want regression test case for EQUALALL

[masinter]

          I wish there were a test case -- just a simple "make three  big bitmaps, the second a copy of the first and the third something different, and test that the first two are equalall and the third is different.

Originally posted by @masinter in #1302 (review)