This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The sequence normalization process described in section 2 of the serialization spec claims to be the same as the process for constructing document nodes in XSLT or XQuery. However, it isn't the same. XSLT, and XQuery after making the change agreed in bug #3637, both replace document nodes by their contents before deciding whether two atomic values are adjacent and therefore need to be space-separated. This affects the outcome of XQuery test case K2-ancestor-or-selfAxis-5, referenced in bug #4624.
This bug was reviewed in meeting 335 and I was actioned (action A-335-13) to look at the specs again. I have done so and I have concluded that they are indeed consistent in this area. I am therefore marking the bug as invalid and closed.