matrix: make accurate inv version default

* now the glm_mat4_inv_precise is deault, because I don't think all people are aware of this func
* the old behavior (fast) replaced with new func: glm_mat4_inv_fast

if fast version is desired then glm_mat4_inv_fast must be used.

include/cglm/mat4.h

@@ -36,7 +36,7 @@
    CGLM_INLINE void  glm_mat4_scale(mat4 m, float s);
    CGLM_INLINE float glm_mat4_det(mat4 mat);
    CGLM_INLINE void  glm_mat4_inv(mat4 mat, mat4 dest);
-   CGLM_INLINE void  glm_mat4_inv_precise(mat4 mat, mat4 dest);
+   CGLM_INLINE void  glm_mat4_inv_fast(mat4 mat, mat4 dest);
    CGLM_INLINE void  glm_mat4_swap_col(mat4 mat, int col1, int col2);
    CGLM_INLINE void  glm_mat4_swap_row(mat4 mat, int row1, int row2);
  */
@@ -78,6 +78,9 @@
 #define glm_mat4_udup(mat, dest) glm_mat4_ucopy(mat, dest)
 #define glm_mat4_dup(mat, dest)  glm_mat4_copy(mat, dest)
 
+/* DEPRECATED! default is precise now. */
+#define glm_mat4_inv_precise(mat, dest) glm_mat4_inv(mat, dest)
+
 /*!
  * @brief copy all members of [mat] to [dest]
  *
@@ -448,10 +451,6 @@ glm_mat4_det(mat4 mat) {
 /*!
  * @brief inverse mat4 and store in dest
  *
- * this func uses reciprocal approximation without extra corrections 
- * e.g Newton-Raphson. this should work faster than _precise, 
- * to get precise value use _precise version
- *
  * @param[in]  mat  matrix
  * @param[out] dest inverse matrix
  */
@@ -504,22 +503,23 @@ glm_mat4_inv(mat4 mat, mat4 dest) {
 #endif
 }
 
-
 /*!
- * @brief inverse mat4 precisely and store in dest
+ * @brief inverse mat4 and store in dest
  *
- * this do same thing as glm_mat4_inv did. the only diff is this func uses
- * division instead of reciprocal approximation. Due to division this might
- * work slower than glm_mat4_inv
+ * this func uses reciprocal approximation without extra corrections
+ * e.g Newton-Raphson. this should work faster than normal,
+ * to get more precise use glm_mat4_inv version.
+ *
+ * NOTE: You will lose precision, glm_mat4_inv is more accurate
  *
  * @param[in]  mat  matrix
  * @param[out] dest inverse matrix
  */
 CGLM_INLINE
 void
-glm_mat4_inv_precise(mat4 mat, mat4 dest) {
+glm_mat4_inv_fast(mat4 mat, mat4 dest) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  glm_mat4_inv_precise_sse2(mat, dest);
+  glm_mat4_inv_fast_sse2(mat, dest);
 #else
   glm_mat4_inv(mat, dest);
 #endif

include/cglm/simd/sse2/mat4.h

@@ -12,6 +12,8 @@
 #include "../../common.h"
 #include "../intrin.h"
 
+#define glm_mat4_inv_precise_sse2(mat, dest) glm_mat4_inv_sse2(mat, dest)
+
 CGLM_INLINE
 void
 glm_mat4_scale_sse2(mat4 m, float s){
@@ -157,7 +159,7 @@ glm_mat4_det_sse2(mat4 mat) {
 
 CGLM_INLINE
 void
-glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
+glm_mat4_inv_fast_sse2(mat4 mat, mat4 dest) {
   __m128 r0, r1, r2, r3,
          v0, v1, v2, v3,
          t0, t1, t2, t3, t4, t5,
@@ -281,7 +283,7 @@ glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
 
 CGLM_INLINE
 void
-glm_mat4_inv_precise_sse2(mat4 mat, mat4 dest) {
+glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
   __m128 r0, r1, r2, r3,
          v0, v1, v2, v3,
          t0, t1, t2, t3, t4, t5,

include/cglm/vec3.h

@@ -226,7 +226,7 @@ glm_vec_inv(vec3 v) {
 /*!
  * @brief inverse/opposite vector
  *
- * @param[in]  vec  source
+ * @param[in]  v    source
  * @param[out] dest destination
  */
 CGLM_INLINE

include/cglm/vec4.h

@@ -251,7 +251,7 @@ glm_vec4_inv(vec4 v) {
 /*!
  * @brief inverse/opposite vector
  *
- * @param[in]  vec  source
+ * @param[in]  v    source
  * @param[out] dest destination
  */
 CGLM_INLINE

test/src/test_mat4.c

@@ -70,8 +70,8 @@ test_mat4(void **state) {
     test_rand_mat4(m3);
     test_rand_mat4(m4);
     
-    glm_mat4_inv(m3, m4);
-    glm_mat4_inv(m4, m5);
+    glm_mat4_inv_fast(m3, m4);
+    glm_mat4_inv_fast(m4, m5);
     test_assert_mat4_eq2(m3, m5, 0.0009f);
     
     test_rand_mat4(m3);