Other Language - A problem with Lex

Hi,

I have some problem with Lex

if I write 1 + 1 I'm getting:
Found token INT ('1')
Found token OP ('+')
Found token INT ('1')

If I write 1+1 I'm getting:
Found token INT ('1')
Found token INT ('+1')

+1 is part of the language, what do I need to change so if I will write 1+1 I will get same as if I will write 1 + 1?

this is my regular expressions:
SIGN [+-]
DIGIT [0-9]
DIGIT_NZ [1-9]
INTEGER ([0])|(({SIGN}{0,1})({DIGIT_NZ}{1})({DIGIT}*))

Thanks

Sounds like you need to ignore whitespace when you parse the string.

You should not parse the sign with INTEGER and resolve this with your parser by having rule like

NUMBER ::= INTEGER | ADDOP INTEGER

Lex only deals with tokens in ignorance of the surrounding context. It can't tell when a '-' is a subtraction operator vs a unary negation (well you could have your parser screw around with the lexer mid-parse like perl does, but I don't even if that's possible in lex and in any case it would be horribly complicated).