Side navigation
#14736 closed bug (fixed)
Opened January 28, 2014 06:29PM UTC
Closed March 20, 2014 06:43PM UTC
window.getDefaultComputedStyle(elem) may return null
| Reported by: | rr.rosas@gmail.com | Owned by: | rr.rosas@gmail.com |
|---|---|---|---|
| Priority: | low | Milestone: | 1.12/2.2 |
| Component: | css | Version: | 1.11.0 |
| Keywords: | Cc: | ||
| Blocked by: | Blocking: |
Description
In Iceweasel 24.2.0 it happens here when I call layout from jQuery layout plugin in an iframe from an invisible tab (from jQuery UI tabs).
I've created this ticket after the PR:
https://github.com/rosenfeld/jquery/commit/ab632f868e620246edbc11564f4bb83b68042ed0
Attachments (0)
Change History (13)
Changed January 28, 2014 06:33PM UTC by comment:1
| owner: | → rr.rosas@gmail.com |
|---|---|
| status: | new → pending |
Changed January 28, 2014 06:39PM UTC by comment:2
| status: | pending → new |
|---|
Sorry, but I don't think I'll have time to work on a project and try to replicate the issue using jQuery core only. It only seems to happen in Firefox. Please believe on me when I say that window.getDefaultComputedStyle may return null (or undefined, don't remember) for this browser.
Also, unfortunately the application I maintain is not open-source so I can't point it to you either (it's available in the web, but require user/password).
Is that really mandatory?
Changed January 28, 2014 06:42PM UTC by comment:3
It returns null actually (not undefined), I just checked.
Changed January 28, 2014 06:45PM UTC by comment:4
| status: | new → pending |
|---|
I understand, we all have work to do. Yes, we need a test case to understand what is happening. We can't just add a bunch of tests for bad return values and then wonder a year later why they are there, with no way to know whether the problem is resolved. We don't need (or want!) to see the whole application or code that involves jQuery UI and third-party plugins, just a test case that shows a valid call to a jQuery core method that causes the problem.
Changed January 28, 2014 06:45PM UTC by comment:5
In Firefox console I could get it to return null this way:
$('iframe:first')[0].contentWindow.getDefaultComputedStyle($('body')[0])
It doesn't matter if the element is on the parent window or in the iframe itself.
Not sure if this is a required condition, but $('iframe:first').is(':visible') returns false.
Changed January 28, 2014 06:48PM UTC by comment:6
You're using the getDefaultComputedStyle of the iframe's window but passing it the <body> element from the parent. I wouldn't expect that to work.
Changed January 28, 2014 06:51PM UTC by comment:7
| status: | pending → new |
|---|
How about this? http://jsfiddle.net/b6Kv5/1/
$(function(){
var iframe = $('iframe:first')[0];
$('body').text(iframe.contentWindow.getDefaultComputedStyle(iframe.contentDocument.body));
});
It still output nulls, as I said before. It doesn't matter if the element belongs to the iframe or master, getDefaultComputedStyle from an iframe doesn't seem to work in Firefox.
Changed January 28, 2014 06:53PM UTC by comment:8
And yes, the iframe being hidden is a required condition for that function to return null.
Changed January 28, 2014 07:05PM UTC by comment:9
| component: | unfiled → css |
|---|---|
| priority: | undecided → low |
| status: | new → open |
Ok, that's enough info for us to look into it.
Changed January 28, 2014 07:20PM UTC by comment:10
Ok, thank you.
Changed January 28, 2014 10:07PM UTC by comment:11
Oops, I posted it too soon. The return value was wrong in the previous examples. Fixed version of workaround:
(function(){
if (window.getDefaultComputedStyle) {
var old = getDefaultComputedStyle;
window.getDefaultComputedStyle = function(elem) {
if (var style = old(elem)) return style.display;
return {display: $.css(elem, "display")};
};
}
})();
Changed February 07, 2014 02:15AM UTC by comment:12
| milestone: | None → 1.12/2.2 |
|---|
Changed March 20, 2014 06:43PM UTC by comment:13
| resolution: | → fixed |
|---|---|
| status: | open → closed |
CSS: window.getDefaultComputedStyle may return null
Fixes #14736
Closes gh-1501
(cherry picked from commit 51910ba8a5aff4cd1a811f1fdb76a6c62a09a666)
Changeset: e547a2775fc79e4b2eb807da9d30719afc570458
The description sounds rather complicated. Can you post a simplified test case that only uses jQuery core? If it's only happening in one browser it could be a browser bug, but we'd need a test case to be sure. We'll also need one anyway to include in our unit tests.