vadgl 0.2.2

Warning

This library is still in alpha states and will be actively changing. If you have any change in mind put it as an issue here.

About

D abstraction over OpenGL. Planned to be used with any OpenGL bindings. None of the functions in this library throw. If compiled with version VADGL_DisableChecks or in release version ~~all~~ most of the functions in the library do not allocate (The ones that do work with strings and there are const(char*) overloads that are @nogc).

Features

• @nothrow error handling system
• Does not allocate (in release and VADGL_DisableChecks)
• Debugging all OpenGL calls with (VADGL_DebugGLCalls)
• Works with any bindings
• Shader, Program, VBO, VAO, Attribute and Uniform abstractions

Why?

OpenGL is not known for having the best error handling. This library provides function overrides, a type safe D-like interface for OpenGL with it's own error handling model for better debugging plus abstractions over Shaders, VAOs, VBOs, Attributes and Uniforms.

How can I use this?

OpenGL functions in VADGL use snakecase instead of camelCase. For example the **VADGL** equivalent of `glCreateShader` is `glcreate_shader`. So programs like:

uint vert_sh = glCreateShader(GL_VERTEX_SHADER);
assert(glGetError() == GL_NO_ERROR);

becomes:

uint vert_hs = gl_create_shader(Shader.Type.VERTEX).throw_on_error();

or for @nothrow

uint vert_hs =
        gl_create_shader(Shader.Type.VERTEX)
            .unwrap_or_else!((result) => assert(0, result.error.to_error_msg));

With the additional benefit that if an error happens. You will get an error message like this:

[GL_ERROR]: glCreateShader: INVALID_ENUM

Using abstractions

Using Shader and Program abstractions you can create a Program like so:

string vertex_source = ... // vertex source
string frag_source = ... //  fragment source

Shader vert_sh, frag_sh;
vert_sh = Shader.from_src("vertex", Shader.Type.VERTEX, vertex_source) .throw_on_error();
frag_sh = Shader.from_src("frag", Shader.Type.FRAGMENT, frag_source) .throw_on_error();

program = Program.create_program("program").throw_on_error();

// Compile shaders
foreach (shader; [&vert_sh, &frag_sh]) {
    shader.compile().unwrap_or_else((res) {
        stderr.writeln(res.error.to_error_msg()); // Print compilation error
        // Get compilation error message if any
        stderr.writeln(vert_sh.get_info_log().throw_on_error);
        assert(0);
    });
}
program.attach(vert_sh, frag_sh).throw_on_error(); // attach
program.link().throw_on_error(); // link
program.validate().throw_on_error(); // validate

program.use().throw_on_error();

With this we create and compile shaders vert_sh and frag_sh, then attach them to the program program, and finally link and validate the program.

All of this with error handling. In case of a compilation error you will get an error message like this:

        HINT: Run get_info_log() to get extra information

Error on "vertex" shader
0:36(33): error: `model_pos' undeclared
0:36(27): error: cannot construct `float' from a non-numeric data type
0:37(33): error: `model_pos' undeclared
0:37(27): error: cannot construct `float' from a non-numeric data type

[email protected](15): Assertion failure

And just like that without any extra code we get so much information of what went wrong.

Extra examples (On construction):

TODO:

• [PRIORITY]: Write testing framework (Should be easy, we have a way to log GL)
• [PRIORITY]: Write documentation
• [PRIORITY]: ~~Maybe~~ use GLInternalError for functions that map directly to OpenGL
• Use GL types
• Rework result.d, keep API the same if possible
• Add more version blocks [important]
• Add abstraction over textures
• Maybe add runtime checks
• Add more OpenGL functions