Irakli Gozalishvili comments

Results 654 comments of


                                            Irakli Gozalishvili

Support for plain array buffer

> It doesn’t even need to be another module, it could just be something like require(‘multihashing-async/core’). The reason I thought it was better to go with separate package is so...

Mutable namespace using "Mutex blocks"

> Bare in mind #256 (comment). If we use a DHT (or some key-value store) to store the mapping from identifier -> latest then the storer of the record needs...

Mutable namespace using "Mutex blocks"

> Bare in mind [#256 (comment)](https://github.com/ipfs/specs/issues/256#issuecomment-969217984). If we use a DHT (or some key-value store) to store the mapping from identifier -> latest then the storer of the record needs...

Define less ambiguous UnixFS spec in TS types

Got no permissions to request review so instead I'm tagging people by names @aschmahmann @achingbrain @rvagg @alanshaw @ribasushi

Define less ambiguous UnixFS spec in TS types

> I'm not sure introducing typescript into the spec is a good idea? We already have prose and protobuf as definition languages, adding a third just increases the probability of...

Define less ambiguous UnixFS spec in TS types

Few things I'd discovered in the realm of unspecified dags ``` ipfs ls /ipfs/Qme1Cyu7ujqn3dRkRGmeTLpHJgbGHFjKmud48XK5W8qA6h Error: only murmur3 supported as hash function ipfs dag get /ipfs/Qme1Cyu7ujqn3dRkRGmeTLpHJgbGHFjKmud48XK5W8qA6h {"Data":{"/":{"bytes":"CAU"}},"Links":[]} # this is what...

Define less ambiguous UnixFS spec in TS types

Here is another fun example, illustrating that filesize can't be trusted and probably should not be even encoded as it could be derived from Data + blocksizes. ``` echo '{...

Handle duplicate dependencies with same/different version in post-install

You could link them not at the top package, but rather at the local level, in fact that would have another benefit of making pachake that isn’t listed as dependency...

Separate private keys from the peer id representation

@vasco-santos why do we even have private key on PeerId ? It feels kind wrong, I think rust-libp2p takes a better approach by having a [node identity](https://docs.rs/libp2p/0.38.0/libp2p/identity/index.html) encapsulating a key...

Use `XHR` instead of `fetch` to provide progress updates

This would imply swapping following code path: https://github.com/ipfs/js-ipfs-utils/blob/61c7fe23acc8b00627d303cbf8c948e0a03c2b48/src/http.js#L147-L151 with something along the lines of https://github.com/Gozala/js-ipfs-lite-http-client/blob/66ef08a39aa8ba9bda4406d32583ec3c80850e8a/src/client.js#L27-L99