On Thu, Apr 3, 2014 at 12:21 PM, Xidorn Quan <quanxunzhen@gmail.com> wrote: > > On Thu, Apr 3, 2014 at 12:01 PM, Tab Atkins Jr. <jackalmage@gmail.com>wrote: > >> >> On the other hand, in the call today dbaron pointed out that >> animation/transition/list-style all have ambiguity issues that they >> solve by just trying as hard as they can to parse - you're allowed to >> name an animation "backwards", for instance. >> > > I see what dbaron said. However, I don't think it is a good idea. Although > it is easy to implement, it is dangerous for forward compatibility. For > example, if someone defines a keyframes called 'move', and uses it in > shorthand animation property like 'animation: move ease 1s', and we later > defines a new timing-function called 'move' as well, the meaning of the > value will be changed. > > IMO, If we decide to allow trying to parse as hard as possible, name must > be parsed in the highest priority, and then moved to other slot if possible > or necessary. > Things could be even more complex for list-style. As animation is still in WD, though very dangerous, it is possible for us to announce that the first none or identifier-like value will always be parsed into animation-name instead of any keywords. But for list-style, we cannot do this, as it already accepts position keywords to be the first part. That is to say, if any new position keyword is defined, meaning of value of list-style may be affected. - Xidorn
Received on Thursday, 3 April 2014 02:08:44 UTC