attr('href') returns object

[[email protected]…]

jQuery 1.7.1

aTag.attr('href') return the URL of aTag
aTag.prop('href') return the full URL of aTag

jQuery 1.7.2

aTag.attr('href') return aTag
aTag.prop('href') return the full URL of aTag

Example

​http://jsfiddle.net/VZLsH/1/

[sindresorhus]

Can you please state the browser and version you're experiencing this in?

I've tried the testcase in IE6-9, Firefox 36, 11, Chrome, Opera 11.62 with jQuery 1.7.1 and 1.7.2 and I couldn't reproduce the described issue.

Duplicate of #11516, but keeping this, since it's the only one with testcase.

[sindresorhus]

#11516 is a duplicate of this ticket.

[Rick Waldron]

In support of @sindresorhus

Chrome...

IE8...

I can post more if nec

[[email protected]…]

It doesn't work with Firefox 11 on Windows XP.

[Rick Waldron]

Works there as well:

[sindresorhus]

#11529 is a duplicate of this ticket.

[mikesherov]

Just tossing this out there randomly: might compatibility view have something to do with this? Whenever I encounter an IE that's sometimes reproducible (maybe), I make sure I check with compatibility view on and compatibility view off.

Maybe it's just noise I'm adding, but it's worth a shot

[Rick Waldron]

@[email protected]…

When you created the jsfiddle - did the bug reproduce?

[Rick Waldron]

#11512 is a duplicate of this ticket.

[Rick Waldron]

#11519 is a duplicate of this ticket.

[ajpiano]

I have a feeling this might be TinyMCE. Given both the nature of the several reports we've gotten, and also because of the fact that TinyMCE is a widely used WYSIWYG editor with a jQuery adapter that duckpunches attr. Take a look at line 215 of this gist ​https://gist.github.com/92a2c3631be1d35bb5e2, which is the beautified version of the minified jQuery plugin they ship in their zip file.

[ajpiano]

For anyone reading this's reference, research earlier today by mikesherov yielded this fiddle: ​http://jsfiddle.net/7u7cN/2/ which demonstrated that plugins that are redefining .attr and then calling it later on with .call would created unexpected results for people using attr in chains between 1.7.1 and 1.7.2. I believe rwaldron and davemethvin have come up with a path forward here that I'll leave it to them to document

[ajpiano]

#11531 is a duplicate of this ticket.

[gnarf]

Left a comment on the gist for anyone interested in the actual fix in TinyMCE's punch.

[[email protected]…]

Checkout "Version 3.5b3" @ ​http://www.tinymce.com/develop/changelog/index.php?type=tinymce "Fixed bug where the jQuery plugin would break the attr method of jQuery 1.7.2.".

[Rick Waldron]

Awesome!

[anonymous]

jQuery.fn.attr looks at arguments.length regardless of whether the last argument was undefined, so if you make a function that passes all of it's arguments to .attr, it would fail.

[dmethvin]

Yes, so if you want .attr() to act as a getter, pass it *one* string argument representing the attribute to be retrieved. That is the way it is documented. It was never documented to act as a getter when undefined was passed as a second argument. Behavior for that case is not specified, and may change from version to version as it did here.

[sindresorhus]

#11562 is a duplicate of this ticket.

[Rick Waldron]

#11572 is a duplicate of this ticket.

[dmethvin]

Turns out TinyMCE broke this before in #7284, and as before they have fixed it again in Version 3.5b3. Lurkers and Google searchers, please update your TinyMCE and hope they don't break it a third time. ​http://www.tinymce.com/download/download.php

[dmethvin]

#11615 is a duplicate of this ticket.

[mikesherov]

#11665 is a duplicate of this ticket.

[Rick Waldron]

#11683 is a duplicate of this ticket.