feat: add new dot-path array helpers

[michalsn]

Description
This PR:

• Adds five new dot-syntax helper functions: dot_array_has(), dot_array_set(), dot_array_unset(), dot_array_only(), and dot_array_except(), each backed by a corresponding ArrayHelper static method.
• Renames ArrayHelper::dotKeyExists() to dotHas() and updates all callers, including Validation/Rules.php and Validation/StrictRules/Rules.php. This is a safe change since ArrayHelper is marked as @internal.
• Adds a fast path in ArrayHelper::convertToArray() that uses explode() for the common case (no escaped dots), avoiding PCRE engine overhead.

Checklist:

• Securely signed commits
• Component(s) with PHPDoc blocks, only if necessary or adds value (without duplication)
• Unit testing, with >80% coverage
• User guide updated
• Conforms to style guide

[datamweb]

For micro-optimization, how about using the first-class callable strlen(...) instead of a closure to filter out empty strings while safely keeping '0'?

Suggested change

	explode('.', $trimmed),
	static fn ($s): bool => $s !== '',
	explode('.', $trimmed),
	strlen(...),

[michalsn]

In this context, static fn ($s): bool => $s !== '' is explicit, while strlen(...) makes developers figure out that it means "non-empty string".

Thanks for the suggestion, but I will keep the closure here for readability reasons.

[datamweb]

Since the array becomes completely empty here, wouldn’t using break instead of continue be more efficient to avoid processing any remaining keys?

Suggested change

	continue;
	break;

[michalsn]

In theory, you're right, but then the exception depends entirely on argument order: ['users.*.*', '*'] throws, ['*', 'users.*.*'] does not.

[datamweb]

Thanks! That’s a brilliant catch!

[michalsn]

Thank you for the reviews.