11765 2011-01-14 19:41:39 +0000 Definitions of \i and \c in regular expressions 2012-01-20 17:48:31 +0000 1 1 1 Unclassified XML Schema Datatypes: XSD Part 2 1.0 only PC Windows NT RESOLVED DUPLICATE 15646 P2 normal --- 15646 11421 1 cmsmcq David_E3 cmsmcq mike www-xml-schema-comments oldest_to_newest 44358 0 cmsmcq 2011-01-14 19:41:39 +0000 +++ This bug was initially created as a clone of Bug #11421, which raised the issue against 1.1. This bug is to enable us to track the issue w.r.t. 1.0 and a possible 1.0 3E +++ The definition of \c is straightforward enough: the set of name characters, those ·matched· by NameChar where NameChar is a link to a production in the XML 1.1 specification. The definition of \i by contrast is rather strange: the set of initial name characters, those ·matched· by NameStartChar in [XML] or by Letter | '_' | ':' in [XML 1.0] How are we to read this? I don't think "or" here means "the union of these two sets of characters"; I think it means use one definition if you're using XML 1.0, a different definition if you are using XML 1.1. But why doesn't it use NameStartChar in both cases? What seems to have happened is that in XML 1.0 ed 4 and earlier, names were defined to start with (Letter | '_' | ':'), but in XML 1.0 ed 5, they are defined to start with NameStartChar (which is a larger set of characters). So we have chosen a definition that fixes \i to the pre-5th-edition of XML names, while moving \c forward to the definition used in 1.0ed5 and 1.1. This can't be right. I would suggest aligning both character classes with the definitions of XML names as they appear in XML 1.0 ed 5 and XML 1.1. 62945 1 cmsmcq 2012-01-20 17:48:31 +0000 *** This bug has been marked as a duplicate of bug 15646 ***