Add complete IPFS path validation

[Arjentix]

Complete validation requires checking if path is composed only from valid hashes

ipfs crate source code can be helpfull

Alternative implementations

#1606

[ilchu]

After some research I am doubtful if this is actually feasible in the scope of data_model crate. There are two ways which I see could be possible for file fetching after decoding the CID hash, namely either bootstrapping a whole IPFS node and running some Kademlia FIND_VALUE queries or trying to fetch it through an HTTP gateway.

The first one has the overhead of bringing in at least the ipfs and tokio or maybe warp, as the
ipfs crate example for fetching shows, which I believe is out of scope of our data model.

The second one would only require something like reqwest, but is still likely to fail miserably, as IPFS currently doesn't seem to have a mechanism for validating file absence (because even if the closest discovered peer lacks the content, we still have to go deeper down the chain hoping some next one might have it), and basically the only way to signal it is request timeout. An example I've found is from an Infura user of the gateway, and there doesn't seem to have been any progress on that from their part from what I can gather. When done manually with an invalid path (e.g. https://ipfs.io/ipfs/Qme7ss3ARVgxv6rXqVPiikMJ8u2NLgmgszg13pYrDKEoii) via httpie, it just expires in a minute with HTTP/1.1 504 Gateway Time-out as expected.

What would be your thoughts on this?

[appetrosyan]

I think we should re-visit this in the Blockchain explorer context. I'll add the relevant issue to the board, and re-assign work.

[appetrosyan]

Closed in favour of soramitsu/iroha2-block-explorer-backend#20