Help understand glemu blend modes (like GL_ADD)

[ericoporto]

Hi,

I was trying to understand how glemu works for something like GL_ADD

emscripten/src/library_glemu.js

Line 1372 in 1171ada

case GL_ADD: {

Afaict it needs, in the case of GL_ADD, use the color of the pixel of the texture and the resulting color of all pixels direct below to be able to calculate the value that pixel should be.

I was trying to understand how this is achieved. The issue I can’t understand is how it samples what is not the current texture, does it flatten all that is below? How does it work? Is there a command that has to be done before to raster everything and push it into a texture (the back buffer?).

If someone could explain this to me it would be helpful, both conceptually and any relevant implementation detail on how this is done.