3211 2006-05-09 07:59:49 +0000 [FS] editorial: B.2 Mapping of Overloaded Internal Functions 2007-11-04 06:00:05 +0000 1 1 1 Unclassified XPath / XQuery / XSLT Formal Semantics 1.0 Candidate Recommendation All All ASSIGNED P2 minor --- 1 jmdyck jmdyck public-qt-comments oldest_to_newest 9623 0 jmdyck 2006-05-09 07:59:49 +0000 B.2 Mapping of Overloaded Internal Functions STA "The following static typing rules apply generically to all the fs: special functions." This should be conveyed by the inference rules too. I.e., add expanded-QName = (FS-URI,String) String in {"plus","minus","times","idiv","div", ...} (directly or indirectly) to each rule. (leftover from Bug 1580) STA / rule 1 The top set of premises is only satisfied if every Typei is a union of the same number (m) of OptAtomicTypes, which is not what you want. You could fix this by making the m's distinct: ... OptAtomicType_m₁,1 ... OptAtomicType_{m_n,1} But then what would you do in the second set of premises? If you want the rule to consider all possible combinations of argument types, you probably need a recursive rule, e.g. Typei = (OptAtomicType1 | ... | OptAtomicTypem) statEnv |- expanded-QName(Type1, ..., OptAtomicType1, ..., Typen) : Type1' ... statEnv |- expanded-QName(Type1, ..., OptAtomicTypem, ..., Typen) : Typem' ----------------------------------------------- statEnv |- expanded-QName(Type1, ..., Typen) : (Type1' |...| Typem') If you concede that the only possibilities are n=1 or n=2, you could split this rule into three simpler rules. STA / rule (2|3|4|5|6|7) / conclusion HTML source: <em><a href="#id-expanded-qnames"><em>expanded-QName</em></a></em> Remove outer <em></em>. STA / rule (4|5|6|8|9) "AtomicType" AtomicType is a SequenceType, not a Formal Type. Instead, you probably mean AtomicTypeName, which *is* a Formal Type. STA / rule 7 The rule needs a premise: Type1 = empty Notation 2 / rule (1|2) "AtomicType" As above, change to AtomicTypeName. 12334 1 jmdyck 2006-10-06 07:31:56 +0000 > "The following static typing rules apply generically to all the fs: > special functions." > This should be conveyed by the inference rules too. I.e., add > expanded-QName = (FS-URI,String) > String in {"plus","minus","times","idiv","div", ...} > (directly or indirectly) to each rule. (leftover from Bug 1580) Or, each rule could have a premise saying expanded-QName is an overloaded internal function where you define that with a rule like: LocalPart in { 'plus', 'minus', 'times', 'idiv', 'div', 'mod', 'eq', 'ne', 'gt', 'lt', 'ge', 'le', 'is-same-node', 'node-before', 'node-after', 'unary-plus, 'unary-minus' } ---------------- (FS-URI,LocalPart) is an overloaded internal function 14202 2 jim.melton 2007-02-26 00:13:27 +0000 The fix for this bug does not appear in the Recommendation of 23 January 2007. It will be considered for a future publication (either an Errata document or some possible future version of the specification).