This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
As written, stds with patterns need not be in the lexical space of the relevant type, so e.g. a type restricting an integer with a pattern "a" would validate an "a". See also R-98 and R-221
There seem to be other problems in the "Datatype Valid" constraint in 4.1.4 of the datatypes spec: - It assumes that there could be literals in the lexical space that don't correspond to any value in the value space. But as pointed out earlier (2.4.2) "Constraining the ·value space· consequently constrains the ·lexical space·." - It ignores the fact that the order of member types in union types is significant. Suggestion: something along the following line: 1. The string must be pattern valid if there is a pattern 2. The string must denote a value: 2.1 Primitive type 2.2 Derived atomic type (must be datatype-valid wrt the base/primitive type 2.3 List type (must be a sequence of items, each of which is datatype-valid wrt the item type) 2.4 Union type (must be datatype-valid wrt to at least one of the member types) (ues the first one in the list to determine the "denoted value") 3. The denoted value must be facet valid wrt all non-pattern facets