bug: setUnion returns wrong element order when sets overlap (intersection placed at end)

## Bug Description

`math.setUnion()` produces incorrect results whenever the two input sets share any elements. The intersection is placed at the **end** of the output rather than in its correct sorted position.

## Reproduction

```js
math.setUnion([1, 2, 3, 4], [3, 4, 5, 6])
// actual: [1, 2, 5, 6, 3, 4]
// expected: [1, 2, 3, 4, 5, 6] ← as documented in the JSDoc

math.setUnion([1, 2], [2, 3])
// actual: [1, 3, 2]
// expected: [1, 2, 3]
```

The JSDoc example explicitly documents `setUnion([1, 2, 3, 4], [3, 4, 5, 6])` → `[1, 2, 3, 4, 5, 6]`.

## Root Cause

The implementation uses:

```js
return concat(setSymDifference(b1, b2), setIntersect(b1, b2))
```

`setSymDifference` returns `[a-only elements, b-only elements]`, so the result is
`[a-only, b-only, a∩b]` — the intersection lands at the end instead of between the two halves.

## Expected Behaviour

Union should return the sorted result: **a-only elements** ++ **common elements** ++ **b-only elements**, i.e.:

```js
return concat(concat(setDifference(b1, b2), setIntersect(b1, b2)), setDifference(b2, b1))
```

## Environment

- mathjs version: develop branch

Start at the math.setUnion implementation and compare its current concatenation order with the JSDoc example and the two reproductions in the issue. Verify that overlapping inputs produce a sorted union with exclusive elements before, between, and after the common elements.

Written by the indexing model from the issue text.