1100611 - json.stringify for [object DOMRect] evaluates to "{}"

Status: RESOLVED INVALID

(Core :: DOM: Core & HTML, defect)

Description

[ganna.khabibullina]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/38.0.2125.111 Safari/537.36

Steps to reproduce:

Open Firefox browser
Go to any page. Example: https://www.mozilla.org
Activate the Web Dev Tool by pressing F12
Run the following command in console: 
JSON.stringify(document.querySelector('body').getBoundingClientRect())


Actual results:

"{}"


Expected results:

"{"height":3462,"width":1263,"left":0,"bottom":3462,"right":1263,"top":0}"

Comment 1

[:Gijs (he/him)]

IE11 does the same as us, Chrome doesn't. Chrome breaks if you pass it a DOM node, we output "{}" just like in this case.

Not sure what the spec says about things like this. Reading http://www.ecma-international.org/ecma-262/5.1/#sec-15.12.3 it seems that this depends on whether the properties of DOMRect are enumerable or not. This is backed up by the behaviour of Object.keys:

Object.keys(document.body.getBoundingClientRect())

gives an empty array on Gecko, and a list of the properties on Blink.

Comment 2

[Boris Zbarsky [:bzbarsky]]

> Not sure what the spec says about things like this.

The spec says our behavior is correct.

> it seems that this depends on whether the properties of DOMRect are enumerable or not.

No, it depends on whether they're own properties or not.  And per spec they're not; Blink gets this wrong.

It may be worth adding a jsonifier to DOMRect, though.  That's a spec issue for http://dev.w3.org/fxtf/geometry/

Comment 3

[Boris Zbarsky [:bzbarsky]]

> No, it depends on whether they're own properties or not. 

Well, more precisely it depends on both:

    Let K be an internal List of Strings consisting of the names of all the own properties
    of value whose [[Enumerable]] attribute is true. The ordering of the Strings should be
    the same as that used by the Object.keys standard built-in function.

(in step 6a of the abstract operation JO).

Comment 4

[Boris Zbarsky [:bzbarsky]]

http://lists.w3.org/Archives/Public/public-fx/2014OctDec/0075.html sent to the standards list responsible for DOMRect.

Comment 5

[Boris Zbarsky [:bzbarsky]]

The spec discussion is trying to figure out what the exact use cases are here, so we can figure out the right serializer behavior.  So what _was_ the use case for JSON-stringifying the DOMRect?