Fix buggy test assertion code

[foxbenjaminfox]

When comparing maps of equal size, the value returned from the do_deep_equal/2 helper is the value of the for loop—a list. That's a problem, given that do_deep_equals/2 is expected to return a boolean: the one place where do_deep_equals/2 is used is as the condition for an if.

When comparing maps of equal size, do_deep_equals/2 will thus always return a truthy value, as all lists are truthy in elixir. This means that the test suite treats any map with the correct number of keys as deeply equal to the result, which causes tests that rightly should fail to pass.

A similar issue arises for lists, where two lists of the same length are always treated as deeply equal, for the same reason.

I have fixed both of these, by piping the for loop evaluates into Enum.all?/1—only if each iteration of the for loop returned true should do_deep_equals/2 return true.

Unfortunately, this PR causes the test suite to fail. That should be expected: the test suite is in fact broken, and tests should be failing with a fixed do_deep_equals/2.

[bitwalker]

Thanks! This was fixed in d10ff17