One might want to use Unicode::UTF8[1] instead of the hand-rolled helper:
#!/usr/bin/perl
use strict;
use warnings FATAL => 'utf8'
use open ':std', ':encoding(UTF-8)';
use Time::Piece;
use Unicode::UTF8 qw( decode_utf8 );
for my $month (1 .. 12) {
my $date = sprintf '2021-%02d-01', $month;
my $tp = Time::Piece->strptime($date, '%Y-%m-%d');
print decode_utf8($tp->strftime('%B')), ' ' x ($month != 12);
}
print "\n";
> Here is a summary of features for comparison with Encode's UTF-8 implementation:
> * Simple API which makes use of Perl's standard warning categories.
> * Recognizes all noncharacters regardless of Perl version
> * Implements Unicode's recommended practice for using U+FFFD.
> * Better diagnostics in warning messages
> * Detects and reports inconsistency in Perl's internal representation of wide characters (UTF-X)
> * Preserves taintedness of decoded $octets or encoded $string
2 comments
[ 3.8 ms ] story [ 17.5 ms ] thread> * Simple API which makes use of Perl's standard warning categories.
> * Recognizes all noncharacters regardless of Perl version
> * Implements Unicode's recommended practice for using U+FFFD.
> * Better diagnostics in warning messages
> * Detects and reports inconsistency in Perl's internal representation of wide characters (UTF-X)
> * Preserves taintedness of decoded $octets or encoded $string
> * Better performance ~ 600% - 1200% (JA: 600%, AR: 700%, SV: 900%, EN: 1200%, see benchmarks directory in git repository)
[1]: https://metacpan.org/pod/Unicode::UTF8