mirror of
https://github.com/brl/mutter.git
synced 2024-11-29 03:20:46 -05:00
72f4ddf532
Since using addresses that might change is something that finally the FSF acknowledge as a plausible scenario (after changing address twice), the license blurb in the source files should use the URI for getting the license in case the library did not come with it. Not that URIs cannot possibly change, but at least it's easier to set up a redirection at the same place. As a side note: this commit closes the oldes bug in Clutter's bug report tool. http://bugzilla.openedhand.com/show_bug.cgi?id=521
349 lines
9.3 KiB
C
349 lines
9.3 KiB
C
/*
|
|
* Cogl
|
|
*
|
|
* An object oriented GL/GLES Abstraction/Utility Layer
|
|
*
|
|
* Copyright (C) 2008,2009 Intel Corporation.
|
|
*
|
|
* This library is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
* License as published by the Free Software Foundation; either
|
|
* version 2 of the License, or (at your option) any later version.
|
|
*
|
|
* This library is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
* Lesser General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU Lesser General Public
|
|
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
|
|
*
|
|
*
|
|
*
|
|
* Authors:
|
|
* Robert Bragg <robert@linux.intel.com>
|
|
*/
|
|
|
|
#define USE_MESA_MATRIX_API
|
|
|
|
#include <cogl.h>
|
|
#include <cogl-matrix.h>
|
|
#include <cogl-matrix-private.h>
|
|
#ifdef USE_MESA_MATRIX_API
|
|
#include <cogl-matrix-mesa.h>
|
|
#endif
|
|
|
|
#include <glib.h>
|
|
#include <math.h>
|
|
#include <string.h>
|
|
|
|
void
|
|
_cogl_matrix_print (CoglMatrix *matrix)
|
|
{
|
|
float *m = (float *)matrix;
|
|
int y;
|
|
|
|
for (y = 0; y < 4; y++)
|
|
g_print ("\t%6.4f %6.4f %6.4f %6.4f\n", m[y], m[4+y], m[8+y], m[12+y]);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_init_identity (CoglMatrix *matrix)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
matrix->xx = 1; matrix->xy = 0; matrix->xz = 0; matrix->xw = 0;
|
|
matrix->yx = 0; matrix->yy = 1; matrix->yz = 0; matrix->yw = 0;
|
|
matrix->zx = 0; matrix->zy = 0; matrix->zz = 1; matrix->zw = 0;
|
|
matrix->wx = 0; matrix->wy = 0; matrix->wz = 0; matrix->ww = 1;
|
|
#else
|
|
_math_matrix_init_identity (matrix);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_multiply (CoglMatrix *result,
|
|
const CoglMatrix *a,
|
|
const CoglMatrix *b)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
CoglMatrix r;
|
|
|
|
/* row 0 */
|
|
r.xx = a->xx * b->xx + a->xy * b->yx + a->xz * b->zx + a->xw * b->wx;
|
|
r.xy = a->xx * b->xy + a->xy * b->yy + a->xz * b->zy + a->xw * b->wy;
|
|
r.xz = a->xx * b->xz + a->xy * b->yz + a->xz * b->zz + a->xw * b->wz;
|
|
r.xw = a->xx * b->xw + a->xy * b->yw + a->xz * b->zw + a->xw * b->ww;
|
|
|
|
/* row 1 */
|
|
r.yx = a->yx * b->xx + a->yy * b->yx + a->yz * b->zx + a->yw * b->wx;
|
|
r.yy = a->yx * b->xy + a->yy * b->yy + a->yz * b->zy + a->yw * b->wy;
|
|
r.yz = a->yx * b->xz + a->yy * b->yz + a->yz * b->zz + a->yw * b->wz;
|
|
r.yw = a->yx * b->xw + a->yy * b->yw + a->yz * b->zw + a->yw * b->ww;
|
|
|
|
/* row 2 */
|
|
r.zx = a->zx * b->xx + a->zy * b->yx + a->zz * b->zx + a->zw * b->wx;
|
|
r.zy = a->zx * b->xy + a->zy * b->yy + a->zz * b->zy + a->zw * b->wy;
|
|
r.zz = a->zx * b->xz + a->zy * b->yz + a->zz * b->zz + a->zw * b->wz;
|
|
r.zw = a->zx * b->xw + a->zy * b->yw + a->zz * b->zw + a->zw * b->ww;
|
|
|
|
/* row 3 */
|
|
r.wx = a->wx * b->xx + a->wy * b->yx + a->wz * b->zx + a->ww * b->wx;
|
|
r.wy = a->wx * b->xy + a->wy * b->yy + a->wz * b->zy + a->ww * b->wy;
|
|
r.wz = a->wx * b->xz + a->wy * b->yz + a->wz * b->zz + a->ww * b->wz;
|
|
r.ww = a->wx * b->xw + a->wy * b->yw + a->wz * b->zw + a->ww * b->ww;
|
|
|
|
/* The idea was that having this unrolled; it might be easier for the
|
|
* compiler to vectorize, but that's probably not true. Mesa does it
|
|
* using a single for (i=0; i<4; i++) approach, maybe that's better...
|
|
*/
|
|
|
|
*result = r;
|
|
#else
|
|
_math_matrix_multiply (result, a, b);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (result);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_rotate (CoglMatrix *matrix,
|
|
float angle,
|
|
float x,
|
|
float y,
|
|
float z)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
CoglMatrix rotation;
|
|
CoglMatrix result;
|
|
float c, s;
|
|
|
|
angle *= G_PI / 180.0f;
|
|
c = cosf (angle);
|
|
s = sinf (angle);
|
|
|
|
rotation.xx = x * x * (1.0f - c) + c;
|
|
rotation.yx = y * x * (1.0f - c) + z * s;
|
|
rotation.zx = x * z * (1.0f - c) - y * s;
|
|
rotation.wx = 0.0f;
|
|
|
|
rotation.xy = x * y * (1.0f - c) - z * s;
|
|
rotation.yy = y * y * (1.0f - c) + c;
|
|
rotation.zy = y * z * (1.0f - c) + x * s;
|
|
rotation.wy = 0.0f;
|
|
|
|
rotation.xz = x * z * (1.0f - c) + y * s;
|
|
rotation.yz = y * z * (1.0f - c) - x * s;
|
|
rotation.zz = z * z * (1.0f - c) + c;
|
|
rotation.wz = 0.0f;
|
|
|
|
rotation.xw = 0.0f;
|
|
rotation.yw = 0.0f;
|
|
rotation.zw = 0.0f;
|
|
rotation.ww = 1.0f;
|
|
|
|
cogl_matrix_multiply (&result, matrix, &rotation);
|
|
*matrix = result;
|
|
#else
|
|
_math_matrix_rotate (matrix, angle, x, y, z);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_translate (CoglMatrix *matrix,
|
|
float x,
|
|
float y,
|
|
float z)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
matrix->xw = matrix->xx * x + matrix->xy * y + matrix->xz * z + matrix->xw;
|
|
matrix->yw = matrix->yx * x + matrix->yy * y + matrix->yz * z + matrix->yw;
|
|
matrix->zw = matrix->zx * x + matrix->zy * y + matrix->zz * z + matrix->zw;
|
|
matrix->ww = matrix->wx * x + matrix->wy * y + matrix->wz * z + matrix->ww;
|
|
#else
|
|
_math_matrix_translate (matrix, x, y, z);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_scale (CoglMatrix *matrix,
|
|
float sx,
|
|
float sy,
|
|
float sz)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
matrix->xx *= sx; matrix->xy *= sy; matrix->xz *= sz;
|
|
matrix->yx *= sx; matrix->yy *= sy; matrix->yz *= sz;
|
|
matrix->zx *= sx; matrix->zy *= sy; matrix->zz *= sz;
|
|
matrix->wx *= sx; matrix->wy *= sy; matrix->wz *= sz;
|
|
#else
|
|
_math_matrix_scale (matrix, sx, sy, sz);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_frustum (CoglMatrix *matrix,
|
|
float left,
|
|
float right,
|
|
float bottom,
|
|
float top,
|
|
float z_near,
|
|
float z_far)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
float x, y, a, b, c, d;
|
|
CoglMatrix frustum;
|
|
|
|
x = (2.0f * z_near) / (right - left);
|
|
y = (2.0f * z_near) / (top - bottom);
|
|
a = (right + left) / (right - left);
|
|
b = (top + bottom) / (top - bottom);
|
|
c = -(z_far + z_near) / ( z_far - z_near);
|
|
d = -(2.0f * z_far* z_near) / (z_far - z_near);
|
|
|
|
frustum.xx = x;
|
|
frustum.yx = 0.0f;
|
|
frustum.zx = 0.0f;
|
|
frustum.wx = 0.0f;
|
|
|
|
frustum.xy = 0.0f;
|
|
frustum.yy = y;
|
|
frustum.zy = 0.0f;
|
|
frustum.wy = 0.0f;
|
|
|
|
frustum.xz = a;
|
|
frustum.yz = b;
|
|
frustum.zz = c;
|
|
frustum.wz = -1.0f;
|
|
|
|
frustum.xw = 0.0f;
|
|
frustum.yw = 0.0f;
|
|
frustum.zw = d;
|
|
frustum.ww = 0.0f;
|
|
|
|
cogl_matrix_multiply (matrix, matrix, &frustum);
|
|
#else
|
|
_math_matrix_frustum (matrix, left, right, bottom, top, z_near, z_far);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_perspective (CoglMatrix *matrix,
|
|
float fov_y,
|
|
float aspect,
|
|
float z_near,
|
|
float z_far)
|
|
{
|
|
float ymax = z_near * tan (fov_y * G_PI / 360.0);
|
|
|
|
cogl_matrix_frustum (matrix,
|
|
-ymax * aspect, /* left */
|
|
ymax * aspect, /* right */
|
|
-ymax, /* bottom */
|
|
ymax, /* top */
|
|
z_near,
|
|
z_far);
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_ortho (CoglMatrix *matrix,
|
|
float left,
|
|
float right,
|
|
float bottom,
|
|
float top,
|
|
float near_val,
|
|
float far_val)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
CoglMatrix ortho;
|
|
|
|
/* column 0 */
|
|
ortho.xx = 2.0 / (right - left);
|
|
ortho.yx = 0.0;
|
|
ortho.zx = 0.0;
|
|
ortho.wx = 0.0;
|
|
|
|
/* column 1 */
|
|
ortho.xy = 0.0;
|
|
ortho.yy = 2.0 / (top - bottom);
|
|
ortho.zy = 0.0;
|
|
ortho.wy = 0.0;
|
|
|
|
/* column 2 */
|
|
ortho.xz = 0.0;
|
|
ortho.yz = 0.0;
|
|
ortho.zz = -2.0 / (far_val - near_val);
|
|
ortho.wz = 0.0;
|
|
|
|
/* column 3 */
|
|
ortho.xw = -(right + left) / (right - left);
|
|
ortho.yw = -(top + bottom) / (top - bottom);
|
|
ortho.zw = -(far_val + near_val) / (far_val - near_val);
|
|
ortho.ww = 1.0;
|
|
|
|
cogl_matrix_multiply (matrix, matrix, &ortho);
|
|
#else
|
|
_math_matrix_ortho (matrix, left, right, bottom, top, near_val, far_val);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
void
|
|
cogl_matrix_init_from_array (CoglMatrix *matrix, const float *array)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
memcpy (matrix, array, sizeof (float) * 16);
|
|
#else
|
|
_math_matrix_init_from_array (matrix, array);
|
|
#endif
|
|
_COGL_MATRIX_DEBUG_PRINT (matrix);
|
|
}
|
|
|
|
const float *
|
|
cogl_matrix_get_array (const CoglMatrix *matrix)
|
|
{
|
|
return (float *)matrix;
|
|
}
|
|
|
|
gboolean
|
|
cogl_matrix_get_inverse (const CoglMatrix *matrix, CoglMatrix *inverse)
|
|
{
|
|
#ifndef USE_MESA_MATRIX_API
|
|
#warning "cogl_matrix_get_inverse not supported without Mesa matrix API"
|
|
cogl_matrix_init_identity (inverse);
|
|
return FALSE;
|
|
#else
|
|
if (_math_matrix_update_inverse ((CoglMatrix *)matrix))
|
|
{
|
|
cogl_matrix_init_from_array (inverse, matrix->inv);
|
|
return TRUE;
|
|
}
|
|
else
|
|
{
|
|
cogl_matrix_init_identity (inverse);
|
|
return FALSE;
|
|
}
|
|
#endif
|
|
}
|
|
|
|
void
|
|
cogl_matrix_transform_point (const CoglMatrix *matrix,
|
|
float *x,
|
|
float *y,
|
|
float *z,
|
|
float *w)
|
|
{
|
|
float _x = *x, _y = *y, _z = *z, _w = *w;
|
|
|
|
*x = matrix->xx * _x + matrix->xy * _y + matrix->xz * _z + matrix->xw * _w;
|
|
*y = matrix->yx * _x + matrix->yy * _y + matrix->yz * _z + matrix->yw * _w;
|
|
*z = matrix->zx * _x + matrix->zy * _y + matrix->zz * _z + matrix->zw * _w;
|
|
*w = matrix->wx * _x + matrix->wy * _y + matrix->wz * _z + matrix->ww * _w;
|
|
}
|
|
|
|
|