* avatar
* refactor avatar display: unify rendering and resolve at use time
- Extract buildAvatarWidget() in common.dart to share avatar rendering
logic across desktop settings, desktop CM and mobile CM
- Add resolve_avatar_url() in Rust, exposed via FFI (SyncReturn),
to resolve relative avatar paths (e.g. "/avatar/xxx") to absolute URLs
- Store avatar as-is in local config, only resolve when displaying
(settings page) or sending (LoginRequest)
- Resolve avatar in LoginRequest before sending to remote peer
- Add error handling for network image load failures
- Guard against empty client.name[0] crash
- Show avatar in mobile settings page account tile
Signed-off-by: 21pages <sunboeasy@gmail.com>
* web: implement mainResolveAvatarUrl via js getByName
Signed-off-by: 21pages <sunboeasy@gmail.com>
* increase ipc Data enum size limit to 120 bytes
Signed-off-by: 21pages <sunboeasy@gmail.com>
---------
Signed-off-by: 21pages <sunboeasy@gmail.com>
Co-authored-by: 21pages <sunboeasy@gmail.com>
* feat(terminal): add reconnection buffer support for persistent sessions
Fix two related issues:
1. Reconnecting to persistent sessions shows blank screen - server now
automatically sends historical buffer on reconnection via SessionState
machine with pending_buffer, eliminating the need for client-initiated
buffer requests.
2. Terminal output before view ready causes NaN errors - buffer output
chunks on client side until terminal view has valid dimensions, then
flush in order on first valid resize.
Rust side:
- Introduce SessionState enum (Closed/Active) replacing bool is_opened
- Auto-attach pending buffer on reconnection in handle_open()
- Always drain output channel in read_outputs() to prevent overflow
- Increase channel buffer from 100 to 500
- Optimize get_recent() to collect whole chunks (avoids ANSI truncation)
- Extract create_terminal_data_response() helper (DRY)
- Add reconnected flag to TerminalOpened protobuf message
Flutter side:
- Buffer output chunks until terminal view has valid dimensions
- Flush buffered output on first valid resize via _markViewReady()
- Clear terminal on reconnection to avoid duplicate output from buffer replay
- Fix max_bytes type (u32) to match protobuf definition
- Pass reconnected field through FlutterHandler event
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(terminal): add two-phase SIGWINCH for TUI app redraw and session remap on reconnection
Fix TUI apps (top, htop) not redrawing after reconnection. A single
resize-then-restore is too fast for ncurses to detect a size change,
so split across two read_outputs() polling cycles (~30ms apart) to
force a full redraw.
Also fix reconnection failure when client terminal_id doesn't match
any surviving server-side session ID by remapping the lowest surviving
session to the requested ID.
Rust side:
- Add two-phase SIGWINCH state machine (SigwinchPhase: TempResize →
Restore → Idle) with retry logic (max 3 attempts per phase)
- Add do_sigwinch_resize() for cross-platform PTY resize (direct PTY
and Windows helper mode)
- Add session remap logic for non-contiguous terminal_id reconnection
- Extract try_send_output() helper with rate-limited drop logging (DRY)
- Add 3-byte limit to UTF-8 continuation byte skipping in get_recent()
to prevent runaway on non-UTF-8 binary data
- Remove reconnected flag from flutter.rs (unused on client side)
Flutter side:
- Add reconnection screen clear and deferred flush logic
- Filter self from persistent_sessions restore list
- Add comments for web-related changes
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
* fix(terminal): ensure tab close is resilient to session cleanup failures
- Wrap _closeTerminalSessionIfNeeded in isolated try/catch so that
tabController.closeBy always executes even if FFI calls throw
- Add clarifying comment in handleWindowCloseButton for single-tab
audit dialog flow
* fix(terminal): fix session reconnect ID mismatch and tab close race condition
Remap surviving persistent sessions to client-requested terminal IDs on
reconnect, preventing new shell creation when IDs are non-contiguous.
Snapshot peerTabCount before async operations in _closeTab to avoid race
with concurrent _closeAllTabs clearing the tab controller.
Remove debug log statements.
Signed-off-by: fufesou <linlong1266@gmail.com>
---------
Signed-off-by: fufesou <linlong1266@gmail.com>
* feat(macos): add privacy mode support for macOS
## Summary
Add privacy mode functionality for macOS platform, allowing remote
desktop sessions to hide the screen content from local users.
## Changes
### Core Implementation (src/platform/macos.mm)
- Implement screen blackout using CGDisplayGammaTable API
- Implement input blocking using CGEventTap to intercept keyboard/mouse
- Store and restore original gamma values for proper cleanup
### Privacy Mode Integration (src/privacy_mode.rs, src/privacy_mode/macos.rs)
- Add macOS privacy mode implementation with PrivacyMode trait
- Register macOS privacy mode in PRIVACY_MODE_CREATOR
- Set DEFAULT_PRIVACY_MODE_IMPL for macOS platform
- Implement get_supported_privacy_mode_impl() for macOS
### Connection Handling (src/server/connection.rs)
- Add supported_privacy_mode_impl to platform_additions for macOS
- Enable privacy mode toggle in client UI when connecting via LAN IP
### Localization (src/lang/*.rs)
- Add "privacy_mode_impl_macos_tip" translation for en/cn/tw
## Safety & Security
- Implements Drop trait to ensure cleanup on normal exit
- macOS system automatically restores gamma table on process termination
- CGEventTap is automatically released when process terminates
- Tested with SIGKILL to verify crash recovery
## Testing
- Verified privacy mode toggle works via both ID and LAN IP connection
- Verified screen recovery after process crash (kill -9)
- Verified input restoration after process termination
* refactor: use existing 'Privacy mode' translation key
* refactor: rename gamma channel variables for better readability - rename r/g/b to red/green/blue to avoid variable shadowing confusion
* fix: add error handling for gamma table restoration with fallback to system reset
* fix: add error handling for CGEventTapCreate failure in privacy mode
* fix: only set display to black if original gamma was saved successfully
* fix: add error handling for CGSetDisplayTransferByTable when setting display to black
* fix: improve event tap callback to properly distinguish remote input from local input
* fix: missing macos.rs
* Fix: Add display validation before restoring gamma values
* Fix: Add mutex lock for thread safety in MacSetPrivacyMode
* Fix: Handle return values and add missing mouse events in macos privacy mode
* fix: only set conn_id after privacy mode is successfully turned on
* fix: reimplement privacy mode with stable display identification
Address code review concern: original gamma values stored with DisplayID
as key could become stale if display list changes between privacy mode
activations (e.g., display reconnected with different ID).
Solution:
- Use UUID instead of DisplayID as storage key (stable across reconnections)
- Clear g_originalGammas when privacy mode is turned off
- Register CGDisplayReconfigurationCallback to handle hot-plug events
- Validate display state via FindDisplayIdByUUID() before restoration
Key features:
- UUID-based display identification (stable across reconnections)
- Hot-plug support via CGDisplayReconfigurationCallback
- EventTap auto re-enable on system timeout
- Fallback to CGDisplayRestoreColorSyncSettings() for recovery
- Detailed error logging with display name/ID/UUID
* fix: ensure EventTap runs on main thread and improve gamma restore error handling
- Add SetupEventTapOnMainThread() to create EventTap on main thread using dispatch_sync, avoiding potential issues when called from background threads
- Add TeardownEventTapOnMainThread() for consistent cleanup on main thread
- Check [NSThread isMainThread] to avoid deadlock when already on main thread
- Add error tracking for gamma restoration during cleanup
- Use CGDisplayRestoreColorSyncSettings() as fallback when individual gamma restoration fails
* fix: remove invalid eventMask bits that caused undefined behavior in input blocking
* fix: address code review comments for macos privacy mode implementation
Changes to src/privacy_mode/macos.rs:
- Add check_on_conn_id() in turn_on_privacy() to prevent duplicate activation
- Add check_off_conn_id() in turn_off_privacy() to validate connection ID
- Add self.conn_id = 0 in clear() to reset connection state
Changes to src/platform/macos.mm:
- Add link comment for ENIGO_INPUT_EXTRA_VALUE referencing libs/enigo/src/macos/macos_impl.rs
- Fix NSLog format string mismatch (5 placeholders vs 4 values)
- Make ApplyBlackoutToDisplay() return bool for proper error handling
- Return false when UUID is empty since privacy mode requires ALL displays
- Add else branches with logging for:
- CGGetDisplayTransferByTable failures
- Zero gamma table capacity (not supported)
- Zero blackout capacity
- Remove unused g_uuidToDisplayId variable (was only written, never read)
* fix(macos): add early return with privacy mode exit on display hotplug failures
Why large-scale changes are needed:
The code review suggested adding early return when errors occur in
DisplayReconfigurationCallback. However, simply returning early is not
enough - when a newly connected display cannot be blacked out, we must
exit privacy mode entirely to maintain security guarantees.
The challenge is that DisplayReconfigurationCallback already holds
g_privacyModeMutex, so calling MacSetPrivacyMode(false) directly would
cause a deadlock. This necessitated:
1. Extract TurnOffPrivacyModeInternal() - a lock-free internal function
that can be safely called from within the callback
2. Refactor MacSetPrivacyMode(false) branch to use this internal function
3. Add early returns with TurnOffPrivacyModeInternal() calls at each
failure point in DisplayReconfigurationCallback
Changes in DisplayReconfigurationCallback:
- UUID empty: log + exit privacy mode + early return
- Gamma table capacity zero: log + exit privacy mode + early return
- CGGetDisplayTransferByTable fails: log + exit privacy mode + early return
- ApplyBlackoutToDisplay fails: log + exit privacy mode + early return
* fix(macos): address code review feedback and improve privacy mode stability
Code Review Fixes:
- Add detailed comments for potential deadlock scenarios in dispatch_sync
with g_privacyModeMutex (SetupEventTapOnMainThread/TeardownEventTapOnMainThread)
- Use async dispatch for privacy mode shutdown from DisplayReconfigurationCallback
to avoid unregistering callback from within itself
- Extract RestoreAllGammas() helper function to reduce code duplication
- Fix Drop implementation in macos.rs to call self.clear() for consistency
- Add comment explaining why _state parameter is ignored on macOS
- Define DISPLAY_RECONFIG_MONITOR_DURATION_MS and GAMMA_CHECK_INTERVAL_MS constants
- Add gamma restoration when UUID retrieval fails during privacy mode activation
Privacy Mode Stability Improvements (Continuous Resolution Changes):
- Implement continuous gamma value monitoring with timer polling after display
reconfiguration to handle rapid successive resolution changes
- Monitor gamma values every 200ms for 5 seconds after each resolution change
- Automatically reapply blackout if system (ColorSync) restores gamma
- Add IsDisplayBlackedOut() to detect if display gamma has been restored
- Use timestamp-based debouncing: monitoring period automatically extends
when new reconfig events occur during active monitoring
- Ensure blackout remains effective even under continuous resolution changes
where macOS may asynchronously restore gamma values multiple times
This ensures privacy mode remains stable and effective when users rapidly
change display resolution multiple times in succession.
---------
Co-authored-by: libin <libin.chat@outlook.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>
1. `TERM` on linux terminal.
2. `htop` command not found on macOS.
3. `vim` and `claude code cli` hung up on windows.
Signed-off-by: fufesou <linlong1266@gmail.com>
- Route Windows server-to-client file reads through CM instead of the connection layer
- Add FS IPC commands (ReadFile, CancelRead, SendConfirmForRead, ReadAllFiles) and CM data messages
(ReadJobInitResult, FileBlockFromCM, FileReadDone, FileReadError, FileDigestFromCM, AllFilesResult)
- Track pending read validations and read jobs to coordinate CM-driven file transfers and clean them up
on completion, cancellation, and errors
- Enforce a configurable file-transfer-max-files limit for ReadAllFiles and add stronger file name/path
validation on the CM side
- Improve Flutter file transfer UX and robustness:
- Use explicit percent/percentText progress fields
- Derive speed and cancel actions from the active job
- Handle job errors via FileModel.handleJobError and complete pending recursive tasks on failure
- Wrap recursive directory operations in try/catch and await sendRemoveEmptyDir when removing empty directories
Signed-off-by: fufesou <linlong1266@gmail.com>
Set TERM=xterm-256color when spawning PTY shell to ensure proper
handling of control sequences. This fixes the issue where Delete/
Backspace keys were not working in terminal connections, particularly
from iPad to Linux.
Fixes#13621
Enhance security by implementing rate limiting on IPv6 prefixes (/64, /56, /48)
to prevent brute force attacks that exploit cheap IPv6 address generation.
* Add private get_ipv6_prefixes() to calculate network prefixes
* Implement private check_failure_ipv6_prefix() for prefix-specific limits
on IPv6 addresses
* Refactor check_failure() and update_failure() to support both IPs and prefixes
* Add ExceedIPv6PrefixAttempts to AlarmAuditType enum
Signed-off-by: Michael Bacarella <m@bacarella.com>
* Add Wayland multi-monitor screen capture functionality
* fix wayland capture issues by reverting to CapturerPtr, the problem was that calling Display::all in get_capturer_for_display was dropping the pipewire capturer and causing the video to freeze.
* If running as AppImage or flatpak, ignore the 'multiple' argument
* Comment out warning log with unclear purpose Comment out warning log with unclear purpose
---------
Co-authored-by: fufesou <13586388+fufesou@users.noreply.github.com>
1. Update kcp-sys to send KCP in frames to avoid potential crashes.
2. Fix the issue when the controling side is closed, the kcp connection close is not immediately recognized by the controlled end.
* Unless the controling side receives the close reason, force the sending of the close reason to the controlled end when using KCP, and delay for 30ms to ensure the message is sent successfully.
* Move the CloseReason receiving forward, as this message needs to be received when unauthorized, especially for kcp.
Signed-off-by: 21pages <sunboeasy@gmail.com>
