feat(es/utils): Support for arrays using cast_to_number

[levi-nz]

Description:
This PR allows ArrayLits to be converted to numbers in the cast_to_number function. This allows expressions using arrays to be converted to numbers. See some example expressions below that were previously not able to be computed, but are now able to due to this change.

+[] // 0
+[[]] // 0
+[1] // 1
+[undefined] // 0
+[null] // 0
+[[1]] // 1
+[,] // 0
+[,,] // NaN

Regarding the implementation, arrays are converted to strings, and the string is then parsed as a number. So arrays like [] and [undefined] return "" which then return 0 when parsed as a string. This is also why arrays with more than one element can't be parsed because e.g. [1, 2] returns "1,2". This procedure follows the ECMAScript specification.
https://262.ecma-international.org/6.0/#sec-tonumber
https://262.ecma-international.org/6.0/#sec-toprimitive

BREAKING CHANGE:

Related issue (if exists):

[codspeed-hq]

CodSpeed Performance Report

Merging #9212 will degrade performances by 38.94%

_{Comparing levi-nz:arrays-to-numbers (bb7ea9d) with main (92879b1)}

Summary

❌ 6 regressions
✅ 168 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	levi-nz:arrays-to-numbers	Change
❌	es/full/bugs-1	587.4 µs	612 µs	-4.03%
❌	es/full/base/resolver_and_hygiene	328.7 µs	538.3 µs	-38.94%
❌	es/visitor/base-perf/boxing_boxed	3.8 µs	4 µs	-3.66%
❌	es/visitor/base-perf/boxing_boxed_clone	2.6 µs	2.8 µs	-5.25%
❌	es/visitor/base-perf/boxing_unboxed	3.7 µs	3.9 µs	-5.18%
❌	es/visitor/base-perf/boxing_unboxed_clone	2.6 µs	2.8 µs	-8.35%

[levi-nz]

ff2749d might be better for performance, and does the same thing as far as I'm aware, but the current code is more correct.

The performance change might also simply be because of arrays not being casted to numbers before.

[kdy1]

No worries, those performance issues are not related to this PR

[kdy1]

swc-bump:

• swc_ecma_utils
• swc_core

[levi-nz]

Sounds good, also I'm not sure if the return statements are correct regarding purity but I assume you've looked over that already