Saxon

Legacy ID: #4000376 Legacy Poster: Michael Kay (mhkay)

It's not clear to me why you would expect this to work. It's failing for a fairly random reason and could be made cleaner, but I think it should fail. I think it should probably fail under the rule "It is a static error if the expanded QName and arity of a function declared in an imported module are respectively equal to the expanded QName and arity of a function declared in the importing module or in another imported module (even if the declarations are consistent) [err:XQST0034]. ") As far as I can see, cases that ought to work do, for example: main.xq import module namespace p = "cyclic" at "cyclic.xq"; p:foo() cyclic.xq module namespace p = "cyclic"; import module "cyclic" at "cyclic1.xq"; declare function p:foo() {1 + p:bar()}; declare function p:baz() {17}; cyclic1.xq module namespace p = "cyclic"; import module "cyclic" at "cyclic.xq"; declare function p:bar() {1 + p:baz()}; Incidentally, the XQuery WG is considering making some last-minute changes in this area because the formal semantics covering module imports is broken. I guess under a strict reading of the rules it's legal for a module to import itself if it contains no variables or functions, but that case is so pointless I think I will still let it fail under the rule that the interpretation of the location URI is implementation-defined.