Re: Acid3 and :first-child

Well if the real issue here is that the specification change means the test
is counting the wrong behavior as true, there seem to be 4 possible courses
of action.

1.  Leave the test as is and just advertise it as testing the specification
as of blah date!

2.  Change the test to require the new behavior.

3.  disable the test altogether and mark everyone as passing.

4.  Permit both the old spec and new spec behavior for some reasonable
extent of time.  Like 2 or three months to get to implement and test the
new spec behavior. and then after that period require the new behavior for
a a passing grade.

I would think the fourth behavior is the correct way to address this.

Of course if the Mozilla interpretation of the new spec is incorrect, then
this is all moot.