Only ASCII 32 and '\t' are currently matched as Spacechar. ASCII 160 and other characters in the Unicode whitespace category should also be recognised as Spacechar tokens.
peg-markdown operates on strings of bytes, and knows nothing of unicode. Unicode nonbreaking space 160 is encoded in UTF-8 as two bytes. So there are some difficulties here.
Thanks for the peg upgrade. However, I don't agree that the nonbreaking space should be counted as a space character in peg-markdown. It is very handy to be able to insert spaces that should be treated just as normal spaces. For example, if you want a paragraph to start with four spaces, you can start with four 160's, and the paragraph won't be treated as a code block. Furthermore, spaces are treated as word boundaries in the writers, and lines may break on them