Wakeup gained a hook to
  wake the renderer thread.
- apprt-gtk: GL proc addresses are resolved through libepoxy's
  epoxy_glFoo pointer variables (plain glFoo symbols resolved to null),
  default DPI fixed to 96, libadwaita version read without GTK init so
  --version works headless.

The root binary dispatches +actions to ghostty-cli (list-fonts,
show-face and new-window are now wired) and otherwise runs the GTK app
with the real core.

Verified at runtime on Wayland and X11: shell prompt and colored output
render, child exit closes the window cleanly. Not yet verified: keyboard
and mouse input, clipboard, IME, tabs and splits (no key injection was
possible in this environment). GHOSTTY_DEBUG_SCREENSHOT=<png> dumps the
renderer's last frame for headless checks.

Test Plan:
- cargo test --workspace (2890 passed)
- cargo clippy --workspace --all-targets -- -D warnings
- GSETTINGS_BACKEND=memory GHOSTTY_RESOURCES_DIR=/usr/share/ghostty \
  cargo run

2026-10-12 01:33:44 +02:00

5.7 KiB

Raw Blame History

ghostty-core notes

Keyboard (fork K) - crates/ghostty-core/src/surface/keyboard.rs

Ported: Keyboard state, keyEventIsBinding, keyCallback, maybeHandleBinding (sequences/leader, key tables incl. once, catch_all ignore, chained actions, performable, consumed/global/all, last_trigger release suppression), endKeySequence, encodeKey (+ encode_options_from_terminal = key_encode.Options.fromTerminal), key-table helpers (activate_key_table, deactivate_key_table, deactivate_all_key_tables), release_pressed_key_on_unfocus, SurfaceMouse.keyToMouseShape (private copy, because it is only used here at the moment). Deviations: sequences/tables hold paths/names resolved against the current config (no raw pointers); global/all chained actions are re-implemented locally (perform_all_chained_action) as the surface has no App handle; no inspector; macOS option-as-alt skipped; once table deactivation calls deactivate_key_table directly. Needs from the mouse module: Mouse.{mods: Mods, over_link: bool, hidden: bool}, mods_changed(Mods), mouse_shift_capture(bool) -> bool, mouse_refresh_links(CursorPos, Coordinate, bool) -> Result<(), CoreError>. Tests: 8 (Zig has none for this code): OwnedKeyEvent round trip, closing actions, sequence path resolution, table lookup, leader path, encode options from terminal modes, mouse shape for keys.

Binding actions (fork B) - surface/binding.rs

Ported: Search (+ search-thread event callback), Text/dump_text/dump_text_locked (viewport math in free fn text_viewport), perform_binding_action for every action, closing_action, writeScreenFile (free fns screen_file_contents / write_screen_temp_file; the temp dir is deliberately left on disk like Zig), adjust_selection (free fn on &mut Terminal).
App-scoped actions are mapped locally (perform_app_action) because App::perform_action needs an App, not the shared AppState; new_window from a surface targets the surface.
toggle_mouse_reporting rebuilds the DerivedConfig with a cloned Config (Zig mutated in place).
Inspector action only forwards RtAction::Inspector (inspector is a no-op).
Cross-crate change: ghostty_terminal::search::thread::Options.terminal is now Arc<dyn TerminalLock> (trait implemented for Mutex<Terminal>; core implements it for the shared render state) because the search thread must lock the terminal inside ghostty_renderer::state::State.
Requires from the mouse module: link_url_at_cursor(&self) -> Option<String> (returns the link text under the cursor incl. clipboard trimming and OSC8 URI; None when not over a link); also mods_changed, mouse_shift_capture, mouse_refresh_links used by keyboard.rs.
Tests: 7 (no Zig tests in this range).

Mouse (fork M)

Files: crates/ghostty-core/src/surface/mouse.rs (Mouse state, mouse reporting x10/utf8/sgr/urxvt/sgr-pixels, button/cursor/scroll/pressure callbacks, click-to-move, multi-click select, right-click actions, middle-click paste, links incl. OSC8, selection scrolling) and src/surface_mouse.rs.

Tests: key_to_mouse_shape, surface_selection_logic, surface_rectangle_selection_logic ported 1:1 (selection math is the pure mouse_selection(&PageList, ..)); 3 extra.
Deviations: mouse reports/IO messages/runtime actions are collected and run after the renderer lock is released (Zig held the lock and used .locked sends). process_links takes the lock itself (Zig called it unlocked). Left-release copy-on-select copies via copy_on_select directly: Zig's setSelection(prev-equal) early-returns on eql, which would never copy.
Extra API for binding fork: link_url_at_cursor().

Core overview (integration, agent CORE)

Done: trait boundary moved to ghostty_apprt::runtime_api (apprt now depends on terminal+renderer and re-exports the real Scrollbar/RendererHealth); termio uses ghostty_renderer::state::{State,SharedState} (SharedStateExt::lock_terminal); renderer update_kitty calls image::State::kitty_update; termio applies image-storage-limit (init + config change) and deletes kitty images on clear-screen; app.rs (App.zig, CoreApp impl, app mailbox), global.rs (logging in Zig format, GHOSTTY_LOG, rlimits, resources dir, i18n), surface/mod.rs (init/teardown with renderer+IO threads, message handler, config updates, sizing, fonts, focus, preedit, selection, CoreSurface impl), surface/{clipboard,keyboard,mouse,binding,config}.rs, surface_mouse.rs; root src/main.rs. Wiring details: renderer->app via Host (scrollbar/health/redraw as app mailbox messages); termio->renderer via RendererLink + Wakeup::with_hook; apprt-gtk got RtSurface::gl_proc_address (libepoxy epoxy_* pointer variables) and create_scope_fn (cgroup scope over D-Bus from the IO thread). Deviations/stubs: inspector no-op; locale only via GTK's setlocale (no unsafe in core/root); crash reporting (sentry) not ported; Rust version row of +version empty; default DPI 96 (Zig linux). How to run: GSETTINGS_BACKEND=memory GHOSTTY_RESOURCES_DIR=/usr/share/ghostty cargo run -- -e sh -c '...' (debug builds only find resources via the env var here). Dev screenshot: GHOSTTY_DEBUG_SCREENSHOT=/abs/out.png [GHOSTTY_DEBUG_SCREENSHOT_MS=2500] dumps the renderer's last frame (compositor screenshots under niri come out black / are blocked). Verified manually: window opens on Wayland and X11, shell prompt and SGR colored output render (target/screenshots/frame.png, shell.png), child exit closes window, clean shutdown. NOT verified: keyboard input end-to-end (no way to inject keys into the window here; xdotool/niri did not deliver), mouse, clipboard, IME, splits/tabs. Known warnings at runtime: unimplemented action=CellSize (same as Zig GTK), cgroup scope creation fails with UnixProcessIdUnknown in this sandbox (non-fatal).

5.7 KiB Raw Blame History