Oops oops oops. Thanks for catching that--I guess I must not always use that regex. :)

But here's your one-liner:

$str = join "\n<p>\n", grep s/\n/<br>\n/g || 1, split /\n\n/, $str;
[download]

(The "|| 1" in there makes it so that even paragraphs that don't contain any carriage returns inside of them, and thus don't match in the substitution, still get included in the final list of paragraphs.)