W3C home > Mailing lists > Public > www-style@w3.org > January 2015

Re: [css-grid] Updates and Issues

From: Sergio Villar Senin <svillar@igalia.com>
Date: Tue, 13 Jan 2015 15:47:53 +0100
Message-ID: <54B53019.3090707@igalia.com>
To: www-style@w3.org
On 12/01/15 21:49, fantasai wrote:
> Tab and Rossen and Greg and I recently went through a pile of
> CSS Grid issues and tried to fix them all. There were some major
> ones that warrant WG discussion, here's a summary:
> 
> 
> Clamping overlarge grids
>   http://lists.w3.org/Archives/Public/www-style/2015Jan/0093.html
>   A grid could, through explicit grid declarations using repeat(),
>   or through positioning and spanning, become too large to handle.
>   We added some text about clamping grid size and positioning.
> 
>   DISCUSS: Whether this text is acceptable, and whether or not we
>   want to clamp repeats per repeat or just truncate the grid.

I think the new text is OK. Clamping repeats by repeat is the right
thing to do IMO and it was mentioned also in previous versions. Also it
seems that you also removed the 1 million minimum size from previous
versions, didn't you?

> 
> Dropping grid-auto-flow: stack
>   http://lists.w3.org/Archives/Public/www-style/2014Sep/0214.html
>   On reviewing the values of grid-auto-flow, we concluded that the
>   'stack' value is not really very intelligent. Our proposal is
>      * Drop the 'stack' value (which puts everything into the
>        first empty slot) because it's not great.
>      * Microsoft keeps -ms-none as the UA-stylesheet value for
>        backwards compat for their apps. (This value puts everything
>        into slot 1,1.)
>   http://lists.w3.org/Archives/Public/www-style/2014Sep/0214.html
> 
>   DISCUSS: Is everyone happy with the proposal to drop 'stack'?

Yeah, as mentioned some other times, it's difficult to find a real use
case for the 'stack' case, at least using the behavior defined in the spec.

> Repeat-to-fill
>   http://lists.w3.org/Archives/Public/www-style/2014Oct/0108.html
>   A very common grid-based layout (used for e.g. online catalogs,
>   photo albums, article summaries, etc) is to have as many columns
>   as there is space. Right now the grid spec only allows fixed
>   repetitions; we propose to add a repeat-to-fill function, which
>   basically does multi-col arithmetic against the grid container's
>   size.
> 
>   DISCUSS: Adding repeat(auto, <track-list>)?

I'm a bit skeptic about this, mainly because I see many potential issues
with content-sized tracks and flex tracks. The whole track sizing
algorithm is full of stuff to allow tracks to grow to their limits and
even beyond with special cases for definite/indefinite sizes. Having
something like that will likely complicate the algorithm a lot. Apart
from that how do you see it working with flex tracks for example, will
we disallow them to grow to create more tracks (then they won't actually
work as flex tracks)?

Thanks for the good work.

BR
Received on Tuesday, 13 January 2015 14:48:20 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:50 UTC