add support for binary (0b[0|1...]) and octal (0o[0-7...]) integers

[tinygiant98]

This PR is an answer to a question nobody asked and resolves no known problems. This change allows the use of binary (0[b|B]) and octal (0[o|O]) integers in nwscript. Additionally, it adds a check for the maximum length of the token based on type (8 characters after 0x for hexadecimal, 31 characters after 0b for binary, and 11 characters after 0o for octal).

For the octal integers, I went with the 0o prefix instead of the potentially more traditional numbers-that-begin-with-0 because numbers that begin with 0 are currently valid integers in nwscript and I wanted to ensure backward compatibility.

The whole token-length checking may be superfluous as normal integers are not checked. Binary and Octal code has been kept in their own blocks instead of integrated into hex code blocks to make future modification or removal easier.

Testing

Successfully tested against the following constructs: int h = [+-]0[xX][A-Fa-f0-9...] int b = [+-]0[bB][01...] int o = [+-]0[oO][0-7...]

Example tests: my_function(h) my_function(b + o) int x = 0xff04ab + 0b000000100010001 + -0o03372 my_function(0xff04ab + 0b000000100010001 + -0o03372) if(h) if(0xff04ab) while(b) while(0xff04ab) my_function(++a)

Errors: Compiler will error with UNEXPECTED CHARACTER if a hex number includes characters other than [A-Fa-f0-9], a binary number includes characters other than [01], or an octal number contains characters other than [0-7]. The error TOKEN TOO LONG will be issued if a hex integer has more than 8 digits after 0x, a binary integer has more than 31 characters after 0b, or an octal integer has more than 11 characters after 0o.

Changelog

Added

• Support for binary (prefix 0b) and octal (prefix 0o) integers.
• Check for character length after prefix for hexadecimal, binary and octal integers.

Licence

• [x] I am licencing my change under the project's MIT licence, including all changes to GPL-3.0 licenced parts of the codebase.