Using subroutines could be something as simple as

sub test_encodings {
    my ($fh, $fn) = tempfile();
    $fh->print($file_string);
    $fh->close();

    is_deeply(
      [ Fancy::Open::fancy_open($fn) ],
      [ map encode('utf8', $_), @wanted_array ],
      "testing an array opened with no encoding"
    );

    my @encodings = qw(UTF-8 ascii cp1252 iso-8859-1);

    for my $encoding (@encodings) {
      is_deeply(
        [ Fancy::Open::fancy_open($fn, { 'encoding' => $encoding }) ],
        [ map encode($encoding, $_), @wanted_array ],
        "testing an array opened with $encoding encoding"
      );
    }
}

sub test_with_file_that_ends_with_a_newline {
    my ($fh, $fn) = tempfile();
    $fh->print("$file_string\n");
    $fh->close();

    is_deeply(
      [ Fancy::Open::fancy_open($fn) ],
      [ @wanted_array ],
      "testing a plain array with file that ends with a newline"
    );

    is_deeply(
      [ Fancy::Open::fancy_open($fn, { 'before' => 'solid ' }) ],
      [ @solid_array ],
      "testing an array with before option with file that ends with a 
+newline"
    );

    is_deeply(
      [ Fancy::Open::fancy_open($fn, { 'after' => ' bead' }) ],
      [ @bead_array ],
      "testing an array with after option with file that ends with a n
+ewline"
    );

    is_deeply(
      [ Fancy::Open::fancy_open($fn, { 'before' => 'solid ', 'after' =
+> ' bead' }) ],
      [ @solid_bead_array ],
      "testing an array with before and after options with file that e
+nds with a newline"
    );
}
[download]

As bonuses, the file name/handle variable names are uniform, and the section comments are made redundant by the sub names.

On second thought, "slurp" (in Perl land) means reading a file into a scalar, whereas you're reading into an array, so "fancy_read_lines" is still more accurate than the current name.

I didn't realise GitHub could automatically render .pod files - that is cool.