fix(es/minifier): mark properties assigned through destructuring targets as mutated

[Baltasar Blanco (baltasarblanco)]

Description:

With hoist_props, a property that is only written through a destructuring assignment
target ([o.n] = [2], ({ x: o.n } = obj)) was treated as never written, so it was
hoisted and its initial value inlined into the assignment target itself, producing invalid
JavaScript ([1]=[2]).

The usage analyzer already marks o as property-mutated for o.n = v via
mark_mutation_if_member, but the AssignTarget::Pat branch of visit_assign_expr and
the for-in/for-of heads never did. This PR walks the pattern and marks every member target
at any depth (array, object, default, rest) the same way. It only ever over-preserves.

The extra work runs only for assignments with a pattern on the left and for-in/for-of
heads, as a single walk of the pattern.

The existing issues/11084 fixture runs without hoist_props, which is why the original
report still reproduced. With @swc/core 1.16.2 default options it breaks inside any
function body (e.g. an IIFE), and at the top level once toplevel is enabled.

This also updates issues/9739: its snapshot had frozen the same bug through variable
inlining ([arr[0], arr[1]] = … became writes to fresh array literals, so arr was never
swapped). The new output keeps the swap.

Testing:

• issue_12397_hoist_props_destructuring_member_targets (exec test, 9 variants) fails
  without the change and passes with it.
• cargo test -p swc_ecma_minifier --no-fail-fast: 0 failures across all test binaries.
• cargo clippy -p swc_ecma_minifier --all-targets -- -D warnings: clean.

Fixes #12397. Follow-up of #11084 / #11221.

[codspeed]

Merging this PR will not alter performance

✅ 61 untouched benchmarks
⏩ 207 skipped benchmarks¹

---

_{Comparing baltasarblanco:fix/12397-hoist-props-destructuring-targets (601e1db) with main (4b42cd5)}

Footnotes

1. 207 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[Donny/강동윤 (kdy1)]

Thank you!