Different behavior of go 1.3 and go 1.2.2

[stretchr]

Different behavior of go 1.3 and go 1.2.2

[vkryukov]

This is an interesting example. Indeed, these two values will fail all of ObjectsAreEqual tests:

a b: []byte{} []byte(nil)
a==nil b==nil: false true
a DeepEqual b: false
aValue == bValue: false
aValue convertible to bValue: true
bValue convertible to aValue: true
aValue converted to bValue == bValue: false
bValue converted to aValue == aValue: false

BTW, fmt.Sprintf("%#v", expected) == fmt.Sprintf("%#v", actual) is not covered by tests, and when I was working on #76, I couldn't come up with anything that would fail all the previous comparisons in ObjectsAreEqual but not this one.

Hopefully someone more enlightened about Go reflection could provide some insight here.

[vkryukov]

The Nil section at http://blog.golang.org/slices describes the difference between empty and nil slices of the same type, specifically:

As should be clear, an empty slice can grow (assuming it has non-zero capacity), but a nil slice has no array to put values in and can never grow to hold even one element.

That said, a nil slice is functionally equivalent to a zero-length slice, even though it points to nothing. It has length zero and can be appended to, with allocation.

Based on the above, I would say the ObjectsAreEqual behavior is correct in this case.

[vkryukov]

See also this: https://groups.google.com/d/msg/golang-nuts/QrItXLz3jis/NubiAVy9F6sJ

[m0sth8]

Victor, I agree with you. It was a mistake in go and they fixed it. But behavior of testify is different on go 1.2 and 1.3. I think I'll close this bug and leave it for the history.