feat(parser/html): lex and parse unquoted attribute values

[dyc3]

Summary

This PR lets the HTML parser correctly lex and parse unquoted attribute values.

Test Plan

Added a test. See also: https://html.spec.whatwg.org/#attributes-2

[codspeed-hq]

CodSpeed Performance Report

Merging #3951 will degrade performances by 6.44%

_{Comparing 09-16-feat_parser_html_lex_and_parse_unquoted_attribute_values (4b03d2f) with main (2c8ff7f)}

Summary

❌ 1 regressions
✅ 106 untouched benchmarks

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

Benchmarks breakdown

	Benchmark	main	09-16-feat_parser_html_lex_and_parse_unquoted_attribute_values	Change
❌	react.production.min_3378072959512366797.js[cached]	1.9 ms	2 ms	-6.44%

[ematipico]

Looks to me, however we should:

• Add more test cases, for example, cases where the attribute value has spaces (\n, \t) and </>, because they are part of the new logic.
• Add an invalid test case that covers the new error you added

[ematipico]

Thank you, there's one case to cover and we can merge it