weird bug in union I think

gistfile1.txt

.:15:24:31:. <geoff> which, I had an interesting time meshing that union with correlated_resultset, 
                     incidentally
.:15:24:41:. frew I made a thing that will correctly sort on a correlated column when the user asks to
.:15:24:51:. frew oh?
.:15:24:52:. frew why?
.:15:24:53:. <geoff> oooo
.:15:24:58:. <geoff> I just had to do
.:15:25:05:. <geoff> $rs1->with_correlation
.:15:25:10:. <geoff> and $rs2->with_correlation
.:15:25:13:. <geoff> then union
.:15:25:14:. <geoff> rather than
.:15:25:16:. <geoff> union
.:15:25:22:. <geoff> $rs3->with_correlation
.:15:25:23:. frew huh
.:15:25:28:. frew I wonder why?
.:15:25:33:. frew I mean, I'm sure you're right
.:15:25:36:. <geoff> the later only generated the subquery
.:15:25:37:. frew but I *think* both should work
.:15:25:39:. <geoff> on the first one
.:15:25:45:. frew huh...
.:15:25:47:. <geoff> so obviously the union then fails
.:15:25:54:. frew I'm gonna copy paste this to myself
.:15:25:57:. frew I think you found a bug
.:15:25:58:. <geoff> that might be a small bug