limit allowed characters to UTF-8 range (0x10FFFF)

[IonBazan]

Requirements

This change limits valid characters recognized to 0x10FFFF according to UTF-8 specification.

Description of the Change

This is to make the regular expressions PCRE-compliant, where matched characters should not fall out of UTF-8 bounds: https://www.pcre.org/original/doc/html/pcreunicode.html
Since UTF-8 is a standard encoding for PHP files according to PSR-1, I don't see a point supporting any invalid Unicode characters.

Alternate Designs

Benefits

The reason for this change is actually to make GitHub Linguist support this grammar as it sticks to strict PCRE rules.

Before

After

Possible Drawbacks

Any invalid character will stop being recognized as a variable name but that shouldn't occur in real world.

Applicable Issues

github-linguist/linguist#5522

While awaiting for workflow run approval, let me just confirm that tests are passing locally.

[KapitanOczywisty]

@sadick254 This is ready to be merged. Other PRs could introduce more 7fffffff, so this should probably go after them and some "replace all" might be needed.

[darangi]

@IonBazan could you take a look at the conflicts?

[IonBazan]

@darangi fixed 😉

[darangi]

Thanks for the contribution @IonBazan 🙇🏾