in reply to The Relation Between Lexers and Parsers

This is a great question. (Well, in my mind, any question that I've struggled with is a great question :)

FWIW, here is a working lexer I wrote for use with Parse::Yapp, for a simple OO class-specification/templating language... if you have any questions about it, let me know:

sub Lexer { my $parser = shift; defined $parser->YYData->{INPUT} or return('',undef); if (not defined $parser->YYData->{LINE}) { $parser->YYData->{LINE} = 0; $parser->YYData->{SLINE} = 0; $parser->YYData->{ADDLINES} = 1; } $parser->YYData->{INPUT} =~ s/^[ \t\r]*//; # remove whitespace $parser->YYData->{INPUT} =~ s/^#[^\n]*//; # remove comments for ($parser->YYData->{INPUT}) { s/^([\n{}:])// and do { return($1,$1); # character tokens }; s/^(\w+)\b// and do { if (exists $modifiers{$1}) { return('MODIFIER', $1); } elsif (exists $keywords{$1}) { return('KEYWORD',$1); } else { return('IDENT', $1); } }; s/^('(.*?)')// and do { return('STRLIT',$2); }; s/^("(.*?)")// and do { return('STRLIT',$2); }; s/^(<{(.*?)}>)//s and do { my $heredoc = $2; $parser->YYData->{ADDLINES} += () = $heredoc =~ m/\n/g; return('HEREDOC',$heredoc); # woohoo i love heredoc's! }; } }