>BTW, will '.' in mnemonic be a problem? The only thing I'm afraid of here (but that applied already when . was added to identifier_chars) is that in Intel syntax, . is an operator (although I think the support for this is currently at best partly working). Jan