ng-click not working in HTML loaded into div - angularjs-ng-click

I've seen a couple answers but can't seem to figure out how to apply them.
I have a page that's two divs. A side nav div and a main div. On the page load, the js below loads up the html into the main div. The same function is used to jump around (via the nav) within the doc once loaded.
The problem is, there are also links within the HTML to the function since there are some self-referencing points, etc.
The side nav ones work fine, the ones within the HTML don't. I'm assuming it's a compiling issue of sorts, but I can't quite figure out how to compile it correctly.
Here's the JS:
Ctrler.loadPage = function(hash){
if(!hash){
var url = "filename.html";
var xhr= new XMLHttpRequest();
xhr.open('GET', url, true);
xhr.onreadystatechange= function() {
if (this.readyState!==4) return;
if (this.status!==200) return;
cur = this.responseText;
cur = cur.replace(/{{sitePath}}/g, jsPath.path);
document.getElementById('maindiv').innerHTML= cur;
$('#maindiv').html = this.responseText;
};
xhr.send();
}
else{
document.getElementById(hash).scrollIntoView();
window.scrollBy(0, -90);
}
}
EDIT: I changed the lines that populate the div to this:
$compile(cur)($scope);
$('#maindiv').append(cur);
but it still doesn't work. It loads the first time, but ng-clicks still don't work.


I figured it out.
I needed to compile $('#maindiv') after I added cur. Not cur

Related

Using dojo dom.byId is not getting an element added programmatically

I'm creating a dom element programatically using dojo and I can "see" it in the dom with its id, but when I attempt a dom.byId("myId") it returns null.
I have a similar jsfiddle that is actually working (so it doesn't reproduce my problem, but it gives an idea of what I'm trying to do): if you click the button (ignore the lack of styling) in the run output panel, it alerts the content of the element retrieved by dom.byId. But similar code within my dojo widget is not working. Here's the code:
var content = lang.replace(selectFilterTemplate, {
"layer-id": layer.id,
"layer-index": idx,
"filter-name": filter.name
}); // this gets template HTML code similar to what's in the HTML panel of the jsfiddle, only it has placeholder tags {} instead of literals, and the tags are replaced with the attributes of the layer, idx, and filter objects here
// Use dojo dom-construct to create a div with the HTML from above
var node = domConstruct.create("div", { "innerHTML": content });
// put the new div into a dojo ContentPane
var filterPanel = new ContentPane({
"id": layer.id + "-filter-" + idx + "-panel",
"content": node,
"style": "width: 200px; float: left;"
});
// Get the dom element:
var mstag = dom.byId(layer.id + "-filter-" + idx + "-ms-tag")
// this is the same as the "var ms = dom.byId("IssuePoints-filter-1-ms-tag")" in the jsfiddle, but this one returns null. If I view the contents of the 'node' variable in the browser debugging console at this point, I can see the <select> tag with the id I'm referencing.
Why would I be getting null in my dom.byId() if I can see that element in the dom in the debugging console?

It seems that the element is added to the dom at a later point. You may see it with the debugger but it is not yet available the moment you call byId().
In the code you posted you create the filterPanel element but you do not place it in the dom. I assume this happens at a later stage. In contrast, the jsfiddle places the Button element with placeAt() directly after constructing it.

inject a javascript function into an Iframe

This link (archived version) describes how to inject code from a script into an iframe:
function injectJS() {
var iFrameHead = window.frames["myiframe"].document.getElementsByTagName("head")[0];
var myscript = document.createElement('script');
myscript.type = 'text/javascript';
myscript.src = 'myscript.js'; // replace this with your SCRIPT
iFrameHead.appendChild(myscript);
}
That's ok, but what if I want to insert a function object into an iframe and get it executed in the iframe context? Let's say I have:
function foo () {
console.log ("Look at me, executed inside an iframe!", window);
}
and I want to insert foo's code inside an iframe? (function foo could be something loaded dynamically, I can't just wrap it in quotes)
I naively tried:
var scriptFooString = "<script>" + foo.toString() + "</script>"
to get the code inside function, but
I don't know how to insert it in the iframe HEAD (maybe with jquery?)
I don't know if it's the right way
I don't know what happens when if function is way more complex than that
I don't know what happens with double and single quotes in scriptFooString
Any hint?

First of all you can only accomplish this if your frame and the page displaying it is within the same domain (Due to cross-domain rules)
secondly you can manipulate dom and window objects of the frame directly through JS:
frames[0].window.foo = function(){
console.log ("Look at me, executed inside an iframe!", window);
}
to get your frame from a DOMElement object you can use:
var myFrame = document.getElementById('myFrame');
myFrame.contentWindow.foo = function(){
console.log ("Look at me, executed inside an iframe!");
}
Note that the scope in foo is NOT changed, so window is still the parent window etc. inside foo.
If you want to inject some code that needs to be run in the context of the other frame you could inject a script tag, or eval it:
frames[0].window.eval('function foo(){ console.log("Im in a frame",window); }');
Though the general consensus is to never use eval, I think its a better alternative than DOM injection if you REALLY need to accomplish this.
So in your specific case you could do something like:
frames[0].window.eval(foo.toString());

This code is the result of my research. The accepted answer also helped me a lot.
First of all, I create a simple iframe:
<iframe id="myiframe" width="200" height="200" srcdoc="<h1 id='title'>Hello from Iframe</h1><button type='button' id='fire'>Click Me!</button>
"></iframe>
For access to iframe's window and document I used this code.
const iframe = document.getElementById('myiframe');
const iframeWin = iframe.contentWindow || iframe;
const iframeDoc = iframe.contentDocument || iframeWin.document;
Finally I injected js codes into iframe:
var script = iframeDoc.createElement("script");
script.append(`
window.onload = function() {
document.getElementById("fire").addEventListener('click', function() {
const text = document.getElementById('title').innerText;
alert(text);
})
}
`);
iframeDoc.documentElement.appendChild(script);
Demo:
https://jsfiddle.net/aliam/1z8f7awk/2/

Here's my solution. I'm using jquery to insert the content, then using eval to execute the script tags in the context of that iframe:
var content = $($.parseHTML(source, document, true));
$("#content").contents().find("html").html(content);
var cw = document.getElementById("content").contentWindow;
[].forEach.call(cw.document.querySelectorAll("script"), function (el, idx) {
cw.eval(el.textContent);
});

Meteor: Elements from CollectionA re-rendering when I insert to CollectionB

I'm attempting to fade-in new elements in a reactive {{#each}} of the comments posted by users.
I have a code sample at https://gist.github.com/3119147 of a very simple comments section (textarea and new comment insert code not included, but it's very boilerplate.). Included is a snippet of CSS where I give .comment.fresh { opacity: 0; }, and then in my script, I have:
Template.individual_comment.postedago_str = function() {
var id = this._id;
Meteor.defer(function() {
$('#commentid_'+id+'.fresh').animate({'opacity':'1'}, function() {
$(this).removeClass('fresh');
});
});
return new Date(this.time).toString();
};
Which seems like a terrible place to execute an animation. My thinking is that each time a new comment is rendered, it will need to call all my Template.individual_comment.* functions, so that's why my animation defers from one of those. However, Meteor is calling Template.individual_comment.postedago_str() each time a different collection (Likes) is inserted to. This means I click the Like button, and my whole list of comments flashes white and fades back in (very annoying!).
I read the Meteor documentation and tried to figure out how to better slice up my templates so only chunks will update, and I added id="" attributes everywhere that seemed reasonable.. still this bug. Anyone know what's going on?
TIA!

As a workaround, you could wrap an {{if}} block around the fresh class on individual comments, that would check the comment's creation time and only add the fresh class in the first place if the comment is actually recent. Something like:
<div class="comment{{#if isActuallyFresh}} fresh{{/if}}" id="commentid_{{_id}}">
And then define the isActuallyFresh function:
Template.individual_comment.isActuallyFresh = function() {
if ((new Date().getTime() - this.time) < 300000) // less than 5 minutes old
return true;
else
return false;

iscroll rubber band effect in jQTouch

I am a new developer and am trying to create a jQTouch application to display some scrollable content throughout multiple pages. I've decided to use iscroll and it only works fine on the home page. I've read that I need to refresh iscroll after each page but I am completely lost on how to do this. Here is my script:
<script type="text/javascript">
var myScroll, myScroll2;
function loaded() {
setTimeout(function () {
myScroll = new iScroll('wrapper1');
}, 100);
setTimeout(function () {
myScroll2 = new iScroll('wrapper2');
}, 100);
}
document.addEventListener('touchmove', function (e) { e.preventDefault(); }, false);
document.addEventListener('DOMContentLoaded', loaded, false);
</script>
In my html I have a div id="wrapper1" which works fine until I navigate to the second page where the div id="wrapper2" has the rubber band effect.

In case you haven't figured this out yet (although I'm sure you have), you want:
myScroll.refresh()
or
myScroll2.refresh()

Ok finally got this working. To get jQTOuch and iScroll to play nice with each other, the scrolling areas on the page need to be reset each time JQTouch makes them disappear. In other words, once you hide the div, iScroll doesn't know what to scroll the next time it's made visible. So as a result, you get the infamous rubberband effect. To solve this, just add an event listener that resets the scrolling area right after the div is called. Make sure you give it 100 to 300ms delay. This code below assumes your variable is called myScroll:
$(".about").tap(function(){
setTimeout(function(){myScroll.refresh()},300);
});
And on a side note, here's how to establish multiple scrollers using iScroll:
var scroll1, scroll2;
function loaded() {
scroll1 = new iScroll('wrapper1');
scroll2 = new iScroll('wrapper2');
}

iOS 5 fixed positioning and virtual keyboard

I have a mobile website which has a div pinned to the bottom of the screen via position:fixed. All works fine in iOS 5 (I'm testing on an iPod Touch) until I'm on a page with a form. When I tap into an input field and the virtual keyboard appears, suddenly the fixed position of my div is lost. The div now scrolls with the page as long as the keyboard is visible. Once I click Done to close the keyboard, the div reverts to its position at the bottom of the screen and obeys the position:fixed rule.
Has anyone else experienced this sort of behavior? Is this expected? Thanks.

I had this problem in my application. Here's how I'm working around it:
input.on('focus', function(){
header.css({position:'absolute'});
});
input.on('blur', function(){
header.css({position:'fixed'});
});
I'm just scrolling to the top and positioning it there, so the iOS user doesn't notice anything odd going on. Wrap this in some user agent detection so other users don't get this behavior.

I had a slightly different ipad issue where the virtual keyboard pushed my viewport up offscreen. Then after the user closed the virtual keyboard my viewport was still offscreen. In my case I did something like the following:
var el = document.getElementById('someInputElement');
function blurInput() {
window.scrollTo(0, 0);
}
el.addEventListener('blur', blurInput, false);

This is the code we use to fix problem with ipad. It basically detect discrepancies between offset and scroll position - which means 'fixed' isn't working correctly.
$(window).bind('scroll', function () {
var $nav = $(".navbar")
var scrollTop = $(window).scrollTop();
var offsetTop = $nav.offset().top;
if (Math.abs(scrollTop - offsetTop) > 1) {
$nav.css('position', 'absolute');
setTimeout(function(){
$nav.css('position', 'fixed');
}, 1);
}
});

The position fixed elements simply don't update their position when the keyboard is up. I found that by tricking Safari into thinking that the page has resized, though, the elements will re-position themselves. It's not perfect, but at least you don't have to worry about switching to 'position: absolute' and tracking changes yourself.
The following code just listens for when the user is likely to be using the keyboard (due to an input being focused), and until it hears a blur it just listens for any scroll events and then does the resize trick. Seems to be working pretty well for me thus far.
var needsScrollUpdate = false;
$(document).scroll(function(){
if(needsScrollUpdate) {
setTimeout(function() {
$("body").css("height", "+=1").css("height", "-=1");
}, 0);
}
});
$("input, textarea").live("focus", function(e) {
needsScrollUpdate = true;
});
$("input, textarea").live("blur", function(e) {
needsScrollUpdate = false;
});

Just in case somebody happens upon this thread as I did while researching this issue. I found this thread helpful in stimulating my thinking on this issue.
This was my solution for this on a recent project. You just need to change the value of "targetElem" to a jQuery selector that represents your header.
if(navigator.userAgent.match(/iPad/i) != null){
var iOSKeyboardFix = {
targetElem: $('#fooSelector'),
init: (function(){
$("input, textarea").on("focus", function() {
iOSKeyboardFix.bind();
});
})(),
bind: function(){
$(document).on('scroll', iOSKeyboardFix.react);
iOSKeyboardFix.react();
},
react: function(){
var offsetX = iOSKeyboardFix.targetElem.offset().top;
var scrollX = $(window).scrollTop();
var changeX = offsetX - scrollX;
iOSKeyboardFix.targetElem.css({'position': 'fixed', 'top' : '-'+changeX+'px'});
$('input, textarea').on('blur', iOSKeyboardFix.undo);
$(document).on('touchstart', iOSKeyboardFix.undo);
},
undo: function(){
iOSKeyboardFix.targetElem.removeAttr('style');
document.activeElement.blur();
$(document).off('scroll',iOSKeyboardFix.react);
$(document).off('touchstart', iOSKeyboardFix.undo);
$('input, textarea').off('blur', iOSKeyboardFix.undo);
}
};
};
There is a little bit of a delay in the fix taking hold because iOS stops DOM manipulation while it is scrolling, but it does the trick...

None of the other answers I've found for this bug have worked for me. I was able to fix it simply by scrolling the page back up by 34px, the amount mobile safari scrolls it down. with jquery:
$('.search-form').on('focusin', function(){
$(window).scrollTop($(window).scrollTop() + 34);
});
This obviously will take effect in all browsers, but it prevents it breaking in iOS.

This issue is really annoying.
I combined some of the above mentioned techniques and came up with this:
$(document).on('focus', 'input, textarea', function() {
$('.YOUR-FIXED-DIV').css('position', 'static');
});
$(document).on('blur', 'input, textarea', function() {
setTimeout(function() {
$('.YOUR-FIXED-DIV').css('position', 'fixed');
$('body').css('height', '+=1').css('height', '-=1');
}, 100);
});
I have two fixed navbars (header and footer, using twitter bootstrap).
Both acted weird when the keyboard is up and weird again after keyboard is down.
With this timed/delayed fix it works. I still find a glitch once in a while, but it seems to be good enough for showing it to the client.
Let me know if this works for you. If not we might can find something else. Thanks.

I was experiencing same issue with iOS7. Bottom fixed elements would mess up my view not focus properly.
All started working when I added this meta tag to my html.
<meta name="viewport" content="width=device-width,initial-scale=1.0,maximum-scale=1.0,user-scalable=no,height=device-height" >
The part which made the difference was:
height=device-height
Hope that helps someone.

I've taken Jory Cunningham answer and improved it:
In many cases, it's not just one element who goes crazy, but several fixed positioned elements, so in this case, targetElem should be a jQuery object which has all the fixed elements you wish to "fix". Ho, this seems to make the iOS keyboard go away if you scroll...
Needless to mention you should use this AFTER document DOM ready event or just before the closing </body> tag.
(function(){
var targetElem = $('.fixedElement'), // or more than one
$doc = $(document),
offsetY, scrollY, changeY;
if( !targetElem.length || !navigator.userAgent.match(/iPhone|iPad|iPod/i) )
return;
$doc.on('focus.iOSKeyboardFix', 'input, textarea, [contenteditable]', bind);
function bind(){
$(window).on('scroll.iOSKeyboardFix', react);
react();
}
function react(){
offsetY = targetElem.offset().top;
scrollY = $(window).scrollTop();
changeY = offsetY - scrollY;
targetElem.css({'top':'-'+ changeY +'px'});
// Instead of the above, I personally just do:
// targetElem.css('opacity', 0);
$doc.on('blur.iOSKeyboardFix', 'input, textarea, [contenteditable]', unbind)
.on('touchend.iOSKeyboardFix', unbind);
}
function unbind(){
targetElem.removeAttr('style');
document.activeElement.blur();
$(window).off('scroll.iOSKeyboardFix');
$doc.off('touchend.iOSKeyboardFix blur.iOSKeyboardFix');
}
})();

I have a solution similar to #NealJMD except mine only executes for iOS and correctly determines the scroll offset by measuring the scollTop before and after the native keyboard scrolling as well as using setTimeout to allow the native scrolling to occur:
var $window = $(window);
var initialScroll = $window.scrollTop();
if (navigator.userAgent.match(/iPhone|iPad|iPod/i)) {
setTimeout(function () {
$window.scrollTop($window.scrollTop() + (initialScroll - $window.scrollTop()));
}, 0);
}

I have fixed my Ipad main layout content fixed position this way:
var mainHeight;
var main = $('.main');
// hack to detects the virtual keyboard close action and fix the layout bug of fixed elements not being re-flowed
function mainHeightChanged() {
$('body').scrollTop(0);
}
window.setInterval(function () {
if (mainHeight !== main.height())mainHeightChanged();
mainHeight = main.height();
}, 100);

I had a similar problem to #ds111 s. My website was pushed up by the keyboard but didn't move down when the keyboard closed.
First I tried #ds111 solution but I had two input fields. Of course, first the keyboard goes away, then the blur happens (or something like that). So the second input was under the keyboard, when the focus switched directly from one input to the other.
Furthermore, the "jump up" wasn't good enough for me as the whole page only has the size of the ipad. So I made the scroll smooth.
Finally, I had to attach the event listener to all inputs, even those, that were currently hidden, hence the live.
All together I can explain the following javascript snippet as:
Attach the following blur event listener to the current and all future input and textarea (=live): Wait a grace period (= window.setTimeout(..., 10)) and smoothly scroll to top (= animate({scrollTop: 0}, ...)) but only if "no keyboard is shown" (= if($('input:focus, textarea:focus').length == 0)).
$('input, textarea').live('blur', function(event) {
window.setTimeout(function() {
if($('input:focus, textarea:focus').length == 0) {
$("html, body").animate({ scrollTop: 0 }, 400);
}
}, 10)
})
Be aware, that the grace period (= 10) may be too short or the keyboard may still be shown although no input or textarea is focused. Of course, if you want the scrolling faster or slower, you may adjust the duration (= 400)

really worked hard to find this workaround, which in short looks for focus and blur events on inputs, and scrolling to selectively change the positioning of the fixed bar when the events happen. This is bulletproof, and covers all cases (navigating with <>, scroll, done button). Note id="nav" is my fixed footer div. You can easily port this to standard js, or jquery. This is dojo for those who use power tools ;-)
define([
"dojo/ready",
"dojo/query",
], function(ready, query){
ready(function(){
/* This addresses the dreaded "fixed footer floating when focusing inputs and keybard is shown" on iphone
*
*/
if(navigator.userAgent.match(/iPhone/i)){
var allInputs = query('input,textarea,select');
var d = document, navEl = "nav";
allInputs.on('focus', function(el){
d.getElementById(navEl).style.position = "static";
});
var fixFooter = function(){
if(d.activeElement.tagName == "BODY"){
d.getElementById(navEl).style.position = "fixed";
}
};
allInputs.on('blur', fixFooter);
var b = d.body;
b.addEventListener("touchend", fixFooter );
}
});
}); //end define

This is a difficult problem to get 'right'. You can try and hide the footer on input element focus, and show on blur, but that isn't always reliable on iOS. Every so often (one time in ten, say, on my iPhone 4S) the focus event seems to fail to fire (or maybe there is a race condition), and the footer does not get hidden.
After much trial and error, I came up with this interesting solution:
<head>
...various JS and CSS imports...
<script type="text/javascript">
document.write( '<style>#footer{visibility:hidden}#media(min-height:' + ($( window ).height() - 10) + 'px){#footer{visibility:visible}}</style>' );
</script>
</head>
Essentially: use JavaScript to determine the window height of the device, then dynamically create a CSS media query to hide the footer when the height of the window shrinks by 10 pixels. Because opening the keyboard resizes the browser display, this never fails on iOS. Because it's using the CSS engine rather than JavaScript, it's much faster and smoother too!
Note: I found using 'visibility:hidden' less glitchy than 'display:none' or 'position:static', but your mileage may vary.

Works for me
if (navigator.userAgent.match(/iPhone|iPad|iPod/i)) {
$(document).on('focus', 'input, textarea', function() {
$('header').css({'position':'static'});
});
$(document).on('blur', 'input, textarea', function() {
$('header').css({'position':'fixed'});
});
}

In our case this would fix itself as soon as user scrolls. So this is the fix we've been using to simulate a scroll on blur on any input or textarea:
$(document).on('blur', 'input, textarea', function () {
setTimeout(function () {
window.scrollTo(document.body.scrollLeft, document.body.scrollTop);
}, 0);
});

My answer is that it can't be done.
I see 25 answers but none work in my case. That's why Yahoo and other pages hide the fixed header when the keyboard is on. And Bing turns the whole page non-scrollable (overflow-y: hidden).
The cases discussed above are different, some have issues when scrolling, some on focus or blur. Some have fixed footer, or header. I can't test now each combination, but you might end up realizing that it can't be done in your case.

Found this solution on Github.
https://github.com/Simbul/baker/issues/504#issuecomment-12821392
Make sure you have scrollable content.
// put in your .js file
$(window).load(function(){
window.scrollTo(0, 1);
});
// min-height set for scrollable content
<div id="wrap" style="min-height: 480px">
// website goes here
</div>
The address bar folds up as an added bonus.

In case anyone wanted to try this. I got the following working for me on a fixed footer with an inputfield in it.
<script>
$('document').ready(
function() {
if (navigator.userAgent.match(/Android/i) || navigator.userAgent.match(/webOS/i) || navigator.userAgent.match(/iPhone/i) || navigator.userAgent.match(/iPad/i)
|| navigator.userAgent.match(/iPod/i) || navigator.userAgent.match(/BlackBerry/i) || navigator.userAgent.match(/Windows Phone/i)) {
var windowHeight = $(window).height();
var documentHeight = $(document).height();
$('#notes').live('focus', function() {
if (documentHeight > windowHeight) {
$('#controlsContainer').css({
position : 'absolute'
});
$("html, body").animate({
scrollTop : $(document).height()
}, 1);
}
});
$('#notes').live('blur', function() {
$('#controlsContainer').css({
position : 'fixed'
});
$("html, body").animate({
scrollTop : 0
}, 1);
});
}
});
</script>

I have the same issue. But I realized that the fixed position is just delayed and not broken (at least for me). Wait 5-10 seconds and see if the div adjusts back to the bottom of the screen. I believe it's not an error but a delayed response when the keyboard is open.

I tried all the approaches from this thread, but if they didn't help, they did even worse.
In the end, I decided force device to loose focus:
$(<selector to your input field>).focus(function(){
var $this = $(this);
if (<user agent target check>) {
function removeFocus () {
$(<selector to some different interactive element>).focus();
$(window).off('resize', removeFocus);
}
$(window).on('resize', removeFocus);
}
});
and it worked like a charm and fixed my sticky login-form.
Please NOTE:
The JS code above is only to present my idea, to execute this snippet please replace values in angular braces (<>) with appropriate values for your situation.
This code is designed to work with jQuery v1.10.2

This is still a large bug for for any HTML pages with taller Bootstrap Modals in iOS 8.3. None of the proposed solutions above worked and after zooming in on any field below the fold of a tall modal, Mobile Safari and/or WkWebView would move the fixed elements to where the HTML body's scroll was situated, leaving them misaligned with where they actually where laid out.
To workaround the bug, add an event listener to any of your modal inputs like:
$(select.modal).blur(function(){
$('body').scrollTop(0);
});
I'm guessing this works because forcing the HTML body's scroll height re-aligns the actual view with where the iOS 8 WebView expects the fixed modal div's contents to be.

If anybody was looking for a completely different route (like you are not even looking to pin this "footer" div as you scroll but you just want the div to stay at the bottom of the page), you can just set the footer position as relative.
That means that even if the virtual keyboard comes up on your mobile browser, your footer will just stay anchored to the bottom of the page, not trying to react to virtual keyboard show or close.
Obviously it looks better on Safari if position is fixed and the footer follows the page as you scroll up or down but due to this weird bug on Chrome, we ended up switching over to just making the footer relative.

None of the scrolling solutions seemed to work for me. Instead, what worked is to set the position of the body to fixed while the user is editing text and then restore it to static when the user is done. This keeps safari from scrolling your content on you. You can do this either on focus/blur of the element(s) (shown below for a single element but could be for all input, textareas), or if a user is doing something to begin editing like opening a modal, you can do it on that action (e.g. modal open/close).
$("#myInput").on("focus", function () {
$("body").css("position", "fixed");
});
$("#myInput").on("blur", function () {
$("body").css("position", "static");
});

iOS9 - same problem.
TLDR - source of the problem. For solution, scroll to bottom
I had a form in a position:fixed iframe with id='subscribe-popup-frame'
As per the original question, on input focus the iframe would go to the top of the document as opposed to the top of the screen.
The same problem did not occur in safari dev mode with user agent set to an idevice. So it seems the problem is caused by iOS virtual keyboard when it pops up.
I got some visibility into what was happening by console logging the iframe's position (e.g. $('#subscribe-popup-frame', window.parent.document).position() ) and from there I could see iOS seemed to be setting the position of the element to {top: -x, left: 0} when the virtual keyboard popped up (i.e. focussed on the input element).
So my solution was to take that pesky -x, reverse the sign and then use jQuery to add that top position back to the iframe. If there is a better solution I would love to hear it but after trying a dozen different approaches it was the only one that worked for me.
Drawback: I needed to set a timeout of 500ms (maybe less would work but I wanted to be safe) to make sure I captured the final x value after iOS had done its mischief with the position of the element. As a result, the experience is very jerky . . . but at least it works
Solution
var mobileInputReposition = function(){
//if statement is optional, I wanted to restrict this script to mobile devices where the problem arose
if(screen.width < 769){
setTimeout(function(){
var parentFrame = $('#subscribe-popup-frame',window.parent.document);
var parentFramePosFull = parentFrame.position();
var parentFramePosFlip = parentFramePosFull['top'] * -1;
parentFrame.css({'position' : 'fixed', 'top' : parentFramePosFlip + 'px'});
},500);
}
}
Then just call mobileInputReposition in something like $('your-input-field).focus(function(){}) and $('your-input-field).blur(function(){})