- From: Benjamin Poulain <benjamin@webkit.org>
- Date: Tue, 14 Apr 2015 13:01:06 -0700
- To: Simon Pieters <simonp@opera.com>, www-style@w3.org
On 4/13/15 2:50 PM, Simon Pieters wrote:
> On Mon, 13 Apr 2015 23:26:41 +0200, Boris Zbarsky <bzbarsky@mit.edu> wrote:
>
>> On 4/13/15 5:13 PM, Simon Pieters wrote:
>>> Yeah, it should be null to match scrollTop's behavior. This was an
>>> oversight.
>>>
>>> https://github.com/w3c/csswg-drafts/commit/ad098d66cfc02a4952e8af2eeea742c520d10c2b
>>>
>>
>> So what _is_ scrollTop's behavior, exactly? Consider this testcase
>> (quirks mode):
>>
>> <html>
>> <body style="display: none">
>> </body>
>> <script>
>> var body = document.createElement("body");
>> document.documentElement.appendChild(body);
>> var div = document.createElement("div");
>> div.style.height = "5000px";
>> body.appendChild(div);
>> </script>
>> </html>
>>
>>
>> scroll down and check document.body.scrollTop.
>>
>> I believe the current spec says that this should return window.scrollY
>> (because the element in question is the body element and does not have
>> a scrolling box), but in Firefox it does not (because it doesn't have
>> a box at all).
>
> Do you think we should change the spec? I'd be OK to match Firefox for
> this case.
Why? The spec seems better than Firefox's behavior in this case. It also
better matches the other engines.
>> In IE11, .scrollTop on _both_ of the <body> elements in that testcase
>> returns the scroll position. So does
>> document.documentElement.scrollTop...
>
> This seems a bit weird. :-)
>
Received on Tuesday, 14 April 2015 20:05:53 UTC