Tue Mar 31 17:40:16 CEST 2009


I think I'm just going to remove the lex/yacc parser.  Anything I want
to do with syntax in ramblings files is going to be completely ad-hoc:
datamining based on existing text files instead of any kind of sane
structure.  So let's just use a direct tokenizer.

What the current lexer produces is a list of strings and xhtml
elements in xexpr form.

OK: replaced with whitespace/workd tokenizer state machine +
individual word matcher.