list_field_update_forbid logic is backwards

Odd. There are tests for the behavior, that currently pass. My manual testing, back in #932502: Changing allowed values in "List" fields, worked ok too.

However, it does seem that manually testing the behavior now indeed doesn't work.
Got to run right now, no time to apply the patch or check the tests.

My manual testing worked ok too

An additional validation handler outside of list_field_update_forbid() also checks for removal of an in-use list value, so your manual test might have correctly blocked you from removing a value prior to hook_field_update_forbid being invoked:

list_allowed_values_settings_validate() in list.module:161

    // Prevent removing values currently in use.
    if ($has_data) {
      $lost_keys = array_diff(array_keys($field['settings']['allowed_values']), array_keys($values));
      if (_list_values_in_use($field, $lost_keys)) {
        form_error($element, t('Allowed values list: some values are being removed while currently in use.'));
      }
    }

When I disable the above form validation, I can successfully remove an allowed value even though it's in use, so Doug's assessment is correct. The automated tests also seem to only cover the form validation, not the list_field_update_forbid() invocation.

Doug's patch looks correct to me (although we should probably add a test against hook_field_update_forbid() for list.module), but leaving for someone with more field-experience to handle this...

The patch is correct. And this means we have major bugs in our tests, in case this is indeed tested already. :(

Part of the issue is that the tests did not create any field data to actually perform tests against so the logic was never being run. 2 patches attached first is test only and second is test plus fix above.

Looks great, one line fix, nice test.

Sorry, please don't commit just yet, something bugs me in the tests; just checking.

Same fix, with a test that is slightly better integrated with the existing ones.
I dare say this is still RTBC, but I'll leave someone else push it back.

Egad! Nice find, and thank you SO much for the tests on this bad boy.

Committed and pushed to 8.x and 7.x. Thanks!

Just came across this bug, I couldn't believe it. :) Guess these things happen.