1658 2005-07-16 04:30:55 +0000 [FS] editorial: 4.7.2 Other Direct Constructors 2007-02-25 23:15:30 +0000 1 1 1 Unclassified XPath / XQuery / XSLT Formal Semantics 1.0 Last Call drafts All All CLOSED FIXED P2 minor --- 1 jmdyck simeon public-qt-comments oldest_to_newest 4742 0 jmdyck 2005-07-16 04:30:55 +0000 4.7.2 Other Direct Constructors (PI and Comment) "its character content is converted to a string as in attribute content." Then it's odd that there isn't an invocation of [[]]_AttributeContent. But really, you don't want it converted as in attribute content, because that would (e.g.) map '&' to '&' and '{{' to '{', neither of which are correct for PI or Comment content. Norm / rule 1 Char* ?>" ]]_Expr Delete stray doublequote. Norm / rule (1|2) "processing-instruction NCName { Char* }" "comment { Char* }" In general, Char* does not constitute a valid Expr, and even when it does, it doesn't denote what you want it to. Putting Char* within quotes would be a good start, but then you have to worry about quotes in the Char* that then need to be escaped. Maybe introduce a []-form to accomplish the conversion. DErr "for direct processing-instruction constructors." Append "or comment constructors." 9277 1 simeon 2006-04-17 03:24:34 +0000 Fixed as suggested. Added an auxiliary normalization rule to convert Char* into an appropriate string. - Jerome 12776 2 jmdyck 2006-11-02 06:21:55 +0000 re the auxiliary normalization rule... [Char*]_Characters == fn:codepoints-to-string(Char*) As in Bug 1629 (4.7.1) and Bug 3896 (4.7.1.1), this doesn't work. Instead, just define []_Characters informally, e.g. "[]_Characters takes the character content of a PI or comment constructor and yields a corresponding StringLiteral." 13595 3 simeon 2007-01-16 17:56:31 +0000 Fixed as suggested. - Jerome