I've no idea why CVS thinks that the .mcp files have been changed - they haven't. git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@12589 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
		
			
				
	
	
		
			325 lines
		
	
	
		
			8.3 KiB
		
	
	
	
		
			C
		
	
	
	
	
	
			
		
		
	
	
			325 lines
		
	
	
		
			8.3 KiB
		
	
	
	
		
			C
		
	
	
	
	
	
| /*
 | |
|  * (c) Copyright 1993, 1994, Silicon Graphics, Inc.
 | |
|  * ALL RIGHTS RESERVED
 | |
|  * Permission to use, copy, modify, and distribute this software for
 | |
|  * any purpose and without fee is hereby granted, provided that the above
 | |
|  * copyright notice appear in all copies and that both the copyright notice
 | |
|  * and this permission notice appear in supporting documentation, and that
 | |
|  * the name of Silicon Graphics, Inc. not be used in advertising
 | |
|  * or publicity pertaining to distribution of the software without specific,
 | |
|  * written prior permission.
 | |
|  *
 | |
|  * THE MATERIAL EMBODIED ON THIS SOFTWARE IS PROVIDED TO YOU "AS-IS"
 | |
|  * AND WITHOUT WARRANTY OF ANY KIND, EXPRESS, IMPLIED OR OTHERWISE,
 | |
|  * INCLUDING WITHOUT LIMITATION, ANY WARRANTY OF MERCHANTABILITY OR
 | |
|  * FITNESS FOR A PARTICULAR PURPOSE.  IN NO EVENT SHALL SILICON
 | |
|  * GRAPHICS, INC.  BE LIABLE TO YOU OR ANYONE ELSE FOR ANY DIRECT,
 | |
|  * SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY
 | |
|  * KIND, OR ANY DAMAGES WHATSOEVER, INCLUDING WITHOUT LIMITATION,
 | |
|  * LOSS OF PROFIT, LOSS OF USE, SAVINGS OR REVENUE, OR THE CLAIMS OF
 | |
|  * THIRD PARTIES, WHETHER OR NOT SILICON GRAPHICS, INC.  HAS BEEN
 | |
|  * ADVISED OF THE POSSIBILITY OF SUCH LOSS, HOWEVER CAUSED AND ON
 | |
|  * ANY THEORY OF LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE
 | |
|  * POSSESSION, USE OR PERFORMANCE OF THIS SOFTWARE.
 | |
|  *
 | |
|  * US Government Users Restricted Rights
 | |
|  * Use, duplication, or disclosure by the Government is subject to
 | |
|  * restrictions set forth in FAR 52.227.19(c)(2) or subparagraph
 | |
|  * (c)(1)(ii) of the Rights in Technical Data and Computer Software
 | |
|  * clause at DFARS 252.227-7013 and/or in similar or successor
 | |
|  * clauses in the FAR or the DOD or NASA FAR Supplement.
 | |
|  * Unpublished-- rights reserved under the copyright laws of the
 | |
|  * United States.  Contractor/manufacturer is Silicon Graphics,
 | |
|  * Inc., 2011 N.  Shoreline Blvd., Mountain View, CA 94039-7311.
 | |
|  *
 | |
|  * OpenGL(TM) is a trademark of Silicon Graphics, Inc.
 | |
|  */
 | |
| /*
 | |
|  * Trackball code:
 | |
|  *
 | |
|  * Implementation of a virtual trackball.
 | |
|  * Implemented by Gavin Bell, lots of ideas from Thant Tessman and
 | |
|  *   the August '88 issue of Siggraph's "Computer Graphics," pp. 121-129.
 | |
|  *
 | |
|  * Vector manip code:
 | |
|  *
 | |
|  * Original code from:
 | |
|  * David M. Ciemiewicz, Mark Grossman, Henry Moreton, and Paul Haeberli
 | |
|  *
 | |
|  * Much mucking with by:
 | |
|  * Gavin Bell
 | |
|  */
 | |
| #include <math.h>
 | |
| #include "trackball.h"
 | |
| 
 | |
| /*
 | |
|  * This size should really be based on the distance from the center of
 | |
|  * rotation to the point on the object underneath the mouse.  That
 | |
|  * point would then track the mouse as closely as possible.  This is a
 | |
|  * simple example, though, so that is left as an Exercise for the
 | |
|  * Programmer.
 | |
|  */
 | |
| #define TRACKBALLSIZE  (0.8f)
 | |
| 
 | |
| /*
 | |
|  * Local function prototypes (not defined in trackball.h)
 | |
|  */
 | |
| static float tb_project_to_sphere(float, float, float);
 | |
| static void normalize_quat(float [4]);
 | |
| 
 | |
| void
 | |
| vzero(float *v)
 | |
| {
 | |
|     v[0] = 0.0;
 | |
|     v[1] = 0.0;
 | |
|     v[2] = 0.0;
 | |
| }
 | |
| 
 | |
| void
 | |
| vset(float *v, float x, float y, float z)
 | |
| {
 | |
|     v[0] = x;
 | |
|     v[1] = y;
 | |
|     v[2] = z;
 | |
| }
 | |
| 
 | |
| void
 | |
| vsub(const float *src1, const float *src2, float *dst)
 | |
| {
 | |
|     dst[0] = src1[0] - src2[0];
 | |
|     dst[1] = src1[1] - src2[1];
 | |
|     dst[2] = src1[2] - src2[2];
 | |
| }
 | |
| 
 | |
| void
 | |
| vcopy(const float *v1, float *v2)
 | |
| {
 | |
|     register int i;
 | |
|     for (i = 0 ; i < 3 ; i++)
 | |
|         v2[i] = v1[i];
 | |
| }
 | |
| 
 | |
| void
 | |
| vcross(const float *v1, const float *v2, float *cross)
 | |
| {
 | |
|     float temp[3];
 | |
| 
 | |
|     temp[0] = (v1[1] * v2[2]) - (v1[2] * v2[1]);
 | |
|     temp[1] = (v1[2] * v2[0]) - (v1[0] * v2[2]);
 | |
|     temp[2] = (v1[0] * v2[1]) - (v1[1] * v2[0]);
 | |
|     vcopy(temp, cross);
 | |
| }
 | |
| 
 | |
| float
 | |
| vlength(const float *v)
 | |
| {
 | |
|     return (float) sqrt(v[0] * v[0] + v[1] * v[1] + v[2] * v[2]);
 | |
| }
 | |
| 
 | |
| void
 | |
| vscale(float *v, float div)
 | |
| {
 | |
|     v[0] *= div;
 | |
|     v[1] *= div;
 | |
|     v[2] *= div;
 | |
| }
 | |
| 
 | |
| void
 | |
| vnormal(float *v)
 | |
| {
 | |
|     vscale(v, 1.0f/vlength(v));
 | |
| }
 | |
| 
 | |
| float
 | |
| vdot(const float *v1, const float *v2)
 | |
| {
 | |
|     return v1[0]*v2[0] + v1[1]*v2[1] + v1[2]*v2[2];
 | |
| }
 | |
| 
 | |
| void
 | |
| vadd(const float *src1, const float *src2, float *dst)
 | |
| {
 | |
|     dst[0] = src1[0] + src2[0];
 | |
|     dst[1] = src1[1] + src2[1];
 | |
|     dst[2] = src1[2] + src2[2];
 | |
| }
 | |
| 
 | |
| /*
 | |
|  * Ok, simulate a track-ball.  Project the points onto the virtual
 | |
|  * trackball, then figure out the axis of rotation, which is the cross
 | |
|  * product of P1 P2 and O P1 (O is the center of the ball, 0,0,0)
 | |
|  * Note:  This is a deformed trackball-- is a trackball in the center,
 | |
|  * but is deformed into a hyperbolic sheet of rotation away from the
 | |
|  * center.  This particular function was chosen after trying out
 | |
|  * several variations.
 | |
|  *
 | |
|  * It is assumed that the arguments to this routine are in the range
 | |
|  * (-1.0 ... 1.0)
 | |
|  */
 | |
| void
 | |
| trackball(float q[4], float p1x, float p1y, float p2x, float p2y)
 | |
| {
 | |
|     float a[3]; /* Axis of rotation */
 | |
|     float phi;  /* how much to rotate about axis */
 | |
|     float p1[3], p2[3], d[3];
 | |
|     float t;
 | |
| 
 | |
|     if (p1x == p2x && p1y == p2y) {
 | |
|         /* Zero rotation */
 | |
|         vzero(q);
 | |
|         q[3] = 1.0;
 | |
|         return;
 | |
|     }
 | |
| 
 | |
|     /*
 | |
|      * First, figure out z-coordinates for projection of P1 and P2 to
 | |
|      * deformed sphere
 | |
|      */
 | |
|     vset(p1, p1x, p1y, tb_project_to_sphere(TRACKBALLSIZE, p1x, p1y));
 | |
|     vset(p2, p2x, p2y, tb_project_to_sphere(TRACKBALLSIZE, p2x, p2y));
 | |
| 
 | |
|     /*
 | |
|      *  Now, we want the cross product of P1 and P2
 | |
|      */
 | |
|     vcross(p2,p1,a);
 | |
| 
 | |
|     /*
 | |
|      *  Figure out how much to rotate around that axis.
 | |
|      */
 | |
|     vsub(p1, p2, d);
 | |
|     t = vlength(d) / (2.0f*TRACKBALLSIZE);
 | |
| 
 | |
|     /*
 | |
|      * Avoid problems with out-of-control values...
 | |
|      */
 | |
|     if (t > 1.0) t = 1.0;
 | |
|     if (t < -1.0) t = -1.0;
 | |
|     phi = 2.0f * (float) asin(t);
 | |
| 
 | |
|     axis_to_quat(a,phi,q);
 | |
| }
 | |
| 
 | |
| /*
 | |
|  *  Given an axis and angle, compute quaternion.
 | |
|  */
 | |
| void
 | |
| axis_to_quat(float a[3], float phi, float q[4])
 | |
| {
 | |
|     vnormal(a);
 | |
|     vcopy(a, q);
 | |
|     vscale(q, (float) sin(phi/2.0));
 | |
|     q[3] = (float) cos(phi/2.0);
 | |
| }
 | |
| 
 | |
| /*
 | |
|  * Project an x,y pair onto a sphere of radius r OR a hyperbolic sheet
 | |
|  * if we are away from the center of the sphere.
 | |
|  */
 | |
| static float
 | |
| tb_project_to_sphere(float r, float x, float y)
 | |
| {
 | |
|     float d, t, z;
 | |
| 
 | |
|     d = (float) sqrt(x*x + y*y);
 | |
|     if (d < r * 0.70710678118654752440) {    /* Inside sphere */
 | |
|         z = (float) sqrt(r*r - d*d);
 | |
|     } else {           /* On hyperbola */
 | |
|         t = r / 1.41421356237309504880f;
 | |
|         z = t*t / d;
 | |
|     }
 | |
|     return z;
 | |
| }
 | |
| 
 | |
| /*
 | |
|  * Given two rotations, e1 and e2, expressed as quaternion rotations,
 | |
|  * figure out the equivalent single rotation and stuff it into dest.
 | |
|  *
 | |
|  * This routine also normalizes the result every RENORMCOUNT times it is
 | |
|  * called, to keep error from creeping in.
 | |
|  *
 | |
|  * NOTE: This routine is written so that q1 or q2 may be the same
 | |
|  * as dest (or each other).
 | |
|  */
 | |
| 
 | |
| #define RENORMCOUNT 97
 | |
| 
 | |
| void
 | |
| add_quats(float q1[4], float q2[4], float dest[4])
 | |
| {
 | |
|     static int count=0;
 | |
|     float t1[4], t2[4], t3[4];
 | |
|     float tf[4];
 | |
| 
 | |
|     vcopy(q1,t1);
 | |
|     vscale(t1,q2[3]);
 | |
| 
 | |
|     vcopy(q2,t2);
 | |
|     vscale(t2,q1[3]);
 | |
| 
 | |
|     vcross(q2,q1,t3);
 | |
|     vadd(t1,t2,tf);
 | |
|     vadd(t3,tf,tf);
 | |
|     tf[3] = q1[3] * q2[3] - vdot(q1,q2);
 | |
| 
 | |
|     dest[0] = tf[0];
 | |
|     dest[1] = tf[1];
 | |
|     dest[2] = tf[2];
 | |
|     dest[3] = tf[3];
 | |
| 
 | |
|     if (++count > RENORMCOUNT) {
 | |
|         count = 0;
 | |
|         normalize_quat(dest);
 | |
|     }
 | |
| }
 | |
| 
 | |
| /*
 | |
|  * Quaternions always obey:  a^2 + b^2 + c^2 + d^2 = 1.0
 | |
|  * If they don't add up to 1.0, dividing by their magnitued will
 | |
|  * renormalize them.
 | |
|  *
 | |
|  * Note: See the following for more information on quaternions:
 | |
|  *
 | |
|  * - Shoemake, K., Animating rotation with quaternion curves, Computer
 | |
|  *   Graphics 19, No 3 (Proc. SIGGRAPH'85), 245-254, 1985.
 | |
|  * - Pletinckx, D., Quaternion calculus as a basic tool in computer
 | |
|  *   graphics, The Visual Computer 5, 2-13, 1989.
 | |
|  */
 | |
| static void
 | |
| normalize_quat(float q[4])
 | |
| {
 | |
|     int i;
 | |
|     float mag;
 | |
| 
 | |
|     mag = (q[0]*q[0] + q[1]*q[1] + q[2]*q[2] + q[3]*q[3]);
 | |
|     for (i = 0; i < 4; i++) q[i] /= mag;
 | |
| }
 | |
| 
 | |
| /*
 | |
|  * Build a rotation matrix, given a quaternion rotation.
 | |
|  *
 | |
|  */
 | |
| void
 | |
| build_rotmatrix(float m[4][4], float q[4])
 | |
| {
 | |
|     m[0][0] = 1.0f - 2.0f * (q[1] * q[1] + q[2] * q[2]);
 | |
|     m[0][1] = 2.0f * (q[0] * q[1] - q[2] * q[3]);
 | |
|     m[0][2] = 2.0f * (q[2] * q[0] + q[1] * q[3]);
 | |
|     m[0][3] = 0.0f;
 | |
| 
 | |
|     m[1][0] = 2.0f * (q[0] * q[1] + q[2] * q[3]);
 | |
|     m[1][1]= 1.0f - 2.0f * (q[2] * q[2] + q[0] * q[0]);
 | |
|     m[1][2] = 2.0f * (q[1] * q[2] - q[0] * q[3]);
 | |
|     m[1][3] = 0.0f;
 | |
| 
 | |
|     m[2][0] = 2.0f * (q[2] * q[0] - q[1] * q[3]);
 | |
|     m[2][1] = 2.0f * (q[1] * q[2] + q[0] * q[3]);
 | |
|     m[2][2] = 1.0f - 2.0f * (q[1] * q[1] + q[0] * q[0]);
 | |
|     m[2][3] = 0.0f;
 | |
| 
 | |
|     m[3][0] = 0.0f;
 | |
|     m[3][1] = 0.0f;
 | |
|     m[3][2] = 0.0f;
 | |
|     m[3][3] = 1.0f;
 | |
| }
 | |
| 
 |