feat: multibase integration

[asmit27rai]

Multibase Integration

Integration of py-multibase into py-libp2p as discussed here #1173.

---

Phase-wise Plan

Phase 1: Add Multibase Support (Completed)

• Add py-multibase as a dependency.
• Extend Peer ID class with multibase encode/decode methods (backward compatible).
• Add multibase-aware utility functions for DHT keys.
• Update pubsub message ID encoding to use multibase.
• Preserve legacy base58/base64 support for compatibility.
• Add and update tests for multibase, compatibility, and error handling.

Phase 2: Internal Migration (Completed)

• Migrate internal code to use multibase by default.
• Update peer ID parsing, DHT, and pubsub internals.
• Update examples and documentation.

Phase 3: External API & Config (Completed)

• Consider making multibase the default for external APIs.
• Provide configuration for encoding preferences.

---

Testing

• All new and existing tests pass for Phase 3.

[seetadev]

@asmit27rai : Thanks for the detailed breakdown and the solid work on the multibase integration — the phased approach and backward compatibility are especially well thought out 👏

Inviting @lla-dane, @yashksaini-coder, @sumanjeet0012, @acul71 to review the changes, particularly around the internal migration, API defaults, and config surface.

Noting that a few CI/CD checks are currently failing (docs, lint across multiple Python versions, and Windows 3.13 core). Could you please take a pass at fixing these so we can get this into a mergeable state? Happy to help debug if needed.

Overall, this is a strong contribution and a meaningful step forward for py-libp2p. Looking forward to the reviews and CI being green ✅

[yashksaini-coder]

Multibase Integration (PR #1179)

Issues that I Found

Critical Blockers

• **Missing newsfragment file in newsfragments/.

Major Issues

1. Inconsistent Return Type for multibase.decode(): In some versions, this returns bytes, in others (encoding, payload). Handling exists, but not explained.

2. Potential Race in encoding_override: In the encoding_config.py, context manager, The global lock is released after context init, not held for whole override duration. Docstring/doc comment should clarify race condition possibility.

3. Unclear Error in ID.from_string():

   • In the main id.py, method ID.from_string, When decoding fails, users cannot tell if it was invalid base58 or multibase without better error messages.

---

Security Risks & Risk Mitigations

• Overall Security Risk: LOW
• Input Validation: Uses py-multibase, validates encodings, and falls back to base58 for compatibility.
• Thread Safety: Protected by a lock, but encoding_override() not thread-safe under concurrency.