style cleanup: issues missed last commit

This commit is contained in:
2012-03-25 15:56:17 +00:00
parent fffe342d87
commit 3f56ee3da7
7 changed files with 70 additions and 66 deletions

View File

@@ -245,9 +245,9 @@ void mul_m4_m3m4(float (*m1)[4], float (*m3)[3], float (*m2)[4])
}
void mul_serie_m3(float answ[][3],
float m1[][3], float m2[][3], float m3[][3],
float m4[][3], float m5[][3], float m6[][3],
float m7[][3], float m8[][3])
float m1[][3], float m2[][3], float m3[][3],
float m4[][3], float m5[][3], float m6[][3],
float m7[][3], float m8[][3])
{
float temp[3][3];
@@ -278,9 +278,9 @@ float m7[][3], float m8[][3])
}
void mul_serie_m4(float answ[][4], float m1[][4],
float m2[][4], float m3[][4], float m4[][4],
float m5[][4], float m6[][4], float m7[][4],
float m8[][4])
float m2[][4], float m3[][4], float m4[][4],
float m5[][4], float m6[][4], float m7[][4],
float m8[][4])
{
float temp[4][4];
@@ -542,9 +542,9 @@ int invert_m4(float m[4][4])
/*
* invertmat -
* computes the inverse of mat and puts it in inverse. Returns
* TRUE on success (i.e. can always find a pivot) and FALSE on failure.
* Uses Gaussian Elimination with partial (maximal column) pivoting.
* computes the inverse of mat and puts it in inverse. Returns
* TRUE on success (i.e. can always find a pivot) and FALSE on failure.
* Uses Gaussian Elimination with partial (maximal column) pivoting.
*
* Mark Segal - 1992
*/
@@ -589,7 +589,7 @@ int invert_m4_m4(float inverse[4][4], float mat[4][4])
temp = tempmat[i][i];
if (temp == 0)
return 0; /* No non-zero pivot */
return 0; /* No non-zero pivot */
for (k = 0; k < 4; k++) {
tempmat[i][k] = (float)(tempmat[i][k] / temp);
inverse[i][k] = (float)(inverse[i][k] / temp);
@@ -1568,7 +1568,7 @@ void svd_m4(float U[4][4], float s[4], float V[4][4], float A_[4][4])
break;
}
t = (ks != p ? fabsf(e[ks]) : 0.f) +
(ks != k + 1 ? fabsf(e[ks - 1]) : 0.0f);
(ks != k + 1 ? fabsf(e[ks - 1]) : 0.0f);
if (fabsf(s[ks]) <= eps * t) {
s[ks] = 0.0f;
break;
@@ -1617,7 +1617,7 @@ void svd_m4(float U[4][4], float s[4], float V[4][4], float A_[4][4])
break;
}
// Split at negligible s(k).
// Split at negligible s(k).
case 2:
{
@@ -1641,7 +1641,7 @@ void svd_m4(float U[4][4], float s[4], float V[4][4], float A_[4][4])
break;
}
// Perform one qr step.
// Perform one qr step.
case 3:
{
@@ -1713,7 +1713,7 @@ void svd_m4(float U[4][4], float s[4], float V[4][4], float A_[4][4])
iter = iter + 1;
break;
}
// Convergence.
// Convergence.
case 4:
{