Simplify date parser

[CrazyboyQCD]

• Avoid allocation if string is ascii
• Return early if string from utf16 encoded JsString is not ascii
• Replace Char<'_> with slice::Iter<'_, u8>
• Add next_n_digits helper method to save calls of next_digit
• Merge duplicate codes in matches

[codecov]

Codecov Report

Attention: Patch coverage is 64.78873% with 25 lines in your changes missing coverage. Please review.

Project coverage is 53.56%. Comparing base (6ddc2b4) to head (d430a23).
Report is 357 commits behind head on main.

Files with missing lines	Patch %	Lines
core/engine/src/builtins/date/utils.rs	64.78%	25 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4143      +/-   ##
==========================================
+ Coverage   47.24%   53.56%   +6.31%     
==========================================
  Files         476      487      +11     
  Lines       46892    48863    +1971     
==========================================
+ Hits        22154    26172    +4018     
+ Misses      24738    22691    -2047     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[nekevss]

Looks good to me at a high level. I like the move from char to u8 ascii. I have a suggestion.

[nekevss]

suggestion: Use next_digit over using next_n_digits where possible.

As tempting as next_n_digits is to write, most the lengths should be relatively stable here. We can then also avoid using unsafe code.

[CrazyboyQCD]

I put next_digit in next_n_digits now because I found Peekable<slice::Iter> doesn't give the len of slice::Iter under serveral conditions, so the length check is useless.
But I prefer using next_n_digits as it is not that noisy here.