- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Wed, 7 Dec 2011 09:54:19 -0800
- To: Simon Sapin <simon.sapin@kozea.fr>
- Cc: www-style@w3.org
On Wed, Dec 7, 2011 at 9:40 AM, Simon Sapin <simon.sapin@kozea.fr> wrote: > Le 07/12/2011 17:54, Tab Atkins Jr. a écrit : >> Indeed, you're right. Fixed now - I explain precisely what it means >> for a list item to "increment the list-item counter". > > > """ > increment the list item counter ‘list-item’: if the element doesn't > otherwise have a ‘counter-increment’ declaration, it must act as if > ‘counter-increment: list-item’ were specified. (This does not affect the > specified or computed values of the counter properties.) > """ > > I’m not sure about other implementation, but for the "not otherwise having a > declaration" test in WeasyPrint I’d have to record more information from the > cascade than I currently do. > > `counter-increment: none`, `counter-increment: initial` and no declaration > all result in having a computed value of 'none', which is the only result > currently kept after the cascade. > > I think that this kind of thing in other properties ends up being resolved > with a 'auto' properties. Here 'auto' would be the same as 'none' except > when 'display' is 'list-item' Unfortunately, the counter-* properties come from 2.1, and have had their definitions and their implementations for some time. I would thus prefer not to change the initial value. However, it makes perfectly good sense, from an implementation perspective, to have a hidden 'auto' value in the UA stylesheet that serializes as 'none' when you query for the computed value. You can't query for the specified value of a UA stylesheet, so no one would be able to tell this from the specified behavior. We only care about the black-box author-detectable behavior. ~TJ
Received on Wednesday, 7 December 2011 17:55:10 UTC