Support microsoft-conform user class option in DHCPv6

[damyan]

This pull request addresses an issue in the DHCPv6 User Class option (OptUserClass) where the FromBytes method failed to parse data in the Microsoft-compatible format. The existing implementation adhered strictly to the RFC 8415 specification, which requires each user class to be preceded by a 16-bit length field. However, Microsoft DHCPv6 implementations deviate from this by providing a single user class as raw bytes without a length prefix. This mismatch caused a "buffer too short" error when parsing Microsoft-formatted data, as observed in the otherwise failing test TestOptUserClassBroken.

Changes Made

• Updated FromBytes Method:
  • The method now attempts to parse the data as multiple user classes with length prefixes (per RFC standards).
  • If a length field exceeds the remaining buffer size, parsing halts at that point.
  • If no valid user classes are parsed (e.g., due to missing or invalid length prefixes), the entire data buffer is treated as a single user class, aligning with Microsoft’s format.
  • This dual-parsing approach ensures compatibility with both RFC-compliant and Microsoft-specific inputs without raising errors.

Why This Fix Is Necessary

• Interoperability: Supporting both RFC and Microsoft formats is essential for seamless interaction with a wide range of DHCPv6 clients, including those in Microsoft environments.
• Test Reliability: The previous implementation caused TestOptUserClassBroken to fail unnecessarily. This change eliminates the error while maintaining functionality for standard inputs.

[pmazzini]

Related for DHCPv4: #114