[css3-transitions] Spec should mention that simultaneity of ?property changes is undefined

I thought I sent comments about this issue before, but I can't find
them.

The transitions spec depends on the relative timing of CSS property
changes in various ways.  For example, if a script changes the
'color' property, and then, later in the same script, changes the
'transition-delay' property from zero to nonzero, whether the color
property transitions or not.  This is because implementations
coalesce (and flush) these changes in different ways.

I think the spec should avoid mandating a specific coalescing
algorithm, but should instead note that some behaviors are
explicitly undefined: in particular, if a change to the value of one
of the transition-* properties happens a small amount of time after
a change to a property whose transition would be different depending
on the first change, it is undefined which transition actually
occurs.

-David

-- 
L. David Baron                                 http://dbaron.org/
Mozilla Corporation                       http://www.mozilla.com/