GL/GLES/EGL/GLX/WGL Loader-Generator based on the official specs.
Use the webservice to generate the files you need!
IMPORTANT: If you're experiencing errors like identifier "GLintptr" is undefined
,
update to the latest glad version!
#include <glad/glad.h>
int main(int argc, char **argv)
{
// .. setup the context
if(!gladLoadGL()) {
printf("Something went wrong!\n");
exit(-1);
}
printf("OpenGL %d.%d\n", GLVersion.major, GLVersion.minor);
// .. render here ..
}
Examples:
- simple.c
- hellowindow2.cpp using GLFW:
If you don't want to install glad you can use the webservice
Otherwise either install glad via pip:
# Windows
pip install glad
# Linux
pip install --user glad
# Linux global (root)
pip install glad
glad --help
To install the most recent version from Github:
pip install --upgrade git+https://github.com/dav1dde/glad.git#egg=glad
Or launch glad directly (after cloning the repository):
python -m glad --help
Possible commandline options:
usage: glad [-h] [--profile {core,compatibility}] --out-path OUT
[--api API] --generator {c,d,volt}
[--extensions EXTENSIONS] [--spec {gl,egl,glx,wgl}]
[--no-loader]
Uses the official Khronos-XML specs to generate a GL/GLES/EGL/GLX/WGL Loader
made for your needs. Glad currently supports the languages C, D and Volt.
optional arguments:
-h, --help show this help message and exit
--profile {core,compatibility}
OpenGL profile (defaults to compatibility)
--out-path OUT Output path for loader
--api API API type/version pairs, like "gl=3.2,gles=", no
version means latest
--generator {c,c-debug,d,volt}
Language to generate the binding for
--extensions EXTENSIONS
Path to extensions file or comma separated list of
extensions, if missing all extensions are included
--spec {gl,egl,glx,wgl}
Name of the spec
--no-loader
--omit-khrplatform Omits inclusion of the khrplatform.h file which is
often unnecessary. Only has an effect if used
together with c generators.
--local-files Forces every file directly into the output directory.
No src or include subdirectories are generated. Only
has an effect if used together with c generators.
To generate a loader for C with two extensions, it could look like this:
python main.py --generator=c --extensions=GL_EXT_framebuffer_multisample,GL_EXT_texture_filter_anisotropic --out-path=GL
--out-path
and --generator
are required!
If the --extensions
option is missing, glad adds support for all extensions found in the OpenGL spec.
struct gladGLversionStruct {
int major;
int minor;
};
extern struct gladGLversionStruct GLVersion;
typedef void* (* GLADloadproc)(const char *name);
/*
* Load OpenGL using the internal loader.
* Returns the true/1 if loading succeeded.
*
*/
int gladLoadGL(void);
/*
* Load OpenGL using an external loader like SDL_GL_GetProcAddress.
*
* Substitute GL with the API you generated
*
*/
int gladLoadGLLoader(GLADloadproc);
glad.h
completely replaces any gl.h
or gl3.h
only include glad.h
.
if(!gladLoadGL()) { exit(-1); }
printf("OpenGL Version %d.%d loaded", GLVersion.major, GLVersion.minor);
if(GLAD_GL_EXT_framebuffer_multisample) {
/* GL_EXT_framebuffer_multisample is supported */
}
if(GLAD_GL_VERSION_3_0) {
/* We support at least OpenGL version 3 */
}
On non-Windows platforms glad requires libdl
, make sure to link with it (-ldl
for gcc)!
Note, there are two kinds of extension/version symbols, e.g. GL_VERSION_3_0
and
GLAD_VERSION_3_0
. Latter is a runtime boolean (represented as integer), whereas
the first (not prefixed with GLAD_
) is a compiletime-constant, indicating that this
header supports this version (the official headers define these symbols as well).
The runtime booleans are only valid after a succesful call to gladLoadGL
or gladLoadGLLoader
.
The C-Debug generator extends the API by these two functions:
// this symbol only exists if generated with the c-debug generator
#define GLAD_DEBUG
typedef void (* GLADcallback)(const char *name, void *funcptr, int len_args, ...);
/*
* Sets a callback which will be called before every function call
* to a function loaded by glad.
*
*/
GLAPI void glad_set_pre_callback(GLADcallback cb);
/*
* Sets a callback which will be called after every function call
* to a function loaded by glad.
*
*/
GLAPI void glad_set_post_callback(GLADcallback cb);
To call a function like glGetError
in a callback prefix it with glad_
, e.g.
the default post callback looks like this:
void _post_call_callback_default(const char *name, void *funcptr, int len_args, ...) {
GLenum error_code;
error_code = glad_glGetError();
if (error_code != GL_NO_ERROR) {
fprintf(stderr, "ERROR %d in %s\n", error_code, name);
}
}
You can also submit own implementations for every call made by overwriting
the function pointer with the name of the function prefixed by glad_debug_
.
E.g. you could disable the callbacks for glClear with glad_debug_glClear = glad_glClear
, where
glad_glClear
is the function pointer loaded by glad.
The glClear
macro is defined as #define glClear glad_debug_glClear
,
glad_debug_glClear
is initialized with a default implementation, which calls
the two callbacks and the real function, in this case glad_glClear
.
Import glad.gl
for OpenGL functions/extensions, import glad.loader
to import
the functions needed to initialize glad and load the OpenGL functions.
enforce(gladLoadGL()); // optionally you can pass a loader to this function
writefln("OpenGL Version %d.%d loaded", GLVersion.major, GLVersion.minor);
if(GL_EXT_framebuffer_multisample) {
/* GL_EXT_framebuffer_multisample is supported */
}
if(GL_VERSION_3_0) {
/* We support at least OpenGL version 3 */
}
On non-Windows platforms glad requires libdl
, make sure to link with it (L-ldl
for dmd)!
Easiest way of using glad is through the webservice.
Alternatively glad integrates with:
CMake
- Conan
Thanks for all the help and support maintaining those!
glad includes windows.h which breaks my code! #42
Defining APIENTRY
before including glad.h
solves this problem:
#ifdef _WIN32
#define APIENTRY __stdcall
#endif
#include <glad/glad.h>
But make sure you have the correct definition of APIENTRY
for platforms which define _WIN32
but don't use __stdcall
What's the license of glad generated code? #101
Any of Public Domain, WTFPL or CC0.
Contributing is easy! Found a bug? Message me or make a pull request! Added a new generator backend? Make a pull request!
Special thanks for all the people who contributed and are going to contribute! Also to these who helped me solve a problem when I simply could not think of a solution.