IE9-10 curCSS => "Interface not supported" in for popups (and probably frames)

[[email protected]…]

The fix for bug #10373 in commit f7ee1f6e59f0b465f5f64bf6ac52108e445efaac broke the getComputedStyle functionality for elements residing in popups or frames accessed from the main document.

The following jsfiddle demonstrates the problem, choose jQuery 1.7.2 and the problem disappears.

​http://jsfiddle.net/VYNj9/

This was also reported by a comment on #10373 ​http://bugs.jquery.com/ticket/10373#comment:7.

[gibson042]

Seems like a reasonable ​one-line change to me, though I wish we had better non-browser testing to be sure.

[anonymous]

I've patched it on the 1.8-stable branch on my fork on GitHub (chrisprice) but I'm not sure how to add tests. It doesn't seem to be reproducible in an iframe.

[dmethvin]

On the general need, since we've talked about requiring a browser-like window object for environments like node.js it does seem rather annoying. However, is there any browser-like simulator for node that could handle $(elem).width() anyway? In the cases where I've used jQuery I've just wanted selection and traversing, perhaps you could add manipulation to that but the rest seems too browser specific to be used.

It sounds like the crash only happens in a window.open window, I'm not sure we want to add a test that requires calling it. Ensuring that document.defaultView works may be enough?

The currently active branches are 1.x-master and master (which is 2.x). The PR should be against one of those, ideally both but if it's easy to cherry pick there should be no problem with just having one PR for now.

[[email protected]…]

PR: ​https://github.com/jquery/jquery/pull/1311

[gibson042]

@dmethvin That's fair, but we probably don't want to crash on .css( property ). As for reproducibility, I can confirm that the error does not manifest in iframes: ​http://jsfiddle.net/VYNj9/2/

As for testing... man, that's tough. If we don't put in a window.open, we're likely to see a regression at some point, but if we do, we're likely to have other problems with it.

[gibson042]

#12235 is the original reason for window.getComputedStyle( elem, null ), but we never questioned the reporter about what ey was doing with CSS on jsdom to run afoul of the bug.

[Chris Price]

Fix #14150: revert to ownerDocument.defaultView.getComputedStyle. Close gh-1311.

Changeset: 3e3680c2cdc524ee9194ef27fdce318cf301144c

[dmethvin]

Bulk update for new milestone

[Oleg Gaidarenko]

CSS: Do not throw on frame elements in FF

IE9-10 throws on elements created in popups (see #14150), FF meanwhile throws on frame elements through "defaultView.getComputedStyle" (see #15098)

Use "defaultView" if in the popup which would fix IE issue, use "window.getComputedStyle" which would fix FF issue.

And everybody wins, except performance, but who cares right?

Fixes #15098 Closes gh-1583

Changeset: e488d985cfb10ab8c684bbc6a9b8ff3eae23bf83