chore: ignore arbitrary data stored in multiaddrs enr key

[richard-ramos]

Description

This would allow reading the multiaddrs key and not fail if one of the values encoded in that key is not a multiaddress.
The reason why this change is being proposed is that it would allow storing some data which can be translated to multiaddresses, for example circuit relay addresses which could be stored with this format:

- 4 bytes for the IP address
- 2 bytes for the Port
- 33 bytes for a secp256 compressed public key

which would mean storing 39 bytes compared to storing 50 bytes which would be the bytes encoding of a multiaddress containing the same information. (In ENRs the less bytes we use the better)

cc: @fryorcraken

[github-actions]

You can find the image built from this PR at

quay.io/wakuorg/nwaku-pr:2853

Built from afacddb

[fryorcraken]

I don't think we should store arbitrary data in mutliaddr, if we go for the new format, we should use a new key.

However, I agree with this change in the way that if not able to read a field, we shouldn't fully fail the enr decoding and just attempt to use the other fields to connect to a node.

[jm-clius]

This makes sense to me. We shouldn't fail on multiaddrs decode failure.

[gabrielmer]

LGTM! Thank you!