At 10:48 AM 4/25/2001, you wrote:
> > I believe if you add the (g)lobal modifier and optionally(i), Paul's
> > line of code may work:
> >
> > $line =~ /(\b\w+\b).*\1/ogi;
You don't need /o, because there are no variables in the pattern, so Perl
will compile the regex at compile time anyway. I guess it doesn't hurt,
though.
I looked through some stuff for items like this, and found something in the
Perl Cookbook that almost fit the bill. I modified it *very* slightly so
that it can catch punctuation between words. Take a look at the comments
and you'll know what to pull out if you don't want that behavior.
And before anyone goes thinking I'm a regex guru, I'm not. I really
literally copied 99% of this out of the Perl Cookbook. (BTW, the three
guys who wrote it are *really* good people to copy off of. :-) ) I know
Perl kinda well, but of all it's features, I'm weakest on re's.
Anyway, here it be.
while (<>) {
while ( m{
\b # Start at a word boundary
(\S+) # \S means non-whitespace who's
# beginning and end are alphanums
# if you put \b's around it.
\b # until you hit another boundary
\W? # match a non-word char
(
\s+ # Then some space
\1 # Then what we matched in (\S+) above
\b # with a boundary after it
)
}xig # eXtended (Comments), case Insensitive
# Global
)
{
print "Dupe word '$1' at line $.\n";
}
}
Thank you for your time,
Sean.
