Eric Myhre

I like that this issue also touches on the user story of wanting to actively remove specific content. However, I wonder if that's directly connected to GC. I think we...

I'm not sure what the standards are in this repo or who should review, but fwiw, I would love to see this being mergeable. ISTM some useful data was gathered...

This probably happens because the "prelude" was revamped to use `Any`, despite the codegen's lack of support for it. You can see where that's introduced here: https://github.com/ipld/go-ipld-prime/blob/65bfa53512f2328d19273e471ce4fd6d964055a2/schema/dmt/compile.go#L38 The codegen should...

Just to describe the space and range of options a bit: Giving responsibilities to the Encode step towards the end there is possible. However, it would have high implementation cost:...

I otherwise agree with the strong hunch about the possibilities :)

I made [some comments in the PR exploring using Patch for this](https://github.com/smrz2001/go-ipld-prime/pull/1#issuecomment-1076421107) -- summary: something in this direction could be very cool, but I wonder if we should make a...

One other thought on where this API should attach: I think it should probably be on the `NodeAssembler`. Not the `NodeBuilder` -- because then amends would only work at the...

Hm, that sounds like a bug. You're sure an empty (rather than nil) list doesn't do what you want?

I'm fairly sure the relevant branches are here: https://github.com/ipld/go-ipld-prime/blob/4acc5447a2efc98fa91a2396c43f570299b4a782/traversal/walk.go#L104-L108 and that returning an empty list should jump down to [here](https://github.com/ipld/go-ipld-prime/blob/4acc5447a2efc98fa91a2396c43f570299b4a782/traversal/walk.go#L148) and cause "don't explore any sub-paths" as desired.

(Also kinda weird if a selector was going to statically return nothing for interests though, isn't it?)