Nested comparisons happen in unexpected order

[jvoigtlaender]

In comparing, say, tuples (a,b,c) vs. (d,e,f), the natural expectation is that first a and d are compared and only if these are equal, b and e are compared, etc.

But that's not what is currently happening. That it's happening in reverse at the moment can be seen from the facts that ( True , identity ) == ( False , \x -> x ) does trigger a runtime error but ( identity , True ) == ( \x -> x , False ) doesn't.

In more realistic code it will more likely not be about runtime errors but about efficiency. But still, the natural expectation would be that (a,b) == (c,d) is equivalent to a==c && b==d, not to b==d && a==c.

[evancz]

This should be "fixed" by 6f5fbb8.

[evancz]

This should be "fixed" by 6f5fbb8.