A simple test of NapariNavigationToolbar._update_buttons_checked.

[samcunliffe]

Inspired by the codecov report, the _update_buttons_checked method of the NapariNavigationToolbar is untested. So here's a (slightly) dumb UI test.

If anyone knows a better way to compare two QImages I'm really open to it.

Buuuut it does seem like a nested for loop is the simplest way to do this. (For reference: there seems to be an unlooping way to shove the data into a numpy array, but it involves pointers, and I'm not sure it's more readable. Readability counts?) I did check the time and the loopy way in this PR is ones of milliseconds.

Testing:

Things I've done:

• Used QImage.save to dump the images.
• Replace the quick return True with

  print(f"a[{x}, {y}] = {a.pixel(x, y)}, \t b[{x}, {y}] = {b.pixel(x, y)}")

  to verify that a copy of the same image is pixel-by-pixel the same, and that the checked and unchecked are pixel-by-pixel different.

Notes for reviewers:

As mentioned: I am not a fan of the nested for loops.

Low priority and independent test. So don't rush to review this. It's unlikely to conflict 👍 .

[samcunliffe]

Suggested change

	Check that a and b are identical, pixel by pixel. Via a stupid nested for loop.
	Check that a and b are not identical, pixel by pixel. Via a stupid nested for loop.

[samcunliffe]

98.8% 🚀