Bug 245123 – REGRESSION (?) (iOS 16): Form element stealing focus without user input

• Status: NEW
• Component: Forms

Created attachment 462307 [details]
Image showing an autofocus'd form element that results in the keyboard being displayed.

Users on iOS16 have reported that Safari is now autofocusing on a form field when a page loads.

I do not have any autofocus JS in the page... and I also do not possess any Apple device so cannot verify the claim (also cannot provide Safari version... but "whatever is the latest safari after iOS16 upgrade" is the Safari version).

Example of bug report: "Minor bug but in iOS 16 Safari, the number input in the pagination captures focus on page load, so whenever you arrive at a forum list page the keyboard is launched." and the attached image by the user who reported that shows focus given to the keyboard.

Background:

I run a website with a pagination capability, here is a public example: https://www.lfgss.com/microcosms/509/

The HTML for the pagination renders this form:

```
Page
<form name="paginationByOffset" method="GET" data-initial="1" data-limit="25" data-max="391">
<input type="text" min="1" value="1" max="391" step="1" pattern="\d*" title="Enter the page number to jump to" class="form-control" />
<input name="offset" type="hidden" />
</form>
of 391
```

Some JavaScript does exist to process this when the form is submitted, i.e. go into the form, change the number, hit enter...

```
////////////////////
//	pagination    //
////////////////////
(function(){
	$('form[name=paginationByOffset]').submit(function(e){
		console.log("Page jump requested")

		var self    = $(this),
			initial = self.attr('data-initial'),
			limit   = self.attr('data-limit'),
			max     = self.attr('data-max'),
			value   = parseInt(self.find('input[type=text]').val()),
			hidden  = self.find('input[name=offset]');

		console.log("limit = " + limit + " , max = " + max + " , value = " + value);

		if (!isNaN(value) && value >= 1 && value <= max && value != initial) {
			console.log("Jump")
			if (limit && value){
				hidden.val(limit * (value-1));
			}
		} else {
			console.log("Cancel jump")
			e.preventDefault();
		}
	});
	$('form[name=paginationByOffset] > input[type=text]').blur(function() {
		$(this).parent().submit();
	});
})();
```

It is claimed that on iOS16 *only* that this will autofocus.

I know nothing and this is hypothesis... but is it suspect that code that affects focus triggering was recently changed? https://github.com/WebKit/WebKit/pull/2907

Found someone with an iOS 16 iPhone... they have *not* been able to reproduce this.

I've asked for more info from the people who originally reported it to me.

[Radar WebKit Bug Importer]

<rdar://problem/99881328>

[Aditya Keerthi]

(In reply to webkit from comment #2)
> Found someone with an iOS 16 iPhone... they have *not* been able to
> reproduce this.
> 
> I've asked for more info from the people who originally reported it to me.

I am also unable to reproduce this issue on iOS 16, visiting https://www.lfgss.com/microcosms/509/.

Please let us know if you hear back with additional information from the original reporter(s).

[Aditya Keerthi]

(In reply to webkit from comment #1)
> I know nothing and this is hypothesis... but is it suspect that code that
> affects focus triggering was recently changed?
> https://github.com/WebKit/WebKit/pull/2907

This change is not in iOS 16.

Apparently it's reproducible.

The video of this is 1.6MB which is too large as an attachment so I have uploaded it to here https://david.kitchen/webkit245123/ and it is viewable at the URL here https://david.kitchen/webkit245123/Video.mov

The only thing that stands out is that the Spanish language is default. Aside from that my understanding is that this is a standard iOS 16 iPhone.

The person who recorded the video is not a technical user or engineer, but I have asked him if he's comfortable with my sharing his email and him working with you directly to try and identify whether this is an issue with the browser. As soon as he replies to that I shall update this ticket.

The user who can reproduce the issue is happy to help debug.

Contact: <[e-mail]>

Unclear to me how to add his email to the CC list of this.

[Ahmad Saleem]

(In reply to webkit from comment #7)
> The user who can reproduce the issue is happy to help debug.
> 
> Contact: <[e-mail]>
> 
> Unclear to me how to add his email to the CC list of this.

We can only them in CC, if they create an account on bugs.webkit.org and then they can comment and keep themselves updated on the bug progress as well.

[Aditya Keerthi]

(In reply to webkit from comment #6)
> Apparently it's reproducible.
> 
> The video of this is 1.6MB which is too large as an attachment so I have
> uploaded it to here https://david.kitchen/webkit245123/ and it is viewable
> at the URL here https://david.kitchen/webkit245123/Video.mov
> 
> The only thing that stands out is that the Spanish language is default.
> Aside from that my understanding is that this is a standard iOS 16 iPhone.
> 
> The person who recorded the video is not a technical user or engineer, but I
> have asked him if he's comfortable with my sharing his email and him working
> with you directly to try and identify whether this is an issue with the
> browser. As soon as he replies to that I shall update this ticket.

Thank for the video, and for reaching out to the user.

In the video, I observed that there are some extensions installed. It would be useful to know what extensions are installed, as that could be one reason why I am unable to reproduce.