If it would be unreliable, it should fail sometimes - but it doesn't, ....

I am pretty sure that Toby meant "unreliable" as in "the return value depends on whether that file had already been required before".

Let me expand his example a bit:

use strict;
use warnings;
use autodie;
use File::Temp qw/ tempfile /;
use Data::Dump;

my ($fh, $filename) = tempfile();

print $fh qq('Hello, world';\n);
close $fh;

my $first = require $filename;
dd $first;

my $second = require $filename;
dd $second;
[download]

Output:

"Hello, world"
1

This is all stated in the documentation for require, inclusive the role of %INC:

if (exists $INC{$filename}) {
    return 1 if $INC{$filename};
    croak "Compilation failed in require";
}
[download]

To import a fle more than once you have to delete from %INC of course.

perl -le'print map{pack c,($-++?1:13)+ord}split//,ESEL'

> To import a fle more than once you have to delete from %INC of course.

Or use a different path to it (I've seen this in production, don't ask).

My.pm:

package My;
use warnings;
use strict;

our $counter //= 10;
++$counter
[download]

script.pl

#!/usr/bin/perl
use warnings;
use strict;
use feature qw{ say };
use lib './';

say require My;         # 11
say require 'My.pm';    # In %INC, returns 1.
say require './My.pm';  # Not in %INC, thus 12!
[download]

map{substr$_->[0],$_->[1]||0,1}[\*||{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^*ARGV,3]