- From: fantasai <fantasai.lists@inkedblade.net>
- Date: Wed, 06 May 2015 11:46:36 -0700
- To: www-style@w3.org
On 05/05/2015 11:47 PM, Daniel Holbert wrote: > On 05/05/2015 05:31 PM, fantasai wrote: >>> All things being equal, I'd suggest erring on the side of simplicity >>> (avoiding inter-property dependencies) -- so, I'd lean towards your >>> second option, i.e. having stretch compute to itself, and simply >>> interpret the value differently for flex containers. >> >> We currently have the 'auto' values in css-align compute differently >> depending on the layout mode... that would be an argument for changing >> that behavior, too? > > Sort of -- though, as you noted later, there's a semi-compelling reason > that 'auto' needs to be magical & compute to different things: to > provide different sane defaults, w/ backwards-compatibility. Whereas, > there's no strong reason that 'stretch' needs this computed-value-time > magic. We resolved to have the alignment properties' auto (and 'stretch') compute to their resulting behavior based on the layout mode. This puts a dependency from the alignment properties to 'position' on the element itself and 'display' on its parent'. We can revisit later if there's problems. > (I suppose if we take this complexity for 'auto', then it's not > significantly worse to take it for 'stretch' as well, though.) Exactly. :) ~fantasai
Received on Wednesday, 6 May 2015 18:47:11 UTC