* fix web break introduced in 38f130071 fix(linux): enable mouse side buttons in remote sessions (#14848)
* feat(shortcuts): user-configurable keyboard shortcuts for session actions
Adds a keyboard shortcut feature (Rust matcher + Dart UI + cross-language
parity tests) that lets users bind combinations like Ctrl+Alt+Shift+P to
session actions. Bindings are stored in LocalConfig under
`keyboard-shortcuts`; the matcher gates dispatch on `enabled` and
`pass_through` flags so flipping the master switch off is a hard stop.
Wire-up summary:
- src/keyboard/shortcuts.rs: matcher, default bindings, parity test against
flutter/test/fixtures/default_keyboard_shortcuts.json
- src/keyboard.rs: shortcut intercept in process_event{,_with_session},
feature-gated to `flutter`; runs before key swapping so users bind to
physical keys
- src/flutter_ffi.rs: main_reload_keyboard_shortcuts +
main_get_default_keyboard_shortcuts; reload_from_config seeded in main_init
- flutter/lib/common/widgets/keyboard_shortcuts/: shared config page body,
recording dialog, shortcut display formatter, action group registry
- flutter/lib/desktop/pages/desktop_keyboard_shortcuts_page.dart and
flutter/lib/mobile/pages/mobile_keyboard_shortcuts_page.dart: platform
shells around the shared body
- flutter/lib/models/shortcut_model.dart: per-session ShortcutModel +
registerSessionShortcutActions for actions with no toolbar TToggleMenu /
TRadioMenu (fullscreen, switch display/tab, close tab, voice call, etc.)
- flutter/lib/common/widgets/toolbar.dart: optional `actionId` field on
TToggleMenu / TRadioMenu, plus per-helper auto-register pass that wires
tagged entries' existing onChanged into the ShortcutModel
- flutter/test/keyboard_shortcuts_test.dart + fixtures: cross-language
parity (default bindings, supported key vocabulary)
Design principles applied during review:
1. Additions are fine; modifications to original logic must be deliberate.
Tagging an existing TToggleMenu entry with `actionId:` is an addition.
Rewriting its onChanged to satisfy a new contract is a modification —
and was reverted for every case where the original click behavior was
working. Four closures were touched and then reverted (mobile View
Mode, Privacy mode multi-impl, Relative mouse mode, Reverse mouse
wheel); their shortcuts are wired via standalone closures in
shortcut_model.dart instead.
2. Toolbar auto-register is reserved for entries whose onChanged is
inherently self-flipping — typically `sessionToggleOption(name)` where
the named option is flipped in place and the input bool is unused. The
register pass passes `!menu.value` from registration time, which is
harmless under self-flipping but wrong for closures that consume the
input bool directly. Tagging a non-self-flipping entry forces a closure
rewrite; choose non-toolbar registration in that case.
3. When shortcuts are disabled, toolbar behavior must be bit-for-bit
unchanged. The matcher's `enabled`-gate already guarantees no
dispatch; the auto-register pass is left unconditional (its only effect
is HashMap operations on a separate ShortcutModel) so mid-session
enable works without a reconnect. The trade-off is intentional and
documented at the top of toolbarControls.
4. Comments stay terse. Rationale lives in one place — the doc comment of
the helper or registration site, not duplicated at every call site.
5. Where an existing helper needs a new optional behavior (e.g.
`_OptionCheckBox` gaining a tooltip slot), the new branch must reduce
to byte-identical output for existing callers (`trailing == null`
case → original `Expanded(Text)` layout). Verified.
6. Action IDs and labels stay consistent. Renamed `reset_cursor` →
`reset_canvas` so the action ID matches its user-facing label
("Reset canvas") and capability flag.
Out-of-scope but included:
- AGENTS.md: documents flutter_rust_bridge no-codegen workflow and the
Web target's hand-written TS client, since both are load-bearing for
any new FFI work.
- remote_toolbar.dart: i18n fix for the per-monitor tooltip ("All
monitors" / "Monitor #N"), unrelated to shortcuts but kept here.
* langs
* fix(keyboard): shortcuts, harden config and callback lifecycle
Signed-off-by: fufesou <linlong1266@gmail.com>
* feat(keyboard): shortcuts, color of "Reset to defaults"
Signed-off-by: fufesou <linlong1266@gmail.com>
* feat(keyboard): shortcuts, release keys before shortcut callback
Signed-off-by: fufesou <linlong1266@gmail.com>
* feat(keyboard): shortcuts, debug web
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(keyboard): iPad, icon
Signed-off-by: fufesou <linlong1266@gmail.com>
* langs: add shortcut keys to az, gl, ur
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, keep local modifiers and release keys on the right session
When a chord fired, the shortcut path reused release_remote_keys(), which
replays synthetic releases through process_event(). That cleared the local
MODIFIERS_STATE while the user still held the modifiers, so the next key of
a held chord no longer matched and leaked to the remote, and it sent the
releases to the globally current session instead of the one that typed the
chord.
The shortcut path now releases the remote keys itself: it sends through the
caller's session, and restores the modifiers it released on the remote as
still held locally. The press that fired is remembered until its release,
so its auto repeats and its release are consumed instead of reaching the
remote.
Also state the contract for unavailable actions: a bound chord always
belongs to RustDesk and is a no-op when the session has no handler.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, record the physical key
The recording dialog and the Web Flutter matcher named keys by
LogicalKeyboardKey, while the native matcher sees the physical key through
its USB HID usage and the Web JS matcher through KeyboardEvent.code. On
AZERTY, QWERTZ, Dvorak and similar layouts a recorded binding then did not
fire, or fired from another key.
Name keys by their physical position everywhere: physicalKeyName maps the
USB HID usage to the stored name. A new fixture pins every name to its USB
HID usage, checked on the Dart side and against rdev::usb_hid_key_from_code
on the Rust side, and a test covers AZERTY and QWERTZ events.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, keep an empty binding list on re-enable
setEnabled seeded the defaults whenever the binding list was empty, so a
user who cleared every binding got all defaults back after disable and
enable. Seed only when the config has never held a bindings list.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, track every fired key until its release
FIRED_KEY held a single key, so overlapping shortcut keys broke it: with the
chord held, P then C overwrote P, P's release reached the remote, and a P
repeat could fire the screenshot again. A repeat after the modifiers were
released was also treated as a new press and leaked to the remote.
Keep a set of fired keys instead. Once a press fires, every event of that
key belongs to RustDesk until its release: repeats are consumed without
dispatching, whatever the modifiers do, and the release removes the key.
Entries whose release never arrives are dropped when a session view is
entered or a session is closed, not by guessing from the modifiers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, keep fired keys per session
FIRED_KEYS was process-global while dispatch is per session. Closing or
entering any session cleared the fired keys of all of them, so a key still
held in another session leaked its release to that remote, and a key fired
in one session made the same key in another session look like a repeat.
Key the fired set by session: try_dispatch resolves the session first and
only looks at that session's keys, and session_close clears only the
closing session. Entering a session view no longer clears anything: it is
a pointer enter, and a fired key may still be physically held across it.
A key whose release is missed otherwise heals itself: its next press is
consumed as a repeat and that release removes it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(shortcuts): preserve input state and refresh action availability
* fix(keyboard): shortcuts, keep a fired key owned across session close
Close tab is itself a shortcut, so the session that consumed the key press
is gone before the key is released. Both matchers dropped the key with the
session: the Flutter dispatcher removed its own entries in clear(), and the
Rust set was keyed by session and cleared in session_close. The release,
and the repeats while the key stayed held, then reached the next session's
remote as ordinary input.
Ownership of a fired key is physical, so it is now one set shared by every
session on each side, and closing a session no longer touches it: clear()
only drops the pending action and modifier replay, and the Rust session
state only tracks released modifiers. A key whose release is missed heals
itself on its next press. Entering view-only with the Flutter input source
still clears the Rust set, since that source stops routing key events to it.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, drop the Toggle Input Source action
The action switches the key capture backend between key down and key up.
Fired-key ownership lives in whichever matcher saw the press, the Flutter
dispatcher or the Rust set, and there is no handoff between them: after the
switch the repeats and the release of the held key reach the other backend,
which forwards them to the remote and, while the chord is held, matches the
action again and toggles the source back. Remove the shortcut instead of
building a handoff for one action. The toolbar radio menu is unchanged.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* perf(keyboard): shortcuts, parse the config once per stored value
The Flutter matcher read the local config three times and parsed its JSON
three times on every key press in legacy mode, on Web and in view-only.
Keep one parsed snapshot keyed by the raw stored string, so a key press
costs one config read and no parsing while the config is unchanged. The
existing readers are thin views of the same snapshot.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix(keyboard): shortcuts, run focus-changing actions on key release
Close tab and tab switching move keyboard focus to another session before
the key that fired them is released. On Linux the Flutter input source
routes map-mode sessions through the Rust matcher and legacy-mode sessions
through the Dart matcher, and neither knows the other's fired keys, so
after such an action the repeats and the release of the held key reached
the next session's remote, and the key stayed recorded as fired in the
matcher that saw the press.
Routing Linux legacy mode through the Rust FFI path instead is not a drop-in
change: the raw key handler builds its event with no unicode, so Rust's
legacy mode would fall back to US key names and lose the layout-aware
characters the Dart path sends today, and it would take modifiers from the
X11 key state instead of Flutter's.
Run these three actions when the fired key is released. The matcher that
consumed the press then also sees the repeats and the release, the remote
modifiers are still released on the press, and a session that closes drops
the action it owed. Both matchers keep the same list of actions.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* test(keyboard): shortcuts, pin the key-up action list on both sides
kShortcutActionsRunOnKeyUp and runs_on_release are two hand-kept lists of
the actions that run on key release. Add a shared fixture and check each
side against it, the way the key vocabulary and the default bindings are
already checked, so adding a focus-changing action on one side breaks a
test until the other side follows. The Dart test also checks that every
listed id is a configurable action.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ra4my61t8q1FN5n59wB16D
* fix: prevent lost input after remote session shortcuts
Avoid synthetic Alt presses that activate the Windows system menu
and swallow the next character.
Trigger restart, switch-sides, and chat actions on key release so
focus changes cannot leave subsequent shortcut presses ignored.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(mobile): correct Chat keyboard and input helper behavior
- Keep soft keyboard and helper hidden when Chat opens via shortcut
- Restore soft input on touch and remote focus after Chat closes
- Handle modifier releases while Chat owns focus
- Clear hidden virtual modifiers while preserving held keys and Pin behavior
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): prevent recording non-simultaneous key combinations
Build new captures from currently held keys instead of reusing released
letters. Preserve the last valid chord after release so it can be saved.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): isolate window actions and prevent duplicate Alt presses
- Dispatch shortcut actions only to the originating session window.
- Clear local modifier state on focus loss without resending modifiers
already released to Windows peers, preventing lost following input.
- Strengthen the regression assertion against extra Alt keydown events
Signed-off-by: fufesou <linlong1266@gmail.com>
* refact(shortcuts): empty default shortcuts for Cmd+Option+Shift+Del on macOS and iOS
Because Del may be ambiguous for new macOS users.
Cmd+Option+Shift+Backspace is "Empty Trash without confirmation" shortcut on macOS.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: use session peer platform for Ctrl+Alt+Del events
Fix Android-to-Linux Ctrl+Alt+Del delivery by using the session's peer
platform instead of the desktop global session, which defaults to Windows
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: hide relative mouse mode shortcuts on Wayland
Separate relative mouse mode availability from screenshot support.
Filter unsupported actions and defaults while preserving saved bindings
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: align shortcut availability with platform support
- Hide unsupported mobile and web shortcut actions
- Enable mobile scaling and Android voice call shortcuts
- Use openMonitorInTheSameTab() for display switching
- Preserve cursor position when switching displays on desktop and web
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: prevent held-key shortcut recording bypass
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: preserve keyboard input after shortcut dialogs
Run native clipboard shortcuts on key release and wait for remote modifier releases before dispatching.
Release remote modifiers when mobile local UI takes focus, and document the tested iPad Bluetooth Mute shortcut limitation.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: preserve remote key state during local shortcuts
Keep unrelated held keys pressed for session actions that retain input. Preserve key release when focus or input handling changes, and avoid synthetic Alt taps during shortcut dispatch.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: suspend saved relative mouse shortcuts on Wayland
Filter the native active binding cache without changing stored configuration, so the binding becomes available again under X11.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: exclude unsupported Wayland shortcuts from Flutter matching
Keep saved bindings intact while filtering the active config used by the Flutter matcher. Cover stored bindings and enable/pass-through flags in the existing shortcut tests.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: retain forwarded key ownership across shortcut repeats
Ignore repeat presses for keys already forwarded to the peer, and release ordinary held keys before Show My Cursor enables view-only. Cover both paths with converter-level regressions.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: release Flutter-held keys before view-only shortcuts
Track ordinary keys forwarded by the Dart matcher paths and await their releases before View Only or Show My Cursor disables input. Keep Mute key ownership and existing empty-cleanup timing unchanged.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: filter active Flutter shortcuts by platform capabilities
Reuse the Settings action filter for saved bindings while preserving storage, default-seeding behavior and configurable Apple Ctrl+Alt+Del bindings.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: filter native shortcut caches for mobile controllers
Suspend saved desktop-only actions on Android and iOS and hide unavailable iOS actions without changing persisted bindings. Keep native cache filtering aligned with Flutter platform capabilities.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: share filtered shortcut bindings with the Web matcher
Pass the active platform config to JavaScript at startup and after settings changes, preserving unsupported saved bindings in local storage.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: release Flutter-held keys before focus-changing shortcuts
Drain ordinary keys owned by the original session before run-on-key-up
actions and native clipboard callbacks. Preserve Mute holds, pending-action
cancellation and Web clipboard key-down activation.
Document the required Web matcher counterpart in the bridge:
rustdesk/rustdesk-web-js@94abe431d6
Validation: 61 native Flutter tests, 32 Chrome/HTML tests, targeted analysis
and Windows debug build.
Signed-off-by: fufesou <linlong1266@gmail.com>
* refact: remove the useless comments
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: prevent Windows menu activation from Alt-only shortcuts
Neutralize the forwarded Alt with a balanced temporary Ctrl press when a consumed local shortcut releases its only modifier in Windows Map mode. Keep physical modifier tracking and normal Alt pass-through unchanged.
Add a regression covering both Alt keys, trigger repeats and the eventual physical release. Verified with 36 native shortcut tests and real Windows menu/input checks.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: prevent Alt menu activation in all keyboard modes
Extend Windows Alt-only shortcut neutralization to Translate and Legacy. Keep the temporary Ctrl packets in Map mode so Legacy modifier synchronization cannot release Alt before Ctrl.
Extend the existing regression across keyboard modes and decode scan, virtual and control-key events. Verify 36 native tests and 34 Windows menu replay cases.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: initialize outgoing chat before shortcut dispatch
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: guard Windows Alt release in Flutter legacy shortcuts
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: coordinate Web shortcut dispatch and Alt modifier cleanup
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: distinguish AltGraph from ordinary Alt shortcuts
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: dispatch macOS Web shortcuts without waiting for keyup
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: support legacy peers in privacy mode shortcut
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: exclude AltGraph from Flutter shortcut matching
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: exclude GTK AltGr input from raw shortcut matching
Signed-off-by: fufesou <linlong1266@gmail.com>
* test: trim redundant keyboard shortcut coverage
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(keyboard): prevent Cmd shortcuts from opening remote Start
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): guard Windows menu keys after Ctrl/Cmd swapping
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): align recording and display routing
Reject non-Apple AltGraph captures to match runtime filtering. Honor individual-display windows for next, previous, and all-display shortcuts.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): reject GTK AltGr chords in recorder
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(android): restore keyboard focus for modifier-first input
Focus the existing FlutterView when a modifier is pressed without native view focus. This restores shortcut-first input after returning to RustDesk while preserving focused inputs and normal key dispatch.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): release held keys before keyboard mode changes
Drain Flutter-forwarded keys before keyboard-mode shortcut actions so releases use the original encoding. This prevents held keys from remaining down after a Web Input source 2 mode change.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(i18n): use sentence case for shortcut labels
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(mobile): preserve touch modifiers in the input helper
Keep touch-selected modifiers visible after hiding the keyboard. Clear physical modifier releases locally in view-only mode so they do not reopen the helper.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(web): preserve modifiers across Legacy-to-Map shortcuts
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(shortcuts): check support before registering all displays view
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
Co-authored-by: fufesou <linlong1266@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* fix(linux): respect Wayland keyboard layouts for uinput
Resolve printable keys from GNOME input sources or the Wayland keymap and KDE active layout group. Refresh layout metadata on input at most once per second, and preserve the keycode and owned modifiers until release.
Preserve the original Legacy character on client key release and handle Caps Lock shortcuts without introducing an unwanted Shift modifier.
Validated with GNOME/KDE full-service input and disconnect checks, plus Windows/Linux native keyboard capture and release checks.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): harden Wayland keyboard layout detection
Keep the last usable map on transient discovery failures and preserve the
legacy character path when no map is available, with explicit warnings.
Retain each injected key representation through release and handle IBus
default layout and variant metadata without compiling a layout named default.
Use native GNOME and Plasma sources, including KWin sessions without the
optional layout service. Read the Xwayland keymap and effective group from
the same keyboard on other desktops. Load XCB/XKB libraries dynamically and
accept the unix/:N local display spelling.
Validation: 24/24 temporary Linux production-module and IPC checks passed,
including an actual read-only Xwayland query. The two latest routing/display
regressions failed before their fixes. Rustfmt and git diff --check passed.
No Cargo/Flutter commands or full remote application tests were rerun.
Xwayland state freshness and synchronous desktop I/O remain known limits.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): bound input waits for layout refresh
Run throttled Wayland layout discovery in a single background worker and give the triggering input a 25 ms wait budget. Later input uses the existing map without waiting or queuing more queries. Preserve source error handling and skip known Legacy clipboard-only text.
Document Xwayland freshness and per-character modifier ownership limits. Desktop calls remain uncancelled; the budget bounds input waiting, subject to OS scheduling.
Validation: Linux module compilation with cached dependencies, 23 existing temporary module/IPC checks, and stalled-XCB timing probes passed. Rechecked cache/error/release cases after the final lock-scope change. No permanent regression tests or Cargo/Flutter commands.
Signed-off-by: fufesou <linlong1266@gmail.com>
* docs(linux): record layout query latency measurements
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): reply to failed uinput key state queries
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): prefer non-keypad character mappings
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): preserve text after unmappable uinput characters
Continue processing a text sequence when one character cannot be resolved
in the host layout. Return the first mapping error after the supported
characters are injected, so the existing throttled log still reports failure.
Keep modifier-query, injection and release errors fail-fast. This fixes
German a^bc losing bc without adding clipboard or fixed-US fallback.
The unsupported caret remains an explicit mapping error.
Validation: three temporary Linux checks (nine inputs) passed using the
production resolver and sequence methods, real framed IPC and native XKB
decoding. The suffix check failed before the fix. Formatting and diff checks
passed; no permanent tests, Cargo/Flutter commands or full app build added.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): preserve Latin shortcuts on non-Latin layouts
Detect Ctrl/Alt/Meta independently of CapsLock. If the active XKB map lacks
an ASCII shortcut letter, reuse the existing physical letter mapping while
preserving explicit Shift and the press-time key for release. Existing
layout mappings and unsupported ordinary text retain their behavior.
Honor disable-uinput-layout-fallback and log this compatibility path with
the existing throttle.
Validation: 12 temporary Linux production-module/IPC checks passed,
including 43 non-Latin shortcut/layout/lock/Shift cases. The new checks
failed before the fix. Formatting and git diff --check passed. No real
mobile/compositor session or Cargo/Flutter build was run.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): preserve missing punctuation shortcuts
Allow missing ASCII graphic shortcuts to reuse the existing physical key
table when Ctrl, Alt or Meta is active. This restores Ctrl+[ and Ctrl+]
with a valid Russian layout instead of discarding their character events.
Retain the table's Shift requirement for punctuation and use existing
modifier ownership and press-time release tracking. Uppercase letters do
not synthesize Shift. Active-layout mappings, ordinary text errors and
disable-uinput-layout-fallback retain their behavior.
Validation: 14 temporary Linux production-module/IPC checks passed after
recompiling the final source with rustc and cached dependencies. Includes
19 punctuation/lock/Shift combinations and 43 existing Latin-shortcut
cases. The two new checks failed before the fix. Formatting and
git diff --check passed. No Cargo/Flutter build or real mobile/native
Wayland session was run.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): respect held Shift when resolving shortcuts
Derive Shift-held candidates from the same XKB keymap and query the
caller's Shift state for character shortcuts. On US layouts, Ctrl+Shift+<
now uses the comma key instead of the extra ISO key that produces >.
Keep caller-held Shift out of synthesized modifier ownership. Preserve
ordinary input and letter-shortcut identity, and do not replace an
ordinary shortcut with a keypad alias to satisfy the held-Shift lookup.
Validation: 16 temporary Linux production-module/IPC checks passed with
native XKB event replay and cached dependencies via rustc. Includes
30 US text/shortcut cases and 40 UK/DE/FR character/lock cases. The new
Ctrl+Shift+< check failed before the fix. Changed sections pass formatting
and git diff --check. No Cargo/Flutter build or real mobile/native Wayland
session was run; device-state and packet-modifier handling use fixtures.
Signed-off-by: fufesou <linlong1266@gmail.com>
* docs(android): document paired-punctuation input limitation
Signed-off-by: fufesou <linlong1266@gmail.com>
* docs(linux): explain Xwayland layout synchronization limits
Document the focus-scoped modifier updates observed on GNOME and KWin. Core and physical X11 keyboards can retain an old layout group while native Wayland windows have focus; reconnecting or waiting cannot request the missing update.
Validation: 16 existing module/IPC probes and the live GNOME mapping probe passed. Live KDE queries reproduced native French versus stale Xwayland US. Formatting and diff checks passed. This is a comment-only change; strict Xwayland-only behavior is preserved.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): query native Sway and Hyprland keyboard layouts
Read the uinput keyboard's active layout through compositor IPC and
validate its group against the native Wayland keymap. Match the IPC
server to the Wayland session, and accept numbered Hyprland device
names while rejecting ambiguous keyboard selections.
Validation: 21 temporary production-module and framed-IPC checks passed.
Live Sway US/FR group and a/q mapping checks passed with Xwayland disabled;
the GNOME French layout probe also passed. Formatting checks passed.
A live Hyprland session and a full application build were not tested.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): defer unverified native compositor layout queries
Remove the Sway/Hyprland adapter that paired a uinput device's layout
group with a seat keymap based only on matching layout names. Different
device options can make a synthesized AltGr key toggle CapsLock instead.
Restore the existing best-effort Xwayland source for other desktops and
document the device-keymap information missing from the examined native
queries. GNOME/KDE discovery and injection/cache behavior are unchanged.
Validation: 16 temporary production-module/IPC checks passed after the
removal, plus read-only GNOME French and Xwayland mapping checks. Rust
formatting and diff checks passed. No full application build or mobile
end-to-end test was run for this follow-up.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): reject unreliable GNOME per-window layout sources
GNOME restores per-window layouts without persisting MRU. Reject MRU
selection when per-window input sources are enabled with multiple sources,
and invalidate the cached map and source identity for this typed error.
Transient discovery failures still retain the last valid mapping.
Reuse the existing logged/disableable compatibility policy. This contains
stale-map acceptance; it does not add effective per-window layout discovery.
Single-source and global-source GNOME configurations retain their path.
Validation: the old code failed the temporary stale-cache regression.
All 3 focused production-module checks pass after the fix, covering cache
invalidation/recovery, compatibility modes, single/default source mapping,
UK punctuation, and last-good retention after an IBus connection failure.
Compiled with rustc and cached Linux dependencies; formatting/diff checks
passed. No permanent tests, Cargo/Flutter build or mobile session added.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): reject inactive IBus layout sources
Read the current GlobalEngine description instead of looking up the
persisted MRU engine by name. GNOME can suppress IBus for password entry
without updating MRU, leaving the old engine's layout valid but inactive.
Report an engine-ID mismatch as UnreliableSource so the existing cache
invalidation path drops that mapping. Keep transient query-error handling
and layout/variant normalization. Do not infer a US compositor layout from
GNOME's xkb:us::eng echo engine.
Validation: the old code failed the inactive-engine cache and default-layout
checks. All 3 temporary production-module/D-Bus checks pass with this fix,
covering invalidation/recovery, compatibility policies, metadata sentinels,
query failures and ordinary French XKB selection. Compiled with rustc and
cached Linux dependencies; formatting/diff checks pass. No permanent tests,
Cargo/Flutter build or live password/mobile session added.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): prewarm uinput layout discovery at startup
Start the existing layout worker when the uinput keyboard client connects,
before the server accepts remote input. Reuse its throttle and single-worker
logic; keep the existing 25 ms input wait and background completion policy.
Prewarming reduces cold-start fallback without introducing input queues or
promising readiness while a desktop query remains outstanding.
Verified with three temporary Linux production-module checks using cached
dependencies: a 125 ms query resolves UK @ before first input, a 200 ms
outstanding query stays nonblocking and publishes its map, and a failed
startup query retries successfully. All three fail their startup assertions
before the change. No permanent tests or Cargo/Flutter builds were added.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): preserve first input wait during layout prewarming
Keep the startup query's completion receiver until the first relevant input
takes it. That input can wait for the already-running query within the
existing 25 ms deadline, without holding the receiver handoff lock. Later
input remains nonblocking and the existing single-worker guard is unchanged.
Prefer a due refresh to an old completed startup notification.
Verified with four temporary Linux production-module checks compiled with
rustc -C opt-level=3 and cached dependencies: joining a pending UK query,
one bounded wait followed by nonblocking input, due refresh after prewarm,
and recovery after query failure. Both lost-wait checks fail before the fix.
Formatting and git diff --check pass. No permanent tests or full build added.
An unoptimized delayed-query probe still exceeded 25 ms and used fallback;
this restores the wait opportunity, not an initial-readiness guarantee.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(uinput): remove unnecessary change
Signed-off-by: fufesou <linlong1266@gmail.com>
* refactor(linux): simplify uinput character releases
Release the mapping stored for the same character. Linux character input already pairs down/up events, so opposite-case matching is unnecessary.
Validation: git diff --check passed. Static comparisons confirmed that Map-mode dispatch, raw-key injection and keycode offsets match the PR base. No Cargo/Flutter commands or live-session tests were run.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): normalize Unicode CapsLock shortcuts
Normalize single-scalar Unicode lowercase for CapsLock shortcuts without adding Shift. For lowercase expansions, resolve the original symbol with the actual CapsLock state instead of truncating it. Document the existing first-seat discovery limitation.
Validation: two temporary Rust context checks against lookup fixtures failed before the fix and passed afterward. No Cargo/Flutter build or end-to-end input test was run.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): derive CapsLock shortcut keys from XKB
Keep the actual CapsLock state when resolving shortcut characters. Use the existing Caps-off and Caps+Shift mappings to omit generated Shift only when XKB confirms the same physical key and remaining modifiers, preserving the Turkish I/ı and İ/i key identities.
Validation: three temporary production-module checks passed on 192.168.5.32 with native libxkbcommon; two failed before this fix. Covered Turkish, German, Cyrillic and ASCII shortcuts, explicit Shift, ordinary text and UK punctuation. No full application build, IPC or end-to-end input test was run.
Signed-off-by: fufesou <linlong1266@gmail.com>
* refactor(linux): remove CapsLock shortcut special cases
Use the existing lock-state synchronization and ordinary XKB mappings for mobile soft-keyboard input. Remove CapsLock-specific shortcut remapping and compatibility-path lowercasing while retaining held-Shift punctuation handling.
Validation: rebuilt the Linux Rust library and reused the unchanged Flutter UI. Actual Android Gboard input matched UK punctuation on GNOME and KDE; French/German switching on KDE passed after the refresh interval. The user confirmed real iPad input on both hosts, including smart quotes. Formatting and whitespace checks passed. The optional Ctrl/Shift toolbar automation was inconclusive and is not counted as passing coverage.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): press Level3 before Shift for uinput characters
Preserve mapping candidate preference while storing generated modifiers in Level3-before-Shift order. This avoids activating Compose with lv3:ralt_switch_multikey; releases retain their existing reverse order.
Verified on Linux with temporary production-mapper/libxkbcommon tests: Lithuanian ! fails before this change and passes afterward; all 24 UK punctuation/lock-state cases pass with released modifiers and unchanged lock state. Formatting and diff checks pass. No full application build or mobile end-to-end test was run for this change.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): preserve legacy uinput IPC without a keymap
Send sequences and character clicks through their original IPC messages when the layout cache is unavailable, avoiding unnecessary synchronous key-state queries. Keep throttled fallback warnings and leave layout discovery, retry policy, and key-down/up pairing unchanged.
Verified with three temporary production-method checks on Linux using a recording transport and injected query errors. The unavailable-map routing check fails before the fix and passes afterward; raw-click routing and available-map query-error handling remain unchanged. Changed-code formatting and diff checks pass. No full application build or end-to-end input test was run.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(linux): skip layout queries for legacy uinput fallback
When no keymap is available, single-character input must use the legacy
KeyDown/KeyUp path without querying lock or modifier state first. Retain
the legacy key in the existing press record so discovery completing
before a repeat or release cannot change the selected representation.
Validated with four temporary production-method checks on Linux using a
controlled cache/resolver and recording transport: missing-map fallback
with query failures, press/release across cache transitions, unchanged
raw-key dispatch, and Ctrl+C without extra Shift or queries. The initial
fallback check failed before this change. Valid-map errors still surface.
Changed-code formatting and git diff --check pass. No full Cargo/Flutter
build or live service/mobile test was run for this change.
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
RemotePage.dispose() only reaches sessionClose at the tail of gFFI.close(),
behind several awaits (canvas save, image update, the enable_soft_keyboard
platform call). If the app is backgrounded while the page is disposing,
dispose can be suspended before that runs, so the session is never torn down.
The next reconnect re-attaches to the leaked session (mobile reuses a constant
sessionId) and is stuck on "Connecting..." forever while the orphaned io_loop
keeps streaming.
Dispatch sessionClose at the start of dispose so teardown happens synchronously
on route pop, before backgrounding can interrupt it. The sessionClose in
gFFI.close() becomes a no-op once the session is already removed.
Fixes#15060
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat: Add relative mouse mode
- Add "Relative Mouse Mode" toggle in desktop toolbar and bind to InputModel
- Implement relative mouse movement path: Flutter pointer deltas -> `type: move_relative` -> new `MOUSE_TYPE_MOVE_RELATIVE` in Rust
- In server input service, simulate relative movement via Enigo and keep latest cursor position in sync
- Track pointer-lock center in Flutter (local widget + screen coordinates) and re-center OS cursor after each relative move
- Update pointer-lock center on window move/resize/restore/maximize and when remote display geometry changes
- Hide local cursor when relative mouse mode is active (both Flutter cursor and OS cursor), restore on leave/disable
- On Windows, clip OS cursor to the window rect while in relative mode and release clip when leaving/turning off
- Implement platform helpers: `get_cursor_pos`, `set_cursor_pos`, `show_cursor`, `clip_cursor` (no-op clip/hide on Linux for now)
- Add keyboard shortcut Ctrl+Alt+Shift+M to toggle relative mode (enabled by default, works on all platforms)
- Remove `enable-relative-mouse-shortcut` config option - shortcut is now always available when keyboard permission is granted
- Handle window blur/focus/minimize events to properly release/restore cursor constraints
- Add MOUSE_TYPE_MASK constant and unit tests for mouse event constants
Note: Relative mouse mode state is NOT persisted to config (session-only).
Note: On Linux, show_cursor and clip_cursor are no-ops; cursor hiding is handled by Flutter side.
Signed-off-by: fufesou <linlong1266@gmail.com>
* feat(mouse): relative mouse mode, exit hint
Signed-off-by: fufesou <linlong1266@gmail.com>
* refact(relative mouse): shortcut
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: prevent custom scale dialog from closing when interacting with slider
Wrapped MobileCustomScaleControls in GestureDetector with opaque behavior
to prevent touch events from propagating to parent dialog's clickMaskDismiss
handler. The slider now works correctly without closing the dialog.
Signed-off-by: Alessandro De Blasis <alex@deblasis.net>
* Update flutter/lib/mobile/widgets/custom_scale_widget.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update flutter/lib/mobile/widgets/custom_scale_widget.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update flutter/lib/mobile/widgets/custom_scale_widget.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update flutter/lib/mobile/widgets/custom_scale_widget.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Revert "fix: mobile remove "Scale custom" (#13323)"
This reverts commit 265d08fc3b.
* chore: keep remote_toolbar.dart cleanup (remove dead code)
The dead code removed in 265d08fc3 hasn't been used since Aug 2023.
Only reverting toolbar.dart is needed for the mobile Scale custom fix.
* Update flutter/lib/mobile/pages/remote_page.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* refactor: Implement CustomScaleControlsMixin for shared scaling logic across mobile and desktop widgets
- Introduced a new mixin `CustomScaleControlsMixin` to encapsulate custom scale control logic, allowing for code reuse in both mobile and desktop widgets.
- Refactored `_CustomScaleMenuControlsState` and `_MobileCustomScaleControlsState` to utilize the new mixin, simplifying the scaling logic and reducing code duplication.
- Updated slider handling and state management to leverage the mixin's methods for improved maintainability.
Signed-off-by: Alessandro De Blasis <alex@deblasis.net>
* Update flutter/lib/desktop/widgets/remote_toolbar.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update flutter/lib/mobile/widgets/custom_scale_widget.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update flutter/lib/mobile/pages/remote_page.dart
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* refactor: changed from mixin to abstract class
Signed-off-by: Alessandro De Blasis <alex@deblasis.net>
* Revert "Update flutter/lib/mobile/pages/remote_page.dart"
This reverts commit 7c35897408.
* refactor: remove unnecessary tap event handling in custom scale controls
- Removed the `onTap` handler from the
Signed-off-by: Alessandro De Blasis <alex@deblasis.net>
* refactor: simplify MobileCustomScaleControls usage in remote_page.dart
- Removed unnecessary GestureDetector wrapper around MobileCustomScaleControls for cleaner code.
Signed-off-by: Alessandro De Blasis <alex@deblasis.net>
---------
Signed-off-by: Alessandro De Blasis <alex@deblasis.net>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* fix: Android, keyboard, map mode, workaround
The `KeyEvent.physicalKey.usbHidUsage` are wrong if using Microsoft
SwiftKey keyboard.
`window.addFlags(WindowManager.LayoutParams.FLAG_ALT_FOCUSABLE_IM)` is a workaround for this issue.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: clear callback on first image
Signed-off-by: fufesou <linlong1266@gmail.com>
* Android disable soft keyboard in remote page if not editing.
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
1. Map mode. Check if the KeyEvent's usbHidUsage is correct.
2. Korean input, use listener to handle composing state.
Signed-off-by: fufesou <linlong1266@gmail.com>
* Add option auto record outgoing session
* In the same connection, all displays and all windows share the same
recording state.
todo:
Android check external storage permission
Known issue:
* Sciter old issue, stop the process directly without stop record, the record file can't play.
Signed-off-by: 21pages <sunboeasy@gmail.com>
* fix: TextFormField onChanged event triggered multiple times when Korean input
* logic fix for iOS
* add comments
* move 'onChanged' logic to handleSoftKeyboardInput