Fix utf8 display width

[Satyakam-Mishra]

Description

Fixes incorrect display width calculation for UTF-8 characters in formatted
strings.

Characters with East Asian Width properties W and F are treated as
double-width when calculating display width. The implementation uses a binary
search over a generated Unicode range table.

This fixes alignment issues for characters such as emoji and CJK characters.

Regional Indicator symbols are not included as double-width ranges, so flag
sequences such as 🇺🇸 are handled correctly.

Testing

Added regression tests covering:

• ❌
• ✅
• 😀
• 🇺🇸
• Müller
• 我

All tests pass:

• 21/21 tests passed
• git diff --check passes

[vitaut]

Please don't add new headers.

[Satyakam-Mishra]

Thank you, I have removed the separate unicode_width_table.h header and moved the unicode with table to format.h as suggested. I have also rerun the tests and it still passes. Pushing the update now.

[vitaut]

Closing in favor of #4896 but thanks anyway.