summaryrefslogtreecommitdiff
path: root/packages/xorg-xserver/xserver-kdrive-1.4.0.90/xorg-1.4-kdrive-rotation.patch
blob: 5e3eb8ee146392ae2c7f6f104d6a93c23829e55b (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
diff -rup a/hw/kdrive/src/kinput.c b/hw/kdrive/src/kinput.c
--- a/hw/kdrive/src/kinput.c	2007-08-23 21:04:53.000000000 +0200
+++ b/hw/kdrive/src/kinput.c	2007-12-22 12:53:27.679853402 +0100
@@ -2075,19 +2075,25 @@ KdEnqueuePointerEvent(KdPointerInfo *pi,
 
     /* we don't need to transform z, so we don't. */
     if (flags & KD_MOUSE_DELTA) {
+/* does it really make sense to transform relative coordinates depending on screen rotation??
 	if (pi->transformCoordinates) {
 	    x = matrix[0][0] * rx + matrix[0][1] * ry;
 	    y = matrix[1][0] * rx + matrix[1][1] * ry;
 	}
-	else {
+	else {*/
 	    x = rx;
 	    y = ry;
-	}
+/*	}*/
     }
     else {
 	if (pi->transformCoordinates) {
 	    x = matrix[0][0] * rx + matrix[0][1] * ry;
 	    y = matrix[1][0] * rx + matrix[1][1] * ry;
+/* negative absolute values indicate calculation from the oposite end of the axis */
+            if (x < 0)
+                x += matrix[0][2];
+            if (y < 0)
+                y += matrix[1][2];
 	}
 	else {
 	    x = rx;