Detect embedded IPv4 in non-canonical 6to4 addresses

[zanbaldwin]

Derived::isEmbedded() only matched canonical 6to4 addresses with the trailing 80 bits zeroed, even though every address within 2002::/16 carries an IPv4 address in bits 16-47, as per RFC 3056. Widen detection to the whole 6to4 block; extraction keeps only the embedded IPv4 address, so an extract-pack round-trip reconstructs the canonical form (2002:XXXX:XXXX::) rather than the original.

[greptile-apps]

Greptile Summary

This PR widens Derived::isEmbedded() to detect any address in the 2002::/16 block (not just canonical forms with zeroed bits 48–127), aligning with RFC 3056 § 2 which embeds the IPv4 address in bits 16–47 regardless of the remaining fields. The companion docblocks explicitly document that extract→pack is a one-way canonicalisation for non-canonical addresses.

• Derived::isEmbedded() now returns true for the entire 2002::/16 range; the trailing-zero guard is removed. Any Multi IP built from a non-canonical 6to4 address will now report getVersion() === 4 and expose the embedded IPv4 via getDotAddress() / getProtocolAppropriateAddress().
• Compatible and Mapped receive doc-comment additions with RFC references; no logic changes.
• The documentation table is updated to use ???? for the variable bits of the 6to4 format.

Confidence Score: 4/5

The core logic change is small, well-reasoned against RFC 3056, and the docblocks clearly warn about the non-roundtrip extraction. The only gap is that the new non-canonical detection path has no test coverage.

The implementation correctly drops a guard that made detection too narrow, and pack/extract behaviour is unchanged. The sole concern is that the test data only contains canonical 6to4 forms, so no test actually exercises the newly-widened isEmbedded() path with a non-zero SLA or interface ID.

tests/DataProvider/Strategy/Derived.php — getValidSequences() and getInvalidSequences() should include at least one non-canonical 6to4 address to cover the new detection logic.

Important Files Changed

Filename	Overview
src/Strategy/Derived.php	Drops the trailing-zero guard from isEmbedded() so any 2002::/16 address is detected; adds docblocks clarifying the canonical-only roundtrip semantics. Logic is correct per RFC 3056.
src/Strategy/Compatible.php	Doc-comment addition only; no logic changes.
src/Strategy/Mapped.php	Doc-comment addition only; no logic changes.
docs/05-strategies.md	Table updated to show wildcard bits (????) in the 6to4 format column, accurately reflecting the relaxed detection.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[16-byte binary input] --> B{First 2 bytes == 0x2002?}
    B -- No --> C[isEmbedded = false]
    B -- Yes --> D[isEmbedded = true]
    D --> E[extract: return bytes 2-5 as IPv4]
    E --> F{Canonical?\nbytes 6-15 == 0x00...?}
    F -- Yes --> G[roundtrip safe:\npack extract = original]
    F -- No --> H[lossy: SLA ID + Interface ID\ndiscarded by pack]
    H --> I[pack produces canonical\n2002:XXXX:XXXX::]

Loading

_{Reviews (1): Last reviewed commit: "bugfix(strategy): 🐛 detect embedded IPv..." | Re-trigger Greptile}