Parser is unaware of system-dependent newlines

[mmichaelis]

The parser is unaware of system-dependent newlines such as CRLF (Windows) or CR (classic MacOS).

My suggestion is, to document this behavior, rather than to change the whole processing. Anyone having issues with that (like we) could then just pre-process the BBCode to normalize newlines.

Test matrix:

Input	Parsed
A\nB	["A", "\n", "B"]
A\r\nB	["A\r", "\n", "B"]
A\rB	["A\rB"]

If instead supporting other line separators, helper methods such as isEOL must be adapted, too.