At 09:58 AM 11/29/00 -0800, Steve Fink wrote: >As soon as the lexer sees "s#", it >starts treating # as a delimiter -- it doesn't need to conditionally >treat the # as either a delimiter or a comment. (Especially since >there's nothing following it that could resolve the ambiguity!) One thing I've been thinking about is setting up a host of symbolic representations of things like comment characters and end-of-statement delimiters and things. That way, rather than hardcoding # as the comment character (which it sometimes isn't) we look for the generic "comment sequence". Makes things like q## more generically handleable, and it also means that if we put in, say, c_mode.pm, it can set the close-comment sequence to */. It seems we'll need to do is weight various lexical constructs, so that the open-quote sequence is checked for before the comment sequence. (Unless we want to do some sort of probabalistic thing, or turn the whole parser into a big neural net or something...) Dan --------------------------------------"it's like this"------------------- Dan Sugalski even samurai dan@sidhe.org have teddy bears and even teddy bears get drunkThread Previous | Thread Next