Fix highlighting for non-decimal numbers with underscores

[smercer10]

Resolves #71

Made some minor changes to the match regexes for binary, octal, and hexadecimal numbers. The syntax highlighting should now be indicative of valid Gleam number formats:

• A single underscore in the middle of a number (including between the base prefix and first digit) ✅
• More than one contiguous underscore ❌
• An underscore at the start or end of a number ❌

Before

After

[smercer10]

Is it also worth supporting numbers with uppercase base prefixes? E.g., 0B, 0O, 0X. Although the formatter automatically converts them to the lowercase variants, perhaps in the intermediate stage it's worth indicating (through highlighting) that they're valid number formats and will compile if left as is?

[PastMoments]

Is it also worth supporting numbers with uppercase base prefixes? E.g., 0B, 0O, 0X. Although the formatter automatically converts them to the lowercase variants, perhaps in the intermediate stage it's worth indicating (through highlighting) that they're valid number formats and will compile if left as is?

I think it's worth it, also because there may be workflows that run the formatter very infrequently

[smercer10]

Is it also worth supporting numbers with uppercase base prefixes? E.g., 0B, 0O, 0X. Although the formatter automatically converts them to the lowercase variants, perhaps in the intermediate stage it's worth indicating (through highlighting) that they're valid number formats and will compile if left as is?

For now I've made it so that numbers with uppercase base prefixes are highlighted as valid numbers, but will happily revert if there are strong reasons against it. I think if the compiler supports it, then syntax highlighting should reflect that, even if it's not particularly idiomatic.

Before

After

[lpil]

Thank you!!