Meta: Investigate viability of insertAdjacentHTML use

[Rick Waldron]

The latest Firefox will have support for this, I'm wondering if support is wide enough now to consider adding it to any of the manipulation logic paths.

[aymeric@…]

It's an awesome idea. The Mozilla Team wrote an article about the better performance of this function: ​http://hacks.mozilla.org/2011/11/insertadjacenthtml-enables-faster-html-snippet-injection/

[dmethvin]

It would be nice if we could totally replace our fragment cache with insertAdjacentHTML instead. That would mitigate the code bloat and eliminate the pseudo-leak that the frag cache causes. However we'd need to see how iAH handles the cases already in .clean() which are primarily IE 6/7/8 workarounds.

[Timmy Willison]

If it removes the need for clean, it would be worth the extra codepath.

[Timmy Willison]

#10804 is a duplicate of this ticket.

[dmethvin]

(In #10903) None of the HTML processing code in jQuery.clean or jQuery.buildFragment currently attaches content to the document, that's done later when the method-specific callback (e.g., append, prepend, before, after) is called. So this isn't a simple change and a very cornery case.

However, we've been talking about using insertAdjacentHTML in #10515 which might solve this problem so I'll leave it open. If it can't be worked in there I think we will close this wontfix.

[dmethvin]

Bulk change from enhancement to feature.

[dmethvin]

Pull request at ​https://github.com/jquery/jquery/pull/1095 for one narrow but common case, $(selector).append(html), but I'd like to see perfs before landing it.

[dmethvin]

Thoroughly investigated by @orkel in ​https://github.com/jquery/jquery/pull/1200 and we came to the conclusion that there is no significant benefit.