Irakli Gozalishvili

It may also be worth considering how we could integrate capability parsers into the system, such that they could be leveraged instead of having to redefine all the input types...

I have no strong feeling on any of this. I would however propose another (perhaps) controversial solution to the problem. What if we just use prettier and and call it...

I really would love web components support, it's almost framework agnostic. I think does pretty good job doing something along these lines and demoing integrations with frameworks as well https://shoelace.style/

I have not considered `upload/add` and now I wonder if `index/add` subsumes that functionality or if we do need both 🤔 In terms of what to do I see following...

Thinking bit more about it I think in the future upload list should simply be a list of CBOR objects like `{ root: Link, parts: Link[] }` where root is...

@joaosa I suspect that w3up-client tests are failing because the upload-api in-memory service isn't wired with new capabilities, I'm not exactly sure why though Actually looks like failing tests also...

@vasco-santos I think it would make most sense to just extend support to `raw` cids (as opposed to adding multihash support) mostly because I expect it to be less work...

@vasco-santos I think you could help fill this out with more details as I'm getting bit lost

@vasco-santos It occured to me that we could validate that input to piece CID derivation is a valid CAR since we have to read CAR bytes anyway to derive piece...

> Yes, we could check that there today. If we would move to a decentralized run of piece computation, where this would run in a Filecoin Station for example, this...