test(fuzz): Lua encode/roundtrip property fuzz (sibling of #63)

[membphis]

Background

The qjson cdylib is a decoder; cargo-fuzz (#63) covers it. But encoding
and materialize live in Lua (qjson/table.lua, surfaced as qjson.encode
/ qjson.materialize) and are unreachable by cargo-fuzz — a different
toolchain with zero shared code.

#71 fixed a stack overflow in the encoder; that class of bug only gets
fuzz coverage through the Lua layer. This issue is the sibling of #63
(decoder-side fuzz).

Goal

Property / fuzz coverage for the Lua encode + roundtrip path:
decode(encode(x)) == x, plus encoder stack-depth safety.

Scope (this issue = one PR)

• Evaluate luzer (LuaJIT libFuzzer binding) vs a busted
  property-test approach; pick one.
• Roundtrip property: parse(x) → materialize → encode → re-parse,
  assert structural equality.
• Encoder stack-depth property under deep nesting (relates fix(encode): add depth counter to prevent stack overflow #71).

Notes

• Distinct from test: Lua-level encode error coverage (NaN, type errors, sparse arrays) #65 (fixed-case encode error coverage: NaN, type errors,
  sparse arrays); this issue is fuzz / property-based, not fixed cases.
• Zero code shared with test: fuzz harness (cargo-fuzz / libFuzzer) #63's fuzz/ crate (different language + toolchain) —
  which is why it is a separate issue rather than a test: fuzz harness (cargo-fuzz / libFuzzer) #63 checklist item.

Affected files

• Lua-side fuzz / property harness (location TBD by the luzer-vs-busted choice)

Sibling of #63. Relates to #65, #71.

[coderabbitai]

🔗 Related PRs

#43 - test: add cjson and simdjson fixture coverage [merged]
#54 - perf: Lua encode + cache + scanner + validation stack optimizations [merged]
#60 - feat: preserve key order in decode-modify-encode workflow [merged]
#71 - fix(encode): add depth counter to prevent stack overflow [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!