tiny_psnr: fix range calculation for sample size of 32 bits
authorMans Rullgard <mans@mansr.com>
Sun, 14 Oct 2012 23:21:53 +0000 (00:21 +0100)
committerMans Rullgard <mans@mansr.com>
Mon, 15 Oct 2012 20:41:25 +0000 (21:41 +0100)
commitbf868c4a9bc6c1ac4c4753822a7c652b7f7c8ff7
tree1169250ef2fe42130023a1db8d49dd7a5c331787
parent3dc06b6972cf389269e9c36ff0a4373f80f7149b
tiny_psnr: fix range calculation for sample size of 32 bits

For a sample size of 32 bits, the shift would overflow producing
undefined results.  Incidentally, in the only test currently using
32-bit samples, the output matches the reference exactly on most
systems meaning the bad 'max' value is never used.

Signed-off-by: Mans Rullgard <mans@mansr.com>
tests/tiny_psnr.c