S-Expression Comments can get Windows/DOS carriage returns embedded in them

[tealeg]

In the event that someone submits a regula rule that includes Windows/DOS line endings. We'll cope with it without error (hopefully). However, the lexer's scanComment function will currently break on \n and read the preceding \r into the comment string. This doesn't cause any practical problems, but still, it's not quite the right behaviour. We should, instead, explicitly ignore the \r character when lexing a comment.