W3C home > Mailing lists > Public > www-style@w3.org > January 2012

Re: [css3-transitions] Transitions from display:none

From: Lea Verou <leaverou@gmail.com>
Date: Tue, 17 Jan 2012 01:52:36 +0200
Message-ID: <4F14B844.6090606@gmail.com>
To: Sylvain Galineau <sylvaing@microsoft.com>
CC: Øyvind Stenhaug <oyvinds@opera.com>, "www-style@w3.org" <www-style@w3.org>
On 17/1/12 01:47, Lea Verou wrote:
> On 17/1/12 01:12, Sylvain Galineau wrote:
>> [Lea Verou:]
>>> On 13/12/11 18:43, Øyvind Stenhaug wrote:
>>>> Hi,
>>>> I see it was resolved at TPAC that "CSS animations do not start or
>>>> continue running on elements that are display:none or inside
>>>> display:none elements"
>>>> (<http://lists.w3.org/Archives/Public/www-style/2011Nov/0709.html>).
>>>> However, transitions need to be considered too, and I couldn't find a
>>>> similar issue reported on the list. And even if the transitions spec
>>>> were to get some similar wording, it would be necessary to define in
>>>> what order simultaneous style changes are considered to happen for
>>>> this purpose. This also seems somewhat connected with the plans to
>>>> make all properties interpolable.
>>>> For instance, one might expect margin-left to transition in this case
>>>> (and maybe especially so if 'display' were to be interpolated
>>>> similarly to 'visibility'):
>>>> #test { transition-duration: 0.5s; }
>>>> #test.before { display: none; margin-left: 100px; } #test.after {
>>>> display: block; margin-left: 0px; }
>>>> However, we have already seen a case relying on the opposite, and thus
>>>> looking buggy in Opera.
>>> Why not handle all non-interpolable values like visibility and 
>>> interpolate
>>> them through a discrete step?
>> What are 'all non-interpolable values like visibility'? We should be 
>> specific
>> as to which cases we want to talk about as there are far more that 
>> just can't
>> really be reasonably defined.
> I was referring to every value for which interpolation isn't 
> explicitly defined. Roughly anything that's not a number, integer, 
> percentage, length, angle, time, color, image (in L4) or a functional 
> notation with parameters of these types. Basically, as a fallback kind 
> of interpolation definition, when there's nothing more specific. I 
> never got what was so special about visibility and it was the only 
> property that got that privileged treatment.
Also, that would have the **amazing** side effect of being able to track 
style changes through the transitionEnd event, as in that case it could 
fire for every single property when transition-property: all;. A bit 
hacky way to achieve that, sure, but it's better than nothing.

Lea Verou (http://lea.verou.me | @LeaVerou)
Received on Monday, 16 January 2012 23:53:08 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:09 UTC