Extract String#to_i impl to scolapasta-int-parse

[lopopolo]

Reuse the radix type. Add tests.

From what I can tell, ASCII whitespace is trimmed from front and back like Kernel#Integer.

[3.1.2] > "  \t\n\u000A\u000C            123abc                \t\n\u000A\u000C       ".to_i(16)
=> 1194684
[3.1.2] > "  \t\n\u000A\u000C            123abc                \t\n\u000A\u000C       ".to_i
=> 123

0x-style prefixes are allowed but only if they match the given base or x is in the given base's alphabet. If the given base does not match the 0x-style prefix, 0 is returned.

[3.1.2] > "0x123abc".to_i
=> 0
[3.1.2] > "0x123abc".to_i(16)
=> 1194684
[3.1.2] > "0x123abc".to_i(24)
=> 0
[3.1.2] > "0x123abc".to_i(36)
=> 71897795832

Numbers are parsed from the front and parsing stops when encountering an ASCII alphanumeric not in the given base's char set:

[3.1.2] > "123abc".to_i
=> 123
[3.1.2] > "123abc".to_i(16)
=> 1194684