perl: how does `$1` behave in a loop? - regex

Question

I'm getting some behavior I don't understand in perl:
>>> my #words = ('hello', 'there');
>>> $words[0] =~ /(el)/; print $1;
el
>>> $words[1] =~ /(el)/; print $1;
undef
but in a loop:
>>> my #words = ('hello', 'there');
>>> foreach my $word (#words){
>>> $word =~ /(el)/;
>>> print "$1\n";
>>> }
el
el
What's going on here? And how can I, in the loop, get $1 to be undefined when it doesn't match in the most recent regex, so that something like this will work:
foreach my $word (#words) {
$word =~ /(el)/;
if ($1) {
print "$word matched\n";
} else {
print "$word did not match\n";
}
}

Answer 1

There's nothing special about the loop.
use strict;
use warnings;
use feature qw( say );
my #words = ('hello', 'there');
$words[0] =~ /(el)/; say $1 // "[undef]";
$words[1] =~ /(el)/; say $1 // "[undef]";
my #words = ('hello', 'there');
foreach my $word (#words){
$word =~ /(el)/;
say $1 // "[undef]";
}
Output:
el
el
el
el
$1 and friends are only changed on a successful match, so you want
for my $word (#words) {
if ( $word =~ /el/ ) {
print "$word matched\n";
} else {
print "$word did not match\n";
}
}

Answer 2

One way would be to avoid the special numbered variables entirely, which don't get reset in between runs. Instead, use a local variable, and reset it at the beginning of each loop:
use warnings;
use strict;
my #words = qw(
one
two
three
);
for my $w (#words){
my $store;
if (($store = $w) =~ /(t)/){
print "$store\n";
}
else {
print "no match\n";
}
}
Output:
two
three

Answer 3

Check the return of the match:
if ($word =~ /(el)/) {
print "$word matched with [$1]\n";
} else {
print "$word did not match\n";
}
I suspect your test environment is doing a bit more when you run them separately than your script does, including reseting $1, among other things.