Here's the second shortest one I could come up with:

#!/usr/bin/perl -l

print for map { $_->[0] }
    sort { $a->[1] cmp $b->[1] }
    map { [$_, ( /blahblah/ ? 0 : 1 ) . $_] }
    <>;
[download]

It uses a Schwartzian Transform that compares the second elements of an array looking like this:

(
    ["bat 1 blahblah", "1bat 1 blahblah"],
    ["foo 2 blahblah", "1foo 2 blahblah"],
    ["bar 1 zot", "0bar 1 zot"],
    ["bar 2 zot", "0bar 2 zot"]
)
[download]

The lines that don't match the regex get prepended with a 0 and the ones that do match get a 1. That way the matching lines always win a cmp with the non-matching ones.

Update: Fixed a typo in the array. I had written "0bat 1 blahblah" and "0foo 2 blahblah"

Here's a shorter, faster one using substr instead of arrays:

#!/usr/bin/perl -l

print for map { substr($_, 1) }
    sort
    map { ( /blahblah/ ? 0 : 1 ) . $_ }
    <>;
[download]

-Matt