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

RE: [css3-flexbox] Changing abspos placeholders to atomic inlines

From: Alex Mogilevsky <alexmog@microsoft.com>
Date: Thu, 31 May 2012 21:34:17 +0000
To: fantasai <fantasai.lists@inkedblade.net>, "www-style@w3.org" <www-style@w3.org>
Message-ID: <2C86A15F63CD734EB1D846A0BA4E0FC823C9FC37@CH1PRD0310MB381.namprd03.prod.outlook.com>
± From: fantasai [mailto:fantasai.lists@inkedblade.net] 
± Sent: Wednesday, May 30, 2012 7:18 PM
± 
± On 05/30/2012 06:06 PM, Tab Atkins Jr. wrote:
± > On Wed, May 30, 2012 at 11:54 AM, Alex Mogilevsky<alexmog@microsoft.com>  wrote:
± >>
± >> It looks like at least two implementations don't like the change to atomic inline. 
± Just in favor of stability that should be enough to revert the change and get to what 
± it has been for last 7 years or so.
± >
± > Hm, I didn't realize the old flexbox did have that behavior in practice.
± >
± > The reason we made the change is because someone (Anton?) pointed out 
± > that inlines don't respect 'height', so the placeholder's actual 
± > dimensions are 0px wide and 'line-height' tall.
± 
± The other reason was that the behavior is more predictable for authors by making them 
± atomic inlines. One of the weird things about placeholders is that
± 
± <div>
±    <div/>
±    <div/>
± </div>
± 
± makes 2 flexboxes (or table cells),
± 
± <div>
±    <div/>
±    <abspos/>
±    <div/>
± </div>
± 
± Adding an abspos element makes 3 flexboxes.
± 
± <div>
±    <div/>
±    <abspos/>
±    <abspos/>
±    <div/>
± </div>
± 
± But adding another abspos element makes still only 3 flexboxes.
± 
± 
± <div>
±    <div/>
±    <abspos/>
±    <abspos/>
±    text
±    <div/>
± </div>
± 
± And here it doesn't add any! For something that's invisible and generally behaves like 
± it doesn't exist, placeholders are weird. Making them atomic makes their behavior a 
± little bit more predictable: they always add an invisible flex item.

I am firmly convinced that the only people who care are the people writing tests for test suite. 

Flexboxes shouldn't have either absolute positioned children or plain text. We can't ignore them because that confuses authors for real (we tried). Otherwise it doesn't really matter what happens as long as it is somehow defined and helps people find bugs in their markup.

To me it seems that if this can be close to any of existing anonymous blocks it is goodness...
Received on Thursday, 31 May 2012 21:40:05 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:54 GMT