Re: [CSSWG] Resolutions 2007-08-28

On 29.08.2007, at 03:46, Boris Zbarsky wrote:

> Timur Mehrvarz wrote:
>> Let's say, one UA does support aspect-ratio, but does not re- 
>> evaluate MQ's in response to aspect-ratio changes. In such a case,  
>> I may want to help myself, with code, similar to the one shown below.
>
> That's exactly what sites did to work around NS4 not reapplying CSS  
> in some cases.  The result is that some UAs don't actually reload  
> when you do location.reload() from a resize event handler.  The  
> relevant Gecko code comment, for example, is:
>
>      // location.reload() was called on a window that is handling a
>      // resize event. Sites do this since Netscape 4.x needed it, but
>      // we don't, and it's a horrible experience for nothing. In stead
>      // of reloading the page, just clear style data and reflow the
>      // page since some sites may use this trick to work around gecko
>      // reflow bugs, and this should have the same effect.
>
> In other words, UAs already have to do with this situation....
>

Interesting. But don't get me wrong. I would really prefer to not  
have to trigger the reload for this. I just wanted to point out, what  
the likely consequences of a should requirement are, in regard to  
aspect-ratio. (As mentioned before, a full reload() is also bad,  
because of it's heavyweight nature and because of the possible loss  
of state.)

Only the agent can provide a true solution to this. Agents (those,  
that run in a resizable window) /could/ load all listed aspect-ratio  
criterias in a HashMap on document load - and do the proper (rather  
lightweight) reflow, whenever necessary. While resize - or after.  
That would be doable. Or not?

Timur

Received on Wednesday, 29 August 2007 02:29:32 UTC