W3C home > Mailing lists > Public > www-style@w3.org > April 2009

Re: [CSS3] Flexible Flow Module, proposal.

From: David Hyatt <hyatt@apple.com>
Date: Sun, 12 Apr 2009 17:54:23 -0500
Cc: "Tab Atkins Jr." <jackalmage@gmail.com>, Andrew Fedoniouk <news@terrainformatica.com>, www-style <www-style@w3.org>
Message-id: <357C9480-C160-48BC-80EC-5A8A6E9C9D81@apple.com>
To: robert@ocallahan.org
On Apr 12, 2009, at 5:45 PM, Robert O'Callahan wrote:

> On Sun, Apr 12, 2009 at 5:14 AM, Tab Atkins Jr.  
> <jackalmage@gmail.com> wrote:
> Robert, what's your opinion on this vs the existing XUL flexboxes in  
> FF?
> I mostly agree with Hyatt's latest message.
> Flex units are cool and give some extra power, and avoid the need  
> for alignment and packing properties. But they're not yet a  
> replacement for what XUL flexboxes can do. I think the biggest  
> limitation is not being able to set width to "intrinsic width plus  
> flex". The limitations can probably be fixed, although the fixes  
> might add complexity or ugliness.
> XUL flexbox ordinals are very rarely used, as far as I can tell, and  
> their omission is probably OK. We don't actually support flex-groups  
> in Gecko today so their omission is OK for us.
> I want to be able to map XUL flexboxes naturally into any 'flex'  
> proposal the group comes up with, partly for the selfish reason of  
> not having to maintain two flex layout systems, but also because XUL- 
> style flexboxes are quite well known among authors (e.g. Firefox  
> extension developers and Adobe "Flex" developers). If Andrew's  
> proposal reaches that point, it could be a good option.

Same.  XUL flexboxes are used in WebKit now as the foundation of some  
of our form controls (HTML buttons for example), as well as used  
heavily in Safari RSS.  It's very important to Apple that we not have  
to maintain two flex systems in WebKit going forward, so any new  
standard needs to be something we can map our existing uses onto.

> We need to do major surgery on our flexbox code anyway (see http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/layout/xul/base/src/nsSprocketLayout.cpp&rev=1.69&mark=695-696#695 
>  ...) Implementing a new flex module based on a solid WG draft and  
> then mapping XUL onto it could be attractive.

I wrote that comment after spending days cleaning up the original code  
(that I cannot be blamed for!!!!).  I'm happy to see that it is still  
in the source tree. ;)


Received on Sunday, 12 April 2009 22:55:11 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:07:35 UTC