questions about text rendering

[dwuggh]

I'm trying to implement a text renderer, After spending days looking into the code, I've managed to write down a simple helloworld program here. (TBH, The official examples are way too complex lol) Feel free to check out my mistakes!
And I got several questions.

1. The architexture of compute-centric rendering. I'm new to this, and still a bit confused looking at this shaders. If I understand correctly, these shaders should implement transformation and rasterization, to replace the render pipeline. But how to transform a render shader(has vertices, indices and fragments) to a compute shader? Should sampler be used(e.g. textureSampleLevel)? Also, there's no sampler in vello::BindType TODO here, is this still needed?
2. glyph raterization and font shaping. vello currently only perform glyph rasterization directly in gpu, is that right? If so, how is this compared to swash or zeno? I notice swash has cache for rasterization, didn't find one in vello though. What's more, is it possible to directly use swash's API to do shaping for this?
3. relationships between swash, skrifa and peniko. Are them designed to be working together? It seems some types are esstentially the same, but not exactly.

Thanks in advance!