rayelliott
/
barrier
mirror of https://github.com/debauchee/barrier.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Changed win32 client side cursor warping to be all relative motion 

when not on the primary monitor.  This should eliminate the flicker
between virtual display 0,0 and the correct position.  While this
allows the user to confuse synergy by using the client's mouse,
synergy recovers quickly and easily from any confusion.
This commit is contained in:
crs 2003-02-16 19:55:54 +00:00
parent 2fc8780285
commit dd339fe375
1 changed files with 25 additions and 80 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

105
lib/platform/CMSWindowsSecondaryScreen.cpp
View File

@ -350,35 +350,16 @@ CMSWindowsSecondaryScreen::hideWindow()
void
CMSWindowsSecondaryScreen::warpCursor(SInt32 x, SInt32 y)
{
typedef UINT (WINAPI *SendInput_t)(UINT, LPINPUT, int);
static bool gotSendInput = false;
static SendInput_t SendInput = NULL;
// motion is simple (i.e. it's on the primary monitor) if there
// is only one monitor.
bool simple = !m_screen->isMultimon();
/* disable attempts to use simple motion with multiple monitors for now
if (!simple) {
// also simple if motion is within the primary monitor
simple = (x >= 0 && x < GetSystemMetrics(SM_CXSCREEN) &&
y >= 0 && y < GetSystemMetrics(SM_CYSCREEN));
if (!simple && !m_is95Family) {
// also simple if not on windows 95 family since the
// NT family mouse_event() allows absolute moves to
// any monitor.
//
// note -- this is possibly untrue if the primary
// monitor isn't upper-left most so limit it to that
// situation.
SInt32 x0, y0, w, h;
m_screen->getShape(x0, y0, w, h);
simple = (x0 == 0 && y0 == 0);
}
}
*/
// move the mouse directly to target position on NT family or if
// not using multiple monitors.
// move the mouse directly to target position if motion is simple
if (simple) {
SInt32 x0, y0, w, h;
m_screen->getShape(x0, y0, w, h);
@ -394,39 +375,29 @@ CMSWindowsSecondaryScreen::warpCursor(SInt32 x, SInt32 y)
// words, "by design." apparently the designers of windows 2000
// we're a little less lazy and did it right.
//
// we use the microsoft recommendation (Q193003): set the absolute
// position on the primary monitor, disable mouse acceleration,
// relative move the mouse to the final location, restore mouse
// acceleration. to avoid one kind of race condition (the user
// clicking the mouse or pressing a key between the absolute and
// relative move) we'll use SendInput() which guarantees that the
// events are delivered uninterrupted. we cannot prevent changes
// to the mouse acceleration at inopportune times, though. if
// SendInput() is unavailable then use mouse_event(); SendInput()
// is not available on Windows 95 and NT 4.0 prior to SP3.
// microsoft recommends in Q193003 to absolute position the cursor
// somewhere on the primary monitor then relative move to the
// desired location. this doesn't work for us because when the
// user drags a scrollbar, a window, etc. it causes the dragged
// item to jump back a forth between the position on the primary
// monitor and the desired position. while it always ends up in
// the right place, the effect is disconcerting.
//
// point-to-activate (x-mouse) doesn't seem to be bothered by the
// absolute/relative combination. a window over the absolute
// position (0,0) does *not* get activated (at least not on win2k)
// if the relative move puts the cursor elsewhere. similarly, the
// app under the final mouse position does *not* get deactivated
// by the absolute move to 0,0.
// instead we'll get the cursor's current position and do just a
// relative move from there to the desired position. relative
// moves are subject to cursor acceleration which we don't want.
// so we disable acceleration, do the relative move, then restore
// acceleration. there's a slight chance we'll end up in the
// wrong place if the user moves the cursor using this system's
// mouse while simultaneously moving the mouse on the server
// system. that defeats the purpose of synergy so we'll assume
// that won't happen. even if it does, the next mouse move will
// correct the position.
else {
// lookup SendInput() function
if (!gotSendInput) {
gotSendInput = true;
HINSTANCE user32 = LoadLibrary("user32.dll");
if (user32 != NULL) {
SendInput = reinterpret_cast<SendInput_t>(
GetProcAddress(user32, "SendInput"));
FreeLibrary(user32);
}
}
// save mouse speed & acceleration
int oldSpeed[4];
bool accelChanged =
SystemParametersInfo(SPI_GETMOUSE,0, oldSpeed,0) &&
SystemParametersInfo(SPI_GETMOUSE,0, oldSpeed, 0) &&
SystemParametersInfo(SPI_GETMOUSESPEED, 0, oldSpeed + 3, 0);
// use 1:1 motion
@ -437,38 +408,12 @@ CMSWindowsSecondaryScreen::warpCursor(SInt32 x, SInt32 y)
SystemParametersInfo(SPI_SETMOUSESPEED, 0, newSpeed + 3, 0);
}
// send events
INPUT events[2];
events[0].type = INPUT_MOUSE;
events[0].mi.dx = 0;
events[0].mi.dy = 0;
events[0].mi.mouseData = 0;
events[0].mi.dwFlags = MOUSEEVENTF_ABSOLUTE | MOUSEEVENTF_MOVE;
events[0].mi.time = GetTickCount();
events[0].mi.dwExtraInfo = 0;
events[1].type = INPUT_MOUSE;
events[1].mi.dx = x;
events[1].mi.dy = y;
events[1].mi.mouseData = 0;
events[1].mi.dwFlags = MOUSEEVENTF_MOVE;
events[1].mi.time = events[0].mi.time;
events[1].mi.dwExtraInfo = 0;
if (SendInput != NULL) {
SendInput(sizeof(events) / sizeof(events[0]),
events, sizeof(events[0]));
}
else {
mouse_event(events[0].mi.dwFlags,
events[0].mi.dx,
events[0].mi.dy,
events[0].mi.mouseData,
events[0].mi.dwExtraInfo);
mouse_event(events[1].mi.dwFlags,
events[1].mi.dx,
events[1].mi.dy,
events[1].mi.mouseData,
events[1].mi.dwExtraInfo);
}
// get current mouse position
POINT pos;
GetCursorPos(&pos);
// move relative to mouse position
mouse_event(MOUSEEVENTF_MOVE, x - pos.x, y - pos.y, 0, 0);
// restore mouse speed & acceleration
if (accelChanged) {