> <@aschmahmann:matrix.org> dadepo: > > I'm pretty sure Bitswap can cancel WANTs so that it doesn't receive a block it has already received (e.g. I asked Alice and Bob for X. Alice gave it to me, so I send a cancel to Bob so that I don't get X a second time). If you're thinking about deny lists (e.g. I download a DAG starting at X, but it points eventually to Y, which is a block I know is bad, and therefore want to screen out the Y block) then that's not implemented natively in go-ipfs (maybe you're thinking of something like this https://github.com/ipfs/notes/issues/284). > > For groups that run gateways you can always block certain URLs which would prevent you from loading certain content. e.g. Cloudflare could just block https://cloudflare-ipfs.com/ipfs/QmXoypizjW3WknFiJnKLwHCnL72vedxjQkDDP1mXWo6uco/wiki/ and then that page would not render on their gateway (you could of course download it on your own node though). > > > I am doing a survey of IPFS for a school assignment so basically need to understand as much as possible of the moving parts > > Cool, well feel free to ask questions people here are generally pretty helpful and responsive. > maybe you're thinking of something like this https://github.com/ipfs/notes/issues/284) Yup, that is exactly what I am looking for. On my forked branch, I already modified the code to have some rudiment/crude black-listing functionality. It works but it became apparent that the approach won't scale. This is due to the fact that different chunk size can be used to import the same data into IPFS, leading to different CIDs. This means you need to have blacklists for all CIDs possible for a particular content you want to deny on the network (If i remember correctly the minimum chunk size is 256k while the maximum allowed in 1MiB). So this will become unwieldy pretty quick. Another possibility is to have a separate hashing layer for contents that would always be unique and use this for the deny lists...but this won't be an idea I will be exploring further based on the scope of my assignment. > Cool, well feel free to ask questions people here are generally pretty helpful and responsive. Indeed! :) And I find this very pleasant :)