hook_ctools_plugin_api generated code throws notices during tests

Ctools plugin api code generated by features is - well - less than ideal ;). So this is an example of the hook_ctools_plugin_api code generated by features:

/**
 * Implements hook_ctools_plugin_api().
 */
function file_types_ctools_plugin_api() {
  list($module, $api) = func_get_args();
  if ($module == "file_entity" && $api == "file_default_displays") {
    return array("version" => "1");
  }
}

This issue has to do with this code: list($module, $api) = func_get_args(); The issue here is that features is not using the method arguments in the method, but rather getting them after the fact. The ctools API is currently very stable, so why this is done I have no idea. The issue was discovered when I was doing simple tests, and i just started throwing no arguments, one argument and so on to see what would happen. The result was that my test started throwing notices when i threw no arguments, or just one.

Fixing this would be a matter of actually putting the arguments in the method, rather than getting them after the fact. This way you ensure that there are always two arguments being thrown. The ctools people themselves - in their tutorial - would have written the method like this:

/**
 * Implements hook_ctools_plugin_api().
 */
function file_types_ctools_plugin_api($module, $api) {
  if ($module == "file_entity" && $api == "file_default_displays") {
    return array("version" => "1");
  }
}

To make this properly testable, and avoid it throwing these notices, I submit that we start using the ctools way of doing it. If not, I'd like to hear arguments of why we don't want to do it the way the ctools developers themselves say we should.