Regex for weird characters

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Regex for weird characters by DrHyde (Prior) on Sep 27, 2004 at 14:56 UTC
The easiest way is not to try to look for all the weird characters, but instead to look for only the permitted characters and invert the test. If, for example, you only permit letters, numbers and commas, you would do something like ... `if($text !~ /^[a-z0-9,]$/i) { die "bad characters\n"; }` [download] Doing it this way ensures that you only permit what you want. This is good practice, because you won't accidentally let bad stuff through. If you try to think of all the weird characters that someone can type you're bound* to forget some of them. For example, you're probably thinking of banning é, è, ñ and ç. But had you remembered æ, œ, ß, ð, ø, å, ł and þ?	[reply] [d/l]
Re^2: Regex for weird characters by Skeeve (Parson) on Sep 27, 2004 at 15:06 UTC
The idea is okay, but why match the whole string if it's sufficiant to fail as soon as one character fails? So instead of `$text !~ /^[a-z0-9,]$/i` [download] this should give the same result: `$text =~ /[^a-z0-9,]/i` [download] `$\=~s;s.*;q^\|D9JYJ^^qq^\//\\\///^;ex;print`	[reply] [d/l] [select]
Re^3: Regex for weird characters by kutsu (Priest) on Sep 27, 2004 at 15:37 UTC
you could shorten Skeeve's regex further using character classes, acutally it's the same length cause I added spaces ;) `$text =~ /[^\w\d\s,]/;` [download] Update: switched /w for \w (++ to graff for pointing out my typo) "Cogito cogito ergo cogito sum - I think that I think, therefore I think that I am." Ambrose Bierce	[reply] [d/l]
Re^4: Regex for weird characters by DrHyde (Prior) on Sep 28, 2004 at 09:00 UTC
Re^2: Regex for weird characters by sulfericacid (Deacon) on Sep 27, 2004 at 15:23 UTC
This goes with what DrHyde was showing with as a bunch of the normal characters you're looking for. It might take more characers depending on what else you want to add, but here you go. `if($text !~ m/^[a-z0-9,!,@,#,$,%,^,&,,9,\,,\.\?,\~,:,\+,\-, ,\"]$/i)` [download] "Age is nothing more than an inaccurate number bestowed upon us at birth as just another means for others to judge and classify us" sulfericacid	[reply] [d/l]
Re^3: Regex for weird characters by !1 (Hermit) on Sep 27, 2004 at 15:54 UTC
Why do you have so many commas in your character class?	[reply]
Re^4: Regex for weird characters by sulfericacid (Deacon) on Sep 27, 2004 at 16:02 UTC
Re^5: Regex for weird characters by !1 (Hermit) on Sep 27, 2004 at 16:16 UTC
Some notes below your chosen depth have not been shown here
Re: Regex for weird characters by sgifford (Prior) on Sep 27, 2004 at 15:43 UTC
If what you want to allow is all printable ASCII characters, you can use regex character classes to specify that a string containing any non-ASCII characters or control characters is "funny": `print "funny" if ($c =~ /[[:^ascii:][:cntrl:]]/);` [download] This is equivalent to checking that no character has an integer value less than or equal to 31 (1f hex), or greater than or equal to 127 (7f hex): `print "funny" if ($c =~ /[\x00-\x1f\x7f-\xff]/s);` [download] See perlre(1) for more details. Both of these solutions disallow newlines and tabs; you can allow them with the `[:isspace:]` character class.	[reply] [d/l] [select]
Re: Regex for weird characters by Skeeve (Parson) on Sep 27, 2004 at 14:32 UTC
Just to get you started... `/[^a-zA-Z0-9_]/ && print "funny things appear in $_\n" for qw/not_funn +y funny1-/;` [download]	[reply] [d/l]
Re^2: Regex for weird characters by JediWizard (Deacon) on Sep 27, 2004 at 14:46 UTC
`/[^\w\x21-\x26\x28-\x2A\x5E\x40]/ && print "funny things appear in $_\n" for qw/not_funny funny1-/;` But that still doesn't account for things like ? or > or <. I'm taking you too litterally when you say shitf-# keys... :-) May the Force be with you	[reply] [d/l]