perf(parser): remove redundant checks

[ulrichstark]

1. is_number in class.rs can be removed, because is_identifier_or_keyword implies is_literal_property_name implying is_number
2. Avoids checking for is_reserved_keyword one way and then the other way.
3. is_all_keyword in is_identifier_or_keyword can be removed, because is_literal_property_name implies is_identifier_name implying is_all_keyword

Let's see if those changes yield any improvements in the benchmark or if the compiler was smart enough to optimize that.

Edit: removed the second change because it's not worth the sacrificed readability

[codspeed-hq]

CodSpeed Instrumentation Performance Report

Merging #11207 will not alter performance

_{Comparing ulrichstark:remove-redundant-checks-in-lexer (281dd3c) with main (8e8dea5)}

Summary

✅ 38 untouched benchmarks

[ulrichstark]

Doesn't look like there are any substantial improvements. Feel free to close my PR or pick the changes you want. Two of my three changes are simple removals and make the code simpler. is_identifier was honestly more readable before I changed it, but I wanted to try it anyway for potentially better performance.

Edit: removed the second change because it's not worth the sacrificed readability

[Boshen]

Less code is always 👍