Re: CSS Viewport proposal

On Mon, 09 Aug 2010 17:50:43 +0200, Kenneth Christiansen  
<kenneth.christiansen@openbossa.org> wrote:

> Hi again,
>
> You seem to be missing the fit-to-contents part. Try opening
> www.uol.com.br on the iPhone. As it has no viewport tag, 980 is
> assumed as the layout width, but as the contents is 990 in width, the
> iPhone actually modifies the scale value to fit to contents.

OK. That means it needs to be recalculated on reflow - the scale value,  
that is. I don't have an iPhone where I am now, but my initial guess would  
be that it will not affect the calculations of width and height in the  
procedure in the spec. I would guess that it will not change if user  
interaction has already changed the scale value. I need to do further  
testing.

Can you say something about how it works in your implementation?

> As I said in private email, I have some tests that I would like to
> contribute, nothing like your 120 tests, but I at least have two cases
> where the iPhone shown inconsistent behavior which my implementation
> doesn't.

Could you provide those two cases?

> I'm also happy contributing our current code so that you can compare
> the algorithm.

Good.

-- 
Rune Lillesveen
Senior Core Developer / Architect
Opera Software ASA