Add support for the HTML5 `oninput` event

[Andy E]

All the edge major browsers appear to support the oninput event, which fires for text input into an element. The event itself is superior to key events because it fires for all forms of text input.

A while ago, I wrote a plugin that normalizes the event for browsers that don't support it, notably IE6-8. I believe this works well enough, despite a few limitations. Despite the existence of this event (and the plugin), every day I see people recommending the use of onkeyup for acting on user input. Integrating this event into jQuery may help these users learn to handle user input more appropriately.

The following fiddle demonstrates the event (and the plugin) very basically: All it does is copies the text as you type, but as you can tell, there is no visual delay.

Further reading:

All it does is copies the text as you type, but as you can tell, there is no visual delay.

Further reading:

[dmethvin]

This is quite a bit of code so I don't think we're in a hurry to include it into core. I'd suggest making it available as a plugin for now. If you'd like to suggest it for 1.7: ​https://spreadsheets.google.com/viewform?hl=en&authkey=CPmgicsO&formkey=dG0yTEs2ZTFWQUhDRUp5dzRyc3NwV2c6MA#gid=0

[john]

This was suggested for 1.7.

[john]

Nominating ticket for 1.7 discussion.

[jaubourg]

-1, Let's start as a plugin so that we can assess the limitations.

[Timmy Willison]

-1, plugin

[dmethvin]

-1, Plugin first.

[anonymous]

I wrote a plugin (as I said in the original filing). I also discuss some of the limitations in the same post.

​http://whattheheadsaid.com/projects/input-special-event

Also, I've just noticed a bug in IE 9 where the oninput event doesn't fire when the backspace or delete keys are pressed - my plugin doesn't yet deal with that.

[john]

-1, Would be good to prove as a plugin.

[Andy E]

I'm not sure I understand this, do you mean an "official" plugin of some sort, or are people just not reading the ticket properly? Despite the ticket body and my other comment saying "I have already provided a plugin", everyone is still saying "-1 plugin first".

This has left me rather confused.

[Rick Waldron]

hey Andy,

The comments you see are votes by core, bus and ui team members - this ticket was elevated to 1.7 proposal status.

As for official plugin, that's not likely - but if you write a solid plugin that has a high adoption rate, there is a better chance for core consideration, though rare

hope that cleared things up for you. Feel free to ask any other questions!

[ajpiano]

+0, Seems the full ticket proposal got truncated and the fiddle and links got cut off - I've amended the description accordingly, and taken a look at your plugin. At the moment, it seems to have a number of shortcomings - the lack of support for delegation, for one, so it seems there's more research necessary into the problem. I like the idea of providing a cross-browser solution to this, however - have you heard of much adoption of the plugin thus far?

[andy@…]

I know several people have commented on my blog post and I regularly direct people to it from Stack Overflow. The plugin is the most visited page on my blog, but I only installed analytics a couple of weeks ago so there's not really enough data to say anything solid.

Telling people doesn't always work, though; there are a lot of confused individuals out there promoting events like keyup for input counters and the like. Most jQuery plugins detecting input use keydown, keypress or keyup too. input would serve them much better, catching cutting, pasting, dropping, spell checker corrections and others. IMO, having the event supported by jQuery would immediately increase awareness of the event, improving the quality of input detection scripts around the web.

As for delegation, it would work for the browsers that support the event natively, since it bubbles but, as for the older browsers (namely IE6-8), it would be difficult to get the event working for "future elements". I did think about this to some extent, though, and may be able to come up with a solution based on focusin or mouseover -- when an input element receives either event it is checked against the delegate selector and then the input event is applied to the element if necessary.

[Rick Waldron]

Mathias also has an implementation: ​http://mathiasbynens.be/notes/oninput

I believe he is going to have this, with unit tests, on github today.

[Andy E]

OK, so I managed to put together an experimental version of the plugin that supports delegate(). It only increased the size by a few lines.

​http://jsfiddle.net/AndyE/pLUY4/1/

I haven't added this to my blog post yet because I just realized I forgot to unset some data in the event teardown and I also need to implement a workaround for the bug in IE 9 where it doesn't fire oninput for backspace or delete. I've run out of time at the moment, though, so just posting what I've got so far.

[mathias]

Note that my plugin is very barebones by design. jQuery probably needs something more consistent across browsers.

I just wanted a simple, minimal solution that gracefully degrades in older browsers. (Also, just like Andy’s special event plugin, it doesn’t support event delegation.)

[scottgonzalez]

-1, perhaps a good next step would be showing how this would improve some jQuery UI plugins

[dmethvin]

No reliable feature detect is available according to ​https://github.com/Modernizr/Modernizr/issues/210

[Andy E]

dmethvin,

That's not the case. According to several users on that ticket, the bug has been fixed in the latest version of Firefox. Aside from that, it's detected by my plugin using a few fallback techniques, including a "last resort" technique that works in Firefox, which was outlined on Daniel Friesen's blog.

​http://blog.danielfriesen.name/2010/02/16/html5-browser-maze-oninput-support/

scott.gonzalez,

The only jQuery plugin that appears (from a glance) to make use of key events for text input is Autocomplete. As far as I can tell, it doesn't support spell-checker corrections, cut, paste, undo, redo, drag and drop, voice control/speech-to-text or any other form of text input.

Supporting the event would be a boost for Opera, which doesn't support several text input events that other browsers do like cut, paste, textInput, etc.

[anonymous]

The input event is not supported for textareas on Safari < 5. You'll need to fall back to the DOM 3 textInput event, which Safari does support from 3.1 (and even then you're not covering Safari 3.0).

[Tim Down]

Previous comment is mine.

[Andy E]

That's a good point, Tim. I believe Safari 3.0 supports oncut, onpaste and drag/drop events (assuming you cancel the ondragstart event for the text box) -- all that's needed is to detect the bug where oninput doesn't fire for textarea elements. I'll give it some thought and see if I can come up with a solution.

As for Safari 3.1 and later, you could bind to just textInput. Or bind to both events and set a flag to suppress one when the other fires (depending on order of precedence).

If I get some time, I'll have a crack at a solution.

[Andy E]

Right, got a bit of time and I've come up with a proof of concept that appears to work well in all browsers:

​http://jsfiddle.net/AndyE/sdkBs/

It works by binding to all known input events on <input> and <textarea> elements, but utilizes only the most "useful" event which is usually one of the following:

• input in browsers that support it (except IE9 due to the delete bug)
• onpropertychange in all versions of IE
• textInput on <textarea> elements in Safari < 5

I've tested it on <textarea> in Safari 3.0 and 4.05. It works with delegate using the same method as before - binding to mouseover, focusin and dragstart on the delegated element and binding to <input> and <textarea> elements that fire any of those events. keydown might also be required for older browsers that don't support focusin - I need to do a little research to see if this is necessary.

One of the interesting advantages of this approach is that it trades in feature detection for graceful degradation - no more awkward checks in Firefox to find out if the event is supported and no more false positives from Safari <5.0 on the <textarea> element.

If you ask me, the bugs in vendor implementations makes it all the more important that one of the major libraries normalizes this event for us. It was introduced in HTML5 to fill an obvious deficiency in input detection and, with the influence jQuery has with web developers, including support for it will help spread the word that input detection doesn't have to stop at the keyboard.

[Andy E]

It was late last night when I wrote that, so I forgot to add that IE 9 does use the input event, except for when the bug is in effect and it doesn't fire for the backspace or delete keys.

Also, because the custom version of the event relies on the input event, I had to rename the special event to input2 - haven't yet figured how to work around this with the jQuery event api.

[dmethvin]

Per the 1.7 meeting this won't be implemented.

[Timmy Willison]

#10971 is a duplicate of this ticket.