Binary Serialization And Codecs

What's being tested

These problems test reversible serialization: converting maps, strings, or KV-store state into bytes and reconstructing the exact original data. Interviewers probe whether you can design binary-safe codecs that handle arbitrary bytes, Unicode, empty values, large inputs, malformed payloads, and format evolution without relying on fragile delimiters.

Patterns & templates

• Length-prefix encoding — write len(key) | key | len(value) | value; O(total_bytes) time, delimiter-free, handles arbitrary characters.

• Fixed-width integer headers — use uint32 or uint64 lengths with explicit endianness; reject negative, overflowed, or truncated lengths.

• Type-length-value (TLV) layout — encode type | length | payload for extensible records; useful for versioned KV entries and metadata.

• Round-trip invariant — always test deserialize(serialize(x)) == x; include empty map, empty string, Unicode, null bytes, duplicates, and huge values.

• Streaming parser — maintain an offset, bounds-check before every read, and fail fast on trailing bytes or incomplete records.

• Versioned format header — prefix with magic | version | flags; enables backward-compatible changes and early rejection of invalid data.

• Persistent mutation log — append PUT / DELETE records and replay into a map; optionally add checksums and compaction for robustness.

Common pitfalls

Pitfall: Using : or , delimiters breaks as soon as keys or values contain the delimiter; prefer length-prefixing or escaping with strict decoding.

Pitfall: Forgetting bounds checks lets malformed input panic, over-read, or allocate massive buffers; validate every declared length before slicing.

Pitfall: Treating strings as characters instead of bytes causes Unicode bugs; serialize UTF-8 bytes and define whether keys/values are str or raw bytes.

Practice these

The practice cards below cover the canonical variants — solve all of them and time yourself.