small updates to README to get demo to work

[DougAnderson444]

Cool library, just need some readme tweaks for those trying to follow along:

Chg blockcodec-to-ipld-format to ipld-format-to-blockcodec to get it to work properly
Chg formats to codecs as per docs

[rvagg]

I'm pretty sure this is not correct, you're converting it the wrong way. But I think what you're encountering is that the lastest js-ipfs now uses the new BlockCodec style of codecs natively so I don't believe you need the adapter now. Try getting rid of the conversion code entirely and pass [dagJose] in to the codecs option and see if that works for you. If so, maybe update the PR to remove references to the legacy conversion entirely.

[DougAnderson444]

Thanks Rod! Yep, what you suggested totally works. Don't even need to load the codecs, looks like Multicodecs auto loads that for us?

import * as dagJose from 'dag-jose';

let ipfs = await createIpfs();

const cid = await this.ipfs.dag.put(jwe, {
		format: dagJose.codec,
		hashAlg: 'sha2-256'
	});

I'll modify the PR to just remove that config code that doesn't work anymore.

[rvagg]

Yeah, blockcodec-to-ipld-format and ipld-format-to-blockcodec were for bridging old and new codec formats, but we've since entirely deprecated the old codec formats, even in js-ipfs, and just use the new ones now which follow the js-multiformats BlockCodec pattern. So 🥳 .

[rvagg]

Suggested change

import { convert } from 'ipld-format-to-blockcodec';

[rvagg]

Suggested change

When using DAG-JOSE (for JWE or JWS) with js-IPFS, you will need to convert it from a raw multiformats style codec to a legacy IPLD codec using [ipld-format-to-blockcodec](https://github.com/ipld/js-ipld-format-to-blockcodec).

[rvagg]

Also note that even this change is going to be breaking soon thanks to a change in the DAG API in go-ipfs v0.10 where format is no longer accepted as an argument, you can't run this code against a go-ipfs v0.10 backend with the JS ipfs http client. That should get fixed via ipfs/js-ipfs#3917. Although this current pattern could be addressed by using the BLOCK API instead of the DAG and just feeding it the raw block data, maybe I should have written the example that way instead..

[DougAnderson444]

Perhaps it should wait until #3917 is merged then? My whole point with this note was to bring the README up to speed for anyone else looking at this so they can understand what's going on