Given that most printable characters appear in consecutive runs in most popular character sets, it's easy to build a character class which only includes them. for example, for 7-bit ASCII you can write something like this to only print character between the space (0x20) and the tilde (0x7E).

while (<>) {
  s/[^ -~]//g;
  print;
}
[download]

Another way of doing that that usually benchmarks faster is a simple tr:

while (<>) {
  tr/ -~//cd;
}
[download]

-- Randal L. Schwartz, Perl hacker

Yeah, but I _always_ forget about the /c modifier to tr :(

--
<http://www.dave.org.uk>

European Perl Conference - Sept 22/24 2000, ICA, London
<http://www.yapc.org/Europe/>

Not wishing to split hairs with a luminary and abbot but doesn't the script remove the characters within the range?

Brother Frankus.

The ^ inside the character class makes it negative. So Dave's regex says every character which is not between a space and a tilde should be replaced with nothing.

Thanx for taking time to enlighten a muppet perler like me.

Brother Frankus.