Add jQuery.support.fixedPosition

[scott.gonzalez]

jQuery already detects support for fixed positioning in the offset code. jQuery.offset.supportsFixedPosition should be moved to jQuery.support.fixedPosition and calculated immediately.

[dmethvin]

Would it make sense to move the others such as doesNotAddBorder as well?

[borisov.gleb@…]

I think that other properties should live in jQuery.offset, just because they're internal and used only in offset.js. I have just forked jquery on github and pushed changed to my master-branch. Also created pull-request for this change.

[rwaldron]

I'm not sure about this particular patch (no offense) but I definitely think this should be reviewed further.

[paul.irish]

Here is the relevant jQuery.offset.supportsFixedPosition code fwiw:  https://github.com/jquery/jquery/blob/eed380/src/offset.js#L125-143

It should certainly execute lazily, but augmenting jQuery.support proper makes sense to me.

[john]

Let's look at this for 1.7.

[john]

Possible patch:  https://github.com/monsterzz/jquery/commit/7ab7f285127fe29107a47fe486e3ea40ff51aec9

[john]

Nominating ticket for 1.7 discussion.

[rwaldron]

+1,

[jaubourg]

+1, There is no reason why offset should keep a private repository of support properties... ajax does its own support tests yet publishes the result on the main support object.

[ajpiano]

+1,

[timmywil]

+1, Makes sense.

[paul.irish]

+1,

though.. iOS false positives on nearly all written fixed pos tests (incl this one);  https://github.com/Modernizr/Modernizr/issues/167

Might be unsolvable and leave it up to docs to explain

[dmethvin]

+1, If some code can benefit from precomputing it, rather than on-demand, fine.

[john]

+1, Seems easy enough.

[scott.gonzalez]

+1

[jzaefferer]

+1

[iMarc]

This false-positive iOS < 5 bug was driving me crazy. The only difference between normal browsers that do support position: fixed and iOS < 5 that I found was that offsetParent is null inside a relative element (code below). Not sure if that is at all related, but I was simply obsessed with locating any difference I could find that might be related to the iOS rendering issue of position fixed..

var test  = document.createElement('div'),
    control = test.cloneNode(false),
    wrap = test.cloneNode(false),
    fake = false,
    root = document.body || (function () {
      fake = true;
      return document.documentElement.appendChild(document.createElement('body'));
    }());

root.appendChild(wrap);
wrap.appendChild(test);
wrap.appendChild(control);

wrap.style.cssText = 'position:relative;';
test.style.cssText = 'position:fixed;';
control.style.cssText = 'position:absolute;';

alert(!!control.offsetParent); // alerts true in all browsers, even iOS
alert(!!test.offsetParent); // alerts false in iOS

if (fake) {
  document.documentElement.removeChild(root);
}

[rwaldron]

The problem I'm initially running into is that support is generally run before before the real body actually exists... I think I can work around that, draft implementation to follow.

[rwaldron]

Working draft:  https://github.com/rwldrn/jquery/tree/6809

Currently the only failing tests are:

offset: body (2, 0, 2)

[john]

Isn't this the case where you can do:

body = document.createElement("body");

to get the body element, even if it doesn't exist yet?

[rwaldron]

Yes - I'm using document.body if it happens to exist early enough or the fake body that the support module has already created as a backup

@  https://github.com/rwldrn/jquery/commit/1f1990431a9abb8e964825e5bf7bd7c1eb28fa1b

[anonymous]

This works:  http://mnobeta.no/2011/09/test-position-fixed-for-iphone/ On iphone. Android needs to be taken care of.

[Rick Waldron]

Landing pull request 477. 1.7 jQuery.offset.supportsFixedPosition. Fixes #6809.

More Details:

•  https://github.com/jquery/jquery/pull/477
• http://bugs.jquery.com/ticket/6809 Changeset: f60213648cefa9b53289ad01a55ead22a15e6ee1