Shell Programming and Scripting

Originally Posted by jimmy_y

Thanks Kevin,

But when i use:
#!/usr/bin/perl
$a = "a";
#if ($a =~ s/[A-Za-z0-9_-.]//g) {
if ($a =~ s/[A-Za-z0-9_\-.]//g) {
print "invalid file name\n";
} else {
print "valid file name\n";
}

i got
invalid file name

Originally Posted by pludi

First, please use [code][/code] tags for sources, code/output listings, ...
Second, [A-Za-z0-9_] has it's own character class in Perl, \w
Third, if you want to match a literal dot, escape it, otherwise it'll match any character
Fourth, if you want to match something, don't run substitution on it.
Fifth, use appropriate logic. In your example the regex matches, but your logic says it's invalid.

This will work:

Code:

#!/usr/bin/perl
$a = "a";
if ( $a =~ /[-\w\.]/g ) {
    print "valid file name\n";
}
else {
    print "invalid file name\n";
}

Originally Posted by KevinADC

Of course that's what you get. Your script is saying:

Code:

if (we substitute one of these characters for nothing) {
   then the filename is invalid
{
else {
   it is valid
}

Since $a has an 'a' character in it, the condition is true because the regexp substitutes the 'a' for nothing then enters the block and tells you the filename is invalid. I think what you are trying to do is validate that a filename only has valid characters which would be done like this:

Code:

#!/usr/bin/perl
$a = "ABCtest-12_3.txt";
if ($a =~ /^[\w.-]+$/) {
        print "valid file name\n";
} else {
        print "invalid file name\n";
}

Although that is not very good validation for a filename since it will pass a single dot or single dash. But that is the basics.

---------- Post updated at 01:49 PM ---------- Previous update was at 01:46 PM ----------




The dot in a character class is just a dot, it is not a wild card match and does not need escaping inside a character class. Test it for yourself.