Simplify test code

[radarhere]

Given

Pillow/Tests/helper.py

Lines 87 to 90 in d3b755d

	def assert_image_equal(a: Image.Image, b: Image.Image, msg: str | None = None) -> None:
	assert a.mode == b.mode, msg or f"got mode {repr(a.mode)}, expected {repr(b.mode)}"
	assert a.size == b.size, msg or f"got size {repr(a.size)}, expected {repr(b.size)}"
	if a.tobytes() != b.tobytes():

Pillow/Tests/helper.py

Lines 112 to 116 in d3b755d

	def assert_image_similar(
	a: Image.Image, b: Image.Image, epsilon: float, msg: str | None = None
	) -> None:
	assert a.mode == b.mode, msg or f"got mode {repr(a.mode)}, expected {repr(b.mode)}"
	assert a.size == b.size, msg or f"got size {repr(a.size)}, expected {repr(b.size)}"

looking at

Pillow/Tests/test_image_resize.py

Lines 233 to 236 in d3b755d

	with pytest.raises(pytest.fail.Exception):
	assert_image_equal(ref, im)
	assert_image_similar(ref, im, epsilon)

assert_image_similar asserts that the mode and size are the same, meaning that assert_image_equal should only fail when comparing im.tobytes().

So let's just simplify things by testing im.tobytes() directly. I find this clearer.

assert_image_similar(ref, im, epsilon)
assert ref.tobytes() != im.tobytes()

[akx]

It does read simpler, yep!

I wonder how much Image.__eq__ is being used in real code in the wild. 53c9a80 changed it to short-circuit but it's still (like this code) paying for serializing both images to bytes (#9938 😉) even if, say, the first byte differs.