.has() method ignores child ( > ) selector

[goozak]

There is a difference between the .has() method and the :has() selector with regards to the child selector (>)

​http://jsfiddle.net/wMBBc/

The .has() method seems to ignore the presence of the child selector (>).

[timmywil]

This is valid. .has() selects its targets then independently does a .contains() check. However, this is probably the only use case I can think of that would fail.

[dmethvin]

It seems consistent with the docs:

The supplied selector is tested against the descendants of the matching elements; the element will be included in the result if any of its descendant elements matches the selector. -- ​http://api.jquery.com/has/

One of the descendants of the outer td has a descendant with an input element as a child.

Also note that the selector string matching the .has() method case would be #rec td *:has(> input) and not the one you've used, since it's testing the descendants of the set and not the set elements themselves.

Does that make sense?

[timmywil]

@dmethvin:

You're right, forgive my confusion.

[dmethvin]

Now, does it make sense for these two cases to be different? Is the selector case wrong?

[gibson042]

I stand behind the selector behavior, and would argue that our documentation text is sufficiently vague regarding implied scope to support either method implementation, though I personally prefer it to match the selector.

At any rate, this fix can be had for 15 min+gz bytes if we come down in favor of it.

[timmywil]

Continuing to work as documented seems fine with me. No change required. As Dave points out, it _does_ match the selector, just depends on which selector you're talking about.

[goozak]

Duplicate of ​http://bugs.jquery.com/ticket/7205

Also kind of duplicate of ​http://bugs.jquery.com/ticket/8384

[dmethvin]

Bulk update for new milestone