Pierre Krieger

I'm going to close this as done. While there might be some occasional panics remaining, saying that the implementation is unfinished is clearly not true anymore.

This can't be done in we want to support syncing from scratch. We can, however, implement this if we force the full node to warp sync.

Maybe a simple work-around is to introduce, in smoldot, a maximum gap between the finalized and new block. If a block needs to be verified and it is too far...

Still blocked on `snow`. At the moment, the compilation fails on `subtle`, but it's because its `std` feature is pulled by `snow`.

> When the example is moved over to vulkano-examples, should it include its own perspective matrix function? That's a good question. It's cumbersome to have your own function.

> Why the AutoCommandBufferBuilder doesn't expose the vkCmdPipelineBarrier function ? Per design, the AutoCommandBufferBuilder is supposed to automatically handle pipeline barriers for you.

The same problem arises for `fill_buffer`. We could accept any buffer whose content implements the `NoInvalidValue` trait, but it would be too cumbersome to forbid f32s.

TL;DR: - Right now you can trigger undefined behavior by creating a buffer that contains bools, enums, strs, etc. and writing to it with vulkano. That's not reaaally problem because...

> Also, I noticed that Vulkano would display an error message that the type of the uniform buffer was too large after I updated Vulkano to the master branch, so...

Actually we don't need to check the limits when a shader module is loaded but only when the pipeline layout is created (that's what the specs say). Therefore it is...