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
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
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
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
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.
Prevent the inner terminal scrollable from accepting user scrolling in the alternate buffer, so xterm's existing outer handler receives trackpad gestures after main-screen history.
Keep wheel encoding and drag reporting unchanged. Cover trackpad forwarding, subsequent drag coordinates, arrow fallback, and return to local scrollback.
* fix(flutter): shrink the unzoomed remote cursor by DPR on macOS and Linux
With "Zoom cursor" off in Adaptive or Custom view, the remote cursor
bitmap was registered at scale 1.0. NSCursor and GdkCursor treat the
bitmap size as logical pixels, so on a HiDPI controller the cursor was
drawn DPR times larger than in Original view (which already passes
1/DPR) and than on Windows (whose cursor path is in physical pixels).
A HiDPI remote such as KDE Wayland sends a 48-64 px bitmap, which then
showed up 3-4x too big on a Retina Mac.
Scale the bitmap by 1/DPR in that case, and scale the Flutter-painted
cursor used while the peer moves the mouse the same way so its size
does not jump. The new branch is an identity at DPR 1 and the Windows
paths are untouched.
Fixes https://github.com/rustdesk/rustdesk/discussions/15363
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K8HSHTEHo27mDcJXVyVfSP
* fix(flutter): check the cursor height against the min cursor size
`_checkUpdateScale` computed the scaled height from `width`, so the
min-size clamp never looked at the height.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K8HSHTEHo27mDcJXVyVfSP
* fix(flutter): keep the painted cursor hotspot in place when zoom cursor is off
`CursorPaint` subtracted the hotspot in remote pixels and then scaled it
by the canvas scale, but drew the image at scale 1.0, so the hotspot
landed hotx * (1 - scale) logical pixels away from the remote cursor
position. Cursors with a centered hotspot (I-beam, crosshair) were off
by up to half their size in Adaptive view. Subtract the hotspot after
scaling the position instead.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K8HSHTEHo27mDcJXVyVfSP
* fix(flutter): read the live DPR and clamp the painted cursor like the native one
CanvasModel caches devicePixelRatio and only refreshes it when the view
style changes, so after the window moves to a monitor with a different
DPR the unzoomed cursor kept the previous monitor's scale. Read it from
MediaQuery instead, which also rebuilds the cursor when it changes.
The native path clamps the scaled bitmap to kMinCursorSize; apply the
same clamp to the painted cursor so a small cursor does not change size
when the peer moves the mouse.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01K8HSHTEHo27mDcJXVyVfSP
* build(cursor): declare existing Zstd dependency for bounded decoding
* fix(cursor): validate and bound received cursor images
* fix(cursor): preserve thin cursor sizes when scaling
* fix(cursor): limit view changes to native cursor sizing
* fix(cursor): apply long-edge minimum to Web cursor sizing
* fix: preserve Linux cursor alpha and match Windows Custom scale
* fix(cursor): keep scaled buffers and raster dimensions in sync
* fix(cursor): preserve Windows peer alpha when resizing
* fix(cursor): preserve mixed alpha during downsampling
* comments
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(cursor): match desktop cursor size to peer display density
* refactor(cursor): clarify desktop and web scale branches
* fix(cursor): correct adaptive display scaling and fixed cursor size
* fix(cursor): apply all-display adaptive density on every desktop
* fix(cursor): normalize unzoomed all-display cursor density
* fix(cursor): synchronize original view on DPR changes
* fix(cursor): match original zoom to the active renderer
* fix: align Wayland software scrolling with input coordinates
* fix: preserve scroll offsets when switching scrolling modes
* fix: correct cursor size and pointer mapping with custom scale
Apply the hovered Linux display density to Custom cursor zoom in All Displays. Refresh scroll fractions after layout so changing the Custom percentage uses current scrollbar extents and detached controllers.
Validated with macOS and Windows component tests, formatting, and static analysis.
* fix(linux): pad tall native cursors to prevent clipping
* fix(cursor): preserve macOS point size in unzoomed views
* fix(linux): pad rectangular cursors to square canvases
* fix(cursor): divide dpr on Linux -> macOS
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix: cursor size test
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(cursor): cursor size of controlled side macOS
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(cursor): limit received cursor allocations
* test(cursor): retain reverse raster transition coverage
* fix(cursor): bound compressed cursor input
* fix(cursor): restrict the scaled size of the cursor
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(cursor): align hotspots with resized raster dimensions
Calculate each hotspot axis from the actual raster-to-source ratio.
Update existing boundary tests and run cursor tests in Flutter CI.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(cursor): align Sciter hotspots with raster dimensions
Calculate native and overlay hotspots from the final raster size.
Preserve input coordinate scaling and cursor refresh order.
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(cursor): use `1.0` instead of `1.0/dpr` on Linux -> macOS, Zoom off, Scale adaptive
The mouse cursor currently appears somewhat large.
However, this is difficult to adjust because the cursor size
is fixed while the window size varies; its relative size
depends on the specific desktop environment.
We can modify it if users actually provide feedback.
Ideally, we should check the "Zoom cursor" option.
Further adjustments may also be needed later based on cursor density.
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
Co-authored-by: rustdesk <info@rustdesk.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
The row was added for the web client, which has no session tab to name
the transport on, but nothing gated it: a desktop session over WebRTC
showed it too, duplicating the tab tooltip's "(WebRTC)". The getter now
answers only on the web, as its own comment intended.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QcJZZeJ3Nqb2MHxXuUadkb
* fix(terminal): send SGR mouse wheel reports with the button codes apps expect
xterm.dart 4.0.0 encodes the wheel buttons as 64+4..64+7 rather than
64+0..64+3, so the low bits land on the modifier field and every wheel
report the terminal emits reads as wheel-with-Shift. Strict full-screen
applications reject the modified event, which is why neither the mouse
wheel nor the trackpad scrolls anything once the peer application takes
over the alternate screen.
Install a mouse handler that keeps every upstream reporting decision and
only re-encodes the wheel buttons as 64..67. Non-wheel reports pass
through untouched, and the emitted bytes stay identical once upstream
ships the same fix, so this can be dropped without a behavior change.
Upstream: TerminalStudio/xterm.dart#238
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(terminal): correct the wheel report row, drop the wasted report build
Address review feedback on the wheel button fix:
- The X10/utf row was encoded as `32 + y + 1` while y is already 1-based, so
every normal-mode report pointed one row too low and the `y > limit` guard
disagreed with what it emitted.
- Gate the wheel path on `mouseMode.reportScroll` and the button state instead
of building and discarding a full report string from `defaultMouseHandler`
on every scroll tick. This also makes the hardcoded SGR 'M' provably right,
since a wheel release now returns before the report is built.
- Derive the wire code as `id - 4` and drop `_wheelButtonId`, whose `default`
branch was unreachable and defeated enum exhaustiveness.
- Assign `mouseHandler` after construction so the `Terminal(...)` line stays
untouched.
Cover the utf, urxvt, null-byte overflow and click-only branches, and assert
that TerminalModel actually installs the handler.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
