Challenge 2: A Second Reference Format — Possible Solution
====================================================================

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

my $eml_body = "<b>Reference Number</b>: JO-2601-11029";

(my $plain_body = $eml_body) =~ s/<[^>]+>//g;

if ($plain_body =~ /Reference(?: Number)?:\s*(JO-\d+-\d+)/) {
    my $job_ref = $1;
    say "Found: $job_ref";
}

Output:
Found: JO-2601-11029

WHY THIS WORKS AS AN ANSWER
------------------------------
(my $plain_body = $eml_body) =~ s/<[^>]+>//g; reuses the chapter's own
strip-first approach exactly, unchanged — it doesn't matter that this
example uses <b> tags instead of <strong>, since s/<[^>]+>//g matches
ANY HTML tag generically (any characters between < and >), exactly
the chapter's own point that this approach avoids needing to guess
which specific tag might appear.

The label text here is "Reference Number" rather than "Job
Reference" — a genuinely different wording than the chapter's own
example. The pattern is adjusted to Reference(?: Number)? to match
this challenge's specific label, reusing the chapter's own (?: ...)?
non-capturing-optional-group technique, just moved to make "Number"
optional after "Reference" instead of after "Job Reference".

\s*(JO-\d+-\d+) is reused completely unchanged from the chapter's own
pattern — the same requirement of one-or-more digits in each group,
correctly matching "2601" and "11029" here exactly as it matched
"2405" and "98673" in the chapter's own worked example, proving the
capture-group portion of the pattern generalizes across different
label wording without needing any changes itself.