Fix median sort, set-op duplicates, and Unescape Unicode range (#2239, #2241, #2242)

[vigneshrajan94]

Summary

Fixes three independent bugs in a single PR.

#2239 — Median returns wrong result for unsorted odd-length inputs

src/core/lib/Arithmetic.mjs — median()

The .sort() call was inside if (even), so odd-length arrays were never sorted before picking the middle element. Median([10, 1, 2]) returned 1 instead of 2.

Fix: sort unconditionally before branching on even/odd length.

---

#2241 — Set Difference / Set Intersection preserve duplicates from first input

src/core/operations/SetDifference.mjs, src/core/operations/SetIntersection.mjs

Both operations used a plain .filter() on array a. If a contained duplicate entries that passed the filter, all copies appeared in the output, violating set semantics. SetDifference(["x","x","y"], ["y"]) returned "x,x".

Fix: deduplicate a via [...new Set(a)] before filtering.

---

#2242 — Unescape Unicode Characters rejects code points > 4 hex digits under U+

src/core/operations/UnescapeUnicodeCharacters.mjs — run()

The regex quantifier was hardcoded to {4} for all prefix types. The U+ notation supports 4–6 hex digits (e.g. U+1F600 for 😀, U+000041 zero-padded). \u and %u remain at {4} per their respective specs.

Fix: use {4,6} when the selected prefix is U+, {4} otherwise.

---

Test plan

• npm test — all 2178 tests pass (243 Node API + 1935 operations), zero failures
• Median with input 10,1,2 (Comma delimiter) → 2
• Set Difference with sample 1 x,x,y, sample 2 y → x
• Set Intersection with sample 1 y,y,z, sample 2 y → y
• Unescape Unicode Characters (prefix U+), input U+1F600 → 😀
• Unescape Unicode Characters (prefix U+), input U+000041 → A
• Unescape Unicode Characters (prefix \u), input A → A (4-digit still works)

[HackingRepo]

@vigneshrajan94, @GCHQDeveloper58, will say you what ai you used?

[C85297]

@vigneshrajan94 thank you for your pull request. In general, we would prefer minimally scoped individual pull requests, rather than pull requests incorporating multiple features/bug fixes. Would you mind splitting the pull request into one for each issue? This makes it easier for us to review and to track changes.