But when you actually use it, you have /$regex/, so $regex gets interpolated into the pattern you said it was. The dollar signs are end-of-string indicators. What you really want is to eval it.

sub findString($) {
    my $line = shift;
    print "found\n" if $line =~ eval "qr/$regex/";
}
[download]

You can also use the experimental delayed-evaluation feature, though it still requires the variables to be in scope:

my $s = 'Tue Feb  8 11:11:11 2005: blah blah blah';

my ($wday, $mon, $mday, $year);

my $regex = qr/^(??{$wday}) (??{$mon})\s+(??{$mday}) \d{2}:\d{2}:\d{2}
+ (??{$year}):/;
$wday = 'Tue';
$mon = 'Feb';
$mday = '8';
$year = '2005';

sub findString;
findString($s);

sub findString($) {
    my $line = shift;

    print "found\n" if $line =~ /$regex/;
}
[download]

Okay, utilizing the code suggested above, I get the expected behaviour. If I change the if-statement to the following, I get the errors noted further below.

#!/usr/bin/perl

my $s = 'Tue Feb  8 11:11:11 2005: blah blah blah';

my $regex = '^$wday $mon\s+$mday \d{2}:\d{2}:\d{2} $year:';
my $wday = 'Tue';
my $mon = 'Feb';
my $mday = '8';
my $year = '2005';

findString($s);

sub findString($) {
    my $line = shift;

#   print "found\n" if $line =~ eval "qr/$regex/";
    if ($line =~ eval "qr/$regex/") {
        print "found\n";
    }
}
[download]

Do the following errors make sense? Yes, I'm using Perl 5.005, but I cannot change this at this point.

Scalar found where operator expected at (eval 1) line 1, at end of lin
+e
        (Missing operator before ?)
Backslash found where operator expected at (eval 1) line 1, near "$mon
+\"
        (Missing operator before \?)
found
[download]

Placing curly braces around the qr() string pacifies the interpreter, but I no longer see the expected answer. Thanks for any insight.

The errors don't make any sense to me, given that it works in the non-block if and only fails with block-form if. Does this work any better? Does it fail before the if?

sub findString($) {
    my $line = shift;
    my $compiled_regex = eval "qr/$regex/";

    if ($line =~ /$compiled_regex/) {
        print "found\n";
    }
}
[download]

---

Caution: Contents may have been coded under pressure.

Thanks for your suggestions. In regards to your first answer using eval, isn't the e switch to regular expressions provide the same functionality?

There is no e switch for regular expressions. The e switch is for substitutions, and applies to the replacement, not to the pattern.

---

Caution: Contents may have been coded under pressure.

$regex = eval $regex;
[download]

my $s = 'Tue Feb  8 11:11:11 2005: blah blah blah';

my $regex = "^%s %s\s+%s \d{2}:\d{2}:\d{2} %s:';

my $wday = 'Tue';
my $mon = 'Feb';
my $mday = '8';
my $year = '2005';

findString($s, $wday, $mon, $mday, $year);

sub findString {
  my $line = shift;
  my $pattern = sprintf($regex, @_);
  print "found\n" if $line =~ /$pattern/;
}
[download]

And if the code you posted is merely an attempt to show us concisely your desire, please ignore this... but if it's reflective of your actual plan of attack, the reference to "dynamically" suggests (to me, anyway) that you need to think thru the meaning of "dynamically" or maybe provide a different term which reflects your actual intent.

The elements of the regex -- $wday, $mon, $mday and $year -- are hard coded here. Only if you actually intend to get their values from STDIN or somesuch, this would qualify (in my book, which may not be yours) as "dynamic."

and suggest perlretut and perlrequick for more on regexen.

my $s = 'Tue Feb  8 11:11:11 2005: blah blah blah';

my $regex = {
  'wday' => qr{^\s*(\w+)\s+},
  'mon'  => qr{^\w+\s+(\w+)},
  'mday' => qr{\s+(\d{1,2})\s+},
  'year' => qr{\s+(\d{4})},
  'time' => qr{((?:\d+:){2}\d+)}
};


my %dt = findString($s);

print "$dt{year} $dt{time} $dt{mday}"; #...

sub findString($) {
    my $line = shift;
    my %dt; # date/time
    for(keys %$regex){
      $line =~ /$regex->{$_}/ and $dt{$_} = $1
    }
    #print "found year\n" if exists$dt{'year'}
    #or just return hash
    %dt
}
[download]

my $s='Tue Feb  8 11:11:11 2005: blah blah blah';
my $regex=qr(^$wday $mon\s+$mday \d{2}:\d{2}:\d{2} $year:);

findstring($s);

sub findstring($)
{
  my $line=shift;

  print "found\n" if $line=~/$regex/;
}
[download]

# interpolate the variables (you have to define your
# variables to interpolate first so they work).
my $regex="^$wday $mon\s+$mday \d{2}:\d{2}:\d{2} $year:";

$regex=qr($regex); # compile the regular expression
[download]

my $s='Tue Feb  8 11:11:11 2005: blah blah blah';
my $regex=createRegEx('Tue','Feb','8','2005');

findstring($s);

sub findstring($)
{
  my $line=shift;

  print "found\n" if $line=~/$regex/;
}
sub createRegEx($$$$)
{
  my ($wday,$mon,$mday,$year)=@_;

  my $regex="^$wday $mon\s+$mday \d{2}:\d{2}:\d{2} $year:";

  return(qr($regex));
}
[download]