What is in your  @Goodlines array? The code posted by stevieb seems to work exactly as advertised:

c:\@Work\Perl>perl -wMstrict -MData::Dump -le
"my @lines = ('', ' ', '  ', 'x', ' x', 'x ', ' x ', '  x  ',);
 ;;
 my @no_empties = grep { $_ !~ /^(?:\s+|)$/ } @lines;
 ;;
 dd \@no_empties;
"
["x", " x", "x ", " x ", "  x  "]
[download]

Incidentally, I would prefer an inverse approach to the regex: a non-blank line is one that has at least one non-blank (i.e., non-whitespace, or \S) character in it:

c:\@Work\Perl>perl -wMstrict -MData::Dump -le
"my @lines = ('', ' ', '  ', 'x', ' x', 'x ', ' x ', '  x  ',);
 ;;
 my @no_empties = grep m{\S}xms, @lines;
 ;;
 dd \@no_empties;
"
["x", " x", "x ", " x ", "  x  "]
[download]

I humbly refer you to my previous reply which solves namelessjoe's problem. It works because the @goodlines array is empty on those occasions where the blank lines are output.

In short, there's no point trying to remove some entries from the array which are themselves empty or just whitespace because there are no entries in the array. The thread title is a bit misleading in that respect. I initially started down the route which you and stevieb took, but a quick test determined that the output would not match what was reported in the original node. Hence the conclusion/supposition that the "bad" lines were ones where the array was entirely empty.

Hope this helps and that it avoids anybody spending more time on blind alleys.

Nice observation. Usually 'good post' type posts aren't presented here, but you had a good eye to suit the OP with what they were specifically asking.

Short-circuiting definitely works better than iteration and removal (in the specific case presented).

-stevieb