peer-id: support CIDv1 peer IDs in all recognized multibase encodings

[adust09]

Spec

specs/peer-ids/peer-ids.md, String Representation / Decoding.

Implementations MUST parse both legacy raw base58btc multihashes and multibase-encoded CIDv1 peer IDs. CID peer IDs must use the libp2p-key codec (0x72).

Current state

parsePeerId recognizes only the lowercase base32 multibase prefix b as a CID. Every other input is passed to the legacy raw-base58 decoder. Valid CIDv1 representations using another multibase encoding, such as base58btc (z) or base32upper (B), are therefore rejected.

Issue #117 added base32lower CID support but did not implement the spec's general multibase decoding requirement.

Scope

• Dispatch legacy raw peer IDs only for their defined 1... and Qm... forms.
• Decode CIDv1 using its multibase prefix rather than assuming base32lower.
• Require CID version 1 and multicodec libp2p-key (0x72).
• Validate the extracted multihash with the existing peer-ID constraints.
• Keep toCIDv1 base32lower as the default output format.
• Reject unknown prefixes, CIDv0-as-CID, wrong codecs, trailing bytes, and non-canonical varints.

Acceptance tests

• The same PeerId parses from base32lower (b), base32upper (B), and base58btc CIDv1 (z).
• Legacy 1... and Qm... forms still parse.
• An arbitrary base58 string is not silently treated as a legacy peer ID unless it has a valid legacy form.
• Wrong CID version/codec and malformed multibase inputs fail cleanly.