c++ – The right way to load shaders in a recreation engine asset loading system?

I’m making an asset loading system for my recreation engine, which is written in C++, and the thought I had in thoughts is that for every asset there’s a corresponding loader class. Though not totally the identical the thought was sparked from this thread.

static std::shared_ptr load(const fs::path& path) {
    AssetManager::getAssetData(); // hundreds the uncooked knowledge both from disk, community, and so forth
    AssetManager::create() // instantiates the asset and caches/registers it so it may be retrieved later AssetManager::get()
}

std::shared_ptr myMesh = MeshLoader::load("path/to/mesh.obj");

I’ve tried to make each loader type of align with this construction and for essentially the most half it was working fantastic, however I’m dealing with some challenges making use of it to shaders. The issue that I am discovering with them is that since they require a number of recordsdata how do you…deal with that? I may do ShaderLoader::load(path_to_vertex, path_to_fragment), however the issue with that is that I’m making an editor and I wish to load property by drag and dropping the file which so far as I do know can solely processes 1 file at a time.

The one answer I can consider is as a substitute of making an attempt to load and create the asset in a single step, I make it a 2-step course of by loading the asset knowledge after which creating the asset.

auto vertexSource = ShaderLoader::load("path/to/vertex.vert");
auto fragmentSource = ShaderLoader::load("path/to/fragment.frag");
std::shared_ptr shaderProgram = AssetManager::create();
shaderProgram->setVertexSource(vertexSource); 
shaderProgram->setFragmentSource(fragmentSource);

Finally I might identical to to know if that is sensible or is there a extra sensible approach that I’ve not thought of?