No use case? Well then... what's the use?

Principle of least surprise and largest utility. Being able to write this:

$buttons = $('ul.buttons li a');
$buttons.click(function() {
  var clicked = this;
  $buttons.toggleClass('active', function() { return this === clicked });
});

instead of (as you currently need to) this:

$buttons = $('ul.buttons li a');
$buttons.click(function() {
  var clicked = this;
  $buttons.each(function() {
    $(this).toggleClass('active', this === clicked);
  });
});

What is particularly unfortunate about the current behaviour is that first code example doesn't throw any error - it just behaves as if the function passed wasn't even there, plain toggling the class for all elements in the jQuery selection.

That use case is usually handled with something along the lines of:

$buttons.not( this ).removeClass( "active" );
$( this ).addClass( "active" );

To me the point of using toggleClass in the first place is to write DRY code.

Anyway, I am not here to argue or press my point – I thought it was just an accidental bug or missed corner case, especially as I've been using the great d3.js idiom selection.classed('active', bool_or_function_returning_bool) long enough to have that hard-wired, rather than breaking it down into multiple discrete manual steps in separate chained commands, or using a more complex chain, as in the fallback I opted for above.

It would just delight me if jQuery matched intuition.

https://github.com/mbostock/d3/wiki/Selections#wiki-classed if unfamiliar - it's super useful.

or

$buttons.removeClass('active').siblings().addClass('active');

+1

+0, What gets passed into the function... index, hasClass ???

-1, not needed