Question

I need to find identifiers in a text file:

But I don’t want a match if the identifier is a keyword. For example, if I have “for” as a keyword, in the following:

for (i=0 ; i< max ; i++)

I should get:

Found: i
Found: i
Found: max
Found: i

I looked into look-ahead assertion, but I wan’t able to make it work:

$IDENTIFIER="(?!(for|while|do))[a-zA-Z_]+[a-zA-Z0-9_]*"
    while ($entireFile =~ /($IDENTIFIER)/g)
    {
          print "Found ($1)" . "\n"; 
    } 

I get:

Found: or
Found: i
Found: i
Found: max
Found: i

This is not quiet what I want! I do understand why I get “or”, but how can I make it smarter and exclude “for” entirely?

Answer 1

You need anchoring to make sure that you’re matching an entire word (potential identifier). To a first approximation, /\b(?!(?:for|while|do)\b)[A-Za-z_][A-Za-z0-9_]*\b/ actually does what you want.