From: "Jason D. Clinton" <me@...>
Date: 2006-06-29T02:14:16+09:00
Subject: Re: remove all illegal chars form string

--=-HdN8U/UDiOdK1Bn5vKK0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Wed, 2006-06-28 at 22:09 +0900, thomas coopman wrote:
> Is there a simple way to remove all but the legal chars from a string.
> where the legal chars are for example: a-z A-Z 0-9
> So everything should be removed from the string but these characters.
> "exam@p  Le3|=C2=A7" --> "exampLe"
>=20
> I don't know very much about regular expressions, so I don't know if
> it's possible with sub or gsub.  My first Idea was to loop over the
> string and check every character but I wondered if there is something
> more simple or better.

An except from my upcoming book, Ruby Phrasebook:

"""
new_password =3D gets
if new_password.count '^A-Za-z._' !=3D 0 then
  puts "Bad Password"
else
  #do something
end

This works by using a special syntax that's shared by .count, .tr,
delete, and squeeze. A parameter beginning with ^ negates the list; the
list consists of any valid characters in the active character set and
may contain ranges formed with -. If more than one parameter list is
given to these functions, the lists of characters are intersected using
set logic[md]that is, only characters in both lists are used for
filtering.

You might also want to simply replace all "evil" characters with _ (such
as perhaps from a CGI form post):

evil_input =3D '`cat /etc/passwd`'=20

evil_input.tr('./\`', '_')

#=3D> "_cat _etc_passwd_"
"""

In your specific question, you will want to use .delete:

'exam@p  Le3|=C2=A7'.delete '^A-Za-z'
  #=3D> "exampLe"



--=-HdN8U/UDiOdK1Bn5vKK0
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQBEopdotSqjk42zvwkRAuvSAKCwr2yIpsj9q5dnUm8xhu+ZFUrFwgCfcbLE
4RfRlK7kGae/tbe7EhO2SdU=
=CfA1
-----END PGP SIGNATURE-----

--=-HdN8U/UDiOdK1Bn5vKK0--