Merge pull request #56 from recp/sphere

aabb and sphere intersect functions

.gitignore

@@ -60,3 +60,12 @@ cglm_test_iosTests/*
 docs/build/*
 win/cglm_test_*
 * copy.*
+*.o
+*.obj
+*codeanalysis.*.xml
+*codeanalysis.xml
+*.lib
+*.tlog
+win/x64
+win/x85
+win/Debug

.travis.yml

@@ -49,7 +49,7 @@ script:
 
 after_success:
   - if [[ "$CC" == "gcc" && "$CODE_COVERAGE" == "ON" ]]; then
-      pip install --user cpp-coveralls
+      pip install --user cpp-coveralls && 
       coveralls
         --build-root .
         --exclude lib

CREDITS

@@ -1,7 +1,7 @@
 This library [initially] used some [piece of] implementations
 (may include codes) from these open source projects/resources:
 
-1. Affine Transforms
+1. Initial Affine Transforms
 The original glm repo (g-truc), url: https://github.com/g-truc/glm
 
 LICENSE[S]:
@@ -11,7 +11,7 @@ LICENSE[S]:
 
 FULL LICENSE: https://github.com/g-truc/glm/blob/master/copying.txt
 
-2. Quaternions
+2. Initial Quaternions
 Anton's OpenGL 4 Tutorials book source code:
 
 LICENSE:
@@ -43,3 +43,12 @@ https://github.com/erich666/GraphicsGems/blob/master/gems/TransBox.c
 6. Cull frustum
 http://www.txutxi.com/?p=584
 http://old.cescg.org/CESCG-2002/DSykoraJJelinek/
+
+7. Quaternions
+Initial mat4_quat is borrowed from Apple's simd library
+
+8. Vector Rotation using Quaternion
+https://gamedev.stackexchange.com/questions/28395/rotating-vector3-by-a-quaternion
+
+9. Sphere AABB intersect
+https://github.com/erich666/GraphicsGems/blob/master/gems/BoxSphere.c

README.md

@@ -1,9 +1,11 @@
 # 🎥 OpenGL Mathematics (glm) for `C`
 [![Build Status](https://travis-ci.org/recp/cglm.svg?branch=master)](https://travis-ci.org/recp/cglm)
-[![Build status](https://ci.appveyor.com/api/projects/status/av7l3gc0yhfex8y4/branch/master?svg=true)](https://ci.appveyor.com/project/recp/cglm/branch/master)
+ [![Build status](https://ci.appveyor.com/api/projects/status/av7l3gc0yhfex8y4/branch/master?svg=true)](https://ci.appveyor.com/project/recp/cglm/branch/master)
 [![Documentation Status](https://readthedocs.org/projects/cglm/badge/?version=latest)](http://cglm.readthedocs.io/en/latest/?badge=latest)
 [![Coverage Status](https://coveralls.io/repos/github/recp/cglm/badge.svg?branch=master)](https://coveralls.io/github/recp/cglm?branch=master)
 [![Codacy Badge](https://api.codacy.com/project/badge/Grade/6a62b37d5f214f178ebef269dc4a6bf1)](https://www.codacy.com/app/recp/cglm?utm_source=github.com&utm_medium=referral&utm_content=recp/cglm&utm_campaign=Badge_Grade)
+[![Backers on Open Collective](https://opencollective.com/cglm/backers/badge.svg)](#backers) 
+[![Sponsors on Open Collective](https://opencollective.com/cglm/sponsors/badge.svg)](#sponsors)
 
 The original glm library is for C++ only (templates, namespaces, classes...), this library targeted to C99 but currently you can use it for C89 safely by language extensions e.g `__restrict`
 
@@ -17,6 +19,11 @@ Complete documentation: http://cglm.readthedocs.io
 - _dup (duplicate) is changed to _copy. For instance `glm_vec_dup -> glm_vec_copy`
 - OpenGL related functions are dropped to make this lib platform/third-party independent
 - make sure you have latest version and feel free to report bugs, troubles
+- **[bugfix]** euler angles was implemented in reverse order (extrinsic) it was fixed, now they are intrinsic. Make sure that 
+you have the latest version
+- **[major change]** by starting v0.4.0, quaternions are stored as [x, y, z, w], it was [w, x, y, z] in v0.3.5 and earlier versions
+- **[api rename]** by starting v0.4.5, **glm_simd** functions are renamed to **glmm_**  
+- **[new option]** by starting v0.4.5, you can disable alignment requirement, check options in docs.  
 
 #### Note for C++ developers:
 If you don't aware about original GLM library yet, you may also want to look at:
@@ -73,6 +80,7 @@ Currently *cglm* uses default clip space configuration (-1, 1) for camera functi
 - inline or pre-compiled function call
 - frustum (extract view frustum planes, corners...)
 - bounding box  (AABB in Frustum (culling), crop, merge...)
+- project, unproject
 
 <hr />
 
@@ -114,6 +122,36 @@ glm_mul(T, R, modelMat);
 glm_inv_tr(modelMat);
 ```
 
+## Contributors
+
+This project exists thanks to all the people who contribute. [[Contribute](CONTRIBUTING.md)].
+<a href="graphs/contributors"><img src="https://opencollective.com/cglm/contributors.svg?width=890&button=false" /></a>
+
+
+## Backers
+
+Thank you to all our backers! 🙏 [[Become a backer](https://opencollective.com/cglm#backer)]
+
+<a href="https://opencollective.com/cglm#backers" target="_blank"><img src="https://opencollective.com/cglm/backers.svg?width=890"></a>
+
+
+## Sponsors
+
+Support this project by becoming a sponsor. Your logo will show up here with a link to your website. [[Become a sponsor](https://opencollective.com/cglm#sponsor)]
+
+<a href="https://opencollective.com/cglm/sponsor/0/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/0/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/1/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/1/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/2/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/2/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/3/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/3/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/4/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/4/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/5/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/5/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/6/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/6/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/7/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/7/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/8/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/8/avatar.svg"></a>
+<a href="https://opencollective.com/cglm/sponsor/9/website" target="_blank"><img src="https://opencollective.com/cglm/sponsor/9/avatar.svg"></a>
+
+
+
 ## License
 MIT. check the LICENSE file
 
@@ -161,7 +199,7 @@ If you want to use inline versions of funcstions then; include main header
 ```C
 #include <cglm/cglm.h>
 ```
-the haeder will include all headers. Then call func you want e.g. rotate vector by axis:
+the header will include all headers. Then call func you want e.g. rotate vector by axis:
 ```C
 glm_vec_rotate(v1, glm_rad(45), (vec3){1.0f, 0.0f, 0.0f});
 ```
@@ -180,7 +218,7 @@ to call pre-compiled versions include header with `c` postfix, c means call. Pre
 ```C
 #include <cglm/call.h>
 ```
-this header will include all heaers with c postfix. You need to call functions with c posfix:
+this header will include all headers with c postfix. You need to call functions with c posfix:
 ```C
 glmc_vec_normalize(vec);
 ```
@@ -194,6 +232,27 @@ glm_mat4_mul(m1, m1, m1);
 ```
 the first two parameter are **[in]** and the last one is **[out]** parameter. After multiplied *m1* and *m2* the result is stored in *m1*. This is why we send *m1* twice. You may store result in different matrix, this just an example.
 
+### Example: Computing MVP matrix
+
+#### Option 1
+```C
+mat4 proj, view, model, mvp;
+
+/* init proj, view and model ... */
+
+glm_mat4_mul(proj, view, viewProj);
+glm_mat4_mul(viewProj, model, mvp);
+```
+
+#### Option 2
+```C
+mat4 proj, view, model, mvp;
+
+/* init proj, view and model ... */
+
+glm_mat4_mulN((mat4 *[]){&proj, &view, &model}, 3, mvp);
+```
+
 ## How to send matrix to OpenGL
 
 mat4 is array of vec4 and vec4 is array of floats. `glUniformMatrix4fv` functions accecpts `float*` as `value` (last param), so you can cast mat4 to float* or you can pass first column of matrix as beginning of memory of matrix:

autogen.sh

@@ -8,17 +8,14 @@
 
 cd $(dirname "$0")
 
-if [ "$(uname)" = "Darwin" ]; then
-libtoolBin=$(which glibtoolize)
-libtoolBinDir=$(dirname "${libtoolBin}")
-
-if [ ! -f "${libtoolBinDir}/libtoolize" ]; then
-ln -s $libtoolBin "${libtoolBinDir}/libtoolize"
-fi
-fi
-
 autoheader
-libtoolize
+
+if [ "$(uname)" = "Darwin" ]; then
+  glibtoolize
+else
+  libtoolize
+fi
+
 aclocal -I m4
 autoconf
 automake --add-missing --copy

build-deps.sh

@@ -9,21 +9,14 @@
 # check if deps are pulled
 git submodule update --init --recursive
 
-# fix glibtoolize
-
 cd $(dirname "$0")
 
-if [ "$(uname)" = "Darwin" ]; then
-  libtoolBin=$(which glibtoolize)
-  libtoolBinDir=$(dirname "${libtoolBin}")
-  ln -s $libtoolBin "${libtoolBinDir}/libtoolize"
-fi
-
 # general deps: gcc make autoconf automake libtool cmake
 
 # test - cmocka
 cd ./test/lib/cmocka
-mkdir build
+rm -rf build
+mkdir -p build
 cd build
 cmake -DCMAKE_INSTALL_PREFIX=/usr -DCMAKE_BUILD_TYPE=Debug ..
 make -j8

cglm.podspec

@@ -0,0 +1,28 @@
+Pod::Spec.new do |s|
+
+  # Description
+  s.name         = "cglm"
+  s.version      = "0.4.6"
+  s.summary      = "📽 Optimized OpenGL/Graphics Math (glm) for C"
+  s.description  = <<-DESC
+cglm is math library for graphics programming for C. It is similar to original glm but it is written for C instead of C++ (you can use here too). See the documentation or README for all features.
+                   DESC
+
+  s.documentation_url = "http://cglm.readthedocs.io"
+
+  # Home
+  s.homepage     = "https://github.com/recp/cglm"
+  s.license      = { :type => "MIT", :file => "LICENSE" }
+  s.author       = { "Recep Aslantas" => "recp@acm.org" }
+
+  # Sources
+  s.source               = { :git => "https://github.com/recp/cglm.git", :tag => "v#{s.version}" }
+  s.source_files         = "src", "include/cglm/**/*.h"
+  s.public_header_files  = "include", "include/cglm/**/*.h"
+  s.exclude_files        = "src/win/*", "src/dllmain.c", "src/**/*.h"
+  s.preserve_paths       = "include", "src"
+  s.header_mappings_dir  = "include"
+
+  # Linking
+  s.library = "m"
+end

configure.ac

@@ -7,7 +7,7 @@
 #*****************************************************************************
 
 AC_PREREQ([2.69])
-AC_INIT([cglm], [0.3.5], [info@recp.me])
+AC_INIT([cglm], [0.4.7], [info@recp.me])
 AM_INIT_AUTOMAKE([-Wall -Werror foreign subdir-objects])
 
 AC_CONFIG_MACRO_DIR([m4])

docs/source/affine-mat.rst

@@ -33,6 +33,7 @@ Table of contents (click func go):
 Functions:
 
 1. :c:func:`glm_mul`
+#. :c:func:`glm_mul_rot`
 #. :c:func:`glm_inv_tr`
 
 Functions documentation
@@ -59,6 +60,27 @@ Functions documentation
       | *[in]*  **m2**    affine matrix 2
       | *[out]* **dest**  result matrix
 
+.. c:function:: void  glm_mul_rot(mat4 m1, mat4 m2, mat4 dest)
+
+    | this is similar to glm_mat4_mul but specialized to rotation matrix
+
+    Right Matrix format should be (left is free):
+
+    .. code-block:: text
+
+       R  R  R  0
+       R  R  R  0
+       R  R  R  0
+       0  0  0  1
+
+    this reduces some multiplications. It should be faster than mat4_mul.
+    if you are not sure about matrix format then DON'T use this! use mat4_mul
+
+    Parameters:
+      | *[in]*  **m1**    affine matrix 1
+      | *[in]*  **m2**    affine matrix 2
+      | *[out]* **dest**  result matrix
+
 .. c:function:: void  glm_inv_tr(mat4 mat)
 
     | inverse orthonormal rotation + translation matrix (ridig-body)

docs/source/affine.rst

@@ -5,6 +5,8 @@ affine transforms
 
 Header: cglm/affine.h
 
+Initialize Transform Matrices
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Functions with **_make** prefix expect you don't have a matrix and they create
 a matrix for you. You don't need to pass identity matrix.
 
@@ -15,6 +17,107 @@ before sending to transfrom functions.
 There are also functions to decompose transform matrix. These functions can't
 decompose matrix after projected.
 
+Rotation Center
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Rotating functions uses origin as rotation center (pivot/anchor point),
+since scale factors are stored in rotation matrix, same may also true for scalling.
+cglm provides some functions for rotating around at given point e.g.
+**glm_rotate_at**, **glm_quat_rotate_at**. Use them or follow next section for algorihm ("Rotate or Scale around specific Point (Pivot Point / Anchor Point)").
+
+Rotate or Scale around specific Point (Anchor Point)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+If you want to rotate model around arbibtrary point follow these steps:
+
+1. Move model from pivot point to origin: **translate(-pivot.x, -pivot.y, -pivot.z)**
+2. Apply rotation (or scaling maybe)
+3. Move model back from origin to pivot (reverse of step-1): **translate(pivot.x, pivot.y, pivot.z)**
+
+**glm_rotate_at**, **glm_quat_rotate_at** and their helper functions works that way.
+
+The implementation would be:
+
+.. code-block:: c
+  :linenos:
+
+  glm_translate(m, pivot);
+  glm_rotate(m, angle, axis);
+  glm_translate(m, pivotInv); /* pivotInv = -pivot */
+
+Transforms Order
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+It is important to understand this part especially if you call transform
+functions multiple times
+
+`glm_translate`, `glm_rotate`, `glm_scale` and `glm_quat_rotate` and their
+helpers functions works like this (cglm may provide reverse order too as alternative in the future):
+
+.. code-block:: c
+  :linenos:
+
+  TransformMatrix = TransformMatrix * TraslateMatrix; // glm_translate()
+  TransformMatrix = TransformMatrix * RotateMatrix;   // glm_rotate(), glm_quat_rotate()
+  TransformMatrix = TransformMatrix * ScaleMatrix;    // glm_scale()
+
+As you can see it is multipled as right matrix. For instance what will happen if you call `glm_translate` twice?
+
+.. code-block:: c
+  :linenos:
+
+  glm_translate(transform, translate1); /* transform = transform * translate1 */
+  glm_translate(transform, translate2); /* transform = transform * translate2 */
+  glm_rotate(transform, angle, axis)    /* transform = transform * rotation   */
+
+Now lets try to understand this:
+
+1. You call translate using `translate1` and you expect it will be first transform
+because you call it first, do you?
+
+Result will be **`transform = transform * translate1`**
+
+2. Then you call translate using `translate2` and you expect it will be second transform?
+
+Result will be **`transform = transform * translate2`**. Now lets expand transform,
+it was `transform * translate1` before second call.
+
+Now it is **`transform = transform * translate1 * translate2`**, now do you understand what I say?
+
+3. After last call transform will be:
+
+**`transform = transform * translate1 * translate2 * rotation`**
+
+The order will be; **rotation will be applied first**, then **translate2** then **translate1**
+
+It is all about matrix multiplication order. It is similar to MVP matrix:
+`MVP = Projection * View * Model`, model will be applied first, then view then projection.
+
+**Confused?**
+
+In the end the last function call applied first in shaders.
+
+As alternative way, you can create transform matrices individually then combine manually,
+but don't forget that `glm_translate`, `glm_rotate`, `glm_scale`... are optimized and should be faster (an smaller assembly output) than manual multiplication
+
+.. code-block:: c
+  :linenos:
+
+  mat4 transform1, transform2, transform3, finalTransform;
+
+  glm_translate_make(transform1, translate1);
+  glm_translate_make(transform2, translate2);
+  glm_rotate_make(transform3, angle, axis);
+
+  /* first apply transform1, then transform2, thentransform3 */
+  glm_mat4_mulN((mat4 *[]){&transform3, &transform2, &transform1}, 3, finalTransform);
+
+  /* if you don't want to use mulN, same as above */
+  glm_mat4_mul(transform3, transform2, finalTransform);
+  glm_mat4_mul(finalTransform, transform1, finalTransform);
+
+Now transform1 will be applied first, then transform2 then transform3
+
 Table of contents (click to go):
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
@@ -29,15 +132,14 @@ Functions:
 #. :c:func:`glm_scale_to`
 #. :c:func:`glm_scale_make`
 #. :c:func:`glm_scale`
-#. :c:func:`glm_scale1`
 #. :c:func:`glm_scale_uni`
 #. :c:func:`glm_rotate_x`
 #. :c:func:`glm_rotate_y`
 #. :c:func:`glm_rotate_z`
-#. :c:func:`glm_rotate_ndc_make`
 #. :c:func:`glm_rotate_make`
-#. :c:func:`glm_rotate_ndc`
 #. :c:func:`glm_rotate`
+#. :c:func:`glm_rotate_at`
+#. :c:func:`glm_rotate_atm`
 #. :c:func:`glm_decompose_scalev`
 #. :c:func:`glm_uniscaled`
 #. :c:func:`glm_decompose_rs`
@@ -122,10 +224,6 @@ Functions documentation
       | *[in, out]* **m** affine transfrom
       | *[in]*      **v** scale vector [x, y, z]
 
-.. c:function:: void  glm_scale1(mat4 m, float s)
-
-    DEPRECATED! Use glm_scale_uni
-
 .. c:function:: void  glm_scale_uni(mat4 m, float s)
 
     applies uniform scale to existing transform matrix v = [s, s, s]
@@ -165,16 +263,6 @@ Functions documentation
       | *[in]*  **angle** angle (radians)
       | *[out]* **dest**  rotated matrix
 
-.. c:function:: void  glm_rotate_ndc_make(mat4 m, float angle, vec3 axis_ndc)
-
-    creates NEW rotation matrix by angle and axis
-    this name may change in the future. axis must be is normalized
-
-    Parameters:
-      | *[out]* **m**        affine transfrom
-      | *[in]*  **angle**    angle (radians)
-      | *[in]*  **axis_ndc** normalized axis
-
 .. c:function:: void  glm_rotate_make(mat4 m, float angle, vec3 axis)
 
     creates NEW rotation matrix by angle and axis,
@@ -185,16 +273,6 @@ Functions documentation
       | *[in]*  **axis** angle (radians)
       | *[in]*  **axis** axis
 
-.. c:function:: void  glm_rotate_ndc(mat4 m, float angle, vec3 axis_ndc)
-
-    rotate existing transform matrix around Z axis by angle and axis
-    this name may change in the future, axis must be normalized.
-
-    Parameters:
-      | *[out]* **m**        affine transfrom
-      | *[in]*  **angle**    angle (radians)
-      | *[in]*  **axis_ndc** normalized axis
-
 .. c:function:: void  glm_rotate(mat4 m, float angle, vec3 axis)
 
     rotate existing transform matrix around Z axis by angle and axis
@@ -204,6 +282,29 @@ Functions documentation
       | *[in]*      **angle** angle (radians)
       | *[in]*      **axis**  axis
 
+.. c:function:: void  glm_rotate_at(mat4 m, vec3 pivot, float angle, vec3 axis)
+
+    rotate existing transform around given axis by angle at given pivot point (rotation center)
+
+    Parameters:
+      | *[in, out]* **m**     affine transfrom
+      | *[in]*      **pivot** pivot, anchor point, rotation center
+      | *[in]*      **angle** angle (radians)
+      | *[in]*      **axis**  axis
+
+.. c:function:: void  glm_rotate_atm(mat4 m, vec3 pivot, float angle, vec3 axis)
+
+    | creates NEW rotation matrix by angle and axis at given point
+    | this creates rotation matrix, it assumes you don't have a matrix
+
+    | this should work faster than glm_rotate_at because it reduces one glm_translate.
+
+    Parameters:
+      | *[in, out]* **m**     affine transfrom
+      | *[in]*      **pivot** pivot, anchor point, rotation center
+      | *[in]*      **angle** angle (radians)
+      | *[in]*      **axis**  axis
+
 .. c:function:: void  glm_decompose_scalev(mat4 m, vec3 s)
 
     decompose scale vector

docs/source/api.rst

@@ -41,6 +41,8 @@ Follow the :doc:`build` documentation for this
    vec4-ext
    color
    plane
+   project
    util
    io
    call
+   sphere

docs/source/box.rst

@@ -24,6 +24,15 @@ Functions:
 #. :c:func:`glm_aabb_crop`
 #. :c:func:`glm_aabb_crop_until`
 #. :c:func:`glm_aabb_frustum`
+#. :c:func:`glm_aabb_invalidate`
+#. :c:func:`glm_aabb_isvalid`
+#. :c:func:`glm_aabb_size`
+#. :c:func:`glm_aabb_radius`
+#. :c:func:`glm_aabb_center`
+#. :c:func:`glm_aabb_aabb`
+#. :c:func:`glm_aabb_sphere`
+#. :c:func:`glm_aabb_point`
+#. :c:func:`glm_aabb_contains`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
@@ -91,3 +100,82 @@ Functions documentation
     Parameters:
       | *[in]*   **box**     bounding box
       | *[out]*  **planes**  frustum planes
+
+.. c:function:: void  glm_aabb_invalidate(vec3 box[2])
+
+    | invalidate AABB min and max values
+
+    | It fills *max* values with -FLT_MAX and *min* values with +FLT_MAX
+
+    Parameters:
+      | *[in, out]*   **box**     bounding box
+
+.. c:function:: bool  glm_aabb_isvalid(vec3 box[2])
+
+    | check if AABB is valid or not
+
+    Parameters:
+      | *[in]*   **box**     bounding box
+
+    Returns:
+      returns true if aabb is valid otherwise false
+
+.. c:function:: float  glm_aabb_size(vec3 box[2])
+
+    | distance between of min and max
+
+    Parameters:
+      | *[in]*   **box**     bounding box
+
+    Returns:
+      distance between min - max
+
+.. c:function:: float  glm_aabb_radius(vec3 box[2])
+
+    | radius of sphere which surrounds AABB
+
+    Parameters:
+      | *[in]*   **box**     bounding box
+
+.. c:function:: void  glm_aabb_center(vec3 box[2], vec3 dest)
+
+    | computes center point of AABB
+
+    Parameters:
+      | *[in]*    **box**      bounding box
+      | *[out]*   **dest**     center of bounding box
+
+.. c:function:: bool  glm_aabb_aabb(vec3 box[2], vec3 other[2])
+
+    | check if two AABB intersects
+
+    Parameters:
+      | *[in]*    **box**     bounding box
+      | *[out]*   **other**   other bounding box
+
+.. c:function:: bool  glm_aabb_sphere(vec3 box[2], vec4 s)
+
+    | check if AABB intersects with sphere
+
+    | https://github.com/erich666/GraphicsGems/blob/master/gems/BoxSphere.c
+    | Solid Box - Solid Sphere test.
+
+    Parameters:
+      | *[in]*    **box**     solid bounding box
+      | *[out]*   **s**       solid sphere
+
+.. c:function:: bool  glm_aabb_point(vec3 box[2], vec3 point)
+
+    | check if point is inside of AABB
+
+    Parameters:
+      | *[in]*    **box**     bounding box
+      | *[out]*   **point**   point
+
+.. c:function:: bool  glm_aabb_contains(vec3 box[2], vec3 other[2])
+
+    | check if AABB contains other AABB
+
+    Parameters:
+      | *[in]*    **box**     bounding box
+      | *[out]*   **other**   other bounding box

docs/source/cam.rst

@@ -167,6 +167,8 @@ Functions documentation
 
     | set up view matrix
 
+    **NOTE:** The UP vector must not be parallel to the line of sight from the eye point to the reference point.
+
     Parameters:
       | *[in]*  **eye**     eye vector
       | *[in]*  **center**  center vector
@@ -181,6 +183,8 @@ Functions documentation
     target self then this might be useful. Because you need to get target
     from direction.
 
+    **NOTE:** The UP vector must not be parallel to the line of sight from the eye point to the reference point.
+
     Parameters:
       | *[in]*  **eye**     eye vector
       | *[in]*  **center**  direction vector

docs/source/conf.py

@@ -62,9 +62,9 @@ author = u'Recep Aslantas'
 # built documents.
 #
 # The short X.Y version.
-version = u'0.3.4'
+version = u'0.4.7'
 # The full version, including alpha/beta/rc tags.
-release = u'0.3.4'
+release = u'0.4.7'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

docs/source/euler.rst

@@ -70,6 +70,7 @@ Functions:
 1. :c:func:`glm_euler_order`
 #. :c:func:`glm_euler_angles`
 #. :c:func:`glm_euler`
+#. :c:func:`glm_euler_xyz`
 #. :c:func:`glm_euler_zyx`
 #. :c:func:`glm_euler_zxy`
 #. :c:func:`glm_euler_xzy`
@@ -115,8 +116,18 @@ Functions documentation
 
     | build rotation matrix from euler angles
 
+    this is alias of glm_euler_xyz function
+
     Parameters:
-      | *[in]*  **angles**  angles as vector [Ex, Ey, Ez]
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
+      | *[in]*  **dest**    rotation matrix
+
+.. c:function:: void  glm_euler_xyz(vec3 angles, mat4 dest)
+
+    | build rotation matrix from euler angles
+
+    Parameters:
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **dest**    rotation matrix
 
 .. c:function:: void  glm_euler_zyx(vec3 angles, mat4 dest)
@@ -124,7 +135,7 @@ Functions documentation
     | build rotation matrix from euler angles
 
     Parameters:
-      | *[in]*  **angles**  angles as vector [Ez, Ey, Ex]
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **dest**    rotation matrix
 
 .. c:function:: void  glm_euler_zxy(vec3 angles, mat4 dest)
@@ -132,7 +143,7 @@ Functions documentation
     | build rotation matrix from euler angles
 
     Parameters:
-      | *[in]*  **angles**  angles as vector [Ez, Ex, Ey]
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **dest**    rotation matrix
 
 .. c:function:: void  glm_euler_xzy(vec3 angles, mat4 dest)
@@ -140,7 +151,7 @@ Functions documentation
     | build rotation matrix from euler angles
 
     Parameters:
-      | *[in]*  **angles**  angles as vector [Ex, Ez, Ey]
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **dest**    rotation matrix
 
 .. c:function:: void  glm_euler_yzx(vec3 angles, mat4 dest)
@@ -148,7 +159,7 @@ Functions documentation
     build rotation matrix from euler angles
 
     Parameters:
-      | *[in]*  **angles**  angles as vector [Ey, Ez, Ex]
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **dest**    rotation matrix
 
 .. c:function:: void  glm_euler_yxz(vec3 angles, mat4 dest)
@@ -156,7 +167,7 @@ Functions documentation
     | build rotation matrix from euler angles
 
     Parameters:
-      | *[in]*  **angles**  angles as vector [Ey, Ex, Ez]
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **dest**    rotation matrix
 
 .. c:function:: void glm_euler_by_order(vec3 angles, glm_euler_sq ord, mat4 dest)
@@ -166,6 +177,6 @@ Functions documentation
     Use :c:func:`glm_euler_order` function to build *ord* parameter
 
     Parameters:
-      | *[in]*  **angles**  angles as vector (ord parameter spceifies angles order)
+      | *[in]*  **angles**  angles as vector [Xangle, Yangle, Zangle]
       | *[in]*  **ord**     euler order
       | *[in]*  **dest**    rotation matrix

docs/source/getting_started.rst

@@ -21,17 +21,24 @@ Types:
 As you can see types don't store extra informations in favor of space.
 You can send these values e.g. matrix to OpenGL directly without casting or calling a function like *value_ptr*
 
-Aligment is Required:
+Alignment is Required:
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-**vec4** and **mat4** requires 16 byte aligment because vec4 and mat4 operations are
+**vec4** and **mat4** requires 16 byte alignment because vec4 and mat4 operations are
 vectorized by SIMD instructions (SSE/AVX).
 
+**UPDATE:**
+  By starting v0.4.5 cglm provides an option to disable alignment requirement, it is enabled as default
+
+  | Check :doc:`opt` page for more details
+
+  Also alignment is disabled for older msvc verisons as default. Now alignment is only required in Visual Studio 2017 version 15.6+ if CGLM_ALL_UNALIGNED macro is not defined.
+
 Allocations:
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 *cglm* doesn't alloc any memory on heap. So it doesn't provide any allocator.
 You must allocate memory yourself. You should alloc memory for out parameters too if you pass pointer of memory location.
-When allocating memory don't forget that **vec4** and **mat4** requires aligment.
+When allocating memory don't forget that **vec4** and **mat4** requires alignment.
 
 **NOTE:** Unaligned vec4 and unaligned mat4 operations will be supported in the future. Check todo list.
 Because you may want to multiply a CGLM matrix with external matrix.

docs/source/index.rst

@@ -40,6 +40,8 @@ Also currently only **float** type is supported for most operations.
    getting_started
    opengl
    api
+   opt
+   troubleshooting
 
 Indices and tables
 ==================

docs/source/io.rst

@@ -39,6 +39,7 @@ Functions:
 #. :c:func:`glm_vec3_print`
 #. :c:func:`glm_ivec3_print`
 #. :c:func:`glm_versor_print`
+#. :c:func:`glm_aabb_print`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
@@ -90,3 +91,12 @@ Functions documentation
     Parameters:
       | *[in]*  **vec**      quaternion
       | *[in]*  **ostream**  FILE to write
+
+.. c:function:: void  glm_aabb_print(versor vec, const char * __restrict tag, FILE * __restrict ostream)
+
+    | print aabb to given stream
+
+    Parameters:
+      | *[in]*  **vec**      aabb (axis-aligned bounding box)
+      | *[in]*  **tag**      tag to find it more easly in logs
+      | *[in]*  **ostream**  FILE to write

docs/source/mat3.rst

@@ -24,6 +24,7 @@ Functions:
 #. :c:func:`glm_mat3_transpose_to`
 #. :c:func:`glm_mat3_transpose`
 #. :c:func:`glm_mat3_mulv`
+#. :c:func:`glm_mat3_quat`
 #. :c:func:`glm_mat3_scale`
 #. :c:func:`glm_mat3_det`
 #. :c:func:`glm_mat3_inv`
@@ -89,6 +90,14 @@ Functions documentation
       | *[in]*  **v**     vec3 (right, column vector)
       | *[out]* **dest**  destination (result, column vector)
 
+.. c:function:: void  glm_mat3_quat(mat3 m, versor dest)
+
+    convert mat3 to quaternion
+
+    Parameters:
+      | *[in]*  **m**     rotation matrix
+      | *[out]* **dest**  destination quaternion
+
 .. c:function:: void  glm_mat3_scale(mat3 m, float s)
 
     multiply matrix with scalar

docs/source/mat4.rst

@@ -32,6 +32,7 @@ Functions:
 #. :c:func:`glm_mat4_mulN`
 #. :c:func:`glm_mat4_mulv`
 #. :c:func:`glm_mat4_mulv3`
+#. :c:func:`glm_mat4_quat`
 #. :c:func:`glm_mat4_transpose_to`
 #. :c:func:`glm_mat4_transpose`
 #. :c:func:`glm_mat4_scale_p`
@@ -146,6 +147,14 @@ Functions documentation
     | *[in]*  **v**     vec3 (right, column vector)
     | *[out]* **dest**  vec3 (result, column vector)
 
+.. c:function:: void  glm_mat4_quat(mat4 m, versor dest)
+
+    convert mat4's rotation part to quaternion
+
+    Parameters:
+    | *[in]*  **m**     affine matrix
+    | *[out]* **dest**  destination quaternion
+
 .. c:function:: void  glm_mat4_transpose_to(mat4 m, mat4 dest)
 
     transpose mat4 and store in dest

docs/source/opt.rst

@@ -0,0 +1,42 @@
+.. default-domain:: C
+
+Options
+===============================================================================
+
+A few options are provided via macros.
+
+Alignment Option
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+As default, cglm requires types to be aligned. Alignment requirements:
+
+vec3:   8 byte
+vec4:   16 byte
+mat4:   16 byte
+versor: 16 byte
+
+By starting **v0.4.5** cglm provides an option to disable alignment requirement.
+To enable this option define **CGLM_ALL_UNALIGNED** macro before all headers.
+You can define it in Xcode, Visual Studio (or other IDEs) or you can also prefer
+to define it in build system. If you use pre-compiled verisons then you
+have to compile cglm with **CGLM_ALL_UNALIGNED** macro.
+
+**VERY VERY IMPORTANT:** If you use cglm in multiple projects and
+ those projects are depends on each other, then
+
+ | *ALWAYS* or *NEVER USE* **CGLM_ALL_UNALIGNED** macro in linked projects
+
+ if you do not know what you are doing. Because a cglm header included
+ via 'project A' may force types to be aligned and another cglm header
+ included via 'project B' may not require alignment. In this case
+ cglm functions will read from and write to **INVALID MEMORY LOCATIONs**.
+
+ ALWAYS USE SAME CONFIGURATION / OPTION for **cglm** if you have multiple projects.
+
+ For instance if you set CGLM_ALL_UNALIGNED in a project then set it in other projects too
+
+SSE and SSE2 Shuffle Option
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+**_mm_shuffle_ps** generates **shufps** instruction even if registers are same.
+You can force it to generate **pshufd** instruction by defining
+**CGLM_USE_INT_DOMAIN** macro. As default it is not defined.

docs/source/project.rst

@@ -0,0 +1,102 @@
+.. default-domain:: C
+
+Project / UnProject
+================================================================================
+
+Header: cglm/project.h
+
+Viewport is required as *vec4* **[X, Y, Width, Height]** but this doesn't mean
+that you should store it as **vec4**. You can convert your data representation
+to vec4 before passing it to related functions.
+
+Table of contents (click to go):
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Functions:
+
+1. :c:func:`glm_unprojecti`
+#. :c:func:`glm_unproject`
+#. :c:func:`glm_project`
+
+Functions documentation
+~~~~~~~~~~~~~~~~~~~~~~~
+
+.. c:function:: void  glm_unprojecti(vec3 pos, mat4 invMat, vec4 vp, vec3 dest)
+
+    | maps the specified viewport coordinates into specified space [1]
+      the matrix should contain projection matrix.
+
+    if you don't have ( and don't want to have ) an inverse matrix then use
+    glm_unproject version. You may use existing inverse of matrix in somewhere
+    else, this is why glm_unprojecti exists to save save inversion cost
+
+    [1] space:
+      - if m = invProj:     View Space
+      - if m = invViewProj: World Space
+      - if m = invMVP:      Object Space
+
+    You probably want to map the coordinates into object space
+    so use invMVP as m
+
+    Computing viewProj:
+
+    .. code-block:: c
+
+       glm_mat4_mul(proj, view, viewProj);
+       glm_mat4_mul(viewProj, model, MVP);
+       glm_mat4_inv(viewProj, invMVP);
+
+    Parameters:
+      | *[in]*  **pos**     point/position in viewport coordinates
+      | *[in]*  **invMat**  matrix (see brief)
+      | *[in]*  **vp**      viewport as [x, y, width, height]
+      | *[out]* **dest**    unprojected coordinates
+
+.. c:function:: void  glm_unproject(vec3 pos, mat4 m, vec4 vp, vec3 dest)
+
+    | maps the specified viewport coordinates into specified space [1]
+      the matrix should contain projection matrix.
+
+    this is same as glm_unprojecti except this function get inverse matrix for
+    you.
+
+    [1] space:
+      - if m = proj:     View Space
+      - if m = viewProj: World Space
+      - if m = MVP:      Object Space
+
+    You probably want to map the coordinates into object space so use MVP as m
+
+    Computing viewProj and MVP:
+
+    .. code-block:: c
+
+       glm_mat4_mul(proj, view, viewProj);
+       glm_mat4_mul(viewProj, model, MVP);
+
+    Parameters:
+      | *[in]*  **pos**  point/position in viewport coordinates
+      | *[in]*  **m**    matrix (see brief)
+      | *[in]*  **vp**   viewport as [x, y, width, height]
+      | *[out]* **dest** unprojected coordinates
+
+.. c:function:: void  glm_project(vec3 pos, mat4 m, vec4 vp, vec3 dest)
+
+    | map object coordinates to window coordinates
+
+    Computing MVP:
+
+    .. code-block:: c
+
+       glm_mat4_mul(proj, view, viewProj);
+       glm_mat4_mul(viewProj, model, MVP);
+
+    this could be useful for gettng a bbox which fits with view frustum and
+    object bounding boxes. In this case you crop view frustum box with objects
+    box
+
+    Parameters:
+      | *[in]*  **pos**      object coordinates
+      | *[in]*  **m**        MVP matrix
+      | *[in]*  **vp**       viewport as [x, y, width, height]
+      | *[out]* **dest**     projected coordinates

docs/source/quat.rst

@@ -5,17 +5,16 @@ quaternions
 
 Header: cglm/quat.h
 
- **Important:** *cglm* stores quaternion as [w, x, y, z] in memory, don't
- forget that when changing quaternion items manually. For instance *quat[3]*
- is *quat.z* and *quat[0*] is *quat.w*. This may change in the future if *cglm*
- will got enough request to do that. Probably it will not be changed in near
- future
+ **Important:** *cglm* stores quaternion as **[x, y, z, w]** in memory
+ since **v0.4.0** it was **[w, x, y, z]**
+ before v0.4.0 ( **v0.3.5 and earlier** ). w is real part.
 
-There are some TODOs for quaternions check TODO list to see them.
+What you can do with quaternions with existing functions is (Some of them):
 
-Also **versor** is identity quaternion so the type may change to **vec4** or
-something else. This will not affect existing functions for your engine because
-*versor* is alias of *vec4*
+- You can rotate transform matrix using quaterion
+- You can rotate vector using quaterion
+- You can create view matrix using quaterion
+- You can create a lookrotation (from source point to dest)
 
 Table of contents (click to go):
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -28,14 +27,38 @@ Macros:
 Functions:
 
 1. :c:func:`glm_quat_identity`
+#. :c:func:`glm_quat_init`
 #. :c:func:`glm_quat`
 #. :c:func:`glm_quatv`
+#. :c:func:`glm_quat_copy`
 #. :c:func:`glm_quat_norm`
 #. :c:func:`glm_quat_normalize`
+#. :c:func:`glm_quat_normalize_to`
 #. :c:func:`glm_quat_dot`
-#. :c:func:`glm_quat_mulv`
+#. :c:func:`glm_quat_conjugate`
+#. :c:func:`glm_quat_inv`
+#. :c:func:`glm_quat_add`
+#. :c:func:`glm_quat_sub`
+#. :c:func:`glm_quat_real`
+#. :c:func:`glm_quat_imag`
+#. :c:func:`glm_quat_imagn`
+#. :c:func:`glm_quat_imaglen`
+#. :c:func:`glm_quat_angle`
+#. :c:func:`glm_quat_axis`
+#. :c:func:`glm_quat_mul`
 #. :c:func:`glm_quat_mat4`
+#. :c:func:`glm_quat_mat4t`
+#. :c:func:`glm_quat_mat3`
+#. :c:func:`glm_quat_mat3t`
+#. :c:func:`glm_quat_lerp`
 #. :c:func:`glm_quat_slerp`
+#. :c:func:`glm_quat_look`
+#. :c:func:`glm_quat_for`
+#. :c:func:`glm_quat_forp`
+#. :c:func:`glm_quat_rotatev`
+#. :c:func:`glm_quat_rotate`
+#. :c:func:`glm_quat_rotate_at`
+#. :c:func:`glm_quat_rotate_atm`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
@@ -47,10 +70,23 @@ Functions documentation
     Parameters:
       | *[in, out]*  **q**    quaternion
 
+.. c:function:: void  glm_quat_init(versor q, float x, float y, float z, float w)
+
+    | inits quaternion with given values
+
+    Parameters:
+      | *[out]* **q**      quaternion
+      | *[in]*  **x**      imag.x
+      | *[in]*  **y**      imag.y
+      | *[in]*  **z**      imag.z
+      | *[in]*  **w**      w (real part)
+
 .. c:function:: void  glm_quat(versor q, float  angle, float  x, float  y, float  z)
 
     | creates NEW quaternion with individual axis components
 
+    | given axis will be normalized
+
     Parameters:
       | *[out]* **q**      quaternion
       | *[in]*  **angle**  angle (radians)
@@ -58,14 +94,24 @@ Functions documentation
       | *[in]*  **y**      axis.y
       | *[in]*  **z**      axis.z
 
-.. c:function:: void  glm_quatv(versor q, float  angle, vec3   v)
+.. c:function:: void  glm_quatv(versor q, float  angle, vec3   axis)
 
     | creates NEW quaternion with axis vector
 
+    | given axis will be normalized
+
     Parameters:
       | *[out]* **q**      quaternion
       | *[in]*  **angle**  angle (radians)
-      | *[in]*  **v**      axis
+      | *[in]*  **axis**   axis (will be normalized)
+
+.. c:function:: void  glm_quat_copy(versor q, versor dest)
+
+    | copy quaternion to another one
+
+    Parameters:
+      | *[in]*  **q**     source quaternion
+      | *[out]* **dest**  destination quaternion
 
 .. c:function:: float  glm_quat_norm(versor q)
 
@@ -77,6 +123,14 @@ Functions documentation
     Returns:
       norm (magnitude)
 
+.. c:function:: void  glm_quat_normalize_to(versor q, versor dest)
+
+    | normalize quaternion and store result in dest, original one will not be normalized
+
+    Parameters:
+      | *[in]*  **q**    quaternion to normalize into
+      | *[out]* **dest** destination quaternion
+
 .. c:function:: void  glm_quat_normalize(versor q)
 
     | normalize quaternion
@@ -84,24 +138,118 @@ Functions documentation
     Parameters:
       | *[in, out]*  **q** quaternion
 
-.. c:function:: float  glm_quat_dot(versor q, versor r)
+.. c:function:: float  glm_quat_dot(versor p, versor q)
 
     dot product of two quaternion
 
     Parameters:
-      | *[in]*  **q1**   quaternion 1
-      | *[in]*  **q2**   quaternion 2
+      | *[in]*  **p**   quaternion 1
+      | *[in]*  **q**   quaternion 2
 
     Returns:
       dot product
 
-.. c:function:: void  glm_quat_mulv(versor q1, versor q2, versor dest)
+.. c:function:: void  glm_quat_conjugate(versor q, versor dest)
+
+    conjugate of quaternion
+
+    Parameters:
+      | *[in]*  **q**      quaternion
+      | *[in]*  **dest**   conjugate
+
+.. c:function:: void  glm_quat_inv(versor q, versor dest)
+
+    inverse of non-zero quaternion
+
+    Parameters:
+      | *[in]*  **q**      quaternion
+      | *[in]*  **dest**   inverse quaternion
+
+.. c:function:: void  glm_quat_add(versor p, versor q, versor dest)
+
+    add (componentwise) two quaternions and store result in dest
+
+    Parameters:
+      | *[in]*  **p**      quaternion 1
+      | *[in]*  **q**      quaternion 2
+      | *[in]*  **dest**   result quaternion
+
+.. c:function:: void  glm_quat_sub(versor p, versor q, versor dest)
+
+    subtract (componentwise) two quaternions and store result in dest
+
+    Parameters:
+      | *[in]*  **p**      quaternion 1
+      | *[in]*  **q**      quaternion 2
+      | *[in]*  **dest**   result quaternion
+
+.. c:function:: float  glm_quat_real(versor q)
+
+    returns real part of quaternion
+
+    Parameters:
+      | *[in]*  **q**   quaternion
+
+    Returns:
+      real part (quat.w)
+
+.. c:function:: void  glm_quat_imag(versor q, vec3 dest)
+
+    returns imaginary part of quaternion
+
+    Parameters:
+      | *[in]*   **q**      quaternion
+      | *[out]*  **dest**   imag
+
+.. c:function:: void  glm_quat_imagn(versor q, vec3 dest)
+
+    returns normalized imaginary part of quaternion
+
+    Parameters:
+      | *[in]*   **q**      quaternion
+      | *[out]*  **dest**   imag
+
+.. c:function:: float  glm_quat_imaglen(versor q)
+
+    returns length of imaginary part of quaternion
+
+    Parameters:
+      | *[in]*   **q**      quaternion
+
+    Returns:
+      norm of imaginary part
+
+.. c:function:: float  glm_quat_angle(versor q)
+
+    returns angle of quaternion
+
+    Parameters:
+      | *[in]*  **q**   quaternion
+
+    Returns:
+      angles of quat (radians)
+
+.. c:function:: void  glm_quat_axis(versor q, versor dest)
+
+    axis of quaternion
+
+    Parameters:
+      | *[in]*   **p**      quaternion
+      | *[out]*  **dest**   axis of quaternion
+
+.. c:function:: void  glm_quat_mul(versor p, versor q, versor dest)
 
     | multiplies two quaternion and stores result in dest
 
+    | this is also called Hamilton Product
+
+    | According to WikiPedia:
+    | The product of two rotation quaternions [clarification needed] will be
+      equivalent to the rotation q followed by the rotation p
+
     Parameters:
-      | *[in]*  **q1**    quaternion 1
-      | *[in]*  **q2**    quaternion 2
+      | *[in]*  **p**     quaternion 1 (first rotation)
+      | *[in]*  **q**     quaternion 2 (second rotation)
       | *[out]* **dest**  result quaternion
 
 .. c:function:: void  glm_quat_mat4(versor q, mat4 dest)
@@ -112,13 +260,121 @@ Functions documentation
       | *[in]*  **q**     quaternion
       | *[out]* **dest**  result matrix
 
+.. c:function:: void  glm_quat_mat4t(versor q, mat4 dest)
+
+    | convert quaternion to mat4 (transposed). This is transposed version of glm_quat_mat4
+
+    Parameters:
+      | *[in]*  **q**     quaternion
+      | *[out]* **dest**  result matrix
+
+.. c:function:: void  glm_quat_mat3(versor q, mat3 dest)
+
+    | convert quaternion to mat3
+
+    Parameters:
+      | *[in]*  **q**     quaternion
+      | *[out]* **dest**  result matrix
+
+.. c:function:: void  glm_quat_mat3t(versor q, mat3 dest)
+
+    | convert quaternion to mat3 (transposed). This is transposed version of glm_quat_mat3
+
+    Parameters:
+      | *[in]*  **q**     quaternion
+      | *[out]* **dest**  result matrix
+
+.. c:function:: void  glm_quat_lerp(versor from, versor to, float t, versor dest)
+
+    | interpolates between two quaternions
+    | using spherical linear interpolation (LERP)
+
+    Parameters:
+      | *[in]*  **from**  from
+      | *[in]*  **to**    to
+      | *[in]*  **t**     interpolant (amount) clamped between 0 and 1
+      | *[out]* **dest**  result quaternion
+
 .. c:function:: void glm_quat_slerp(versor q, versor r, float  t, versor dest)
 
     | interpolates between two quaternions
     | using spherical linear interpolation (SLERP)
 
     Parameters:
-      | *[in]*  **q**     from
-      | *[in]*  **r**     to
-      | *[in]*  **t**     amout
+      | *[in]*  **from**  from
+      | *[in]*  **to**    to
+      | *[in]*  **t**     interpolant (amount) clamped between 0 and 1
       | *[out]* **dest**  result quaternion
+
+.. c:function:: void  glm_quat_look(vec3 eye, versor ori, mat4 dest)
+
+    | creates view matrix using quaternion as camera orientation
+
+    Parameters:
+      | *[in]*  **eye**   eye
+      | *[in]*  **ori**   orientation in world space as quaternion
+      | *[out]* **dest**  result matrix
+
+.. c:function:: void  glm_quat_for(vec3 dir, vec3 fwd, vec3 up, versor dest)
+
+    | creates look rotation quaternion
+
+    Parameters:
+      | *[in]*  **dir**   direction to look
+      | *[in]*  **fwd**   forward vector
+      | *[in]*  **up**    up vector
+      | *[out]* **dest**  result matrix
+
+.. c:function:: void  glm_quat_forp(vec3 from, vec3 to, vec3 fwd, vec3 up, versor dest)
+
+    | creates look rotation quaternion using source and  destination positions p suffix stands for position
+
+    | this is similar to glm_quat_for except this computes direction for glm_quat_for for you.
+
+    Parameters:
+      | *[in]*  **from**  source point
+      | *[in]*  **to**    destination point
+      | *[in]*  **fwd**   forward vector
+      | *[in]*  **up**    up vector
+      | *[out]* **dest**  result matrix
+
+.. c:function:: void  glm_quat_rotatev(versor q, vec3 v, vec3 dest)
+
+    | crotate vector using using quaternion
+
+    Parameters:
+      | *[in]*  **q**     quaternion
+      | *[in]*  **v**     vector to rotate
+      | *[out]* **dest**  rotated vector
+
+.. c:function:: void glm_quat_rotate(mat4 m, versor q, mat4 dest)
+
+    | rotate existing transform matrix using quaternion
+
+    instead of passing identity matrix, consider to use quat_mat4 functions
+
+    Parameters:
+      | *[in]*  **m**     existing transform matrix to rotate
+      | *[in]*  **q**     quaternion
+      | *[out]* **dest**  rotated matrix/transform
+
+.. c:function:: void glm_quat_rotate_at(mat4 m, versor q, vec3 pivot)
+
+    | rotate existing transform matrix using quaternion at pivot point
+
+    Parameters:
+      | *[in, out]*  **m**      existing transform matrix to rotate
+      | *[in]*       **q**      quaternion
+      | *[in]*       **pivot**  pivot
+
+.. c:function:: void glm_quat_rotate(mat4 m, versor q, mat4 dest)
+
+    | rotate NEW transform matrix using quaternion at pivot point
+    | this creates rotation matrix, it assumes you don't have a matrix
+
+    | this should work faster than glm_quat_rotate_at because it reduces one glm_translate.
+
+    Parameters:
+      | *[in, out]*  **m**      existing transform matrix to rotate
+      | *[in]*       **q**      quaternion
+      | *[in]*       **pivot**  pivot

docs/source/sphere.rst

@@ -0,0 +1,74 @@
+.. default-domain:: C
+
+Sphere
+================================================================================
+
+Header: cglm/sphere.h
+
+**Definition of sphere:**
+
+Sphere Representation in cglm is *vec4*: **[center.x, center.y, center.z, radii]**
+
+You can call any vec3 function by pasing sphere. Because first three elements
+defines center of sphere.
+
+Table of contents (click to go):
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Functions:
+
+1. :c:func:`glm_sphere_radii`
+#. :c:func:`glm_sphere_transform`
+#. :c:func:`glm_sphere_merge`
+#. :c:func:`glm_sphere_sphere`
+#. :c:func:`glm_sphere_point`
+
+Functions documentation
+~~~~~~~~~~~~~~~~~~~~~~~
+
+.. c:function:: float  glm_sphere_radii(vec4 s)
+
+    | helper for getting sphere radius
+
+    Parameters:
+      | *[in]*  **s**   sphere
+
+    Returns:
+       returns radii
+
+.. c:function:: void  glm_sphere_transform(vec4 s, mat4 m, vec4 dest)
+
+    | apply transform to sphere, it is just wrapper for glm_mat4_mulv3
+
+    Parameters:
+      | *[in]*  **s**    sphere
+      | *[in]*  **m**    transform matrix
+      | *[out]* **dest** transformed sphere
+
+.. c:function:: void  glm_sphere_merge(vec4 s1, vec4 s2, vec4 dest)
+
+    | merges two spheres and creates a new one
+
+    two sphere must be in same space, for instance if one in world space then
+    the other must be in world space too, not in local space.
+
+    Parameters:
+      | *[in]*  **s1**      sphere 1
+      | *[in]*  **s2**      sphere 2
+      | *[out]* **dest**    merged/extended sphere
+
+.. c:function:: bool  glm_sphere_sphere(vec4 s1, vec4 s2)
+
+    | check if two sphere intersects
+
+    Parameters:
+      | *[in]*  **s1**      sphere
+      | *[in]*  **s2**      other sphere
+
+.. c:function:: bool  glm_sphere_point(vec4 s, vec3 point)
+
+    | check if sphere intersects with point
+
+    Parameters:
+      | *[in]*  **s**       sphere
+      | *[in]*  **point**   point

docs/source/troubleshooting.rst

@@ -0,0 +1,79 @@
+.. default-domain:: C
+
+Troubleshooting
+================================================================================
+
+It is possible that sometimes you may get crashes or wrong results.
+Follow these topics
+
+Memory Allocation:
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Again, **cglm** doesn't alloc any memory on heap.
+cglm functions works like memcpy; it copies data from src,
+makes calculations then copy the result to dest.
+
+You are responsible for allocation of **src** and **dest** parameters.
+
+Alignment:
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+**vec4** and **mat4** types requires 16 byte alignment.
+These types are marked with align attribute to let compiler know about this
+requirement.
+
+But since MSVC (Windows) throws the error:
+
+**"formal parameter with requested alignment of 16 won't be aligned"**
+
+The alignment attribute has been commented for MSVC
+
+.. code-block:: c
+
+   #if defined(_MSC_VER)
+   #  define CGLM_ALIGN(X) /* __declspec(align(X)) */
+   #else
+   #  define CGLM_ALIGN(X) __attribute((aligned(X)))
+   #endif.
+
+So MSVC may not know about alignment requirements when creating variables.
+The interesting thing is that, if I remember correctly Visual Studio 2017
+doesn't throw the above error. So we may uncomment that line for Visual Studio 2017,
+you may do it yourself.
+
+**This MSVC issue is still in TODOs.**
+
+**UPDATE:** By starting v0.4.5 cglm provides an option to disable alignment requirement.
+Also alignment is disabled for older msvc verisons as default. Now alignment is only required in Visual Studio 2017 version 15.6+ if CGLM_ALL_UNALIGNED macro is not defined.
+
+Crashes, Invalid Memory Access:
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Probably you are trying to write to invalid memory location.
+
+You may used wrong function for what you want to do.
+
+For instance you may called **glm_vec4_** functions for **vec3** data type.
+It will try to write 32 byte but since **vec3** is 24 byte it should throw
+memory access error or exit the app without saying anything.
+
+Wrong Results:
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Again, you may used wrong function.
+
+For instance if you use **glm_normalize()** or **glm_vec_normalize()** for **vec4**,
+it will assume that passed param is **vec3** and will normalize it for **vec3**.
+Since you need to **vec4** to be normalized in your case, you will get wrong results.
+
+Accessing vec4 type with vec3 functions is valid, you will not get any error, exception or crash.
+You only get wrong results if you don't know what you are doing!
+
+So be carefull, when your IDE (Xcode, Visual Studio ...) tried to autocomplete function names, READ IT :)
+
+**Also implementation may be wrong please let us know by creating an issue on Github.**
+
+Other Issues?
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+**Please let us know by creating an issue on Github.**

docs/source/util.rst

@@ -13,18 +13,25 @@ Table of contents (click to go):
 Functions:
 
 1. :c:func:`glm_sign`
+#. :c:func:`glm_signf`
 #. :c:func:`glm_rad`
 #. :c:func:`glm_deg`
 #. :c:func:`glm_make_rad`
 #. :c:func:`glm_make_deg`
 #. :c:func:`glm_pow2`
+#. :c:func:`glm_min`
+#. :c:func:`glm_max`
+#. :c:func:`glm_clamp`
+#. :c:func:`glm_lerp`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
 
 .. c:function:: int  glm_sign(int val)
 
-    | returns sign of 32 bit integer as +1 or -1
+    | returns sign of 32 bit integer as +1, -1, 0
+
+    | **Important**: It returns 0 for zero input
 
     Parameters:
       | *[in]*  **val**   an integer
@@ -32,6 +39,18 @@ Functions documentation
     Returns:
       sign of given number
 
+.. c:function:: float  glm_signf(float val)
+
+    | returns sign of 32 bit integer as +1.0, -1.0, 0.0
+
+    | **Important**: It returns 0.0f for zero input
+
+    Parameters:
+      | *[in]*  **val**   a float
+
+    Returns:
+      sign of given number
+
 .. c:function:: float  glm_rad(float deg)
 
     | convert degree to radians
@@ -91,3 +110,29 @@ Functions documentation
 
     Returns:
       maximum value
+
+.. c:function:: void  glm_clamp(float val, float minVal, float maxVal)
+
+    constrain a value to lie between two further values
+
+    Parameters:
+      | *[in]*  **val**     input value
+      | *[in]*  **minVal**  minimum value
+      | *[in]*  **maxVal**  maximum value
+
+    Returns:
+      clamped value
+
+.. c:function:: float  glm_lerp(float from, float to, float t)
+
+    linear interpolation between two number
+
+    | formula:  from + s * (to - from)
+
+    Parameters:
+      | *[in]*  **from**   from value
+      | *[in]*  **to**     to value
+      | *[in]*  **t**      interpolant (amount) clamped between 0 and 1
+
+    Returns:
+       interpolated value

docs/source/vec3-ext.rst

@@ -23,6 +23,11 @@ Functions:
 #. :c:func:`glm_vec_eqv_eps`
 #. :c:func:`glm_vec_max`
 #. :c:func:`glm_vec_min`
+#. :c:func:`glm_vec_isnan`
+#. :c:func:`glm_vec_isinf`
+#. :c:func:`glm_vec_isvalid`
+#. :c:func:`glm_vec_sign`
+#. :c:func:`glm_vec_sqrt`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
@@ -96,3 +101,43 @@ Functions documentation
 
     Parameters:
       | *[in]*  **v**  vector
+
+.. c:function:: bool glm_vec_isnan(vec3 v)
+
+    | check if one of items is NaN (not a number)
+    | you should only use this in DEBUG mode or very critical asserts
+
+    Parameters:
+      | *[in]*  **v**  vector
+
+.. c:function:: bool glm_vec_isinf(vec3 v)
+
+    | check if one of items is INFINITY
+    | you should only use this in DEBUG mode or very critical asserts
+
+    Parameters:
+      | *[in]*  **v**  vector
+
+.. c:function:: bool glm_vec_isvalid(vec3 v)
+
+    | check if all items are valid number
+    | you should only use this in DEBUG mode or very critical asserts
+
+    Parameters:
+      | *[in]*  **v**  vector
+
+.. c:function:: void glm_vec_sign(vec3 v, vec3 dest)
+
+    get sign of 32 bit float as +1, -1, 0
+
+    Parameters:
+      | *[in]*   **v**     vector
+      | *[out]*  **dest**  sign vector (only keeps signs as -1, 0, -1)
+
+.. c:function:: void glm_vec_sqrt(vec3 v, vec3 dest)
+
+    square root of each vector item
+
+    Parameters:
+      | *[in]*   **v**     vector
+      | *[out]*  **dest**  destination vector (sqrt(v))

docs/source/vec3.rst

@@ -31,28 +31,44 @@ Functions:
 
 1. :c:func:`glm_vec3`
 #. :c:func:`glm_vec_copy`
+#. :c:func:`glm_vec_zero`
+#. :c:func:`glm_vec_one`
 #. :c:func:`glm_vec_dot`
 #. :c:func:`glm_vec_cross`
 #. :c:func:`glm_vec_norm2`
 #. :c:func:`glm_vec_norm`
 #. :c:func:`glm_vec_add`
+#. :c:func:`glm_vec_adds`
 #. :c:func:`glm_vec_sub`
+#. :c:func:`glm_vec_subs`
+#. :c:func:`glm_vec_mul`
 #. :c:func:`glm_vec_scale`
 #. :c:func:`glm_vec_scale_as`
+#. :c:func:`glm_vec_div`
+#. :c:func:`glm_vec_divs`
+#. :c:func:`glm_vec_addadd`
+#. :c:func:`glm_vec_subadd`
+#. :c:func:`glm_vec_muladd`
+#. :c:func:`glm_vec_muladds`
 #. :c:func:`glm_vec_flipsign`
+#. :c:func:`glm_vec_flipsign_to`
 #. :c:func:`glm_vec_inv`
 #. :c:func:`glm_vec_inv_to`
 #. :c:func:`glm_vec_normalize`
 #. :c:func:`glm_vec_normalize_to`
+#. :c:func:`glm_vec_distance2`
 #. :c:func:`glm_vec_distance`
 #. :c:func:`glm_vec_angle`
 #. :c:func:`glm_vec_rotate`
 #. :c:func:`glm_vec_rotate_m4`
+#. :c:func:`glm_vec_rotate_m3`
 #. :c:func:`glm_vec_proj`
 #. :c:func:`glm_vec_center`
 #. :c:func:`glm_vec_maxv`
 #. :c:func:`glm_vec_minv`
 #. :c:func:`glm_vec_ortho`
+#. :c:func:`glm_vec_clamp`
+#. :c:func:`glm_vec_lerp`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
@@ -73,6 +89,20 @@ Functions documentation
       | *[in]*  **a**     source
       | *[out]* **dest**  destination
 
+.. c:function:: void  glm_vec_zero(vec3 v)
+
+    makes all members 0.0f (zero)
+
+    Parameters:
+      | *[in, out]*  **v**     vector
+
+.. c:function:: void  glm_vec_one(vec3 v)
+
+    makes all members 1.0f (one)
+
+    Parameters:
+      | *[in, out]*  **v**     vector
+
 .. c:function:: float  glm_vec_dot(vec3 a, vec3 b)
 
     dot product of vec3
@@ -114,24 +144,51 @@ Functions documentation
     Parameters:
       | *[in]*  **vec**   vector
 
-.. c:function:: void  glm_vec_add(vec3 v1, vec3 v2, vec3 dest)
+.. c:function:: void  glm_vec_add(vec3 a, vec3 b, vec3 dest)
 
-    add v2 vector to v1 vector store result in dest
+    add a vector to b vector store result in dest
 
     Parameters:
-      | *[in]*  **v1**    vector1
-      | *[in]*  **v2**    vector2
+      | *[in]*  **a**     vector1
+      | *[in]*  **b**     vector2
+      | *[out]* **dest**  destination vector
+
+.. c:function:: void  glm_vec_adds(vec3 a, float s, vec3 dest)
+
+    add scalar to v vector store result in dest (d = v + vec(s))
+
+    Parameters:
+      | *[in]*  **v**     vector
+      | *[in]*  **s**     scalar
       | *[out]* **dest**  destination vector
 
 .. c:function:: void  glm_vec_sub(vec3 v1, vec3 v2, vec3 dest)
 
-    subtract v2 vector from v1 vector store result in dest
+    subtract b vector from a vector store result in dest (d = v1 - v2)
 
     Parameters:
-      | *[in]*  **v1**    vector1
-      | *[in]*  **v2**    vector2
+      | *[in]*  **a**     vector1
+      | *[in]*  **b**     vector2
       | *[out]* **dest**  destination vector
 
+.. c:function:: void  glm_vec_subs(vec3 v, float s, vec3 dest)
+
+    subtract scalar from v vector store result in dest (d = v - vec(s))
+
+    Parameters:
+      | *[in]*  **v**     vector
+      | *[in]*  **s**     scalar
+      | *[out]* **dest**  destination vector
+
+.. c:function:: void  glm_vec_mul(vec3 a, vec3 b, vec3 d)
+
+    multiply two vector (component-wise multiplication)
+
+    Parameters:
+      | *[in]*  **a**     vector
+      | *[in]*  **b**     scalar
+      | *[out]* **d**     result = (a[0] * b[0], a[1] * b[1], a[2] * b[2])
+
 .. c:function:: void glm_vec_scale(vec3 v, float s, vec3 dest)
 
      multiply/scale vec3 vector with scalar: result = v * s
@@ -151,6 +208,64 @@ Functions documentation
       | *[in]*  **s**     scalar
       | *[out]* **dest**  destination vector
 
+.. c:function:: void  glm_vec_div(vec3 a, vec3 b, vec3 dest)
+
+    div vector with another component-wise division: d = a / b
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  result = (a[0] / b[0], a[1] / b[1], a[2] / b[2])
+
+.. c:function:: void  glm_vec_divs(vec3 v, float s, vec3 dest)
+
+    div vector with scalar: d = v / s
+
+    Parameters:
+      | *[in]*  **v**     vector
+      | *[in]*  **s**     scalar
+      | *[out]* **dest**  result = (a[0] / s, a[1] / s, a[2] / s])
+
+.. c:function:: void  glm_vec_addadd(vec3 a, vec3 b, vec3 dest)
+
+    | add two vectors and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  dest += (a + b)
+
+.. c:function:: void  glm_vec_subadd(vec3 a, vec3 b, vec3 dest)
+
+    | sub two vectors and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  dest += (a - b)
+
+.. c:function:: void  glm_vec_muladd(vec3 a, vec3 b, vec3 dest)
+
+    | mul two vectors and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  dest += (a * b)
+
+.. c:function:: void  glm_vec_muladds(vec3 a, float s, vec3 dest)
+
+    | mul vector with scalar and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector
+      | *[in]*  **s**     scalar
+      | *[out]* **dest**  dest += (a * b)
+
 .. c:function:: void  glm_vec_flipsign(vec3 v)
 
     flip sign of all vec3 members
@@ -158,6 +273,14 @@ Functions documentation
     Parameters:
       | *[in, out]*  **v**    vector
 
+.. c:function:: void  glm_vec_flipsign_to(vec3 v, vec3 dest)
+
+    flip sign of all vec3 members and store result in dest
+
+    Parameters:
+      | *[in]*  **v**       vector
+      | *[out]* **dest**    negated vector
+
 .. c:function:: void  glm_vec_inv(vec3 v)
 
     make vector as inverse/opposite of itself
@@ -205,7 +328,7 @@ Functions documentation
 
     Parameters:
       | *[in, out]*  **v**      vector
-      | *[in]*       **axis**   axis vector (must be unit vector)
+      | *[in]*       **axis**   axis vector (will be normalized)
       | *[out]*      **angle**  angle (radians)
 
 .. c:function:: void  glm_vec_rotate_m4(mat4 m, vec3 v, vec3 dest)
@@ -217,6 +340,15 @@ Functions documentation
       | *[in]*  **v**     vector
       | *[out]* **dest**  rotated vector
 
+.. c:function:: void  glm_vec_rotate_m3(mat3 m, vec3 v, vec3 dest)
+
+    apply rotation matrix to vector
+
+    Parameters:
+      | *[in]*  **m**     affine matrix or rot matrix
+      | *[in]*  **v**     vector
+      | *[out]* **dest**  rotated vector
+
 .. c:function:: void  glm_vec_proj(vec3 a, vec3 b, vec3 dest)
 
     project a vector onto b vector
@@ -235,6 +367,17 @@ Functions documentation
       | *[in]*  **v2**    vector2
       | *[out]* **dest**  center point
 
+.. c:function:: float  glm_vec_distance2(vec3 v1, vec3 v2)
+
+    squared distance between two vectors
+
+    Parameters:
+      | *[in]*  **mat**   vector1
+      | *[in]*  **row1**  vector2
+
+    Returns:
+      | squared distance (distance * distance)
+
 .. c:function:: float  glm_vec_distance(vec3 v1, vec3 v2)
 
     distance between two vectors
@@ -271,3 +414,24 @@ Functions documentation
     Parameters:
       | *[in]*  **mat**   vector
       | *[out]* **dest**  orthogonal/perpendicular vector
+
+.. c:function:: void  glm_vec_clamp(vec3 v, float minVal, float maxVal)
+
+    constrain a value to lie between two further values
+
+    Parameters:
+      | *[in, out]*  **v**       vector
+      | *[in]*       **minVal**  minimum value
+      | *[in]*       **maxVal**  maximum value
+
+.. c:function:: void  glm_vec_lerp(vec3 from, vec3 to, float t, vec3 dest)
+
+    linear interpolation between two vector
+
+    | formula:  from + s * (to - from)
+
+    Parameters:
+      | *[in]*  **from**   from value
+      | *[in]*  **to**     to value
+      | *[in]*  **t**      interpolant (amount) clamped between 0 and 1
+      | *[out]* **dest**   destination

docs/source/vec4-ext.rst

@@ -96,3 +96,43 @@ Functions documentation
 
     Parameters:
       | *[in]*  **v**  vector
+
+.. c:function:: bool glm_vec4_isnan(vec4 v)
+
+    | check if one of items is NaN (not a number)
+    | you should only use this in DEBUG mode or very critical asserts
+
+    Parameters:
+      | *[in]*  **v**  vector
+
+.. c:function:: bool glm_vec4_isinf(vec4 v)
+
+    | check if one of items is INFINITY
+    | you should only use this in DEBUG mode or very critical asserts
+
+    Parameters:
+      | *[in]*  **v**  vector
+
+.. c:function:: bool glm_vec4_isvalid(vec4 v)
+
+    | check if all items are valid number
+    | you should only use this in DEBUG mode or very critical asserts
+
+    Parameters:
+      | *[in]*  **v**  vector
+
+.. c:function:: void glm_vec4_sign(vec4 v, vec4 dest)
+
+    get sign of 32 bit float as +1, -1, 0
+
+    Parameters:
+      | *[in]*   **v**     vector
+      | *[out]*  **dest**  sign vector (only keeps signs as -1, 0, -1)
+
+.. c:function:: void glm_vec4_sqrt(vec4 v, vec4 dest)
+
+    square root of each vector item
+
+    Parameters:
+      | *[in]*   **v**     vector
+      | *[out]*  **dest**  destination vector (sqrt(v))

docs/source/vec4.rst

@@ -24,14 +24,26 @@ Functions:
 1. :c:func:`glm_vec4`
 #. :c:func:`glm_vec4_copy3`
 #. :c:func:`glm_vec4_copy`
+#. :c:func:`glm_vec4_zero`
+#. :c:func:`glm_vec4_one`
 #. :c:func:`glm_vec4_dot`
 #. :c:func:`glm_vec4_norm2`
 #. :c:func:`glm_vec4_norm`
 #. :c:func:`glm_vec4_add`
+#. :c:func:`glm_vec4_adds`
 #. :c:func:`glm_vec4_sub`
+#. :c:func:`glm_vec4_subs`
+#. :c:func:`glm_vec4_mul`
 #. :c:func:`glm_vec4_scale`
 #. :c:func:`glm_vec4_scale_as`
+#. :c:func:`glm_vec4_div`
+#. :c:func:`glm_vec4_divs`
+#. :c:func:`glm_vec4_addadd`
+#. :c:func:`glm_vec4_subadd`
+#. :c:func:`glm_vec4_muladd`
+#. :c:func:`glm_vec4_muladds`
 #. :c:func:`glm_vec4_flipsign`
+#. :c:func:`glm_vec_flipsign_to`
 #. :c:func:`glm_vec4_inv`
 #. :c:func:`glm_vec4_inv_to`
 #. :c:func:`glm_vec4_normalize`
@@ -39,6 +51,13 @@ Functions:
 #. :c:func:`glm_vec4_distance`
 #. :c:func:`glm_vec4_maxv`
 #. :c:func:`glm_vec4_minv`
+#. :c:func:`glm_vec4_clamp`
+#. :c:func:`glm_vec4_lerp`
+#. :c:func:`glm_vec4_isnan`
+#. :c:func:`glm_vec4_isinf`
+#. :c:func:`glm_vec4_isvalid`
+#. :c:func:`glm_vec4_sign`
+#. :c:func:`glm_vec4_sqrt`
 
 Functions documentation
 ~~~~~~~~~~~~~~~~~~~~~~~
@@ -70,6 +89,13 @@ Functions documentation
       | *[in]*  **v**     source
       | *[in]*  **dest**  destination
 
+.. c:function:: void  glm_vec4_zero(vec4 v)
+
+    makes all members zero
+
+    Parameters:
+      | *[in, out]*  **v**     vector
+
 .. c:function:: float  glm_vec4_dot(vec4 a, vec4 b)
 
     dot product of vec4
@@ -102,24 +128,51 @@ Functions documentation
     Parameters:
       | *[in]*  **vec**   vector
 
-.. c:function:: void  glm_vec4_add(vec4 v1, vec4 v2, vec4 dest)
+.. c:function:: void  glm_vec4_add(vec4 a, vec4 b, vec4 dest)
 
-    add v2 vector to v1 vector store result in dest
+    add a vector to b vector store result in dest
 
     Parameters:
-      | *[in]*  **v1**    vector1
-      | *[in]*  **v2**    vector2
+      | *[in]*  **a**     vector1
+      | *[in]*  **b**     vector2
       | *[out]* **dest**  destination vector
 
-.. c:function:: void  glm_vec4_sub(vec4 v1, vec4 v2, vec4 dest)
+.. c:function:: void  glm_vec4_adds(vec4 v, float s, vec4 dest)
 
-    subtract v2 vector from v1 vector store result in dest
+    add scalar to v vector store result in dest (d = v + vec(s))
 
     Parameters:
-      | *[in]*  **v1**    vector1
-      | *[in]*  **v2**    vector2
+      | *[in]*  **v**     vector
+      | *[in]*  **s**     scalar
       | *[out]* **dest**  destination vector
 
+.. c:function:: void  glm_vec4_sub(vec4 a, vec4 b, vec4 dest)
+
+    subtract b vector from a vector store result in dest (d = v1 - v2)
+
+    Parameters:
+      | *[in]*  **a**     vector1
+      | *[in]*  **b**     vector2
+      | *[out]* **dest**  destination vector
+
+.. c:function:: void  glm_vec4_subs(vec4 v, float s, vec4 dest)
+
+    subtract scalar from v vector store result in dest (d = v - vec(s))
+
+    Parameters:
+      | *[in]*  **v**     vector
+      | *[in]*  **s**     scalar
+      | *[out]* **dest**  destination vector
+
+.. c:function:: void  glm_vec4_mul(vec4 a, vec4 b, vec4 d)
+
+    multiply two vector (component-wise multiplication)
+
+    Parameters:
+      | *[in]*  **a**     vector1
+      | *[in]*  **b**     vector2
+      | *[out]* **dest**  result = (a[0] * b[0], a[1] * b[1], a[2] * b[2], a[3] * b[3])
+
 .. c:function:: void glm_vec4_scale(vec4 v, float s, vec4 dest)
 
      multiply/scale vec4 vector with scalar: result = v * s
@@ -138,6 +191,64 @@ Functions documentation
       | *[in]*  **s**     scalar
       | *[out]* **dest**  destination vector
 
+.. c:function:: void  glm_vec4_div(vec4 a, vec4 b, vec4 dest)
+
+    div vector with another component-wise division: d = v1 / v2
+
+    Parameters:
+      | *[in]*  **a**     vector1
+      | *[in]*  **b**     vector2
+      | *[out]* **dest**  result = (a[0] / b[0], a[1] / b[1], a[2] / b[2], a[3] / b[3])
+
+.. c:function:: void  glm_vec4_divs(vec4 v, float s, vec4 dest)
+
+    div vector with scalar: d = v / s
+
+    Parameters:
+      | *[in]*  **v**     vector
+      | *[in]*  **s**     scalar
+      | *[out]* **dest**  result = (a[0] / s, a[1] / s, a[2] / s, a[3] / s)
+
+.. c:function:: void  glm_vec4_addadd(vec4 a, vec4 b, vec4 dest)
+
+    | add two vectors and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  dest += (a + b)
+
+.. c:function:: void  glm_vec4_subadd(vec4 a, vec4 b, vec4 dest)
+
+    | sub two vectors and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  dest += (a - b)
+
+.. c:function:: void  glm_vec4_muladd(vec4 a, vec4 b, vec4 dest)
+
+    | mul two vectors and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector 1
+      | *[in]*  **b**     vector 2
+      | *[out]* **dest**  dest += (a * b)
+
+.. c:function:: void  glm_vec4_muladds(vec4 a, float s, vec4 dest)
+
+    | mul vector with scalar and add result to sum
+    | it applies += operator so dest must be initialized
+
+    Parameters:
+      | *[in]*  **a**     vector
+      | *[in]*  **s**     scalar
+      | *[out]* **dest**  dest += (a * b)
+
 .. c:function:: void  glm_vec4_flipsign(vec4 v)
 
     flip sign of all vec4 members
@@ -145,6 +256,14 @@ Functions documentation
     Parameters:
     | *[in, out]*  **v**    vector
 
+.. c:function:: void  glm_vec4_flipsign_to(vec4 v, vec4 dest)
+
+    flip sign of all vec4 members and store result in dest
+
+    Parameters:
+      | *[in]*  **v**       vector
+      | *[out]* **dest**    negated vector
+
 .. c:function:: void  glm_vec4_inv(vec4 v)
 
     make vector as inverse/opposite of itself
@@ -203,3 +322,24 @@ Functions documentation
       | *[in]*  **v1**    vector1
       | *[in]*  **v2**    vector2
       | *[out]* **dest**  destination
+
+.. c:function:: void  glm_vec4_clamp(vec4 v, float minVal, float maxVal)
+
+    constrain a value to lie between two further values
+
+    Parameters:
+      | *[in, out]*  **v**       vector
+      | *[in]*       **minVal**  minimum value
+      | *[in]*       **maxVal**  maximum value
+
+.. c:function:: void  glm_vec4_lerp(vec4 from, vec4 to, float t, vec4 dest)
+
+    linear interpolation between two vector
+
+    | formula:  from + s * (to - from)
+
+    Parameters:
+      | *[in]*  **from**   from value
+      | *[in]*  **to**     to value
+      | *[in]*  **t**      interpolant (amount) clamped between 0 and 1
+      | *[out]* **dest**   destination

include/cglm/affine-mat.h

@@ -16,6 +16,7 @@
 
 #include "common.h"
 #include "mat4.h"
+#include "mat3.h"
 
 #ifdef CGLM_SSE_FP
 #  include "simd/sse2/affine.h"
@@ -81,6 +82,59 @@ glm_mul(mat4 m1, mat4 m2, mat4 dest) {
 #endif
 }
 
+/*!
+ * @brief this is similar to glm_mat4_mul but specialized to affine transform
+ *
+ * Right Matrix format should be:
+ *   R  R  R  0
+ *   R  R  R  0
+ *   R  R  R  0
+ *   0  0  0  1
+ *
+ * this reduces some multiplications. It should be faster than mat4_mul.
+ * if you are not sure about matrix format then DON'T use this! use mat4_mul
+ *
+ * @param[in]   m1    affine matrix 1
+ * @param[in]   m2    affine matrix 2
+ * @param[out]  dest  result matrix
+ */
+CGLM_INLINE
+void
+glm_mul_rot(mat4 m1, mat4 m2, mat4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glm_mul_rot_sse2(m1, m2, dest);
+#else
+  float a00 = m1[0][0], a01 = m1[0][1], a02 = m1[0][2], a03 = m1[0][3],
+        a10 = m1[1][0], a11 = m1[1][1], a12 = m1[1][2], a13 = m1[1][3],
+        a20 = m1[2][0], a21 = m1[2][1], a22 = m1[2][2], a23 = m1[2][3],
+        a30 = m1[3][0], a31 = m1[3][1], a32 = m1[3][2], a33 = m1[3][3],
+
+        b00 = m2[0][0], b01 = m2[0][1], b02 = m2[0][2],
+        b10 = m2[1][0], b11 = m2[1][1], b12 = m2[1][2],
+        b20 = m2[2][0], b21 = m2[2][1], b22 = m2[2][2];
+
+  dest[0][0] = a00 * b00 + a10 * b01 + a20 * b02;
+  dest[0][1] = a01 * b00 + a11 * b01 + a21 * b02;
+  dest[0][2] = a02 * b00 + a12 * b01 + a22 * b02;
+  dest[0][3] = a03 * b00 + a13 * b01 + a23 * b02;
+
+  dest[1][0] = a00 * b10 + a10 * b11 + a20 * b12;
+  dest[1][1] = a01 * b10 + a11 * b11 + a21 * b12;
+  dest[1][2] = a02 * b10 + a12 * b11 + a22 * b12;
+  dest[1][3] = a03 * b10 + a13 * b11 + a23 * b12;
+
+  dest[2][0] = a00 * b20 + a10 * b21 + a20 * b22;
+  dest[2][1] = a01 * b20 + a11 * b21 + a21 * b22;
+  dest[2][2] = a02 * b20 + a12 * b21 + a22 * b22;
+  dest[2][3] = a03 * b20 + a13 * b21 + a23 * b22;
+
+  dest[3][0] = a30;
+  dest[3][1] = a31;
+  dest[3][2] = a32;
+  dest[3][3] = a33;
+#endif
+}
+
 /*!
  * @brief inverse orthonormal rotation + translation matrix (ridig-body)
  *

include/cglm/affine.h

@@ -16,15 +16,14 @@
    CGLM_INLINE void glm_scale_to(mat4 m, vec3 v, mat4 dest);
    CGLM_INLINE void glm_scale_make(mat4 m, vec3 v);
    CGLM_INLINE void glm_scale(mat4 m, vec3 v);
-   CGLM_INLINE void glm_scale1(mat4 m, float s);
    CGLM_INLINE void glm_scale_uni(mat4 m, float s);
    CGLM_INLINE void glm_rotate_x(mat4 m, float angle, mat4 dest);
    CGLM_INLINE void glm_rotate_y(mat4 m, float angle, mat4 dest);
    CGLM_INLINE void glm_rotate_z(mat4 m, float angle, mat4 dest);
-   CGLM_INLINE void glm_rotate_ndc_make(mat4 m, float angle, vec3 axis_ndc);
    CGLM_INLINE void glm_rotate_make(mat4 m, float angle, vec3 axis);
-   CGLM_INLINE void glm_rotate_ndc(mat4 m, float angle, vec3 axis);
    CGLM_INLINE void glm_rotate(mat4 m, float angle, vec3 axis);
+   CGLM_INLINE void glm_rotate_at(mat4 m, vec3 pivot, float angle, vec3 axis);
+   CGLM_INLINE void glm_rotate_atm(mat4 m, vec3 pivot, float angle, vec3 axis);
    CGLM_INLINE void glm_decompose_scalev(mat4 m, vec3 s);
    CGLM_INLINE bool glm_uniscaled(mat4 m);
    CGLM_INLINE void glm_decompose_rs(mat4 m, mat4 r, vec3 s);
@@ -35,51 +34,15 @@
 #define cglm_affine_h
 
 #include "common.h"
-#include "vec4.h"
-#include "affine-mat.h"
 #include "util.h"
+#include "vec3.h"
+#include "vec4.h"
+#include "mat4.h"
+#include "affine-mat.h"
 
-/*!
- * @brief translate existing transform matrix by v vector
- *        and store result in dest
- *
- * @param[in]  m    affine transfrom
- * @param[in]  v    translate vector [x, y, z]
- * @param[out] dest translated matrix
- */
 CGLM_INLINE
 void
-glm_translate_to(mat4 m, vec3 v, mat4 dest) {
-  mat4 t = GLM_MAT4_IDENTITY_INIT;
-
-#if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(dest[3],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_load_ps(t[0]),
-                                                _mm_set1_ps(v[0])),
-                                     _mm_mul_ps(_mm_load_ps(t[1]),
-                                                _mm_set1_ps(v[1]))),
-                          _mm_add_ps(_mm_mul_ps(_mm_load_ps(t[2]),
-                                                _mm_set1_ps(v[2])),
-                                     _mm_load_ps(t[3]))))
-  ;
-
-  _mm_store_ps(dest[0], _mm_load_ps(m[0]));
-  _mm_store_ps(dest[1], _mm_load_ps(m[1]));
-  _mm_store_ps(dest[2], _mm_load_ps(m[2]));
-#else
-  vec4 v1, v2, v3;
-
-  glm_vec4_scale(t[0], v[0], v1);
-  glm_vec4_scale(t[1], v[1], v2);
-  glm_vec4_scale(t[2], v[2], v3);
-
-  glm_vec4_add(v1, t[3], t[3]);
-  glm_vec4_add(v2, t[3], t[3]);
-  glm_vec4_add(v3, t[3], t[3]);
-
-  glm__memcpy(float, dest, t, sizeof(mat4));
-#endif
-}
+glm_mat4_mul(mat4 m1, mat4 m2, mat4 dest);
 
 /*!
  * @brief translate existing transform matrix by v vector
@@ -92,14 +55,14 @@ CGLM_INLINE
 void
 glm_translate(mat4 m, vec3 v) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(m[3],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_load_ps(m[0]),
+  glmm_store(m[3],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_load(m[0]),
                                               _mm_set1_ps(v[0])),
-                                     _mm_mul_ps(_mm_load_ps(m[1]),
+                                   _mm_mul_ps(glmm_load(m[1]),
                                               _mm_set1_ps(v[1]))),
-                          _mm_add_ps(_mm_mul_ps(_mm_load_ps(m[2]),
+                        _mm_add_ps(_mm_mul_ps(glmm_load(m[2]),
                                               _mm_set1_ps(v[2])),
-                                     _mm_load_ps(m[3]))))
+                                   glmm_load(m[3]))))
   ;
 #else
   vec4 v1, v2, v3;
@@ -114,6 +77,23 @@ glm_translate(mat4 m, vec3 v) {
 #endif
 }
 
+/*!
+ * @brief translate existing transform matrix by v vector
+ *        and store result in dest
+ *
+ * source matrix will remain same
+ *
+ * @param[in]  m    affine transfrom
+ * @param[in]  v    translate vector [x, y, z]
+ * @param[out] dest translated matrix
+ */
+CGLM_INLINE
+void
+glm_translate_to(mat4 m, vec3 v, mat4 dest) {
+  glm_mat4_copy(m, dest);
+  glm_translate(dest, v);
+}
+
 /*!
  * @brief translate existing transform matrix by x factor
  *
@@ -124,10 +104,10 @@ CGLM_INLINE
 void
 glm_translate_x(mat4 m, float x) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(m[3],
-               _mm_add_ps(_mm_mul_ps(_mm_load_ps(m[0]),
+  glmm_store(m[3],
+             _mm_add_ps(_mm_mul_ps(glmm_load(m[0]),
                                    _mm_set1_ps(x)),
-                          _mm_load_ps(m[3])))
+                        glmm_load(m[3])))
   ;
 #else
   vec4 v1;
@@ -146,10 +126,10 @@ CGLM_INLINE
 void
 glm_translate_y(mat4 m, float y) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(m[3],
-               _mm_add_ps(_mm_mul_ps(_mm_load_ps(m[1]),
+  glmm_store(m[3],
+             _mm_add_ps(_mm_mul_ps(glmm_load(m[1]),
                                    _mm_set1_ps(y)),
-                          _mm_load_ps(m[3])))
+                        glmm_load(m[3])))
   ;
 #else
   vec4 v1;
@@ -168,10 +148,10 @@ CGLM_INLINE
 void
 glm_translate_z(mat4 m, float z) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(m[3],
-               _mm_add_ps(_mm_mul_ps(_mm_load_ps(m[2]),
+  glmm_store(m[3],
+             _mm_add_ps(_mm_mul_ps(glmm_load(m[2]),
                                    _mm_set1_ps(z)),
-                          _mm_load_ps(m[3])))
+                        glmm_load(m[3])))
   ;
 #else
   vec4 v1;
@@ -189,8 +169,8 @@ glm_translate_z(mat4 m, float z) {
 CGLM_INLINE
 void
 glm_translate_make(mat4 m, vec3 v) {
-  mat4 t = GLM_MAT4_IDENTITY_INIT;
-  glm_translate_to(t, v, m);
+  glm_mat4_identity(m);
+  glm_vec_copy(v, m[3]);
 }
 
 /*!
@@ -220,8 +200,10 @@ glm_scale_to(mat4 m, vec3 v, mat4 dest) {
 CGLM_INLINE
 void
 glm_scale_make(mat4 m, vec3 v) {
-  mat4 t = GLM_MAT4_IDENTITY_INIT;
-  glm_scale_to(t, v, m);
+  glm_mat4_identity(m);
+  m[0][0] = v[0];
+  m[1][1] = v[1];
+  m[2][2] = v[2];
 }
 
 /*!
@@ -237,16 +219,6 @@ glm_scale(mat4 m, vec3 v) {
   glm_scale_to(m, v, m);
 }
 
-/*!
- * @brief DEPRECATED! Use glm_scale_uni
- */
-CGLM_INLINE
-void
-glm_scale1(mat4 m, float s) {
-  vec3 v = { s, s, s };
-  glm_scale_to(m, v, m);
-}
-
 /*!
  * @brief applies uniform scale to existing transform matrix v = [s, s, s]
  *        and stores result in same matrix
@@ -257,7 +229,7 @@ glm_scale1(mat4 m, float s) {
 CGLM_INLINE
 void
 glm_scale_uni(mat4 m, float s) {
-  vec3 v = { s, s, s };
+  CGLM_ALIGN(8) vec3 v = { s, s, s };
   glm_scale_to(m, v, m);
 }
 
@@ -272,19 +244,18 @@ glm_scale_uni(mat4 m, float s) {
 CGLM_INLINE
 void
 glm_rotate_x(mat4 m, float angle, mat4 dest) {
-  float cosVal;
-  float sinVal;
-  mat4  t = GLM_MAT4_IDENTITY_INIT;
+  CGLM_ALIGN(16) mat4 t = GLM_MAT4_IDENTITY_INIT;
+  float c, s;
 
-  cosVal = cosf(angle);
-  sinVal = sinf(angle);
+  c = cosf(angle);
+  s = sinf(angle);
 
-  t[1][1] =  cosVal;
-  t[1][2] =  sinVal;
-  t[2][1] = -sinVal;
-  t[2][2] =  cosVal;
+  t[1][1] =  c;
+  t[1][2] =  s;
+  t[2][1] = -s;
+  t[2][2] =  c;
 
-  glm_mat4_mul(m, t, dest);
+  glm_mul_rot(m, t, dest);
 }
 
 /*!
@@ -298,19 +269,18 @@ glm_rotate_x(mat4 m, float angle, mat4 dest) {
 CGLM_INLINE
 void
 glm_rotate_y(mat4 m, float angle, mat4 dest) {
-  float cosVal;
-  float sinVal;
-  mat4  t = GLM_MAT4_IDENTITY_INIT;
+  CGLM_ALIGN(16) mat4 t = GLM_MAT4_IDENTITY_INIT;
+  float c, s;
 
-  cosVal = cosf(angle);
-  sinVal = sinf(angle);
+  c = cosf(angle);
+  s = sinf(angle);
 
-  t[0][0] =  cosVal;
-  t[0][2] = -sinVal;
-  t[2][0] =  sinVal;
-  t[2][2] =  cosVal;
+  t[0][0] =  c;
+  t[0][2] = -s;
+  t[2][0] =  s;
+  t[2][2] =  c;
 
-  glm_mat4_mul(m, t, dest);
+  glm_mul_rot(m, t, dest);
 }
 
 /*!
@@ -324,61 +294,18 @@ glm_rotate_y(mat4 m, float angle, mat4 dest) {
 CGLM_INLINE
 void
 glm_rotate_z(mat4 m, float angle, mat4 dest) {
-  float cosVal;
-  float sinVal;
-  mat4  t = GLM_MAT4_IDENTITY_INIT;
-
-  cosVal = cosf(angle);
-  sinVal = sinf(angle);
-
-  t[0][0] =  cosVal;
-  t[0][1] =  sinVal;
-  t[1][0] = -sinVal;
-  t[1][1] =  cosVal;
-
-  glm_mat4_mul(m, t, dest);
-}
-
-/*!
- * @brief creates NEW rotation matrix by angle and axis
- *
- * this name may change in the future. axis must be is normalized
- *
- * @param[out] m        affine transfrom
- * @param[in]  angle    angle (radians)
- * @param[in]  axis_ndc normalized axis
- */
-CGLM_INLINE
-void
-glm_rotate_ndc_make(mat4 m, float angle, vec3 axis_ndc) {
-  /* https://www.opengl.org/sdk/docs/man2/xhtml/glRotate.xml */
-
-  vec3 v, vs;
-  float c;
+  CGLM_ALIGN(16) mat4 t = GLM_MAT4_IDENTITY_INIT;
+  float c, s;
 
   c = cosf(angle);
+  s = sinf(angle);
 
-  glm_vec_scale(axis_ndc, 1.0f - c, v);
-  glm_vec_scale(axis_ndc, sinf(angle), vs);
+  t[0][0] =  c;
+  t[0][1] =  s;
+  t[1][0] = -s;
+  t[1][1] =  c;
 
-  glm_vec_scale(axis_ndc, v[0], m[0]);
-  glm_vec_scale(axis_ndc, v[1], m[1]);
-  glm_vec_scale(axis_ndc, v[2], m[2]);
-
-  m[0][0] += c;
-  m[0][1] += vs[2];
-  m[0][2] -= vs[1];
-
-  m[1][0] -= vs[2];
-  m[1][1] += c;
-  m[1][2] += vs[0];
-
-  m[2][0] += vs[1];
-  m[2][1] -= vs[0];
-  m[2][2] += c;
-
-  m[0][3] = m[1][3] = m[2][3] = m[3][0] = m[3][1] = m[3][2] = 0.0f;
-  m[3][3] = 1.0f;
+  glm_mul_rot(m, t, dest);
 }
 
 /*!
@@ -393,53 +320,29 @@ glm_rotate_ndc_make(mat4 m, float angle, vec3 axis_ndc) {
 CGLM_INLINE
 void
 glm_rotate_make(mat4 m, float angle, vec3 axis) {
-  vec3 axis_ndc;
+  CGLM_ALIGN(8) vec3 axisn, v, vs;
+  float c;
 
-  glm_vec_normalize_to(axis, axis_ndc);
-  glm_rotate_ndc_make(m, angle, axis_ndc);
+  c = cosf(angle);
+
+  glm_vec_normalize_to(axis, axisn);
+  glm_vec_scale(axisn, 1.0f - c, v);
+  glm_vec_scale(axisn, sinf(angle), vs);
+
+  glm_vec_scale(axisn, v[0], m[0]);
+  glm_vec_scale(axisn, v[1], m[1]);
+  glm_vec_scale(axisn, v[2], m[2]);
+
+  m[0][0] += c;       m[1][0] -= vs[2];   m[2][0] += vs[1];
+  m[0][1] += vs[2];   m[1][1] += c;       m[2][1] -= vs[0];
+  m[0][2] -= vs[1];   m[1][2] += vs[0];   m[2][2] += c;
+
+  m[0][3] = m[1][3] = m[2][3] = m[3][0] = m[3][1] = m[3][2] = 0.0f;
+  m[3][3] = 1.0f;
 }
 
 /*!
- * @brief rotate existing transform matrix around Z axis by angle and axis
- *
- * this name may change in the future, axis must be normalized.
- *
- * @param[in, out]  m         affine transfrom
- * @param[in]       angle     angle (radians)
- * @param[in]       axis_ndc  normalized axis
- */
-CGLM_INLINE
-void
-glm_rotate_ndc(mat4 m, float angle, vec3 axis_ndc) {
-  mat4 rot, tmp;
-
-  glm_rotate_ndc_make(rot, angle, axis_ndc);
-
-  glm_vec4_scale(m[0], rot[0][0], tmp[1]);
-  glm_vec4_scale(m[1], rot[0][1], tmp[0]);
-  glm_vec4_add(tmp[1], tmp[0],    tmp[1]);
-  glm_vec4_scale(m[2], rot[0][2], tmp[0]);
-  glm_vec4_add(tmp[1], tmp[0],    tmp[1]);
-
-  glm_vec4_scale(m[0], rot[1][0], tmp[2]);
-  glm_vec4_scale(m[1], rot[1][1], tmp[0]);
-  glm_vec4_add(tmp[2], tmp[0],    tmp[2]);
-  glm_vec4_scale(m[2], rot[1][2], tmp[0]);
-  glm_vec4_add(tmp[2], tmp[0],    tmp[2]);
-
-  glm_vec4_scale(m[0], rot[2][0], tmp[3]);
-  glm_vec4_scale(m[1], rot[2][1], tmp[0]);
-  glm_vec4_add(tmp[3], tmp[0],    tmp[3]);
-  glm_vec4_scale(m[2], rot[2][2], tmp[0]);
-  glm_vec4_add(tmp[3], tmp[0],    tmp[3]);
-
-  glm_vec4_copy(tmp[1], m[0]);
-  glm_vec4_copy(tmp[2], m[1]);
-  glm_vec4_copy(tmp[3], m[2]);
-}
-
-/*!
- * @brief rotate existing transform matrix around Z axis by angle and axis
+ * @brief rotate existing transform matrix around given axis by angle
  *
  * @param[in, out]  m      affine transfrom
  * @param[in]       angle  angle (radians)
@@ -448,10 +351,55 @@ glm_rotate_ndc(mat4 m, float angle, vec3 axis_ndc) {
 CGLM_INLINE
 void
 glm_rotate(mat4 m, float angle, vec3 axis) {
-  vec3 axis_ndc;
+  CGLM_ALIGN(16) mat4 rot;
+  glm_rotate_make(rot, angle, axis);
+  glm_mul_rot(m, rot, m);
+}
 
-  glm_vec_normalize_to(axis, axis_ndc);
-  glm_rotate_ndc(m, angle, axis_ndc);
+/*!
+ * @brief rotate existing transform
+ *        around given axis by angle at given pivot point (rotation center)
+ *
+ * @param[in, out]  m      affine transfrom
+ * @param[in]       pivot  rotation center
+ * @param[in]       angle  angle (radians)
+ * @param[in]       axis   axis
+ */
+CGLM_INLINE
+void
+glm_rotate_at(mat4 m, vec3 pivot, float angle, vec3 axis) {
+  CGLM_ALIGN(8) vec3 pivotInv;
+
+  glm_vec_inv_to(pivot, pivotInv);
+
+  glm_translate(m, pivot);
+  glm_rotate(m, angle, axis);
+  glm_translate(m, pivotInv);
+}
+
+/*!
+ * @brief creates NEW rotation matrix by angle and axis at given point
+ *
+ * this creates rotation matrix, it assumes you don't have a matrix
+ *
+ * this should work faster than glm_rotate_at because it reduces
+ * one glm_translate.
+ *
+ * @param[out] m      affine transfrom
+ * @param[in]  pivot  rotation center
+ * @param[in]  angle  angle (radians)
+ * @param[in]  axis   axis
+ */
+CGLM_INLINE
+void
+glm_rotate_atm(mat4 m, vec3 pivot, float angle, vec3 axis) {
+  CGLM_ALIGN(8) vec3 pivotInv;
+
+  glm_vec_inv_to(pivot, pivotInv);
+
+  glm_translate_make(m, pivot);
+  glm_rotate(m, angle, axis);
+  glm_translate(m, pivotInv);
 }
 
 /*!
@@ -479,9 +427,8 @@ glm_decompose_scalev(mat4 m, vec3 s) {
 CGLM_INLINE
 bool
 glm_uniscaled(mat4 m) {
-  vec3 s;
+  CGLM_ALIGN(8) vec3 s;
   glm_decompose_scalev(m, s);
-
   return glm_vec_eq_all(s);
 }
 
@@ -496,8 +443,8 @@ glm_uniscaled(mat4 m) {
 CGLM_INLINE
 void
 glm_decompose_rs(mat4 m, mat4 r, vec3 s) {
-  vec4 t = {0.0f, 0.0f, 0.0f, 1.0f};
-  vec3 v;
+  CGLM_ALIGN(16) vec4 t = {0.0f, 0.0f, 0.0f, 1.0f};
+  CGLM_ALIGN(8)  vec3 v;
 
   glm_vec4_copy(m[0], r[0]);
   glm_vec4_copy(m[1], r[1]);

include/cglm/box.h

@@ -11,6 +11,7 @@
 #include "common.h"
 #include "vec3.h"
 #include "vec4.h"
+#include "util.h"
 
 /*!
  * @brief apply transform to Axis-Aligned Bounding Box
@@ -153,4 +154,130 @@ glm_aabb_frustum(vec3 box[2], vec4 planes[6]) {
   return true;
 }
 
+/*!
+ * @brief invalidate AABB min and max values
+ *
+ * @param[in, out]  box bounding box
+ */
+CGLM_INLINE
+void
+glm_aabb_invalidate(vec3 box[2]) {
+  glm_vec_broadcast(FLT_MAX,  box[0]);
+  glm_vec_broadcast(-FLT_MAX, box[1]);
+}
+
+/*!
+ * @brief check if AABB is valid or not
+ *
+ * @param[in]  box bounding box
+ */
+CGLM_INLINE
+bool
+glm_aabb_isvalid(vec3 box[2]) {
+  return glm_vec_max(box[0]) != FLT_MAX
+         && glm_vec_min(box[1]) != -FLT_MAX;
+}
+
+/*!
+ * @brief distance between of min and max
+ *
+ * @param[in]  box bounding box
+ */
+CGLM_INLINE
+float
+glm_aabb_size(vec3 box[2]) {
+  return glm_vec_distance(box[0], box[1]);
+}
+
+/*!
+ * @brief radius of sphere which surrounds AABB
+ *
+ * @param[in]  box bounding box
+ */
+CGLM_INLINE
+float
+glm_aabb_radius(vec3 box[2]) {
+  return glm_aabb_size(box) * 0.5f;
+}
+
+/*!
+ * @brief computes center point of AABB
+ *
+ * @param[in]   box  bounding box
+ * @param[out]  dest center of bounding box
+ */
+CGLM_INLINE
+void
+glm_aabb_center(vec3 box[2], vec3 dest) {
+  glm_vec_center(box[0], box[1], dest);
+}
+
+/*!
+ * @brief check if two AABB intersects
+ *
+ * @param[in]   box    bounding box
+ * @param[in]   other  other bounding box
+ */
+CGLM_INLINE
+bool
+glm_aabb_aabb(vec3 box[2], vec3 other[2]) {
+  return (box[0][0] <= other[1][0] && box[1][0] >= other[0][0])
+      && (box[0][1] <= other[1][1] && box[1][1] >= other[0][1])
+      && (box[0][2] <= other[1][2] && box[1][2] >= other[0][2]);
+}
+
+/*!
+ * @brief check if AABB intersects with sphere
+ *
+ * https://github.com/erich666/GraphicsGems/blob/master/gems/BoxSphere.c
+ * Solid Box - Solid Sphere test.
+ *
+ * @param[in]   box    solid bounding box
+ * @param[in]   s      solid sphere
+ */
+CGLM_INLINE
+bool
+glm_aabb_sphere(vec3 box[2], vec4 s) {
+  float dmin;
+  int   a, b, c;
+
+  a = s[0] >= box[0][0];
+  b = s[1] >= box[0][1];
+  c = s[2] >= box[0][2];
+
+  dmin  = glm_pow2(s[0] - box[a][0])
+        + glm_pow2(s[1] - box[b][1])
+        + glm_pow2(s[2] - box[c][2]);
+
+  return dmin <= glm_pow2(s[3]);
+}
+
+/*!
+ * @brief check if point is inside of AABB
+ *
+ * @param[in]   box    bounding box
+ * @param[in]   point  point
+ */
+CGLM_INLINE
+bool
+glm_aabb_point(vec3 box[2], vec3 point) {
+  return (point[0] >= box[0][0] && point[0] <= box[1][0])
+      && (point[1] >= box[0][1] && point[1] <= box[1][1])
+      && (point[2] >= box[0][2] && point[2] <= box[1][2]);
+}
+
+/*!
+ * @brief check if AABB contains other AABB
+ *
+ * @param[in]   box    bounding box
+ * @param[in]   other  other bounding box
+ */
+CGLM_INLINE
+bool
+glm_aabb_contains(vec3 box[2], vec3 other[2]) {
+  return (box[0][0] <= other[0][0] && box[1][0] >= other[1][0])
+      && (box[0][1] <= other[0][1] && box[1][1] >= other[1][1])
+      && (box[0][2] <= other[0][2] && box[1][2] >= other[1][2]);
+}
+
 #endif /* cglm_box_h */

include/cglm/call.h

@@ -24,6 +24,8 @@ extern "C" {
 #include "call/frustum.h"
 #include "call/box.h"
 #include "call/io.h"
+#include "call/project.h"
+#include "call/sphere.h"
 
 #ifdef __cplusplus
 }

include/cglm/call/affine.h

@@ -13,6 +13,10 @@ extern "C" {
 
 #include "../cglm.h"
 
+CGLM_EXPORT
+void
+glmc_translate_make(mat4 m, vec3 v);
+
 CGLM_EXPORT
 void
 glmc_translate_to(mat4 m, vec3 v, mat4 dest);
@@ -33,6 +37,10 @@ CGLM_EXPORT
 void
 glmc_translate_z(mat4 m, float to);
 
+CGLM_EXPORT
+void
+glmc_scale_make(mat4 m, vec3 v);
+
 CGLM_EXPORT
 void
 glmc_scale_to(mat4 m, vec3 v, mat4 dest);
@@ -43,7 +51,7 @@ glmc_scale(mat4 m, vec3 v);
 
 CGLM_EXPORT
 void
-glmc_scale1(mat4 m, float s);
+glmc_scale_uni(mat4 m, float s);
 
 CGLM_EXPORT
 void
@@ -57,26 +65,30 @@ CGLM_EXPORT
 void
 glmc_rotate_z(mat4 m, float rad, mat4 dest);
 
-CGLM_EXPORT
-void
-glmc_rotate_ndc_make(mat4 m, float angle, vec3 axis_ndc);
-
 CGLM_EXPORT
 void
 glmc_rotate_make(mat4 m, float angle, vec3 axis);
 
-CGLM_EXPORT
-void
-glmc_rotate_ndc(mat4 m, float angle, vec3 axis_ndc);
-
 CGLM_EXPORT
 void
 glmc_rotate(mat4 m, float angle, vec3 axis);
 
+CGLM_EXPORT
+void
+glmc_rotate_at(mat4 m, vec3 pivot, float angle, vec3 axis);
+
+CGLM_EXPORT
+void
+glmc_rotate_atm(mat4 m, vec3 pivot, float angle, vec3 axis);
+
 CGLM_EXPORT
 void
 glmc_decompose_scalev(mat4 m, vec3 s);
 
+CGLM_EXPORT
+bool
+glmc_uniscaled(mat4 m);
+
 CGLM_EXPORT
 void
 glmc_decompose_rs(mat4 m, mat4 r, vec3 s);
@@ -85,6 +97,20 @@ CGLM_EXPORT
 void
 glmc_decompose(mat4 m, vec4 t, mat4 r, vec3 s);
 
+/* affine-mat */
+
+CGLM_EXPORT
+void
+glmc_mul(mat4 m1, mat4 m2, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_mul_rot(mat4 m1, mat4 m2, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_inv_tr(mat4 mat);
+
 #ifdef __cplusplus
 }
 #endif

include/cglm/call/box.h

@@ -32,6 +32,46 @@ glmc_aabb_crop_until(vec3 box[2],
                      vec3 clampBox[2],
                      vec3 dest[2]);
 
+CGLM_EXPORT
+bool
+glmc_aabb_frustum(vec3 box[2], vec4 planes[6]);
+
+CGLM_EXPORT
+void
+glmc_aabb_invalidate(vec3 box[2]);
+
+CGLM_EXPORT
+bool
+glmc_aabb_isvalid(vec3 box[2]);
+
+CGLM_EXPORT
+float
+glmc_aabb_size(vec3 box[2]);
+
+CGLM_EXPORT
+float
+glmc_aabb_radius(vec3 box[2]);
+
+CGLM_EXPORT
+void
+glmc_aabb_center(vec3 box[2], vec3 dest);
+
+CGLM_EXPORT
+bool
+glmc_aabb_aabb(vec3 box[2], vec3 other[2]);
+
+CGLM_EXPORT
+bool
+glmc_aabb_point(vec3 box[2], vec3 point);
+
+CGLM_EXPORT
+bool
+glmc_aabb_contains(vec3 box[2], vec3 other[2]);
+
+CGLM_EXPORT
+bool
+glmc_aabb_sphere(vec3 box[2], vec4 s);
+
 #ifdef __cplusplus
 }
 #endif

include/cglm/call/cam.h

@@ -33,6 +33,26 @@ glmc_ortho(float left,
            float farVal,
            mat4 dest);
 
+CGLM_EXPORT
+void
+glmc_ortho_aabb(vec3 box[2], mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_ortho_aabb_p(vec3 box[2], float padding, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_ortho_aabb_pz(vec3 box[2], float padding, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_ortho_default(float aspect, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_ortho_default_s(float aspect, float size, mat4 dest);
+
 CGLM_EXPORT
 void
 glmc_perspective(float fovy,
@@ -41,6 +61,14 @@ glmc_perspective(float fovy,
                  float farVal,
                  mat4 dest);
 
+CGLM_EXPORT
+void
+glmc_perspective_default(float aspect, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_perspective_resize(float aspect, mat4 proj);
+
 CGLM_EXPORT
 void
 glmc_lookat(vec3 eye, vec3 center, vec3 up, mat4 dest);
@@ -53,6 +81,58 @@ CGLM_EXPORT
 void
 glmc_look_anyup(vec3 eye, vec3 dir, mat4 dest);
 
+CGLM_EXPORT
+void
+glmc_persp_decomp(mat4 proj,
+                  float * __restrict nearVal,
+                  float * __restrict farVal,
+                  float * __restrict top,
+                  float * __restrict bottom,
+                  float * __restrict left,
+                  float * __restrict right);
+
+CGLM_EXPORT
+void
+glmc_persp_decompv(mat4 proj, float dest[6]);
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_x(mat4 proj,
+                    float * __restrict left,
+                    float * __restrict right);
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_y(mat4 proj,
+                    float * __restrict top,
+                    float * __restrict bottom);
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_z(mat4 proj,
+                    float * __restrict nearVal,
+                    float * __restrict farVal);
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_far(mat4 proj, float * __restrict farVal);
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_near(mat4 proj, float * __restrict nearVal);
+
+CGLM_EXPORT
+float
+glmc_persp_fovy(mat4 proj);
+
+CGLM_EXPORT
+float
+glmc_persp_aspect(mat4 proj);
+
+CGLM_EXPORT
+void
+glmc_persp_sizes(mat4 proj, float fovy, vec4 dest);
+
 #ifdef __cplusplus
 }
 #endif

include/cglm/call/euler.h

@@ -21,6 +21,10 @@ CGLM_EXPORT
 void
 glmc_euler(vec3 angles, mat4 dest);
 
+CGLM_EXPORT
+void
+glmc_euler_xyz(vec3 angles,  mat4 dest);
+
 CGLM_EXPORT
 void
 glmc_euler_zyx(vec3 angles,  mat4 dest);

include/cglm/call/mat3.h

@@ -40,6 +40,10 @@ CGLM_EXPORT
 void
 glmc_mat3_mulv(mat3 m, vec3 v, vec3 dest);
 
+CGLM_EXPORT
+void
+glmc_mat3_quat(mat3 m, versor dest);
+
 CGLM_EXPORT
 void
 glmc_mat3_scale(mat3 m, float s);

include/cglm/call/mat4.h

@@ -47,12 +47,20 @@ glmc_mat4_mul(mat4 m1, mat4 m2, mat4 dest);
 
 CGLM_EXPORT
 void
-glmc_mat4_mulN(mat4 * __restrict matrices[], int len, mat4 dest);
+glmc_mat4_mulN(mat4 * __restrict matrices[], uint32_t len, mat4 dest);
 
 CGLM_EXPORT
 void
 glmc_mat4_mulv(mat4 m, vec4 v, vec4 dest);
 
+CGLM_EXPORT
+void
+glmc_mat4_mulv3(mat4 m, vec3 v, float last, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_mat4_quat(mat4 m, versor dest);
+
 CGLM_EXPORT
 void
 glmc_mat4_transpose_to(mat4 m, mat4 dest);
@@ -81,6 +89,10 @@ CGLM_EXPORT
 void
 glmc_mat4_inv_precise(mat4 mat, mat4 dest);
 
+CGLM_EXPORT
+void
+glmc_mat4_inv_fast(mat4 mat, mat4 dest);
+
 CGLM_EXPORT
 void
 glmc_mat4_swap_col(mat4 mat, int col1, int col2);

include/cglm/call/project.h

@@ -0,0 +1,33 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#ifndef cglmc_project_h
+#define cglmc_project_h
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include "../cglm.h"
+
+CGLM_EXPORT
+void
+glmc_unprojecti(vec3 pos, mat4 invMat, vec4 vp, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_unproject(vec3 pos, mat4 m, vec4 vp, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_project(vec3 pos, mat4 m, vec4 vp, vec3 dest);
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* cglmc_project_h */
+
+

include/cglm/call/quat.h

@@ -19,33 +19,79 @@ glmc_quat_identity(versor q);
 
 CGLM_EXPORT
 void
-glmc_quat(versor q,
-          float angle,
-          float x,
-          float y,
-          float z);
+glmc_quat_init(versor q, float x, float y, float z, float w);
 
 CGLM_EXPORT
 void
-glmc_quatv(versor q,
-          float  angle,
-          vec3   v);
+glmc_quat(versor q, float angle, float x, float y, float z);
+
+CGLM_EXPORT
+void
+glmc_quatv(versor q, float angle, vec3 axis);
+
+CGLM_EXPORT
+void
+glmc_quat_copy(versor q, versor dest);
 
 CGLM_EXPORT
 float
 glmc_quat_norm(versor q);
 
+CGLM_EXPORT
+void
+glmc_quat_normalize_to(versor q, versor dest);
+
 CGLM_EXPORT
 void
 glmc_quat_normalize(versor q);
 
 CGLM_EXPORT
 float
-glmc_quat_dot(versor q, versor r);
+glmc_quat_dot(versor p, versor q);
 
 CGLM_EXPORT
 void
-glmc_quat_mulv(versor q1, versor q2, versor dest);
+glmc_quat_conjugate(versor q, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_inv(versor q, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_add(versor p, versor q, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_sub(versor p, versor q, versor dest);
+
+CGLM_EXPORT
+float
+glmc_quat_real(versor q);
+
+CGLM_EXPORT
+void
+glmc_quat_imag(versor q, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_imagn(versor q, vec3 dest);
+
+CGLM_EXPORT
+float
+glmc_quat_imaglen(versor q);
+
+CGLM_EXPORT
+float
+glmc_quat_angle(versor q);
+
+CGLM_EXPORT
+void
+glmc_quat_axis(versor q, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_mul(versor p, versor q, versor dest);
 
 CGLM_EXPORT
 void
@@ -53,10 +99,51 @@ glmc_quat_mat4(versor q, mat4 dest);
 
 CGLM_EXPORT
 void
-glmc_quat_slerp(versor q,
-                versor r,
-                float  t,
-                versor dest);
+glmc_quat_mat4t(versor q, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_mat3(versor q, mat3 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_mat3t(versor q, mat3 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_lerp(versor from, versor to, float t, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_slerp(versor q, versor r, float t, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_look(vec3 eye, versor ori, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_for(vec3 dir, vec3 fwd, vec3 up, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_forp(vec3 from, vec3 to, vec3 fwd, vec3 up, versor dest);
+
+CGLM_EXPORT
+void
+glmc_quat_rotatev(versor from, vec3 to, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_rotate(mat4 m, versor q, mat4 dest);
+
+CGLM_EXPORT
+void
+glmc_quat_rotate_at(mat4 model, versor q, vec3 pivot);
+
+CGLM_EXPORT
+void
+glmc_quat_rotate_atm(mat4 m, versor q, vec3 pivot);
 
 #ifdef __cplusplus
 }

include/cglm/call/sphere.h

@@ -0,0 +1,36 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#ifndef cglmc_sphere_h
+#define cglmc_sphere_h
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include "../cglm.h"
+
+CGLM_EXPORT
+float
+glmc_sphere_radii(vec4 s);
+
+CGLM_EXPORT
+void
+glmc_sphere_transform(vec4 s, mat4 m, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_sphere_merge(vec4 s1, vec4 s2, vec4 dest);
+
+CGLM_EXPORT
+bool
+glmc_sphere_sphere(vec4 s1, vec4 s2);
+
+CGLM_EXPORT
+bool
+glmc_sphere_point(vec4 s, vec3 point);
+
+#endif /* cglmc_sphere_h */

include/cglm/call/vec3.h

@@ -16,10 +16,22 @@ extern "C" {
 /* DEPRECATED! use _copy, _ucopy versions */
 #define glmc_vec_dup(v, dest) glmc_vec_copy(v, dest)
 
+CGLM_EXPORT
+void
+glmc_vec3(vec4 v4, vec3 dest);
+
 CGLM_EXPORT
 void
 glmc_vec_copy(vec3 a, vec3 dest);
 
+CGLM_EXPORT
+void
+glmc_vec_zero(vec3 v);
+
+CGLM_EXPORT
+void
+glmc_vec_one(vec3 v);
+
 CGLM_EXPORT
 float
 glmc_vec_dot(vec3 a, vec3 b);
@@ -50,7 +62,19 @@ glmc_vec_add(vec3 v1, vec3 v2, vec3 dest);
 
 CGLM_EXPORT
 void
-glmc_vec_sub(vec3 v1, vec3 v2, vec3 dest);
+glmc_vec_adds(vec3 v, float s, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_sub(vec3 a, vec3 b, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_subs(vec3 v, float s, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_mul(vec3 a, vec3 b, vec3 d);
 
 CGLM_EXPORT
 void
@@ -60,10 +84,38 @@ CGLM_EXPORT
 void
 glmc_vec_scale_as(vec3 v, float s, vec3 dest);
 
+CGLM_EXPORT
+void
+glmc_vec_div(vec3 a, vec3 b, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_divs(vec3 a, float s, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_addadd(vec3 a, vec3 b, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_subadd(vec3 a, vec3 b, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_muladd(vec3 a, vec3 b, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_muladds(vec3 a, float s, vec3 dest);
+
 CGLM_EXPORT
 void
 glmc_vec_flipsign(vec3 v);
 
+CGLM_EXPORT
+void
+glmc_vec_flipsign_to(vec3 v, vec3 dest);
+
 CGLM_EXPORT
 void
 glmc_vec_inv(vec3 v);
@@ -84,6 +136,10 @@ CGLM_EXPORT
 void
 glmc_vec_rotate_m4(mat4 m, vec3 v, vec3 dest);
 
+CGLM_EXPORT
+void
+glmc_vec_rotate_m3(mat3 m, vec3 v, vec3 dest);
+
 CGLM_EXPORT
 void
 glmc_vec_proj(vec3 a, vec3 b, vec3 dest);
@@ -92,6 +148,10 @@ CGLM_EXPORT
 void
 glmc_vec_center(vec3 v1, vec3 v2, vec3 dest);
 
+CGLM_EXPORT
+float
+glmc_vec_distance2(vec3 v1, vec3 v2);
+
 CGLM_EXPORT
 float
 glmc_vec_distance(vec3 v1, vec3 v2);
@@ -104,6 +164,76 @@ CGLM_EXPORT
 void
 glmc_vec_minv(vec3 v1, vec3 v2, vec3 dest);
 
+CGLM_EXPORT
+void
+glmc_vec_clamp(vec3 v, float minVal, float maxVal);
+
+CGLM_EXPORT
+void
+glmc_vec_ortho(vec3 v, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_lerp(vec3 from, vec3 to, float t, vec3 dest);
+
+/* ext */
+
+CGLM_EXPORT
+void
+glmc_vec_mulv(vec3 a, vec3 b, vec3 d);
+
+CGLM_EXPORT
+void
+glmc_vec_broadcast(float val, vec3 d);
+
+CGLM_EXPORT
+bool
+glmc_vec_eq(vec3 v, float val);
+
+CGLM_EXPORT
+bool
+glmc_vec_eq_eps(vec3 v, float val);
+
+CGLM_EXPORT
+bool
+glmc_vec_eq_all(vec3 v);
+
+CGLM_EXPORT
+bool
+glmc_vec_eqv(vec3 v1, vec3 v2);
+
+CGLM_EXPORT
+bool
+glmc_vec_eqv_eps(vec3 v1, vec3 v2);
+
+CGLM_EXPORT
+float
+glmc_vec_max(vec3 v);
+
+CGLM_EXPORT
+float
+glmc_vec_min(vec3 v);
+
+CGLM_EXPORT
+bool
+glmc_vec_isnan(vec3 v);
+
+CGLM_EXPORT
+bool
+glmc_vec_isinf(vec3 v);
+
+CGLM_EXPORT
+bool
+glmc_vec_isvalid(vec3 v);
+
+CGLM_EXPORT
+void
+glmc_vec_sign(vec3 v, vec3 dest);
+
+CGLM_EXPORT
+void
+glmc_vec_sqrt(vec3 v, vec3 dest);
+
 #ifdef __cplusplus
 }
 #endif

include/cglm/call/vec4.h

@@ -17,6 +17,18 @@ extern "C" {
 #define glmc_vec4_dup3(v, dest) glmc_vec4_copy3(v, dest)
 #define glmc_vec4_dup(v, dest)  glmc_vec4_copy(v, dest)
 
+CGLM_EXPORT
+void
+glmc_vec4(vec3 v3, float last, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_zero(vec4 v);
+
+CGLM_EXPORT
+void
+glmc_vec4_one(vec4 v);
+
 CGLM_EXPORT
 void
 glmc_vec4_copy3(vec4 a, vec3 dest);
@@ -47,11 +59,23 @@ glmc_vec4_normalize(vec4 v);
 
 CGLM_EXPORT
 void
-glmc_vec4_add(vec4 v1, vec4 v2, vec4 dest);
+glmc_vec4_add(vec4 a, vec4 b, vec4 dest);
 
 CGLM_EXPORT
 void
-glmc_vec4_sub(vec4 v1, vec4 v2, vec4 dest);
+glmc_vec4_adds(vec4 v, float s, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_sub(vec4 a, vec4 b, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_subs(vec4 v, float s, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_mul(vec4 a, vec4 b, vec4 d);
 
 CGLM_EXPORT
 void
@@ -61,10 +85,38 @@ CGLM_EXPORT
 void
 glmc_vec4_scale_as(vec3 v, float s, vec3 dest);
 
+CGLM_EXPORT
+void
+glmc_vec4_div(vec4 a, vec4 b, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_divs(vec4 v, float s, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_addadd(vec4 a, vec4 b, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_subadd(vec4 a, vec4 b, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_muladd(vec4 a, vec4 b, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_muladds(vec4 a, float s, vec4 dest);
+
 CGLM_EXPORT
 void
 glmc_vec4_flipsign(vec4 v);
 
+CGLM_EXPORT
+void
+glmc_vec4_flipsign_to(vec4 v, vec4 dest);
+
 CGLM_EXPORT
 void
 glmc_vec4_inv(vec4 v);
@@ -85,6 +137,72 @@ CGLM_EXPORT
 void
 glmc_vec4_minv(vec4 v1, vec4 v2, vec4 dest);
 
+CGLM_EXPORT
+void
+glmc_vec4_clamp(vec4 v, float minVal, float maxVal);
+
+CGLM_EXPORT
+void
+glmc_vec4_lerp(vec4 from, vec4 to, float t, vec4 dest);
+
+/* ext */
+
+CGLM_EXPORT
+void
+glmc_vec4_mulv(vec4 a, vec4 b, vec4 d);
+
+CGLM_EXPORT
+void
+glmc_vec4_broadcast(float val, vec4 d);
+
+CGLM_EXPORT
+bool
+glmc_vec4_eq(vec4 v, float val);
+
+CGLM_EXPORT
+bool
+glmc_vec4_eq_eps(vec4 v, float val);
+
+CGLM_EXPORT
+bool
+glmc_vec4_eq_all(vec4 v);
+
+CGLM_EXPORT
+bool
+glmc_vec4_eqv(vec4 v1, vec4 v2);
+
+CGLM_EXPORT
+bool
+glmc_vec4_eqv_eps(vec4 v1, vec4 v2);
+
+CGLM_EXPORT
+float
+glmc_vec4_max(vec4 v);
+
+CGLM_EXPORT
+float
+glmc_vec4_min(vec4 v);
+
+CGLM_EXPORT
+bool
+glmc_vec4_isnan(vec4 v);
+
+CGLM_EXPORT
+bool
+glmc_vec4_isinf(vec4 v);
+
+CGLM_EXPORT
+bool
+glmc_vec4_isvalid(vec4 v);
+
+CGLM_EXPORT
+void
+glmc_vec4_sign(vec4 v, vec4 dest);
+
+CGLM_EXPORT
+void
+glmc_vec4_sqrt(vec4 v, vec4 dest);
+
 #ifdef __cplusplus
 }
 #endif

include/cglm/cam.h

@@ -321,6 +321,9 @@ glm_perspective_resize(float aspect,
 /*!
  * @brief set up view matrix
  *
+ * NOTE: The UP vector must not be parallel to the line of sight from
+ *       the eye point to the reference point
+ *
  * @param[in]  eye    eye vector
  * @param[in]  center center vector
  * @param[in]  up     up vector
@@ -332,7 +335,7 @@ glm_lookat(vec3 eye,
            vec3 center,
            vec3 up,
            mat4 dest) {
-  vec3 f, u, s;
+  CGLM_ALIGN(8) vec3 f, u, s;
 
   glm_vec_sub(center, eye, f);
   glm_vec_normalize(f);
@@ -364,6 +367,9 @@ glm_lookat(vec3 eye,
  * convenient wrapper for lookat: if you only have direction not target self
  * then this might be useful. Because you need to get target from direction.
  *
+ * NOTE: The UP vector must not be parallel to the line of sight from
+ *       the eye point to the reference point
+ *
  * @param[in]  eye    eye vector
  * @param[in]  dir    direction vector
  * @param[in]  up     up vector
@@ -372,7 +378,7 @@ glm_lookat(vec3 eye,
 CGLM_INLINE
 void
 glm_look(vec3 eye, vec3 dir, vec3 up, mat4 dest) {
-  vec3 target;
+  CGLM_ALIGN(8) vec3 target;
   glm_vec_add(eye, dir, target);
   glm_lookat(eye, target, up, dest);
 }
@@ -390,7 +396,7 @@ glm_look(vec3 eye, vec3 dir, vec3 up, mat4 dest) {
 CGLM_INLINE
 void
 glm_look_anyup(vec3 eye, vec3 dir, mat4 dest) {
-  vec3 up;
+  CGLM_ALIGN(8) vec3 up;
   glm_vec_ortho(dir, up);
   glm_look(eye, dir, up, dest);
 }

include/cglm/cglm.h

@@ -23,5 +23,7 @@
 #include "color.h"
 #include "util.h"
 #include "io.h"
+#include "project.h"
+#include "sphere.h"
 
 #endif /* cglm_h */

include/cglm/common.h

@@ -14,7 +14,7 @@
 #include <math.h>
 #include <float.h>
 
-#if defined(_WIN32)
+#if defined(_MSC_VER)
 #  ifdef CGLM_DLL
 #    define CGLM_EXPORT __declspec(dllexport)
 #  else

include/cglm/euler.h

@@ -5,6 +5,14 @@
  * Full license can be found in the LICENSE file
  */
 
+/*
+ NOTE:
+  angles must be passed as [X-Angle, Y-Angle, Z-angle] order
+  For instance you don't pass angles as [Z-Angle, X-Angle, Y-angle] to
+  glm_euler_zxy funciton, All RELATED functions accept angles same order
+  which is [X, Y, Z].
+ */
+
 /*
  Types:
    enum glm_euler_sq
@@ -13,6 +21,7 @@
    CGLM_INLINE glm_euler_sq glm_euler_order(int newOrder[3]);
    CGLM_INLINE void glm_euler_angles(mat4 m, vec3 dest);
    CGLM_INLINE void glm_euler(vec3 angles, mat4 dest);
+   CGLM_INLINE void glm_euler_xyz(vec3 angles, mat4 dest);
    CGLM_INLINE void glm_euler_zyx(vec3 angles, mat4 dest);
    CGLM_INLINE void glm_euler_zxy(vec3 angles, mat4 dest);
    CGLM_INLINE void glm_euler_xzy(vec3 angles, mat4 dest);
@@ -61,205 +70,115 @@ glm_euler_order(int ord[3]) {
 CGLM_INLINE
 void
 glm_euler_angles(mat4 m, vec3 dest) {
-  if (m[0][2] < 1.0f) {
-    if (m[0][2] > -1.0f) {
-      vec3  a[2];
-      float cy1, cy2;
-      int   path;
+  float m00, m01, m10, m11, m20, m21, m22;
+  float thetaX, thetaY, thetaZ;
 
-      a[0][1] = asinf(-m[0][2]);
-      a[1][1] = CGLM_PI - a[0][1];
+  m00 = m[0][0];  m10 = m[1][0];  m20 = m[2][0];
+  m01 = m[0][1];  m11 = m[1][1];  m21 = m[2][1];
+                                  m22 = m[2][2];
 
-      cy1 = cosf(a[0][1]);
-      cy2 = cosf(a[1][1]);
-
-      a[0][0] = atan2f(m[1][2] / cy1, m[2][2] / cy1);
-      a[1][0] = atan2f(m[1][2] / cy2, m[2][2] / cy2);
-
-      a[0][2] = atan2f(m[0][1] / cy1, m[0][0] / cy1);
-      a[1][2] = atan2f(m[0][1] / cy2, m[0][0] / cy2);
-
-      path = (fabsf(a[0][0]) + fabsf(a[0][1]) + fabsf(a[0][2])) >=
-               (fabsf(a[1][0]) + fabsf(a[1][1]) + fabsf(a[1][2]));
-
-      glm_vec_copy(a[path], dest);
-    } else {
-      dest[0] = atan2f(m[1][0], m[2][0]);
-      dest[1] = CGLM_PI_2;
-      dest[2] = 0.0f;
+  if (m20 < 1.0f) {
+    if (m20 > -1.0f) {
+      thetaY = asinf(m20);
+      thetaX = atan2f(-m21, m22);
+      thetaZ = atan2f(-m10, m00);
+    } else { /* m20 == -1 */
+      /* Not a unique solution */
+      thetaY = -CGLM_PI_2;
+      thetaX = -atan2f(m01, m11);
+      thetaZ =  0.0f;
     }
-  } else {
-    dest[0] = atan2f(-m[1][0], -m[2][0]);
-    dest[1] =-CGLM_PI_2;
-    dest[2] = 0.0f;
+  } else { /* m20 == +1 */
+    thetaY = CGLM_PI_2;
+    thetaX = atan2f(m01, m11);
+    thetaZ = 0.0f;
   }
+
+  dest[0] = thetaX;
+  dest[1] = thetaY;
+  dest[2] = thetaZ;
 }
 
 /*!
  * @brief build rotation matrix from euler angles
  *
- * @param[in]  angles angles as vector [Ex, Ey, Ez]
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
+ * @param[out] dest   rotation matrix
+ */
+CGLM_INLINE
+void
+glm_euler_xyz(vec3 angles, mat4 dest) {
+  float cx, cy, cz,
+        sx, sy, sz, czsx, cxcz, sysz;
+
+  sx   = sinf(angles[0]); cx = cosf(angles[0]);
+  sy   = sinf(angles[1]); cy = cosf(angles[1]);
+  sz   = sinf(angles[2]); cz = cosf(angles[2]);
+
+  czsx = cz * sx;
+  cxcz = cx * cz;
+  sysz = sy * sz;
+
+  dest[0][0] =  cy * cz;
+  dest[0][1] =  czsx * sy + cx * sz;
+  dest[0][2] = -cxcz * sy + sx * sz;
+  dest[1][0] = -cy * sz;
+  dest[1][1] =  cxcz - sx * sysz;
+  dest[1][2] =  czsx + cx * sysz;
+  dest[2][0] =  sy;
+  dest[2][1] = -cy * sx;
+  dest[2][2] =  cx * cy;
+  dest[0][3] =  0.0f;
+  dest[1][3] =  0.0f;
+  dest[2][3] =  0.0f;
+  dest[3][0] =  0.0f;
+  dest[3][1] =  0.0f;
+  dest[3][2] =  0.0f;
+  dest[3][3] =  1.0f;
+}
+
+/*!
+ * @brief build rotation matrix from euler angles
+ *
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
  * @param[out] dest   rotation matrix
  */
 CGLM_INLINE
 void
 glm_euler(vec3 angles, mat4 dest) {
-  float cx, cy, cz,
-        sx, sy, sz;
-
-  sx = sinf(angles[0]); cx = cosf(angles[0]);
-  sy = sinf(angles[1]); cy = cosf(angles[1]);
-  sz = sinf(angles[2]); cz = cosf(angles[2]);
-
-  dest[0][0] = cy * cz;
-  dest[0][1] = cy * sz;
-  dest[0][2] =-sy;
-  dest[1][0] = cz * sx * sy - cx * sz;
-  dest[1][1] = cx * cz + sx * sy * sz;
-  dest[1][2] = cy * sx;
-  dest[2][0] = cx * cz * sy + sx * sz;
-  dest[2][1] =-cz * sx + cx * sy * sz;
-  dest[2][2] = cx * cy;
-  dest[0][3] = 0.0f;
-  dest[1][3] = 0.0f;
-  dest[2][3] = 0.0f;
-  dest[3][0] = 0.0f;
-  dest[3][1] = 0.0f;
-  dest[3][2] = 0.0f;
-  dest[3][3] = 1.0f;
+  glm_euler_xyz(angles, dest);
 }
 
 /*!
  * @brief build rotation matrix from euler angles
  *
- * @param[in]  angles angles as vector [Ez, Ey, Ex]
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
  * @param[out] dest   rotation matrix
  */
 CGLM_INLINE
 void
-glm_euler_zyx(vec3 angles,
-              mat4 dest) {
+glm_euler_xzy(vec3 angles, mat4 dest) {
   float cx, cy, cz,
-        sx, sy, sz;
+  sx, sy, sz, sxsy, cysx, cxsy, cxcy;
 
   sx   = sinf(angles[0]); cx = cosf(angles[0]);
   sy   = sinf(angles[1]); cy = cosf(angles[1]);
   sz   = sinf(angles[2]); cz = cosf(angles[2]);
 
+  sxsy = sx * sy;
+  cysx = cy * sx;
+  cxsy = cx * sy;
+  cxcy = cx * cy;
+
   dest[0][0] =  cy * cz;
-  dest[0][1] = cz * sx * sy + cx * sz;
-  dest[0][2] =-cx * cz * sy + sx * sz;
-  dest[1][0] =-cy * sz;
-  dest[1][1] = cx * cz - sx * sy * sz;
-  dest[1][2] = cz * sx + cx * sy * sz;
-  dest[2][0] = sy;
-  dest[2][1] =-cy * sx;
-  dest[2][2] = cx * cy;
-  dest[0][3] = 0.0f;
-  dest[1][3] = 0.0f;
-  dest[2][3] = 0.0f;
-  dest[3][0] = 0.0f;
-  dest[3][1] = 0.0f;
-  dest[3][2] = 0.0f;
-  dest[3][3] = 1.0f;
-}
-
-/*!
- * @brief build rotation matrix from euler angles
- *
- * @param[in]  angles angles as vector [Ez, Ex, Ey]
- * @param[out] dest   rotation matrix
- */
-CGLM_INLINE
-void
-glm_euler_zxy(vec3 angles,
-              mat4 dest) {
-  float cx, cy, cz,
-        sx, sy, sz;
-
-  sx = sinf(angles[0]); cx = cosf(angles[0]);
-  sy = sinf(angles[1]); cy = cosf(angles[1]);
-  sz = sinf(angles[2]); cz = cosf(angles[2]);
-
-  dest[0][0] = cy * cz + sx * sy * sz;
-  dest[0][1] = cx * sz;
-  dest[0][2] =-cz * sy + cy * sx * sz;
-  dest[1][0] = cz * sx * sy - cy * sz;
-  dest[1][1] = cx * cz;
-  dest[1][2] = cy * cz * sx + sy * sz;
-  dest[2][0] = cx * sy;
-  dest[2][1] =-sx;
-  dest[2][2] = cx * cy;
-  dest[0][3] = 0.0f;
-  dest[1][3] = 0.0f;
-  dest[2][3] = 0.0f;
-  dest[3][0] = 0.0f;
-  dest[3][1] = 0.0f;
-  dest[3][2] = 0.0f;
-  dest[3][3] = 1.0f;
-}
-
-/*!
- * @brief build rotation matrix from euler angles
- *
- * @param[in]  angles angles as vector [Ex, Ez, Ey]
- * @param[out] dest   rotation matrix
- */
-CGLM_INLINE
-void
-glm_euler_xzy(vec3 angles,
-              mat4 dest) {
-  float cx, cy, cz,
-        sx, sy, sz;
-
-  sx = sinf(angles[0]); cx = cosf(angles[0]);
-  sy = sinf(angles[1]); cy = cosf(angles[1]);
-  sz = sinf(angles[2]); cz = cosf(angles[2]);
-
-  dest[0][0] = cy * cz;
-  dest[0][1] = sz;
-  dest[0][2] =-cz * sy;
-  dest[1][0] = sx * sy - cx * cy * sz;
-  dest[1][1] = cx * cz;
-  dest[1][2] = cy * sx + cx * sy * sz;
-  dest[2][0] = cx * sy + cy * sx * sz;
-  dest[2][1] =-cz * sx;
-  dest[2][2] = cx * cy - sx * sy * sz;
-  dest[0][3] = 0.0f;
-  dest[1][3] = 0.0f;
-  dest[2][3] = 0.0f;
-  dest[3][0] = 0.0f;
-  dest[3][1] = 0.0f;
-  dest[3][2] = 0.0f;
-  dest[3][3] = 1.0f;
-}
-
-/*!
- * @brief build rotation matrix from euler angles
- *
- * @param[in]  angles angles as vector [Ey, Ez, Ex]
- * @param[out] dest   rotation matrix
- */
-CGLM_INLINE
-void
-glm_euler_yzx(vec3 angles,
-              mat4 dest) {
-  float cx, cy, cz,
-        sx, sy, sz;
-
-  sx = sinf(angles[0]); cx = cosf(angles[0]);
-  sy = sinf(angles[1]); cy = cosf(angles[1]);
-  sz = sinf(angles[2]); cz = cosf(angles[2]);
-
-  dest[0][0] = cy * cz;
-  dest[0][1] = sx * sy + cx * cy * sz;
-  dest[0][2] =-cx * sy + cy * sx * sz;
-  dest[1][0] =-sz;
+  dest[0][1] =  sxsy + cxcy * sz;
+  dest[0][2] = -cxsy + cysx * sz;
+  dest[1][0] = -sz;
   dest[1][1] =  cx * cz;
   dest[1][2] =  cz * sx;
   dest[2][0] =  cz * sy;
-  dest[2][1] =-cy * sx + cx * sy * sz;
-  dest[2][2] = cx * cy + sx * sy * sz;
+  dest[2][1] = -cysx + cxsy * sz;
+  dest[2][2] =  cxcy + sxsy * sz;
   dest[0][3] =  0.0f;
   dest[1][3] =  0.0f;
   dest[2][3] =  0.0f;
@@ -269,31 +188,36 @@ glm_euler_yzx(vec3 angles,
   dest[3][3] =  1.0f;
 }
 
+
 /*!
  * @brief build rotation matrix from euler angles
  *
- * @param[in]  angles angles as vector [Ey, Ex, Ez]
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
  * @param[out] dest   rotation matrix
  */
 CGLM_INLINE
 void
-glm_euler_yxz(vec3 angles,
-              mat4 dest) {
+glm_euler_yxz(vec3 angles, mat4 dest) {
   float cx, cy, cz,
-        sx, sy, sz;
+        sx, sy, sz, cycz, sysz, czsy, cysz;
 
   sx   = sinf(angles[0]); cx = cosf(angles[0]);
   sy   = sinf(angles[1]); cy = cosf(angles[1]);
   sz   = sinf(angles[2]); cz = cosf(angles[2]);
 
-  dest[0][0] = cy * cz - sx * sy * sz;
-  dest[0][1] = cz * sx * sy + cy * sz;
-  dest[0][2] =-cx * sy;
-  dest[1][0] =-cx * sz;
+  cycz = cy * cz;
+  sysz = sy * sz;
+  czsy = cz * sy;
+  cysz = cy * sz;
+
+  dest[0][0] =  cycz + sx * sysz;
+  dest[0][1] =  cx * sz;
+  dest[0][2] = -czsy + cysz * sx;
+  dest[1][0] = -cysz + czsy * sx;
   dest[1][1] =  cx * cz;
-  dest[1][2] = sx;
-  dest[2][0] = cz * sy + cy * sx * sz;
-  dest[2][1] =-cy * cz * sx + sy * sz;
+  dest[1][2] =  cycz * sx + sysz;
+  dest[2][0] =  cx * sy;
+  dest[2][1] = -sx;
   dest[2][2] =  cx * cy;
   dest[0][3] =  0.0f;
   dest[1][3] =  0.0f;
@@ -307,7 +231,122 @@ glm_euler_yxz(vec3 angles,
 /*!
  * @brief build rotation matrix from euler angles
  *
- * @param[in]  angles angles as vector (ord parameter spceifies angles order)
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
+ * @param[out] dest   rotation matrix
+ */
+CGLM_INLINE
+void
+glm_euler_yzx(vec3 angles, mat4 dest) {
+  float cx, cy, cz,
+        sx, sy, sz, sxsy, cxcy, cysx, cxsy;
+
+  sx   = sinf(angles[0]); cx = cosf(angles[0]);
+  sy   = sinf(angles[1]); cy = cosf(angles[1]);
+  sz   = sinf(angles[2]); cz = cosf(angles[2]);
+
+  sxsy = sx * sy;
+  cxcy = cx * cy;
+  cysx = cy * sx;
+  cxsy = cx * sy;
+
+  dest[0][0] =  cy * cz;
+  dest[0][1] =  sz;
+  dest[0][2] = -cz * sy;
+  dest[1][0] =  sxsy - cxcy * sz;
+  dest[1][1] =  cx * cz;
+  dest[1][2] =  cysx + cxsy * sz;
+  dest[2][0] =  cxsy + cysx * sz;
+  dest[2][1] = -cz * sx;
+  dest[2][2] =  cxcy - sxsy * sz;
+  dest[0][3] =  0.0f;
+  dest[1][3] =  0.0f;
+  dest[2][3] =  0.0f;
+  dest[3][0] =  0.0f;
+  dest[3][1] =  0.0f;
+  dest[3][2] =  0.0f;
+  dest[3][3] =  1.0f;
+}
+
+/*!
+ * @brief build rotation matrix from euler angles
+ *
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
+ * @param[out] dest   rotation matrix
+ */
+CGLM_INLINE
+void
+glm_euler_zxy(vec3 angles, mat4 dest) {
+  float cx, cy, cz,
+        sx, sy, sz, cycz, sxsy, cysz;
+
+  sx   = sinf(angles[0]); cx = cosf(angles[0]);
+  sy   = sinf(angles[1]); cy = cosf(angles[1]);
+  sz   = sinf(angles[2]); cz = cosf(angles[2]);
+
+  cycz = cy * cz;
+  sxsy = sx * sy;
+  cysz = cy * sz;
+
+  dest[0][0] =  cycz - sxsy * sz;
+  dest[0][1] =  cz * sxsy + cysz;
+  dest[0][2] = -cx * sy;
+  dest[1][0] = -cx * sz;
+  dest[1][1] =  cx * cz;
+  dest[1][2] =  sx;
+  dest[2][0] =  cz * sy + cysz * sx;
+  dest[2][1] = -cycz * sx + sy * sz;
+  dest[2][2] =  cx * cy;
+  dest[0][3] =  0.0f;
+  dest[1][3] =  0.0f;
+  dest[2][3] =  0.0f;
+  dest[3][0] =  0.0f;
+  dest[3][1] =  0.0f;
+  dest[3][2] =  0.0f;
+  dest[3][3] =  1.0f;
+}
+
+/*!
+ * @brief build rotation matrix from euler angles
+ *
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
+ * @param[out] dest   rotation matrix
+ */
+CGLM_INLINE
+void
+glm_euler_zyx(vec3 angles, mat4 dest) {
+  float cx, cy, cz,
+        sx, sy, sz, czsx, cxcz, sysz;
+
+  sx   = sinf(angles[0]); cx = cosf(angles[0]);
+  sy   = sinf(angles[1]); cy = cosf(angles[1]);
+  sz   = sinf(angles[2]); cz = cosf(angles[2]);
+
+  czsx = cz * sx;
+  cxcz = cx * cz;
+  sysz = sy * sz;
+
+  dest[0][0] =  cy * cz;
+  dest[0][1] =  cy * sz;
+  dest[0][2] = -sy;
+  dest[1][0] =  czsx * sy - cx * sz;
+  dest[1][1] =  cxcz + sx * sysz;
+  dest[1][2] =  cy * sx;
+  dest[2][0] =  cxcz * sy + sx * sz;
+  dest[2][1] = -czsx + cx * sysz;
+  dest[2][2] =  cx * cy;
+  dest[0][3] =  0.0f;
+  dest[1][3] =  0.0f;
+  dest[2][3] =  0.0f;
+  dest[3][0] =  0.0f;
+  dest[3][1] =  0.0f;
+  dest[3][2] =  0.0f;
+  dest[3][3] =  1.0f;
+}
+
+/*!
+ * @brief build rotation matrix from euler angles
+ *
+ * @param[in]  angles angles as vector [Xangle, Yangle, Zangle]
  * @param[in]  ord    euler order
  * @param[out] dest   rotation matrix
  */
@@ -332,71 +371,71 @@ glm_euler_by_order(vec3 angles, glm_euler_sq ord, mat4 dest) {
   sysz = sy * sz;
 
   switch (ord) {
-    case GLM_EULER_XYZ:
-      dest[0][0] = cycz;
-      dest[0][1] = cysz;
-      dest[0][2] =-sy;
-      dest[1][0] = czsx * sy - cxsz;
-      dest[1][1] = cxcz + sx * sysz;
-      dest[1][2] = cysx;
-      dest[2][0] = cx * czsy + sx * sz;
-      dest[2][1] =-czsx + cx * sysz;
-      dest[2][2] = cxcy;
-      break;
     case GLM_EULER_XZY:
       dest[0][0] =  cycz;
-      dest[0][1] = sz;
-      dest[0][2] =-czsy;
-      dest[1][0] = sx * sy - cx * cysz;
+      dest[0][1] =  sx * sy + cx * cysz;
+      dest[0][2] = -cx * sy + cysx * sz;
+      dest[1][0] = -sz;
       dest[1][1] =  cxcz;
-      dest[1][2] = cysx + cx * sysz;
-      dest[2][0] = cx * sy + cysx * sz;
-      dest[2][1] =-czsx;
-      dest[2][2] = cxcy - sx * sysz;
+      dest[1][2] =  czsx;
+      dest[2][0] =  czsy;
+      dest[2][1] = -cysx + cx * sysz;
+      dest[2][2] =  cxcy + sx * sysz;
       break;
-    case GLM_EULER_ZXY:
+    case GLM_EULER_XYZ:
+      dest[0][0] =  cycz;
+      dest[0][1] =  czsx * sy + cxsz;
+      dest[0][2] = -cx * czsy + sx * sz;
+      dest[1][0] = -cysz;
+      dest[1][1] =  cxcz - sx * sysz;
+      dest[1][2] =  czsx + cx * sysz;
+      dest[2][0] =  sy;
+      dest[2][1] = -cysx;
+      dest[2][2] =  cxcy;
+      break;
+    case GLM_EULER_YXZ:
       dest[0][0] =  cycz + sx * sysz;
       dest[0][1] =  cxsz;
-      dest[0][2] =-czsy + cysx * sz;
+      dest[0][2] = -czsy + cysx * sz;
       dest[1][0] =  czsx * sy - cysz;
       dest[1][1] =  cxcz;
       dest[1][2] =  cycz * sx + sysz;
       dest[2][0] =  cx * sy;
-      dest[2][1] =-sx;
-      dest[2][2] = cxcy;
-      break;
-    case GLM_EULER_ZYX:
-      dest[0][0] = cycz;
-      dest[0][1] = czsx * sy + cxsz;
-      dest[0][2] =-cx * czsy + sx * sz;
-      dest[1][0] =-cysz;
-      dest[1][1] = cxcz - sx * sysz;
-      dest[1][2] = czsx + cx * sysz;
-      dest[2][0] = sy;
-      dest[2][1] =-cysx;
-      dest[2][2] = cxcy;
-      break;
-    case GLM_EULER_YXZ:
-      dest[0][0] = cycz - sx * sysz;
-      dest[0][1] = czsx * sy + cysz;
-      dest[0][2] =-cx * sy;
-      dest[1][0] =-cxsz;
-      dest[1][1] = cxcz;
-      dest[1][2] = sx;
-      dest[2][0] = czsy + cysx * sz;
-      dest[2][1] =-cycz * sx + sysz;
+      dest[2][1] = -sx;
       dest[2][2] =  cxcy;
       break;
     case GLM_EULER_YZX:
       dest[0][0] =  cycz;
-      dest[0][1] = sx * sy + cx * cysz;
-      dest[0][2] =-cx * sy + cysx * sz;
-      dest[1][0] =-sz;
+      dest[0][1] =  sz;
+      dest[0][2] = -czsy;
+      dest[1][0] =  sx * sy - cx * cysz;
       dest[1][1] =  cxcz;
-      dest[1][2] = czsx;
-      dest[2][0] = czsy;
-      dest[2][1] =-cysx + cx * sysz;
-      dest[2][2] = cxcy + sx * sysz;
+      dest[1][2] =  cysx + cx * sysz;
+      dest[2][0] =  cx * sy + cysx * sz;
+      dest[2][1] = -czsx;
+      dest[2][2] =  cxcy - sx * sysz;
+      break;
+    case GLM_EULER_ZXY:
+      dest[0][0] =  cycz - sx * sysz;
+      dest[0][1] =  czsx * sy + cysz;
+      dest[0][2] = -cx * sy;
+      dest[1][0] = -cxsz;
+      dest[1][1] =  cxcz;
+      dest[1][2] =  sx;
+      dest[2][0] =  czsy + cysx * sz;
+      dest[2][1] = -cycz * sx + sysz;
+      dest[2][2] =  cxcy;
+      break;
+    case GLM_EULER_ZYX:
+      dest[0][0] =  cycz;
+      dest[0][1] =  cysz;
+      dest[0][2] = -sy;
+      dest[1][0] =  czsx * sy - cxsz;
+      dest[1][1] =  cxcz + sx * sysz;
+      dest[1][2] =  cysx;
+      dest[2][0] =  cx * czsy + sx * sz;
+      dest[2][1] = -czsx + cx * sysz;
+      dest[2][2] =  cxcy;
       break;
   }
 

include/cglm/frustum.h

@@ -10,6 +10,9 @@
 
 #include "common.h"
 #include "plane.h"
+#include "vec3.h"
+#include "vec4.h"
+#include "mat4.h"
 
 #define GLM_LBN 0 /* left  bottom near */
 #define GLM_LTN 1 /* left  top    near */

include/cglm/io.h

@@ -171,4 +171,33 @@ glm_versor_print(versor vec,
 #undef m
 }
 
+CGLM_INLINE
+void
+glm_aabb_print(vec3                    bbox[2],
+               const char * __restrict tag,
+               FILE       * __restrict ostream) {
+  int i, j;
+
+#define m 3
+
+  fprintf(ostream, "AABB (%s):\n", tag ? tag: "float");
+
+  for (i = 0; i < 2; i++) {
+    fprintf(ostream, "\t|");
+
+    for (j = 0; j < m; j++) {
+      fprintf(ostream, "%0.4f", bbox[i][j]);
+
+      if (j != m - 1)
+        fprintf(ostream, "\t");
+    }
+
+    fprintf(ostream, "|\n");
+  }
+
+  fprintf(ostream, "\n");
+
+#undef m
+}
+
 #endif /* cglm_io_h */

include/cglm/mat3.h

@@ -31,6 +31,7 @@
 #define cglm_mat3_h
 
 #include "common.h"
+#include "vec3.h"
 
 #ifdef CGLM_SSE_FP
 #  include "simd/sse2/mat3.h"
@@ -80,7 +81,7 @@ glm_mat3_copy(mat3 mat, mat3 dest) {
 CGLM_INLINE
 void
 glm_mat3_identity(mat3 mat) {
-  mat3 t = GLM_MAT3_IDENTITY_INIT;
+  CGLM_ALIGN(16) mat3 t = GLM_MAT3_IDENTITY_INIT;
   glm_mat3_copy(t, mat);
 }
 
@@ -154,7 +155,7 @@ glm_mat3_transpose_to(mat3 m, mat3 dest) {
 CGLM_INLINE
 void
 glm_mat3_transpose(mat3 m) {
-  mat3 tmp;
+  CGLM_ALIGN(16) mat3 tmp;
 
   tmp[0][1] = m[1][0];
   tmp[0][2] = m[2][0];
@@ -186,6 +187,56 @@ glm_mat3_mulv(mat3 m, vec3 v, vec3 dest) {
   dest[2] = m[0][2] * v[0] + m[1][2] * v[1] + m[2][2] * v[2];
 }
 
+
+/*!
+ * @brief convert mat3 to quaternion
+ *
+ * @param[in]  m    rotation matrix
+ * @param[out] dest destination quaternion
+ */
+CGLM_INLINE
+void
+glm_mat3_quat(mat3 m, versor dest) {
+  float trace, r, rinv;
+
+  /* it seems using like m12 instead of m[1][2] causes extra instructions */
+
+  trace = m[0][0] + m[1][1] + m[2][2];
+  if (trace >= 0.0f) {
+    r       = sqrtf(1.0f + trace);
+    rinv    = 0.5f / r;
+
+    dest[0] = rinv * (m[1][2] - m[2][1]);
+    dest[1] = rinv * (m[2][0] - m[0][2]);
+    dest[2] = rinv * (m[0][1] - m[1][0]);
+    dest[3] = r    * 0.5f;
+  } else if (m[0][0] >= m[1][1] && m[0][0] >= m[2][2]) {
+    r       = sqrtf(1.0f - m[1][1] - m[2][2] + m[0][0]);
+    rinv    = 0.5f / r;
+
+    dest[0] = r    * 0.5f;
+    dest[1] = rinv * (m[0][1] + m[1][0]);
+    dest[2] = rinv * (m[0][2] + m[2][0]);
+    dest[3] = rinv * (m[1][2] - m[2][1]);
+  } else if (m[1][1] >= m[2][2]) {
+    r       = sqrtf(1.0f - m[0][0] - m[2][2] + m[1][1]);
+    rinv    = 0.5f / r;
+
+    dest[0] = rinv * (m[0][1] + m[1][0]);
+    dest[1] = r    * 0.5f;
+    dest[2] = rinv * (m[1][2] + m[2][1]);
+    dest[3] = rinv * (m[2][0] - m[0][2]);
+  } else {
+    r       = sqrtf(1.0f - m[0][0] - m[1][1] + m[2][2]);
+    rinv    = 0.5f / r;
+
+    dest[0] = rinv * (m[0][2] + m[2][0]);
+    dest[1] = rinv * (m[1][2] + m[2][1]);
+    dest[2] = r    * 0.5f;
+    dest[3] = rinv * (m[0][1] - m[1][0]);
+  }
+}
+
 /*!
  * @brief scale (multiply with scalar) matrix
  *

include/cglm/mat4.h

@@ -45,6 +45,8 @@
 #define cglm_mat_h
 
 #include "common.h"
+#include "vec4.h"
+#include "vec3.h"
 
 #ifdef CGLM_SSE_FP
 #  include "simd/sse2/mat4.h"
@@ -58,7 +60,9 @@
 #  include "simd/neon/mat4.h"
 #endif
 
-#include <assert.h>
+#ifdef DEBUG
+# include <assert.h>
+#endif
 
 #define GLM_MAT4_IDENTITY_INIT  {{1.0f, 0.0f, 0.0f, 0.0f},                    \
                                  {0.0f, 1.0f, 0.0f, 0.0f},                    \
@@ -106,13 +110,13 @@ CGLM_INLINE
 void
 glm_mat4_copy(mat4 mat, mat4 dest) {
 #ifdef __AVX__
-  _mm256_store_ps(dest[0], _mm256_load_ps(mat[0]));
-  _mm256_store_ps(dest[2], _mm256_load_ps(mat[2]));
+  glmm_store256(dest[0], glmm_load256(mat[0]));
+  glmm_store256(dest[2], glmm_load256(mat[2]));
 #elif defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(dest[0], _mm_load_ps(mat[0]));
-  _mm_store_ps(dest[1], _mm_load_ps(mat[1]));
-  _mm_store_ps(dest[2], _mm_load_ps(mat[2]));
-  _mm_store_ps(dest[3], _mm_load_ps(mat[3]));
+  glmm_store(dest[0], glmm_load(mat[0]));
+  glmm_store(dest[1], glmm_load(mat[1]));
+  glmm_store(dest[2], glmm_load(mat[2]));
+  glmm_store(dest[3], glmm_load(mat[3]));
 #else
   glm_mat4_ucopy(mat, dest);
 #endif
@@ -135,7 +139,7 @@ glm_mat4_copy(mat4 mat, mat4 dest) {
 CGLM_INLINE
 void
 glm_mat4_identity(mat4 mat) {
-  mat4 t = GLM_MAT4_IDENTITY_INIT;
+  CGLM_ALIGN(16) mat4 t = GLM_MAT4_IDENTITY_INIT;
   glm_mat4_copy(t, mat);
 }
 
@@ -281,19 +285,17 @@ glm_mat4_mul(mat4 m1, mat4 m2, mat4 dest) {
  */
 CGLM_INLINE
 void
-glm_mat4_mulN(mat4 * __restrict matrices[], int len, mat4 dest) {
-  int i;
+glm_mat4_mulN(mat4 * __restrict matrices[], uint32_t len, mat4 dest) {
+  uint32_t i;
 
+#ifdef DEBUG
   assert(len > 1 && "there must be least 2 matrices to go!");
+#endif
 
-  glm_mat4_mul(*matrices[0],
-               *matrices[1],
-               dest);
+  glm_mat4_mul(*matrices[0], *matrices[1], dest);
 
   for (i = 2; i < len; i++)
-    glm_mat4_mul(dest,
-                 *matrices[i],
-                 dest);
+    glm_mat4_mul(dest, *matrices[i], dest);
 }
 
 /*!
@@ -319,20 +321,69 @@ glm_mat4_mulv(mat4 m, vec4 v, vec4 dest) {
 }
 
 /*!
- * @brief multiply vector with mat4's mat3 part(rotation)
+ * @brief convert mat4's rotation part to quaternion
  *
- * @param[in]  m    mat4(affine transform)
- * @param[in]  v    vec3
- * @param[out] dest vec3
+ * @param[in]  m    affine matrix
+ * @param[out] dest destination quaternion
  */
 CGLM_INLINE
 void
-glm_mat4_mulv3(mat4 m, vec3 v, vec3 dest) {
-  vec3 res;
-  res[0] = m[0][0] * v[0] + m[1][0] * v[1] + m[2][0] * v[2];
-  res[1] = m[0][1] * v[0] + m[1][1] * v[1] + m[2][1] * v[2];
-  res[2] = m[0][2] * v[0] + m[1][2] * v[1] + m[2][2] * v[2];
-  glm_vec_copy(res, dest);
+glm_mat4_quat(mat4 m, versor dest) {
+  float trace, r, rinv;
+
+  /* it seems using like m12 instead of m[1][2] causes extra instructions */
+
+  trace = m[0][0] + m[1][1] + m[2][2];
+  if (trace >= 0.0f) {
+    r       = sqrtf(1.0f + trace);
+    rinv    = 0.5f / r;
+
+    dest[0] = rinv * (m[1][2] - m[2][1]);
+    dest[1] = rinv * (m[2][0] - m[0][2]);
+    dest[2] = rinv * (m[0][1] - m[1][0]);
+    dest[3] = r    * 0.5f;
+  } else if (m[0][0] >= m[1][1] && m[0][0] >= m[2][2]) {
+    r       = sqrtf(1.0f - m[1][1] - m[2][2] + m[0][0]);
+    rinv    = 0.5f / r;
+
+    dest[0] = r    * 0.5f;
+    dest[1] = rinv * (m[0][1] + m[1][0]);
+    dest[2] = rinv * (m[0][2] + m[2][0]);
+    dest[3] = rinv * (m[1][2] - m[2][1]);
+  } else if (m[1][1] >= m[2][2]) {
+    r       = sqrtf(1.0f - m[0][0] - m[2][2] + m[1][1]);
+    rinv    = 0.5f / r;
+
+    dest[0] = rinv * (m[0][1] + m[1][0]);
+    dest[1] = r    * 0.5f;
+    dest[2] = rinv * (m[1][2] + m[2][1]);
+    dest[3] = rinv * (m[2][0] - m[0][2]);
+  } else {
+    r       = sqrtf(1.0f - m[0][0] - m[1][1] + m[2][2]);
+    rinv    = 0.5f / r;
+
+    dest[0] = rinv * (m[0][2] + m[2][0]);
+    dest[1] = rinv * (m[1][2] + m[2][1]);
+    dest[2] = r    * 0.5f;
+    dest[3] = rinv * (m[0][1] - m[1][0]);
+  }
+}
+
+/*!
+ * @brief multiply vector with mat4
+ *
+ * @param[in]  m    mat4(affine transform)
+ * @param[in]  v    vec3
+ * @param[in]  last 4th item to make it vec4
+ * @param[out] dest result vector (vec3)
+ */
+CGLM_INLINE
+void
+glm_mat4_mulv3(mat4 m, vec3 v, float last, vec3 dest) {
+  vec4 res;
+  glm_vec4(v, last, res);
+  glm_mat4_mulv(m, res, res);
+  glm_vec3(res, dest);
 }
 
 /*!
@@ -535,7 +586,7 @@ glm_mat4_inv_fast(mat4 mat, mat4 dest) {
 CGLM_INLINE
 void
 glm_mat4_swap_col(mat4 mat, int col1, int col2) {
-  vec4 tmp;
+  CGLM_ALIGN(16) vec4 tmp;
   glm_vec4_copy(mat[col1], tmp);
   glm_vec4_copy(mat[col2], mat[col1]);
   glm_vec4_copy(tmp, mat[col2]);
@@ -551,7 +602,7 @@ glm_mat4_swap_col(mat4 mat, int col1, int col2) {
 CGLM_INLINE
 void
 glm_mat4_swap_row(mat4 mat, int row1, int row2) {
-  vec4 tmp;
+  CGLM_ALIGN(16) vec4 tmp;
   tmp[0] = mat[0][row1];
   tmp[1] = mat[1][row1];
   tmp[2] = mat[2][row1];
@@ -568,5 +619,4 @@ glm_mat4_swap_row(mat4 mat, int row1, int row2) {
   mat[3][row2] = tmp[3];
 }
 
-#else
 #endif /* cglm_mat_h */

include/cglm/plane.h

@@ -9,9 +9,7 @@
 #define cglm_plane_h
 
 #include "common.h"
-#include "mat4.h"
 #include "vec4.h"
-#include "vec3.h"
 
 /*
  Plane equation:  Ax + By + Cz + D = 0;

include/cglm/project.h

@@ -0,0 +1,117 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#ifndef cglm_project_h
+#define cglm_project_h
+
+#include "vec3.h"
+#include "vec4.h"
+#include "mat4.h"
+
+/*!
+ * @brief maps the specified viewport coordinates into specified space [1]
+ *        the matrix should contain projection matrix.
+ *
+ * if you don't have ( and don't want to have ) an inverse matrix then use
+ * glm_unproject version. You may use existing inverse of matrix in somewhere
+ * else, this is why glm_unprojecti exists to save save inversion cost
+ *
+ * [1] space:
+ *  1- if m = invProj:     View Space
+ *  2- if m = invViewProj: World Space
+ *  3- if m = invMVP:      Object Space
+ *
+ * You probably want to map the coordinates into object space
+ * so use invMVP as m
+ *
+ * Computing viewProj:
+ *   glm_mat4_mul(proj, view, viewProj);
+ *   glm_mat4_mul(viewProj, model, MVP);
+ *   glm_mat4_inv(viewProj, invMVP);
+ *
+ * @param[in]  pos      point/position in viewport coordinates
+ * @param[in]  invMat   matrix (see brief)
+ * @param[in]  vp       viewport as [x, y, width, height]
+ * @param[out] dest     unprojected coordinates
+ */
+CGLM_INLINE
+void
+glm_unprojecti(vec3 pos, mat4 invMat, vec4 vp, vec3 dest) {
+  vec4 v;
+
+  v[0] = 2.0f * (pos[0] - vp[0]) / vp[2] - 1.0f;
+  v[1] = 2.0f * (pos[1] - vp[1]) / vp[3] - 1.0f;
+  v[2] = 2.0f *  pos[2]                  - 1.0f;
+  v[3] = 1.0f;
+
+  glm_mat4_mulv(invMat, v, v);
+  glm_vec4_scale(v, 1.0f / v[3], v);
+  glm_vec3(v, dest);
+}
+
+/*!
+ * @brief maps the specified viewport coordinates into specified space [1]
+ *        the matrix should contain projection matrix.
+ *
+ * this is same as glm_unprojecti except this function get inverse matrix for
+ * you.
+ *
+ * [1] space:
+ *  1- if m = proj:     View Space
+ *  2- if m = viewProj: World Space
+ *  3- if m = MVP:      Object Space
+ *
+ * You probably want to map the coordinates into object space
+ * so use MVP as m
+ *
+ * Computing viewProj and MVP:
+ *   glm_mat4_mul(proj, view, viewProj);
+ *   glm_mat4_mul(viewProj, model, MVP);
+ *
+ * @param[in]  pos      point/position in viewport coordinates
+ * @param[in]  m        matrix (see brief)
+ * @param[in]  vp       viewport as [x, y, width, height]
+ * @param[out] dest     unprojected coordinates
+ */
+CGLM_INLINE
+void
+glm_unproject(vec3 pos, mat4 m, vec4 vp, vec3 dest) {
+  mat4 inv;
+  glm_mat4_inv(m, inv);
+  glm_unprojecti(pos, inv, vp, dest);
+}
+
+/*!
+ * @brief map object coordinates to window coordinates
+ *
+ * Computing MVP:
+ *   glm_mat4_mul(proj, view, viewProj);
+ *   glm_mat4_mul(viewProj, model, MVP);
+ *
+ * @param[in]  pos      object coordinates
+ * @param[in]  m        MVP matrix
+ * @param[in]  vp       viewport as [x, y, width, height]
+ * @param[out] dest     projected coordinates
+ */
+CGLM_INLINE
+void
+glm_project(vec3 pos, mat4 m, vec4 vp, vec3 dest) {
+  CGLM_ALIGN(16) vec4 pos4, vone = GLM_VEC4_ONE_INIT;
+
+  glm_vec4(pos, 1.0f, pos4);
+
+  glm_mat4_mulv(m, pos4, pos4);
+  glm_vec4_scale(pos4, 1.0f / pos4[3], pos4); /* pos = pos / pos.w */
+  glm_vec4_add(pos4, vone, pos4);
+  glm_vec4_scale(pos4, 0.5f, pos4);
+
+  dest[0] = pos4[0] * vp[2] + vp[0];
+  dest[1] = pos4[1] * vp[3] + vp[1];
+  dest[2] = pos4[2];
+}
+
+#endif /* cglm_project_h */

include/cglm/quat.h

@@ -12,40 +12,83 @@
 
  Functions:
    CGLM_INLINE void glm_quat_identity(versor q);
+   CGLM_INLINE void glm_quat_init(versor q, float x, float y, float z, float w);
    CGLM_INLINE void glm_quat(versor q, float angle, float x, float y, float z);
-   CGLM_INLINE void  glm_quatv(versor q, float angle, vec3 v);
+   CGLM_INLINE void glm_quatv(versor q, float angle, vec3 axis);
+   CGLM_INLINE void glm_quat_copy(versor q, versor dest);
    CGLM_INLINE float glm_quat_norm(versor q);
    CGLM_INLINE void glm_quat_normalize(versor q);
-   CGLM_INLINE float glm_quat_dot(versor q, versor r);
-   CGLM_INLINE void  glm_quat_mulv(versor q1, versor q2, versor dest);
+   CGLM_INLINE void glm_quat_normalize_to(versor q, versor dest);
+   CGLM_INLINE float glm_quat_dot(versor q1, versor q2);
+   CGLM_INLINE void glm_quat_conjugate(versor q, versor dest);
+   CGLM_INLINE void glm_quat_inv(versor q, versor dest);
+   CGLM_INLINE void glm_quat_add(versor p, versor q, versor dest);
+   CGLM_INLINE void glm_quat_sub(versor p, versor q, versor dest);
+   CGLM_INLINE float glm_quat_real(versor q);
+   CGLM_INLINE void glm_quat_imag(versor q, vec3 dest);
+   CGLM_INLINE void glm_quat_imagn(versor q, vec3 dest);
+   CGLM_INLINE float glm_quat_imaglen(versor q);
+   CGLM_INLINE float glm_quat_angle(versor q);
+   CGLM_INLINE void glm_quat_axis(versor q, versor dest);
+   CGLM_INLINE void glm_quat_mul(versor p, versor q, versor dest);
    CGLM_INLINE void glm_quat_mat4(versor q, mat4 dest);
+   CGLM_INLINE void glm_quat_mat4t(versor q, mat4 dest);
+   CGLM_INLINE void glm_quat_mat3(versor q, mat3 dest);
+   CGLM_INLINE void glm_quat_mat3t(versor q, mat3 dest);
+   CGLM_INLINE void glm_quat_lerp(versor from, versor to, float t, versor dest);
    CGLM_INLINE void glm_quat_slerp(versor q, versor r, float t, versor dest);
+   CGLM_INLINE void glm_quat_look(vec3 eye, versor ori, mat4 dest);
+   CGLM_INLINE void glm_quat_for(vec3 dir, vec3 fwd, vec3 up, versor dest);
+   CGLM_INLINE void glm_quat_forp(vec3 from,
+                                  vec3 to,
+                                  vec3 fwd,
+                                  vec3 up,
+                                  versor dest);
+   CGLM_INLINE void glm_quat_rotatev(versor q, vec3 v, vec3 dest);
+   CGLM_INLINE void glm_quat_rotate(mat4 m, versor q, mat4 dest);
  */
 
 #ifndef cglm_quat_h
 #define cglm_quat_h
 
 #include "common.h"
+#include "vec3.h"
 #include "vec4.h"
+#include "mat4.h"
+#include "mat3.h"
+#include "affine-mat.h"
 
 #ifdef CGLM_SSE_FP
 #  include "simd/sse2/quat.h"
 #endif
 
+CGLM_INLINE
+void
+glm_mat4_identity(mat4 mat);
+
+CGLM_INLINE
+void
+glm_mat4_mulv(mat4 m, vec4 v, vec4 dest);
+
+CGLM_INLINE
+void
+glm_mul_rot(mat4 m1, mat4 m2, mat4 dest);
+
+CGLM_INLINE
+void
+glm_translate(mat4 m, vec3 v);
+
 /*
- * IMPORTANT! cglm stores quat as [w, x, y, z]
+ * IMPORTANT:
+ * ----------------------------------------------------------------------------
+ * cglm stores quat as [x, y, z, w] since v0.3.6
  *
- * Possible changes (these may be changed in the future):
- *  - versor is identity quat, we can define new type for quat.
- *    it can't be quat or quaternion becuase someone can use that name for
- *    variable name. maybe just vec4.
- *  - it stores [w, x, y, z] but it may change to [x, y, z, w] if we get enough
- *    feedback to change it.
- *  - in general we use last param as dest, but this header used first param
- *    as dest this may be changed but decided yet
+ * it was [w, x, y, z] before v0.3.6 it has been changed to [x, y, z, w]
+ * with v0.3.6 version.
+ * ----------------------------------------------------------------------------
  */
 
-#define GLM_QUAT_IDENTITY_INIT  {1.0f, 0.0f, 0.0f, 0.0f}
+#define GLM_QUAT_IDENTITY_INIT  {0.0f, 0.0f, 0.0f, 1.0f}
 #define GLM_QUAT_IDENTITY       ((versor)GLM_QUAT_IDENTITY_INIT)
 
 /*!
@@ -56,10 +99,53 @@
 CGLM_INLINE
 void
 glm_quat_identity(versor q) {
-  versor v = GLM_QUAT_IDENTITY_INIT;
+  CGLM_ALIGN(16) versor v = GLM_QUAT_IDENTITY_INIT;
   glm_vec4_copy(v, q);
 }
 
+/*!
+ * @brief inits quaterion with raw values
+ *
+ * @param[out]  q     quaternion
+ * @param[in]   x     x
+ * @param[in]   y     y
+ * @param[in]   z     z
+ * @param[in]   w     w (real part)
+ */
+CGLM_INLINE
+void
+glm_quat_init(versor q, float x, float y, float z, float w) {
+  q[0] = x;
+  q[1] = y;
+  q[2] = z;
+  q[3] = w;
+}
+
+/*!
+ * @brief creates NEW quaternion with axis vector
+ *
+ * @param[out]  q     quaternion
+ * @param[in]   angle angle (radians)
+ * @param[in]   axis  axis
+ */
+CGLM_INLINE
+void
+glm_quatv(versor q, float angle, vec3 axis) {
+  CGLM_ALIGN(8) vec3 k;
+  float a, c, s;
+
+  a = angle * 0.5f;
+  c = cosf(a);
+  s = sinf(a);
+
+  glm_normalize_to(axis, k);
+
+  q[0] = s * k[0];
+  q[1] = s * k[1];
+  q[2] = s * k[2];
+  q[3] = c;
+}
+
 /*!
  * @brief creates NEW quaternion with individual axis components
  *
@@ -71,45 +157,21 @@ glm_quat_identity(versor q) {
  */
 CGLM_INLINE
 void
-glm_quat(versor q,
-         float  angle,
-         float  x,
-         float  y,
-         float  z) {
-  float a, c, s;
-
-  a = angle * 0.5f;
-  c = cosf(a);
-  s = sinf(a);
-
-  q[0] = c;
-  q[1] = s * x;
-  q[2] = s * y;
-  q[3] = s * z;
+glm_quat(versor q, float angle, float x, float y, float z) {
+  CGLM_ALIGN(8) vec3 axis = {x, y, z};
+  glm_quatv(q, angle, axis);
 }
 
 /*!
- * @brief creates NEW quaternion with axis vector
+ * @brief copy quaternion to another one
  *
- * @param[out]  q     quaternion
- * @param[in]   angle angle (radians)
- * @param[in]   v     axis
+ * @param[in]  q     quaternion
+ * @param[out] dest  destination
  */
 CGLM_INLINE
 void
-glm_quatv(versor q,
-          float  angle,
-          vec3   v) {
-  float a, c, s;
-
-  a = angle * 0.5f;
-  c = cosf(a);
-  s = sinf(a);
-
-  q[0] = c;
-  q[1] = s * v[0];
-  q[2] = s * v[1];
-  q[3] = s * v[2];
+glm_quat_copy(versor q, versor dest) {
+  glm_vec4_copy(q, dest);
 }
 
 /*!
@@ -123,6 +185,43 @@ glm_quat_norm(versor q) {
   return glm_vec4_norm(q);
 }
 
+/*!
+ * @brief normalize quaternion and store result in dest
+ *
+ * @param[in]   q     quaternion to normalze
+ * @param[out]  dest  destination quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_normalize_to(versor q, versor dest) {
+#if defined( __SSE2__ ) || defined( __SSE2__ )
+  __m128 xdot, x0;
+  float  dot;
+
+  x0   = glmm_load(q);
+  xdot = glmm_dot(x0, x0);
+  dot  = _mm_cvtss_f32(xdot);
+
+  if (dot <= 0.0f) {
+    glm_quat_identity(dest);
+    return;
+  }
+
+  glmm_store(dest, _mm_div_ps(x0, _mm_sqrt_ps(xdot)));
+#else
+  float dot;
+
+  dot = glm_vec4_norm2(q);
+
+  if (dot <= 0.0f) {
+    glm_quat_identity(q);
+    return;
+  }
+
+  glm_vec4_scale(q, 1.0f / sqrtf(dot), dest);
+#endif
+}
+
 /*!
  * @brief normalize quaternion
  *
@@ -131,45 +230,178 @@ glm_quat_norm(versor q) {
 CGLM_INLINE
 void
 glm_quat_normalize(versor q) {
-  float sum;
-
-  sum = q[0] * q[0] + q[1] * q[1]
-          + q[2] * q[2] + q[3] * q[3];
-
-  if (fabs(1.0f - sum) < 0.0001f)
-    return;
-
-  glm_vec4_scale(q, 1.0f / sqrtf(sum), q);
+  glm_quat_normalize_to(q, q);
 }
 
 /*!
  * @brief dot product of two quaternion
  *
- * @param[in]  q  quaternion 1
- * @param[in]  r  quaternion 2
+ * @param[in]  p  quaternion 1
+ * @param[in]  q  quaternion 2
  */
 CGLM_INLINE
 float
-glm_quat_dot(versor q, versor r) {
-  return glm_vec4_dot(q, r);
+glm_quat_dot(versor p, versor q) {
+  return glm_vec4_dot(p, q);
 }
 
 /*!
- * @brief multiplies two quaternion and stores result in dest
+ * @brief conjugate of quaternion
  *
- * @param[in]   q1    quaternion 1
- * @param[in]   q2    quaternion 2
+ * @param[in]   q     quaternion
+ * @param[out]  dest  conjugate
+ */
+CGLM_INLINE
+void
+glm_quat_conjugate(versor q, versor dest) {
+  glm_vec4_flipsign_to(q, dest);
+  dest[3] = -dest[3];
+}
+
+/*!
+ * @brief inverse of non-zero quaternion
+ *
+ * @param[in]   q    quaternion
+ * @param[out]  dest inverse quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_inv(versor q, versor dest) {
+  CGLM_ALIGN(8) versor conj;
+  glm_quat_conjugate(q, conj);
+  glm_vec4_scale(conj, 1.0f / glm_vec4_norm2(q), dest);
+}
+
+/*!
+ * @brief add (componentwise) two quaternions and store result in dest
+ *
+ * @param[in]   p    quaternion 1
+ * @param[in]   q    quaternion 2
  * @param[out]  dest result quaternion
  */
 CGLM_INLINE
 void
-glm_quat_mulv(versor q1, versor q2, versor dest) {
-  dest[0] = q2[0] * q1[0] - q2[1] * q1[1] - q2[2] * q1[2] - q2[3] * q1[3];
-  dest[1] = q2[0] * q1[1] + q2[1] * q1[0] - q2[2] * q1[3] + q2[3] * q1[2];
-  dest[2] = q2[0] * q1[2] + q2[1] * q1[3] + q2[2] * q1[0] - q2[3] * q1[1];
-  dest[3] = q2[0] * q1[3] - q2[1] * q1[2] + q2[2] * q1[1] + q2[3] * q1[0];
+glm_quat_add(versor p, versor q, versor dest) {
+  glm_vec4_add(p, q, dest);
+}
 
-  glm_quat_normalize(dest);
+/*!
+ * @brief subtract (componentwise) two quaternions and store result in dest
+ *
+ * @param[in]   p    quaternion 1
+ * @param[in]   q    quaternion 2
+ * @param[out]  dest result quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_sub(versor p, versor q, versor dest) {
+  glm_vec4_sub(p, q, dest);
+}
+
+/*!
+ * @brief returns real part of quaternion
+ *
+ * @param[in]   q    quaternion
+ */
+CGLM_INLINE
+float
+glm_quat_real(versor q) {
+  return q[3];
+}
+
+/*!
+ * @brief returns imaginary part of quaternion
+ *
+ * @param[in]   q    quaternion
+ * @param[out]  dest imag
+ */
+CGLM_INLINE
+void
+glm_quat_imag(versor q, vec3 dest) {
+  dest[0] = q[0];
+  dest[1] = q[1];
+  dest[2] = q[2];
+}
+
+/*!
+ * @brief returns normalized imaginary part of quaternion
+ *
+ * @param[in]   q    quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_imagn(versor q, vec3 dest) {
+  glm_normalize_to(q, dest);
+}
+
+/*!
+ * @brief returns length of imaginary part of quaternion
+ *
+ * @param[in]   q    quaternion
+ */
+CGLM_INLINE
+float
+glm_quat_imaglen(versor q) {
+  return glm_vec_norm(q);
+}
+
+/*!
+ * @brief returns angle of quaternion
+ *
+ * @param[in]   q    quaternion
+ */
+CGLM_INLINE
+float
+glm_quat_angle(versor q) {
+  /*
+   sin(theta / 2) = length(x*x + y*y + z*z)
+   cos(theta / 2) = w
+   theta          = 2 * atan(sin(theta / 2) / cos(theta / 2))
+   */
+  return 2.0f * atan2f(glm_quat_imaglen(q), glm_quat_real(q));
+}
+
+/*!
+ * @brief axis of quaternion
+ *
+ * @param[in]   q    quaternion
+ * @param[out]  dest axis of quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_axis(versor q, versor dest) {
+  glm_quat_imagn(q, dest);
+}
+
+/*!
+ * @brief multiplies two quaternion and stores result in dest
+ *        this is also called Hamilton Product
+ *
+ * According to WikiPedia:
+ * The product of two rotation quaternions [clarification needed] will be
+ * equivalent to the rotation q followed by the rotation p
+ *
+ * @param[in]   p     quaternion 1
+ * @param[in]   q     quaternion 2
+ * @param[out]  dest  result quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_mul(versor p, versor q, versor dest) {
+  /*
+    + (a1 b2 + b1 a2 + c1 d2 − d1 c2)i
+    + (a1 c2 − b1 d2 + c1 a2 + d1 b2)j
+    + (a1 d2 + b1 c2 − c1 b2 + d1 a2)k
+       a1 a2 − b1 b2 − c1 c2 − d1 d2
+   */
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glm_quat_mul_sse2(p, q, dest);
+#else
+  dest[0] = p[3] * q[0] + p[0] * q[3] + p[1] * q[2] - p[2] * q[1];
+  dest[1] = p[3] * q[1] - p[0] * q[2] + p[1] * q[3] + p[2] * q[0];
+  dest[2] = p[3] * q[2] + p[0] * q[1] - p[1] * q[0] + p[2] * q[3];
+  dest[3] = p[3] * q[3] - p[0] * q[0] - p[1] * q[1] - p[2] * q[2];
+#endif
 }
 
 /*!
@@ -181,19 +413,22 @@ glm_quat_mulv(versor q1, versor q2, versor dest) {
 CGLM_INLINE
 void
 glm_quat_mat4(versor q, mat4 dest) {
-  float w, x, y, z;
-  float xx, yy, zz;
-  float xy, yz, xz;
-  float wx, wy, wz;
+  float w, x, y, z,
+        xx, yy, zz,
+        xy, yz, xz,
+        wx, wy, wz, norm, s;
 
-  w = q[0];
-  x = q[1];
-  y = q[2];
-  z = q[3];
+  norm = glm_quat_norm(q);
+  s    = norm > 0.0f ? 2.0f / norm : 0.0f;
 
-  xx = 2.0f * x * x;   xy = 2.0f * x * y;   wx = 2.0f * w * x;
-  yy = 2.0f * y * y;   yz = 2.0f * y * z;   wy = 2.0f * w * y;
-  zz = 2.0f * z * z;   xz = 2.0f * x * z;   wz = 2.0f * w * z;
+  x = q[0];
+  y = q[1];
+  z = q[2];
+  w = q[3];
+
+  xx = s * x * x;   xy = s * x * y;   wx = s * w * x;
+  yy = s * y * y;   yz = s * y * z;   wy = s * w * y;
+  zz = s * z * z;   xz = s * x * z;   wz = s * w * z;
 
   dest[0][0] = 1.0f - yy - zz;
   dest[1][1] = 1.0f - xx - zz;
@@ -207,8 +442,8 @@ glm_quat_mat4(versor q, mat4 dest) {
   dest[2][1] = yz - wx;
   dest[0][2] = xz - wy;
 
-  dest[1][3] = 0.0f;
   dest[0][3] = 0.0f;
+  dest[1][3] = 0.0f;
   dest[2][3] = 0.0f;
   dest[3][0] = 0.0f;
   dest[3][1] = 0.0f;
@@ -216,69 +451,343 @@ glm_quat_mat4(versor q, mat4 dest) {
   dest[3][3] = 1.0f;
 }
 
+/*!
+ * @brief convert quaternion to mat4 (transposed)
+ *
+ * @param[in]   q     quaternion
+ * @param[out]  dest  result matrix as transposed
+ */
+CGLM_INLINE
+void
+glm_quat_mat4t(versor q, mat4 dest) {
+  float w, x, y, z,
+        xx, yy, zz,
+        xy, yz, xz,
+        wx, wy, wz, norm, s;
+
+  norm = glm_quat_norm(q);
+  s    = norm > 0.0f ? 2.0f / norm : 0.0f;
+
+  x = q[0];
+  y = q[1];
+  z = q[2];
+  w = q[3];
+
+  xx = s * x * x;   xy = s * x * y;   wx = s * w * x;
+  yy = s * y * y;   yz = s * y * z;   wy = s * w * y;
+  zz = s * z * z;   xz = s * x * z;   wz = s * w * z;
+
+  dest[0][0] = 1.0f - yy - zz;
+  dest[1][1] = 1.0f - xx - zz;
+  dest[2][2] = 1.0f - xx - yy;
+
+  dest[1][0] = xy + wz;
+  dest[2][1] = yz + wx;
+  dest[0][2] = xz + wy;
+
+  dest[0][1] = xy - wz;
+  dest[1][2] = yz - wx;
+  dest[2][0] = xz - wy;
+
+  dest[0][3] = 0.0f;
+  dest[1][3] = 0.0f;
+  dest[2][3] = 0.0f;
+  dest[3][0] = 0.0f;
+  dest[3][1] = 0.0f;
+  dest[3][2] = 0.0f;
+  dest[3][3] = 1.0f;
+}
+
+/*!
+ * @brief convert quaternion to mat3
+ *
+ * @param[in]   q     quaternion
+ * @param[out]  dest  result matrix
+ */
+CGLM_INLINE
+void
+glm_quat_mat3(versor q, mat3 dest) {
+  float w, x, y, z,
+        xx, yy, zz,
+        xy, yz, xz,
+        wx, wy, wz, norm, s;
+
+  norm = glm_quat_norm(q);
+  s    = norm > 0.0f ? 2.0f / norm : 0.0f;
+
+  x = q[0];
+  y = q[1];
+  z = q[2];
+  w = q[3];
+
+  xx = s * x * x;   xy = s * x * y;   wx = s * w * x;
+  yy = s * y * y;   yz = s * y * z;   wy = s * w * y;
+  zz = s * z * z;   xz = s * x * z;   wz = s * w * z;
+
+  dest[0][0] = 1.0f - yy - zz;
+  dest[1][1] = 1.0f - xx - zz;
+  dest[2][2] = 1.0f - xx - yy;
+
+  dest[0][1] = xy + wz;
+  dest[1][2] = yz + wx;
+  dest[2][0] = xz + wy;
+
+  dest[1][0] = xy - wz;
+  dest[2][1] = yz - wx;
+  dest[0][2] = xz - wy;
+}
+
+/*!
+ * @brief convert quaternion to mat3 (transposed)
+ *
+ * @param[in]   q     quaternion
+ * @param[out]  dest  result matrix
+ */
+CGLM_INLINE
+void
+glm_quat_mat3t(versor q, mat3 dest) {
+  float w, x, y, z,
+        xx, yy, zz,
+        xy, yz, xz,
+        wx, wy, wz, norm, s;
+
+  norm = glm_quat_norm(q);
+  s    = norm > 0.0f ? 2.0f / norm : 0.0f;
+
+  x = q[0];
+  y = q[1];
+  z = q[2];
+  w = q[3];
+
+  xx = s * x * x;   xy = s * x * y;   wx = s * w * x;
+  yy = s * y * y;   yz = s * y * z;   wy = s * w * y;
+  zz = s * z * z;   xz = s * x * z;   wz = s * w * z;
+
+  dest[0][0] = 1.0f - yy - zz;
+  dest[1][1] = 1.0f - xx - zz;
+  dest[2][2] = 1.0f - xx - yy;
+
+  dest[1][0] = xy + wz;
+  dest[2][1] = yz + wx;
+  dest[0][2] = xz + wy;
+
+  dest[0][1] = xy - wz;
+  dest[1][2] = yz - wx;
+  dest[2][0] = xz - wy;
+}
+
+/*!
+ * @brief interpolates between two quaternions
+ *        using linear interpolation (LERP)
+ *
+ * @param[in]   from  from
+ * @param[in]   to    to
+ * @param[in]   t     interpolant (amount) clamped between 0 and 1
+ * @param[out]  dest  result quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_lerp(versor from, versor to, float t, versor dest) {
+  glm_vec4_lerp(from, to, t, dest);
+}
+
 /*!
  * @brief interpolates between two quaternions
  *        using spherical linear interpolation (SLERP)
  *
- * @param[in]   q     from
- * @param[in]   r     to
+ * @param[in]   from  from
+ * @param[in]   to    to
  * @param[in]   t     amout
  * @param[out]  dest  result quaternion
  */
 CGLM_INLINE
 void
-glm_quat_slerp(versor q,
-               versor r,
-               float  t,
-               versor dest) {
-  /* https://en.wikipedia.org/wiki/Slerp */
-#if defined( __SSE__ ) || defined( __SSE2__ )
-  glm_quat_slerp_sse2(q, r, t, dest);
-#else
-  float cosTheta, sinTheta, angle, a, b, c;
+glm_quat_slerp(versor from, versor to, float t, versor dest) {
+  CGLM_ALIGN(16) vec4 q1, q2;
+  float cosTheta, sinTheta, angle;
 
-  cosTheta = glm_quat_dot(q, r);
-  if (cosTheta < 0.0f) {
-    q[0] *= -1.0f;
-    q[1] *= -1.0f;
-    q[2] *= -1.0f;
-    q[3] *= -1.0f;
+  cosTheta = glm_quat_dot(from, to);
+  glm_quat_copy(from, q1);
 
-    cosTheta = -cosTheta;
-  }
-
-  if (fabs(cosTheta) >= 1.0f) {
-    dest[0] = q[0];
-    dest[1] = q[1];
-    dest[2] = q[2];
-    dest[3] = q[3];
+  if (fabsf(cosTheta) >= 1.0f) {
+    glm_quat_copy(q1, dest);
     return;
   }
 
-  sinTheta = sqrt(1.0f - cosTheta * cosTheta);
+  if (cosTheta < 0.0f) {
+    glm_vec4_flipsign(q1);
+    cosTheta = -cosTheta;
+  }
 
-  c = 1.0f - t;
+  sinTheta = sqrtf(1.0f - cosTheta * cosTheta);
 
-  /* LERP */
-  /* TODO: FLT_EPSILON vs 0.001? */
-  if (sinTheta < 0.001f) {
-    dest[0] = c * q[0] + t * r[0];
-    dest[1] = c * q[1] + t * r[1];
-    dest[2] = c * q[2] + t * r[2];
-    dest[3] = c * q[3] + t * r[3];
+  /* LERP to avoid zero division */
+  if (fabsf(sinTheta) < 0.001f) {
+    glm_quat_lerp(from, to, t, dest);
     return;
   }
 
   /* SLERP */
   angle = acosf(cosTheta);
-  a = sinf(c * angle);
-  b = sinf(t * angle);
+  glm_vec4_scale(q1, sinf((1.0f - t) * angle), q1);
+  glm_vec4_scale(to, sinf(t * angle), q2);
 
-  dest[0] = (q[0] * a + r[0] * b) / sinTheta;
-  dest[1] = (q[1] * a + r[1] * b) / sinTheta;
-  dest[2] = (q[2] * a + r[2] * b) / sinTheta;
-  dest[3] = (q[3] * a + r[3] * b) / sinTheta;
-#endif
+  glm_vec4_add(q1, q2, q1);
+  glm_vec4_scale(q1, 1.0f / sinTheta, dest);
+}
+
+/*!
+ * @brief creates view matrix using quaternion as camera orientation
+ *
+ * @param[in]   eye   eye
+ * @param[in]   ori   orientation in world space as quaternion
+ * @param[out]  dest  view matrix
+ */
+CGLM_INLINE
+void
+glm_quat_look(vec3 eye, versor ori, mat4 dest) {
+  /* orientation */
+  glm_quat_mat4t(ori, dest);
+
+  /* translate */
+  glm_mat4_mulv3(dest, eye, 1.0f, dest[3]);
+  glm_vec_flipsign(dest[3]);
+}
+
+/*!
+ * @brief creates look rotation quaternion
+ *
+ * @param[in]   dir   direction to look
+ * @param[in]   fwd   forward vector
+ * @param[in]   up    up vector
+ * @param[out]  dest  destination quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_for(vec3 dir, vec3 fwd, vec3 up, versor dest) {
+  CGLM_ALIGN(8) vec3 axis;
+  float dot, angle;
+
+  dot = glm_vec_dot(dir, fwd);
+  if (fabsf(dot + 1.0f)  < 0.000001f) {
+    glm_quat_init(dest, up[0], up[1], up[2], CGLM_PI);
+    return;
+  }
+
+  if (fabsf(dot - 1.0f) < 0.000001f) {
+    glm_quat_identity(dest);
+    return;
+  }
+
+  angle = acosf(dot);
+  glm_cross(fwd, dir, axis);
+  glm_normalize(axis);
+
+  glm_quatv(dest, angle, axis);
+}
+
+/*!
+ * @brief creates look rotation quaternion using source and
+ *        destination positions p suffix stands for position
+ *
+ * @param[in]   from  source point
+ * @param[in]   to    destination point
+ * @param[in]   fwd   forward vector
+ * @param[in]   up    up vector
+ * @param[out]  dest  destination quaternion
+ */
+CGLM_INLINE
+void
+glm_quat_forp(vec3 from, vec3 to, vec3 fwd, vec3 up, versor dest) {
+  CGLM_ALIGN(8) vec3 dir;
+  glm_vec_sub(to, from, dir);
+  glm_quat_for(dir, fwd, up, dest);
+}
+
+/*!
+ * @brief rotate vector using using quaternion
+ *
+ * @param[in]   q     quaternion
+ * @param[in]   v     vector to rotate
+ * @param[out]  dest  rotated vector
+ */
+CGLM_INLINE
+void
+glm_quat_rotatev(versor q, vec3 v, vec3 dest) {
+  CGLM_ALIGN(16) versor p;
+  CGLM_ALIGN(8)  vec3   u, v1, v2;
+  float s;
+
+  glm_quat_normalize_to(q, p);
+  glm_quat_imag(p, u);
+  s = glm_quat_real(p);
+
+  glm_vec_scale(u, 2.0f * glm_vec_dot(u, v), v1);
+  glm_vec_scale(v, s * s - glm_vec_dot(u, u), v2);
+  glm_vec_add(v1, v2, v1);
+
+  glm_vec_cross(u, v, v2);
+  glm_vec_scale(v2, 2.0f * s, v2);
+
+  glm_vec_add(v1, v2, dest);
+}
+
+/*!
+ * @brief rotate existing transform matrix using quaternion
+ *
+ * @param[in]   m     existing transform matrix
+ * @param[in]   q     quaternion
+ * @param[out]  dest  rotated matrix/transform
+ */
+CGLM_INLINE
+void
+glm_quat_rotate(mat4 m, versor q, mat4 dest) {
+  CGLM_ALIGN(16) mat4 rot;
+  glm_quat_mat4(q, rot);
+  glm_mul_rot(m, rot, dest);
+}
+
+/*!
+ * @brief rotate existing transform matrix using quaternion at pivot point
+ *
+ * @param[in, out]   m     existing transform matrix
+ * @param[in]        q     quaternion
+ * @param[out]       pivot pivot
+ */
+CGLM_INLINE
+void
+glm_quat_rotate_at(mat4 m, versor q, vec3 pivot) {
+  CGLM_ALIGN(8) vec3 pivotInv;
+
+  glm_vec_inv_to(pivot, pivotInv);
+
+  glm_translate(m, pivot);
+  glm_quat_rotate(m, q, m);
+  glm_translate(m, pivotInv);
+}
+
+/*!
+ * @brief rotate NEW transform matrix using quaternion at pivot point
+ *
+ * this creates rotation matrix, it assumes you don't have a matrix
+ *
+ * this should work faster than glm_quat_rotate_at because it reduces
+ * one glm_translate.
+ *
+ * @param[out]  m     existing transform matrix
+ * @param[in]   q     quaternion
+ * @param[in]   pivot pivot
+ */
+CGLM_INLINE
+void
+glm_quat_rotate_atm(mat4 m, versor q, vec3 pivot) {
+  CGLM_ALIGN(8) vec3 pivotInv;
+
+  glm_vec_inv_to(pivot, pivotInv);
+
+  glm_translate_make(m, pivot);
+  glm_quat_rotate(m, q, m);
+  glm_translate(m, pivotInv);
 }
 
 #endif /* cglm_quat_h */

include/cglm/simd/avx/affine.h

@@ -21,11 +21,11 @@ glm_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
 
   __m256 y0, y1, y2, y3, y4, y5, y6, y7, y8, y9;
 
-  y0 = _mm256_load_ps(m2[0]); /* h g f e d c b a */
-  y1 = _mm256_load_ps(m2[2]); /* p o n m l k j i */
+  y0 = glmm_load256(m2[0]); /* h g f e d c b a */
+  y1 = glmm_load256(m2[2]); /* p o n m l k j i */
 
-  y2 = _mm256_load_ps(m1[0]); /* h g f e d c b a */
-  y3 = _mm256_load_ps(m1[2]); /* p o n m l k j i */
+  y2 = glmm_load256(m1[0]); /* h g f e d c b a */
+  y3 = glmm_load256(m1[2]); /* p o n m l k j i */
 
   y4 = _mm256_permute2f128_ps(y2, y2, 0b00000011); /* d c b a h g f e */
   y5 = _mm256_permute2f128_ps(y3, y3, 0b00000000); /* l k j i l k j i */
@@ -37,7 +37,7 @@ glm_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
   y6 = _mm256_permutevar_ps(y0, _mm256_set_epi32(1, 1, 1, 1, 0, 0, 0, 0));
   y8 = _mm256_permutevar_ps(y0, _mm256_set_epi32(0, 0, 0, 0, 1, 1, 1, 1));
 
-  _mm256_store_ps(dest[0],
+  glmm_store256(dest[0],
                 _mm256_add_ps(_mm256_add_ps(_mm256_mul_ps(y2, y6),
                                             _mm256_mul_ps(y4, y8)),
                               _mm256_mul_ps(y5, y7)));
@@ -52,7 +52,7 @@ glm_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
   y8 = _mm256_permutevar_ps(y1, _mm256_set_epi32(0, 0, 0, 0, 1, 1, 1, 1));
   y9 = _mm256_permutevar_ps(y1, _mm256_set_epi32(2, 2, 2, 2, 3, 3, 3, 3));
 
-  _mm256_store_ps(dest[2],
+  glmm_store256(dest[2],
                 _mm256_add_ps(_mm256_add_ps(_mm256_mul_ps(y2, y6),
                                             _mm256_mul_ps(y3, y7)),
                               _mm256_add_ps(_mm256_mul_ps(y4, y8),

include/cglm/simd/avx/mat4.h

@@ -21,11 +21,11 @@ glm_mat4_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
 
   __m256 y0, y1, y2, y3, y4, y5, y6, y7, y8, y9;
 
-  y0 = _mm256_load_ps(m2[0]); /* h g f e d c b a */
-  y1 = _mm256_load_ps(m2[2]); /* p o n m l k j i */
+  y0 = glmm_load256(m2[0]); /* h g f e d c b a */
+  y1 = glmm_load256(m2[2]); /* p o n m l k j i */
 
-  y2 = _mm256_load_ps(m1[0]); /* h g f e d c b a */
-  y3 = _mm256_load_ps(m1[2]); /* p o n m l k j i */
+  y2 = glmm_load256(m1[0]); /* h g f e d c b a */
+  y3 = glmm_load256(m1[2]); /* p o n m l k j i */
 
   y4 = _mm256_permute2f128_ps(y2, y2, 0b00000011); /* d c b a h g f e */
   y5 = _mm256_permute2f128_ps(y3, y3, 0b00000011); /* l k j i p o n m */
@@ -39,7 +39,7 @@ glm_mat4_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
   y8 = _mm256_permutevar_ps(y0, _mm256_set_epi32(0, 0, 0, 0, 1, 1, 1, 1));
   y9 = _mm256_permutevar_ps(y0, _mm256_set_epi32(2, 2, 2, 2, 3, 3, 3, 3));
 
-  _mm256_store_ps(dest[0],
+  glmm_store256(dest[0],
                 _mm256_add_ps(_mm256_add_ps(_mm256_mul_ps(y2, y6),
                                             _mm256_mul_ps(y3, y7)),
                               _mm256_add_ps(_mm256_mul_ps(y4, y8),
@@ -54,7 +54,7 @@ glm_mat4_mul_avx(mat4 m1, mat4 m2, mat4 dest) {
   y8 = _mm256_permutevar_ps(y1, _mm256_set_epi32(0, 0, 0, 0, 1, 1, 1, 1));
   y9 = _mm256_permutevar_ps(y1, _mm256_set_epi32(2, 2, 2, 2, 3, 3, 3, 3));
 
-  _mm256_store_ps(dest[2],
+  glmm_store256(dest[2],
                 _mm256_add_ps(_mm256_add_ps(_mm256_mul_ps(y2, y6),
                                             _mm256_mul_ps(y3, y7)),
                               _mm256_add_ps(_mm256_mul_ps(y4, y8),

include/cglm/simd/intrin.h

@@ -8,28 +8,83 @@
 #ifndef cglm_intrin_h
 #define cglm_intrin_h
 
-#if defined( _WIN32 )
+#if defined( _MSC_VER )
 #  if (defined(_M_AMD64) || defined(_M_X64)) || _M_IX86_FP == 2
+#    ifndef __SSE2__
 #      define __SSE2__
+#    endif
 #  elif _M_IX86_FP == 1
+#    ifndef __SSE__
 #      define __SSE__
 #    endif
+#  endif
+/* do not use alignment for older visual studio versions */
+#  if _MSC_VER < 1913     /* Visual Studio 2017 version 15.6 */
+#    define CGLM_ALL_UNALIGNED
+#  endif
 #endif
 
 #if defined( __SSE__ ) || defined( __SSE2__ )
 #  include <xmmintrin.h>
 #  include <emmintrin.h>
 
-/* float */
-#  define _mm_shuffle1_ps(a, z, y, x, w)                                       \
-     _mm_shuffle_ps(a, a, _MM_SHUFFLE(z, y, x, w))
+/* OPTIONAL: You may save some instructions but latency (not sure) */
+#ifdef CGLM_USE_INT_DOMAIN
+#  define glmm_shuff1(xmm, z, y, x, w)                                        \
+     _mm_castsi128_ps(_mm_shuffle_epi32(_mm_castps_si128(xmm),                \
+                                        _MM_SHUFFLE(z, y, x, w)))
+#else
+#  define glmm_shuff1(xmm, z, y, x, w)                                        \
+     _mm_shuffle_ps(xmm, xmm, _MM_SHUFFLE(z, y, x, w))
+#endif
 
-#  define _mm_shuffle1_ps1(a, x)                                               \
-     _mm_shuffle_ps(a, a, _MM_SHUFFLE(x, x, x, x))
-
-#  define _mm_shuffle2_ps(a, b, z0, y0, x0, w0, z1, y1, x1, w1)                \
-     _mm_shuffle1_ps(_mm_shuffle_ps(a, b, _MM_SHUFFLE(z0, y0, x0, w0)),        \
+#define glmm_shuff1x(xmm, x) glmm_shuff1(xmm, x, x, x, x)
+#define glmm_shuff2(a, b, z0, y0, x0, w0, z1, y1, x1, w1)                     \
+     glmm_shuff1(_mm_shuffle_ps(a, b, _MM_SHUFFLE(z0, y0, x0, w0)),           \
                  z1, y1, x1, w1)
+
+static inline
+__m128
+glmm_dot(__m128 a, __m128 b) {
+  __m128 x0;
+  x0 = _mm_mul_ps(a, b);
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 1, 0, 3, 2));
+  return _mm_add_ps(x0, glmm_shuff1(x0, 0, 1, 0, 1));
+}
+
+static inline
+__m128
+glmm_norm(__m128 a) {
+  return _mm_sqrt_ps(glmm_dot(a, a));
+}
+
+static inline
+__m128
+glmm_load3(float v[3]) {
+  __m128i xy;
+  __m128  z;
+
+  xy = _mm_loadl_epi64((const __m128i *)v);
+  z  = _mm_load_ss(&v[2]);
+
+  return _mm_movelh_ps(_mm_castsi128_ps(xy), z);
+}
+
+static inline
+void
+glmm_store3(__m128 vx, float v[3]) {
+  _mm_storel_pi((__m64 *)&v[0], vx);
+  _mm_store_ss(&v[2], glmm_shuff1(vx, 2, 2, 2, 2));
+}
+
+#ifdef CGLM_ALL_UNALIGNED
+#  define glmm_load(p)      _mm_loadu_ps(p)
+#  define glmm_store(p, a)  _mm_storeu_ps(p, a)
+#else
+#  define glmm_load(p)      _mm_load_ps(p)
+#  define glmm_store(p, a)  _mm_store_ps(p, a)
+#endif
+
 #endif
 
 /* x86, x64 */
@@ -39,6 +94,15 @@
 
 #ifdef __AVX__
 #  define CGLM_AVX_FP 1
+
+#ifdef CGLM_ALL_UNALIGNED
+#  define glmm_load256(p)      _mm256_loadu_ps(p)
+#  define glmm_store256(p, a)  _mm256_storeu_ps(p, a)
+#else
+#  define glmm_load256(p)      _mm256_load_ps(p)
+#  define glmm_store256(p, a)  _mm256_store_ps(p, a)
+#endif
+
 #endif
 
 /* ARM Neon */

include/cglm/simd/sse2/affine.h

@@ -18,35 +18,67 @@ glm_mul_sse2(mat4 m1, mat4 m2, mat4 dest) {
   /* D = R * L (Column-Major) */
   __m128 l0, l1, l2, l3, r;
 
-  l0 = _mm_load_ps(m1[0]);
-  l1 = _mm_load_ps(m1[1]);
-  l2 = _mm_load_ps(m1[2]);
-  l3 = _mm_load_ps(m1[3]);
+  l0 = glmm_load(m1[0]);
+  l1 = glmm_load(m1[1]);
+  l2 = glmm_load(m1[2]);
+  l3 = glmm_load(m1[3]);
 
-  r = _mm_load_ps(m2[0]);
-  _mm_store_ps(dest[0],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2)));
+  r = glmm_load(m2[0]);
+  glmm_store(dest[0],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_mul_ps(glmm_shuff1x(r, 2), l2)));
 
-  r = _mm_load_ps(m2[1]);
-  _mm_store_ps(dest[1],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2)));
+  r = glmm_load(m2[1]);
+  glmm_store(dest[1],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_mul_ps(glmm_shuff1x(r, 2), l2)));
 
-  r = _mm_load_ps(m2[2]);
-  _mm_store_ps(dest[2],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2)));
+  r = glmm_load(m2[2]);
+  glmm_store(dest[2],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_mul_ps(glmm_shuff1x(r, 2), l2)));
 
-  r = _mm_load_ps(m2[3]);
-  _mm_store_ps(dest[3],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 3), l3))));
+  r = glmm_load(m2[3]);
+  glmm_store(dest[3],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 2), l2),
+                                   _mm_mul_ps(glmm_shuff1x(r, 3), l3))));
+}
+
+CGLM_INLINE
+void
+glm_mul_rot_sse2(mat4 m1, mat4 m2, mat4 dest) {
+  /* D = R * L (Column-Major) */
+  __m128 l0, l1, l2, l3, r;
+
+  l0 = glmm_load(m1[0]);
+  l1 = glmm_load(m1[1]);
+  l2 = glmm_load(m1[2]);
+  l3 = glmm_load(m1[3]);
+
+  r = glmm_load(m2[0]);
+  glmm_store(dest[0],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_mul_ps(glmm_shuff1x(r, 2), l2)));
+
+  r = glmm_load(m2[1]);
+  glmm_store(dest[1],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_mul_ps(glmm_shuff1x(r, 2), l2)));
+
+  r = glmm_load(m2[2]);
+  glmm_store(dest[2],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_mul_ps(glmm_shuff1x(r, 2), l2)));
+
+  glmm_store(dest[3], l3);
 }
 
 CGLM_INLINE
@@ -54,25 +86,25 @@ void
 glm_inv_tr_sse2(mat4 mat) {
   __m128 r0, r1, r2, r3, x0, x1;
 
-  r0 = _mm_load_ps(mat[0]);
-  r1 = _mm_load_ps(mat[1]);
-  r2 = _mm_load_ps(mat[2]);
-  r3 = _mm_load_ps(mat[3]);
+  r0 = glmm_load(mat[0]);
+  r1 = glmm_load(mat[1]);
+  r2 = glmm_load(mat[2]);
+  r3 = glmm_load(mat[3]);
   x1 = _mm_set_ps(1.0f, 0.0f, 0.0f, 0.0f);
 
   _MM_TRANSPOSE4_PS(r0, r1, r2, x1);
 
-  x0 = _mm_add_ps(_mm_mul_ps(r0, _mm_shuffle1_ps(r3, 0, 0, 0, 0)),
-                  _mm_mul_ps(r1, _mm_shuffle1_ps(r3, 1, 1, 1, 1)));
-  x0 = _mm_add_ps(x0, _mm_mul_ps(r2, _mm_shuffle1_ps(r3, 2, 2, 2, 2)));
+  x0 = _mm_add_ps(_mm_mul_ps(r0, glmm_shuff1(r3, 0, 0, 0, 0)),
+                  _mm_mul_ps(r1, glmm_shuff1(r3, 1, 1, 1, 1)));
+  x0 = _mm_add_ps(x0, _mm_mul_ps(r2, glmm_shuff1(r3, 2, 2, 2, 2)));
   x0 = _mm_xor_ps(x0, _mm_set1_ps(-0.f));
 
   x0 = _mm_add_ps(x0, x1);
 
-  _mm_store_ps(mat[0], r0);
-  _mm_store_ps(mat[1], r1);
-  _mm_store_ps(mat[2], r2);
-  _mm_store_ps(mat[3], x0);
+  glmm_store(mat[0], r0);
+  glmm_store(mat[1], r1);
+  glmm_store(mat[2], r2);
+  glmm_store(mat[3], x0);
 }
 
 #endif

include/cglm/simd/sse2/mat3.h

@@ -27,27 +27,25 @@ glm_mat3_mul_sse2(mat3 m1, mat3 m2, mat3 dest) {
   r1 = _mm_loadu_ps(&m2[1][1]);
   r2 = _mm_set1_ps(m2[2][2]);
 
-  x1 = _mm_shuffle2_ps(l0, l1, 1, 0, 3, 3, 0, 3, 2, 0);
-  x2 = _mm_shuffle2_ps(l1, l2, 0, 0, 3, 2, 0, 2, 1, 0);
+  x1 = glmm_shuff2(l0, l1, 1, 0, 3, 3, 0, 3, 2, 0);
+  x2 = glmm_shuff2(l1, l2, 0, 0, 3, 2, 0, 2, 1, 0);
 
-  x0 = _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps(l0, 0, 2, 1, 0),
-                             _mm_shuffle1_ps(r0, 3, 0, 0, 0)),
-                  _mm_mul_ps(x1,
-                             _mm_shuffle2_ps(r0, r1, 0, 0, 1, 1, 2, 0, 0, 0)));
+  x0 = _mm_add_ps(_mm_mul_ps(glmm_shuff1(l0, 0, 2, 1, 0),
+                             glmm_shuff1(r0, 3, 0, 0, 0)),
+                  _mm_mul_ps(x1, glmm_shuff2(r0, r1, 0, 0, 1, 1, 2, 0, 0, 0)));
 
   x0 = _mm_add_ps(x0,
-                  _mm_mul_ps(x2,
-                             _mm_shuffle2_ps(r0, r1, 1, 1, 2, 2, 2, 0, 0, 0)));
+                  _mm_mul_ps(x2, glmm_shuff2(r0, r1, 1, 1, 2, 2, 2, 0, 0, 0)));
 
   _mm_storeu_ps(dest[0], x0);
 
-  x0 = _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps(l0, 1, 0, 2, 1),
+  x0 = _mm_add_ps(_mm_mul_ps(glmm_shuff1(l0, 1, 0, 2, 1),
                              _mm_shuffle_ps(r0, r1, _MM_SHUFFLE(2, 2, 3, 3))),
-                  _mm_mul_ps(_mm_shuffle1_ps(x1, 1, 0, 2, 1),
-                             _mm_shuffle1_ps(r1, 3, 3, 0, 0)));
+                  _mm_mul_ps(glmm_shuff1(x1, 1, 0, 2, 1),
+                             glmm_shuff1(r1, 3, 3, 0, 0)));
 
   x0 = _mm_add_ps(x0,
-                  _mm_mul_ps(_mm_shuffle1_ps(x2, 1, 0, 2, 1),
+                  _mm_mul_ps(glmm_shuff1(x2, 1, 0, 2, 1),
                              _mm_shuffle_ps(r1, r2, _MM_SHUFFLE(0, 0, 1, 1))));
 
   _mm_storeu_ps(&dest[1][1], x0);

include/cglm/simd/sse2/mat4.h

@@ -20,10 +20,10 @@ glm_mat4_scale_sse2(mat4 m, float s){
   __m128 x0;
   x0 = _mm_set1_ps(s);
 
-  _mm_store_ps(m[0], _mm_mul_ps(_mm_load_ps(m[0]), x0));
-  _mm_store_ps(m[1], _mm_mul_ps(_mm_load_ps(m[1]), x0));
-  _mm_store_ps(m[2], _mm_mul_ps(_mm_load_ps(m[2]), x0));
-  _mm_store_ps(m[3], _mm_mul_ps(_mm_load_ps(m[3]), x0));
+  glmm_store(m[0], _mm_mul_ps(glmm_load(m[0]), x0));
+  glmm_store(m[1], _mm_mul_ps(glmm_load(m[1]), x0));
+  glmm_store(m[2], _mm_mul_ps(glmm_load(m[2]), x0));
+  glmm_store(m[3], _mm_mul_ps(glmm_load(m[3]), x0));
 }
 
 CGLM_INLINE
@@ -31,17 +31,17 @@ void
 glm_mat4_transp_sse2(mat4 m, mat4 dest){
   __m128 r0, r1, r2, r3;
 
-  r0 = _mm_load_ps(m[0]);
-  r1 = _mm_load_ps(m[1]);
-  r2 = _mm_load_ps(m[2]);
-  r3 = _mm_load_ps(m[3]);
+  r0 = glmm_load(m[0]);
+  r1 = glmm_load(m[1]);
+  r2 = glmm_load(m[2]);
+  r3 = glmm_load(m[3]);
 
   _MM_TRANSPOSE4_PS(r0, r1, r2, r3);
 
-  _mm_store_ps(dest[0], r0);
-  _mm_store_ps(dest[1], r1);
-  _mm_store_ps(dest[2], r2);
-  _mm_store_ps(dest[3], r3);
+  glmm_store(dest[0], r0);
+  glmm_store(dest[1], r1);
+  glmm_store(dest[2], r2);
+  glmm_store(dest[3], r3);
 }
 
 CGLM_INLINE
@@ -51,36 +51,36 @@ glm_mat4_mul_sse2(mat4 m1, mat4 m2, mat4 dest) {
 
   __m128 l0, l1, l2, l3, r;
 
-  l0 = _mm_load_ps(m1[0]);
-  l1 = _mm_load_ps(m1[1]);
-  l2 = _mm_load_ps(m1[2]);
-  l3 = _mm_load_ps(m1[3]);
+  l0 = glmm_load(m1[0]);
+  l1 = glmm_load(m1[1]);
+  l2 = glmm_load(m1[2]);
+  l3 = glmm_load(m1[3]);
 
-  r = _mm_load_ps(m2[0]);
-  _mm_store_ps(dest[0],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 3), l3))));
-  r = _mm_load_ps(m2[1]);
-  _mm_store_ps(dest[1],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 3), l3))));
-  r = _mm_load_ps(m2[2]);
-  _mm_store_ps(dest[2],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 3), l3))));
+  r = glmm_load(m2[0]);
+  glmm_store(dest[0],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 2), l2),
+                                   _mm_mul_ps(glmm_shuff1x(r, 3), l3))));
+  r = glmm_load(m2[1]);
+  glmm_store(dest[1],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 2), l2),
+                                   _mm_mul_ps(glmm_shuff1x(r, 3), l3))));
+  r = glmm_load(m2[2]);
+  glmm_store(dest[2],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 2), l2),
+                                   _mm_mul_ps(glmm_shuff1x(r, 3), l3))));
 
-  r = _mm_load_ps(m2[3]);
-  _mm_store_ps(dest[3],
-               _mm_add_ps(_mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 0), l0),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 1), l1)),
-                          _mm_add_ps(_mm_mul_ps(_mm_shuffle1_ps1(r, 2), l2),
-                                     _mm_mul_ps(_mm_shuffle1_ps1(r, 3), l3))));
+  r = glmm_load(m2[3]);
+  glmm_store(dest[3],
+             _mm_add_ps(_mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 0), l0),
+                                   _mm_mul_ps(glmm_shuff1x(r, 1), l1)),
+                        _mm_add_ps(_mm_mul_ps(glmm_shuff1x(r, 2), l2),
+                                   _mm_mul_ps(glmm_shuff1x(r, 3), l3))));
 }
 
 CGLM_INLINE
@@ -88,18 +88,14 @@ void
 glm_mat4_mulv_sse2(mat4 m, vec4 v, vec4 dest) {
   __m128 x0, x1, x2;
 
-  x0 = _mm_load_ps(v);
-  x1 = _mm_add_ps(_mm_mul_ps(_mm_load_ps(m[0]),
-                             _mm_shuffle1_ps1(x0, 0)),
-                  _mm_mul_ps(_mm_load_ps(m[1]),
-                             _mm_shuffle1_ps1(x0, 1)));
+  x0 = glmm_load(v);
+  x1 = _mm_add_ps(_mm_mul_ps(glmm_load(m[0]), glmm_shuff1x(x0, 0)),
+                  _mm_mul_ps(glmm_load(m[1]), glmm_shuff1x(x0, 1)));
 
-  x2 = _mm_add_ps(_mm_mul_ps(_mm_load_ps(m[2]),
-                             _mm_shuffle1_ps1(x0, 2)),
-                  _mm_mul_ps(_mm_load_ps(m[3]),
-                             _mm_shuffle1_ps1(x0, 3)));
+  x2 = _mm_add_ps(_mm_mul_ps(glmm_load(m[2]), glmm_shuff1x(x0, 2)),
+                  _mm_mul_ps(glmm_load(m[3]), glmm_shuff1x(x0, 3)));
 
-  _mm_store_ps(dest, _mm_add_ps(x1, x2));
+  glmm_store(dest, _mm_add_ps(x1, x2));
 }
 
 CGLM_INLINE
@@ -108,10 +104,10 @@ glm_mat4_det_sse2(mat4 mat) {
   __m128 r0, r1, r2, r3, x0, x1, x2;
 
   /* 127 <- 0, [square] det(A) = det(At) */
-  r0 = _mm_load_ps(mat[0]); /* d c b a */
-  r1 = _mm_load_ps(mat[1]); /* h g f e */
-  r2 = _mm_load_ps(mat[2]); /* l k j i */
-  r3 = _mm_load_ps(mat[3]); /* p o n m */
+  r0 = glmm_load(mat[0]); /* d c b a */
+  r1 = glmm_load(mat[1]); /* h g f e */
+  r2 = glmm_load(mat[2]); /* l k j i */
+  r3 = glmm_load(mat[3]); /* p o n m */
 
   /*
    t[1] = j * p - n * l;
@@ -119,20 +115,20 @@ glm_mat4_det_sse2(mat4 mat) {
    t[3] = i * p - m * l;
    t[4] = i * o - m * k;
    */
-  x0 = _mm_sub_ps(_mm_mul_ps(_mm_shuffle1_ps(r2, 0, 0, 1, 1),
-                             _mm_shuffle1_ps(r3, 2, 3, 2, 3)),
-                  _mm_mul_ps(_mm_shuffle1_ps(r3, 0, 0, 1, 1),
-                             _mm_shuffle1_ps(r2, 2, 3, 2, 3)));
+  x0 = _mm_sub_ps(_mm_mul_ps(glmm_shuff1(r2, 0, 0, 1, 1),
+                             glmm_shuff1(r3, 2, 3, 2, 3)),
+                  _mm_mul_ps(glmm_shuff1(r3, 0, 0, 1, 1),
+                             glmm_shuff1(r2, 2, 3, 2, 3)));
   /*
    t[0] = k * p - o * l;
    t[0] = k * p - o * l;
    t[5] = i * n - m * j;
    t[5] = i * n - m * j;
    */
-  x1 = _mm_sub_ps(_mm_mul_ps(_mm_shuffle1_ps(r2, 0, 0, 2, 2),
-                             _mm_shuffle1_ps(r3, 1, 1, 3, 3)),
-                  _mm_mul_ps(_mm_shuffle1_ps(r3, 0, 0, 2, 2),
-                             _mm_shuffle1_ps(r2, 1, 1, 3, 3)));
+  x1 = _mm_sub_ps(_mm_mul_ps(glmm_shuff1(r2, 0, 0, 2, 2),
+                             glmm_shuff1(r3, 1, 1, 3, 3)),
+                  _mm_mul_ps(glmm_shuff1(r3, 0, 0, 2, 2),
+                             glmm_shuff1(r2, 1, 1, 3, 3)));
 
   /*
      a * (f * t[0] - g * t[1] + h * t[2])
@@ -140,19 +136,19 @@ glm_mat4_det_sse2(mat4 mat) {
    + c * (e * t[1] - f * t[3] + h * t[5])
    - d * (e * t[2] - f * t[4] + g * t[5])
    */
-  x2 = _mm_sub_ps(_mm_mul_ps(_mm_shuffle1_ps(r1, 0, 0, 0, 1),
+  x2 = _mm_sub_ps(_mm_mul_ps(glmm_shuff1(r1, 0, 0, 0, 1),
                              _mm_shuffle_ps(x1, x0, _MM_SHUFFLE(1, 0, 0, 0))),
-                  _mm_mul_ps(_mm_shuffle1_ps(r1, 1, 1, 2, 2),
-                             _mm_shuffle1_ps(x0, 3, 2, 2, 0)));
+                  _mm_mul_ps(glmm_shuff1(r1, 1, 1, 2, 2),
+                             glmm_shuff1(x0, 3, 2, 2, 0)));
 
   x2 = _mm_add_ps(x2,
-                  _mm_mul_ps(_mm_shuffle1_ps(r1, 2, 3, 3, 3),
+                  _mm_mul_ps(glmm_shuff1(r1, 2, 3, 3, 3),
                              _mm_shuffle_ps(x0, x1, _MM_SHUFFLE(2, 2, 3, 1))));
   x2 = _mm_xor_ps(x2, _mm_set_ps(-0.f, 0.f, -0.f, 0.f));
 
   x0 = _mm_mul_ps(r0, x2);
-  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 0, 1, 2, 3));
-  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 1, 3, 3, 1));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 0, 1, 2, 3));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 1, 3, 3, 1));
 
   return _mm_cvtss_f32(x0);
 }
@@ -166,14 +162,14 @@ glm_mat4_inv_fast_sse2(mat4 mat, mat4 dest) {
          x0, x1, x2, x3, x4, x5, x6, x7;
 
   /* 127 <- 0 */
-  r0 = _mm_load_ps(mat[0]); /* d c b a */
-  r1 = _mm_load_ps(mat[1]); /* h g f e */
-  r2 = _mm_load_ps(mat[2]); /* l k j i */
-  r3 = _mm_load_ps(mat[3]); /* p o n m */
+  r0 = glmm_load(mat[0]); /* d c b a */
+  r1 = glmm_load(mat[1]); /* h g f e */
+  r2 = glmm_load(mat[2]); /* l k j i */
+  r3 = glmm_load(mat[3]); /* p o n m */
 
   x0 = _mm_shuffle_ps(r2, r3, _MM_SHUFFLE(3, 2, 3, 2));  /* p o l k */
-  x1 = _mm_shuffle1_ps(x0, 1, 3, 3, 3);                  /* l p p p */
-  x2 = _mm_shuffle1_ps(x0, 0, 2, 2, 2);                  /* k o o o */
+  x1 = glmm_shuff1(x0, 1, 3, 3, 3);                      /* l p p p */
+  x2 = glmm_shuff1(x0, 0, 2, 2, 2);                      /* k o o o */
   x0 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(3, 3, 3, 3));  /* h h l l */
   x3 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(2, 2, 2, 2));  /* g g k k */
 
@@ -184,7 +180,7 @@ glm_mat4_inv_fast_sse2(mat4 mat, mat4 dest) {
   t0 = _mm_sub_ps(_mm_mul_ps(x3, x1), _mm_mul_ps(x2, x0));
 
   x4 = _mm_shuffle_ps(r2, r3, _MM_SHUFFLE(2, 1, 2, 1)); /* o n k j */
-  x4 = _mm_shuffle1_ps(x4, 0, 2, 2, 2);                 /* j n n n */
+  x4 = glmm_shuff1(x4, 0, 2, 2, 2);                     /* j n n n */
   x5 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(1, 1, 1, 1)); /* f f j j */
 
   /* t1[1] = j * p - n * l;
@@ -200,7 +196,7 @@ glm_mat4_inv_fast_sse2(mat4 mat, mat4 dest) {
   t2 = _mm_sub_ps(_mm_mul_ps(x5, x2), _mm_mul_ps(x4, x3));
 
   x6 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(0, 0, 0, 0)); /* e e i i */
-  x7 = _mm_shuffle2_ps(r3, r2, 0, 0, 0, 0, 2, 0, 0, 0); /* i m m m */
+  x7 = glmm_shuff2(r3, r2, 0, 0, 0, 0, 2, 0, 0, 0);     /* i m m m */
 
   /* t1[3] = i * p - m * l;
      t1[3] = i * p - m * l;
@@ -220,10 +216,10 @@ glm_mat4_inv_fast_sse2(mat4 mat, mat4 dest) {
      t3[5] = e * j - i * f; */
   t5 = _mm_sub_ps(_mm_mul_ps(x6, x4), _mm_mul_ps(x7, x5));
 
-  x0 = _mm_shuffle2_ps(r1, r0, 0, 0, 0, 0, 2, 2, 2, 0); /* a a a e */
-  x1 = _mm_shuffle2_ps(r1, r0, 1, 1, 1, 1, 2, 2, 2, 0); /* b b b f */
-  x2 = _mm_shuffle2_ps(r1, r0, 2, 2, 2, 2, 2, 2, 2, 0); /* c c c g */
-  x3 = _mm_shuffle2_ps(r1, r0, 3, 3, 3, 3, 2, 2, 2, 0); /* d d d h */
+  x0 = glmm_shuff2(r1, r0, 0, 0, 0, 0, 2, 2, 2, 0); /* a a a e */
+  x1 = glmm_shuff2(r1, r0, 1, 1, 1, 1, 2, 2, 2, 0); /* b b b f */
+  x2 = glmm_shuff2(r1, r0, 2, 2, 2, 2, 2, 2, 2, 0); /* c c c g */
+  x3 = glmm_shuff2(r1, r0, 3, 3, 3, 3, 2, 2, 2, 0); /* d d d h */
 
   /*
    dest[0][0] =  f * t1[0] - g * t1[1] + h * t1[2];
@@ -271,14 +267,14 @@ glm_mat4_inv_fast_sse2(mat4 mat, mat4 dest) {
   x0 = _mm_shuffle_ps(x0, x1, _MM_SHUFFLE(2, 0, 2, 0));
 
   x0 = _mm_mul_ps(x0, r0);
-  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 0, 1, 2, 3));
-  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 1, 0, 0, 1));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 0, 1, 2, 3));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 1, 0, 0, 1));
   x0 = _mm_rcp_ps(x0);
 
-  _mm_store_ps(dest[0], _mm_mul_ps(v0, x0));
-  _mm_store_ps(dest[1], _mm_mul_ps(v1, x0));
-  _mm_store_ps(dest[2], _mm_mul_ps(v2, x0));
-  _mm_store_ps(dest[3], _mm_mul_ps(v3, x0));
+  glmm_store(dest[0], _mm_mul_ps(v0, x0));
+  glmm_store(dest[1], _mm_mul_ps(v1, x0));
+  glmm_store(dest[2], _mm_mul_ps(v2, x0));
+  glmm_store(dest[3], _mm_mul_ps(v3, x0));
 }
 
 CGLM_INLINE
@@ -290,14 +286,14 @@ glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
          x0, x1, x2, x3, x4, x5, x6, x7;
 
   /* 127 <- 0 */
-  r0 = _mm_load_ps(mat[0]); /* d c b a */
-  r1 = _mm_load_ps(mat[1]); /* h g f e */
-  r2 = _mm_load_ps(mat[2]); /* l k j i */
-  r3 = _mm_load_ps(mat[3]); /* p o n m */
+  r0 = glmm_load(mat[0]); /* d c b a */
+  r1 = glmm_load(mat[1]); /* h g f e */
+  r2 = glmm_load(mat[2]); /* l k j i */
+  r3 = glmm_load(mat[3]); /* p o n m */
 
   x0 = _mm_shuffle_ps(r2, r3, _MM_SHUFFLE(3, 2, 3, 2));  /* p o l k */
-  x1 = _mm_shuffle1_ps(x0, 1, 3, 3, 3);                  /* l p p p */
-  x2 = _mm_shuffle1_ps(x0, 0, 2, 2, 2);                  /* k o o o */
+  x1 = glmm_shuff1(x0, 1, 3, 3, 3);                      /* l p p p */
+  x2 = glmm_shuff1(x0, 0, 2, 2, 2);                      /* k o o o */
   x0 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(3, 3, 3, 3));  /* h h l l */
   x3 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(2, 2, 2, 2));  /* g g k k */
 
@@ -308,7 +304,7 @@ glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
   t0 = _mm_sub_ps(_mm_mul_ps(x3, x1), _mm_mul_ps(x2, x0));
 
   x4 = _mm_shuffle_ps(r2, r3, _MM_SHUFFLE(2, 1, 2, 1)); /* o n k j */
-  x4 = _mm_shuffle1_ps(x4, 0, 2, 2, 2);                 /* j n n n */
+  x4 = glmm_shuff1(x4, 0, 2, 2, 2);                     /* j n n n */
   x5 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(1, 1, 1, 1)); /* f f j j */
 
   /* t1[1] = j * p - n * l;
@@ -324,7 +320,7 @@ glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
   t2 = _mm_sub_ps(_mm_mul_ps(x5, x2), _mm_mul_ps(x4, x3));
 
   x6 = _mm_shuffle_ps(r2, r1, _MM_SHUFFLE(0, 0, 0, 0)); /* e e i i */
-  x7 = _mm_shuffle2_ps(r3, r2, 0, 0, 0, 0, 2, 0, 0, 0); /* i m m m */
+  x7 = glmm_shuff2(r3, r2, 0, 0, 0, 0, 2, 0, 0, 0);     /* i m m m */
 
   /* t1[3] = i * p - m * l;
      t1[3] = i * p - m * l;
@@ -344,10 +340,10 @@ glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
      t3[5] = e * j - i * f; */
   t5 = _mm_sub_ps(_mm_mul_ps(x6, x4), _mm_mul_ps(x7, x5));
 
-  x0 = _mm_shuffle2_ps(r1, r0, 0, 0, 0, 0, 2, 2, 2, 0); /* a a a e */
-  x1 = _mm_shuffle2_ps(r1, r0, 1, 1, 1, 1, 2, 2, 2, 0); /* b b b f */
-  x2 = _mm_shuffle2_ps(r1, r0, 2, 2, 2, 2, 2, 2, 2, 0); /* c c c g */
-  x3 = _mm_shuffle2_ps(r1, r0, 3, 3, 3, 3, 2, 2, 2, 0); /* d d d h */
+  x0 = glmm_shuff2(r1, r0, 0, 0, 0, 0, 2, 2, 2, 0); /* a a a e */
+  x1 = glmm_shuff2(r1, r0, 1, 1, 1, 1, 2, 2, 2, 0); /* b b b f */
+  x2 = glmm_shuff2(r1, r0, 2, 2, 2, 2, 2, 2, 2, 0); /* c c c g */
+  x3 = glmm_shuff2(r1, r0, 3, 3, 3, 3, 2, 2, 2, 0); /* d d d h */
 
   /*
    dest[0][0] =  f * t1[0] - g * t1[1] + h * t1[2];
@@ -395,14 +391,14 @@ glm_mat4_inv_sse2(mat4 mat, mat4 dest) {
   x0 = _mm_shuffle_ps(x0, x1, _MM_SHUFFLE(2, 0, 2, 0));
 
   x0 = _mm_mul_ps(x0, r0);
-  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 0, 1, 2, 3));
-  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 1, 0, 0, 1));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 0, 1, 2, 3));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 1, 0, 0, 1));
   x0 = _mm_div_ps(_mm_set1_ps(1.0f), x0);
 
-  _mm_store_ps(dest[0], _mm_mul_ps(v0, x0));
-  _mm_store_ps(dest[1], _mm_mul_ps(v1, x0));
-  _mm_store_ps(dest[2], _mm_mul_ps(v2, x0));
-  _mm_store_ps(dest[3], _mm_mul_ps(v3, x0));
+  glmm_store(dest[0], _mm_mul_ps(v0, x0));
+  glmm_store(dest[1], _mm_mul_ps(v1, x0));
+  glmm_store(dest[2], _mm_mul_ps(v2, x0));
+  glmm_store(dest[3], _mm_mul_ps(v3, x0));
 }
 
 #endif

include/cglm/simd/sse2/quat.h

@@ -14,56 +14,33 @@
 
 CGLM_INLINE
 void
-glm_quat_slerp_sse2(versor q,
-                    versor r,
-                    float  t,
-                    versor dest) {
-  /* https://en.wikipedia.org/wiki/Slerp */
-  float cosTheta, sinTheta, angle, a, b, c;
+glm_quat_mul_sse2(versor p, versor q, versor dest) {
+  /*
+   + (a1 b2 + b1 a2 + c1 d2 − d1 c2)i
+   + (a1 c2 − b1 d2 + c1 a2 + d1 b2)j
+   + (a1 d2 + b1 c2 − c1 b2 + d1 a2)k
+     a1 a2 − b1 b2 − c1 c2 − d1 d2
+   */
 
-  __m128 xmm_q;
+  __m128 xp, xq, x0, r;
 
-  xmm_q = _mm_load_ps(q);
+  xp = glmm_load(p); /* 3 2 1 0 */
+  xq = glmm_load(q);
 
-  cosTheta = glm_vec4_dot(q, r);
-  if (cosTheta < 0.0f) {
-    _mm_store_ps(q,
-                 _mm_xor_ps(xmm_q,
-                            _mm_set1_ps(-0.f))) ;
+  r  = _mm_mul_ps(glmm_shuff1x(xp, 3), xq);
 
-    cosTheta = -cosTheta;
-  }
+  x0 = _mm_xor_ps(glmm_shuff1x(xp, 0), _mm_set_ps(-0.f, 0.f, -0.f, 0.f));
+  r  = _mm_add_ps(r, _mm_mul_ps(x0, glmm_shuff1(xq, 0, 1, 2, 3)));
 
-  if (cosTheta >= 1.0f) {
-    _mm_store_ps(dest, xmm_q);
-    return;
-  }
+  x0 = _mm_xor_ps(glmm_shuff1x(xp, 1), _mm_set_ps(-0.f, -0.f, 0.f, 0.f));
+  r  = _mm_add_ps(r, _mm_mul_ps(x0, glmm_shuff1(xq, 1, 0, 3, 2)));
 
-  sinTheta = sqrtf(1.0f - cosTheta * cosTheta);
+  x0 = _mm_xor_ps(glmm_shuff1x(xp, 2), _mm_set_ps(-0.f, 0.f, 0.f, -0.f));
+  r  = _mm_add_ps(r, _mm_mul_ps(x0, glmm_shuff1(xq, 2, 3, 0, 1)));
 
-  c = 1.0f - t;
-
-  /* LERP */
-  if (sinTheta < 0.001f) {
-    _mm_store_ps(dest, _mm_add_ps(_mm_mul_ps(_mm_set1_ps(c),
-                                             xmm_q),
-                                  _mm_mul_ps(_mm_set1_ps(t),
-                                             _mm_load_ps(r))));
-    return;
-  }
-
-  /* SLERP */
-  angle = acosf(cosTheta);
-  a = sinf(c * angle);
-  b = sinf(t * angle);
-
-  _mm_store_ps(dest,
-               _mm_div_ps(_mm_add_ps(_mm_mul_ps(_mm_set1_ps(a),
-                                                xmm_q),
-                                     _mm_mul_ps(_mm_set1_ps(b),
-                                                _mm_load_ps(r))),
-                          _mm_set1_ps(sinTheta)));
+  glmm_store(dest, r);
 }
 
+
 #endif
 #endif /* cglm_quat_simd_h */

include/cglm/sphere.h

@@ -0,0 +1,99 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#ifndef cglm_sphere_h
+#define cglm_sphere_h
+
+#include "common.h"
+#include "mat4.h"
+
+/*
+  Sphere Representation in cglm: [center.x, center.y, center.z, radii]
+
+  You could use this representation or you can convert it to vec4 before call
+  any function
+ */
+
+/*!
+ * @brief helper for getting sphere radius
+ *
+ * @param[in]   s  sphere
+ *
+ * @return returns radii
+ */
+CGLM_INLINE
+float
+glm_sphere_radii(vec4 s) {
+  return s[3];
+}
+
+/*!
+ * @brief apply transform to sphere, it is just wrapper for glm_mat4_mulv3
+ *
+ * @param[in]  s    sphere
+ * @param[in]  m    transform matrix
+ * @param[out] dest transformed sphere
+ */
+CGLM_INLINE
+void
+glm_sphere_transform(vec4 s, mat4 m, vec4 dest) {
+  glm_mat4_mulv3(m, s, 1.0f, dest);
+  dest[3] = s[3];
+}
+
+/*!
+ * @brief merges two spheres and creates a new one
+ *
+ * two sphere must be in same space, for instance if one in world space then
+ * the other must be in world space too, not in local space.
+ *
+ * @param[in]  s1   sphere 1
+ * @param[in]  s2   sphere 2
+ * @param[out] dest merged/extended sphere
+ */
+CGLM_INLINE
+void
+glm_sphere_merge(vec4 s1, vec4 s2, vec4 dest) {
+  float dist, radii;
+
+  dist  = glm_vec_distance(s1, s2);
+  radii = dist + s1[3] + s2[3];
+
+  radii = glm_max(radii, s1[3]);
+  radii = glm_max(radii, s2[3]);
+
+  glm_vec_center(s1, s2, dest);
+  dest[3] = radii;
+}
+
+/*!
+ * @brief check if two sphere intersects
+ *
+ * @param[in]   s1  sphere
+ * @param[in]   s2  other sphere
+ */
+CGLM_INLINE
+bool
+glm_sphere_sphere(vec4 s1, vec4 s2) {
+  return glm_vec_distance2(s1, s2) <= glm_pow2(s1[3] + s2[3]);
+}
+
+/*!
+ * @brief check if sphere intersects with point
+ *
+ * @param[in]   s      sphere
+ * @param[in]   point  point
+ */
+CGLM_INLINE
+bool
+glm_sphere_point(vec4 s, vec3 point) {
+  float rr;
+  rr = s[3] * s[3];
+  return glm_vec_distance2(point, s) <= rr;
+}
+
+#endif /* cglm_sphere_h */

include/cglm/types.h

@@ -9,22 +9,35 @@
 #define cglm_types_h
 
 #if defined(_MSC_VER)
-#  define CGLM_ALIGN(X) /* __declspec(align(X)) */
+/* do not use alignment for older visual studio versions */
+#if _MSC_VER < 1913 /*  Visual Studio 2017 version 15.6  */
+#  define CGLM_ALL_UNALIGNED
+#  define CGLM_ALIGN(X) /* no alignment */
+#else
+#  define CGLM_ALIGN(X) __declspec(align(X))
+#endif
 #else
 #  define CGLM_ALIGN(X) __attribute((aligned(X)))
 #endif
 
-typedef float vec3[3];
+#ifndef CGLM_ALL_UNALIGNED
+#  define CGLM_ALIGN_IF(X) CGLM_ALIGN(X)
+#else
+#  define CGLM_ALIGN_IF(X) /* no alignment */
+#endif
+
+typedef float                   vec2[2];
+typedef CGLM_ALIGN_IF(8)  float vec3[3];
 typedef int                    ivec3[3];
-typedef CGLM_ALIGN(16) float vec4[4];
+typedef CGLM_ALIGN_IF(16) float vec4[4];
 
 typedef vec3                    mat3[3];
-typedef vec4 mat4[4];
+typedef CGLM_ALIGN_IF(16) vec4  mat4[4];
 
 typedef vec4                    versor;
 
-#define CGLM_PI    (float)M_PI
-#define CGLM_PI_2  (float)M_PI_2
-#define CGLM_PI_4  (float)M_PI_4
+#define CGLM_PI    ((float)M_PI)
+#define CGLM_PI_2  ((float)M_PI_2)
+#define CGLM_PI_4  ((float)M_PI_4)
 
 #endif /* cglm_types_h */

include/cglm/util.h

@@ -21,7 +21,9 @@
 #include "common.h"
 
 /*!
- * @brief get sign of 32 bit integer as +1 or -1
+ * @brief get sign of 32 bit integer as +1, -1, 0
+ *
+ * Important: It returns 0 for zero input
  *
  * @param val integer value
  */
@@ -31,6 +33,19 @@ glm_sign(int val) {
   return ((val >> 31) - (-val >> 31));
 }
 
+/*!
+ * @brief get sign of 32 bit float as +1, -1, 0
+ *
+ * Important: It returns 0 for zero/NaN input
+ *
+ * @param val float value
+ */
+CGLM_INLINE
+float
+glm_signf(float val) {
+  return (float)((val > 0.0f) - (val < 0.0f));
+}
+
 /*!
  * @brief convert degree to radians
  *
@@ -83,7 +98,6 @@ glm_make_deg(float *rad) {
 CGLM_INLINE
 float
 glm_pow2(float x) {
-
   return x * x;
 }
 
@@ -115,4 +129,32 @@ glm_max(float a, float b) {
   return b;
 }
 
+/*!
+ * @brief clamp a number between min and max
+ *
+ * @param[in] val    value to clamp
+ * @param[in] minVal minimum value
+ * @param[in] maxVal maximum value
+ */
+CGLM_INLINE
+float
+glm_clamp(float val, float minVal, float maxVal) {
+  return glm_min(glm_max(val, minVal), maxVal);
+}
+
+/*!
+ * @brief linear interpolation between two number
+ *
+ * formula:  from + s * (to - from)
+ *
+ * @param[in]   from from value
+ * @param[in]   to   to value
+ * @param[in]   t    interpolant (amount) clamped between 0 and 1
+ */
+CGLM_INLINE
+float
+glm_lerp(float from, float to, float t) {
+  return from + glm_clamp(t, 0.0f, 1.0f) * (to - from);
+}
+
 #endif /* cglm_util_h */

include/cglm/vec3-ext.h

@@ -26,12 +26,13 @@
 #define cglm_vec3_ext_h
 
 #include "common.h"
+#include "util.h"
 #include <stdbool.h>
 #include <math.h>
 #include <float.h>
 
 /*!
- * @brief multiplies individual items, just for convenient like SIMD
+ * @brief DEPRECATED! use glm_vec_mul
  *
  * @param[in]  a vec1
  * @param[in]  b vec2
@@ -160,4 +161,69 @@ glm_vec_min(vec3 v) {
   return min;
 }
 
+/*!
+ * @brief check if all items are NaN (not a number)
+ *        you should only use this in DEBUG mode or very critical asserts
+ *
+ * @param[in] v vector
+ */
+CGLM_INLINE
+bool
+glm_vec_isnan(vec3 v) {
+  return isnan(v[0]) || isnan(v[1]) || isnan(v[2]);
+}
+
+/*!
+ * @brief check if all items are INFINITY
+ *        you should only use this in DEBUG mode or very critical asserts
+ *
+ * @param[in] v vector
+ */
+CGLM_INLINE
+bool
+glm_vec_isinf(vec3 v) {
+  return isinf(v[0]) || isinf(v[1]) || isinf(v[2]);
+}
+
+/*!
+ * @brief check if all items are valid number
+ *        you should only use this in DEBUG mode or very critical asserts
+ *
+ * @param[in] v vector
+ */
+CGLM_INLINE
+bool
+glm_vec_isvalid(vec3 v) {
+  return !glm_vec_isnan(v) && !glm_vec_isinf(v);
+}
+
+/*!
+ * @brief get sign of 32 bit float as +1, -1, 0
+ *
+ * Important: It returns 0 for zero/NaN input
+ *
+ * @param v vector
+ */
+CGLM_INLINE
+void
+glm_vec_sign(vec3 v, vec3 dest) {
+  dest[0] = glm_signf(v[0]);
+  dest[1] = glm_signf(v[1]);
+  dest[2] = glm_signf(v[2]);
+}
+
+/*!
+ * @brief square root of each vector item
+ *
+ * @param[in]  v    vector
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec_sqrt(vec3 v, vec3 dest) {
+  dest[0] = sqrtf(v[0]);
+  dest[1] = sqrtf(v[1]);
+  dest[2] = sqrtf(v[2]);
+}
+
 #endif /* cglm_vec3_ext_h */

include/cglm/vec3.h

@@ -28,10 +28,18 @@
    CGLM_INLINE void  glm_vec_cross(vec3 a, vec3 b, vec3 d);
    CGLM_INLINE float glm_vec_norm2(vec3 v);
    CGLM_INLINE float glm_vec_norm(vec3 vec);
-   CGLM_INLINE void  glm_vec_add(vec3 v1, vec3 v2, vec3 dest);
-   CGLM_INLINE void  glm_vec_sub(vec3 v1, vec3 v2, vec3 dest);
+   CGLM_INLINE void  glm_vec_add(vec3 a, vec3 b, vec3 dest);
+   CGLM_INLINE void  glm_vec_adds(vec3 a, float s, vec3 dest);
+   CGLM_INLINE void  glm_vec_sub(vec3 a, vec3 b, vec3 dest);
+   CGLM_INLINE void  glm_vec_subs(vec3 a, float s, vec3 dest);
+   CGLM_INLINE void  glm_vec_mul(vec3 a, vec3 b, vec3 dest);
    CGLM_INLINE void  glm_vec_scale(vec3 v, float s, vec3 dest);
    CGLM_INLINE void  glm_vec_scale_as(vec3 v, float s, vec3 dest);
+   CGLM_INLINE void  glm_vec_div(vec3 a, vec3 b, vec3 dest);
+   CGLM_INLINE void  glm_vec_divs(vec3 a, float s, vec3 dest);
+   CGLM_INLINE void  glm_vec_addadd(vec3 a, vec3 b, vec3 dest);
+   CGLM_INLINE void  glm_vec_subadd(vec3 a, vec3 b, vec3 dest);
+   CGLM_INLINE void  glm_vec_muladd(vec3 a, vec3 b, vec3 dest);
    CGLM_INLINE void  glm_vec_flipsign(vec3 v);
    CGLM_INLINE void  glm_vec_inv(vec3 v);
    CGLM_INLINE void  glm_vec_inv_to(vec3 v, vec3 dest);
@@ -46,6 +54,7 @@
    CGLM_INLINE void  glm_vec_maxv(vec3 v1, vec3 v2, vec3 dest);
    CGLM_INLINE void  glm_vec_minv(vec3 v1, vec3 v2, vec3 dest);
    CGLM_INLINE void  glm_vec_ortho(vec3 v, vec3 dest);
+   CGLM_INLINE void  glm_vec_clamp(vec3 v, float minVal, float maxVal);
 
  Convenient:
    CGLM_INLINE void  glm_cross(vec3 a, vec3 b, vec3 d);
@@ -58,6 +67,7 @@
 #define cglm_vec3_h
 
 #include "common.h"
+#include "vec4.h"
 #include "vec3-ext.h"
 #include "util.h"
 
@@ -102,6 +112,32 @@ glm_vec_copy(vec3 a, vec3 dest) {
   dest[2] = a[2];
 }
 
+/*!
+ * @brief make vector zero
+ *
+ * @param[in, out]  v vector
+ */
+CGLM_INLINE
+void
+glm_vec_zero(vec3 v) {
+  v[0] = 0.0f;
+  v[1] = 0.0f;
+  v[2] = 0.0f;
+}
+
+/*!
+ * @brief make vector one
+ *
+ * @param[in, out]  v vector
+ */
+CGLM_INLINE
+void
+glm_vec_one(vec3 v) {
+  v[0] = 1.0f;
+  v[1] = 1.0f;
+  v[2] = 1.0f;
+}
+
 /*!
  * @brief vec3 dot product
  *
@@ -146,7 +182,7 @@ glm_vec_cross(vec3 a, vec3 b, vec3 d) {
 CGLM_INLINE
 float
 glm_vec_norm2(vec3 v) {
-  return v[0] * v[0] + v[1] * v[1] + v[2] * v[2];
+  return glm_vec_dot(v, v);
 }
 
 /*!
@@ -163,33 +199,78 @@ glm_vec_norm(vec3 vec) {
 }
 
 /*!
- * @brief add v2 vector to v1 vector store result in dest
+ * @brief add a vector to b vector store result in dest
  *
- * @param[in]  v1 vector1
- * @param[in]  v2 vector2
+ * @param[in]  a    vector1
+ * @param[in]  b    vector2
  * @param[out] dest destination vector
  */
 CGLM_INLINE
 void
-glm_vec_add(vec3 v1, vec3 v2, vec3 dest) {
-  dest[0] = v1[0] + v2[0];
-  dest[1] = v1[1] + v2[1];
-  dest[2] = v1[2] + v2[2];
+glm_vec_add(vec3 a, vec3 b, vec3 dest) {
+  dest[0] = a[0] + b[0];
+  dest[1] = a[1] + b[1];
+  dest[2] = a[2] + b[2];
+}
+
+/*!
+ * @brief add scalar to v vector store result in dest (d = v + s)
+ *
+ * @param[in]  v    vector
+ * @param[in]  s    scalar
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec_adds(vec3 v, float s, vec3 dest) {
+  dest[0] = v[0] + s;
+  dest[1] = v[1] + s;
+  dest[2] = v[2] + s;
 }
 
 /*!
  * @brief subtract v2 vector from v1 vector store result in dest
  *
- * @param[in]  v1 vector1
- * @param[in]  v2 vector2
+ * @param[in]  a    vector1
+ * @param[in]  b    vector2
  * @param[out] dest destination vector
  */
 CGLM_INLINE
 void
-glm_vec_sub(vec3 v1, vec3 v2, vec3 dest) {
-  dest[0] = v1[0] - v2[0];
-  dest[1] = v1[1] - v2[1];
-  dest[2] = v1[2] - v2[2];
+glm_vec_sub(vec3 a, vec3 b, vec3 dest) {
+  dest[0] = a[0] - b[0];
+  dest[1] = a[1] - b[1];
+  dest[2] = a[2] - b[2];
+}
+
+/*!
+ * @brief subtract scalar from v vector store result in dest (d = v - s)
+ *
+ * @param[in]  v    vector
+ * @param[in]  s    scalar
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec_subs(vec3 v, float s, vec3 dest) {
+  dest[0] = v[0] - s;
+  dest[1] = v[1] - s;
+  dest[2] = v[2] - s;
+}
+
+/*!
+ * @brief multiply two vector (component-wise multiplication)
+ *
+ * @param a v1
+ * @param b v2
+ * @param d v3 = (a[0] * b[0], a[1] * b[1], a[2] * b[2])
+ */
+CGLM_INLINE
+void
+glm_vec_mul(vec3 a, vec3 b, vec3 d) {
+  d[0] = a[0] * b[0];
+  d[1] = a[1] * b[1];
+  d[2] = a[2] * b[2];
 }
 
 /*!
@@ -220,14 +301,112 @@ glm_vec_scale_as(vec3 v, float s, vec3 dest) {
   float norm;
   norm = glm_vec_norm(v);
 
-  if (norm == 0) {
-    glm_vec_copy(v, dest);
+  if (norm == 0.0f) {
+    glm_vec_zero(dest);
     return;
   }
 
   glm_vec_scale(v, s / norm, dest);
 }
 
+/*!
+ * @brief div vector with another component-wise division: d = a / b
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest result = (a[0]/b[0], a[1]/b[1], a[2]/b[2])
+ */
+CGLM_INLINE
+void
+glm_vec_div(vec3 a, vec3 b, vec3 dest) {
+  dest[0] = a[0] / b[0];
+  dest[1] = a[1] / b[1];
+  dest[2] = a[2] / b[2];
+}
+
+/*!
+ * @brief div vector with scalar: d = v / s
+ *
+ * @param[in]  v    vector
+ * @param[in]  s    scalar
+ * @param[out] dest result = (a[0]/s, a[1]/s, a[2]/s)
+ */
+CGLM_INLINE
+void
+glm_vec_divs(vec3 v, float s, vec3 dest) {
+  dest[0] = v[0] / s;
+  dest[1] = v[1] / s;
+  dest[2] = v[2] / s;
+}
+
+/*!
+ * @brief add two vectors and add result to sum
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest dest += (a + b)
+ */
+CGLM_INLINE
+void
+glm_vec_addadd(vec3 a, vec3 b, vec3 dest) {
+  dest[0] += a[0] + b[0];
+  dest[1] += a[1] + b[1];
+  dest[2] += a[2] + b[2];
+}
+
+/*!
+ * @brief sub two vectors and add result to dest
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest dest += (a + b)
+ */
+CGLM_INLINE
+void
+glm_vec_subadd(vec3 a, vec3 b, vec3 dest) {
+  dest[0] += a[0] - b[0];
+  dest[1] += a[1] - b[1];
+  dest[2] += a[2] - b[2];
+}
+
+/*!
+ * @brief mul two vectors and add result to dest
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest dest += (a * b)
+ */
+CGLM_INLINE
+void
+glm_vec_muladd(vec3 a, vec3 b, vec3 dest) {
+  dest[0] += a[0] * b[0];
+  dest[1] += a[1] * b[1];
+  dest[2] += a[2] * b[2];
+}
+
+/*!
+ * @brief mul vector with scalar and add result to sum
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector
+ * @param[in]  s    scalar
+ * @param[out] dest dest += (a * b)
+ */
+CGLM_INLINE
+void
+glm_vec_muladds(vec3 a, float s, vec3 dest) {
+  dest[0] += a[0] * s;
+  dest[1] += a[1] * s;
+  dest[2] += a[2] * s;
+}
+
 /*!
  * @brief flip sign of all vec3 members
  *
@@ -241,6 +420,20 @@ glm_vec_flipsign(vec3 v) {
   v[2] = -v[2];
 }
 
+/*!
+ * @brief flip sign of all vec3 members and store result in dest
+ *
+ * @param[in]   v     vector
+ * @param[out]  dest  result vector
+ */
+CGLM_INLINE
+void
+glm_vec_flipsign_to(vec3 v, vec3 dest) {
+  dest[0] = -v[0];
+  dest[1] = -v[1];
+  dest[2] = -v[2];
+}
+
 /*!
  * @brief make vector as inverse/opposite of itself
  *
@@ -261,8 +454,7 @@ glm_vec_inv(vec3 v) {
 CGLM_INLINE
 void
 glm_vec_inv_to(vec3 v, vec3 dest) {
-  glm_vec_copy(v, dest);
-  glm_vec_flipsign(dest);
+  glm_vec_flipsign_to(v, dest);
 }
 
 /*!
@@ -299,7 +491,7 @@ glm_vec_normalize_to(vec3 vec, vec3 dest) {
   norm = glm_vec_norm(vec);
 
   if (norm == 0.0f) {
-    dest[0] = dest[1] = dest[2] = 0.0f;
+    glm_vec_zero(dest);
     return;
   }
 
@@ -324,12 +516,6 @@ glm_vec_angle(vec3 v1, vec3 v2) {
   return acosf(glm_vec_dot(v1, v2) * norm);
 }
 
-CGLM_INLINE
-void
-glm_quatv(versor q,
-          float  angle,
-          vec3   v);
-
 /*!
  * @brief rotate vec3 around axis by angle using Rodrigues' rotation formula
  *
@@ -340,31 +526,55 @@ glm_quatv(versor q,
 CGLM_INLINE
 void
 glm_vec_rotate(vec3 v, float angle, vec3 axis) {
-  versor q;
-  vec3   v1, v2, v3;
+  vec3   v1, v2, k;
   float  c, s;
 
   c = cosf(angle);
   s = sinf(angle);
 
+  glm_vec_normalize_to(axis, k);
+
   /* Right Hand, Rodrigues' rotation formula:
         v = v*cos(t) + (kxv)sin(t) + k*(k.v)(1 - cos(t))
    */
-
-  /* quaternion */
-  glm_quatv(q, angle, v);
-
   glm_vec_scale(v, c, v1);
 
-  glm_vec_cross(axis, v, v2);
+  glm_vec_cross(k, v, v2);
   glm_vec_scale(v2, s, v2);
 
-  glm_vec_scale(axis,
-                glm_vec_dot(axis, v) * (1.0f - c),
-                v3);
-
   glm_vec_add(v1, v2, v1);
-  glm_vec_add(v1, v3, v);
+
+  glm_vec_scale(k, glm_vec_dot(k, v) * (1.0f - c), v2);
+  glm_vec_add(v1, v2, v);
+}
+
+/*!
+ * @brief apply rotation matrix to vector
+ *
+ *  matrix format should be (no perspective):
+ *   a  b  c  x
+ *   e  f  g  y
+ *   i  j  k  z
+ *   0  0  0  w
+ *
+ * @param[in]  m    affine matrix or rot matrix
+ * @param[in]  v    vector
+ * @param[out] dest rotated vector
+ */
+CGLM_INLINE
+void
+glm_vec_rotate_m4(mat4 m, vec3 v, vec3 dest) {
+  vec4 x, y, z, res;
+
+  glm_vec4_normalize_to(m[0], x);
+  glm_vec4_normalize_to(m[1], y);
+  glm_vec4_normalize_to(m[2], z);
+
+  glm_vec4_scale(x,   v[0], res);
+  glm_vec4_muladds(y, v[1], res);
+  glm_vec4_muladds(z, v[2], res);
+
+  glm_vec3(res, dest);
 }
 
 /*!
@@ -376,18 +586,22 @@ glm_vec_rotate(vec3 v, float angle, vec3 axis) {
  */
 CGLM_INLINE
 void
-glm_vec_rotate_m4(mat4 m, vec3 v, vec3 dest) {
-  vec3 res, x, y, z;
+glm_vec_rotate_m3(mat3 m, vec3 v, vec3 dest) {
+  vec4 res, x, y, z;
 
-  glm_vec_normalize_to(m[0], x);
-  glm_vec_normalize_to(m[1], y);
-  glm_vec_normalize_to(m[2], z);
+  glm_vec4(m[0], 0.0f, x);
+  glm_vec4(m[1], 0.0f, y);
+  glm_vec4(m[2], 0.0f, z);
 
-  res[0] = x[0] * v[0] + y[0] * v[1] + z[0] * v[2];
-  res[1] = x[1] * v[0] + y[1] * v[1] + z[1] * v[2];
-  res[2] = x[2] * v[0] + y[2] * v[1] + z[2] * v[2];
+  glm_vec4_normalize(x);
+  glm_vec4_normalize(y);
+  glm_vec4_normalize(z);
 
-  glm_vec_copy(res, dest);
+  glm_vec4_scale(x,   v[0], res);
+  glm_vec4_muladds(y, v[1], res);
+  glm_vec4_muladds(z, v[2], res);
+
+  glm_vec3(res, dest);
 }
 
 /*!
@@ -419,6 +633,21 @@ glm_vec_center(vec3 v1, vec3 v2, vec3 dest) {
   glm_vec_scale(dest, 0.5f, dest);
 }
 
+/**
+ * @brief squared distance between two vectors
+ *
+ * @param[in] v1 vector1
+ * @param[in] v2 vector2
+ * @return returns squared distance (distance * distance)
+ */
+CGLM_INLINE
+float
+glm_vec_distance2(vec3 v1, vec3 v2) {
+  return glm_pow2(v2[0] - v1[0])
+       + glm_pow2(v2[1] - v1[1])
+       + glm_pow2(v2[2] - v1[2]);
+}
+
 /**
  * @brief distance between two vectors
  *
@@ -429,9 +658,7 @@ glm_vec_center(vec3 v1, vec3 v2, vec3 dest) {
 CGLM_INLINE
 float
 glm_vec_distance(vec3 v1, vec3 v2) {
-  return sqrtf(glm_pow2(v2[0] - v1[0])
-             + glm_pow2(v2[1] - v1[1])
-             + glm_pow2(v2[2] - v1[2]));
+  return sqrtf(glm_vec_distance2(v1, v2));
 }
 
 /*!
@@ -478,6 +705,43 @@ glm_vec_ortho(vec3 v, vec3 dest) {
   dest[2] = v[0] - v[1];
 }
 
+/*!
+ * @brief clamp vector's individual members between min and max values
+ *
+ * @param[in, out]  v      vector
+ * @param[in]       minVal minimum value
+ * @param[in]       maxVal maximum value
+ */
+CGLM_INLINE
+void
+glm_vec_clamp(vec3 v, float minVal, float maxVal) {
+  v[0] = glm_clamp(v[0], minVal, maxVal);
+  v[1] = glm_clamp(v[1], minVal, maxVal);
+  v[2] = glm_clamp(v[2], minVal, maxVal);
+}
+
+/*!
+ * @brief linear interpolation between two vector
+ *
+ * formula:  from + s * (to - from)
+ *
+ * @param[in]   from from value
+ * @param[in]   to   to value
+ * @param[in]   t    interpolant (amount) clamped between 0 and 1
+ * @param[out]  dest destination
+ */
+CGLM_INLINE
+void
+glm_vec_lerp(vec3 from, vec3 to, float t, vec3 dest) {
+  vec3 s, v;
+
+  /* from + s * (to - from) */
+  glm_vec_broadcast(glm_clamp(t, 0.0f, 1.0f), s);
+  glm_vec_sub(to, from, v);
+  glm_vec_mulv(s, v, v);
+  glm_vec_add(from, v, dest);
+}
+
 /*!
  * @brief vec3 cross product
  *

include/cglm/vec4-ext.h

@@ -32,7 +32,7 @@
 #include <float.h>
 
 /*!
- * @brief multiplies individual items, just for convenient like SIMD
+ * @brief DEPRECATED! use glm_vec4_mul
  *
  * @param a v1
  * @param b v2
@@ -42,7 +42,7 @@ CGLM_INLINE
 void
 glm_vec4_mulv(vec4 a, vec4 b, vec4 d) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(d, _mm_mul_ps(_mm_load_ps(a), _mm_load_ps(b)));
+  glmm_store(d, _mm_mul_ps(glmm_load(a), glmm_load(b)));
 #else
   d[0] = a[0] * b[0];
   d[1] = a[1] * b[1];
@@ -61,7 +61,7 @@ CGLM_INLINE
 void
 glm_vec4_broadcast(float val, vec4 d) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(d, _mm_set1_ps(val));
+  glmm_store(d, _mm_set1_ps(val));
 #else
   d[0] = d[1] = d[2] = d[3] = val;
 #endif
@@ -174,5 +174,88 @@ glm_vec4_min(vec4 v) {
   return min;
 }
 
-#endif /* cglm_vec4_ext_h */
+/*!
+ * @brief check if one of items is NaN (not a number)
+ *        you should only use this in DEBUG mode or very critical asserts
+ *
+ * @param[in] v vector
+ */
+CGLM_INLINE
+bool
+glm_vec4_isnan(vec4 v) {
+  return isnan(v[0]) || isnan(v[1]) || isnan(v[2]) || isnan(v[3]);
+}
 
+/*!
+ * @brief check if one of items is INFINITY
+ *        you should only use this in DEBUG mode or very critical asserts
+ *
+ * @param[in] v vector
+ */
+CGLM_INLINE
+bool
+glm_vec4_isinf(vec4 v) {
+  return isinf(v[0]) || isinf(v[1]) || isinf(v[2]) || isinf(v[3]);
+}
+
+/*!
+ * @brief check if all items are valid number
+ *        you should only use this in DEBUG mode or very critical asserts
+ *
+ * @param[in] v vector
+ */
+CGLM_INLINE
+bool
+glm_vec4_isvalid(vec4 v) {
+  return !glm_vec4_isnan(v) && !glm_vec4_isinf(v);
+}
+
+/*!
+ * @brief get sign of 32 bit float as +1, -1, 0
+ *
+ * Important: It returns 0 for zero/NaN input
+ *
+ * @param v vector
+ */
+CGLM_INLINE
+void
+glm_vec4_sign(vec4 v, vec4 dest) {
+#if defined( __SSE2__ ) || defined( __SSE2__ )
+  __m128 x0, x1, x2, x3, x4;
+
+  x0 = glmm_load(v);
+  x1 = _mm_set_ps(0.0f, 0.0f, 1.0f, -1.0f);
+  x2 = glmm_shuff1x(x1, 2);
+
+  x3 = _mm_and_ps(_mm_cmpgt_ps(x0, x2), glmm_shuff1x(x1, 1));
+  x4 = _mm_and_ps(_mm_cmplt_ps(x0, x2), glmm_shuff1x(x1, 0));
+
+  glmm_store(dest, _mm_or_ps(x3, x4));
+#else
+  dest[0] = glm_signf(v[0]);
+  dest[1] = glm_signf(v[1]);
+  dest[2] = glm_signf(v[2]);
+  dest[3] = glm_signf(v[3]);
+#endif
+}
+
+/*!
+ * @brief square root of each vector item
+ *
+ * @param[in]  v    vector
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec4_sqrt(vec4 v, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_sqrt_ps(glmm_load(v)));
+#else
+  dest[0] = sqrtf(v[0]);
+  dest[1] = sqrtf(v[1]);
+  dest[2] = sqrtf(v[2]);
+  dest[3] = sqrtf(v[3]);
+#endif
+}
+
+#endif /* cglm_vec4_ext_h */

include/cglm/vec4.h

@@ -28,10 +28,18 @@
    CGLM_INLINE float glm_vec4_dot(vec4 a, vec4 b);
    CGLM_INLINE float glm_vec4_norm2(vec4 v);
    CGLM_INLINE float glm_vec4_norm(vec4 vec);
-   CGLM_INLINE void  glm_vec4_add(vec4 v1, vec4 v2, vec4 dest);
-   CGLM_INLINE void  glm_vec4_sub(vec4 v1, vec4 v2, vec4 dest);
+   CGLM_INLINE void  glm_vec4_add(vec4 a, vec4 b, vec4 dest);
+   CGLM_INLINE void  glm_vec4_adds(vec4 v, float s, vec4 dest);
+   CGLM_INLINE void  glm_vec4_sub(vec4 a, vec4 b, vec4 dest);
+   CGLM_INLINE void  glm_vec4_subs(vec4 v, float s, vec4 dest);
+   CGLM_INLINE void  glm_vec4_mul(vec4 a, vec4 b, vec4 dest);
    CGLM_INLINE void  glm_vec4_scale(vec4 v, float s, vec4 dest);
    CGLM_INLINE void  glm_vec4_scale_as(vec4 v, float s, vec4 dest);
+   CGLM_INLINE void  glm_vec4_div(vec4 a, vec4 b, vec4 dest);
+   CGLM_INLINE void  glm_vec4_divs(vec4 v, float s, vec4 dest);
+   CGLM_INLINE void  glm_vec4_addadd(vec4 a, vec4 b, vec4 dest);
+   CGLM_INLINE void  glm_vec4_subadd(vec4 a, vec4 b, vec4 dest);
+   CGLM_INLINE void  glm_vec4_muladd(vec4 a, vec4 b, vec4 dest);
    CGLM_INLINE void  glm_vec4_flipsign(vec4 v);
    CGLM_INLINE void  glm_vec4_inv(vec4 v);
    CGLM_INLINE void  glm_vec4_inv_to(vec4 v, vec4 dest);
@@ -40,6 +48,8 @@
    CGLM_INLINE float glm_vec4_distance(vec4 v1, vec4 v2);
    CGLM_INLINE void  glm_vec4_maxv(vec4 v1, vec4 v2, vec4 dest);
    CGLM_INLINE void  glm_vec4_minv(vec4 v1, vec4 v2, vec4 dest);
+   CGLM_INLINE void  glm_vec4_clamp(vec4 v, float minVal, float maxVal);
+   CGLM_INLINE void  glm_vec4_lerp(vec4 from, vec4 to, float t, vec4 dest)
  */
 
 #ifndef cglm_vec4_h
@@ -101,7 +111,7 @@ CGLM_INLINE
 void
 glm_vec4_copy(vec4 v, vec4 dest) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(dest, _mm_load_ps(v));
+  glmm_store(dest, glmm_load(v));
 #else
   dest[0] = v[0];
   dest[1] = v[1];
@@ -110,6 +120,42 @@ glm_vec4_copy(vec4 v, vec4 dest) {
 #endif
 }
 
+/*!
+ * @brief make vector zero
+ *
+ * @param[in, out]  v vector
+ */
+CGLM_INLINE
+void
+glm_vec4_zero(vec4 v) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(v, _mm_setzero_ps());
+#else
+  v[0] = 0.0f;
+  v[1] = 0.0f;
+  v[2] = 0.0f;
+  v[3] = 0.0f;
+#endif
+}
+
+/*!
+ * @brief make vector one
+ *
+ * @param[in, out]  v vector
+ */
+CGLM_INLINE
+void
+glm_vec4_one(vec4 v) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(v, _mm_set1_ps(1.0f));
+#else
+  v[0] = 1.0f;
+  v[1] = 1.0f;
+  v[2] = 1.0f;
+  v[3] = 1.0f;
+#endif
+}
+
 /*!
  * @brief vec4 dot product
  *
@@ -121,7 +167,14 @@ glm_vec4_copy(vec4 v, vec4 dest) {
 CGLM_INLINE
 float
 glm_vec4_dot(vec4 a, vec4 b) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  __m128 x0;
+  x0 = _mm_mul_ps(glmm_load(a), glmm_load(b));
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 1, 0, 3, 2));
+  return _mm_cvtss_f32(_mm_add_ss(x0, glmm_shuff1(x0, 0, 1, 0, 1)));
+#else
   return a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + a[3] * b[3];
+#endif
 }
 
 /*!
@@ -138,7 +191,15 @@ glm_vec4_dot(vec4 a, vec4 b) {
 CGLM_INLINE
 float
 glm_vec4_norm2(vec4 v) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  __m128 x0;
+  x0 = glmm_load(v);
+  x0 = _mm_mul_ps(x0, x0);
+  x0 = _mm_add_ps(x0, glmm_shuff1(x0, 1, 0, 3, 2));
+  return _mm_cvtss_f32(_mm_add_ss(x0, glmm_shuff1(x0, 0, 1, 0, 1)));
+#else
   return v[0] * v[0] + v[1] * v[1] + v[2] * v[2] + v[3] * v[3];
+#endif
 }
 
 /*!
@@ -151,50 +212,112 @@ glm_vec4_norm2(vec4 v) {
 CGLM_INLINE
 float
 glm_vec4_norm(vec4 vec) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  __m128 x0;
+  x0 = glmm_load(vec);
+  return _mm_cvtss_f32(_mm_sqrt_ss(glmm_dot(x0, x0)));
+#else
   return sqrtf(glm_vec4_norm2(vec));
+#endif
 }
 
 /*!
  * @brief add v2 vector to v1 vector store result in dest
  *
- * @param[in]  v1 vector1
- * @param[in]  v2 vector2
+ * @param[in]  a    vector1
+ * @param[in]  b    vector2
  * @param[out] dest destination vector
  */
 CGLM_INLINE
 void
-glm_vec4_add(vec4 v1, vec4 v2, vec4 dest) {
+glm_vec4_add(vec4 a, vec4 b, vec4 dest) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(dest,
-               _mm_add_ps(_mm_load_ps(v1),
-                          _mm_load_ps(v2)));
+  glmm_store(dest, _mm_add_ps(glmm_load(a), glmm_load(b)));
 #else
-  dest[0] = v1[0] + v2[0];
-  dest[1] = v1[1] + v2[1];
-  dest[2] = v1[2] + v2[2];
-  dest[3] = v1[3] + v2[3];
+  dest[0] = a[0] + b[0];
+  dest[1] = a[1] + b[1];
+  dest[2] = a[2] + b[2];
+  dest[3] = a[3] + b[3];
 #endif
 }
 
 /*!
- * @brief subtract v2 vector from v1 vector store result in dest
+ * @brief add scalar to v vector store result in dest (d = v + vec(s))
  *
- * @param[in]  v1 vector1
- * @param[in]  v2 vector2
+ * @param[in]  v    vector
+ * @param[in]  s    scalar
  * @param[out] dest destination vector
  */
 CGLM_INLINE
 void
-glm_vec4_sub(vec4 v1, vec4 v2, vec4 dest) {
+glm_vec4_adds(vec4 v, float s, vec4 dest) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(dest,
-               _mm_sub_ps(_mm_load_ps(v1),
-                          _mm_load_ps(v2)));
+  glmm_store(dest, _mm_add_ps(glmm_load(v), _mm_set1_ps(s)));
 #else
-  dest[0] = v1[0] - v2[0];
-  dest[1] = v1[1] - v2[1];
-  dest[2] = v1[2] - v2[2];
-  dest[3] = v1[3] - v2[3];
+  dest[0] = v[0] + s;
+  dest[1] = v[1] + s;
+  dest[2] = v[2] + s;
+  dest[3] = v[3] + s;
+#endif
+}
+
+/*!
+ * @brief subtract b vector from a vector store result in dest (d = v1 - v2)
+ *
+ * @param[in]  a    vector1
+ * @param[in]  b    vector2
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec4_sub(vec4 a, vec4 b, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_sub_ps(glmm_load(a), glmm_load(b)));
+#else
+  dest[0] = a[0] - b[0];
+  dest[1] = a[1] - b[1];
+  dest[2] = a[2] - b[2];
+  dest[3] = a[3] - b[3];
+#endif
+}
+
+/*!
+ * @brief subtract scalar from v vector store result in dest (d = v - vec(s))
+ *
+ * @param[in]  v    vector
+ * @param[in]  s    scalar
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec4_subs(vec4 v, float s, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_sub_ps(glmm_load(v), _mm_set1_ps(s)));
+#else
+  dest[0] = v[0] - s;
+  dest[1] = v[1] - s;
+  dest[2] = v[2] - s;
+  dest[3] = v[3] - s;
+#endif
+}
+
+/*!
+ * @brief multiply two vector (component-wise multiplication)
+ *
+ * @param a v1
+ * @param b v2
+ * @param d v3 = (a[0] * b[0], a[1] * b[1], a[2] * b[2], a[3] * b[3])
+ */
+CGLM_INLINE
+void
+glm_vec4_mul(vec4 a, vec4 b, vec4 d) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(d, _mm_mul_ps(glmm_load(a), glmm_load(b)));
+#else
+  d[0] = a[0] * b[0];
+  d[1] = a[1] * b[1];
+  d[2] = a[2] * b[2];
+  d[3] = a[3] * b[3];
 #endif
 }
 
@@ -209,9 +332,7 @@ CGLM_INLINE
 void
 glm_vec4_scale(vec4 v, float s, vec4 dest) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(dest,
-               _mm_mul_ps(_mm_load_ps(v),
-                          _mm_set1_ps(s)));
+  glmm_store(dest, _mm_mul_ps(glmm_load(v), _mm_set1_ps(s)));
 #else
   dest[0] = v[0] * s;
   dest[1] = v[1] * s;
@@ -233,14 +354,148 @@ glm_vec4_scale_as(vec4 v, float s, vec4 dest) {
   float norm;
   norm = glm_vec4_norm(v);
 
-  if (norm == 0) {
-    glm_vec4_copy(v, dest);
+  if (norm == 0.0f) {
+    glm_vec4_zero(dest);
     return;
   }
 
   glm_vec4_scale(v, s / norm, dest);
 }
 
+/*!
+ * @brief div vector with another component-wise division: d = v1 / v2
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest result = (a[0]/b[0], a[1]/b[1], a[2]/b[2], a[3]/b[3])
+ */
+CGLM_INLINE
+void
+glm_vec4_div(vec4 a, vec4 b, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_div_ps(glmm_load(a), glmm_load(b)));
+#else
+  dest[0] = a[0] / b[0];
+  dest[1] = a[1] / b[1];
+  dest[2] = a[2] / b[2];
+  dest[3] = a[3] / b[3];
+#endif
+}
+
+/*!
+ * @brief div vec4 vector with scalar: d = v / s
+ *
+ * @param[in]  v    vector
+ * @param[in]  s    scalar
+ * @param[out] dest destination vector
+ */
+CGLM_INLINE
+void
+glm_vec4_divs(vec4 v, float s, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_div_ps(glmm_load(v), _mm_set1_ps(s)));
+#else
+  glm_vec4_scale(v, 1.0f / s, dest);
+#endif
+}
+
+
+/*!
+ * @brief add two vectors and add result to sum
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest dest += (a + b)
+ */
+CGLM_INLINE
+void
+glm_vec4_addadd(vec4 a, vec4 b, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_add_ps(glmm_load(dest),
+                              _mm_add_ps(glmm_load(a),
+                                         glmm_load(b))));
+#else
+  dest[0] += a[0] + b[0];
+  dest[1] += a[1] + b[1];
+  dest[2] += a[2] + b[2];
+  dest[3] += a[3] + b[3];
+#endif
+}
+
+/*!
+ * @brief sub two vectors and add result to dest
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest dest += (a - b)
+ */
+CGLM_INLINE
+void
+glm_vec4_subadd(vec4 a, vec4 b, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_add_ps(glmm_load(dest),
+                              _mm_sub_ps(glmm_load(a),
+                                         glmm_load(b))));
+#else
+  dest[0] += a[0] - b[0];
+  dest[1] += a[1] - b[1];
+  dest[2] += a[2] - b[2];
+  dest[3] += a[3] - b[3];
+#endif
+}
+
+/*!
+ * @brief mul two vectors and add result to dest
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector 1
+ * @param[in]  b    vector 2
+ * @param[out] dest dest += (a * b)
+ */
+CGLM_INLINE
+void
+glm_vec4_muladd(vec4 a, vec4 b, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_add_ps(glmm_load(dest),
+                              _mm_mul_ps(glmm_load(a),
+                                         glmm_load(b))));
+#else
+  dest[0] += a[0] * b[0];
+  dest[1] += a[1] * b[1];
+  dest[2] += a[2] * b[2];
+  dest[3] += a[3] * b[3];
+#endif
+}
+
+/*!
+ * @brief mul vector with scalar and add result to sum
+ *
+ * it applies += operator so dest must be initialized
+ *
+ * @param[in]  a    vector
+ * @param[in]  s    scalar
+ * @param[out] dest dest += (a * b)
+ */
+CGLM_INLINE
+void
+glm_vec4_muladds(vec4 a, float s, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_add_ps(glmm_load(dest),
+                              _mm_mul_ps(glmm_load(a),
+                                         _mm_set1_ps(s))));
+#else
+  dest[0] += a[0] * s;
+  dest[1] += a[1] * s;
+  dest[2] += a[2] * s;
+  dest[3] += a[3] * s;
+#endif
+}
+
 /*!
  * @brief flip sign of all vec4 members
  *
@@ -250,8 +505,7 @@ CGLM_INLINE
 void
 glm_vec4_flipsign(vec4 v) {
 #if defined( __SSE__ ) || defined( __SSE2__ )
-  _mm_store_ps(v, _mm_xor_ps(_mm_load_ps(v),
-                             _mm_set1_ps(-0.0f)));
+  glmm_store(v, _mm_xor_ps(glmm_load(v), _mm_set1_ps(-0.0f)));
 #else
   v[0] = -v[0];
   v[1] = -v[1];
@@ -260,6 +514,25 @@ glm_vec4_flipsign(vec4 v) {
 #endif
 }
 
+/*!
+ * @brief flip sign of all vec4 members and store result in dest
+ *
+ * @param[in]  v     vector
+ * @param[out] dest  vector
+ */
+CGLM_INLINE
+void
+glm_vec4_flipsign_to(vec4 v, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_xor_ps(glmm_load(v), _mm_set1_ps(-0.0f)));
+#else
+  dest[0] = -v[0];
+  dest[1] = -v[1];
+  dest[2] = -v[2];
+  dest[3] = -v[3];
+#endif
+}
+
 /*!
  * @brief make vector as inverse/opposite of itself
  *
@@ -284,26 +557,6 @@ glm_vec4_inv_to(vec4 v, vec4 dest) {
   glm_vec4_flipsign(dest);
 }
 
-/*!
- * @brief normalize vec4 and store result in same vec
- *
- * @param[in, out] v vector
- */
-CGLM_INLINE
-void
-glm_vec4_normalize(vec4 v) {
-  float norm;
-
-  norm = glm_vec4_norm(v);
-
-  if (norm == 0.0f) {
-    v[0] = v[1] = v[2] = v[3] = 0.0f;
-    return;
-  }
-
-  glm_vec4_scale(v, 1.0f / norm, v);
-}
-
 /*!
  * @brief normalize vec4 to dest
  *
@@ -313,16 +566,43 @@ glm_vec4_normalize(vec4 v) {
 CGLM_INLINE
 void
 glm_vec4_normalize_to(vec4 vec, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  __m128 xdot, x0;
+  float  dot;
+
+  x0   = glmm_load(vec);
+  xdot = glmm_dot(x0, x0);
+  dot  = _mm_cvtss_f32(xdot);
+
+  if (dot == 0.0f) {
+    glmm_store(dest, _mm_setzero_ps());
+    return;
+  }
+
+  glmm_store(dest, _mm_div_ps(x0, _mm_sqrt_ps(xdot)));
+#else
   float norm;
 
   norm = glm_vec4_norm(vec);
 
   if (norm == 0.0f) {
-    dest[0] = dest[1] = dest[2] = dest[3] = 0.0f;
+    glm_vec4_zero(dest);
     return;
   }
 
   glm_vec4_scale(vec, 1.0f / norm, dest);
+#endif
+}
+
+/*!
+ * @brief normalize vec4 and store result in same vec
+ *
+ * @param[in, out] v vector
+ */
+CGLM_INLINE
+void
+glm_vec4_normalize(vec4 v) {
+  glm_vec4_normalize_to(v, v);
 }
 
 /**
@@ -351,10 +631,14 @@ glm_vec4_distance(vec4 v1, vec4 v2) {
 CGLM_INLINE
 void
 glm_vec4_maxv(vec4 v1, vec4 v2, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_max_ps(glmm_load(v1), glmm_load(v2)));
+#else
   dest[0] = glm_max(v1[0], v2[0]);
   dest[1] = glm_max(v1[1], v2[1]);
   dest[2] = glm_max(v1[2], v2[2]);
   dest[3] = glm_max(v1[3], v2[3]);
+#endif
 }
 
 /*!
@@ -367,10 +651,57 @@ glm_vec4_maxv(vec4 v1, vec4 v2, vec4 dest) {
 CGLM_INLINE
 void
 glm_vec4_minv(vec4 v1, vec4 v2, vec4 dest) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(dest, _mm_min_ps(glmm_load(v1), glmm_load(v2)));
+#else
   dest[0] = glm_min(v1[0], v2[0]);
   dest[1] = glm_min(v1[1], v2[1]);
   dest[2] = glm_min(v1[2], v2[2]);
   dest[3] = glm_min(v1[3], v2[3]);
+#endif
+}
+
+/*!
+ * @brief clamp vector's individual members between min and max values
+ *
+ * @param[in, out]  v      vector
+ * @param[in]       minVal minimum value
+ * @param[in]       maxVal maximum value
+ */
+CGLM_INLINE
+void
+glm_vec4_clamp(vec4 v, float minVal, float maxVal) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  glmm_store(v, _mm_min_ps(_mm_max_ps(glmm_load(v), _mm_set1_ps(minVal)),
+                           _mm_set1_ps(maxVal)));
+#else
+  v[0] = glm_clamp(v[0], minVal, maxVal);
+  v[1] = glm_clamp(v[1], minVal, maxVal);
+  v[2] = glm_clamp(v[2], minVal, maxVal);
+  v[3] = glm_clamp(v[3], minVal, maxVal);
+#endif
+}
+
+/*!
+ * @brief linear interpolation between two vector
+ *
+ * formula:  from + s * (to - from)
+ *
+ * @param[in]   from from value
+ * @param[in]   to   to value
+ * @param[in]   t    interpolant (amount) clamped between 0 and 1
+ * @param[out]  dest destination
+ */
+CGLM_INLINE
+void
+glm_vec4_lerp(vec4 from, vec4 to, float t, vec4 dest) {
+  vec4 s, v;
+
+  /* from + s * (to - from) */
+  glm_vec4_broadcast(glm_clamp(t, 0.0f, 1.0f), s);
+  glm_vec4_sub(to, from, v);
+  glm_vec4_mulv(s, v, v);
+  glm_vec4_add(from, v, dest);
 }
 
 #endif /* cglm_vec4_h */

include/cglm/version.h

@@ -9,7 +9,7 @@
 #define cglm_version_h
 
 #define CGLM_VERSION_MAJOR 0
-#define CGLM_VERSION_MINOR 3
-#define CGLM_VERSION_PATCH 5
+#define CGLM_VERSION_MINOR 4
+#define CGLM_VERSION_PATCH 7
 
 #endif /* cglm_version_h */

makefile.am

@@ -54,7 +54,9 @@ cglm_HEADERS = include/cglm/version.h \
                   include/cglm/plane.h \
                   include/cglm/frustum.h \
                   include/cglm/box.h \
-                  include/cglm/color.h
+                  include/cglm/color.h \
+                  include/cglm/project.h \
+                  include/cglm/sphere.h
 
 cglm_calldir=$(includedir)/cglm/call
 cglm_call_HEADERS = include/cglm/call/mat4.h \
@@ -68,7 +70,9 @@ cglm_call_HEADERS = include/cglm/call/mat4.h \
                     include/cglm/call/euler.h \
                     include/cglm/call/plane.h \
                     include/cglm/call/frustum.h \
-                    include/cglm/call/box.h
+                    include/cglm/call/box.h \
+                    include/cglm/call/project.h \
+                    include/cglm/call/sphere.h
 
 cglm_simddir=$(includedir)/cglm/simd
 cglm_simd_HEADERS = include/cglm/simd/intrin.h
@@ -98,13 +102,23 @@ libcglm_la_SOURCES=\
     src/mat4.c \
     src/plane.c \
     src/frustum.c \
-    src/box.c
+    src/box.c \
+    src/project.c \
+    src/sphere.c
 
 test_tests_SOURCES=\
     test/src/test_common.c \
     test/src/test_main.c \
     test/src/test_mat4.c \
-    test/src/test_cam.c
+    test/src/test_cam.c \
+    test/src/test_project.c \
+    test/src/test_clamp.c \
+    test/src/test_euler.c \
+    test/src/test_quat.c \
+    test/src/test_vec4.c \
+    test/src/test_vec3.c \
+    test/src/test_mat3.c \
+    test/src/test_affine.c
 
 all-local:
 	sh ./post-build.sh

src/affine.c

@@ -8,6 +8,12 @@
 #include "../include/cglm/cglm.h"
 #include "../include/cglm/call.h"
 
+CGLM_EXPORT
+void
+glmc_translate_make(mat4 m, vec3 v) {
+  glm_translate_make(m, v);
+}
+
 CGLM_EXPORT
 void
 glmc_translate_to(mat4 m, vec3 v, mat4 dest) {
@@ -38,6 +44,12 @@ glmc_translate_z(mat4 m, float to) {
   glm_translate_z(m, to);
 }
 
+CGLM_EXPORT
+void
+glmc_scale_make(mat4 m, vec3 v) {
+  glm_scale_make(m, v);
+}
+
 CGLM_EXPORT
 void
 glmc_scale_to(mat4 m, vec3 v, mat4 dest) {
@@ -52,8 +64,8 @@ glmc_scale(mat4 m, vec3 v) {
 
 CGLM_EXPORT
 void
-glmc_scale1(mat4 m, float s) {
-  glm_scale1(m, s);
+glmc_scale_uni(mat4 m, float s) {
+  glm_scale_uni(m, s);
 }
 
 CGLM_EXPORT
@@ -74,36 +86,42 @@ glmc_rotate_z(mat4 m, float rad, mat4 dest) {
   glm_rotate_z(m, rad, dest);
 }
 
-CGLM_EXPORT
-void
-glmc_rotate_ndc_make(mat4 m, float angle, vec3 axis_ndc) {
-  glm_rotate_ndc_make(m, angle, axis_ndc);
-}
-
 CGLM_EXPORT
 void
 glmc_rotate_make(mat4 m, float angle, vec3 axis) {
   glm_rotate_make(m, angle, axis);
 }
 
-CGLM_EXPORT
-void
-glmc_rotate_ndc(mat4 m, float angle, vec3 axis_ndc) {
-  glm_rotate_ndc(m, angle, axis_ndc);
-}
-
 CGLM_EXPORT
 void
 glmc_rotate(mat4 m, float angle, vec3 axis) {
   glm_rotate(m, angle, axis);
 }
 
+CGLM_EXPORT
+void
+glmc_rotate_at(mat4 m, vec3 pivot, float angle, vec3 axis) {
+  glm_rotate_at(m, pivot, angle, axis);
+}
+
+CGLM_EXPORT
+void
+glmc_rotate_atm(mat4 m, vec3 pivot, float angle, vec3 axis) {
+  glm_rotate_atm(m, pivot, angle, axis);
+}
+
 CGLM_EXPORT
 void
 glmc_decompose_scalev(mat4 m, vec3 s) {
   glm_decompose_scalev(m, s);
 }
 
+CGLM_EXPORT
+bool
+glmc_uniscaled(mat4 m) {
+  return glm_uniscaled(m);
+}
+
 CGLM_EXPORT
 void
 glmc_decompose_rs(mat4 m, mat4 r, vec3 s) {
@@ -115,3 +133,21 @@ void
 glmc_decompose(mat4 m, vec4 t, mat4 r, vec3 s) {
   glm_decompose(m, t, r, s);
 }
+
+CGLM_EXPORT
+void
+glmc_mul(mat4 m1, mat4 m2, mat4 dest) {
+  glm_mul(m1, m2, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_mul_rot(mat4 m1, mat4 m2, mat4 dest) {
+  glm_mul_rot(m1, m2, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_inv_tr(mat4 mat) {
+  glm_inv_tr(mat);
+}

src/box.c

@@ -34,3 +34,63 @@ glmc_aabb_crop_until(vec3 box[2],
                      vec3 dest[2]) {
   glm_aabb_crop_until(box, cropBox, clampBox, dest);
 }
+
+CGLM_EXPORT
+bool
+glmc_aabb_frustum(vec3 box[2], vec4 planes[6]) {
+  return glm_aabb_frustum(box, planes);
+}
+
+CGLM_EXPORT
+void
+glmc_aabb_invalidate(vec3 box[2]) {
+  glm_aabb_invalidate(box);
+}
+
+CGLM_EXPORT
+bool
+glmc_aabb_isvalid(vec3 box[2]) {
+  return glm_aabb_isvalid(box);
+}
+
+CGLM_EXPORT
+float
+glmc_aabb_size(vec3 box[2]) {
+  return glm_aabb_size(box);
+}
+
+CGLM_EXPORT
+float
+glmc_aabb_radius(vec3 box[2]) {
+  return glm_aabb_radius(box);
+}
+
+CGLM_EXPORT
+void
+glmc_aabb_center(vec3 box[2], vec3 dest) {
+  glm_aabb_center(box, dest);
+}
+
+CGLM_EXPORT
+bool
+glmc_aabb_aabb(vec3 box[2], vec3 other[2]) {
+  return glm_aabb_aabb(box, other);
+}
+
+CGLM_EXPORT
+bool
+glmc_aabb_point(vec3 box[2], vec3 point) {
+  return glm_aabb_point(box, point);
+}
+
+CGLM_EXPORT
+bool
+glmc_aabb_contains(vec3 box[2], vec3 other[2]) {
+  return glm_aabb_contains(box, other);
+}
+
+CGLM_EXPORT
+bool
+glmc_aabb_sphere(vec3 box[2], vec4 s) {
+  return glm_aabb_sphere(box, s);
+}

src/cam.c

@@ -44,6 +44,36 @@ glmc_ortho(float left,
             dest);
 }
 
+CGLM_EXPORT
+void
+glmc_ortho_aabb(vec3 box[2], mat4 dest) {
+  glm_ortho_aabb(box, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_ortho_aabb_p(vec3 box[2], float padding, mat4 dest) {
+  glm_ortho_aabb_p(box, padding, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_ortho_aabb_pz(vec3 box[2], float padding, mat4 dest) {
+  glm_ortho_aabb_pz(box, padding, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_ortho_default(float aspect, mat4 dest) {
+  glm_ortho_default(aspect, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_ortho_default_s(float aspect, float size, mat4 dest) {
+  glm_ortho_default_s(aspect, size, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_perspective(float fovy,
@@ -58,6 +88,18 @@ glmc_perspective(float fovy,
                   dest);
 }
 
+CGLM_EXPORT
+void
+glmc_perspective_default(float aspect, mat4 dest) {
+  glm_perspective_default(aspect, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_perspective_resize(float aspect, mat4 proj) {
+  glm_perspective_resize(aspect, proj);
+}
+
 CGLM_EXPORT
 void
 glmc_lookat(vec3 eye,
@@ -78,3 +120,75 @@ void
 glmc_look_anyup(vec3 eye, vec3 dir, mat4 dest) {
   glm_look_anyup(eye, dir, dest);
 }
+
+CGLM_EXPORT
+void
+glmc_persp_decomp(mat4 proj,
+                  float * __restrict nearVal,
+                  float * __restrict farVal,
+                  float * __restrict top,
+                  float * __restrict bottom,
+                  float * __restrict left,
+                  float * __restrict right) {
+  glm_persp_decomp(proj, nearVal, farVal, top, bottom, left, right);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_decompv(mat4 proj, float dest[6]) {
+  glm_persp_decompv(proj, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_x(mat4 proj,
+                    float * __restrict left,
+                    float * __restrict right) {
+  glm_persp_decomp_x(proj, left, right);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_y(mat4 proj,
+                    float * __restrict top,
+                    float * __restrict bottom) {
+  glm_persp_decomp_y(proj, top, bottom);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_z(mat4 proj,
+                    float * __restrict nearVal,
+                    float * __restrict farVal) {
+  glm_persp_decomp_z(proj, nearVal, farVal);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_far(mat4 proj, float * __restrict farVal) {
+  glm_persp_decomp_far(proj, farVal);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_decomp_near(mat4 proj, float * __restrict nearVal) {
+  glm_persp_decomp_near(proj, nearVal);
+}
+
+CGLM_EXPORT
+float
+glmc_persp_fovy(mat4 proj) {
+  return glm_persp_fovy(proj);
+}
+
+CGLM_EXPORT
+float
+glmc_persp_aspect(mat4 proj) {
+  return glm_persp_aspect(proj);
+}
+
+CGLM_EXPORT
+void
+glmc_persp_sizes(mat4 proj, float fovy, vec4 dest) {
+  glm_persp_sizes(proj, fovy, dest);
+}

src/euler.c

@@ -20,6 +20,12 @@ glmc_euler(vec3 angles, mat4 dest) {
   glm_euler(angles, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_euler_xyz(vec3 angles,  mat4 dest) {
+  glm_euler_xyz(angles, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_euler_zyx(vec3 angles,  mat4 dest) {

src/mat3.c

@@ -44,6 +44,12 @@ glmc_mat3_mulv(mat3 m, vec3 v, vec3 dest) {
   glm_mat3_mulv(m, v, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_mat3_quat(mat3 m, versor dest) {
+  glm_mat3_quat(m, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_mat3_scale(mat3 m, float s) {

src/mat4.c

@@ -52,7 +52,7 @@ glmc_mat4_mul(mat4 m1, mat4 m2, mat4 dest) {
 
 CGLM_EXPORT
 void
-glmc_mat4_mulN(mat4 * __restrict matrices[], int len, mat4 dest) {
+glmc_mat4_mulN(mat4 * __restrict matrices[], uint32_t len, mat4 dest) {
   glm_mat4_mulN(matrices, len, dest);
 }
 
@@ -62,6 +62,18 @@ glmc_mat4_mulv(mat4 m, vec4 v, vec4 dest) {
   glm_mat4_mulv(m, v, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_mat4_mulv3(mat4 m, vec3 v, float last, vec3 dest) {
+  glm_mat4_mulv3(m, v, last, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_mat4_quat(mat4 m, versor dest) {
+  glm_mat4_quat(m, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_mat4_transpose_to(mat4 m, mat4 dest) {
@@ -104,6 +116,12 @@ glmc_mat4_inv_precise(mat4 mat, mat4 dest) {
   glm_mat4_inv_precise(mat, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_mat4_inv_fast(mat4 mat, mat4 dest) {
+  glm_mat4_inv_fast(mat, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_mat4_swap_col(mat4 mat, int col1, int col2) {

src/project.c

@@ -0,0 +1,27 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "../include/cglm/cglm.h"
+#include "../include/cglm/call.h"
+
+CGLM_EXPORT
+void
+glmc_unprojecti(vec3 pos, mat4 invMat, vec4 vp, vec3 dest) {
+  glm_unprojecti(pos, invMat, vp, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_unproject(vec3 pos, mat4 m, vec4 vp, vec3 dest) {
+  glm_unproject(pos, m, vp, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_project(vec3 pos, mat4 m, vec4 vp, vec3 dest) {
+  glm_project(pos, m, vp, dest);
+}

src/quat.c

@@ -8,6 +8,7 @@
 #include "../include/cglm/cglm.h"
 #include "../include/cglm/call.h"
 
+
 CGLM_EXPORT
 void
 glmc_quat_identity(versor q) {
@@ -16,20 +17,26 @@ glmc_quat_identity(versor q) {
 
 CGLM_EXPORT
 void
-glmc_quat(versor q,
-               float angle,
-               float x,
-               float y,
-               float z) {
+glmc_quat_init(versor q, float x, float y, float z, float w) {
+  glm_quat_init(q, x, y, z, w);
+}
+
+CGLM_EXPORT
+void
+glmc_quat(versor q, float angle, float x, float y, float z) {
   glm_quat(q, angle, x, y, z);
 }
 
 CGLM_EXPORT
 void
-glmc_quatv(versor q,
-           float  angle,
-           vec3   v) {
-  glm_quatv(q, angle, v);
+glmc_quatv(versor q, float angle, vec3 axis) {
+  glm_quatv(q, angle, axis);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_copy(versor q, versor dest) {
+  glm_quat_copy(q, dest);
 }
 
 CGLM_EXPORT
@@ -40,20 +47,86 @@ glmc_quat_norm(versor q) {
 
 CGLM_EXPORT
 void
-glmc_quat_normalize(versor q) {
-  glm_quat_normalize(q);
-}
-
-CGLM_EXPORT
-float
-glmc_quat_dot(versor q, versor r) {
-  return glm_quat_dot(q, r);
+glmc_quat_normalize_to(versor q, versor dest) {
+  glm_quat_normalize_to(q, dest);
 }
 
 CGLM_EXPORT
 void
-glmc_quat_mulv(versor q1, versor q2, versor dest) {
-  glm_quat_mulv(q1, q2, dest);
+glmc_quat_normalize(versor q) {
+  glm_quat_norm(q);
+}
+
+CGLM_EXPORT
+float
+glmc_quat_dot(versor p, versor q) {
+  return glm_quat_dot(p, q);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_conjugate(versor q, versor dest) {
+  glm_quat_conjugate(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_inv(versor q, versor dest) {
+  glm_quat_inv(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_add(versor p, versor q, versor dest) {
+  glm_quat_add(p, q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_sub(versor p, versor q, versor dest) {
+  glm_quat_sub(p, q, dest);
+}
+
+CGLM_EXPORT
+float
+glmc_quat_real(versor q) {
+  return glm_quat_real(q);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_imag(versor q, vec3 dest) {
+  glm_quat_imag(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_imagn(versor q, vec3 dest) {
+  glm_quat_imagn(q, dest);
+}
+
+CGLM_EXPORT
+float
+glmc_quat_imaglen(versor q) {
+  return glm_quat_imaglen(q);
+}
+
+CGLM_EXPORT
+float
+glmc_quat_angle(versor q) {
+  return glm_quat_angle(q);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_axis(versor q, versor dest) {
+  glm_quat_axis(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_mul(versor p, versor q, versor dest) {
+  glm_quat_mul(p, q, dest);
 }
 
 CGLM_EXPORT
@@ -64,9 +137,72 @@ glmc_quat_mat4(versor q, mat4 dest) {
 
 CGLM_EXPORT
 void
-glmc_quat_slerp(versor q,
-                versor r,
-                float  t,
-                versor dest) {
-  glm_quat_slerp(q, r, t, dest);
+glmc_quat_mat4t(versor q, mat4 dest) {
+  glm_quat_mat4t(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_mat3(versor q, mat3 dest) {
+  glm_quat_mat3(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_mat3t(versor q, mat3 dest) {
+  glm_quat_mat3t(q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_lerp(versor from, versor to, float t, versor dest) {
+  glm_quat_lerp(from, to, t, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_slerp(versor from, versor to, float t, versor dest) {
+  glm_quat_slerp(from, to, t, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_look(vec3 eye, versor ori, mat4 dest) {
+  glm_quat_look(eye, ori, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_for(vec3 dir, vec3 fwd, vec3 up, versor dest) {
+  glm_quat_for(dir, fwd, up, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_forp(vec3 from, vec3 to, vec3 fwd, vec3 up, versor dest) {
+  glm_quat_forp(from, to, fwd, up, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_rotatev(versor q, vec3 v, vec3 dest) {
+  glm_quat_rotatev(q, v, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_rotate(mat4 m, versor q, mat4 dest) {
+  glm_quat_rotate(m, q, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_rotate_at(mat4 model, versor q, vec3 pivot) {
+  glm_quat_rotate_at(model, q, pivot);
+}
+
+CGLM_EXPORT
+void
+glmc_quat_rotate_atm(mat4 m, versor q, vec3 pivot) {
+  glm_quat_rotate_atm(m, q, pivot);
 }

src/sphere.c

@@ -0,0 +1,39 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "../include/cglm/cglm.h"
+#include "../include/cglm/call.h"
+
+CGLM_EXPORT
+float
+glmc_sphere_radii(vec4 s) {
+  return glm_sphere_radii(s);
+}
+
+CGLM_EXPORT
+void
+glmc_sphere_transform(vec4 s, mat4 m, vec4 dest) {
+  glm_sphere_transform(s, m, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_sphere_merge(vec4 s1, vec4 s2, vec4 dest) {
+  glm_sphere_merge(s1, s2, dest);
+}
+
+CGLM_EXPORT
+bool
+glmc_sphere_sphere(vec4 s1, vec4 s2) {
+  return glm_sphere_sphere(s1, s2);
+}
+
+CGLM_EXPORT
+bool
+glmc_sphere_point(vec4 s, vec3 point) {
+  return glm_sphere_point(s, point);
+}

src/vec3.c

@@ -8,12 +8,30 @@
 #include "../include/cglm/cglm.h"
 #include "../include/cglm/call.h"
 
+CGLM_EXPORT
+void
+glmc_vec3(vec4 v4, vec3 dest) {
+  glm_vec3(v4, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_vec_copy(vec3 a, vec3 dest) {
   glm_vec_copy(a, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_vec_zero(vec3 v) {
+  glm_vec_zero(v);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_one(vec3 v) {
+  glm_vec_one(v);
+}
+
 CGLM_EXPORT
 float
 glmc_vec_dot(vec3 a, vec3 b) {
@@ -58,8 +76,26 @@ glmc_vec_add(vec3 v1, vec3 v2, vec3 dest) {
 
 CGLM_EXPORT
 void
-glmc_vec_sub(vec3 v1, vec3 v2, vec3 dest) {
-  glm_vec_sub(v1, v2, dest);
+glmc_vec_adds(vec3 v, float s, vec3 dest) {
+  glm_vec_adds(v, s, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_sub(vec3 a, vec3 b, vec3 dest) {
+  glm_vec_sub(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_subs(vec3 v, float s, vec3 dest) {
+  glm_vec_subs(v, s, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_mul(vec3 a, vec3 b, vec3 d) {
+  glm_vec_mul(a, b, d);
 }
 
 CGLM_EXPORT
@@ -74,12 +110,54 @@ glmc_vec_scale_as(vec3 v, float s, vec3 dest) {
   glm_vec_scale_as(v, s, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_vec_div(vec3 a, vec3 b, vec3 dest) {
+  glm_vec_div(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_divs(vec3 a, float s, vec3 dest) {
+  glm_vec_divs(a, s, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_addadd(vec3 a, vec3 b, vec3 dest) {
+  glm_vec_addadd(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_subadd(vec3 a, vec3 b, vec3 dest) {
+  glm_vec_subadd(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_muladd(vec3 a, vec3 b, vec3 dest) {
+  glm_vec_muladd(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_muladds(vec3 a, float s, vec3 dest) {
+  glm_vec_muladds(a, s, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_vec_flipsign(vec3 v) {
   glm_vec_flipsign(v);
 }
 
+CGLM_EXPORT
+void
+glmc_vec_flipsign_to(vec3 v, vec3 dest) {
+  glm_vec_flipsign_to(v, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_vec_inv(vec3 v) {
@@ -110,6 +188,12 @@ glmc_vec_rotate_m4(mat4 m, vec3 v, vec3 dest) {
   glm_vec_rotate_m4(m, v, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_vec_rotate_m3(mat3 m, vec3 v, vec3 dest) {
+  glm_vec_rotate_m3(m, v, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_vec_proj(vec3 a, vec3 b, vec3 dest) {
@@ -122,6 +206,12 @@ glmc_vec_center(vec3 v1, vec3 v2, vec3 dest) {
   glm_vec_center(v1, v2, dest);
 }
 
+CGLM_EXPORT
+float
+glmc_vec_distance2(vec3 v1, vec3 v2) {
+  return glm_vec_distance2(v1, v2);
+}
+
 CGLM_EXPORT
 float
 glmc_vec_distance(vec3 v1, vec3 v2) {
@@ -139,3 +229,107 @@ void
 glmc_vec_minv(vec3 v1, vec3 v2, vec3 dest) {
   glm_vec_maxv(v1, v2, dest);
 }
+
+CGLM_EXPORT
+void
+glmc_vec_clamp(vec3 v, float minVal, float maxVal) {
+  glm_vec_clamp(v, minVal, maxVal);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_ortho(vec3 v, vec3 dest) {
+  glm_vec_ortho(v, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_lerp(vec3 from, vec3 to, float t, vec3 dest) {
+  glm_vec_lerp(from, to, t, dest);
+}
+
+/* ext */
+
+CGLM_EXPORT
+void
+glmc_vec_mulv(vec3 a, vec3 b, vec3 d) {
+  glm_vec_mulv(a, b, d);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_broadcast(float val, vec3 d) {
+  glm_vec_broadcast(val, d);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_eq(vec3 v, float val) {
+  return glm_vec_eq(v, val);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_eq_eps(vec3 v, float val) {
+  return glm_vec_eq_eps(v, val);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_eq_all(vec3 v) {
+  return glm_vec_eq_all(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_eqv(vec3 v1, vec3 v2) {
+  return glm_vec_eqv(v1, v2);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_eqv_eps(vec3 v1, vec3 v2) {
+  return glm_vec_eqv_eps(v1, v2);
+}
+
+CGLM_EXPORT
+float
+glmc_vec_max(vec3 v) {
+  return glm_vec_max(v);
+}
+
+CGLM_EXPORT
+float
+glmc_vec_min(vec3 v) {
+  return glm_vec_min(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_isnan(vec3 v) {
+  return glm_vec_isnan(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_isinf(vec3 v) {
+  return glm_vec_isinf(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec_isvalid(vec3 v) {
+  return glm_vec_isvalid(v);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_sign(vec3 v, vec3 dest) {
+  glm_vec_sign(v, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec_sqrt(vec3 v, vec3 dest) {
+  glm_vec_sqrt(v, dest);
+}

src/vec4.c

@@ -8,6 +8,24 @@
 #include "../include/cglm/cglm.h"
 #include "../include/cglm/call.h"
 
+CGLM_EXPORT
+void
+glmc_vec4(vec3 v3, float last, vec4 dest) {
+  glm_vec4(v3, last, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_zero(vec4 v) {
+  glm_vec4_zero(v);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_one(vec4 v) {
+  glm_vec4_one(v);
+}
+
 CGLM_EXPORT
 void
 glmc_vec4_copy3(vec4 a, vec3 dest) {
@@ -52,14 +70,32 @@ glmc_vec4_norm2(vec4 vec) {
 
 CGLM_EXPORT
 void
-glmc_vec4_add(vec4 v1, vec4 v2, vec4 dest) {
-  glm_vec4_add(v1, v2, dest);
+glmc_vec4_add(vec4 a, vec4 b, vec4 dest) {
+  glm_vec4_add(a, b, dest);
 }
 
 CGLM_EXPORT
 void
-glmc_vec4_sub(vec4 v1, vec4 v2, vec4 dest) {
-  glm_vec4_sub(v1, v2, dest);
+glmc_vec4_adds(vec4 v, float s, vec4 dest) {
+  glm_vec4_adds(v, s, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_sub(vec4 a, vec4 b, vec4 dest) {
+  glm_vec4_sub(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_subs(vec4 v, float s, vec4 dest) {
+  glm_vec4_subs(v, s, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_mul(vec4 a, vec4 b, vec4 d) {
+  glm_vec4_mul(a, b, d);
 }
 
 CGLM_EXPORT
@@ -70,16 +106,58 @@ glmc_vec4_scale(vec4 v, float s, vec4 dest) {
 
 CGLM_EXPORT
 void
-glmc_vec4_scale_as(vec3 v, float s, vec3 dest) {
+glmc_vec4_scale_as(vec4 v, float s, vec4 dest) {
   glm_vec4_scale_as(v, s, dest);
 }
 
+CGLM_EXPORT
+void
+glmc_vec4_div(vec4 a, vec4 b, vec4 dest) {
+  glm_vec4_div(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_divs(vec4 v, float s, vec4 dest) {
+  glm_vec4_divs(v, s, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_addadd(vec4 a, vec4 b, vec4 dest) {
+  glm_vec4_addadd(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_subadd(vec4 a, vec4 b, vec4 dest) {
+  glm_vec4_subadd(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_muladd(vec4 a, vec4 b, vec4 dest) {
+  glm_vec4_muladd(a, b, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_muladds(vec4 a, float s, vec4 dest) {
+  glm_vec4_muladds(a, s, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_vec4_flipsign(vec4 v) {
   glm_vec4_flipsign(v);
 }
 
+CGLM_EXPORT
+void
+glmc_vec4_flipsign_to(vec4 v, vec4 dest) {
+  glm_vec4_flipsign_to(v, dest);
+}
+
 CGLM_EXPORT
 void
 glmc_vec4_inv(vec4 v) {
@@ -109,3 +187,101 @@ void
 glmc_vec4_minv(vec4 v1, vec4 v2, vec4 dest) {
   glm_vec4_maxv(v1, v2, dest);
 }
+
+CGLM_EXPORT
+void
+glmc_vec4_clamp(vec4 v, float minVal, float maxVal) {
+  glm_vec4_clamp(v, minVal, maxVal);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_lerp(vec4 from, vec4 to, float t, vec4 dest) {
+  glm_vec4_lerp(from, to, t, dest);
+}
+
+/* ext */
+
+CGLM_EXPORT
+void
+glmc_vec4_mulv(vec4 a, vec4 b, vec4 d) {
+  glm_vec4_mulv(a, b, d);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_broadcast(float val, vec4 d) {
+  glm_vec4_broadcast(val, d);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_eq(vec4 v, float val) {
+  return glm_vec4_eq(v, val);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_eq_eps(vec4 v, float val) {
+  return glm_vec4_eq_eps(v, val);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_eq_all(vec4 v) {
+  return glm_vec4_eq_all(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_eqv(vec4 v1, vec4 v2) {
+  return glm_vec4_eqv(v1, v2);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_eqv_eps(vec4 v1, vec4 v2) {
+  return glm_vec4_eqv_eps(v1, v2);
+}
+
+CGLM_EXPORT
+float
+glmc_vec4_max(vec4 v) {
+  return glm_vec4_max(v);
+}
+
+CGLM_EXPORT
+float
+glmc_vec4_min(vec4 v) {
+  return glm_vec4_min(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_isnan(vec4 v) {
+  return glm_vec4_isnan(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_isinf(vec4 v) {
+  return glm_vec4_isinf(v);
+}
+
+CGLM_EXPORT
+bool
+glmc_vec4_isvalid(vec4 v) {
+  return glm_vec4_isvalid(v);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_sign(vec4 v, vec4 dest) {
+  glm_vec4_sign(v, dest);
+}
+
+CGLM_EXPORT
+void
+glmc_vec4_sqrt(vec4 v, vec4 dest) {
+  glm_vec4_sqrt(v, dest);
+}

test/src/test_affine.c

@@ -0,0 +1,113 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+void
+test_affine(void **state) {
+  mat4 t1, t2, t3, t4, t5;
+
+  /* test translate is postmultiplied */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){34, 57, 36});
+
+  glmc_mat4_mul(t1, t2, t3); /* R * T */
+
+  glm_translate(t1, (vec3){34, 57, 36});
+  test_assert_mat4_eq(t1, t3);
+
+  /* test rotate is postmultiplied */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){34, 57, 36});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+
+  glm_rotate(t2, M_PI_4, GLM_YUP);
+  test_assert_mat4_eq(t2, t3);
+
+  /* test scale is postmultiplied */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){34, 57, 36});
+  glm_scale_make(t4, (vec3){3, 5, 6});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+  glmc_mat4_mul(t3, t4, t5); /* T * R * S */
+
+  glm_scale(t3, (vec3){3, 5, 6});
+  test_assert_mat4_eq(t3, t5);
+
+  /* test translate_x */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){34, 0, 0});
+
+  glmc_mat4_mul(t1, t2, t3); /* R * T */
+  glm_translate_x(t1, 34);
+  test_assert_mat4_eq(t1, t3);
+
+  /* test translate_y */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){0, 57, 0});
+
+  glmc_mat4_mul(t1, t2, t3); /* R * T */
+  glm_translate_y(t1, 57);
+  test_assert_mat4_eq(t1, t3);
+
+  /* test translate_z */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){0, 0, 36});
+
+  glmc_mat4_mul(t1, t2, t3); /* R * T */
+  glm_translate_z(t1, 36);
+  test_assert_mat4_eq(t1, t3);
+
+  /* test rotate_x */
+  glmc_rotate_make(t1, M_PI_4, (vec3){1, 0, 0});
+  glm_translate_make(t2, (vec3){34, 57, 36});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+
+  glm_rotate_x(t2, M_PI_4, t2);
+  test_assert_mat4_eq(t2, t3);
+
+  /* test rotate_y */
+  glmc_rotate_make(t1, M_PI_4, (vec3){0, 1, 0});
+  glm_translate_make(t2, (vec3){34, 57, 36});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+
+  glm_rotate_y(t2, M_PI_4, t2);
+  test_assert_mat4_eq(t2, t3);
+
+  /* test rotate_z */
+  glmc_rotate_make(t1, M_PI_4, (vec3){0, 0, 1});
+  glm_translate_make(t2, (vec3){34, 57, 36});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+
+  glm_rotate_z(t2, M_PI_4, t2);
+  test_assert_mat4_eq(t2, t3);
+
+  /* test rotate */
+  glmc_rotate_make(t1, M_PI_4, (vec3){0, 0, 1});
+  glm_translate_make(t2, (vec3){34, 57, 36});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+  glmc_rotate(t2, M_PI_4, (vec3){0, 0, 1});
+
+  test_assert_mat4_eq(t3, t2);
+
+  /* test scale_uni */
+  glmc_rotate_make(t1, M_PI_4, GLM_YUP);
+  glm_translate_make(t2, (vec3){34, 57, 36});
+  glm_scale_make(t4, (vec3){3, 3, 3});
+
+  glmc_mat4_mul(t2, t1, t3); /* T * R */
+  glmc_mat4_mul(t3, t4, t5); /* T * R * S */
+
+  glm_scale_uni(t3, 3);
+  test_assert_mat4_eq(t3, t5);
+}

test/src/test_clamp.c

@@ -0,0 +1,30 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+void
+test_clamp(void **state) {
+  vec3 v3 = {15.07, 0.4, 17.3};
+  vec4 v4 = {5.07,  2.3, 1.3, 1.4};
+
+  assert_true(glm_clamp(1.6f, 0.0f, 1.0f)  == 1.0f);
+  assert_true(glm_clamp(-1.6f, 0.0f, 1.0f) == 0.0f);
+  assert_true(glm_clamp(0.6f, 0.0f, 1.0f)  == 0.6f);
+
+  glm_vec_clamp(v3, 0.0, 1.0);
+  glm_vec4_clamp(v4, 1.5, 3.0);
+
+  assert_true(v3[0] == 1.0f);
+  assert_true(v3[1] == 0.4f);
+  assert_true(v3[2] == 1.0f);
+
+  assert_true(v4[0] == 3.0f);
+  assert_true(v4[1] == 2.3f);
+  assert_true(v4[2] == 1.5f);
+  assert_true(v4[3] == 1.5f);
+}

test/src/test_common.c

@@ -27,6 +27,50 @@ test_rand_mat4(mat4 dest) {
   /* glm_scale(dest, (vec3){drand48(), drand48(), drand48()}); */
 }
 
+void
+test_rand_mat3(mat3 dest) {
+  mat4 m4;
+
+  srand((unsigned int)time(NULL));
+
+  /* random rotatation around random axis with random angle */
+  glm_rotate_make(m4, drand48(), (vec3){drand48(), drand48(), drand48()});
+  glm_mat4_pick3(m4, dest);
+}
+
+void
+test_rand_vec3(vec3 dest) {
+  srand((unsigned int)time(NULL));
+
+  dest[0] = drand48();
+  dest[1] = drand48();
+  dest[2] = drand48();
+}
+
+void
+test_rand_vec4(vec4 dest) {
+  srand((unsigned int)time(NULL));
+
+  dest[0] = drand48();
+  dest[1] = drand48();
+  dest[2] = drand48();
+  dest[3] = drand48();
+}
+
+float
+test_rand_angle(void) {
+  srand((unsigned int)time(NULL));
+  return drand48();
+}
+
+void
+test_rand_quat(versor q) {
+  srand((unsigned int)time(NULL));
+
+  glm_quat(q, drand48(), drand48(), drand48(), drand48());
+  glm_quat_normalize(q);
+}
+
 void
 test_assert_mat4_eq(mat4 m1, mat4 m2) {
   int i, j, k;
@@ -50,3 +94,52 @@ test_assert_mat4_eq2(mat4 m1, mat4 m2, float eps) {
     }
   }
 }
+
+void
+test_assert_mat3_eq(mat3 m1, mat3 m2) {
+  int i, j, k;
+
+  for (i = 0; i < 3; i++) {
+    for (j = 0; j < 3; j++) {
+      for (k = 0; k < 3; k++)
+        assert_true(fabsf(m1[i][j] - m2[i][j]) <= 0.0000009);
+    }
+  }
+}
+
+void
+test_assert_eqf(float a, float b) {
+  assert_true(fabsf(a - b) <= 0.000009); /* rounding errors */
+}
+
+void
+test_assert_vec3_eq(vec3 v1, vec3 v2) {
+  assert_true(fabsf(v1[0] - v2[0]) <= 0.000009); /* rounding errors */
+  assert_true(fabsf(v1[1] - v2[1]) <= 0.000009);
+  assert_true(fabsf(v1[2] - v2[2]) <= 0.000009);
+}
+
+void
+test_assert_vec4_eq(vec4 v1, vec4 v2) {
+  assert_true(fabsf(v1[0] - v2[0]) <= 0.000009); /* rounding errors */
+  assert_true(fabsf(v1[1] - v2[1]) <= 0.000009);
+  assert_true(fabsf(v1[2] - v2[2]) <= 0.000009);
+  assert_true(fabsf(v1[3] - v2[3]) <= 0.000009);
+}
+
+void
+test_assert_quat_eq_abs(versor v1, versor v2) {
+  assert_true(fabsf(fabsf(v1[0]) - fabsf(v2[0])) <= 0.0009); /* rounding errors */
+  assert_true(fabsf(fabsf(v1[1]) - fabsf(v2[1])) <= 0.0009);
+  assert_true(fabsf(fabsf(v1[2]) - fabsf(v2[2])) <= 0.0009);
+  assert_true(fabsf(fabsf(v1[3]) - fabsf(v2[3])) <= 0.0009);
+}
+
+void
+test_assert_quat_eq(versor v1, versor v2) {
+  assert_true(fabsf(v1[0] - v2[0]) <= 0.000009); /* rounding errors */
+  assert_true(fabsf(v1[1] - v2[1]) <= 0.000009);
+  assert_true(fabsf(v1[2] - v2[2]) <= 0.000009);
+  assert_true(fabsf(v1[3] - v2[3]) <= 0.000009);
+}
+

test/src/test_common.h

@@ -25,10 +25,43 @@
 void
 test_rand_mat4(mat4 dest);
 
+void
+test_rand_mat3(mat3 dest);
+
+void
+test_assert_eqf(float a, float b);
+
 void
 test_assert_mat4_eq(mat4 m1, mat4 m2);
 
 void
 test_assert_mat4_eq2(mat4 m1, mat4 m2, float eps);
 
+void
+test_assert_mat3_eq(mat3 m1, mat3 m2);
+
+void
+test_assert_vec3_eq(vec3 v1, vec3 v2);
+
+void
+test_assert_vec4_eq(vec4 v1, vec4 v2);
+
+void
+test_assert_quat_eq(versor v1, versor v2);
+
+void
+test_assert_quat_eq_abs(versor v1, versor v2);
+
+void
+test_rand_vec3(vec3 dest);
+
+void
+test_rand_vec4(vec4 dest) ;
+
+float
+test_rand_angle(void);
+
+void
+test_rand_quat(versor q);
+
 #endif /* test_common_h */

test/src/test_euler.c

@@ -0,0 +1,44 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+void
+test_euler(void **state) {
+  mat4  rot1, rot2;
+  vec3  inAngles, outAngles;
+
+  inAngles[0] = glm_rad(-45.0f);  /* X angle */
+  inAngles[1] = glm_rad(88.0f);   /* Y angle */
+  inAngles[2] = glm_rad(18.0f);   /* Z angle */
+
+  glm_euler_xyz(inAngles, rot1);
+
+  /* extract angles */
+  glmc_euler_angles(rot1, outAngles);
+
+  /* angles must be equal in that range */
+  test_assert_vec3_eq(inAngles, outAngles);
+
+  /* matrices must be equal */
+  glmc_euler_xyz(outAngles, rot2);
+  test_assert_mat4_eq(rot1, rot2);
+
+  /* change range */
+  inAngles[0] = glm_rad(-145.0f);  /* X angle */
+  inAngles[1] = glm_rad(818.0f);   /* Y angle */
+  inAngles[2] = glm_rad(181.0f);   /* Z angle */
+
+  glm_euler_xyz(inAngles, rot1);
+  glmc_euler_angles(rot1, outAngles);
+
+  /* angles may not be equal but matrices MUST! */
+
+  /* matrices must be equal */
+  glmc_euler_xyz(outAngles, rot2);
+  test_assert_mat4_eq(rot1, rot2);
+}

test/src/test_main.c

@@ -12,9 +12,33 @@ main(int argc, const char * argv[]) {
     /* mat4 */
     cmocka_unit_test(test_mat4),
 
+    /* mat3 */
+    cmocka_unit_test(test_mat3),
+
     /* camera */
     cmocka_unit_test(test_camera_lookat),
-    cmocka_unit_test(test_camera_decomp)
+    cmocka_unit_test(test_camera_decomp),
+
+    /* project */
+    cmocka_unit_test(test_project),
+
+    /* vector */
+    cmocka_unit_test(test_clamp),
+
+    /* euler */
+    cmocka_unit_test(test_euler),
+
+    /* quaternion */
+    cmocka_unit_test(test_quat),
+
+    /* vec4 */
+    cmocka_unit_test(test_vec4),
+
+    /* vec3 */
+    cmocka_unit_test(test_vec3),
+
+    /* affine */
+    cmocka_unit_test(test_affine)
   };
 
   return cmocka_run_group_tests(tests, NULL, NULL);

test/src/test_mat3.c

@@ -0,0 +1,58 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+#define m 3
+#define n 3
+
+void
+test_mat3(void **state) {
+  mat3  m1 = GLM_MAT3_IDENTITY_INIT;
+  mat3  m2 = GLM_MAT3_IDENTITY_INIT;
+  mat3  m3;
+  mat3  m4 = GLM_MAT3_ZERO_INIT;
+  mat3  m5;
+  int   i, j, k;
+
+  /* test identity matrix multiplication */
+  glmc_mat3_mul(m1, m2, m3);
+  for (i = 0; i < m; i++) {
+    for (j = 0; j < n; j++) {
+      if (i == j)
+        assert_true(m3[i][j] == 1.0f);
+      else
+        assert_true(m3[i][j] == 0.0f);
+    }
+  }
+
+  /* test random matrices */
+  /* random matrices */
+  test_rand_mat3(m1);
+  test_rand_mat3(m2);
+
+  glmc_mat3_mul(m1, m2, m3);
+  for (i = 0; i < m; i++) {
+    for (j = 0; j < n; j++) {
+      for (k = 0; k < m; k++)
+        /* column-major */
+        m4[i][j] += m1[k][j] * m2[i][k];
+    }
+  }
+
+  test_assert_mat3_eq(m3, m4);
+
+  for (i = 0; i < 100000; i++) {
+    test_rand_mat3(m3);
+    test_rand_mat3(m4);
+
+    /* test inverse precise */
+    glmc_mat3_inv(m3, m4);
+    glmc_mat3_inv(m4, m5);
+    test_assert_mat3_eq(m3, m5);
+  }
+}

test/src/test_project.c

@@ -0,0 +1,31 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+void
+test_project(void **state) {
+  mat4 model, view, proj, mvp;
+  vec4 viewport = {0.0f, 0.0f, 800.0f, 600.0f};
+  vec3 pos      = {13.0f, 45.0f, 0.74f};
+  vec3 projected, unprojected;
+
+  glm_translate_make(model, (vec3){0.0f, 0.0f, -10.0f});
+  glm_lookat((vec3){0.0f, 0.0f, 0.0f}, pos, GLM_YUP, view);
+
+  glm_perspective_default(0.5f, proj);
+  glm_mat4_mulN((mat4 *[]){&proj, &view, &model}, 3, mvp);
+
+  glmc_project(pos, mvp, viewport, projected);
+  glmc_unproject(projected, mvp, viewport, unprojected);
+
+  /* unprojected of projected vector must be same as original one */
+  /* we used 0.01 because of projection floating point errors */
+  assert_true(fabsf(pos[0] - unprojected[0]) < 0.01);
+  assert_true(fabsf(pos[1] - unprojected[1]) < 0.01);
+  assert_true(fabsf(pos[2] - unprojected[2]) < 0.01);
+}

test/src/test_quat.c

@@ -0,0 +1,199 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+CGLM_INLINE
+void
+test_quat_mul_raw(versor p, versor q, versor dest) {
+  dest[0] = p[3] * q[0] + p[0] * q[3] + p[1] * q[2] - p[2] * q[1];
+  dest[1] = p[3] * q[1] - p[0] * q[2] + p[1] * q[3] + p[2] * q[0];
+  dest[2] = p[3] * q[2] + p[0] * q[1] - p[1] * q[0] + p[2] * q[3];
+  dest[3] = p[3] * q[3] - p[0] * q[0] - p[1] * q[1] - p[2] * q[2];
+}
+
+void
+test_quat(void **state) {
+  mat4   inRot, outRot, view1, view2, rot1, rot2;
+  versor inQuat, outQuat, q3, q4, q5;
+  vec3   eye, axis, imag, v1, v2;
+  int    i;
+
+  /* 0. test identiy quat */
+  glm_quat_identity(q4);
+  assert_true(glm_quat_real(q4) == cosf(glm_rad(0.0f) * 0.5f));
+  glm_quat_mat4(q4, rot1);
+  test_assert_mat4_eq2(rot1, GLM_MAT4_IDENTITY, 0.000009);
+
+  /* 1. test quat to mat and mat to quat */
+  for (i = 0; i < 1000; i++) {
+    test_rand_quat(inQuat);
+
+    glmc_quat_mat4(inQuat, inRot);
+    glmc_mat4_quat(inRot, outQuat);
+    glmc_quat_mat4(outQuat, outRot);
+
+    /* 2. test first quat and generated one equality */
+    test_assert_quat_eq_abs(inQuat, outQuat);
+
+    /* 3. test first rot and second rotation */
+    test_assert_mat4_eq2(inRot, outRot, 0.000009); /* almost equal */
+
+    /* 4. test SSE mul and raw mul */
+    test_quat_mul_raw(inQuat, outQuat, q3);
+    glm_quat_mul_sse2(inQuat, outQuat, q4);
+    test_assert_quat_eq(q3, q4);
+  }
+
+  /* 5. test lookat */
+  test_rand_vec3(eye);
+  glm_quatv(q3, glm_rad(-90.0f), GLM_YUP);
+
+  /* now X axis must be forward axis, Z must be right axis */
+  glm_look(eye, GLM_XUP, GLM_YUP, view1);
+
+  /* create view matrix with quaternion */
+  glm_quat_look(eye, q3, view2);
+
+  test_assert_mat4_eq2(view1, view2, 0.000009);
+
+  /* 6. test quaternion rotation matrix result */
+  test_rand_quat(q3);
+  glm_quat_mat4(q3, rot1);
+
+  /* 6.1 test axis and angle of quat */
+  glm_quat_axis(q3, axis);
+  glm_rotate_make(rot2, glm_quat_angle(q3), axis);
+
+  test_assert_mat4_eq2(rot1, rot2, 0.000009);
+
+  /* 7. test quaternion multiplication (hamilton product),
+        final rotation = first rotation + second = quat1 * quat2
+   */
+  test_rand_quat(q3);
+  test_rand_quat(q4);
+
+  glm_quat_mul(q3, q4, q5);
+
+  glm_quat_axis(q3, axis);
+  glm_rotate_make(rot1, glm_quat_angle(q3), axis);
+
+  glm_quat_axis(q4, axis);
+  glm_rotate(rot1, glm_quat_angle(q4), axis);
+
+  /* rot2 is combine of two rotation now test with quaternion result */
+  glm_quat_mat4(q5, rot2);
+
+  /* result must be same (almost) */
+  test_assert_mat4_eq2(rot1, rot2, 0.000009);
+
+  /* 8. test quaternion for look rotation */
+
+  /* 8.1 same direction */
+  /* look at from 0, 0, 1 to zero, direction = 0, 0, -1 */
+  glm_quat_for((vec3){0, 0, -1}, (vec3){0, 0, -1}, GLM_YUP, q3);
+
+  /* result must be identity */
+  glm_quat_identity(q4);
+  test_assert_quat_eq(q3, q4);
+
+  /* look at from 0, 0, 1 to zero, direction = 0, 0, -1 */
+  glm_quat_forp(GLM_ZUP, GLM_VEC3_ZERO, (vec3){0, 0, -1}, GLM_YUP, q3);
+
+  /* result must be identity */
+  glm_quat_identity(q4);
+  test_assert_quat_eq(q3, q4);
+
+  /* 8.2 perpendicular */
+  glm_quat_for(GLM_XUP, (vec3){0, 0, -1}, GLM_YUP, q3);
+
+  /* result must be -90 */
+  glm_quatv(q4, glm_rad(-90.0f), GLM_YUP);
+  test_assert_quat_eq(q3, q4);
+
+  /* 9. test imag, real */
+
+  /* 9.1 real */
+  assert_true(glm_quat_real(q4) == cosf(glm_rad(-90.0f) * 0.5f));
+
+  /* 9.1 imag */
+  glm_quat_imag(q4, imag);
+
+  /* axis = Y_UP * sinf(angle * 0.5), YUP = 0, 1, 0 */
+  axis[0] = 0.0f;
+  axis[1] = sinf(glm_rad(-90.0f) * 0.5f) * 1.0f;
+  axis[2] = 0.0f;
+
+  assert_true(glm_vec_eqv_eps(imag, axis));
+
+  /* 9.2 axis */
+  glm_quat_axis(q4, axis);
+  imag[0] =  0.0f;
+  imag[1] = -1.0f;
+  imag[2] =  0.0f;
+
+  test_assert_vec3_eq(imag, axis);
+
+  /* 10. test rotate vector using quat */
+  /* (0,0,-1) around (1,0,0) must give (0,1,0) */
+  v1[0] = 0.0f; v1[1] = 0.0f; v1[2] = -1.0f;
+  v2[0] = 0.0f; v2[1] = 0.0f; v2[2] = -1.0f;
+
+  glm_vec_rotate(v1, glm_rad(90.0f), (vec3){1.0f, 0.0f, 0.0f});
+  glm_quatv(q3, glm_rad(90.0f), (vec3){1.0f, 0.0f, 0.0f});
+
+  glm_vec4_scale(q3, 1.5, q3);
+  glm_quat_rotatev(q3, v2, v2);
+
+  /* result must be : (0,1,0) */
+  assert_true(fabsf(v1[0]) <= 0.00009f
+              && fabsf(v1[1] - 1.0f) <= 0.00009f
+              && fabsf(v1[2]) <= 0.00009f);
+
+  test_assert_vec3_eq(v1, v2);
+
+  /* 11. test rotate transform */
+  glm_translate_make(rot1, (vec3){-10.0, 45.0f, 8.0f});
+  glm_rotate(rot1, glm_rad(-90), GLM_ZUP);
+
+  glm_quatv(q3, glm_rad(-90.0f), GLM_ZUP);
+  glm_translate_make(rot2, (vec3){-10.0, 45.0f, 8.0f});
+  glm_quat_rotate(rot2, q3, rot2);
+
+  /* result must be same (almost) */
+  test_assert_mat4_eq2(rot1, rot2, 0.000009);
+
+  glm_rotate_make(rot1, glm_rad(-90), GLM_ZUP);
+  glm_translate(rot1, (vec3){-10.0, 45.0f, 8.0f});
+
+  glm_quatv(q3, glm_rad(-90.0f), GLM_ZUP);
+  glm_mat4_identity(rot2);
+  glm_quat_rotate(rot2, q3, rot2);
+  glm_translate(rot2, (vec3){-10.0, 45.0f, 8.0f});
+
+  /* result must be same (almost) */
+  test_assert_mat4_eq2(rot1, rot2, 0.000009);
+
+  /* reverse */
+  glm_rotate_make(rot1, glm_rad(-90), GLM_ZUP);
+  glm_quatv(q3, glm_rad(90.0f), GLM_ZUP);
+  glm_quat_rotate(rot1, q3, rot1);
+
+  /* result must be identity */
+  test_assert_mat4_eq2(rot1, GLM_MAT4_IDENTITY, 0.000009);
+
+  test_rand_quat(q3);
+
+  /* 12. inverse of quat, multiplication must be IDENTITY */
+  glm_quat_inv(q3, q4);
+  glm_quat_mul(q3, q4, q5);
+
+  glm_quat_identity(q3);
+  test_assert_quat_eq(q3, q5);
+
+  /* TODO: add tests for slerp, lerp */
+}

test/src/test_tests.h

@@ -9,6 +9,9 @@
 /* mat4 */
 void test_mat4(void **state);
 
+/* mat3 */
+void test_mat3(void **state);
+
 /* camera */
 void
 test_camera_lookat(void **state);
@@ -16,4 +19,25 @@ test_camera_lookat(void **state);
 void
 test_camera_decomp(void **state);
 
+void
+test_project(void **state);
+
+void
+test_clamp(void **state);
+
+void
+test_euler(void **state);
+
+void
+test_quat(void **state);
+
+void
+test_vec4(void **state);
+
+void
+test_vec3(void **state);
+
+void
+test_affine(void **state);
+
 #endif /* test_tests_h */

test/src/test_vec3.c

@@ -0,0 +1,78 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+void
+test_vec3(void **state) {
+  mat3 rot1m3;
+  mat4 rot1;
+  vec3 v, v1, v2;
+
+  /* test zero */
+  glm_vec_zero(v);
+  test_assert_vec3_eq(GLM_VEC3_ZERO, v);
+
+  /* test one */
+  glm_vec_one(v);
+  test_assert_vec3_eq(GLM_VEC3_ONE, v);
+
+  /* adds, subs, div, divs, mul */
+  glm_vec_add(v, GLM_VEC3_ONE, v);
+  assert_true(glmc_vec_eq_eps(v, 2));
+
+  glm_vec_adds(v, 10, v);
+  assert_true(glmc_vec_eq_eps(v, 12));
+
+  glm_vec_sub(v, GLM_VEC3_ONE, v);
+  assert_true(glmc_vec_eq_eps(v, 11));
+
+  glm_vec_subs(v, 1, v);
+  assert_true(glmc_vec_eq_eps(v, 10));
+
+  glm_vec_broadcast(2, v1);
+  glm_vec_div(v, v1, v);
+  assert_true(glmc_vec_eq_eps(v, 5));
+
+  glm_vec_divs(v, 0.5, v);
+  assert_true(glmc_vec_eq_eps(v, 10));
+
+  glm_vec_mul(v, v1, v);
+  assert_true(glmc_vec_eq_eps(v, 20));
+
+  glm_vec_scale(v, 0.5, v);
+  assert_true(glmc_vec_eq_eps(v, 10));
+
+  glm_vec_normalize_to(v, v1);
+  glm_vec_scale(v1, 0.8, v1);
+  glm_vec_scale_as(v, 0.8, v);
+  test_assert_vec3_eq(v1, v);
+
+  /* addadd, subadd, muladd */
+  glm_vec_one(v);
+
+  glm_vec_addadd(GLM_VEC3_ONE, GLM_VEC3_ONE, v);
+  assert_true(glmc_vec_eq_eps(v, 3));
+
+  glm_vec_subadd(GLM_VEC3_ONE, GLM_VEC3_ZERO, v);
+  assert_true(glmc_vec_eq_eps(v, 4));
+
+  glm_vec_broadcast(2, v1);
+  glm_vec_broadcast(3, v2);
+  glm_vec_muladd(v1, v2, v);
+  assert_true(glmc_vec_eq_eps(v, 10));
+
+  /* rotate */
+  glm_vec_copy(GLM_YUP, v);
+  glm_rotate_make(rot1, glm_rad(90), GLM_XUP);
+  glm_vec_rotate_m4(rot1, v, v1);
+  glm_mat4_pick3(rot1, rot1m3);
+  glm_vec_rotate_m3(rot1m3, v, v2);
+
+  test_assert_vec3_eq(v1, v2);
+  test_assert_vec3_eq(v1, GLM_ZUP);
+}

test/src/test_vec4.c

@@ -0,0 +1,178 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+CGLM_INLINE
+float
+test_vec4_dot(vec4 a, vec4 b) {
+  return a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + a[3] * b[3];
+}
+
+CGLM_INLINE
+void
+test_vec4_normalize_to(vec4 vec, vec4 dest) {
+  float norm;
+
+  norm = glm_vec4_norm(vec);
+
+  if (norm == 0.0f) {
+    dest[0] = dest[1] = dest[2] = dest[3] = 0.0f;
+    return;
+  }
+
+  glm_vec4_scale(vec, 1.0f / norm, dest);
+}
+
+float
+test_vec4_norm2(vec4 vec) {
+  return test_vec4_dot(vec, vec);
+}
+
+float
+test_vec4_norm(vec4 vec) {
+  return sqrtf(test_vec4_dot(vec, vec));
+}
+
+void
+test_vec4_maxv(vec4 v1, vec4 v2, vec4 dest) {
+  dest[0] = glm_max(v1[0], v2[0]);
+  dest[1] = glm_max(v1[1], v2[1]);
+  dest[2] = glm_max(v1[2], v2[2]);
+  dest[3] = glm_max(v1[3], v2[3]);
+}
+
+void
+test_vec4_minv(vec4 v1, vec4 v2, vec4 dest) {
+  dest[0] = glm_min(v1[0], v2[0]);
+  dest[1] = glm_min(v1[1], v2[1]);
+  dest[2] = glm_min(v1[2], v2[2]);
+  dest[3] = glm_min(v1[3], v2[3]);
+}
+
+void
+test_vec4_clamp(vec4 v, float minVal, float maxVal) {
+  v[0] = glm_clamp(v[0], minVal, maxVal);
+  v[1] = glm_clamp(v[1], minVal, maxVal);
+  v[2] = glm_clamp(v[2], minVal, maxVal);
+  v[3] = glm_clamp(v[3], minVal, maxVal);
+}
+
+void
+test_vec4(void **state) {
+  vec4  v, v1, v2, v3, v4;
+  int   i;
+  float d1, d2;
+
+
+  for (i = 0; i < 1000; i++) {
+    /* 1. test SSE/SIMD dot product */
+    test_rand_vec4(v);
+    d1 = glm_vec4_dot(v, v);
+    d2 = test_vec4_dot(v, v);
+
+    assert_true(fabsf(d1 - d2) <= 0.000009);
+
+    /* 2. test SIMD normalize */
+    test_vec4_normalize_to(v, v1);
+    glm_vec4_normalize_to(v, v2);
+    glm_vec4_normalize(v);
+
+    /* all must be same */
+    test_assert_vec4_eq(v1, v2);
+    test_assert_vec4_eq(v, v2);
+
+    /* 3. test SIMD norm */
+    test_rand_vec4(v);
+    test_assert_eqf(test_vec4_norm(v), glm_vec4_norm(v));
+
+    /* 3. test SIMD norm2 */
+    test_rand_vec4(v);
+    test_assert_eqf(test_vec4_norm2(v), glm_vec4_norm2(v));
+  }
+
+  /* test zero */
+  glm_vec4_zero(v);
+  test_assert_vec4_eq(GLM_VEC4_ZERO, v);
+
+  /* test one */
+  glm_vec4_one(v);
+  test_assert_vec4_eq(GLM_VEC4_ONE, v);
+
+  /* adds, subs, div, divs, mul */
+  glm_vec4_add(v, GLM_VEC4_ONE, v);
+  assert_true(glmc_vec4_eq_eps(v, 2));
+
+  glm_vec4_adds(v, 10, v);
+  assert_true(glmc_vec4_eq_eps(v, 12));
+
+  glm_vec4_sub(v, GLM_VEC4_ONE, v);
+  assert_true(glmc_vec4_eq_eps(v, 11));
+
+  glm_vec4_subs(v, 1, v);
+  assert_true(glmc_vec4_eq_eps(v, 10));
+
+  glm_vec4_broadcast(2, v1);
+  glm_vec4_div(v, v1, v);
+  assert_true(glmc_vec4_eq_eps(v, 5));
+
+  glm_vec4_divs(v, 0.5, v);
+  assert_true(glmc_vec4_eq_eps(v, 10));
+
+  glm_vec4_mul(v, v1, v);
+  assert_true(glmc_vec4_eq_eps(v, 20));
+
+  glm_vec4_scale(v, 0.5, v);
+  assert_true(glmc_vec4_eq_eps(v, 10));
+
+  glm_vec4_normalize_to(v, v1);
+  glm_vec4_scale(v1, 0.8, v1);
+  glm_vec4_scale_as(v, 0.8, v);
+  test_assert_vec4_eq(v1, v);
+
+  /* addadd, subadd, muladd */
+  glm_vec4_one(v);
+
+  glm_vec4_addadd(GLM_VEC4_ONE, GLM_VEC4_ONE, v);
+  assert_true(glmc_vec4_eq_eps(v, 3));
+
+  glm_vec4_subadd(GLM_VEC4_ONE, GLM_VEC4_ZERO, v);
+  assert_true(glmc_vec4_eq_eps(v, 4));
+
+  glm_vec4_broadcast(2, v1);
+  glm_vec4_broadcast(3, v2);
+  glm_vec4_muladd(v1, v2, v);
+  assert_true(glmc_vec4_eq_eps(v, 10));
+
+  /* min, max */
+  test_rand_vec4(v1);
+  test_rand_vec4(v2);
+
+  glm_vec4_maxv(v1, v2, v3);
+  test_vec4_maxv(v1, v2, v4);
+  test_assert_vec4_eq(v3, v4);
+
+  glm_vec4_minv(v1, v2, v3);
+  test_vec4_minv(v1, v2, v4);
+  test_assert_vec4_eq(v3, v4);
+
+  glm_vec4_print(v3, stderr);
+  glm_vec4_print(v4, stderr);
+
+  /* clamp */
+  glm_vec4_clamp(v3, 0.1, 0.8);
+  test_vec4_clamp(v4, 0.1, 0.8);
+  test_assert_vec4_eq(v3, v4);
+
+  glm_vec4_print(v3, stderr);
+  glm_vec4_print(v4, stderr);
+
+  assert_true(v3[0] >= 0.0999 && v3[0] <= 0.80001); /* rounding erros */
+  assert_true(v3[1] >= 0.0999 && v3[1] <= 0.80001);
+  assert_true(v3[2] >= 0.0999 && v3[2] <= 0.80001);
+  assert_true(v3[3] >= 0.0999 && v3[3] <= 0.80001);
+}

win/cglm.vcxproj

@@ -29,7 +29,9 @@
     <ClCompile Include="..\src\mat3.c" />
     <ClCompile Include="..\src\mat4.c" />
     <ClCompile Include="..\src\plane.c" />
+    <ClCompile Include="..\src\project.c" />
     <ClCompile Include="..\src\quat.c" />
+    <ClCompile Include="..\src\sphere.c" />
     <ClCompile Include="..\src\vec3.c" />
     <ClCompile Include="..\src\vec4.c" />
   </ItemGroup>
@@ -47,7 +49,9 @@
     <ClInclude Include="..\include\cglm\call\mat3.h" />
     <ClInclude Include="..\include\cglm\call\mat4.h" />
     <ClInclude Include="..\include\cglm\call\plane.h" />
+    <ClInclude Include="..\include\cglm\call\project.h" />
     <ClInclude Include="..\include\cglm\call\quat.h" />
+    <ClInclude Include="..\include\cglm\call\sphere.h" />
     <ClInclude Include="..\include\cglm\call\vec3.h" />
     <ClInclude Include="..\include\cglm\call\vec4.h" />
     <ClInclude Include="..\include\cglm\cam.h" />
@@ -60,6 +64,7 @@
     <ClInclude Include="..\include\cglm\mat3.h" />
     <ClInclude Include="..\include\cglm\mat4.h" />
     <ClInclude Include="..\include\cglm\plane.h" />
+    <ClInclude Include="..\include\cglm\project.h" />
     <ClInclude Include="..\include\cglm\quat.h" />
     <ClInclude Include="..\include\cglm\simd\avx\affine.h" />
     <ClInclude Include="..\include\cglm\simd\avx\mat4.h" />
@@ -69,6 +74,7 @@
     <ClInclude Include="..\include\cglm\simd\sse2\mat3.h" />
     <ClInclude Include="..\include\cglm\simd\sse2\mat4.h" />
     <ClInclude Include="..\include\cglm\simd\sse2\quat.h" />
+    <ClInclude Include="..\include\cglm\sphere.h" />
     <ClInclude Include="..\include\cglm\types.h" />
     <ClInclude Include="..\include\cglm\util.h" />
     <ClInclude Include="..\include\cglm\vec3-ext.h" />

win/cglm.vcxproj.filters

@@ -75,6 +75,12 @@
     <ClCompile Include="..\src\box.c">
       <Filter>src</Filter>
     </ClCompile>
+    <ClCompile Include="..\src\project.c">
+      <Filter>src</Filter>
+    </ClCompile>
+    <ClCompile Include="..\src\sphere.c">
+      <Filter>src</Filter>
+    </ClCompile>
   </ItemGroup>
   <ItemGroup>
     <ClInclude Include="..\src\config.h">
@@ -206,5 +212,17 @@
     <ClInclude Include="..\include\cglm\color.h">
       <Filter>include\cglm</Filter>
     </ClInclude>
+    <ClInclude Include="..\include\cglm\project.h">
+      <Filter>include\cglm</Filter>
+    </ClInclude>
+    <ClInclude Include="..\include\cglm\call\project.h">
+      <Filter>include\cglm\call</Filter>
+    </ClInclude>
+    <ClInclude Include="..\include\cglm\call\sphere.h">
+      <Filter>include\cglm\call</Filter>
+    </ClInclude>
+    <ClInclude Include="..\include\cglm\sphere.h">
+      <Filter>include\cglm</Filter>
+    </ClInclude>
   </ItemGroup>
 </Project>