Why does JSoup remove element IDs?

Question

I'm using JSoup to sanitize some untrusted HTML. I discovered that if I call

String html = "<div id='foo'><script type='text/javascript'>alert('hello');</script></div>";
String cleanedHtml = Jsoup.clean(html, Whitelist.relaxed());

At this point cleanedHtml is

<div><div>

So the <script> tag has correctly been removed, but mysteriously, so has the id attribute of the <div> . Is there any good reason why this should be removed or is it a bug?

Answer 1

By default the id attribute is removed; add it as an allowable attribute:

Whitelist whitelist = Whitelist.relaxed().addAttributes("div", "id");
System.out.println(Jsoup.clean(html, whitelist));

=> <div id="foo"></div>

Is it a bug? Not AFAIC; it's in the source. IMO there are documentation bugs, though.

Is there "any good reason" why this should be removed? Not sure about that one, but attributes like this aren't structural: removing it doesn't alter the DOM. That's the thing about whitelists–they explicitly allow, and must be curated to match your precise needs.