Yup, the t_cmp arguments were flipped a while back.
FWIW, I think whomever flipped the t_cmp arguments but didn't flip the included test cases at the same time needs a stern talking to. I spent over an hour and a half figuring out why the heck httpd was returning a 200 in that case where a 413 was clearly (or at least more) correct: only to find out that the debug output was swapped. Incredibly, incredibly lame.
All I wanted to do last night was add some caching tests: instead I had to fix our tests to pass at all. *sigh* -- justin
