chickadee » gl-utils


Provides a higher level interface to OpenGL. The following modules are included:

gl-utils is known to work on Linux, Mac OS X, Windows, and with OpenGL ES. gl-utils will automatically compile with ES support on ARM hardware, or when gles is defined during compilation (e.g. chicken-install -D gles).





Performs get-error (glGetError) and prints the error type when an error is returned.

make-shader TYPE SOURCEprocedure

Creates and compiles a shader object given the shader's type (e.g. +vertex-shader+, +geometry-shader+, +fragment-shader+), and a string containing the GLSL source. Returns an integer representing the ID of the shader.

make-program SHADER-LIST #!optional PROGRAM-IDprocedure

Creates and links a program object, given a list of shader objects (i.e. the integers returned by make-shader. Returns an integer representing the ID of the program.

Accepts an optional PROGRAM-ID argument. If given, make-program will use this ID rather than generating a new one.


Analogous to their pluralized counterparts, but only generates and returns one (integer) object.

delete-buffer BUFFERprocedure
delete-framebuffer FRAMEBUFFERprocedure
delete-program-pipeline PROGRAM-PIPELINEprocedure
delete-query QUERYprocedure
delete-renderbuffer RENDERBUFFERprocedure
delete-sampler SAMPLERprocedure
delete-texture TEXTUREprocedure
delete-transform-feedback TRANSFORM-FEEDBACKprocedure
delete-vertex-array VERTEX-ARRAYprocedure

Analogous to their pluralized counterparts, but only accepts and deletes one (integer) object.

(with-buffer TYPE BUFFER BODY ...)syntax
(with-framebuffer FRAMEBUFFER BODY ...)syntax
(with-program-pipeline PROGRAM-PIPELINE BODY ...)syntax
(with-renderbuffer RENDERBUFFER BODY ...)syntax
(with-sampler UNIT SAMPLER BODY ...)syntax
(with-texture TYPE TEXTURE BODY ...)syntax
(with-transform-feedback TRANSFORM-FEEDBACK BODY ...)syntax
(with-vertex-array VERTEX-ARRAY BODY ...)syntax

Equivalent to binding the object, executing the body, then binding 0. E.g. (bind-texture TYPE TEXTURE) BODY ... (bind-texture TYPE 0).

(set-texture-properties ID type: TYPE mag: MAG min: MIN wrap: WRAP wrap-s: WRAP-S wrap-t: WRAP-T wrap-r: WRAP-R)procedure

Conveniently set the most common properties of the texture ID. TYPE is the texture type, defaulting to +texture-2d+. MAG and MIN are the texture magnify and minifying functions, which default to +linear+. WRAP-S, WRAP-T, and WRAP-R set the wrapping parameters which default to +repeat+. WRAP sets all three wrapping parameters to the same value.

If ID is #f, no texture will be bound, and therefore properties of whatever texture is currently bound will be set.

(create-framebuffer WIDTH HEIGHT channels: CHANNELS type: TYPE)procedure

Create a framebuffer with a texture and depth renderbuffer attached. The texture and renderbuffer are given the dimensions WITH and HEIGHT. CHANNELS is the number of channels that the texture has: 1, 2, 3, or 4, corresponding to +red+, +rg+, +rgb+, and+rgba+ respectively, defaulting to 4. TYPE is the type of the texture data which defaults to +unsigned-byte+. Returns three values: The framebuffer, the texture, and the renderbuffer.

->pointer VECTORprocedure

Returns the pointer to a srfi-4 vector or blob.

size VECTORprocedure

Returns the size, in bytes, of a srfi-4 vector or blob.

type->bytes TYPEprocedure

Returns the size of TYPE (as accepted by type->gl) in number of bytes.

type->gl TYPEprocedure

Converts the keyword TYPE into a OpenGL type enum value. Accepted types (grouped by synonyms) are:

  • char: int8: byte:
  • uchar: uint8: unsigned-byte:
  • short: int16:
  • ushort: uint16: unsigned-short:
  • int: int32: integer: integer32:
  • uint: uint32: unsigned-int: unsigned-int32: unsigned-integer: unsigned-integer32:
  • float: float32:
  • double: float64:

Double is not, however, a valid type when using GL ES.


r7rs style bytevectors with unsafe accessors.

bytevector BYTE ...procedure

Returns a newly allocated bytevector containing its arguments. The resulting bytevector will be created in garbage-collected memory.

(make-bytevector K [BYTE] [NONGC] [FINALIZE])procedure

Return a newly-allocated bytevector of length K. If the optional fill BYTE is specified, it specifies the initial value for each slot in the bytevector.

The optional arguments NONGC and FINALIZE define whether the vector should be allocated in a memory area not subject to garbage collection and whether the associated storage should be automatically freed (using finalization) when there are no references from Scheme variables and data. NONGC defaults to #t (the vector will be located in non-garbage-collected memory) and FINALIZE defaults to #t. Note that the FINALIZE argument is only used when NONGC is true.

bytevector-length BYTEVECTORprocedure

Return the length in bytes of BYTEVECTOR.

bytevector? BYTEVECTORprocedure

Returns true if BYTEVECTOR is a bytevector, false otherwise.

bytevector-u8-set! BYTEVECTOR K UNSIGNED-BYTEprocedure
bytevector-s8-set! BYTEVECTOR K BYTEprocedure
bytevector-u16-set! BYTEVECTOR K UNSIGNED-SHORTprocedure
bytevector-s16-set! BYTEVECTOR K SHORTprocedure
bytevector-u32-set! BYTEVECTOR K UNSIGNED-INTprocedure
bytevector-s32-set! BYTEVECTOR K INTprocedure
bytevector-u64-set! BYTEVECTOR K UNSIGNED-LONGprocedure
bytevector-s64-set! BYTEVECTOR K LONGprocedure
bytevector-f32-set! BYTEVECTOR K FLOATprocedure
bytevector-f64-set! BYTEVECTOR K DOUBLEprocedure

Sets the byte K of the given bytevector to be the value of the given fixnum or flonum. These functions are unsafe, so be sure K is a valid location in the bytevector.

bytevector-u8-ref BYTEVECTOR Kprocedure
bytevector-s8-ref BYTEVECTOR Kprocedure
bytevector-u16-ref BYTEVECTOR Kprocedure
bytevector-s16-ref BYTEVECTOR Kprocedure
bytevector-u32-ref BYTEVECTOR Kprocedure
bytevector-s32-ref BYTEVECTOR Kprocedure
bytevector-u64-ref BYTEVECTOR Kprocedure
bytevector-s64-ref BYTEVECTOR Kprocedure
bytevector-f32-ref BYTEVECTOR Kprocedure
bytevector-f64-ref BYTEVECTOR Kprocedure

Returns the fixnum or flonum of the given size located at byte K of the given bytevector. These functions are unsafe, so be sure K is a valid location in the bytevector.

bytevector–>pointer BYTEVECTORprocedure

Returns the pointer to BYTEVECTOR’s data.usage

bytevector-append BYTEVECTOR #!rest BYTEVECTORSSprocedure

Returns a newly allocated bytevector whose elements are the concatenation of the elements in the given bytevectors. If only one argument is passed to bytevector-append, it is assumed that it is a list of bytevectors.

(bytevector-copy BYTEVECTOR [START] [END])procedure

Returns a newly allocated copy of the elements of the given bytevector between START and END.

(bytevector-copy! TO AT FROM [START] [END])procedure

Copies the elements of bytevector FROM between START and END to bytevector TO, starting at AT. It is an error if AT is less than zero or greater than the length of TO. It is also an error if (- (bytevector-length TO) AT) is less than (- END START).


(make-mesh vertices: VERTICES [indices: INDICES] [mode: MODE])procedure

Create a new mesh. VERTICES is a list of key value pairs that specifies the mesh’s vertex data. It should be in the form:

   (attributes: ATTRIBUTES [initial-elements: INITIAL-ELEMENTS] [n-vertices: N-VERTICES])

ATTRIBUTES is a list in the form:

   (NAME TYPE N [normalized: NORMALIZED])

where NAME is the attribute name (as a symbol), TYPE is the type of the attribute as accepted by type->gl, N is the number of elements in the attribute, NORMALIZED is a boolean value indicating whether the attribute’s values should be normalized (defaulting to #f).

INITIAL-ELEMENTS is either a bytevector or a list of (NAME . VALUE) pairs where name is the name of the attribute to set (as per the name given in ATTRIBUTES) and VALUE is the initial contents of that attribute. When a list is given and more than one attribute is given initial-elements, the VALUEs should represent the same number of vertices. Values associated with attributes that are NORMALIZED should be provided as float between 0.0 and 1.0 (for unsigned types) or -1.0 and 1.0, which are then normalized. If INITIAL-ELEMENTS is given as a bytevector, that bytevector is used as the entire mesh’s vertex data and N-VERTICES – the number of vertices – must be provided, otherwise N-VERTICES is ignored.

INDICES is an optional list of key value pairs that specifies the mesh’s index data, if it exists. It should be in the form:

   (type: TYPE [initial-elements: INITIAL-ELEMENTS] [n-indices: N-INDICES])

TYPE is a type keyword (as accepted by type->gl) that must be a valid type for an element array buffer (i.e. an unsigned fixnum). INITIAL-ELEMENTS is either a bytevector or a list of values. If INITIAL-ELEMENTS is given as a bytevector, N-INDICES – the number of indices – must be provided, otherwise N-INDICES is ignored. When no INDICES list is provided, the number of indices (as returned by mesh-n-indices) is set to the number of vertices.

MODE is the keyword (as accepted by mode->gl) that defines what the mesh is supposed to represent. Defaults to #:triangles.


The type of record returned by make-mesh. VERTEX-ATTRIBUTES is a list of vertex-attribute records. INDEX-TYPE is the type given to the indices: argument of make-mesh. VERTEX-DATA and INDEX-DATA are the bytevectors representing the vertex and index data. N-VERTICES and N-INDICES are the number of vertices and indices present in the data. VERTEX-BUFFER, INDEX-BUFFER and VAO are the vertex buffers and VAO created by mesh-make-vao!. STRIDE is the number of bytes between the start of consecutive vertices. MODE is the value of the mode: argument provided to make-mesh. USAGE is the buffer usage that is set with make-mesh-vao!.

vertex-attribute name type number normalizedrecord

The type of record returned by mesh-vertex-attributes. Getters for all of the fields are provided.

mesh-make-vao! MESH LOCATIONS #!optional USAGEprocedure

Create a vertex attribute object (VAO) for MESH. LOCATIONS is a list of (ATTRIBUTE-NAME . LOCATION) pairs. USAGE is the buffer usage hint keyword as accepted by usage->gl, defaulting to #:static. Vertex buffer objects (VBOs) are created for the vertex and index data. The VAO binds these buffers, and sets the vertex attribute pointers of attributes for which locations have been given. If the usage is one of the static types, the vertex and index data of the mesh are deleted, as are the vertex and index buffers. The VBOs and VAO created by make-mesh-vao! are managed and should not be deleted.

mesh-update! MESH VERTICES #!optional INDICESprocedure

Update the entirety of the vertex data and optionally the index data of the MESH. This can be used in order to reuse the memory and vertex buffers of a mesh. VERTICES should be the same sort of list of (NAME . VALUE) attribute pairs that is accepted as the initial-elements: argument to make-mesh’s vertices: (though not a bytevector). The data in VERTICES should match the attributes of the original vector. Similarly, INDICES should be the same kind of list of elements that is accepted to make-mesh’s indices’s initial-elements:. The data in both VERTICES and INDICES must fit within the vertex and index bytevectors of the MESH. This can be called before or after mesh-make-vao!. Do not call inside with-mesh.

mesh-vertex-ref MESH ATTRIBUTE VERTEXprocedure

Return a vector containing the values of the attribute named ATTRIBUTE corresponding to the VERTEXth vertex of the MESH. The result will be a srfi-4 numeric vector, corresponding to the type of the given attribute.

mesh-vertex-set! MESH ATTRIBUTE VERTEX VALUEprocedure

Set the VERTEXth vertex of the MESH’s attribute named ATTRIBUTE to the values provided by the srfi-4 numeric vector VALUE. VALUE must correspond to the type of the attribute, and should be the same length as the attribute. Using a VALUE that is too short is unsafe.

If mesh-make-vao! has been called already, mesh-vertex-set! should be called inside a with-mesh. If mesh-make-vao! was called with a static usage, mesh-vertex-set! will result in an error, since there is no longer any vertex data to set.

with-mesh MESH THUNKprocedure

Calls THUNK with the VBO of the MESH’s vertex buffer bound. If the mesh has been modified by mesh-vertex-set!, the vertex buffer’s data will be updated before the VBO is unbound. When the usage hint given to mesh-make-vao! is dynamic, buffer-sub-data will be used to update the buffer data. When the usage hint is stream, buffer-data will be used.

(mesh-copy! TO AT FROM [START] [END])procedure

Similar to bytevector-copy!, copies the vertices of mesh FROM between vertices START and END to mesh TO, starting at vertex AT.

mesh-copy MESHprocedure

Creates a fresh copy of MESH.

mesh-append MESHESprocedure

Creates a new mesh resulting from appending the vertices of the given meshes together. The indices are also appended and modified so they point to the same vertices. The attributes of all the meshes are assumed to be the same, otherwise bad things will probably happen.

(mesh-transform! MESH TRANSFORM [start: START] [end: END] [position-name: POSITION-NAME] [normal-name: NORMAL-NAME] [normal-transform: NORMAL-TRANSFORM])procedure

Destructively modifies the POSITION-NAME attribute of MESH by the gl-math matrix TRANSFORM. POSITION-NAME defaults to 'position and must be the name of a three element float attribute of MESH. If NORMAL-NAME (defaulting to 'normal) is the name of an attribute in the given mesh, this attribute will be transformed by the inverse-transpose of the transform matrix. If NORMAL-TRANSFORM is present, it will be used instead of the TRANSFORM matrix for normal transformations.

(mesh-transform-append MESH-TRANSFORM-PAIRS [position-name: POSITION-NAME] [normal-name: NORMAL-NAME])procedure

Creates a new mesh resulting by appending all the given meshes together, then transforming the attribute named by POSITION-NAME by the given gl-math transform matrices. MESH-TRANSFORM-PAIRS are (MESH TRANSFORM [NORMAL-TRANSFORM]) tuples. POSITION-NAME defaults to 'position and must be the name of a three element float attribute of MESH. If NORMAL-NAME (defaulting to 'normal) is the name of an attribute in the given mesh, this attribute will be transformed by the inverse-transpose of the transform matrices. If NORMAL-TRANSFORM is present, it will be used instead of the TRANSFORM matrix for normal transformations. The attributes of all the meshes are assumed to be the same, otherwise bad things will probably happen.

usage->gl USAGEprocedure

Converts the keyword USAGE into a OpenGL usage enum value. Accepted usages (grouped by synonyms) are:

  • dynamic: dynamic-draw:
  • stream: stream-draw:
  • static: static-draw:
  • dynamic-read:
  • stream-read:
  • static-read:
  • dynamic-copy:
  • stream-copy:
  • static-copy:

This line is here to prevent a Markdown parsing error :|

mode->gl MODEprocedure

Converts the keyword MODE into a OpenGL mode enum value. Accepted modes are:

  • points:
  • lines:
  • line-strip:
  • line-loop:
  • line-strip-adjacency:**
  • lines-adjacency:**
  • triangle-strip:
  • triangle-fan:
  • triangles:
  • triangle-strip-adjacency:**
  • triangles-adjacency:**
  • patches:**
    • Not available in GL ES


load-ply FILE BUFFER-SPECprocedure

Loads a PLY file. FILE is a path that may be pointing either to a gziped PLY file or a regular PLY file. BUFFER-SPEC is a list in the form ((NAME VARS) ...) where NAME is the name of an element in the PLY file and VARS is either a list of property names or, in the case of a property list, a single name. Two values are returned: a list of bytevectors which correspond to the buffers named in BUFFER-SPEC and a list of the elements that are in the PLY file in the form of:

   (element-name n-elements (property-name property-type))

Or, when an element is a property list:

   (element-name n-elements (property-name (list: list-length-type element-type)))

The buffers returned are packed with the contents of the properties named in the BUFFER-SPEC. Thus, for a PLY file that has element vertex with properties float x, float y, float z, float confidence, uchar r, uchar g, and uchar b, as well as an element face with a property list uchar ushort vertex_index, the following BUFFER-SPEC could be used:

   (load-ply "example.ply.gz" '((vertex: (x y z r g b)) (face: vertex_index)))

This buffer spec would result in a list of two u8vectors being returned: one with the packed elements corresponding to properties x, y, z, r, g, and b (with the corresponding property types), and the second containing the vertex indices.

(load-ply-mesh FILE vertex: VERTEX face: FACE)procedure

Similar to load-ply, but returns a mesh. FILE is a PLY file (which may be gziped). The PLY file must contain at least the elements vertex and face (other elements will be ignored). VERTEX is a list of (ATTRIBUTE-NAME PROPERTY-NAME ...) elements, which specifies which PLY properties are associate with which attribute, in which order. Attributes are all normalized. All properties named by each element of VERTEX must be of the same type. FACE is the name of the face property list.

Again, for a PLY file that has element vertex with properties float x, float y, float z, float confidence, uchar r, uchar g, and uchar b, as well as a element face with a property list uchar ushort vertex_index, the following could be used:

   (load-ply-mesh "example.ply" vertex: `((position x y z) 
                                         (color r g b))
                               face: vertex_index)

This would create a mesh with vertex attributes (position #:float 3) and (color #:unsigned-byte 3 normalized: #t) and the #:unsigned-short indices given by vertex_index.


This module has been deprecated. With gl-utils-mesh and gl-utils-bytevector, this functionality is better served by directly using srfi-4. This module may still be accessed by explicitly using gl-utils-srfi-4 for the time-being, but it will be removed in the future.

gl-utils-srfi-4 reexports a version of srfi-4 that gives preference to vectors being created in non-garbage collected memory. This is useful for use with OpenGL, since it is often desirable to pass vectors to OpenGL that will remain in one place. All srfi-4 functions not mentioned below are reexported without changes.

The NNNvector and list->NNNvector constructors have been modified so that they return vectors in non-garbage collected memory. They will still be freed when no longer used.

The make-NNNvector constructors act as their srfi-4 counterparts, except they return vectors in non-garbage collected memory by default. They will still be freed when non longer used.


This example depends on the opengl-glew egg, the glfw3 egg for window and context creation, and the gl-math egg for matrix math.

For more examples, check out the examples directory.

(import chicken scheme)
(use (prefix glfw3 glfw:) (prefix opengl-glew gl:) gl-math gl-utils)

(define *vertex* 
#version 330
in vec2 position;
in vec3 color;
out vec3 c;
uniform mat4 MVP;

void main(){
   gl_Position = MVP * vec4(position, 0.0, 1.0);
   c = color;

(define *fragment*
#version 330
in vec3 c;
out vec4 fragColor;
void main(){
  fragColor = vec4(c, 1.0);

(define rect (make-mesh
              vertices: '(attributes: ((position #:float 2)
                                       (color #:unsigned-byte 3
                                              normalized: #t))
                          initial-elements: ((position . (-1 -1
                                                           1 -1
                                                           1  1
                                                           -1  1))
                                             (color . (255 0   0
                                                       0   255 0
                                                       0   0   255
                                                       255 0   255))))
              indices: '(type: #:ushort
                         initial-elements: (0 1 2
                                            0 2 3))))

(define program (make-parameter #f))

(define projection-matrix
  (perspective 640 480 0.1 100 70))

(define view-matrix
  (look-at (make-point 1 0 3)
           (make-point 0 0 0)
           (make-point 0 1 0)))

(define model-matrix (mat4-identity))

(define (render)
  (gl:use-program (program))
  (gl:uniform-matrix4fv (gl:get-uniform-location (program) "MVP")
                        1 #f
                        (m* projection-matrix
                            (m* view-matrix model-matrix)))
  (gl:bind-vertex-array (mesh-vao rect))
  (gl:draw-elements-base-vertex (mode->gl (mesh-mode rect))
                                (mesh-n-indices rect)
                                (type->gl (mesh-index-type rect))
                                #f 0)

  (gl:bind-vertex-array 0))

(glfw:with-window (640 480 "Example" resizable: #f
                       context-version-major: 3
                       context-version-minor: 3)

  (print (gl:supported? "GL_ARB_framebuffer_object"))

  (set! *vertex* (make-shader gl:+vertex-shader+ *vertex*))
  (set! *fragment* (make-shader gl:+fragment-shader+ *fragment*))
  (program (make-program (list *vertex* *fragment*)))

  (mesh-make-vao! rect `((position . ,(gl:get-attrib-location
                                       (program) "position"))
                         (color . ,(gl:get-attrib-location
                                    (program) "color"))))
  (let loop ()
    (glfw:swap-buffers (glfw:window))
    (gl:clear (bitwise-ior gl:+color-buffer-bit+ gl:+depth-buffer-bit+))
    (glfw:poll-events) ; Because of the context version, initializing GLEW results in a harmless invalid enum
    (unless (glfw:window-should-close (glfw:window))

Version history

Version 0.7.0

24 April 2016

Version 0.6.0

16 January 2014

Version 0.5.0

23 December 2014

Version 0.4.0

8 December 2014

Version 0.3.0

11 September 2014

Version 0.2.0

10 September 2014

Version 0.1.2

2 September 2014

Version 0.1.1

Source repository

Source available on GitHub.

Bug reports and patches welcome! Bugs can be reported via GitHub or to alex.n.charlton at gmail.


Alex Charlton



Contents »