Your example may be a bit more abstract than you intended, but, in case it is not, why not just test the array?
foreach (@kabluther) {
#...
}
if (@kabluther) {
# the "..." above was executed
# i.e., your $x was modified
$_->change_in_x() for @listeners; # sounds like this is something li
+ke what you want to do.
}
Note that this test is executed once, and only once, each time through this function. The test should be O(1) - that is, independent of the actual size of the list. Even if perl didn't keep track of the actual size of the list separate from the list itself (i.e., not a calculation), which it does, perl would be able to optimise this by noting we're in boolean context and knowing to just check if the list is empty or not, and thereby only count the first element in the list.
Note that this does not (quite) help if you're doing the $x modification inside an "if", e.g.:
foreach (@kabluther) {
$x += $_->getSkookiness() if $_; # ignore undef arguments
# or if ref $_; # only count ref's
# or if ref $_ and $_->can('getSkookiness'); # only count objects t
+hat have this method
}
But, even then, based on your original question ('usually empty'), and your fuzzy operator ('can report true when there is no actual change, but cannot report false when there is a change'), I think this may be close enough.
For more accuracy, I would just create a lexical variable to track changes.
sub onTick {
our $x;
my $changed;
# Initialize and associate $x with some external value
foreach (@kabluther) {
do { # these two lines must be a block - do together, or not at
+all.
$x += $_->getSkookiness();
$changed++;
} # if clause can go here I think, if needed
}
if ($changed) {
$_->ticked() for @listeners;
}
return $x;
}
A bit more overhead when doing something, but about the same overhead when doing nothing, and will get it right all the time, and no funky XS. That said, it requires a change to the loop. Probably not a huge change, but a change nonetheless.