Port Function

• Every RNG call (rn2, rn1, rnd, d, rnz, rne) and its arguments
• Every conditional branch and the order of operations
• Side effects on global state (flags, context, u. fields)
• Early returns and their conditions

3. Check for existing JS equivalent

Search js/*.js for the function name or a camelCase equivalent:

grep -rn "function_name\|functionName" js/

4a. If no JS equivalent exists -- create it

• Identify which JS module it belongs in (match the C source file mapping)
• Translate line-by-line, preserving:
  • RNG call order exactly
  • Conditional structure exactly
  • Side effect order exactly
• Use the project's established patterns (see docs/LORE.md Cardinal Rule 2)
• Mark C-isms that need JS adaptation:
  • char comparisons become string comparisons
  • Pointer arithmetic becomes array indexing
  • struct field access becomes object property access
  • Bitfield flags become numeric bitmask operations

4b. If JS equivalent exists -- compare for divergences

• Diff the C and JS side by side
• Check every RNG call matches (same function, same arguments, same order)
• Check every conditional has the same structure
• Check boolean coercion: C 0 is falsy, but JS '\0' is truthy
• Check signed/unsigned arithmetic: JS numbers are IEEE 754 doubles
• Flag any divergences with exact line references

5. Run relevant session tests

node --test test/comparison/sessions.test.js

If specific sessions exercise this function, run them individually:

node test/comparison/session_test_runner.js --verbose <session-path>

6. Write a LORE entry if anything non-obvious was found

If the port required any non-obvious adaptation (C-ism, RNG ordering subtlety, boolean coercion trap), append a structured entry to docs/LORE.md under "Recent Findings":

### YYYY-MM-DD - short description

- Context: which function / what session exposed this
- Problem: what was wrong or tricky
- Why: root cause in C semantics
- Fix: what the JS port does differently and why

7. Report metrics

Summarize:

• Function name and source file
• Lines of C ported
• RNG calls count
• Divergences found (if auditing existing port)
• Test results (pass/fail, sessions affected)
• LORE entry written (yes/no)

Common Pitfalls

• Cardinal Rule 2: Read the C, not the comments. Port the actual behavior.
• Boolean coercion: vault_occupied() returns '\0' which is truthy in JS.
• mcanmove: Treat as C boolean -- both 0 and false mean immobile.
• Object age: mkobj.newobj() must seed obj.age from current move count.
• Thrown stacks: C computes multishot before splitobj() / next_ident().