Challenge 2: The $@ Capture Pattern — Possible Solution
====================================================================

use strict;
use warnings;
use feature 'say';

eval {
    die "Something broke\n";
};
my $err = $@;

if ($err) {
    say "Caught: $err";
}

# Capturing $@ into my $err immediately, right after the eval block,
# is safer than checking $@ directly later because $@ is a GLOBAL
# variable that anything running between the eval and a later check
# — another unrelated eval, or a DESTROY method firing during garbage
# collection — could silently overwrite in the meantime.

Output:
Caught: Something broke

WHY THIS WORKS AS AN ANSWER
------------------------------
eval { die "..."; }; reuses the chapter's own classic eval/$@ pattern
directly — the die inside the block is caught by eval rather than
terminating the script, and $@ becomes populated with the message.

my $err = $@; immediately after the eval block reuses the chapter's
own recommended safety practice exactly — copying $@'s current value
into a fresh, ordinary lexical variable right away, before anything
else has a chance to run and potentially clobber the global $@ in the
meantime.

The one-sentence explanation reuses the chapter's own warn-box
reasoning almost word for word — "$@ is a global, and anything that
runs between your eval and your check of $@... can silently overwrite
it" — restated here specifically to answer why the my $err = $@;
line matters, rather than just checking if ($@) { ... } directly
several lines later, where more code (and more chances for $@ to
change) could have run in between.